P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML

GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

• •
Human Walking

i
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

ii
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:7

• •
Human Walking
• THIRD EDITION

EDITED BY
•
Jessica Rose, PhD
Assistant Professor, Department of Orthopaedic Surgery
Stanford University School of Medicine
Director, Motion & Gait Analysis Laboratory
Lucile Packard Children’s Hospital
Palo Alto, California

James G. Gamble, MD, PhD

Professor, Department of Orthopaedic Surgery
Stanford University School of Medicine
Medical Director, Motion & Gait Analysis Laboratory
Lucile Packard Children’s Hospital
Palo Alto, California

iii
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

Acquisitions Editor: Robert Hurley

Managing Editor: Michelle LaPlante
Production Manager: Bridgett Dougherty
Senior Manufacturing Manager: Benjamin Rivera
Marketing Director: Sharon Zinner
Design Coordinator: Holly McLaughlin
Production Services: Techbooks
Printer: Edward Brothers

C 2006 by LIPPINCOTT WILLIAMS & WILKINS
530 Walnut Street
Philadelphia, PA 19106 USA
LWW.com

All rights reserved. This book is protected by copyright. No part of this book may be reproduced in
any form or by any means, including photocopying, or utilized by any information storage and
retrieval system without written permission from the copyright owner, except for brief quotations
embodied in critical articles and reviews. Materials appearing in this book prepared by individuals
as part of their official duties as U.S. government employees are not covered by the
above-mentioned copyright.

Printed in the USA

Library of Congress Cataloging-in-Publication Data

Human walking / [edited by] Jessica Rose, James G. Gamble.— 3rd ed.
p. ; cm.
Includes bibliographical references and index.
ISBN 0-7817-5954-4
1. Walking. I. Rose, Jessica. II. Gamble, James Gibson.
[DNLM: 1. Locomotion—physiology. 2. Walking—physiology. 3. Biomechanics.
4. Gait—physiology. WE 103 H918855 2006]
QP310.W3H85 2006
612.7 6—dc22
2005022885

Care has been taken to confirm the accuracy of the information presented and to describe
generally accepted practices. However, the authors, editors, and publisher are not responsible for
errors or omissions or for any consequences from application of the information in this book and
make no warranty, expressed or implied, with respect to the currency, completeness, or accuracy of
the contents of the publication. Application of this information in a particular situation remains
the professional responsibility of the practitioner.
The authors, editors, and publisher have exerted every effort to ensure that drug selection and
dosage set forth in this text are in accordance with current recommendations and practice at the
time of publication. However, in view of ongoing research, changes in government regulations, and
the constant flow of information relating to drug therapy and drug reactions, the reader is urged to
check the package insert for each drug for any change in indications and dosage and for added
warnings and precautions. This is particularly important when the recommended agent is a new or
infrequently employed drug.
Some drugs and medical devices presented in this publication have Food and Drug
Administration (FDA) clearance for limited use in restricted research settings. It is the
responsibility of the health care provider to ascertain the FDA status of each drug or device
planned for use in their clinical practice.
To purchase additional copies of this book, call our customer service department at (800) 638-
3030 or fax orders to (301) 223-2320. International customers should call (301) 223-2300.
Visit Lippincott Williams & Wilkins on the Internet: at LWW.com. Lippincott Williams &
Wilkins customer service representatives are available from 8:30 am to 6 pm, EST.
10 9 8 7 6 5 4 3 2 1

iv
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

To Will, Thomas, Jamie, Justin, Laura, Jeffrey, Jayson, Jared and

all the other children of the world.

v
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

vi
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

• I N M E M O R I A M
•

Henry J. Ralston, Professor of Physiology, University of Just after World War II, Dr. Ralston began the collabo-
California School of Medicine, San Francisco, co-authored ration with Dr. Verne T. Inman that would result in a se-
the 1st edition of Human Walking. Dr. Ralston was an in- ries of major contributions to the field of human locomo-
ternationally renowned investigator in the physiology and tion. Supported by funding from various federal agencies,
biophysics of human locomotion.Dr. Ralston was also a tal- Inman and Ralston began the Lower Extremity Amputee
ented teacher, educating several generations of neuromus- Research Laboratory, which soon evolved into the Biome-
cular physiology students as part of the Physical Therapy chanics Laboratory. With colleagues in bioengineering at
and Medicine Departments at the University of California the University of California at Berkeley, as well as with
at San Francisco as well as teaching general physiology for physicians and scientists at the University of California
nearly three decades. at San Francisco, the laboratory pioneered work that rev-
Known to his friends, family, and colleagues as “Bip,” olutionized the design of lower limb prosthetic devices.
a name given to him as an infant, Ralston was a descen- His physiological investigations focused on neuromuscu-
dant of a family that came to San Francisco from Scotland lar physiology and the energetics of walking and led to
in the early 1860s; the family started an iron works com- improved surgical approaches to lower limb repair and
pany that survived until the Depression. Ralston worked enhanced design of prostheses. Human Walking was a cul-
his way through the University of California at Berkeley as mination of Dr. Ralston’s ground breaking research and his
a newspaper writer and initiated the first regular column collaboration with Dr. Inman.
in San Francisco reviewing and critiquing radio programs.
His PhD thesis concerned the biological effects of ionizing
radiation.

vii
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

viii In Memoriam

Verne T. Inman, Professor Emeritus and former Chairman sics in the field. He was one of the pioneers in the use of
of the Department of Orthopaedic Surgery, University of electromyography to analyze muscle function.
California School of Medicine, San Francisco, co-edited Shortly after World War II, Dr. Inman, along with col-
the 1st edition of Human Walking. leagues in engineering and physiology, became involved
Dr. Inman was born in San Jose, California, Novem- in lower limb prosthetics research. This led to the forma-
ber 6, 1905. He received both his medical education and tion of the Biomechanics Laboratory at the University of
formal training in human anatomy at the University of California at San Francisco and Berkeley, of which he was
California. director from 1957 to 1973 and consultant until his death.
Dr. Inman’s primary research interest since his student Dr. Inman had long expressed the wish to prepare a book
days may be best described as functional anatomy. His summarizing the research studies on human walking in
studies on the actions of the shoulder joint, the clavicle, the Biomechanics Laboratory, and Human Walking is the
the abductor muscles of the hip, and the ankle are clas- culmination of that wish.
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

• C O N T E N T S
•
In Memoriam vii 8 Gait Adaptations in Adulthood: Pregnancy,
Contributor List xi Aging, and Alcoholism 131
Preface to the Third Edition xiii Erin E. Butler, Maurice Druzin, and Edith V. Sullivan
9 Walking for Health 149
1 Human Locomotion 1 William L. Haskell and Leslie Torburn
Verne T. Inman, Henry J. Ralston, and Frank Todd
Commentary by Dudley S. Childress and Steven A. Gard
10 Gait Analysis: Clinical Decision Making 165
Janet M. Adams and Jacquelin Perry
2 The Evolution of Human Walking 23 11 Lower Limb Prostheses: Implications
Timothy D. Weaver and Richard G. Klein and Applications 185
3 Kinematics of Normal Human John W. Michael
Walking 33 12 Simulation of Walking 193
Kenton R. Kaufman and David H. Sutherland Frank C. Anderson, Allison S. Arnold,
Marcus G. Pandy, Saryn R. Goldberg, and
4 Kinetics of Normal Walking 53
Scott L. Delp
Roy B. Davis and Kenton R. Kaufman
13 The Next Step: Restoring Walking After
5 Energetics of Walking 77 Paralysis 209
Jessica Rose, Don W. Morgan, and James G. Gamble Ronald J. Triolo and Rudi Kobetic
6 Muscle Activity During Walking 103 14 Human Walking: Six Take-Home
Jennette L. Boakes and George T. Rab Lessons 223
7 Development of Gait 119 James G. Gamble and Jessica Rose
Rosanne Kermoian, M. Elise Johanson, Erin E. Butler,
and Stephen Skinner Index 229

ix
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

x
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:7

• C O N T R I B U T O R S
•
Janet M. Adams, PT, MS, DPT Steven A. Gard, PhD
Professor and Chair, Department of Physical Therapy Director, Northwestern University Prosthetics
California State University, Northridge Research Laboratory & Rehabilitation
Northridge, California Engineering Research Program
Research Associate Professor, Physical Medicine & Rehabilitation,
Frank C. Anderson, PhD Northwestern University
Engineering Research Associate Research Health Scientist, Jesse Brown VA Medical Center
Division of Biomechanical Engineering Chicago, Illinois
Stanford University
Stanford, California Saryn R. Goldberg, PhD
Clinical Research Scientist
Allison S. Arnold, PhD Physical Disabilities Branch of the
Physical Science Research Associate National Institutes of Health
Division of Biomechanical Engineering Bethesda, Maryland
Stanford University
Stanford, California William L. Haskell, PhD
Professor Emeritus, Department of Medicine
Stanford Prevention Research Center
Jennette L. Boakes, MD
Stanford University School of Medicine
Clinical Associate Professor of Orthopaedic Surgery
Stanford, California
UC Davis School of Medicine
Pediatric Orthopaedic Surgeon
Shriners Hospitals for Children Verne T. Inman, MD, PhD†
Sacramento, California Professor Emeritus and Former Chairman of the
Department of Orthopaedic Surgery
University of California School of Medicine
Erin E. Butler, MS San Francisco, California
Biomechanical Engineer
Motion & Gait Analysis Laboratory
Lucile Packard Children’s Hospital M. Elise Johanson, PT, MS
Palo Alto, California Research Health Scientist
VA Palo Alto Health Care System
Rehabilitation Research & Development Center
Dudley S. Childress, PhD and Motion & Gait Analysis Laboratory
Professor (Emeritus) Lucile Packard Children’s Hospital
Biomedical Engineering and Physical Palo Alto, California
Medicine & Rehabilitation
Northwestern University
Kenton R. Kaufman, PhD
Chicago, Illinois
Director, Orthopaedic Biomechanics Laboratory
Professor, Biomedical Engineering
Roy B. Davis, PhD Mayo Clinic
Director, Motion Analysis Laboratory Rochester, Minnesota
Shriners Hospitals for Children
Greenville, South Carolina Rosanne Kermoian, PhD
Senior Research Scientist, Department of Orthopaedic Surgery
Scott L. Delp, PhD Stanford University School of Medicine
Professor and Chair Motion & Gait Analysis Laboratory
Bioengineering Department Lucile Packard Children’s Hospital
Stanford University Palo Alto, California
Stanford, California
Richard G. Klein, PhD
Maurice Druzin, MD Professor, Department of Anthropological Sciences
Professor, Department of Obstetrics and Gynecology Program in Human Biology
Division of Maternal & Fetal Medicine Stanford University
Stanford University School of Medicine Stanford, California
Stanford, California
Rudi Kobetic, MS
James G. Gamble, MD, PhD Motion Study Laboratory
Professor, Department of Orthopaedic Surgery L. Stokes Cleveland VA Medical Center
Stanford University School of Medicine Cleveland, Ohio
Medical Director, Motion & Gait Analysis Laboratory
Lucile Packard Children’s Hospital
Palo Alto, California †
Deceased

xi
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:7

xii Contributors

John W. Michael, MEd, CPO Stephen Skinner, MD

President, CPO Services, Inc. Clinical Professor, Department of Orthopedic Surgery
Portage, Indiana UC Davis School of Medicine
Adjunct Faculty Chief of Orthopaedics
School of Applied Physiology Shriners Hospitals for Children
Georgia Institute of Technology Sacramento, California
Atlanta, Georgia
Edith V. Sullivan, PhD
Don W. Morgan, PhD Professor, Department of Psychiatry & Behavioral Sciences and
Professor, Department of Health and Human Performance Neurosciences Program
Middle Tennessee State University Stanford University School of Medicine
Murfreesboro, Tennessee Stanford, California

Marcus G. Pandy, PhD David H. Sutherland, MD

Professor, Department of Biomedical Engineering Professor Emeritus,
The University of Texas at Austin Department of Orthopaedic Surgery
Austin, Texas University of California, San Diego
Motion Analysis Laboratory
San Diego Children’s Hospital
Jacquelin Perry, MD San Diego, California
Professor Emeritus, Department of Orthopaedics
and Department of Biokinesiology & Physical Therapy
Frank Todd, BA
University of Southern California
Berkeley, CA
Los Angeles, California

Leslie Torburn, PT, MS

George T. Rab, MD Physical Therapist
Professor and Chair, Department of Orthopaedic Surgery Motion & Gait Analysis Laboratory
UC Davis School of Medicine Lucile Packard Children’s Hospital
Sacramento, California Palo Alto, California

Henry J. Ralston, PhD† Ronald J. Triolo, PhD

Professor of Physiology Associate Professor of Orthopaedics &
University of California School of Medicine Biomedical Engineering
San Francisco, California Case Western Reserve University
VA Rehabilitation Research Career Scientist
Jessica Rose, PhD Cleveland, Ohio
Assistant Professor, Department of Orthopaedic Surgery
Stanford University School of Medicine Timothy D. Weaver, PhD
Director, Motion & Gait Analysis Laboratory Department of Human Evolution
Lucile Packard Children’s Hospital Max Planck Institute for Evolutionary Anthropology
Palo Alto, California Leipzig, Germany

†
Deceased
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:7

•P R E FA C E
•
T he third edition of Human Walking embraces the mul-
tidisciplinary approach and pragmatic spirit that was a
Gait, Clinical Gait Analysis, Lower Limb Prostheses and
Restoring Walking After Paralysis. Furthermore, we have
expanded the multidisciplinary approach to include new
chapters on rapidly developing fields such as The Evolu-
major theme of the first and second editions. The increased
tion of Human Walking, Gait Adaptations in Adulthood,
breadth and depth of material for the third edition re-
Walking for Health, and Simulation of Walking. It was
flects the expanding nature of the field. Our understand-
clear a decade ago that biomechanical modeling would
ing of human walking and the information available has
make interesting contributions to our understanding of
increased exponentially over the past decade. New areas
human walking. However, it was not certain how rapidly
of knowledge have developed over the last several years
these contributions would come and how important they
in fields such as physical anthropology, neuromotor de-
would ultimately be for identifying the sources of patho-
velopment, and biomechanics, with groundbreaking ad-
logical gait. Chapter 12, “Simulation of Walking,” shows
vances in biomechanical modeling and artificial walking.
just how valuable biomechanical modeling can be and pro-
Increasingly precise measurement techniques have made
vides a fresh understanding of the scientific basis for the
it possible to study the neuromuscular activation and intri-
treatment of patients with walking disorders. Biomechani-
cate biomechanics of human walking and have deepened
cal modeling is now used in the clinical setting to plan such
our understanding of the neurological and musculoskele-
surgical procedures as tendon transfers, tendon lengthen-
tal mechanisms underlying walking disorders. The third
ings, and osteotomies. The final chapter, “Six Take-Home
edition summarizes and integrates this new information
Lessons,” summarizes some of the essential elements of
with our classical understanding of human walking.
human walking for students who are new to the field.
The first edition of Human Walking, published in 1981,
Human walking is an extremely complex activity whose
was written by an interdisciplinary team of investiga-
apparent simplicity disappears when one attempts a quan-
tors, composed of Verne T. Inman, an orthopedic surgeon,
titative or even qualitative description of the process.
Henry J. Ralston, a physiologist, and Frank Todd, an en-
Fortunately, the theories and techniques of modern mo-
gineer. In the years following the publication of the first
tion analysis have markedly improved our ability to de-
edition, a generation of students and researchers used
scribe and understand normal and pathological ambula-
Human Walking as both a primary text and a reference
tion. Much of the current success is a result of the wide
as they expanded the available knowledge in the field of
interest in human walking as demonstrated by a diverse
motion analysis. In the second edition, we chose to pre-
group of clinicians and scientists currently working in the
serve the multidisciplinary approach as well as the prag-
field, including orthopaedic surgeons, physical therapists,
matism of the previous edition, while extending the scope
bioengineers, physiatrists, neurologists, orthotists, pros-
and the scale of the book. We invited a diverse group of dis-
thetists and exercise physiologists. The third edition of
tinguished contributors to share their ideas, information,
Human Walking is geared to this diverse group of students,
and expertise. The third edition expands on this theme.
researchers, and clinicians, and continues the pragmatic
We have preserved the classic and original chapter “Hu-
tradition of providing useful information from a broad
man Locomotion” written by Verne T. Inman, Henry J.
spectrum of expertise while offering a springboard for
Ralston, and Frank Todd, and added commentary on
future advances in the field by the next generation.
the determinants of gait that integrates new informa-
tion. There are updated chapters on Kinematics of Nor- Jessica Rose
mal Walking, Kinetics of Normal Walking, Energetics of James G. Gamble
Walking, Muscle Activity During Walking, Development of September 2005

xiii
P1: IML/OVY P2: IML/OVY QC: IML/OVY T1: IML
GRBT092-FM Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:20

xiv
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

• C h a p t e r 1
•
◗ Human Locomotion
Verne T. Inman, Henry J. Ralston, and Frank Todd
Commentary by Dudley S. Childress and Steven A. Gard

Locomotion, a characteristic of animals, is the process by as will be shown later, with surprising efficiency. A con-
which the animal moves itself from one geographic posi- clusion that seems inescapable is that each of us learns
tion to another. Locomotion includes starting, stopping, to integrate the numerous variables that nature has be-
changes in speed, alterations in direction, and modifica- stowed on our individual neuromusculoskeletal systems
tions for changes in slope. These events, however, are tran- into a smoothly functioning whole. Obviously, our bipedal
sitory activities that are superimposed on a basic pattern. plantigrade type of progression imposes gross similarities
In walking and running animals, this pattern can be de- on our manner of walking. These are easily identified. We
fined as a rhythmic displacement of body parts that main- must oscillate our legs, and as we do our bodies rise and
tains the animal in constant forward progression. fall with each step. The movements parallel to the plane
The majority of mammals are quadripedal. When walk- of progression are large, and the individual variations in
ing slowly, quadripeds tend to coordinate their four limbs relation to the size of the total angular displacements are
so that three of their feet are on the ground. A crawl- relatively small. When these aspects of human walking are
ing infant uses its limbs in a sequence that is essentially considered, the use of average values helps to develop a
quadripedal, only advancing one while the other three general understanding of the basic relationships that exist
support its body on the floor. This provides the stabil- between the major segments of the lower limb. Upon these
ity of a tripod. This stability is lost when the animal be- basic activities are superimposed numerous less obvious
comes bipedal, and while bipedal locomotion seems sim- movements of individual parts of the body. These small
pler, it requires greater neural control. The mastering of the movements occur in planes closer to the coronal and trans-
erect bipedal type of locomotion is a relatively prolonged verse planes of the body, and in these small movements, the
affair and appears to be a combination of instinct and individual variations are much greater.
learning. Furthermore, when the locations of axes of movement
If walking is a learned activity, it is not surprising that are determined and ranges of motion measured both in the
each of us displays certain personal peculiarities superim- cadaver and in the living, marked individual differences
posed on the basic pattern of bipedal locomotion. Physical are disclosed. The differences in these small movements
anthropologists have studied the differences between races bestow on each of us a distinctive manner of walking. Here,
and measured the variations in skeletal parts. Anatomists the use of average values can hinder the recognition of cer-
are aware of the presence of individual variations. All of tain interrelationships that must exist between the partic-
us are aware that individuals walk differently; one can ipating joints. This is particularly true when one is trying
often recognize an acquaintance by his manner of walk- to understand the functional behavior of the joints of the
ing even when seen at a distance. Tall, slender people walk ankle and foot.
differently from short, stocky people. People alter their A hypothesis is easily formulated that seems to explain
manner of walking when wearing shoes with different heel most observations, including the peculiar behavior of the
heights. A person walks differently when exhilarated than major segments of the body during walking. This hypoth-
when mentally depressed. With these ideas in mind, one esis states that the human body will integrate the motions
may legitimately question the usefulness of anthropomet- of the various segments and control the activity of the
ric data and averages in furthering our understanding of muscles so that the metabolic energy required for a given
human walking. distance walked is minimized. In later sections, it will be
Certainly everyone has his own idiosyncratic way of shown that any interference with normal relationships be-
walking, and there is no such thing as an average person. tween various segments of the body invariably increases
However, most of us do walk with reasonable facility and, the metabolic cost of walking.

1
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

2 Chapter 1 / Human Locomotion

PROCESS OF WALKING a result of the alternating accelerations and decelerations

of the body.
The term walking is nonspecific. Its connotation is that As the body passes over the supporting leg, it rises until
of a cyclic pattern of body movements that are repeated the foot is directly underneath and then descends again
over and over, step after step. Consequently, descriptions of as the foot passes behind. The highest point in elevation
walking customarily deal with what happens in the course occurs when the speed is lowest, and the lowest point in
of just one cycle, with the assumption that successive cy- elevation occurs when the speed is highest.
cles are all about the same. Although this assumption is During the period of single support, the body also tends
not strictly true, it is a reasonable approximation. Apart to shift laterally over the support limb. The pelvis achieves
from the multiple variations that may occur between dif- its maximal lateral displacement somewhat after mid-
ferent individuals or within the same individual as a result stance and then starts back toward the other side. The
of changes in the speed of walking, or such factors as al- amount of lateral sway increases when the tread width
terations in footwear, there are certain observable events is increased. While individual variations in the measured
that are shared by all. magnitudes of these motions will always be observed in
Human walking is a process of locomotion in which any group of people, the motions will be present to some
the erect, moving body is supported by first one leg and degree in everyone. Normally there is symmetry in the
then the other. As the moving body passes over the sup- movement, and the patterns repeat themselves with each
porting leg, the other leg swings forward in preparation successive cycle.
for its next support phase. One foot or the other is always Although the displacements of the entire body through
on the ground, and during that period when the support of space may be described as translational, this translation
the body is transferred from the trailing to the leading leg is achieved by the angular displacements of various seg-
there is a brief period when both feet are on the ground. ments of the body about axes that lie in the proximity of
As a person walks faster, these periods of double support joints. A principal task in describing human locomotion is
become smaller fractions of the walking cycle until, even- to measure the angular displacements of the various seg-
tually, as a person starts to run, they disappear altogether ments during the translational movement of the body as
and are replaced by brief periods when neither foot is on a whole. Because the translational movements are the fi-
the ground. The cyclic alternations of the support function nal product of the angular displacements of the individual
of each leg and the existence of a transfer period when both segments, and these are easily discernible and measurable,
feet are on the ground are essential features of the locomo- they may be used as one set of parameters for the descrip-
tion process known as walking. tion of the walking cycle. However, a description that deals
In the act of walking there are two basic requisites: solely with movement and ignores the forces that produce
(1) continuing ground reaction forces that support the these motions constitutes only a small part of the entire
body, and (2) periodic movement of each foot from one po- story of human walking.
sition of support to the next in the direction of progression.
These elements are necessary for any form of bipedal walk-
ing no matter how distorted by physical disability. They are MAJOR DISPLACEMENTS OF THE
equally necessary when prosthetic or orthotic devices are BODY DURING WALKING
used.
These two basic requisites of walking give rise to spe- Synchronous movements of nearly all the major parts of
cific body motions that are universally observable during the body occur during walking at moderate speeds. The
walking. As the body passes over the weight-bearing limb, pelvis lists, rotates, and undulates as it moves forward. The
three different deviations occur from uniform progression segments of the lower limb show displacements in all three
in a straight line. With each step, the body speeds up and planes of space, while the shoulders rotate and the arms
slows down slightly, it rises and falls a few centimeters, swing out of phase with the displacements of the pelvis
and it weaves slightly from side to side. These motions are and legs. It seems reasonable to begin the description of
related to one another in a systematic fashion. walking with a discussion of the translation of the body as
The body must slow down and then speed up again dur- a whole through space. To do this, the concept of the path-
ing each step because the support provided by the legs does way of the center of mass of the body will be used. The cen-
not remain directly under the body at all times. A support- ter of mass of any body is a point such that if any plane is
ing foot starts out ahead of the body where it tends to slow passed through it, the mass moments on one side of the
the body down, and then it passes under the body and to plane are equal to the mass moments on the other. If the
the rear, where it tends to speed the body up again. This body is suspended at this center of mass, it will not tend
motion is difficult to see, but easy to sense when a person to tip in any direction. During walking, the center of
carries a shallow pan of water. It is almost impossible to mass of the body, although not remaining in an absolutely
prevent the water from surging backward and forward as fixed position, tends to remain within the pelvis. This is
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 1 / Human Locomotion 3

phase. The center of mass falls to its lowest level during the
middle of double weightbearing, when both feet are in con-
tact with the ground. The curve is remarkably smooth and
is found to fluctuate evenly between maxima and minima
of displacements, with few irregularities. It is interesting
to note that at their maximal vertical displacement, the
head and the center of mass are slightly lower than when
the subject is standing on both feet. In other words, in a
smooth walk, a person is slightly shorter than when he
is standing, so that if he were to walk through a tunnel
the height of which corresponded exactly to his standing
height he could do so without fear of bumping his head.
The center of mass of the body is also displaced later-
ally in the horizontal plane. In this plane, too, it describes a
FIGURE 1-1. Displacements of center of mass in three planes of sinusoidal curve, the maximal values of which alternately
space during single stride (cycle). The actual displacements have pass to the right and to the left in association with the sup-
been greatly exaggerated. (A), Lateral displacement in a horizon- port of the weight-bearing limb. The curve is sinusoidal,
tal plane; (B), vertical displacement. Combined displacements of
A and B as projected onto a plane perpendicular to the plane of at one-half the frequency of the vertical displacement.
progression are shown in C. When viewed from the back, the body is seen to un-
dulate up and down and swing from side to side during
each cycle. If the vertical and lateral displacements are
convenient for two reasons. Measurements of the move- considered as pure sine waves, with the frequency of the
ments of the pelvis in the three planes of space are readily vertical displacements being precisely twice that of the lat-
made, and the pelvis becomes a suitable structure to sep- eral displacements and the peaks being achieved at the
arate the body into upper and lower parts, which behave same time, then the curve of displacement of the center of
differently during walking. mass, as projected onto a plane at right angles to the line
In normal level walking, the center of mass describes of progression, is in the form of a “U.” At higher speeds of
a smooth sinusoidal curve when projected on the plane walking, this situation is approximated; at lower speeds,
of progression (Fig. 1-1). The total amount of vertical dis- however, the peak of the curve for vertical displacement
placement in normal adult men is typically about 5 cm at is reached slightly before the peak of lateral displacement.
the usual speeds of walking. The summits of these oscil- This causes the curve of movement of the center of mass
lations appear at about the middle of the stance (foot on as projected on a coronal plane (a vertical plane at a right
ground) phase of the supporting limb. The opposite limb angle to the line of progression) to resemble a slightly dis-
is at this time in the middle of its swing (foot off ground) torted lazy 8 (Fig. 1-2).

FIGURE 1-2. Effect of variations in speed on displacement of pelvis as projected onto plane
perpendicular to plane of progression (see Fig. 1-1C).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

4 Chapter 1 / Human Locomotion

For the sake of simplicity, a series of models will be the center of mass in the succeeding arc of translation is
employed to illustrate how the smooth sinusoidal displace- less, and the angular displacement at the hip in flexion and
ment pathway is achieved in bipedal locomotion (5,6). The extension is reduced.
first model will show the body as consisting solely of a bar
representing the pelvis, with the center of mass depicted
Pelvic List
as a small block lying midway between the two hips (Fig.
1-3). The legs will be represented as rigid levers without In normal walking, the pelvis lists downward in the coro-
foot, ankle, or knee mechanisms, articulated only at the nal plane on the side opposite to that of the weight-bearing
hip joints, which will permit flexion and extension only. limb (positive Trendelenburg). At moderate speeds, the
Such a system of quasilocomotion would produce some- alternate angular displacement is about 5 degrees. The dis-
thing analogous to the process of stepping off distances placement occurs at the hip joint, producing an equivalent
with a pair of compasses or dividers, the pathway of the relative adduction of the supporting limb and relative ab-
center of mass of such a system being a series of intersect- duction of the other limb, which is in the swing phase of
ing arcs. the cycle. To permit pelvic list, the knee joint of the non–
The radius of each arc would be equal to the length of weight-bearing limb must flex to allow clearance for the
the limbs, and with each step, the extent of flexion and ex- swing-through of that member.
tension of the hip joint would be the same. Locomotion The effects of pelvic list on the pathway of the center
of this type might be imitated, but imperfectly, by walking of mass are evident in the experimental model (Fig. 1-5).
on one’s heels with the knees fixed in extension. Such a type As the lateral list occurs while the body is passing over
of locomotion would require that the center of mass be ele- the vertical supporting member in early stance phase, the
vated to a height equal to the height of the center of mass in center of mass is lowered. Thus, the summit of the arc is
the standing person; it would also result in a severe jolt at lowered, further flattening the pathway. In addition and
the point of intersection of each two arcs, where there is an perhaps more importantly, pelvic list contributes to the
abrupt change in the direction of movement of the center effectiveness of the abductor mechanism of the hip (the
of mass. Decreasing the total elevation, depressing the cen- abductor muscles and iliotibial tract). The latter effect will
ter of mass, and smoothing the series of interrupted arcs be discussed in greater detail in the section on the phasic
require coordinated movements involving all the joints of action of muscles (see Chapter 6).
the lower limb. These individual movements can be con-
sidered as elements that contribute to the total process of
Knee Flexion in Stance Phase
walking. A qualitative description of the principal elements
is presented in the following paragraphs to provide a basis A characteristic of walking at moderate and fast speeds
for the quantitative descriptions in later chapters. is knee flexion of the supporting limb as the body passes
over it. This supporting member enters stance phase at heel
strike with the knee joint in nearly full extension. There-
Pelvic Rotation
after, the knee joint begins to flex and continues to do so
In normal level walking, the pelvis rotates about a vertical until the foot is flat on the ground. A typical magnitude of
axis alternately to the right and to the left, relative to the this flexion is 15 degrees. Just before the middle of the pe-
line of progression. At the customary cadence and stride of riod of full weight bearing, the knee joint once more passes
typical people, the magnitude of this rotation is approxi- into extension, which is immediately followed by the ter-
mately 4 degrees on either side of the central axis, or a total minal flexion of the knee. This begins simultaneously with
of some 8 degrees. This value usually increases markedly heel rise, as the limb is carried into swing phase. During
when speed is increased. Because the pelvis is a rigid struc- this period of stance phase, occupying about 40% of the
ture, the rotations occur alternately at each hip joint and cycle, the knee is first extended, then flexed, and again ex-
require a deviation from pure flexion and extension of the tended before its final flexion.
hips. During the beginning and end of the stance phase, knee
The significance of pelvic rotation can best be appre- flexion contributes to smoothing the abrupt changes at the
ciated by a study of the theoretical model (Fig. 1-4). The intersections of the arcs of translation of the center of mass
effects of pelvic rotation are to flatten somewhat the arc of (Fig. 1-6).
the passage of the center of mass in compass gait by elevat- These three elements of gait—pelvic rotation, pelvic list,
ing the ends of that arc. In consequence, the angles at the and knee flexion during early stance phase, all act in the
intersections of successive arcs are rendered less abrupt same direction by flattening the arc through which the cen-
and, at the same time, are elevated in relation to the sum- ter of mass of the body is translated. The first (pelvic rota-
mits. In this way, the severity of the impact at floor contact tion) elevates the ends of the arc, and the second and third
is reduced. The force required to change the direction of (pelvic list and knee flexion) depress its summit. The net
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 1 / Human Locomotion 5

FIGURE 1-3. Simplified model depicting bipedal locomotion. The pelvis is a double-forked bar artic-
ulating with spheres depicting the hip joints and carrying a small block that represents the center of
mass of the body. The legs are straight members without knee, ankle, or foot components. Note that
the pathway of the center of mass is through a series of intersecting arcs. (From Saunders JB, Inman
VT, Eberhart HD. The major determinants in normal and pathological gait. J Bone Joint Surg 1953;
35–A:543.)

effect is the passage of the center of mass through a seg- body. However, if no additional elements were active, the
ment of a circle, the radius of which is about 2.2 times pathway of the center of mass would still consist of a series
longer than the length of the lower limb. The effective of arcs, and at their intersections the center of mass would
lengthening of the limbs reduces materially the range of be subject to a sudden change in vertical displacement.
flexion and extension at the hip joint required to maintain This would result in a jarring effect on the body. Thus,
the same length of stride. an additional mechanism must be active that smooths the
The three elements so far discussed (pelvic rotation, pathway of the center of mass by a gradual change in the
pelvic list, and knee flexion) act to decrease the magnitude vertical displacement of the center of mass from a down-
of the vertical displacement of the center of mass of the ward to an upward direction, converting what would be a
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

6 Chapter 1 / Human Locomotion

FIGURE 1-4. Effect of pelvic rotation. By permitting the pelvis to rotate in a horizontal plane during
locomotion, the center of mass is prevented from falling as far during the phase of double weight bearing
as was shown in Figure 1-3. The solid line at the top represents the curve shown Figure 1-3. (Adapted
from Saunders JB, Inman VT, Eberhart HD. The major determinants in normal and pathological gait.
J Bone Joint Surg 1953;35–A:543.)

series of intersecting arcs into a sinusoidal path. This is ac- stance phase. This in turn allows the initial knee flexion to
complished by certain movements in the knee, ankle, and act more effectively in smoothing the pathway of the hip.
foot. To understand the mechanics involved, a series of simple
The single most important factor in achieving the con- drawings may be helpful. In Figure 1-7, the actual pathway
version of the pathway of the center of mass from a series of the knee joint during stance phase is shown. Except for
of intersecting arcs to a smooth curve is the presence of an initial rise, the pathway is relatively flat. In Figure 1-8,
a foot attached to the distal end of the limb. Through three other situations are shown. If no foot is attached to
its action, the foot enables the pathway of displacement the shank, the pathway of the knee is an arc whose ra-
of the knee to remain relatively horizontal during the entire dius is the distance from the floor to the knee. By simply
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 1 / Human Locomotion 7

FIGURE 1-5. Effect of pelvic list. Normally the pelvis drops slightly on the non–weight-bearing side
during walking (positive Trendelenburg). The result is that the center of mass need not be elevated as
much when the body passes over the weight-bearing leg during midstance. Because of the pelvic list, the
swinging leg becomes relatively too long to clear the floor during the midswing phase. Flexion of the knee
allows for this clearance. The solid line at the top represents the curve shown in Figure 1-4. (Adapted
from Saunders JB, Inman VT, Eberhart HD. The major determinants in normal and pathological gait.
J Bone Joint Surg 1953;35–A:543.)

adding a foot rigidly fixed to the shank (no ankle), the path- joint, results in achievement of the normal pathway of the
way, although it is composed of two arcs, approaches more knee joint.
closely the normal course. Provision of a flail ankle results At the time of heel strike, the center of mass of the
in a pattern resembling the pathway of the knee without body is falling. This downward movement is decelerated
any foot. Provision of a normal ankle, however (Fig. 1-9), by slight flexion of the knee against the resistance of the
with proper phasing of the extensor and flexor muscles quadriceps. After heel strike, the foot is plantar flexed
and only a minor amount of motion occurring in the ankle against the resisting tibialis anterior muscle. This plantar
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

8 Chapter 1 / Human Locomotion

FIGURE 1-6. Knee flexion during stance. Except at very low speeds of walking, the knee undergoes
approximately 15 degrees of flexion immediately after heel strike and continues to remain flexed until
the center of mass has passed over the weight-bearing leg. The effect of this knee flexion is twofold.
Initially, it absorbs part of the impact of the body at heel strike and later it decreases the amount that
the center of mass must be elevated as it passes over the weight-bearing leg. The solid line at the top
represents the curve shown in Figure 1-5. (Adapted from Saunders JB, Inman VT, Eberhart HD. The
major determinants in normal and pathological gait. J Bone Joint Surg 1953;35–A:543.)

flexion of the foot occurs about a point where the heel con- ing midstance, it pronates to a varying degree. Although
tacts the floor. Rotation about this point causes the leg to this pronation contributes only a few millimeters to the
undergo relative shortening and the ankle to be carried further relative shortening of the leg, the elastic compo-
slightly forward in the direction of progression until the nents in the plantar region of the foot assist in absorbing
foot is flat. Contraction of the quadriceps acting on the the shock of impact.
knee and the tibialis anterior muscle on the foot causes The center of mass begins its upward movement im-
these movements to be slowed, and the downward motion mediately after it has passed in front of the weightbearing
of the center of mass of the body is smoothly decelerated. foot, as the forward momentum of the body carries the
In addition, as the foot receives the weight of the body dur- body up and over the weight-bearing leg. After the center
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 1 / Human Locomotion 9

FIGURE 1-7. Pathway of knee in walking at moderate speed. Note that there is a slight elevation
immediately after heel strike, but for the remainder of stance phase the pathway is relatively straight
and shows only a slight declination from the horizontal. (Reproduced from Saunders JB, Inman VT,
Eberhart HD. The major determinants in normal and pathological gait. J Bone Joint Surg 1953;35–A:543.)

of mass has passed over and in front of the foot, its imme- Lateral Displacement of Body
diate fall is delayed by relative elongation of the weight-
As mentioned previously, the body is shifted slightly over
bearing leg through extension of the knee, plantar flexion
the weight-bearing leg with each step; there is a total lat-
at the ankle, and supination of the foot. All these elements
eral displacement of the body from side to side of approx-
acting in proper relationships lead to the smoothing of the
imately 4 to 5 cm with each complete stride. This lateral
passage of the center of mass into an approximately sinu-
displacement can be increased by walking with the feet
soidal pathway (Figs. 1-10, 1-11, and 1-12).

FIGURE 1-8. Effect of foot on pathway of knee. (A) Arch described when there is no foot. (B) Effect
of foot without ankle. Note that the pathway now comprises two intersecting arcs. However, it does not
fall abruptly at the end of stance and begins to resemble the normal pathway. (C) Effect of foot and flail
ankle. (Reproduced from Saunders JB, Inman VT, Eberhart HD. The major determinants in normal and
pathological gait. J Bone Joint Surg 1953;35–A:543.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

FIGURE 1-9. Effect of ankle motion, controlled by muscle action, on pathway of knee. The smooth
and flattened pathway of the knee during stance phase is achieved by forces acting from the leg on the
foot. Foot slap is restrained during initial lowering of the foot; afterward, the plantar flexors raise the
heel.

FIGURE 1-10. Interrupted light studies. The photograph was obtained by having a subject walk in front
of the open lens of a camera while carrying small light bulbs located at the hip, knee, ankle, and foot. A
slotted disc was rotated in front of the camera producing a series of white dots at equal time intervals.
Note that the curve of displacement at the hip is a smooth curve but is not sinusoidal. This is due to
the differences in phase of the two legs. (From Eberhard HD, Inman VT. An evaluation of experimental
procedures used in a fundamental study of human locomotion. Ann N Y Acad Sci 1951;51:1213.)

10
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 1 / Human Locomotion 11

FIGURE 1-11. Vertical displacements of hip joints. Although the pathways of the hip joints are smooth
curves, they are not sinusoidal and they are 180◦ out of phase. (From Saunders JB, Inman VT, Eberhart
HD. The major determinants in normal and pathological gait. J Bone Joint Surg 1953;35–A:543.)

more widely separated (Fig. 1-13) and decreased by keep- Rotations in Transverse Plane
ing the feet close to the plane of progression (Fig. 1-14). Reference has already been made to the transverse rota-
Normally, the presence of the tibiofemoral angle (slight tions of the pelvis that occur during walking. These rota-
genu valgum) permits the tibia to remain essentially verti- tions are easily seen when attention is called to them. There
cal and the feet close together, while the femurs diverge to are also other transverse rotations, involving the parts of
articulate with the pelvis. the body above and below the pelvis that merit attention.

FIGURE 1-12. Sinusoidal pathway of center of mass. The center of mass, which lies between the hip
joints, is equally affected by the displacements of each hip. The combined effect is a sinusoidal curve of
low amplitude. (From Saunders JB, Inman VT, Eberhart HD. The major determinants in normal and
pathological gait. J Bone Joint Surg 1953;35–A:543.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

12 Chapter 1 / Human Locomotion

FIGURE 1-13. Lateral displacement of center of mass. Dur-

ing each step, the body is displaced over the weight-bearing leg.
With a wide walking base (stride width) the amplitude is large.
(From Saunders JB, Inman VT, Eberhart HD. The major determi- FIGURE 1-14. Effect of narrow walking base. By providing the
nants in normal and pathological gait. J Bone Joint Surg 1953;35– model with a degree of genu valgum, the walking base is nar-
A:543.) rowed and the amplitude of lateral displacement is decreased
while the shank remains vertical. (From Saunders JB, Inman VT,
Eberhart HD. The major determinants in normal and pathologi-
cal gait. J Bone Joint Surg 1953;35–A:543.)

Rotations of Thorax and Shoulders

That the thorax and shoulders rotate during walking is
Rotations of Thigh and Leg (Shank)
easily seen. Interestingly, at moderate speeds of walking,
these rotations are approximately 180◦ out of phase with Less obvious but still considerable are the transverse rota-
the pelvic rotation. The rotation of the shoulders produces tions of the thigh and shank (4). In contrast to the shoul-
arm swing, and even the casual observer recognizes that ders, the rotations of the thigh and shank are in phase
the forward swing of one leg is accompanied by a forward with the pelvic rotation. The interesting fact is that their
swing of the arm on the opposite side. This opposite rota- rotatory displacements increase progressively from pelvis
tion of the pelvis as compared with the shoulders appears to thigh to shank. Thus, the shank on the average rotates
to provide a balancing effect that smooths the forward pro- approximately 3 times as much in a transverse plane as
gression of the body as a whole. Its suppression leads to does the pelvis. These rotations, however, show marked
inability to progress in a straight line at higher speeds of differences and constitute one of the factors that provide
walking and, as will be shown, to increased energy expen- distinctive characteristics to each individual’s appearance
diture through greater muscular effort. when walking.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 1 / Human Locomotion 13

FIGURE 1-15. Rotations of pelvis, femur, and tibia in transverse plane: composite curves of 19 young
male adults. (Adapted from Levens AS, Inman VT, Blosser JA. Transverse rotation of the segments of
the lower extremity in locomotion. J Bone Joint Surg 1948;30–A:859.)

In general, the pelvis, thigh, and leg begin to rotate inter- The axis of the ankle joint has been found to vary in
nally toward the weight-bearing leg at the beginning of its obliquity in the coronal plane from 2 degrees to 23 degrees
swing phase. This rotation is continued during the double (3). The ability of the ankle to participate in the absorption
weight-bearing phase and into midstance. At midstance, of the rotations of the shank depends on the obliquity of the
there is an abrupt change, and the leg begins to rotate ex- ankle axis and the range of flexion and extension used. The
ternally and continues to do so until the beginning of its effect of an oblique axis on the foot during the swing phase
next swing phase (Fig. 1-15). and on the shank during the stance phase is clearly shown
in Figures 1-17 and 1-18. During the swing phase, with the
foot free, the foot toes outward on dorsiflexion and inward
on plantar flexion (Fig. 1-17). During stance, with the foot
Rotations in the Ankle and Foot fixed to the floor, relative dorsiflexion produces internal
During the swing phase of walking, the segments of the rotation of the shank, and relative plantar flexion causes
lower limb (including the foot) are free in space and can ro- external rotation of the shank (Fig. 1-18). Therefore, the
tate internally without restriction (Fig. 1-16A). During the ankle joint, in proportion to the obliquity of its axis and the
stance phase, the foot is on the floor and external rotation amount of dorsiflexion, may participate in the absorption
of the leg occurs because mechanisms exist in the ankle of the transverse rotations of the shank during the stance
and foot that permit the leg to rotate externally while the phase.
foot remains stationary. If such mechanisms did not exist, However, the structure that is principally involved is
the foot would have to slip as shown in Figure 1-16B. the subtalar joint. Its ability to permit transverse rotation
There is an interesting interrelationship between the of the leg without slippage of the foot on the ground is
ankle and the hindfoot. Both the ankle and subtalar joints clearly shown in the following figures. The subtalar joint
are capable to varying degrees of absorbing the trans- is a single-axis joint whose axis is inclined approximately
verse rotations of the shank during the stance phase of 45 degrees to the foot (horizontal) and shank (vertical) in
walking. the standing position. It functions essentially as a mitered
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

14 Chapter 1 / Human Locomotion

FIGURE 1-16. Transverse rotations of pelvis, leg, and foot. A model of the pelvis and lower limb is
viewed from above. Small sticks have been attached to the pelvis and femurs to reveal more clearly that
the leg rotates through a greater range than the pelvis. The transverse rotations are readily seen. The
actual angular displacements have been exaggerated approximately threefold for emphasis. Note that
in (A) the swinging leg is free to rotate internally from toe-off to heel strike. In (B) the leg is in stance
phase and must rotate externally through the same amount. On a slippery surface, the foot would have
to slip as shown in the figure.

hinge; the basic mechanism is shown in Figures 1-19 and surface, eversion of the heel produces what is in essence a
1-20. In the living, the relationship between internal and pronated foot (Fig. 1-23). In this situation, two interesting
external rotation of the leg and pronation and supination observations may be made. The degree of plantar flexion
of the foot may be easily demonstrated by affixing targets and dorsiflexion that is possible in the midfoot is maximal.
on the leg and midfoot (Fig. 1-21). The interaction between The amount of dorsiflexion appears to be limited by the
the ankle and subtalar joints during a single stride in a plantar aponeurosis, which becomes taut when the fore-
young adult male is shown in Figure 1-22. foot is forcibly dorsiflexed against the everted heel. With
The changing of the foot from a mobile structure dur- the foot in this position, dorsiflexion of the great toe pro-
ing the first part of the stance phase into a rigid lever at duces additional tension on the plantar aponeurosis, and
push-off is a complicated and not completely understood the arch rises (Hicks’ windlass action) (2). Support of the
mechanism. That such a mechanism exists is readily longitudinal arch in the pronated foot seems to depend
demonstrated in a normal foot. If the forefoot is grasped predominantly on the plantar fasciae and the aponeuro-
by the examiner and firmly held, as would occur if the sis. If the heel is inverted while the forefoot is fixed, the
forefoot were supporting the body weight on the walking longitudinal arch is seen to rise, the plantar aponeurosis
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

FIGURE 1-17. Simple wooden model of shank and foot with

hinge representing ankle joint. The axis has been inclined 25 de-
grees to the floor. The shank is fixed and the foot is free to move.
Note that on dorsiflexion, the foot toes outward and on plantar
flexion toes inward. (From Inman VT. The Joints of the Ankle.
Baltimore: Williams & Wilkins, 1976.)

FIGURE 1-19. Action of subtalar joint. Two pieces of wood have

been connected by a hinge whose axis is positioned between the
vertical member (shank) and the horizontal member (foot) at
45 degrees. Note that in A, internal and external rotation of the
shank around a vertical axis causes the foot to rotate around
a horizontal axis. In B, a pivot joint has been inserted in the
foot to represent the transverse tarsal joint, thus permitting the
forefoot to remain in a fixed position. (From Inman VT, Mann RA.
Biomechanics of the foot and ankle. In: Mann RA, ed. DuVries’
Surgery of the Foot, 4th ed. St. Louis: The C. V. Mosby Co., 1978.)

FIGURE 1-18. Same model as in Figure 1-17. The foot is fixed and the shank is moved from its
neutral position (A) to a position of relative dorsiflexion (B). Note that internal rotation of the shank
has occurred. In C, the shank has been moved to a position of relative plantar flexion in relation to the
foot. Note that the shank has rotated externally. (From Inman VT. The Joints of the Ankle. Baltimore:
Williams & Wilkins, 1976.)

15
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

16 Chapter 1 / Human Locomotion

FIGURE 1-20. Action of subtalar joint, represented as hinge lo-

cated at angle of 45 degrees to shank and foot. The forefoot has
been divided into two segments. The medial segment represents
the three medial rays articulating with the hindfoot through the
talonavicular joint. The lateral segment represents the two lateral
rays articulating with the hindfoot through the calcaneocuboid
joint. Note that rotation of the shank is accompanied by supina-
tion and pronation of the foot. (From Inman VT, Mann RA.
DuVries’ Surgery of the Foot. St Louis: Mosby, 1978.)

FIGURE 1-22. Synchronous motion in ankle and subtalar joints.

Note that variations in the degree of toe-out and toe-in cause vari-
ations in the magnitude, phasic action, and angular movements
of both ankle and subtalar joints. (From Wright DB, Desai SM,
Henderson WH. Action of the subtalar and ankle-joint complex
during the stance phase of walking. Bone Joint Surg 46–A:361,
1964.)

FIGURE 1-21. Rotation of leg accompanied by pronation and

supination of foot. (A) Internal rotation of leg causing foot to
pronate. (B) External rotation of leg causing foot to supinate. FIGURE 1-23. Change from mobile to rigid foot. With heel ev-
(From Inman VT. The Joints of the Ankle. Baltimore: Williams & erted, maximal midfoot motion is permitted; with heel inverted
Wilkins, 1976.) and forefoot fixed, the foot becomes rigid.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 1 / Human Locomotion 17

FIGURE 1-24. Rising on toes. Note that as the body weight is

transferred to the forefoot, the heels invert, the legs rotate ex-
ternally, and the longitudinal arches rise. The barograph records FIGURE 1-26. Angle of metatarsal heads: the angle between the
the distribution of weight (white areas) between all the metatarsal long axis of the foot and a line connecting the heads of the sec-
heads and the toes. (From Inman VT, Mann RA. DuVries’ Surgery ond and fifth metatarsals, measured in 100 randomly selected
of the Foot. St. Louis: The C.V. Mosby Co., 1978.) x-rays. (Data from Moskowitz PS. Anthropometric studies of the
human metatarsus as seen on x-ray. I. Normal feet. [Thesis]. Un-
published, 1967.)

becomes lax, and the Hicks’ windlass action is suppressed.

However, something occurs in the transverse tarsal artic-
ulations that results in a rigid foot. No fully adequate ex- Inversion of the heel occurs when a standing person
planation for this phenomenon has been published, and a transfers the body weight from the heel to the forefoot and
review of the literature reveals that only a few studies have rises onto the metatarsal heads and toes (Fig. 1-24). This
been undertaken (1). results in an attempt to supinate the foot, with raising of
the longitudinal arch on the medial side and depressing of
the arch on the lateral side. With this movement, the body
weight is shifted to the lateral side of the foot. Fortunately,
the presence of shorter metatarsals on the lateral side
of the foot permits this supination to occur and results in
the lateral metatarsals and toes participating to a greater
degree in bearing the body weight.
To distribute the weight between all metatarsal heads,
the foot must deviate laterally at push-off. That this occurs
normally in walking is confirmed by the inspection of any
well-worn shoe. A crease in the leather is evident in the re-
gion of the juncture of the cap and vamp of the shoe. This
crease is always oblique to the longitudinal axis of the shoe
(Fig. 1-25) and marks the position of the metatarsopha-
langeal articulations. The crease shows individual varia-
FIGURE 1-25. Crease between cap and vamp in well-worn tions in obliquity in accordance with the variations in the
shoes. angle of the metatarsophalangeal break. Figure 1-26 shows
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

18 Chapter 1 / Human Locomotion

the variation in the angle of the metatarsal heads and the 2. Hicks JH. The mechanics of the foot. II. The plantar aponeurosis and
long axis of the foot in 100 randomly selected x-rays. the arch. J Anat 1954;88:25.
3. Inman VT. The Joints of the Ankle. Baltimore: Williams & Wilkins,
1976:29–44.
4. Levens AS, Inman VT, Blosser JA. Transverse rotation of the segments
of the lower extremity in locomotion. J Bone Joint Surg 1948;30–A:859.
REFERENCES 5. Saunders JB, Inman VT, Eberhart HD. The major determinants in
normal and pathological gait. J Bone Joint Surg 1953;35–A:543.
1. Elftman H. The transverse tarsal joint and its control. Clin Orthop 6. Inman VT. Special Article: Human locomotion. Can Med Assoc J
1960;16:41. 1966;94:1047–1054.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

◗ Commentary on the Six

Determinants of Gait
Dudley S. Childress and Steven A. Gard

The original material by Inman, Ralston and Todd in The basic concept proposed was that the first three
Chapter 1 is an overall perspective of walking, providing a determinants—pelvic rotation [first], pelvic list [second]
three-dimensional qualitative description of major angular and knee flexion in stance phase [third]—all presumably
displacements of the body during gait. The determinants act to flatten the trajectory of the body center of mass
that are described originate from two early papers asso- (BCOM) and thereby reduce the vertical translation of the
ciated with Inman, “Major Determinants in Normal and body during able-bodied walking. The thought behind this
Pathological Gait” (1) and “Human Locomotion” (2). The concept was that energy is saved if the vertical transla-
1953 paper “Major Determinants in Normal and Patholog- tion is reduced by each determinant. The fourth and fifth
ical Gait” is considered to be a landmark document in the determinants—foot mechanism [fourth] and ankle mech-
human gait field. The two papers, along with Chapter 1 (3), anism [fifth]—were claimed to smooth the trajectory of
are widely known and frequently referenced. For example, the BCOM, particularly where the trajectories of each step
the determinants are discussed extensively in Muscles, Re- intersect. The sixth determinant is basically the lateral dis-
flexes, and Locomotion (4) where McMahon presents them placement of the BCOM, which typically needs to be kept
as useful because of their simplicity and completeness. narrow for good ambulation. It should be noted that some
Whittle (5) suggests that the six determinants of gait com- authors (4,5,6) describe the action of the six determinants
bine together so that the trajectory of the COM is smooth in slightly different ways than Inman does (1,3).
and has a reasonably large radius of curvature, thereby re- Because of increased understanding of normal walking,
ducing energy expenditure (according to Inman). In the a number of investigators are convinced that several of the
third edition of his book, Whittle (6) acknowledges that original six determinants of gait probably serve functions
the determinants may be questioned. Perry (7) mentions other than those originally claimed. The purpose of the
the six determinants only briefly. six determinants of gait has been called into question by
The concept of the six determinants of gait was an orig- several studies that have critically evaluated their effect in
inal idea that has been appealing to clinicians, investi- normal walking (9–11). Pelvic rotation, the first determi-
gators and educators. Part of the broad general appeal nant, may be used to increase step length, especially at
may have been the beautifully rendered drawings asso- faster walking speeds, but it reportedly has little effect on
ciated with the early papers and with the book Human the vertical displacement of the body’s center of mass (12).
Walking. Inman and his colleagues are to be commended Similarly, pelvic obliquity and stance-phase knee flexion,
for proposing an innovative and forward-thinking theory the second and third determinants of gait, respectively, ap-
and for illustrating the concepts so well graphically. Un- pear to have virtually no effect on the vertical excursion of
fortunately, no empirical data supporting the theory were the body during able-bodied walking (9,10). These results
presented in the original paper, nor were the determi- have been supported by an investigation of Quesada and
nants ever objectively tested or evaluated. The accuracy of Rash (13). Inman’s own data (2) shows one reason why this
stance phase knee flexion in some of the attractive walk- is true: pelvic list and stance-phase knee flexion both occur
ing illustrations can be questioned. Anyway, at least in at the wrong time to have much influence in flattening the
the United States, the six determinants of gait became trajectory of the BCOM.
pervasive in clinical, research, and educational fields in- Vertical motion of the body’s center of mass appears
volved with gait. The six determinants have been taught to to be reduced by foot and ankle rocker mechanisms (14),
prosthetists, orthotists, physical therapists, medical stu- which serve to effectively lengthen the leg, and not by the
dents, kinesiologists, and other persons involved in anal- first three determinants as claimed. Inman et al. (3) ob-
ysis and study of human walking. Bowker (8) in the At- served that during single support the body appears to move
las of Limb Prosthetics illustrates how the determinants along the arc of a circle that has a radius about 2.2 times
have been explained for educational purposes through the longer than the length of the leg. Furthermore, they allude
years. to the important role of foot and ankle rocker mechanisms

19
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

20 Commentary on the Six Determinants of Gait

and their effect on flattening the trajectory of the knee joint

center.

ALTERNATIVE MODELS

Perry (7) has detailed the foot/ankle mechanisms, suggest-

ing that the three foot rocker mechanisms—heel, ankle,
and forefoot—can facilitate forward progression. It is be-
lieved that these three foot rocker mechanisms are inte-
grated during walking to create a single, smooth “roll-
over shape” for the foot/ankle system that facilitates pro-
gression of the center of pressure (COP) distally along
the plantar surface of the foot; this effectively lengthens
the leg and flattens the trajectory of the body center of
mass (14). Using a rocker-based inverted pendulum model
(Figure 1A-1), the vertical displacement of the body can be
accounted for by the geometrical constraints imposed by
the legs and the roll-over shape of the foot (Figure 1A-2).
The flatter trajectory of the BCOM is actually a section of
a prolate cycloid brought about by the circular rocker and FIGURE 1A-2 In a rocker-based inverted pendulum model, the
the attached mass. The results (14) show that the effective foot rocker (with radius r) effectively lengthens the anatomical leg
leg length is approximately 1.5 to 1.8 times as long as the (L) to create an effective leg length (Lv ) that flattens the vertical
displacement of the body (h) for a given step length (Sl ).
actual leg length, compared with the 2.2 value of Inman
(3). Inman was one of the first investigators to suggest the
tive over typical walking speeds, provides some results in
concept of effective leg length or virtual leg length.
closed mathematical form, yet is simple enough to be held
Hansen et al. (15) have shown that the human foot/ankle
together for thinking purposes. For example, it predicts
provides a roll-over shape that is approximately circular
that the vertical movement “h” is inversely proportional to
with a radius of about one-third of the anatomical leg
the virtual leg length and proportional to the step length
length. They have further shown that the roll-over shape
squared (14). It can also be shown that to an approxima-
of the physiologic foot-ankle tends to be relatively invari-
tion, “h” is linearly related to the speed of walking.
ant to changes in walking speed, heel height, and added
The heel-rise characteristic observed by Kerrigan et al.
weight.
(16) can be accounted for using the model of Figure 1A-
The model of Figure 1A-2 is only an approximation of
2 (see the position of the heel near toe-off). The model
reality (e.g. the legs are completely straight, the BCOM lo-
suggests that heel rise is a natural consequence of the ge-
cation is fixed, etc.). Nevertheless, it turns out to be predic-
ometry of the system and not causal. It is conjectured that
differences between the general results of Croce et al. (11)
and those of Gard and Childress (14) may be more related
to differences in language usage and interpretations than
to differences in fact.

CONCLUSIONS

Data suggest that the first three determinants do not re-

duce the vertical displacement of the BCOM during able-
bodied gait. The second and third determinants appear
to provide shock absorption. The shock absorption deter-
minants (second and third) seem likely to smooth the dis-
continuities associated with heel contact (create sinusoidal
trajectories). The fourth and fifth determinants are asso-
ciated with the foot and ankle complex, which is believed
to create a nearly circular rocker to assist with forward
FIGURE 1A-1 Conceptual walking model that incorporates progression and to create a BCOM trajectory that is a sec-
several of the functional aspects of able-bodied gait. tion of a prolate cycloid (the BCOM is outside the circle
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Commentary on the Six Determinants of Gait 21

of rotation of the rocker). The prolate cycloid trajectories 3. Inman, VT, Ralston, HJ, Todd, F. Human Walking. Williams and
appear to provide most of the observed flattening of the Wilkins, 1st edition, p. 6–14, 1981.
4. McMahon, TA. Muscles, Reflexes, and Locomotion. Princeton Univer-
trajectories, rather than the first three determinants. sity Press, p. 192–198, 1984.
It should be noted that the mechanisms of healthy hu- 5. Whittle, MW. Gait Analysis: An Introduction. 1st edition, Butterworth
man walking never try to make the trajectory of the BCOM Heinemann, p. 74–77, 1991.
6. Whittle, MW. Gait Analysis: An Introduction, 3rd edition, Butterworth
a straight line parallel to the walking surface in the direc- Heinemann, p. 77–81, 2002.
tion of progression. Humans do not roll like a ball. The 7. Perry, J. Gait Analysis: Normal and Pathological Function. Thorofare
human walking structure is somewhat similar to an in- N.J.: SLACK Inc., p. 40–42, 1992.
8. Bowker JH. Kinesiology and Functional Characteristics of the
verted pendulum, and walking is often described as falling Lower Limb. Chapter 18. In: American Academy of Orthopaedic
from one foot and leg and being caught by the opposite Surgeons, editors. The Atlas of Limb Prosthetics: Surgical and
foot and leg. Going up and down is not necessarily ineffi- Prosthetic Principles. St. Louis: C.V. Mosby Company, 1981;
261–71.
cient for a walker if the motion can be utilized for storing 9. Gard SA, Childress DS. The Effect of Pelvic List on the Vertical
and returning mechanical energy. As an example, it has Displacement of the Trunk During Normal Walking. Gait Posture.
been suggested (17) that walking is somewhat similar to 1997;5:233–8.
10. Gard SA, Childress DS. The Influence of Stance-Phase Knee Flexion
a hard boiled egg rolling end over end on a rigid surface. on the Vertical Displacement of the Trunk During Normal Walking.
The center of mass of the egg goes up and down as the egg Arch Phys Med Rehabil. 1999;80(1):26–32.
rolls but energy loss is quite small (theoretically zero). The 11. Croce UD, Riley PO, Lelas JL, Kerrigan DC. A Refined View of the
Determinants of Gait. Gait Posture 2001;14(2):79–84.
energy is exchanged between kinetic and potential forms. 12. Kerrigan DC, Riley PO, Lelas JL, Croce UD. Quantification of
There is a scientific aphorism that says, “It had been Pelvic Rotation as a Determinant of Gait. Arch Phys Med Rehabil
seen many times before but never observed.” Over the years 2001;82:217–20.
13. Quesada, PM, Rash GS. Simulation of Walking Without Stance Phase
since the paper of Saunders et al. (1) the six determinants Knee Flexion. Gait and Posture 1998;7(2):151, 152. [Abstracts of the
have been seen often but never closely observed. It appears 3rd Annual North American Society of Gait and Clinical Movement
that close observation of the six determinants over the last Analysis].
14. Gard SA, Childress DS. What Determines the Vertical Displace-
decade has been productive and that further study and rec- ment of the Body During Normal Walking? J Prosthet Orthot.
onciliation may continue to produce surprising findings. 2001;13(3):64–67.
15. Hansen AH, Childress DS, Knox E. Roll-over Shapes of Human Lo-
comotor Systems: Effects of Walking Speed. Clinical Biomechanics
2004;19(4):407–414.
REFERENCES FOR COMMENTARY 16. Kerrigan D, Croce, UD, Marciello M., Riley PO. A Refined View of
the Determinants of Gait: Significance of Heel Rise. Arch Phys Med
1. Saunders JB, Inman VT, Eberhart HD. The Major Determinants in Rehabil 2000;81:1077–1080.
Normal and Pathological Gait. JBJS 1953;35-A(3):543–58. 17. Margaria, R. Biomechanics and Energetics of Muscular Exercise, Ox-
2. Inman VT. Special Article: Human Locomotion. Can Med Assoc J ford University Press, Chapter 3, “Biomechanics of human locomo-
1966;94:1047–54. tion”, p. 86, 1976.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-01 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

22
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

• C h a p t e r 2
•
◗ The Evolution of Human Walking
Timothy D. Weaver and Richard G. Klein

The other chapters in this volume focus on how the hu- and relative to body mass, small canine teeth that do not
man musculoskeletal system is integrated to produce hu- differ substantially in size between the sexes, extended
man walking, which we will refer to as human bipedal- childhoods, long life-spans, and an advanced capacity for
ism. Our focus in this chapter is similar, except that we culture. Other authorities suggested that brain expansion
proceed from an evolutionary perspective, building on the came first and that it was then the catalyst for other dis-
fossil evidence for bipedalism, which spans more than tinctive traits, including bipedalism. However, fossil dis-
4 million years. Because the evolution of bipedalism can- coveries that are made almost yearly in Africa now abun-
not be understood in isolation, we examine its place within dantly demonstrate the temporal primacy of bipedalism.
the broader sweep of human evolution, and we consider The earliest well-known bipedal ancestors of 4 to 3 million
some of the theory behind the behavioral reconstructions years ago retained small, apelike brains and many other
we make from fossils. Our goal is to summarize both apelike characteristics of the head and upper body (46,85).
what we think occurred in human evolution and why we The spotty fossil record from before 4 million years ago
think so. implies that yet earlier humans were even more apelike
Bipedalism intrigues paleoanthropologists (that is, spe- (12,26–28,40,50,66,83). In short, the fossil record demon-
cialists in human evolution) not simply for the anatomy strates conclusively that bipedalism evolved millions of
that allows it, but also because it appears to be the years before large brains and other human specializations.
characteristic that first distinguished humans from apes. The credit for discovering that the first bipedal species
Long before there was a significant fossil record, Charles was otherwise apelike belongs to Raymond Dart, an
Darwin and his key ally, Thomas Huxley, proposed that anatomist at the University of the Witwatersrand Medi-
humans originated in Africa, because it was home to our cal School in Johannesburg, South Africa. In 1925, Dart
closest living relatives, the chimpanzees and the gorilla. described a child’s skull that had been found in 1924 in
The degree of genetic difference between humans and the ancient cave deposits at Taung, about 320 km southwest
African apes suggests that they last shared a common an- of Johannesburg (19). The skull showed that even in adult-
cestor between 8 and 5 million years ago, and the oldest hood the child’s brain would have been scarcely larger than
known human fossils now amply confirm Darwin’s and a chimpanzee’s, but the inferred orientation of the large
Huxley’s prescience that this ancestor lived in Africa. The basal aperture (foramen magnum) through which connec-
fossil record before 5 million years ago is too meager to tions pass from the brain to the spinal column suggested
clearly illuminate the split, but abundant human fossils to Dart that the skull had been balanced above an upright
that date from slightly before 4 million years ago and later spinal column in the typically human (bipedal) manner. He
confirm it. The fossils reveal a variety of human (that is, assigned the skull to the new genus and species, Australop-
bipedal) species that all appear to share a recent common ithecus africanus, and he concluded that it represented “an
ancestor (Fig. 2-1). Thus far, there is no basis for assum- extinct link between man and his simian ancestor.” From
ing that the bipedal adaptation arose more than once, al- the late 1930s onwards, when other fossils that broadly
though prudence demands that we leave the possibility resembled the “Taung child” were found in Africa, paleoan-
open, pending the accumulation of many more fossils from thropologists have commonly grouped them in the “aus-
before 4 million years ago. tralopithecines,” or “australopiths” for short. Most special-
The central place of bipedalism in human origins has ists now recognize at least two genera (Australopithecus
not always been apparent (79). When the fossil evidence and Paranthropus) and multiple species of australopiths,
was much more poorly known, some specialists believed and they believe that one of the known species, or per-
that bipedalism evolved concurrently with other human haps a yet-to-be-discovered close relative, gave rise to the
novelties, such as brains that are large both absolutely first member of our own genus, Homo (see Fig. 2-1). Some

23
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

24 Chapter 2 / The Evolution of Human Walking

FIGURE 2-1. A simplified phylogeny of hominins, illustrating the evolutionary relationships among
the species mentioned in the text.

australopith species persisted after Homo had emerged, relative brain size, large, forwardly projecting faces, rel-
but the last became extinct by 1 million years ago (see (36) atively long arms and short legs, curved finger and toe
for further discussion). bones, and at least in some species, a conically-shaped
Traditionally, the australopiths and Homo have been ribcage, which expanded conspicuously towards the waist.
lumped in the zoological family Hominidae (hominids in They contrasted with apes in the smaller size of their ca-
the vernacular), and the chimpanzees (two species), go- nine teeth and in the larger size and thicker enamel of
rilla, and orangutan have been assigned to a separate fam- their premolars and molars. However, their most strik-
ily, the Pongidae (or pongids). However, it is now clear that ing difference from apes was in the lower body. Aus-
the chimpanzees and gorilla are more closely related to hu- tralopith pelves, legs, and feet were all similarly shaped
mans than they are to the orangutan, and many authorities to those of later humans because they were constructed
thus prefer a scheme that includes the chimpanzees and for habitual bipedalism (46,85). The australopith mix
the gorilla in the Hominidae. In this scheme, which we ac- of apelike and human features is readily apparent in
cept here, the australopiths and Homo are separated from three-way comparisons of the skull and pelvis (Fig. 2-2):
great apes at the tribal level, as the Hominini or hominins. australopith skulls more closely resemble chimpanzee
The australopiths were similar to apes in many im- ones, but their pelves are conspicuously more like those
portant anatomic respects, including small absolute and of living humans. The striking features that distinguish
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 2 / The Evolution of Human Walking 25

Molecular genetics shows that humans, chimpanzees, and

gorillas are so closely related that depending on the ge-
netic region, chimpanzees may appear more like humans
or more like gorillas (49). However, the large majority of
regions link chimpanzees most closely to humans, and
it now appears that the gorilla line separated 8 to 6 mil-
lion years ago and that the chimpanzee and human lines
split 2 to 1 million years later (15). An estimate of roughly
6 million years for the human-chimpanzee split is consis-
tent with the geologic ages of the oldest putative hominin
species (12,27,28,64,66,75,84). We emphasize putative, be-
cause so far, there is no unambiguous skeletal evidence
that any 6 to 5 million-year-old species was bipedal. The
strongest case is for the species Orrorin tugenensis, which
among other parts is represented by a nearly complete
proximal femur with many features that imply bipedal-
ism. These include a femoral neck with a groove on the
rear for the obturator externus tendon, indicating full hip
extension, and cortical bone that is thicker on the infe-
rior margin than the superior one, indicating the action
of human-like hip abductors during the one-legged stance
phase of walking (26,50). The hominin status of the other
species depends mainly on tantalizing characters in their
skulls, jaws, or teeth, along with some ambiguous features
of the lower body (12,27,28). Even with a wider range of
FIGURE 2-2. Chimpanzee, australopith, and modern human fossils, however, it may be difficult to distinguish the ear-
skulls (left) and pelves (right) [pelves redrawn after (72)]. Note
that the australopith skull more closely resembles the chim- liest hominin from its ancestors and contemporaries, be-
panzee one, while the australopith pelvis looks more like the cause the anatomical differences were probably subtle.
modern human one. Before molecular genetics demonstrated the especially
close relationship between chimpanzees and humans,
anatomical analyses tended to link chimpanzees more
people from apes in the head and upper body began to closely to gorillas. The apparent contradiction between
appear only after 2 million years ago in the genus Homo. the data sets means either that chimpanzees and goril-
las evolved many detailed anatomical similarities inde-
pendently, or more likely, that the last shared ancestor
AN OVERVIEW OF HUMAN of chimpanzees, gorillas, and humans resembled chim-
EVOLUTION panzees and gorillas far more closely than it did humans.
Fossil ape species generally support this inference, be-
Our purpose here is to place the evolution of human walk- cause they tend to possess most of the features that unite
ing within the broader framework of human evolution. chimpanzees and gorillas to the exclusion of humans (78).
We proceed in rough chronological order, from the oldest Many of the striking anatomical similarities between chim-
known hominin species to the most recent. Our focus is on panzees and gorillas are thus probably primitive features
the anatomical evidence for bipedalism, but we also sum- that were lost in humans after the human and chimpanzee
marize change in other anatomical regions, particularly lineages diverged.
the brain, that must have impacted bipedalism, and we Like the living apes, the last shared ancestor of humans
touch on the natural selective forces, especially changes and chimpanzees was probably constructed for arm-over-
in climate, which probably drove large-scale anatomical arm climbing and under-branch hanging in the trees (56).
change. We conclude with a summary of some differences Its posture in the trees would thus have been orthograde
among living human populations in the anatomy for like that of bipedal humans, with the spinal column
walking. perpendicular to the ground. In this regard, it would have
In the introduction, we noted that genetic differences, contrasted sharply with tree-living monkeys, which tend
or more precisely, the inferred rate at which such differ- to walk quadrupedally on the surfaces of branches, with
ences accrued, place the last shared ancestor of people the spinal column parallel to the ground in the manner
and the African apes between 8 and 5 million years ago. of strictly terrestrial quadrupeds. On the ground, the last
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

26 Chapter 2 / The Evolution of Human Walking

shared ancestor would have been quadrupedal, and it tralopith legs may reflect their generally short stature, be-
could have walked the way that chimpanzees and gorillas cause in living humans, taller individuals tend to have rel-
do, with the forelimbs resting on the knuckles and the hind atively longer legs. Australopith males averaged less than
limbs on the soles of the feet (17,53,79). Chimpanzees and five feet tall, and females may have been much shorter
gorillas routinely move from place to place on the ground, (16,45). The most complete australopith skeletons are from
and “knuckle-walking” clearly meets their needs. So far, especially small individuals, and this may enhance the im-
there is no unequivocal evidence that hominins descend pression that australopiths were especially short legged
from a knuckle-walking ancestor, but if new fossil finds (24). Still, there is some unequivocal anatomical evidence
show that they did, hypotheses for the origins of bipedal- (see below) that the australopith gait was kinematically dif-
ism would have to explain why there was a shift from one ferent from that of later humans, and short legs may have
form of specialized terrestrial walking to another. On the made long-distance travel energetically expensive and/or
other hand, if future finds suggest that the last shared time consuming (35,71). Short legs would probably also
ancestor of chimpanzees and people probably spent little have forced the transition from walking to running to oc-
time on the ground, we would have to explain only why ho- cur at a fairly low velocity (37,38). These energetic, time,
minins adopted bipedalism rather than knuckle-walking and velocity constraints may mean that australopith de-
or some other form of terrestrial locomotion. pendence on bipedalism was more limited than for later
We will briefly address hypotheses for the origins of hominins or at least that their home ranges were smaller.
bipedalism near the end of this chapter. Here, we stress Arguably, long arms and other apelike upper body reten-
that the immediate impetus for bipedalism may always tions imply that the australopiths continued to rely on trees
be difficult to establish, but the effects of global climate for food and refuge and that they used bipedalism mainly
change on regional environments almost certainly provide to travel short distances between tree patches. Plant or an-
the ultimate cause. In the interval between 8 and 5 mil- imal fossils that accompany australopith fossils at some
lion years ago, when the human, chimpanzee, and gorilla sites show that trees often remained plentiful nearby (51).
lineages diverged, ice sheets expanded, sea level fell, at- Longer legs in an australopith species dated to around
mospheric levels of CO2 decreased, global temperatures 2.5 million years ago (7) may anticipate a shift to the
declined, and aridity increased at low latitudes (52). In more complete investment in bipedalism that indisputably
equatorial Africa, where the hominins emerged, once con- marks Homo after 2 million years ago.
tinuous forests became fragmented into islands separated Sparse fossils indicate that Homo emerged between
by stretches of woodland and grassland. Many mammal 2.5 and 2 million years ago, but it is well known only after
species that were adapted to tropical forests became ex- 2 million years ago. The oldest widely recognized species
tinct, and species that were better suited to more open, is Homo habilis, but variation in skull and tooth size in
grassier habitats burgeoned. Ape species were among the specimens dated to around 1.9 million years ago suggests
losers, and for the first time, monkey species came to that H. habilis could comprise two species—a smaller ver-
outnumber them, reflecting the greater ability of many sion that would be narrowly understood (H. habilis) and a
monkeys to exploit more open settings (14). Hominins pre- larger one for which the name H. rudolfensis has been pro-
sumably derive from an ape that likewise found a way posed. H. habilis/rudolfensis was unquestionably bipedal,
to benefit from more open vegetation. Contrasts between but the details of its bipedalism cannot be established, for
early hominin and chimpanzee teeth suggest that to begin lack of sufficiently complete limb bones that are directly
with, a major difference between the human and chim- associated with diagnostic skulls or jaws.
panzee lineages may have been dietary—early hominins One or both variants of Homo habilis is presumed
may have focused on hard, brittle food items found mainly to have produced the oldest known flaked stone tools,
on or near the ground, while their earliest ancestors, like dated to about 2.5 million years ago. The tools occur in
living chimpanzees, probably concentrated on relatively clusters that mark the oldest known archaeological sites
soft fruits found in trees (73). This difference in dietary and they are often accompanied by fragmentary animal
preference could have produced ecological and geographic bones that provide the oldest direct evidence for human
separation that promoted genetic separation and the for- carnivory (36).
mation of new species. Fresh fossil discoveries are required About 1.8 million years ago Homo habilis was sup-
to determine if this scenario is correct, and if so, to illumi- planted by the more advanced species Homo ergaster (also
nate it in greater detail. sometimes called early African Homo erectus). Specialists
By at least 4 million years ago, the australopiths had generally assume that H. ergaster evolved from H. habilis
evolved indisputable anatomical adaptations for bipedal- (or one of its constituents), but it may actually represent a
ism, but they also preserved numerous apelike character- parallel development that originated before 2 million years
istics that were lost in later people (70). The most conspic- ago. The more important point is that it is known not only
uous retained trait was a combination of relatively long from skulls, jaws, and isolated limb bones, but also from
arms and short legs. To some extent, the shortness of aus- the nearly complete skeleton of an 8- to 11-year-old boy
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 2 / The Evolution of Human Walking 27

spective, the small brain had a distinct advantage, because

full-term fetuses with smaller heads could pass more eas-
ily through the pelvic inlet, mid-plane, and outlet (birth
canal) and they were thus less likely to cause infant and/or
maternal mortality (57,58). Compared to pelvic shape in
living humans then, shape in H. ergaster was less subject
to tension between obstetrical requirements and the de-
mands of bipedalism, and it is thus possible that H. ergaster
was a more efficient biped. The human brain appears to
have remained absolutely and relatively small until about
600,000 years ago, after which it expanded rapidly to ap-
proximate the modern average (46,59). Brain expansion
probably affected the anatomy of the pelvis and femur, but
many more fossils are required to document the pattern
of change after H. ergaster and to determine if there were
implications for bipedalism.
By 500,000 years ago, Homo ergaster had given rise
to at least three geographically distinct evolutionary lin-
eages that culminated in three separate species: Homo ne-
anderthalensis (the Neandertals) in Europe, Homo sapiens
FIGURE 2-3. Differences in body size and proportions be- (modern humans and their immediate ancestors) in Africa,
tween Australopithecus afarensis and Homo ergaster [redrawn and Homo erectus (the relatively poorly known nonmodern
from (61)]. In both species, the brain was significantly smaller contemporaries of the Neandertals and early modern hu-
than in living humans. mans) in eastern Asia (Fig. 2-1). Beginning roughly 50,000
years ago, H. sapiens spread from Africa to replace the last
representatives of H. neanderthalensis and H. erectus. More
who died in northern Kenya about 1.6 million years ago advanced cultural behavior, perhaps grounded in a neu-
(76). Together with less complete specimens, the skeleton rologic change, underlay the spread (36), and there is no
shows that H. ergaster had lost the apelike features that ar- reason to suppose it was assisted by differences in human
guably facilitated tree climbing in the australopiths. In all walking. The skeletal anatomy of the Neandertals is partic-
major respects, including relatively short arms and long ularly well known, and despite some detailed differences
legs, stature and body mass approximating that of living in pelvis and femur shape, Neandertals probably walked
equatorial Africans, a barrel- (vs. cone-) shaped ribcage in exactly the same way as modern humans (80). However,
and relatively small premolars and molars, it contrasted Neandertal legs were shorter than those of their modern
with the australopiths in essentially the same way that liv- European successors (and of most living humans) (31),
ing humans do (Fig. 2-3) (16,44). The leg bones differ in and the Neandertals may thus have expended more energy
small details from those of living humans, but none of the per unit of walking distance (71).
differences indicate a significant difference in bipedalism. Large-scale or macroevolutionary change in the hu-
The change in thorax shape could imply a reduction in man form ceased after 40,000 to 30,000 years ago, when
gut size that was probably tied to a diet of higher qual- modern humans had established themselves everywhere,
ity foods (5). Either meat or subterranean tubers would fit but small-scale, or microevolutionary, change continued.
the bill, and even cooking could be implied, although the Some of the change reflects random genetic drift (chance
archaeological record provides no evidence for fire until changes in gene frequencies that are particularly likely to
after 500,000 years ago. The skeletal changes that mark occur in small populations), but some was adaptive and
H. ergaster, together with archaeological indications that reflects the natural selective effects of climate differences.
it was the first human species to occupy tree-poor en- Average body size and shape varies with latitude in most
vironments, suggest a transformation in thermoregula- widespread mammalian species and humans are no ex-
tion (including the loss of body hair to enhance sweating) ception (8,30,48,63,81). The reason is that differences in
and possibly the development of foraging patterns, group latitude correlate closely with differences in temperature
sizes, and social relations foreshadowing those of historic and humidity, and humans like other animals have evolved
hunter-gatherers (4,23,25). different body shapes in response. Thus, people whose an-
Homo ergaster contrasted most conspicuously with cestry traces to equatorial latitudes tend to have especially
living humans in its much smaller brain and in the low slim trunks with long limbs, while people who come from
ratio of brain size to body mass, in which it hardly dif- arctic or subarctic latitudes tend to have much stockier
fered from earlier species (46). From an obstetrical per- trunks with short limbs. Average male trunk breadth, as
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

28 Chapter 2 / The Evolution of Human Walking

reflected in hip breadth (the maximum distance between that they reflect actual behaviors. In contrast, primitive
the iliac crests of the pelvis), varies from 23 to 30 cm among features might reflect behaviors that are no longer being
living human populations (63), and the variation in aver- performed. Of course, the distinction between primitive
age leg length is similar. The narrow hipped, long-legged and derived features requires a knowledge of evolutionary
equatorial body form facilitates heat dissipation in set- history, and behavioral reconstructions that proceed with-
tings where this is obviously desirable, while the broader out such knowledge will always be questionable. In the
hipped, shorter legged arctic form retards heat loss in set- modern human case, the fossil record shows that adapta-
tings where heat retention is crucial. tions for tree climbing are primitive, while adaptations for
To the extent that climatic differences impact pelvic bipedalism are derived. Even in the absence of eyewitness
breadth and leg length, they also affect the precise im- accounts then, it could be concluded that bipedalism is
plementation of bipedalism in different populations. The important to us, while tree climbing might not be.
wide range in modern body form shows that the modern For the purpose of behavioral reconstruction, it is fur-
skeleton is not optimized for bipedalism, but represents ther important to distinguish between genetically deter-
a compromise between the demands of bipedalism and mined and behaviorally plastic skeletal traits. Plastic traits,
other constraints like climate and of course obstetrics. as we use the term, are ones that change in response to ac-
tivities during an individual’s lifetime, and they are there-
fore not meaningfully characterized as primitive or de-
THE THEORETICAL BASIS FOR rived. The simple presence of a plastic feature shows that
BEHAVIORAL RECONSTRUCTION a particular behavior was being performed (78).
The human bicondylar angle exemplifies a behav-
Reconstructing the locomotor behavior of an extinct iorally plastic skeletal feature that unambiguously implies
species is complicated by the possibility that some anatom- bipedalism, and a fossil femur is sufficient to determine if
ical features represent nonfunctional ancestral retentions. the angle was present. The bicondylar angle is defined as
Thus, in the previous section, we noted that the long arms the angle between the shaft of the femur and a line per-
and other apelike characteristics of the australopiths may pendicular to a plane surface (Fig. 2-3, left), when the fe-
indicate an apelike ability to climb trees, when they may mur is placed with the (distal) condyles flat against this
mean only that the australopiths were directly descended surface. In adult humans, the angle is usually around 9 to
from a tree-climbing ape. To illustrate how evolutionary 10 degrees, while in adult chimpanzees it is generally close
history or phylogeny can complicate behavioral interpre- to 0 degrees (meaning that chimpanzees lack a conspicu-
tations, we consider how we would reconstruct the loco- ous bicondylar angle). The substantial angle in humans
motor behavior of modern humans if all we knew were helps to center the body over one leg while the other is
skeletal anatomy. in swing phase, and without it, bipedal locomotion would
The modern human skeleton has many obvious be energetically expensive to employ over long distances.
anatomical adaptations for bipedal locomotion. The key However, human infants start out like adult chimpanzees
traits include a short pelvis that brings the sacrum very with no observable bicondylar angle, and the angle appears
close to the hip joints, laterally facing iliac blades, a com- only after children begin to walk. Biomechanical studies
plexly curved vertebral column, an angle between the fe- demonstrate that the adult human bicondylar angle is a
mur and tibia at the knee (the bicondylar angle), and a direct consequence of the forces that bipedalism imposes
foot that has a well-developed longitudinal arch with a big during development (67), and this makes it the best avail-
toe that is aligned parallel to the others. However, mod- able skeletal indicator of habitual bipedalism. With this
ern humans also have laterally facing and extremely mo- and our more general theoretical framework in mind, we
bile shoulder joints, which in apes are clearly adaptations now consider the fossils that document the evolution of
for arm-over-arm climbing and under-branch suspension. human walking.
Given anatomy alone then, there would be no reason to
suppose that modern humans were more committed to
bipedalism than to arm-swinging through the trees. We THE FOSSIL EVIDENCE FOR EARLY
know differently mainly because we can observe living hu- HUMAN WALKING
mans directly.
We might also avoid a false conclusion if we distin- The oldest-known hominin species for which bipedalism
guished anatomical features that are primitive (that is, has been unambiguously inferred is Australopithecus ana-
ones that were present in the last shared ancestor of a mensis, dated to between 4.2 and 3.9 million years ago in
species group) from ones that are derived or advanced northern Kenya (40). The fossil sample of this species in-
(that is, ones that developed after various species had di- cludes not only the usual jaws and teeth that tend to dom-
verged from a shared ancestor). Derived features imply a inate the fossil record, but also a fragmentary tibia that
shift in anatomy, and it is therefore reasonable to conclude preserves numerous indications of bipedalism. The most
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 2 / The Evolution of Human Walking 29

important indication is the approximate angle between the faces of the first sacral vertebra) is generally acute, but
shaft and the articular surface for the talus, or astragalus, in the best preserved sacrum of Australopithecus afaren-
at the distal (ankle) end (40,78). In chimpanzees, the shaft sis, it is closer to 90 degrees (2). This suggests that the
slopes away laterally from this surface, and the contrast vertebral column was oriented less vertically in A. afaren-
reflects a difference in the position of the knee, which lies sis, however, the promontory angle varies substantially in
directly above the ankle in humans and lateral to it in apes. living humans (78). An angle near 90 degrees has also been
In the last section, we stressed that the bicondylar angle is found in at least one Neandertal (80). Thus, on present evi-
the least ambiguous skeletal indicator of bipedalism in hu- dence, the promontory angle found in A. afarensis does not
mans, and the right angle between the tibia and the distal necessarily imply a different posture from later humans.
articular surface indirectly indicates it. The bipedal status A quasi-static equilibrium biomechanical model
of A. anamensis is thus secure. founded on observations of living humans and applied to
The fossil samples for most geologically younger aus- the partial skeleton of Australopithecus afarensis referred
tralopiths include a wider range of skeletal parts that bear to above suggests that by comparison to living humans,
upon walking, and they uniformly imply bipedalism. Per- A. afarensis should have had an enlarged femoral head
haps most importantly, the femora always exhibit a typi- and increased mediolateral buttressing of the femoral
cally human bicondylar angle, and the pelvis presents oft- shaft and ilium (62). It does not, and the implication
cited modifications for bipedal walking, including the posi- could be that A. afarensis walked with an elevated hip
tioning of the sacrum close to the hip joints and the lateral on the nonsupport side to minimize abductor and hip
orientation of the iliac blades (34,42,54). In these features joint reaction forces and mediolateral bending of the
and others, the australopiths were unmistakably derived femoral shaft of the stance leg. However, the fit between
away from the apes in the direction of living humans. predicted hip joint reaction force and femoral head size is
However, along with the derived features that demon- loose in living humans. Women, for example, have wide
strate bipedalism, the australopiths exhibited primitive biacetabular breadths and short femoral necks relative
traits that are usually associated with tree climbing and to men, but they do not have disproportionately large
also some unique traits that are unknown in either apes femoral heads (11).
or later humans. We noted previously that the traits The idea that australopiths retained an apelike ability
associated with climbing are often interpreted to mean to climb is more compelling than suggestions that their
that the australopiths depended on trees for food and bipedalism was posturally and kinematically unique. The
refuge. Some researchers have also proposed that these anatomical features that support special climbing abil-
traits and the ones that are unique to the australopiths ity include relatively long arms and short legs, many fea-
imply a posturally and kinematically unique form of tures of the hand and shoulder that reflect well developed
bipedalism (1–3,9,62,69,70). musculature for under-branch grasping and hanging, in-
The most widely discussed proposal stems from the as- dications of longer hamstring moment arms, and a deep
sessment of a partial skeleton and other bones of Aus- peroneal groove on the fibula. Anatomical details of the
tralopithecus afarensis, which lived in eastern Africa be- hip, knee, and ankle joints imply increased mobility (70).
tween roughly 3.9 and 2.9 million years ago. According to Arguably, the dramatic shift in body size and shape that
this analysis, individuals of A. afarensis lacked the mod- marked the emergence of Homo ergaster about 1.8 million
ern human ability to extend the hip and the knee, and they years ago is also pertinent. The shift is difficult to explain if
thus walked with a bent-hip, bent-knee gait (69,70). Critics the australopith locomotor repertoire differed little from
have disputed the anatomical features behind this conclu- that of later humans. Even if the long legs of H. ergaster
sion and some have also noted that walking with flexed were an adaptation only to longer range foraging and do
hip and knee joints would inhibit the pendular transfer not mean that tree climbing had become less important,
of kinetic and potential energy that typifies human walk- the reason that H. ergaster had lost so many other primitive
ing. Without this energy transfer, walking would be ener- features is inexplicable, unless H. ergaster signals a signifi-
getically far more expensive (13,18,77), and studies of liv- cant behavioral shift towards greater reliance on bipedal-
ing humans show that less hip and knee extension would ism and less dependence on tree climbing.
also increase the energetic costs of running (47). However, The problem remains that the supposed australop-
most primates already walk with a compliant gait (65), so ith tree-climbing adaptations are almost all primitive,
bipedalism employing partially flexed hip and knee joints and they could represent nonfunctional retentions inher-
may have been energetically no less efficient than the lo- ited from the last shared human and ape ancestor (39).
comotion of the last shared ancestor of humans and apes. Additionally, some of the anatomical differences from liv-
If this is accepted, then an energetically expensive gait in ing humans could reflect different obstetrical and vis-
australopiths could represent a primitive retention. ceral constraints, given that australopith brains were much
In living humans, the promontory angle of the sacrum smaller relative to body mass and that their intestines may
(that is, the angle between the ventral and superior sur- have been relatively larger (5,41,72). In the absence of
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

30 Chapter 2 / The Evolution of Human Walking

unequivocally derived adaptations or behaviorally plastic The various alternatives are difficult to falsify, and one
traits that indicate tree climbing, it may always be difficult or more of them could have operated in concert. How-
to show that tree climbing was a significant component of ever, it may be possible to narrow the choices somewhat by
australopith behavior. considering the developmental mechanisms that underlie
The locomotor skeleton of Homo ergaster closely resem- bipedalism. For example, as we noted previously, the adult
bled that of living humans, but it was not identical. Most human bicondylar angle develops only in response to the
notably, compared to the femora of living humans, those mechanical forces that bipedalism produces during devel-
of H. ergaster tended to have longer, anteroposteriorly flat- opment. Thus, the presence of a human bicondylar angle
ter necks with lower angles between the neck and the shaft implies that juveniles were habitually walking bipedally.
and more mediolaterally reinforced shafts with thicker cor- This means that a viable explanation for the origins of hu-
tical bone (22,74,76). There are no complete pelves, but man bipedalism must consider bipedal locomotion not just
the femoral characteristics may mean that the pelvis was in adults, but also in juveniles, and any explanation that
exceptionally broad from side to side and flat from front focuses tightly on adult behavior is unlikely.
to back (platypelloid). If so, this would be a feature that
H. ergaster retained from its australopith ancestry. Such
a pelvis may require long femoral necks to keep the hip SUMMARY
abductors in an advantageous position for opposing large
body weight moments during one-legged stance, and medi- Genetic evidence indicates that the human and chim-
olaterally thick femoral shafts may be required to resist panzee lineages diverged about 6 million years ago. It is
the large bending moments that result from long femoral widely assumed that the earliest humans walked bipedally,
necks and an extremely platypelloid pelvis (60). but the oldest indisputable fossil evidence for bipedalism
It was probably much later in human evolution, after is only slightly more than 4 million years old. Bipedalism
brain size had more closely approached the modern aver- is well documented in all human species after 4 million
age, that the pelvis achieved its present configuration. The years ago, including both the australopiths, which domi-
oldest evidence for such a pelvis comes from the Sima de nate the record before 2 million years ago and members of
los Huesos site in northern Spain, dated to about 400,000 the genus Homo, the earliest of which appeared between
years ago (6,10). The associated cranial remains indicate 2.5 and 2 million years ago. Anatomical differences be-
that the Sima people were on or near the evolutionary tween the australopiths and later Homo suggest that
line to the classic Neandertals. The Sima pelvis is large, the australopiths may have employed a kinematically
most likely that of a male, but it is not markedly broad and unique bipedal gait and that they remained proficient tree
flat, and its inlet exhibits a typically human shape. The im- climbers, but the case remains debatable.
plication is that once absolute and relative brain size ap- Homo ergaster, which appeared in eastern Africa about
proximated the values in living humans, basic pelvic form 1.8 million years ago, is the oldest known human species
was fixed, and there were probably no further anatomical to lack the apelike features that the australopiths retained
changes that significantly affected bipedalism. and also the oldest that unquestionably walked in the fully
modern way. Minor differences between H. ergaster and
later humans in locomotor anatomy may reflect the rela-
HYPOTHESES FOR THE ORIGINS tively small size of the H. ergaster brain, which placed more
OF BIPEDALISM limited constraints on the pelvic inlet (birth canal). Be-
tween 600,000 and 300,000 years ago, brain size expanded
In closing, we return to an issue that we deliberately skirted to near the modern average, and the pelvis was modi-
before—an explanation for why bipedalism evolved. The fied to its present form. This is essentially a compromise
initial selective advantages may have included; (a) the abil- between the demands of obstetrics and those of bipedal-
ity to carry meat or other food items to trees, other refuges, ism. After 50,000 years ago, when modern humans spread
or other group members (29,43); (b) an enhanced facility from Africa to predominately replace nonmodern humans
to gather small-diameter fruits from short trees (32); (c) in Eurasia, differences in climate selected for differences
a reduction in the amount of skin surface exposed to di- in body shape and proportions, which affected human
rect sunlight at midday, thus reducing the danger of heat locomotion in minor ways.
stress, particularly on the brain (82); (d) the freeing of the
hands for tool use (20), or for carrying the young long dis-
tances (68); (e) a decrease in the energy required to walk at ACKNOWLEDGMENTS
low speeds (85); (f) the ability to see more clearly or farther
during passage from place to place (21); (g) or the enhance- We are grateful to the curators who generously allowed
ment of threat displays that reduced violent conflicts over us to study comparative and fossil material under their
scarce resources (33). care, to David DeGusta and Teresa Steele for commenting
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

Chapter 2 / The Evolution of Human Walking 31

on drafts of this chapter, and to Kathryn Cruz-Uribe for 22. Day MH. Postcranial remains of Homo erectus from Bed IV, Olduvai
help with the illustrations. The L.S.B. Leakey Founda- Gorge, Tanzania. Nature 1971;232:383–387.
23. Foley RA. Another Unique Species: Patterns in Human Evolutionary
tion, the Morrison Institute for Population and Resource Ecology. Essex: Longman Scientific & Technical; 1987.
Studies at Stanford University, the A.W. Mellon Foun- 24. Franciscus RG, Holliday TW. Hindlimb skeletal allometry in Plio-
dation, the Department of Anthropological Sciences at Pleistocene hominids with special reference to AL-288–1 (“Lucy”).
Bulletins et Mémoires de la Société d’Anthropologie de Paris 1992;4(1–
Stanford University, and a Michael Guyer Postdoctoral 2):5–20.
Fellowship from the University of Wisconsin-Madison 25. Franciscus RG, Trinkaus E. Nasal morphology and the emergence of
provided financial support for research by TW reported Homo erectus. Am J Phys Anthropol 1988;75:517–527.
26. Galik K, Senut B, Pickford M, et al. External and internal morphology
in this chapter. of the BAR 1002’00 Orrorin tugenensis femur. Science 2004;305:1450–
1453.
27. Haile-Selassie Y, Suwa G, White TD. Late Miocene teeth from Mid-
dle Awash, Ethiopia, and early hominid dental evolution. Science
2004;303:1503–1505.
REFERENCES 28. Haile-Selassie Y. Late Miocene hominids from the Middle Awash,
Ethiopia. Nature 2001;412:178–181.
1. Abitbol MM. Evolution of the lumbosacral angle. Am J Phys Anthropol 29. Hewes GW. Food transport and the origin of hominid bipedalism.
1987;72:361–372. Amer Anthropol 1961;63:687–710.
2. Abitbol MM. Lateral view of Australopithecus afarensis: primitive as- 30. Holliday TW. Body proportions in Late Pleistocene Europe and mod-
pects of bipedal positional behavior in the earliest hominids. J Hum ern human origins. J Hum Evol 1997;32:423–447.
Evol 1995;28:211–229. 31. Holliday TW. Brachial and crural indices of European late Up-
3. Abitbol MM. Reconstruction of the STS 14 (Australopithecus per Paleolithic and Mesolithic humans. J Hum Evol 1999;36:549–
africanus) pelvis. Am J Phys Anthropol 1995;96:143–158. 566.
4. Aiello LC, Key C. Energetic consequences of being a Homo erectus 32. Hunt KD. Ecological morphology of Australopithecus afarensis: trav-
female. Am J Hum Biol 2002;14:551–565. eling terrestrially, eating arboreally. In: Strasser E, Fleagle JG,
5. Aiello LC, Wheeler P. The expensive-tissue hypothesis. Curr Anthropol McHenry HM, et al., eds. Primate Locomotion: Recent Advances.
1995;36:199–221. New York: Plenum Press; 1998:397–418.
6. Arsuaga J-L, Lorenzo C, Carretero J-M, Gracia A, Martínez I, García 33. Jablonski NG, Chaplin G. Origin of habitual terrestrial bipedalism in
N, et al. A complete human pelvis from the Middle Pleistocene of the ancestor of the Hominidae. J Hum Evol 1993;24:259–280.
Spain. Nature 1999;399:255–258. 34. Johanson DC, Lovejoy CO, Kimbel WH, et al. Morphology of the
7. Asfaw B, White TD, Lovejoy CO, Latimer BM, Simpson S, Suwa G. Pliocene partial hominid skeleton (A.L. 288-1) from the Hadar For-
Australopithecus garhi: a new species of early hominid from Ethiopia. mation, Ethiopia. Am J Phys Anthropol 1982;57:403–451.
Science 1999;284:629–636. 35. Jungers WL. Lucy’s limbs: skeletal allometry and locomotion in Aus-
8. Ashton KG, Tracy MC, de Queiroz A. Is Bergmann’s rule valid for tralopithecus afarensis. Nature 1982;297:676–678.
mammals? American Naturalist 2000;156:390–415. 36. Klein RG, Edgar B. The Dawn of Human Culture. New York: John
9. Berge C. How did the australopithecines walk? A biomechanical Wiley and Sons; 2002.
study of the hip and thigh of Australopithecus afarensis. J Hum Evol 37. Kramer PA, Eck GG. Locomotor energetics and leg length in hominid
1994;26:259–273. bipedality. J Hum Evol 2000;38:651–666.
10. Bischoff JL, Shamp DD, Aramburu A, et al. The Sima de los Huesos 38. Kramer PA. Modeling the locomotor energetics of extinct hominids.
hominids date to beyond U/Th equilibrium (>350 kyr) and perhaps to J Exp Biol 1999;202:2807–2818.
400–500 kyr: New radiometric dates. J Archaeol Sci 2003;30:275–280. 39. Latimer BM. Locomotor adaptations in Australopithecus afarensis:
11. Brinckmann P, Hoefert H, Jongen HT. Sex differences in the skeletal the issue of arboreality. In: Coppens Y, Senut B, eds. Origine(s) de la
geometry of the human pelvis and hip joint. J Biomech 1981;14:427– Bipédie chez les Hominidés. Paris: CNRS; 1991:169–176.
430. 40. Leakey MG, Feibel CS, McDougall I, Walker A. New four-million-
12. Brunet M, Guy F, Pilbeam D, et al. A new hominid from the Upper year-old hominid species from Kanapoi and Allia Bay, Kenya. Nature
Miocene of Chad, central Africa. Nature 2002;418:145–151. 1995;376:565–571.
13. Cavagna GA, Heglund NC, Taylor CR. Mechanical work in terrestrial 41. Lovejoy CO, Heiple KG, Burstein AH. The gait of Australopithecus.
locomotion: two basic mechanisms for minimizing energy expendi- Am J Phys Anthropol 1973;38(3):757–779.
ture. Am J Physiol 1977;233(5):R243–R261. 42. Lovejoy CO. Evolution of human walking. Sci Am 1988;259:118–125.
14. Cerling TE, Harris JM, MacFadden BJ, et al. Global vege- 43. Lovejoy CO. The origin of man. Science 1981;211:341–350.
tation change through the Miocene/Pliocene boundary. Nature 44. McHenry HM, Coffing K. Australopithecus to Homo: transformations
1997;389:153–158. in body and mind. Ann Rev Anthropol 2000;29:125–146.
15. Chen F-C, Li W-H. Genomic divergences between humans and other 45. McHenry HM. How big were the early hominids? Evol Anthro
hominoids and the effective population size of the common ancestor 1992;1:15–20.
of humans and chimpanzees. Am J Hum Genet 2001;68:444–456. 46. McHenry HM. Tempo and mode in human evolution. Proc Nat Acad
16. Collard M. Grades and transitions in human evolution. In: Crow TJ, Sci USA 1994;91:6780–6786.
ed. The Speciation of Modern Homo sapiens. Oxford: Oxford Univer- 47. McMahon TA, Valiant G, Frederick E. Groucho running. J Appl Phys-
sity Press; 2002:61–100. iol 1987;62(6):2326–2337.
17. Corruccini RS, McHenry HM. Knuckle-walking hominid ancestors. 48. Meiri S, Dayan T. On the validity of Bergmann’s rule. J Biogeograph
J Hum Evol 2001;40:507–511. 2003;30:331–351.
18. Crompton RH, Yu L, Weijie W, Günther M, Savage R. The mechanical 49. Pääbo S. The mosaic that is our genome. Nature 2003;42:409–412.
effectiveness of erect and “bent-hip, bent-knee” bipedal walking in 50. Pickford M, Senut B, Gommery D, Treil J. Bipedalism in Orrorin tu-
Australopithecus afarensis. J Hum Evol 1998;35:55–74. genensis revealed by its femora. Comptes Rendus Palevol 2002;1:191–
19. Dart RA. Australopithecus africanus: The man-ape of South Africa. 203.
Nature 1925;115:195–197. 51. Reed KE. Early hominid evolution and ecological change through
20. Darwin C. The Descent of Man and Selection in Relation to Sex. Lon- the African Plio-Pleistocene. J Hum Evol 1997;32:289–322.
don: John Murray and Sons; 1871. 52. Retallack GJ. Cenozoic expansion of grasslands and climatic cooling.
21. Day MH. Bipedalism: pressures, origins and modes. In: Wood J Geol 2001;109:407–426.
BA, Martin L, Andrews P, eds. Major Topics in Primate and Hu- 53. Richmond BG, Begun DR, Strait DS. Origin of human bipedalism:
man Evolution. Cambridge: Cambridge University Press; 1986:188– the knuckle-walking hypothesis revisited. Yearbook of Physical An-
202. thropology 2001;44:70–105.
P1: NDL/OVY P2: NDL/OVY QC: NDK/OVY T1: NDL
GRBT092-02 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:9

32 Chapter 2 / The Evolution of Human Walking

54. Robinson JT. Early Hominid Posture and Locomotion. Chicago: Uni- 71. Steudel-Numbers KL, Tilkens MJ. The effect of lower limb length on
versity of Chicago Press; 1972. the energetic cost of locomotion: implications for fossil hominins. J
55. Rodman PS, McHenry HM. Energetics and the origin of hominid Hum Evol 2004;47:95–109.
bipedalism. Am J Phys Anthropol 1980;52:103–106. 72. Tague RG, Lovejoy CO. The obstetric pelvis of A.L. 288-1 (Lucy). J
56. Rose MD. The process of bipedalization in hominids. In: Coppens Hum Evol 1986;15:237–255.
Y, Senut B, eds. Origine(s) de la Bipédie chez les Hominidés. Paris: 73. Teaford MF, Ungar PS. Diet and the evolution of the earliest human
CNRS; 1991:37–48. ancestors. Proc Natl Acad Sci USA 2000;97:13506–13511.
57. Rosenberg KR. Reply to Baskerville. Curr Anthropol 1989;30:486– 74. Trinkaus E. Femoral neck-shaft angles of the Qafzeh-Skhul early
488. modern humans, and activity levels among immature Near Eastern
58. Rosenberg KR. The evolution of modern human childbirth. Yearbook Paleolithic hominids. J Hum Evol 1993;25:393–416.
of Physical Anthropology 1992;35:89–124. 75. Vignaud P, Duringer P, Taïsso Mackaye H, et al. Geology and palaeon-
59. Ruff CB, Trinkaus E, Holliday TW. Body mass and encephalization tology of the Upper Miocene Toros-Menalla hominid locality, Chad.
in Pleistocene Homo. Nature 1997;387:173–176. Nature 2002;418:152–155.
60. Ruff CB. Biomechanics of the hip and birth in early Homo. Am J Phys 76. Walker A, Leakey R, eds. The Nariokotome Homo erectus Skeleton.
Anthropol 1995;98:527–574. Cambridge: Harvard University Press; 1993.
61. Ruff CB. Climatic adaptation in hominid evolution: the thermoregu- 77. Wang WJ, Crompton RH, Li Y, Gunther MM. Energy transformation
latory imperative. Evol Anthropol 1993;2:53–60. during erect and ‘bent-hip, bent-knee’ walking by humans with im-
62. Ruff CB. Evolution of the hominid hip. In: Strasser E, Fleagle JG, plications for the evolution of bipedalism. J Hum Evol 2003;44:563–
McHenry HM, et al., eds. Primate Locomotion: Recent Advances. New 579.
York: Plenum Press; 1998:449–469. 78. Ward CV. Interpreting the posture and locomotion of Australopithe-
63. Ruff CB. Morphological adaptation to climate in modern and fossil cus afarensis: where do we stand? Yearbook of Physical Anthropology
hominids. Yearbook of Physical Anthropology 1994;37:65–107. 2002;45:185–215.
64. Sawada Y, Pickford M, Senut B, et al. The age of Orrorin tugenensis, an 79. Washburn SL. Human evolution after Raymond Dart. In: Tobias PV,
early hominid from the Tugen Hills, Kenya. Comptes Rendus Palevol ed. Hominid Evolution Past, Present, and Future. New York: Alan R.
2002;1:293–303. Liss; 1985:3–18.
65. Schmitt D. Insights into the evolution of human bipedalism from 80. Weaver TD. A multi-causal functional analysis of hominid hip mor-
experimental studies of humans and other primates. J Exp Biol phology [Ph.D.]. Stanford: Stanford University; 2002.
2003;206:1437–1448. 81. Weaver TD. The shape of the Neandertal femur is primarily the
66. Senut B, Pickford M, Gommery D, et al. First hominid from the consequence of a hyperpolar body form. Proc Natl Acad Sci USA
Miocene (Lukeino formation, Kenya). Comptes Rendus de L’Academie 2003;100:6926–6929.
des Sciences. Serie II, Fascicule A-Sciences de la Terre et des Planetes 82. Wheeler PE. The thermoregulatory advantages of hominid bipedal-
2001;322(2):137–144. ism in open equatorial environments: the contribution of increased
67. Shefelbine SJ, Tardieu C, Carter DR. Development of the femoral convective heat loss and cutaneous evaporative cooling. J Hum Evol
bicondylar angle in hominid bipedalism. Bone 2002;30:765– 1991;21:107–115.
770. 83. White TD, Suwa G, Asfaw B. Australopithecus ramidus, a new species
68. Sinclair ARE, Leakey MD, Norton-Griffiths M. Migration and ho- of early hominid from Aramis, Ethiopia. Nature 1994;371:306–312.
minid bipedalism. Nature 1986;324:307–308. 84. WoldeGabriel G, Haile-Selassie Y, Renne PR, et al. Geology and pale-
69. Stern JT, Susman RL. The locomotor anatomy of Australopithecus ontology of the Late Miocene Middle Awash valley, Afar rift, Ethiopia.
afarensis. Am J Phys Anthropol 1983;60:279–317. Nature 2001;412:175–177.
70. Stern JT. Climbing to the top: a personal memoir of Australopithecus 85. Wood B, Richmond BG. Human evolution: taxonomy and paleobiol-
afarensis. Evol Anthropol 2000;9(3):113–133. ogy. J Anat 2000;196:19–60.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

• C h a p t e r 3
•
◗ Kinematics of Normal Human Walking
Kenton R. Kaufman and David H. Sutherland

OVERVIEW individual. The patterns of movements of the body during

free speed-walking are remarkably consistent between in-
A principle of science, and of medicine in particular, is dividuals and form the basis of the data to be presented in
that one must understand the normal or natural history of this chapter.
a studied phenomenon before attempting to describe and An important concept to understand is that smooth pas-
study the pathological or abnormal. It was just such desire sage of the common center of mass is essential for efficient
that launched the University of California San Francisco/ walking (13,27). Elevation of the center of mass begins
Berkeley Laboratory investigations into the kinematics of in early single limb support during the phase of forward
normal human walking. Through an unfortunate accident, deceleration. When the center of mass reaches its great-
Professor Howard D. Eberhart, an engineer, became an est height (approximately 30% of gait cycle), potential en-
amputee and was astounded that limb prosthetics had ergy is maximal and kinetic energy is minimal. This re-
never been adequately studied scientifically and was still ciprocal relationship between potential and kinetic energy
basically an art form. When he questioned Dr. Inman, he is retained, but the order is reversed during double limb
was invited to join a collaborative effort to study normal support, in which kinetic energy is maximal and poten-
human walking so that prosthetic research could benefit tial energy is at its lowest level. Muscle action is required,
from a solid background of engineering. A decade of in- and oxygen must be consumed both to initiate and sustain
tense, multidisciplinary investigation followed, much of walking. For the movements of the common center of mass
which is summarized in the first edition of Human Walking to be smooth, the movements of the lower extremities, and
(14). This effort inspired many other scientists and clini- to a lesser extent the movements of the upper extremities,
cians to further investigate and expand our basic knowl- must follow normal or near normal pathways. While our
edge of the kinematics of normal human walking. task in this chapter is to describe kinematics, not kinetics,
Human walking is a relatively unique form of ambu- the two are intertwined. However, accurate measurement
lation in that it is bipedal. While there are animal exam- of kinematics provides a foundation on which to build ki-
ples of bipedal locomotion (bears, primates, marsupials), netic analysis, which will be described in the next chapter.
the bipedal gait in man is uniquely efficient and func- In this chapter, we will provide a fundamental background
tional. Some researchers feel that it is the evolution of the of the engineering principles used to quantify human mo-
foot, allowing a stable platform for bipedal walking, which tion, describe the basic events of the gait cycle, explain how
allowed development of man’s hands as tool users. Hu- these events subdivide the phases and periods of the cycle,
man bipedal walking has been described by many authors and give an in-depth description of the normal average dy-
using simple analogies to make a complex phenomenon namic joint angle measurement of the pelvis and lower ex-
more understandable. The pattern of leg motions has been tremity body segments/joint in the three standard frames
likened to bicycling (2), to pendulum action, and to a con- of reference (sagittal, coronal, and transverse planes).
trolled fall. Lettre and Contini (21) have described human
and animal locomotion in three distinct stages:
PRINCIPLES OF MOTION
1) development stage (from rest to some velocity), MEASUREMENT
2) rhythmic stage (some constant average velocity), and
3) decay stage (coming back to rest) (Fig. 3-1). Motion analysis is the set of methods and techniques
Most research has concentrated on the rhythmic stage aimed at quantitative analysis of human movement. The
of free speed walking. Many observations have confirmed scientific foundation of gait analysis is based on rigid body
that the rhythmic stage of human walking is very consis- mechanics. Bioengineers have contributed the most in this
tent and is directly related to the optimal efficiency for an area. Nonetheless, it is important for all to have some basic

33
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

34 Chapter 3 / Kinematics of Normal Human Walking

FIGURE 3-1. Three phases of walking. (Adapted from

Lettre, Contini, 1967.)

understanding of the concepts that are shaping the disci- Using stereophotogrammetric principles, the planar pro-
pline of gait analysis. Confusion sometimes arises regard- jections of markers viewed by each camera are used to
ing the use of terms in the literature that describe the study reconstruct the 3-dimensional instantaneous position of
of motion. The study and analysis of humans in motion the markers relative to an inertially fixed laboratory coor-
and the forces acting upon humans in motion is known as dinate system. If the position of at least three noncolinear
biomechanics. The prefix “bio” is from the Greek, and it points fixed to the body segment can be obtained (and the
means “life.” The suffix “mechanics” is an area of Newto- body segment is assumed to be rigid) then the six degrees-
nian physics designed to study the effect of forces on bodies of-freedom associated with the position and orientation
and motion. Understanding the biomechanics of human of each segment can be obtained. Initially, a body-fixed
movement has both basic and applied value. There are a coordinate system is computed for each body segment
number of fundamental concepts and definitions that are (Figure 3-2b). For instance, consider the markers on the
required as a foundation for the understanding of human shank at an instant in time. A vector, STZ , can be formed
movement. from the lateral malleolus to the lateral knee marker. An-
There are two types of physical quantities: scalars and other vector can be formed from the lateral malleolus to
vectors. Scalars are those fundamental quantities that re- the marker on the shank wand. The vector cross-product
quire only a single number to specify them. Examples in- of these two vectors is a vector STX , which is perpen-
clude volume, mass, density, electric charge, and speed. dicular to the plane containing all three markers. The
Vectors require specification of both magnitude and direc- unit vector, STY , may be determined as the vector cross-
tion to completely characterize them. Examples include product of STZ and STX . Thus, the vectors STX , STY , and STZ
displacement, velocity, and acceleration. form an orthogonal body fixed coordinate system, called
Kinematics is the subdivision of mechanics that deals a technical coordinate system. In a similar manner, the
with the description of motion without regard to the forces marker based, or technical, coordinate system may be cal-
causing this motion. Kinematics describes motion in terms culated for the thigh, i.e., TTX , TTY , and TTZ . These seg-
of displacement, velocity, and acceleration in space. Kine- ments are linked and thus lack independence of move-
matic analysis relates further to the study of relative mo- ment. Hence their points of attachment, i.e., the joints, are
tion between rigid bodies and finds application in analysis the points of principal kinematic significance. Once the
of gait and other body movements where each limb seg- position of adjacent limb segments has been determined,
ment is considered a rigid body. The body segments are it is possible to determine the relative angle between ad-
usually defined as the HAT (head, arms, and trunk), pelvis, jacent limb segments in three dimensions. This assumes
thigh, shank, and foot. that the technical coordinate systems reasonably approx-
External markers are used to define orthogonal co- imate the anatomic axes of the body segments, e.g., TTZ
ordinate systems affixed to each body segment, whose approximates along the axis of the thigh and STZ approxi-
axes define the position of these body segments. With a mates the long axis of the shank. Alternatively, additional
camera-based system, either passively reflective or actively data can be collected that connects the technical coordi-
illuminated markers are used (Fig. 3-2a). These markers nate system to the underlying anatomic coordinate sys-
are commonly attached to the subjects as either discrete tem. This more rigorous approach adapts a subject calibra-
points or rigid clusters with multiple markers on each clus- tion procedure to relate the technical coordinate systems
ter. Placement of these external markers on the surface with pertinent anatomic landmarks (5). The subject cali-
of the body segments are aligned with particular bony bration is performed as a static trial with the subject stand-
landmarks. As the patient walks along a marked walk- ing. Additional markers are typically added to the medial
way, the cameras track and record the marker trajectories. femoral condyles and the medial malleoli during the static
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 35

FIGURE 3-2. Body-fixed reflective markers used for establishing anatomic coordinate systems. Video
camera motion measurement systems calculate the location of external markers placed on the body
segments and aligned with specific bony landmarks (Fig. 3-2a). A body-fixed external coordinate system
is computed from three or more markers on each body segment (Fig. 3-2b). Subsequently, a subject
calibration relates the external coordinate system with an anatomic coordinate system through the
identification of anatomic landmarks, e.g. the medial and lateral femoral condyles and medial and
lateral malleoli (Fig. 3-2c).

calibration trial. These markers serve as anatomic refer- (and each body segment is assumed to be rigid), it is possi-
ences for the knee axis and ankle axis. The hip center loca- ble to determine the relative angles between adjacent limb
tion is estimated from markers placed on the pelvis (31). segments in three dimensions.
The technical coordinate system is then transformed into These measurements are made in three dimensions
alignment with the anatomic coordinate system for each with reference to standard anatomic planes: sagittal, coro-
limb segment, e.g. SAX , SAY , SAZ (Figure 3-2c). The marker nal, and transverse, utilizing embedded coordinate sys-
system is coupled to a biomechanical model (15). Once the tems (7). Presently, there is a lack of uniformity in the
position of adjacent limb segments has been determined human locomotion literature on the specification of the
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

36 Chapter 3 / Kinematics of Normal Human Walking

local and global coordinate systems. This lack of uni- three general categories (1) which will be described in the
formity leads to ambiguities in the description of loco- next section.
motion performance. For example, one author may de-
scribe a certain motion or joint moment positive while
a second author may term the same quantity negative. ANALYTICAL DESCRIPTION OF JOINT
Standardization is needed. Currently, the International MOVEMENT
Society of Biomechanics is attempting to address this
issue. Planar Motion
Degrees-of-freedom is a term in mechanics that de- A hinge joint is the simplest but most common model used
scribes the ability of an object to move in space. The to represent an anatomic joint in planar motion about
number of degrees-of-freedom of a system is the number a single axis embedded in a fixed segment (i.e., a single
of independent coordinates that must be specified to define degree-of-freedom). In this case, points on the limb de-
the location and orientation of the object. A rigid body in scribe a circular path around a fixed center of rotation.
space without any constraint has six degrees-of-freedom. The limb undergoes simple angular motion. The center of
This means that three coordinates of a reference point rotation is the point that has zero velocity relative to all
on the body are needed to specify its location, and three points on the body that are rotating around it. Assume
angles with respect to a set of reference axes are needed to that from a motion study, the velocity (time and displace-
specify its orientation. In reality, all anatomic joints have ment) is known for the two points A and B on the shank
six degrees-of-freedom that consist of three rotations and (Fig. 3-3), which is behaving as a rigid body. Points A and
three translations about the coordinate axes. However, if B undergo angular displacement in space as shown by the
the motion of the joint is limited by anatomic constraints, two velocity vectors. At any instant there will be a center
the number of degrees-of-freedom is reduced. Depending of rotation for the knee. Surprisingly, it is easy to find the
on the anticipated application, various degrees of simpli- location of the center of rotation, which need not be on the
fication have been considered for kinematic modeling of body in question. Lines are drawn perpendicular to each
joints. The interphalangeal joints of the foot have been as- of the two velocity vectors. A velocity vector has no compo-
sumed to have one degree-of-freedom (flexion/extension). nent perpendicular to itself and therefore velocity is zero
The ankle, able to rotate about two axes, can be assumed along these lines. The point where the two zero velocity
to have two degrees-of-freedom (plantar/dorsiflexion, in/ components intersect is the instant center of rotation. By
eversion). The hip, able to rotate about three axes, has
three degrees-of-freedom (flexion/extension, abduction/
adduction, internal/external rotation).
Displacement is a change of position in space and may
be either linear or angular. Displacement represents the
motion that is measured. Displacements are vector quan-
tities requiring a definition of magnitude as well as direc-
tion. Generally, three-dimensional motion of a rigid body is
defined by six independent quantities, usually three trans-
lational and three rotational degrees-of-freedom. Trans-
lation, or linear displacement, is a special term that de-
scribes motion during which all points on a body or line
describe parallel lines as being either straight or curved.
The term, therefore, can apply to either linear or angular
motions where there is no tendency for rotation. Rotation,
or angular displacement, refers to an angular change in
position of a line in space (or a line drawn on a body) in
relation to some reference in which all points on a body
or line describe circular arcs about a fixed axis or point.
All limb segments undergo angular displacement during
human movement. Points on the limb may also undergo
linear displacement.
Given the wide range of possible types of motion,
care must be taken when describing or interpreting the
biomechanics of human movement. The type of movement
FIGURE 3-3. Determination of the instant center of rotation of
(linear or angular), the reference frame, as well as the di- the knee using the velocities of points A and B. The perpendic-
mensions of the movement (degrees-of-freedom), must be ulars to the velocity vectors of points A and B intersect at the
defined. The motion of a rigid body may be grouped into instant center of rotation.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 37

definition, it has zero velocity relative to the points rotating the movement. Otherwise, loci of the ICR or centrodes will
around it. result. In practice, determination of the ICR by the above
In general planar motion, the moving limb segment can described method is highly sensitive to error in locating
have both translation and rotation about the fixed segment. the points used to define the individual ICRs. The errors
Because of the translational component of motion, the cen- increase exponentially as the individual displacements are
ter of rotation or axis of rotation for the moving segment made smaller. Conversely, increasing the displacement in-
will change through the course of motion. At any point in crements decreases the error in determining the ICR (24).
time, an approximate center of rotation can be determined, However, with larger intervals, the true kinematics are not
which is defined as the instantaneous center of rotation accurately reproduced.
(ICR). As shown in the previous example, the velocity of a
point on a rigid body experiencing rotation must be per-
Rotational Three-Dimensional Motion
pendicular to a line joining the point and the center of ro-
tation. This specific property can be used to determine the Very commonly, when describing or modeling an anatomic
ICR graphically. In experimental measurements, however, joint, only the rotational motions are considered. A spher-
it is difficult to determine the velocity of different points on ical joint is commonly used to represent a joint that al-
a body in motion. An alternate method for approximating lows three degrees-of-freedom in rotation. Three angles
the ICR was described by Franz Reuleaux in 1876 (26). In are required in order to specify the relative position be-
this method, the instantaneous location of two points on tween the moving and fixed segments. For finite spatial ro-
the moving segment are identified from two consecutive tation, the sequence of rotation is extremely important and
positions within a short period of time, and the intersec- must be specified for a unique description of joint motion
tion of the bisectors of the line joining the same points (10). For the same amount of rotation, different final ori-
at the two positions defines the ICR (Fig. 3-4). For a true entations will result from different sequences of rotation
hinged motion, the ICR will be a fixed point throughout (Fig. 3-5). However, with proper selection and definition
of the axes of rotation between two bony segments, it is
possible to make the finite rotation sequence independent
or commutative (6,11).
The concept of Eulerian angles has been adopted in the
field of orthopedic biomechanics to unify the definition of
finite spatial rotation (6,11). In the selection of reference
axes, one axis is fixed to the stationary segment and an-
other axis is fixed to the moving segment (Fig. 3-6). In the
knee joint, for example, the flexion-extension angle ϕ oc-
curs about a mediolaterally directed axis defined by a line
connecting the medial and lateral femoral condyles. The
axial rotation angle ψ is measured about an axis defined
by the line along the shaft of the tibia. The third axis (also
defined as the floating axis) is orthogonal to the other two
axes and defines abduction/adduction θ . These rotations
match the Eulerian angle description and are thought to
be performed in such a way as to bring the moving segment
from the reference orientation into the current orientation.
 J, K ) is attached to a fixed seg-
If a unit vector triad ( I,
ment along the X, Y, Z axis and another triad (i, j , k ) is
fixed to the moving segment along the x, y, z, axes, the re-
lationship between them after any arbitrary finite rotation
can be expressed by a rotational matrix in terms of the
Eulerian angles φ, ψ, θ.
⎡ ⎤ ⎡ ⎤⎡ ⎤
i cψ· cφ + sψ· sθ · sφ sψ· cθ − cψ· sφ + sψ· sθ· cφ I
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ j ⎥ = ⎢−sψ· cφ + cψ· sθ · sφ cψ· cθ sψ· sφ + cψ· sθ· cφ ⎥ ⎢ ⎥
⎣ ⎦ ⎣ ⎦⎣ J ⎦
k cθ · sφ − sθ cθ· cφ K
FIGURE 3-4. Determination of the instant center of rotation
using Reuleaux’s method. The points A and B are displaced to A Where s and c stand for sine and cosine, respectively. The
and B , respectively, during the rotation of the moving segment.
Straight, connecting lines are drawn between A and A , and be- Eulerian angles can be calculated based on the known
tween B and B . Perpendicular bisectors of these displacement orientation of these unit vector triads attached to the prox-
lines intersect at the instant center of rotation. imal and distal body segments.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

38 Chapter 3 / Kinematics of Normal Human Walking

FIGURE 3-5. The sequence of rigid body ro-

tation is extremely important. For the same
amount of rotation different final orientations
will result from different sequences of rotation.

The advantage of using this system for description of problem can be resolved by selecting the proper axis of ro-
the spatial rotation of anatomic joints is that the angular tation or by using special mathematical manipulation (23).
rotations do not have to be referred back to the neutral
position of the joint because the rotation sequence can be
Generalized Six Degree-of-Freedom
independent, and the measurement can be easily related
Joint Motion
to anatomic structures (6,11). However, it is important to
recognize that two of the rotational axes in the system are For a more general unconstrained movement in 3-D space,
nonorthogonal when the joint departs from its neutral po- three translations and three rotations are required to de-
sition. Consequently, the system is difficult to use in ki- scribe the joint motion. The displacement of a rigid body
netic analysis. Angular velocity and acceleration have to be may take place along any one of an infinite number of
transformed into a set of inertial axes in terms of the Eule- paths. It is convenient to describe the displacement in
rian angles defined (17). Another disadvantage is the “gim- terms of the simplest motion that can produce it. The
bal lock” situation (6) when the first and third (embedded) most commonly used analytical method for the descrip-
rotation axes are parallel. In this situation, if the angles are tion of six degree-of-freedom displacement of a rigid body
inexact because of noisy measurements, strong nonorthog- is the screw displacement axis (SDA) (19,20,28,36). This is
onality will result in highly correlated, large errors Chaslés’ Theorem (8). The motion of the moving segment
in the rotational angles of the embedded axes (37). This from one position to another can be defined in terms of a

FIGURE 3-6. Description of knee joint motion

using the Eulerian angle system. The axis fixed
to the distal femur defines the flexion/extension
motion, ϕ. The axis fixed to the proximal tibia
along its anatomic axis defines internal-external
rotation, ψ. The floating axis is orthogonal to
the other two axes and is used to measure
abduction-adduction, θ.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 39

FIGURE 3-7. The screw displacement axis can

be used to describe general spatial motion. The
tibia moves from position 1 to position 2 by ro-
tation about the screw axis by an amount φ and
by translating along the screw axis, ρ, by an
amount τ . (Reproduced with permission from
Mayo Clinic.)

simultaneous rotation  around and a translation t along of translation along the SDA and the position vector of a
a unique axis, called a screw displacement axis, which is point on the SDA can be calculated.
located in the fixed segment (Fig. 3-7). Four additional There are several advantages of using the screw dis-
parameters are required to completely describe the dis- placement axis. First, the orientation of the screw axis
placement of the moving body. These additional parame- remains invariant, regardless of the reference coordinate
ters are the inclination and location of the SDA. The screw axis used. Second, the gimbal lock situation can be avoided
displacement axis is a true vector quantity. Its magnitude when using the SDA approach (30). Third, the SDA can
can be decomposed along any coordinate axes used for be decomposed into orthogonal axes, which may be bet-
the analysis. However, the amount of the finite screw ro- ter suited for decomposing other vectorial entities such
tation is not a vector quantity and the decomposition of as force and moment vectors (37). Unfortunately, just like
it must be carefully interpreted because of the noncom- determination of the center of rotation for planar motion,
mutative nature of finite rotation. Woltring (16) recom- determination of the screw displacement axis is highly sen-
mended that the component rotations (flexion/extension, sitive to measurement error as well. The ratio of error in-
abduction/adduction, endo-/exorotation) can be defined as creases exponentially with decreasing displacement (20).
components of the product Φ =  ρ, where ρ is the screw In practice, use of more than three reference points on
axis unit direction vector. the moving segment helps to minimize experimental error.
The orientation of the SDA and screw rotation is based Implementation of these procedures for defining a screw
on a rotational matrix used to define the location of ref- displacement axis description of human body movement
erence points with respect to a coordinate system after has been explored in the literature (28).
finite rotation. This rotational matrix can be expressed in
terms of directional cosines between the coordinate axes
before and after rotation or in terms of Eulerian angles. For GAIT EVENTS
the SDA, if the direction angles for the SDA are ux , u y , uz
letting  be the screw rotation, the rotation matrix [R] can A gait cycle is defined in terms of an interval of time dur-
also be derived (18). ing which one sequence of regularly recurring succession
 (u2 ver s + cos) (u u ver s − u sin) (u u ver s + u sin) of events is completed. During free speed (self selected),
x y z x z y
x
R = (ux uy ver s + uz sin) (u2y ver s + cos) (uy uz ver s + ux sin) walking a cycle of repeated events has been consistently ob-
(uz ux ver s − u y sin) (u y uz ver s + ux sin) (u2z ver s + cos) served. These events are simply 1) foot strike and 2) foot-off
(30). Since there are two extremities, there are four events:
Where vers  = 1 − cos. The screw rotation  may be
foot strike (FS), opposite foot-off (OFO), opposite foot strike
found from the trace of the matrix [R]
(OFS), and foot-off (FO). Typically, initial foot contact is at
 = cos−1 {(tr[R] − 1)/2} the heel, but in pathologic gait other areas of the foot may
strike first (i.e., toe walkers). Hence the term foot strike
With  known, the components for the orientation of the is used instead of heel strike. Similar reasoning applies to
screw axis can be calculated. From the translation vector, using the term foot-off rather than toe-off. The entire cycle
which can be obtained by comparing the position vectors repeats itself with the second foot strike (Fig. 3-8). The two
of the reference points on the moving body, the amount phases (stance and swing) and most of the periods of the
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

40 Chapter 3 / Kinematics of Normal Human Walking

FIGURE 3-8. Normal walk cycle

illustrating the events of gait.

basic gait cycle can be described in reference to these ba- off (62%). Swing phase starts a foot off (62%) and ends at
sic repeated events (30,34). Today, the commonly accepted second ipsilateral foot strike (100%) (see Fig. 3-8).
convention is to describe the cycle in terms of percentage, The major subdivisions of the cycle/phases (Table 3-1)
rather than the time elapsed, as we have observed that the describe the transitions that have to occur while the body’s
events occur in a remarkably similar sequence and are in- center of mass passes over the oscillating limbs. Stance
dependent of time, thus allowing normalization of the data phase is commonly divided into three periods: 1) initial
for multiple subjects. Therefore, initial foot strike is desig- double limb support (foot strike to opposite foot-off), 2) sin-
nated as 0% and the second ipsilateral foot strike as 100% gle limb support (opposite foot-off to opposite foot strike),
(0%–100%). Furthermore, in normal subjects, the oppo- and 3) second double limb support (opposite foot strike to
site limb repeats the same sequence of events, but is 180 foot-off). These periods are again defined by their respec-
degrees out of phase so that opposite foot strike = 50% tive events of gait. Swing phase can also be subdivided into
and the second opposite foot strike = 150% of the gait three periods: 1) initial swing (foot-off to foot clearance),
cycle. 2) mid swing (foot clearance to tibia vertical), and 3) ter-
The phases of the basic walking cycle are simple: the minal swing (tibia vertical to foot strike). It is important
stance phase is defined by the percentage of the cycle when to recognize that, from the standpoint of function, second
the foot is in contact with the ground and the swing phase double support (preswing) prepares the limb for swing.
by the time when the foot is in the air. The four basic events The average cycle consists of 62% stance phase and 38%
of gait define these phases (30). Stance phase comprises swing phase. Swing time is identical with the time of con-
the period between foot strike (0%) and ipsilateral foot- tralateral single limb stance.

◗ TABLE 3-1 Gait Cycle: Events, Periods, and Phases

Event % Gait Cycle Period Phase

Footstrike 0
Initial double limb support
Opposite foot-off 12 Stance, 62% of cycle
Single limb support
Opposite foot strike 50
Second double limb support
Foot-off 62
Initial swing
Foot clearance 75
Mid swing Swing, 38% of cycle
Tibia vertical 85
Terminal swing
Second foot strike 100
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 41

◗ TABLE 3-2 Gait Cycle: Periods and Functions

Period % Cycle Function Contralateral Limb

1. Initial double limb 0–12% Loading, weight transfer Unloading and preparing
support for swing (preswing)
2. Single limb support 12–50 Support of entire body weight; Swing
center of mass moving forward
3. Second double 50–62 Unloading and preparing for swing Loading, weight transfer
limb support (preswing)
4. Initial swing 62–75 Foot clearance Single limb support
5. Mid swing 75–85 Limb advances in front of body Single limb support
6. Terminal swing 85–100 Limb deceleration, preparation for Single limb support
weight transfer

Initial double limb support (Table 3-2) is characterized weight is transferred rapidly to the forward limb, the trail-
by a very rapid loading onto the forward limb with shock ing limb is ending its extension movement in preparation
absorption and slowing of the body’s forward momentum. to swing forward in front of the body. The preparation for
The foot usually progresses to foot flat, and the knee acts swing limb acceleration actually occurs during the period
as a shock absorber (see specific sagittal motion curves for of second double limb support. At this time, the leg must
more discussion). After opposite foot-off, the opposite leg be flexed at the knee and hip and plantar flexed at the ankle
is in swing, and the weight-bearing limb is in single limb to prepare for “toe off”.
stance. As the body passes over the fixed foot, the center of The subdivision of swing phase can best be understood
mass rises to its peak while both forward and vertical veloc- by comparing the leg to a compound pendulum. Such a
ity decrease. Forward shear then reverses to aft shear, the pendulum is able to change its period through the action
center of mass falls, and forward and vertical velocity in- of muscles and hence the cadence of walking (9,22). The
crease. This transition from fore to aft shear occurs around duration of swing is determined by the mass moment of
30% of the cycle (mid stance in normal subjects). It is very inertia of the parts (body segments) and their configu-
difficult to determine this transition point precisely in the ration in space (12). The critical event of foot clearance
walking cycle without a laboratory environment, while the occurs around 75% of the cycle when the swinging limb
other events are clearly observable. Once this peak in ele- passes the standing limb. The time when the tibia becomes
vation of the center of mass is achieved, the center of mass perpendicular to the floor heralds the beginning of limb
falls until the end of single limb stance at opposite foot deceleration.
strike (50% of the cycle). Second double limb support (op- Terminology has developed to describe the linear mea-
posite foot strike to foot-off) is also defined as preswing. As surement parameters of the gait cycle (Fig. 3-9) (34).

FIGURE 3-9. Linear distance measurements

of the gait cycle.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

42 Chapter 3 / Kinematics of Normal Human Walking

◗ TABLE 3-3 Age-Related Temporal-Distance Factors

Velocity Cadence Opposite Toe-Off Duration of Duration of Stride
Age (yr) (cm/sec) (steps/min) (% of cycle) Stance (%) Swing (%) Length (cm)

1 64 176 17.5 67.3 33.4 43.0

1.5 71 171 17.5 67.6 33.4 4935
2 72 156 16.7 66.8 34.2 54.9
2.5 81 156 153.5 65.7 35.3 61.8
3 86 154 15.5 65.5 35.5 66.8
4 100 152 14.2 63.7 37.3 77.9
5 108 154 13.4 63.5 37.5 84.3
6 109 146 13.4 63.6 37.4 89.3
7 114 143 12.4 62.4 38.6 96.5
20–25 153 115 9.0 60.0 41.0 158.8
30–35 145 111 11.0 61.0 39.0 156.9
40–45 159 122 11.0 61.0 39.0 155.9
50–55 155 118 10.0 60.0 40.0 157.9
60–65 147 115 11.0 61.0 39.0 153.0

Cadence is defined as the number of steps in a standard maturation. Initial heel strike, reciprocal arm/leg swing,
time frame (steps/minute). Step length is defined by the and initial knee flexion wave were present in the majority
distance (in centimeters) between the same point on each of normal 2-year-old children. Maturation of the dynamic
foot (usually the heel), during double limb support. Right joint angles was well-established between 3 and 4 years
step length refers to the distance from the right heel to the of age (4). Five important determinates of mature gait
left heel. The converse is true for left step length. Stride were: 1) duration of single limb stance, 2) walking speed,
length is defined by the distance (in centimeters) trav- 3) cadence, 4) step length, and 5) the ratio of pelvic
eled between two successive foot strikes of the same foot. spread to ankle spread. The ability to walk seemed to
Therefore, each stride is composed of one right and one depend primarily on motor control system maturation.
left step length (measured in centimeters). Walking speed is Myelination is an important element of this process (32).
the average speed attained after approximately three steps In children aged 1 to 7 years there was a linear relationship
(rhythmic stage) expressed in distance covered per unit of between step length and leg length (Fig. 3-10). There was
time (cm/sec or m/minute). also a linear relationship between age and walking speed;
It should be emphasized that these parameters of the however, the slope changed around 4 years of age due to a
human gait cycle are comparable only when limited to change in the rate of growth (Fig. 3-11). Even though the
free speed walking on level ground. The temporal dis- pattern of mature walking was well-established between 3
tance parameters vary with age (Table 3-3). Therefore, age- and 4 years of age, growth changes continue throughout
matched normal controls are required when comparing puberty. The body’s increase in stature continues to
with children or older adults. Introducing other variables influence the temporal/distance factors of step length,
such as fixed slow/fast walking velocities, ramps, stairs, walking speed (increases) and cadence (decreases). The
or even neurologic immaturity can change these relation- temporal/distance parameters stabilize by age 20 and
ships markedly and make comparison with normal data remain largely unchanged throughout most of adult life
difficult or impossible. (see Table 3-3 and Chapter 8 for more details).

GAIT MATURATION MOTION CURVES

An extensive study of gait maturation has been per- The gait kinematics described here are associated with a
formed (31,32) and is reviewed in Chapter 7. The group of twenty adult subjects (9 males and 11 females)
joint angle rotations of 415 children aged 1 to 7 years evaluated at Mayo Clinic. Subjects range in age from 20 to
were measured in the sagittal, coronal, and transverse 42 years and had a mean age of 30 years (± 8). Their mean
planes. When the joint angles were grouped by age, weight was 75 kg (± 17) and their mean height was 173 cm
they were remarkably similar during free speed walk- (± 11). These individuals had normal strength, full range
ing, although they showed evidence of age-related gait of motion of the lower extremities, no neurologic deficits,
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 43

FIGURE 3-10. Scatter plot showing linear re-

lationship between step length and leg length of
subjects aged 1–7 years. (From Sutherland DH,
Olshen RA, Biden EN, Wyatt MP. The develop-
ment of mature walking. London: MacKeith Press;
1988:57.)

no arthritis, and no previous major lower extremity joint SAGITTAL PLANE (SIDE VIEW)
surgery. The subjects walked at a self-selected speed (117
± 14 cm/sec). The user is referred to Kaufman et al. (16) The sagittal plane measurements are probably the most
for a description of the methods used in the data collec- commonly studied, best understood, and most accurately
tion. Description of the lower extremity motion is typically reproduced.
reported using Eulerian angles. The 3-D motion of the
pelvis, hip, knee, and angle is commonly available using Anterior Pelvic Tilt
current technology. In order to remember these curves, we The sagittal movements of the pelvis are controlled by
have found it useful to break down each curve into separate gravity, inertia, and the action of the hip flexor and ex-
recognizable segments and emphasize certain key points tensor muscles (Fig. 3-12). The role of the thoracopelvic
when teaching this material to other colleagues. muscles has produced speculation, but objective data

FIGURE 3-11. The lines in the center connect the 25th ,

50th , and 75th percentiles. The dots represent minimal
and maximal values. (From Sutherland DH, Olshen
RA, Biden EN, Wyatt MP. The development of mature
walking. London: MacKeith Press; 1988:57.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

44 Chapter 3 / Kinematics of Normal Human Walking

FIGURE 3-12. Sagittal movements of the pelvis (ante-

rior pelvic tilt represented by all values above the zero
axis).

relative to their contribution are not available. The cen- decelerated during late single limb stance (following heel
ter of gravity is within the pelvis when the body is in the rise).
anatomic (upright) position; however, dynamic postural
changes can alter the location of the center of mass. It KEY POINTS
is therefore necessary, under dynamic conditions, to cal-
culate the position of the center of mass rather than as-
✔ Pelvic tilt oscillates like body center of mass
cribe it to an anatomic location. The pelvis is inclined for- ✔ Flattest (least tilt at end of double limb support)
ward (anterior) or flexed and moves in a sinusoidal man-
ner with two peaks and two valleys. The pelvis is most
horizontal (least amount of tilt) at foot-off and opposite Hip Flexion/Extension
foot-off, with maximum flexion occurring in mid- to late The sagittal plane motion is very simple and can be seen
stance and terminal swing. The curve is very similar to the to be a single, sinusoidal curve (Fig. 3-14). One leg pro-
sagittal curve of the body center of mass (Figure 3-13) but gresses forward in order to advance the body while the
lags behind it. Forward pelvic tilt decreases during the pe- other remains behind to support the body. The hip is flexed
riod of loading (first and second double limb support) and at initial contact and extends until opposite foot contact.
the increases while the body’s center of mass passes over As soon as the opposite foot strikes the ground, weight is
the fixed foot (mid-stance and terminal swing). The tilt be- transferred to the forward limb, and the trailing leg begins
gins to flatten once again as the body center of mass is to flex at the knee and hip, while pivoting on the forefoot.

FIGURE 3-13. Sagittal curve of

the body center of mass. (From
Saunders JB, Enmanm VT, Eber-
hart HD.The major determinants
in normal and pathological gait.
J Bone Joint Surg 1953; 35A:544–
553.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 45

FIGURE 3-14. Normal hip flexion-extension. A pos-

itive value indicates hip flexion and a negative value
indicates hip extension.

Preswing is synonymous with second double limb support gresses anterior to the knee, creating an extension force
and describes this period. It is followed by slight extension and bringing the knee joint back into extension by mid
just before foot strike. The hip extensor muscles decelerate stance. It should be noted that this passive extension can-
the thigh and diminish the hip flexion in preparation for not occur without the strong eccentric contraction of the
weight acceptance. plantar flexors restraining the shank from progressive for-
ward rotation (29,33). The second flexion wave is necessary
in order to clear the foot in early swing phase. Flexion at
Knee Flexion/Extension
the knee actually precedes the onset of hip flexion at op-
The sagittal motion of the knee is known as the knee flex- posite foot strike. The knee is rapidly flexed beginning just
ion and extension curve (Fig. 3-15). This motion curve can after heel rise to a maximum in swing phase just as the
be described as two flexion waves, each starting in relative swinging foot passes the opposite limb. This flexion effec-
extension, progressing into flexion, then returning again tively shortens the limb allowing for foot clearance of the
to the starting point in extension. The first flexion wave, or swinging limb to prevent foot dragging. The knee joint is
stance phase knee flexion, acts as a shock absorber to aid then rapidly extended by a combination of inertial force
in weight acceptance. This curve peaks in early stance at of the compound inverted pendulum (thigh and shank),
opposite foot-off. The mechanical source for this shock ab- and the activity of the gluteus maximum, hamstrings, and
sorber is the eccentrically contracting quadriceps muscles, quadriceps muscles. Nearly full extension is achieved just
which are active until the ground reaction force line pro- prior to foot strike.

FIGURE 3-15. Normal knee flexion-extension. A pos-

itive value indicates knee flexion.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

46 Chapter 3 / Kinematics of Normal Human Walking

KEY POINTS opposite limb occurs very rapidly, and the plantar flexion
movement occurring after opposite foot strike is passive.
✔ First flexion wave: shock absorber This movement may be entirely due to gravity and inertia
✔ Second flexion wave: foot clearance or there may be passive tension in the plantar flexors after
cessation of the electromyographic signal.
The fourth segment is rapid ankle dorsiflexion. The tim-
Ankle Plantar Flexion/Dorsiflexion ing of this swing phase movement is found to coincide with
This is the most complex of the sagittal curves and can the maximum foot clearance effort and also with the sec-
be broken down into four separate functional segments ond knee flexion wave. Hence this segment is functionally
(Fig. 3-16). connected to foot clearance. The ankle is maintained in
This first segment occurs between foot strike and oppo- this neutral position by isometric contraction of the an-
site foot-off. The ankle is positioned at approximately neu- terior compartment muscles until foot strike when these
tral when foot strike occurs (normally heel first), and the same muscles are again needed to eccentrically restrain the
position of the ground reaction force posterior to the an- plantar flexion that repeats the first segment of the cycle.
kle center causes plantar flexion until foot flat is achieved,
prior to opposite foot-off. This portion of the ankle motion
KEY POINTS
curve is also known as the first rocker (25).
The second segment occurs during single limb stance ✔ Foot flat occurs in initial double support
(between the events of opposite foot-off and opposite foot ✔ Single limb stance contains, in sequence, progressive dor-
strike). It is typically convex superiorly and reflects the siflexion, and reversal of movement towards plantar flex-
body passing over the fixed flat foot (second rocker) (25). ion (due to eccentric then concentric action of the plantar
Toward the end of the single limb stance, at approximately flexors muscles)
40% of the cycle, the heel begins to rise as the plantar flex- ✔ Second double support (preswing), ends in foot-off and is
ors increase their force of contraction and act concentri- passive with respect to the plantar flexor muscles
cally (29,33). This is also known as the third rocker. At
opposite foot strike the ankle has lost some dorsiflexion
✔ Limb shortening to clear the foot begins at foot-off and
peaks during swing when the swinging ankle passes the
but has not returned to neutral.
supporting ankle
The third segment continues with opposite foot strike
and ends with foot-off. Rapid plantar flexion occurs to a
maximum of 20 to 25 degrees just as the foot is lifted off
CORONAL PLANE (FRONT VIEW)
the ground. Some individuals confuse this period with the
period of acceleration of the ankle (actually heel rise) due
Pelvic Obliquity
to the concentric action of the plantar flexors. This asso-
ciation would require simultaneous muscle activity and Pelvic obliquity can be viewed with respect to each lower
plantar flexion movement, but electromyography has con- extremity or with both together (Fig. 3-17). Although the
sistently shown that the plantar flexors are silent after op- same motion curve is associated with the movements
posite foot strike (29,33). The transfer of weight to the of both lower extremities, it is necessary to correlate pelvic

FIGURE 3-16. Normal ankle plantar-dorsiflexion. A

positive value indicates dorsiflexion and a negative
value indicates plantarflexion.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 47

FIGURE 3-17. Normal pelvic obliquity. A positive

value indicates that the pelvis is higher on the ipsilat-
eral side than on the contralateral side.

obliquity functionally with the movement of both sides KEY POINTS

of the pelvis. We find it easier to understand the motion
✔ Some drop of the contralateral pelvis during stance is
curve when we study the events and motion measurements
normal (in Trendelenburg gait the drop is excessive)
of one side at a time rather than with both together. The
key functional correlation is to remember that the motion ✔ Pelvic obliquity serves as a shock absorber and limb length
classically described by Trendelenburg (35) is normal. The adjustment
contralateral pelvis normally drops in early stance phase,
and a rise or positive curve is produced ipsilaterally, where HIP ABDUCTION/ADDUCTION
the measurement is referenced. In Trendelenburg gait the
contralateral pelvic drop is excessive. In normal subjects, In normal subjects, the hip abduction/adduction curve
the pelvic obliquity peak occurs just after opposite foot- (Fig. 3-18) is remarkably similar to the pelvic obliquity
off, corresponding with early single limb stance. There is a curve (Fig. 3-17). This makes perfect sense. Since the foot
small secondary rise during the acceleration period of sin- is basically fixed to the floor, there is very little coro-
gle stance phase and then the pattern inversely duplicates nal plane motion at the ankle or knee, and the pelvis is
itself, beginning at opposite foot strike as the opposite leg acting as a shock absorber. The hip abduction/adduction
repeats the same functional tasks. The rise in the ipsilat- wave gives the coronal plane dynamic alignment of the
eral pelvis functionally shortens the limb and acts like a thigh with respect to the pelvis. The hip, which is neu-
shock absorber for the body center of mass, decreasing tral at foot strike, rapidly adducts during first double sup-
the required vertical translation needed to pass over the port and rapidly abducts during second double support.
fixed foot. In swing phase, the hip adducts from near maximum

FIGURE 3-18. Normal hip abduction (AB)/adduction.

A positive value indicates hip adduction.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

48 Chapter 3 / Kinematics of Normal Human Walking

FIGURE 3-19. Normal pelvic rotation. A positive

value indicates internal pelvic rotation.

abduction at toe off to a neutral or perpendicular position minimum at opposite foot strike (50%). Initially, the pelvis
relative to the pelvis just before foot strike. There is no is internally rotated and it externally rotates until opposite
abduction motion during swing phase, which implies that foot strike when it begins to internally rotate again. Func-
there is no circumduction occurring in normal subjects. tionally, this curve is closely related to the sagittal plane hip
flexion and extension curve. Both pelvic rotation and hip
KEY POINTS flexion serve to effectively lengthen the limb and increase
stride length.
✔ Similar to pelvic obliquity in normal subject
✔ Adduction peaks at opposite foot-off
KEY POINTS
✔ Maximum abduction at toe-off
✔ Relative adduction in swing phase ✔ Single sinusoidal curve similar to hip flexion/extension
✔ Maximum internal rotation at foot strike
TRANSVERSE PLANE (AXIAL VIEW)
✔ Maximal external rotation at opposite foot strike

Pelvic Rotation Hip Rotation

This curve reports movement of the pelvis with respect Hip rotation describes the motion of the femur with re-
to the laboratory coordinates (Fig. 3-19). The motion is a spect to the pelvis (Fig. 3-20). Internal rotation of the hip
single sinusoidal curve with the peak at foot strike and a begins in late swing phase and continues through stance

FIGURE 3-20. Normal hip rotation. A positive value

indicates internal hip rotation.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 49

FIGURE 3-21. Normal knee rotation. A positive value

indicates internal knee rotation.

to opposite foot strike. The hip then externally rotates until (first rocker) and the foot transitions to foot flat, the hind
late swing, completing one sinusoidal motion curve. Hip foot and mid foot must pronate to accommodate the walk-
rotation changes direction from internal to external in late ing surface, producing a coupled internal torsion of the
stance phase. shank, reaching a maximum at opposite toe-off. The shank
begins to externally rotate slowly in early single limb sup-
KEY POINTS port and increases as the plantar flexors increase their ef-
fort, thereby causing heel rise. The hind foot then moves
✔ Sinusoidal curve into supination and stiffens the ankle joint, while the knee
✔ Peak internal rotation coincides with opposite foot strike internally rotates until foot-off.
✔ External rotation in swing
KEY POINTS
Knee Rotation ✔ Internal rotation during pronation
The shank is closely coupled to the fixed foot and the ✔ External rotation during supination
obliquely oriented mitre-hinge apparatus of the subtalar
joint (Fig. 3-21) (14). Pronation of the hind foot produces
Foot Progression Angle
obligatory internal rotation of the shank while supination
causes external rotation (14). At foot strike the tibia is in This measurement is made with respect to the laboratory
neutral rotation in relation to the laboratory coordinates, coordinate frame (Fig. 3-22). It describes the position of
and the foot (heel) is in supination. As loading occurs the foot with respect to the line of walking progression.

FIGURE 3-22. Normal foot progression. The negative

values indicate that the long axis of the foot is externally
rotated with respect to the line of progression.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

50 Chapter 3 / Kinematics of Normal Human Walking

This measurement is closely related to clinical observa- ACKNOWLEDGMENT

tions and is of great clinical importance. The first segment
of the curve shows a relatively constant amount of external We gratefully acknowledge Barbara Iverson-Literski for
rotation. During single limb support the foot externally ro- her careful preparation of this manuscript.
tates further as the opposite limb advances during swing.
Between the event of opposite foot strike and foot-off (the
period of second double support), there is an internal ro-
tation movement. Beginning at foot-off, the foot is rotated REFERENCES
externally as it simultaneously dorsiflexes to clear the foot 1. An KN, Chao EY. Kinematic analysis of human movement. Ann
during swing phase. In late swing the foot internally ro- Biomed Eng 1984;12:585–597.
tates as it prepares for foot contact. 2. Basmajian JV. The human bicycle: an ultimate biological conve-
nience. Orthop Clin North Am, 1976;7:1027–1029.
3. Bell AL, Pederson DR, Brand RA, Prediction of hip joint center loca-
tion from external landmarks. Hum Move Sci 1989;8:3–16.
KEY POINTS 4. Biafore S, Cottrell G, Focht LM, et al. Neural network analysis of
✔ Foot rotation remains in external rotation throughout entire gait dynamics. In: Transactions of the Orthopaedic Research Society;
1991.
gait cycle 5. Cappozzo A, Catani F, Croce UD, et al., Position and orientation in
✔ Further external rotation during single limb support as op- space of bones during movement: Anatomic frame definition and
determination. Clin Biomech 1995;10:171–178.
posite limb is advanced in swing 6. Chao EYS. Justification of triaxial goniometer for the measurement
✔ Small internal rotation during second double support (third of joint rotation. J Biomech 1980;13:989–1006.
7. Chao E. Justification of triaxial electrogoniometer for the measure-
rocker) ment of joint rotation. J Biomech 1980;13:987–1006.
✔ External rotation in initial swing 8. Chasles M. Bulletin de la Universite des Sciences (Ferussac).
1830;XIV:321.
✔ Internal rotation in terminal swing 9. Gage JR. An overview of normal walking. Instructional Course Lec-
tures 1990;39:291–303.
10. Goldstein H. Classical Mechanics. Reading, MA: Addison-Wesley;
The motion curves that we have described are in clini- 1959.
cal use in gait laboratories, and provide an important basis 11. Grood E, Suntay W. A joint coordinate system for the clinical de-
scription of three-dimensional motions: Applications to the knee.
for clinical assessment. They do not replace dynamic elec- J Biomech Eng 1983;105:136–143.
tromyographic or force studies. It is necessary to use all 12. Hicks R, Tashman S, Cary JM, et al. Swing phase control with knee
three types of measurements to arrive at a robust knowl- friction in juvenile amputees. J Orthop Res 1985;3:198–201.
13. Inman VT. The Joints of the Ankle. Baltimore: Williams and Wilkins,
edge of the causal factors in abnormal gait. 1976.
14. Inman VT. Ralston HJ, Todd F. Human Walking. Baltimore,
Williams & Wilkins; 1981.
15. Kadaba MP, Ramakrishnan HK, Wootten ME. Measurement of lower
SUMMARY extremity kinematics during level walking. J Orthop Res 1990;8:383–
392.
16. Kaufman K, Hughes C, Morrey B, et al. Gait characteristics of pa-
The study of movements in human walking has produced tients with knee osteoarthritis. J Biomech 2001;34:907–915.
rich information that lays the cornerstone for kinetic anal- 17. Kaufman KR, An KN, Chao EY. A comparison of intersegmental
ysis. Most of the early clinically oriented gait literature joint dynamics to isokinetic dynamometer measurements. J Biomech
1995;28:1243–1256.
emphasized sagittal motion; however, coronal plane and 18. Kinzel GL, Hall AS, Jr, Hillberry BM. Measurement of the to-
transverse plane movements are now being used with tal motion between two body segments. I. Analytical development.
equal importance as better analytic methods and improved J Biomech 1972;5:93–105.
19. Kinzel GL, Hillberry BM, Hall AS, Jr, et al. Measurement of the total
technology have evolved. Clinicians who provide treatment motion between two body segments. II. Description of application.
to patients with gait disorders expect high quality biome- J Biomech 1972;5:283–293.
chanical analyses to result from gait laboratory referrals. 20. Kinzel GL. On the design of instrumented linkages for the measure-
ment of relative motion between two rigid bodies, in Ph.D. Thesis.
If this expectation is to be met, accurate and meaningful West Lafayette, IN: Purdue University, 1973.
movement measurements must form the foundation for 21. Lettre C, Contini R. Accelerographic analysis of pathological gait. New
clinical analysis and treatment plans. All clinicians who York University School of Engineering and Science Technical Report
1368-01. New York, 1967.
treat patients with disorders of walking should become fa- 22. Mena D, Mansour JM, Simon SR. Analysis and synthesis of human
miliar with at least the rudiments of the scientific basis swing leg motion during gait and its clinical applications. J Biomech
of gait analysis. This is essential if they are to use appro- 1981;14:823–832.
23. Mital NK. Computation of rigid body rotation in three-dimensional
priate biomechanical principles in treatment planning and space from body-fixed acceleration measurements, in Ph.D. Thesis.
post-treatment evaluation. Gait laboratory physicians have Detroit: Wayne State University, 1978.
a particular responsibility for meeting this challenge be- 24. Panjabi MM. Centers and angles of rotation of body joints: a study
of errors and optimization. J Biomech 1979;12:911–920.
cause of their prominent role in the treatment of patients 25. Perry J. Gait Analysis: Normal and Pathological Function. Thorofare,
with musculoskeletal disorders. NJ: Slack; 1992.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

Chapter 3 / Kinematics of Normal Human Walking 51

26. Reuleaux F. The kinematics of machinery: Outline of a theory of ma- of the ankle in normal walking on the level. J Bone Joint Surg Am
chines. London; 1876. 1966;48:66–71.
27. Saunders JB, Inman VT, Eberhart HD. The major determinants 34. Sutherland DH. Gait Disorders in Childhood and Adolescence. Balti-
in normal and pathological gait. J Bone Joint Surg Am 1953:35A: more: Williams & Wilkins; 1984.
543. 35. Trendelenburg FV. Deutsche mechanische Wochenschrift. 1924;2:
28. Spoor C, Veldpaus F. Rigid body motion calculated from spatial co- 21–24.
ordinates of markers. J Biomech 1980;13:391–393. 36. Woltring HJ, Huiskes R, Lange DA, et al. Finite centroid and helical
29. Sutherland DH, Cooper L, Daniel D. The role of the ankle plan- axis estimation from noisy landmark measurements in the study of
tar flexors in normal walking. J Bone Joint Surg Am, 1980;62:354– human joint kinematics. J Biomech 1985;18:379–389.
363. 37. Woltring HJ. Analytical body-segment photogrammetry. In: Mod-
30. Sutherland DH, Cooper L. The events of gait. Bulletin of Prosthetic els, Connection with Experimental Apparatus and Relevant DSP Tech-
Research 1981;10–35:281–282. niques for Functional Movement Analysis. Dipartimento di Elettron-
31. Sutherland DH, Olshen R, Cooper L, et al. The development of mature ica ed Automatica: Universita di Ancona, 1990.
gait. J Bone Joint Surg Am 1980;62:336–353. 38. Woltring HJ. Representation and calculation of 3-D joint movement.
32. Sutherland DH, Olshen RA, Biden EN, et al. The Development of Ma- In: Leo T, Fioretti S, eds. Workshop on Assessment of Clinical Pro-
ture Walking. Oxford, England: Mac Keith Press; 1988. tocols. Internal Report Dipartimento di Elettronica ed Automatica:
33. Sutherland DH. An electromyographic study of the plantar flexors Universita degli Studi do Ancona, 1989.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-03 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:10

52
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

• C h a p t e r 4
•
◗ Kinetics of Normal Walking
Roy B. Davis and Kenton R. Kaufman

A number of clinical centers routinely use information plex three-dimensional (3-D) motion. Gait kinetic results,
derived from quantitative gait analysis in the decision- however, are inherently more difficult to understand be-
making process for the treatment of gait-related abnor- cause they cannot be directly observed. Moreover, while
malities (22,47,51,67,71). Kinematic data augmented by the focus of gait kinematics may be concentrated on vari-
electromyographic (EMG) tracings compliment clinical ables familiar to many readers, such as joint angles, gait
examination and observational gait findings (15). Increas- kinetic results involve perhaps less well-understood con-
ingly, the results of joint kinetic analyses, specifically in- cepts such as intersegmental moments, work, mechanical
tersegmental moments and powers, are also used in the energy, and power. The objectives of this chapter are there-
evaluation and assessment of normal (21,29,46,65,76) and fore to introduce and define kinetic concepts that are rele-
pathologic gait (18,22,34,48,59,73,74). For example, the vant to human locomotion and to present and describe the
evaluation of the relationship of intersegmental power underlying kinetics of the gait of normal ambulators (i.e.,
generation in persons with hemiplegia suggests that the without locomotor impairment).
noninvolved limb shows greater than normal power gener-
ation to compensate for the weaker involved limb (41,42).
In clinical gait analysis, an appreciation of compensatory GROUND REACTIONS
mechanisms and secondary abnormalities is often as valu-
able as an understanding of primary gait deviations. Other Standing quietly, a person’s weight (force of gravity) tends
clinical investigators have used joint kinetic parameters to pull the person down toward the ground (Figure 4-1).
to examine the biomechanical performance of orthoses That person does not move downward because the ground
in children with cerebral palsy and myelomeningocele is pushing upward with a total force equal in magnitude
(26,44,68). to the individual’s weight (Newton’s third law). If that in-
An understanding of knee joint loading is important in dividual’s body is relatively symmetric and the person is
subjects with angular deformities. In many studies, the standing without “leaning to one side,” then their weight
methods for determining joint loads have been based on is evenly shared by each lower extremity (Figure 4-1A).
static analysis when an individual stands on one limb. This The loads under each foot can be represented by one re-
posture is easily obtained from a standing radiograph. It sultant ground reaction force (GRF) that combines the
is implied that this represents single limb support during two loads, i.e., the resultant GRF magnitude equals body
gait. However, in reality, this is not the case. Several stud- weight. Analogously, a bathroom scale combines the load
ies have clearly demonstrated that there is no correlation under each foot when it reports body weight. With one half
between the angular deformity of the knee measured on ra- of body weight supported by each foot/leg, the point of ap-
diographs and the dynamic force distribution in the joint plication of the resultant GRF passes approximately mid-
during gait (24,28,70). Further, it has been shown that dy- way between the subject’s two feet (Figure 4-1B). In this
namic loading during gait is more closely correlated with case, there is no motion because the external forces applied
clinical outcome than static measurements (54,69). to the person, i.e., body weight and the GRF, are balanced,
Gait kinematics can be observed, to a degree, directly. that is, they are equal and opposite in magnitude. When the
Consequently, quantitative measurements can be com- person leans to one side (Figure 4-1C), the point of applica-
pared with observed motions to develop an understand- tion of the GRF shifts in that direction. Clearly, overall body
ing of the quantitative results. This does not suggest that posture can affect the GRF. Additional trunk lean would
understanding the relationship between observed patterns leave that person unstable, i.e., the downward weight vec-
and quantitative measurements is without challenge, par- tor (or weight line) would fall outside of the base of sup-
ticularly as it relates to pathologic gait with its often com- port. To remain in static equilibrium, the subject requires

53
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

54 Chapter 4 / Kinetics of Normal Walking

FIGURE 4-1. (A) Standing qui-

etly with body weight evenly dis-
tributed under each foot, the sub-
ject is in “static equilibrium.” (B)
The loads under each foot are
combined into a single ground
reaction force equal to the sum
of the forces under the two feet
and located approximately mid-
way between the feet. (C) As body
weight is shifted laterally, the
point of application of the ground
reaction force shifts as well. (D)
When the body weight vector is
shifted beyond the base of sup-
port, additional lateral forces are
required to maintain equilibrium.

additional support (Figure 4-1D). The horizontal force ap- zontal force passing through the center of the ball, then the
plied to the upper body by the wall is balanced with an ball accelerates horizontally due to the applied horizontal
equal and opposite horizontal GRF component. Note also force. Newton’s second law,
that the two equal and opposite vertical forces are no
F = ma
longer aligned (i.e., colinear). They form a “force cou-
ple” that would tend to rotate the body in a counterclock- where F represents the imbalanced force vector, m is the
wise direction. A second, clockwise force couple formed by mass of the object, and a is the translational (or linear)
the two equal and opposite horizontal forces balances the acceleration of the object (in the same direction as F), al-
counterclockwise force couple. Consequently, the subject lows a prediction of the motion that results from the force
is in both translational and rotational static equilibrium. application. Consequently, if the horizontal forces due to
A change in motion occurs when a net imbalance of air resistance and the friction between the ball and the icy
external forces acts on a rigid body or a system of rigid ground are assumed to be small, then the kicking force
bodies. Consider a ball that is resting on an icy surface will produce a horizontal acceleration of the ball that is
without motion, as shown in Figure 4-2A. Again, there proportional to the mass of the ball, or
is no motion because the upward ground reaction force,
Fkick = mball aball
FGRF , balances the downward weight (force) of the ball, W.
If the ball is kicked as shown in Figure 4-2B, with a hori- where Fkick represents the kick force, mball is the mass of
the ball, and aball is the acceleration of the ball. If the kick
is stronger (i.e., the magnitude of the force increased), then
the acceleration of the ball will be greater. Note also that
if the ball is sodden with water (i.e., its mass is increased),
then a greater kicking force would be required to produce
the same acceleration realized with a dry ball.
In mechanics, mass represents resistance to a change in
translational motion or a measure of the inertia of the ob-
ject. It is also the property of the object that gives rise to its
gravitational attraction. Taken together, the product of the
mass and acceleration of the object represent an inertial
FIGURE 4-2. Acceleration of a ball due to the force applied force. This inertial force and its associated acceleration are
during a kick. in opposite directions. For example, when a player kicks
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 55

the ball (Figure 4-2) by applying Fkick to the right, the ball greater than body weight (Figure 4-3D, Point C). This verti-
accelerates to the right. The ball simultaneously pushes cal force imbalance (i.e., vertical GRF minus body weight)
back (to the left) on the kicker’s foot with an inertial force accelerates the center of mass of the subject upward. Note
equal in magnitude to mball aball (and to Fkick ). Newton’s that the GRF is not directly applied to the body COM like
third law states this explicitly, i.e., the forces between inter- the kicking force was applied to the ball. The external force
acting bodies (e.g., the player’s foot and the ball) are equal imbalance is transmitted from the ground upward through
in magnitude, opposite in direction, and lie along the same the joints and ultimately applied to the trunk through the
line of action. hip joints. The muscle forces cause the legs to push down
The “center of mass” of an object is that point where on the ground and up on the trunk, simultaneously. Hu-
all of the mass of the object could be concentrated and mans walk using this same action–reaction relationship,
still have the same mechanical effect. In a uniform grav- that is, muscle forces continually modify the magnitude
itational field, the center of mass (COM) and the center and direction of the ground reaction loads to produce and
of gravity of an object are located at the same point. In a control ambulation.
multi-segmented object such as the human body, the lo- During walking, the body COM is accelerated upward
cation of the center of mass of the object is affected by and downward over the gait cycle (Figure 4-4). One might
the location of the center of mass of each of the segments. anticipate that the GRF would vary over the gait cycle be-
When the subject in Figure 4-1C leaned his upper body to cause of this upward and downward acceleration, as was
his right, it shifted the trunk COM, the head COM, and the seen in the previous jumping example. An examination of
right and left arm COM to his right. As a result, his body the ground reaction loads during walking presented in Fig-
COM shifted to his right as well. ure 4-5 demonstrates this relationship.
Note that the application of force causes an immediate The magnitude of the vertical GRF under each foot de-
acceleration of the object. Since acceleration is the rate of pends on whether the limb of interest is in double sup-
change in velocity port, single support, or swing phase. Consider a limb that
v is transitioning from swing phase to stance phase. Dur-
a= ing double support, when both feet are in contact with the
t
ground, the vertical GRF rapidly increases in magnitude
or
as the external load is transferred from one lower extrem-
v = at ity to the other (Figure 4-5, Point A). One might anticipate
that the magnitude of the vertical GRF be equal to the sub-
a change in velocity is not immediately produced. That is, ject’s weight during single support, with a constant value
it takes a change in time to result in a change in velocity. from approximately 12% to 50% of the gait cycle. In real-
Moreover, since velocity is the rate of change in displace- ity, however, the magnitude of the vertical GRF oscillates
ment above and below the value of the subject’s weight due to
s the upward and downward acceleration of the body COM.
v=
t That is, during the first part of single support, the body
or COM translates from its lowest to its highest elevation in
the gait cycle (Figure 4-4). Consequently, this upward ac-
s = vt
celeration coincides with a vertical GRF that is greater
a change in displacement is not immediately produced ei- than body weight (Figure 4-5, Point B). Before reaching
ther. Again, it takes time to result in a change in displace- its highest elevation at about 30% of the gait cycle (Figure
ment. 4-5, Point C), the upward velocity of the body COM begins
In this same way, external forces that are applied to to decrease. This vertical deceleration upward (or acceler-
the human body can also produce motion. For example, ation downward) coincides with a vertical GRF that is less
immediately before a vertical jump, the magnitude of the than body weight. From this highest elevation, the body
vertical ground reaction force equals the subject’s weight COM falls vertically, resulting in an increasing downward
(Figure 4-3A and Figure 4-3D, Point A). As the individual acceleration that coincides with a vertical GRF that con-
begins his downward motion to prepare to jump, i.e., ac- tinues to drop in magnitude. After reaching a relative mini-
celerating downward, the vertical GRF drops in magnitude mum, the vertical GRF increases in magnitude, reflecting a
to less than body weight (Figure 4-3B and Figure 4-3D, downward acceleration that is now decreasing (Figure 4-5,
Point B). With the generation of additional muscle force, Point D). The second rise in the vertical GRF above body
the downward motion is halted and the hips and knees weight, later in single support (Figure 4-5, Point E), co-
begin to extend and ankles begin to plantar flex (Figure incides with a second upward acceleration that slows
4-3C). The subject actively pushes down on the ground and controls the downward movement of the body COM.
and the ground pushes up on the subject with a GRF that is Finally, during second double support body weight is
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

56 Chapter 4 / Kinetics of Normal Walking

FIGURE 4-3. Vertical GRF changes during a vertical jump: (A) subject standing, (B) downward motion,
(C) jumping motion, and (D) trace of the vertical ground reaction force during this movement sequence.
The points A, B, and C in plot (D) correspond to illustrations (A), (B) and (C ).

transferred from one limb to the other and the GRF drops reaction forces reflect forward and backward accelerations
to zero (Figure 4-5, Point F). of the body. During the first double support interval, the
In addition to the vertical GRF, friction between the ground pushes backward on the lower extremity, produc-
foot and the ground produces anterior and posterior shear ing a backward acceleration that decelerates and controls
ground reaction forces (i.e., forces that are parallel to the forward movement. This posterior (Figure 4-5, Point
the ground). These anterior and posterior shear ground G) shear force reaches a maximum at the end of the first

FIGURE 4-4. Vertical displacement of the

body center of mass during ambulation. The
highest elevation occurs at midstance and
the lowest elevation occurs during double
support.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 57

FIGURE 4-5. Ground reaction loads (forces and vertical torque)

for one adult female walking at 71 m/min. Letters A through K
are identified in the text.

double support interval after which it deceases in magni- FIGURE 4-6. Schematic of a person in single support. The body
weight and vertical GRF force couple tend to rotate the body
tude. The relatively small shear force values in midstance toward the swing limb (clockwise). The medial shear GRF and
reflect relatively low backward or forward acceleration of the corresponding lateral inertial force balance this effect (i.e.,
the body during this brief time interval, i.e., the body COM produce a counterclockwise force couple) in order to achieve dy-
moves forward at approximately a constant speed. Later namic equilibrium.
in single support, muscle forces accelerate the body COM
forward as reflected in an anterior (Figure 4-5, Point H)
shear force. During second double support, muscle forces tion, thereby dynamically balancing the weight/GRF force
accelerate the stance limb forward into swing. These re- couple.
lationships between the external shear ground reaction In addition to the ground reaction forces, a vertical
forces and the associated accelerations of the body COM ground reaction torque is also applied to the foot during
hold true for single support. During first double support, gait. There is a tendency for the foot to twist about a verti-
however, a posterior shear force is applied to the leading cal axis while on the ground during normal locomotion.
limb while an anterior shear force is applied to the trailing This tendency is resisted or constrained by the friction
limb. Consequently, depending on the relative magnitudes between the foot and the ground. During the second half
of these two shear forces, the net imbalance in shear force of single support, for example, the advancing swing limb
might be quite small, reflecting little horizontal (forward and other upper body motions tend to rotate externally
or backward) acceleration of the body COM during double the stance limb foot. The vertical ground reaction torque
support. prevents this motion with a resistance that tends to rotate
A third GRF component, a medial shear force (Figure the foot internally (Figure 4-5, Point J). As the magnitude
4-5, Point I) applied to the stance limb, increases during of the vertical ground reaction torque falls off during the
double support to a relatively constant value for all of sin- second double support (Figure 4-5, Point K), the foot is
gle support. This relatively small medial GRF force pro- less constrained by the torque and externally rotates with
duces a medial acceleration of the body COM (along with respect to the direction of progression.
a corresponding lateral inertial force). This mechanism al- During normal ambulation, the foot lands on its heel
lows humans to walk with a step width. Since the vertical and lifts off from its toes. Consequently, the point of ap-
GRF component generally passes lateral to the body COM plication of the ground reaction load, referred to as the
during gait (Figure 4-6), the vertical GRF and the subject’s center of pressure (COP), moves from the heel forward to
weight create a force couple that tends to rotate the body the toes, as illustrated in Figure 4-7. The COP progresses
(i.e., the body COM would fall toward the swing limb). rapidly forward so that by midstance, the ground reac-
The medial shear force and the lateral inertial force, as a tion load is concentrated under the metatarsals. With the
force couple, tend to rotate the body in the opposite direc- COP positioned under the forefoot, the GRF can more
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

58 Chapter 4 / Kinetics of Normal Walking

position by the plantar flexor muscles (gastrocnemius and

soleus) in opposition to external force FGRF . That is, as-
sume that force FGRF is sufficiently strong to dorsiflex the
ankle. Consequently, the plantar flexor muscles must pro-
duce an internal muscle force that is sufficiently strong
to prevent that dorsiflexion. To explore this relationship
between this external force FGRF and the internal muscle
FIGURE 4-7. Center of pressure for one adult female walking at force, a “free-body-diagram” of the foot is created in which
71 m/min. Times during the gait cycle (i.e., 0%, 10%, 20%, etc.) the foot is separated from the leg to reveal the internal
are indicated on the plantar surface of the foot. muscle force, Fmuscle (Figure 4-8B). Note that the internal
force at the ankle joint, Fjoint , is added to the diagram to
represent the internal force that the tibia and fibula would
effectively stabilize the stance limb, as will be described apply to the foot to keep external force FGRF and internal
later. muscle force Fmuscle from translating the foot upward. For
the joint force Fjoint to effectively counter this influence by
INTERSEGMENTAL FORCES AND the external force FGRF and the muscle force Fmuscle , the
MOMENTS magnitude of Fjoint must be equivalent to the combined
magnitudes of FGRF and Fmuscle , expressed mathematically
Muscle forces produce and control joint motion. As was as
seen in the jumping example, muscle forces controlled the Fjoint = FGRF + Fmuscle
lower extremity joints as the subject moved downward and
then provided additional force to extend those joints in his The diagram is simplified further by representing the foot
jump upward. Those muscle forces were also channeled as a rigid lever with a fulcrum at the ankle (Figure 4-8C).
through the joints to allow the subject to “push down” on In this way, one can appreciate that the distances between
the ground and produce an upward acceleration of the GRF FGRF and muscle force Fmuscle and the fulcrum, d1
body COM. Muscles are able to both influence joint mo- and d2 , respectively, are important. Levers work because
tion and channel force through the joints because muscles these distances, called “lever arms,” increase the mechani-
cross those joints at some distance away from their cen- cal effectiveness or advantage of an applied force. The use
ters of rotation. These mechanisms are illustrated in the of common tools illustrates this point. To tighten a nut on
following simple example. a bolt, a turning force is more effectively applied at the end
Assume that a foot is loaded with an external vertical of the wrench away from the nut. To pull a nail, force is
GRF, FGRF , as shown in Figure 4-8A. Other external loads more effectively applied at the end of the hammer handle
such as the weight of the foot are neglected in this ex- away from the head of the hammer. The forces applied to a
ample. Further assume that the foot is held in this static lever produce moments about the fulcrum that reflect both

FIGURE 4-8. Static equilibrium of the lower leg. (A) An external force FGRF is resisted by plantar flexor
muscle force (shaded area). (B) The “free-body-diagram” of the foot illustrates the internal muscle and
joint forces, Fmuscle and Fjoint , respectively, which counterbalance the external force FGRF . The weight of
the foot is neglected. (C) The equivalent loading of the foot is illustrated as a lever with a fulcrum at the
ankle.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 59

the magnitude of the force and the distance between the during gait, thereby changing the mechanical advantage
force and fulcrum. Returning to the example (Figure 4-8C), of the internal muscle forces (63).
the external moment produced by the external force FGRF Muscle forces are generally not produced in isolation.
about the fulcrum is the product of the magnitude of FGRF For walking, Figure 4-9A depicts a more realistic scenario
and d1 . Similarly, the internal moment produced by the where multiple tendons carry internal muscle forces across
muscle force Fmuscle is the product of the magnitude of the ankle joint. Also shown are external loads that are ap-
Fmuscle and d2 . Note that Fjoint does not produce a moment plied to the foot, including the weight of the foot and a
about the fulcrum because it passes through the fulcrum, distributed force (or pressure) on the plantar surface of
i.e., it has no lever arm (also referred to as “moment arm”). the foot. Figure 4-9A also depicts a distributed force load-
Observe that FGRF tends to rotate the foot in a counter- ing on the ankle articular surface that reflects the joint
clockwise direction while Fmuscle tends to rotate the foot in contact force (also referred to as a bone-on-bone force)
a clockwise direction. Consequently, for the foot to remain that is transmitted from the tibia and fibula to the dome
in this position, the counterclockwise moment associated of the talus. Not shown is the distributed shear force un-
with FGRF is balanced by the clockwise moment associated der the foot. Ideally, a mechanical analysis would provide
with Fmuscle , or expressed mathematically, values for the several “unknown” muscle forces and the
“unknown” ankle joint loading if the kinematics of the foot
d1 × FGRF = d2 × Fmuscle and ground reaction loads are measured and the inertial
properties, e.g., mass, mass moment of inertia, of the foot
It can be observed from this relationship that if the ex- are estimated. This is a very challenging analytical prob-
ternal force FGRF increases, then the muscle force Fmuscle lem, however. Fundamentally, there are more unknowns
must increase as well (keeping d1 and d2 constant). Ob- in this scenario than there are equations to solve for the
serve that if both the external force FGRF and the internal unknowns. A good deal of work has been done in this area,
muscle force Fmuscle increase, then the internal joint force employing optimization strategies to predict values for the
Fjoint must increase as well. The converse is also true. That unknown muscle forces (10,12,31). While these techniques
is, if the muscle force Fmuscle changes, then both the exter- may provide some insight in normal locomotion, assump-
nal force FGRF and the joint force Fjoint change in response. tions required in these approaches limit their applicability
In this way, the muscle force influences both the position in pathological gait analysis.
of the joint and the force applied to the joint. Note also More commonly in clinical gait analysis, several muscle
that this relationship is directly related to the magnitudes forces and the distributed load on the ankle are expressed
of the moment arms. For example, if the magnitude of the as a single concentrated intersegmental force vector, FA ,
external force FGRF and the muscle moment arm d2 are and moments produced by the muscle forces (and any
held constant, then the muscle force Fmuscle must increase other structures that cross the joint) are represented as
(or decrease) if d1 is increased (or decreased). One can an- a single net intersegmental moment vector, MA (Figure 4-
ticipate from the previous discussion pertaining to ground 9B). The distributed loads on the plantar surface of the foot
reactions during gait that muscle force magnitudes will are presented as a resultant ground reaction force, FGRF ,
vary to produce changes in the GRF magnitude and COP and torque, TGR , both applied at the COP. The kinematics
location as well as in response to that external loading. The of the foot and ground reaction loads are measured and the
muscle moment arm values also vary with joint position inertial properties, e.g., mass and mass moment of inertia,

FIGURE 4-9. Schematic of the lower leg during

gait. (A) Some of the forces (concentrated and dis-
tributed) that would be applied to the foot during the
stance phase of gait are illustrated. (B) A “free-body-
diagram” of the foot showing the ankle intersegmen-
tal force and moment, weight of the foot, and ground
reactions associated with a three-dimensional in-
verse dynamics analysis.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

60 Chapter 4 / Kinetics of Normal Walking

of the foot are estimated. Newton’s translational equation two inertial properties of the object: the mass of the object
of motion, and the distribution of mass within the object. Commonly
F=m a found objects illustrate this relationship. For example, a
hammer offers a different resistance to rotation depend-
where  F represents the sum of the forces applied to the ing on whether it is held and rotated about its handle or its
body segment, m is the mass of the body segment, and a is head. The mass of the hammer does not change in the two
the linear acceleration vector of the body segment COM, scenarios, but the distribution of that mass about the two
allows for the solution of unknown FA . Newton’s rotational reference axes does change. Consequently, with respect to
equation of motion, mass moment of inertia, the choice of reference axes is
Σ Mg = Ig α important.
where  Mg represents the sum of the moments of force Inertial properties of body segments are often estimated
acting about the body segment COM, Ig is the centroidal from regression models based on human cadaveric stud-
mass moment of inertia of the body segment, and α is the ies (8,11,19). For example, the locations of the COM of
angular acceleration of the body segment, is employed to the thigh, shank, and foot might be expressed as a func-
solve for the unknown ankle moment, MA . tion of segment length, e.g., the shank COM is distal to
This kinetic analysis begins with the foot to solve for the the knee center by approximately 43% of the shank length
intersegmental reactions at the ankle and then moves up (19). Alternatively, other investigators (27,72) map the sur-
from the ground to the hip. For the foot, the translational face geometry of the limb segment and then divide the seg-
equation of motion becomes ment into elliptical slices of constant density. The inertial
properties of the entire segment are then determined by
 F = FA + FGRF + Wfoot = mfoot ag combining the contributions of the individual elements.
where mfoot is the mass of the foot and ag is the linear Still other investigators approximate entire segments as
acceleration of the foot COM. Consequently, the unknown simple geometric solids, e.g., ellipsoids, cylinders, of con-
ankle intersegmental force can be expressed as stant density, such as the Hanavan (23) 15-segment model.
The time required to quantify the surface geometry of indi-
FA = mfoot ag − FGRF − mfoot g vidual segments from photographic images or numerous
where the weight of the foot, Wfoot , is represented by the anatomic measurements, e.g., over 200 in the Hatze (25)
product of the mass of the foot and the gravitational con- detailed model, limits the feasibility of some of these latter
stant g. The expanded rotational equation of motion be- approaches in the clinical setting. Cappozzo and Berme (6)
comes conclude that when extreme accuracy is desired, the Hatze
model appears to offer the best current approach. These
 Mg = MA + r1 × FA + r2 × FGRF + TGR = Ifoot αfoot authors advise that if a statistical model is used, then the
where r1 and r2 are position vectors (not shown in Figure test subject should be consistent with the body composi-
4-9B) that describe the location of the ankle joint center tion and gender of the study population employed to de-
and the COP relative to the foot COM, respectively, TGR is velop the regression relationships. For example, the work
the vertical ground reaction torque, Ifoot is the centroidal of Dempster (19) is based on eight male subjects ranging
mass moment of inertia of the foot, and αfoot is the an- in age from 52 to 83 years. Note that these techniques pro-
gular acceleration of the foot. The moment that force FA vide estimates of the “principal” mass moments of inertia
produces about the foot COM is the vector cross product about principal axes passing through the segment COM.
of r1 and FA . Similarly, the moment that force FGRF pro- A principal axis is based on the symmetric distribution
duces about the foot COM is the vector cross product of of the mass of the segment about that axis. The principal
r2 and FGRF . The unknown ankle reaction moment can be axes of the thigh, for example, might be approximated by
expressed as a longitudinal axis from the knee to hip center and two
transverse axes that are perpendicular to each other and
MA = Ifoot αfoot − r1 × FA − r2 × FGRF − TGR the longitudinal axis. Generally, the anatomically aligned
Mass moment of inertia was referenced in Newton’s ro- coordinate systems constructed to compute the kinemat-
tational equation of motion. Just as the mass of an object ics of the segment approximate the principal axes of the
represents the resistance to a change in translational mo- segment as well.
tion, the mass moment of inertia represents the resistance For greater accuracy, a technique, known variously
to a change in rotational motion. Similarly, as mass is a as stereophotometrics, biostereometrics or stereopho-
measure of the translational inertia of a segment, mass togrammetry, has also been used as a means of mathemati-
moment of inertia is associated with the rotational inertia cal segmentation to obtain mass distribution properties of
of a segment. The product of the mass moment of iner- body segments in a living subject. This technique involves
tia and the angular acceleration may be thought of as an three-dimensional photography of the subject by cameras
inertial moment. The mass moment of inertia combines placed at strategic locations. The coordinates of a number
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 61

of points on the body serve as input to a digital computer shank. The same process is then repeated to solve for the
that can then recreate the geometric subtlety of the body hip intersegmental force and moment.
segments. This technique is used to determine the mass The mechanical power associated with joint kinematics
distribution properties of the body segments, which can and intersegmental moments can be computed from the
be calculated from anthropometric dimensions using re- combination of the intersegmental moment and the joint
gression equations (36). Detailed analysis has shown that angular velocity, i.e.,
the principal axes are not aligned with the anatomic axes.
PJ = MJ · ω J
Nonetheless, a cosine matrix of the principal axes with re-
spect to the anatomic axes is available (36). where PJ is the intersegmental power, MJ is the net in-
To summarize, the unknown intersegmental force and tersegmental moment, and ω J is the joint angular veloc-
moment at the ankle, FA and MA , respectively, can be ity at joint J. Intersegmental power represents the rate at
solved through the application of Newton’s translational which work is done by or on the intersegmental moment
and rotational equations of motion. A number of mea- in producing or controlling joint rotational displacement.
surements are required for this computational process, Intersegmental power “generation” (positive intersegmen-
including: tal power) may be related to concentric muscle contrac-
tion and intersegmental power “absorption” (negative in-
■ The ground reactions, FGRF and TGR , and COP location, tersegmental power) may be related to eccentric muscle
measured using instrumented force platforms, and contraction. However, these relationships are not always
■ The location of the ankle joint center (and other valid. Intersegmental power absorption may also be asso-
anatomic reference points on the foot), measured with ciated with the passive elongation of muscles and other
laboratory-based motion measurement technology, e.g., soft tissue structures that cross the joint.
motion data capture cameras (14,15). Intersegmental power is a scalar quantity that reflects
the sum of the products of the vector components of the in-
A number of intermediate computational estimates and
tersegmental moment and joint angular velocity. As such,
results are provided as well, including:
power does not have directionality. In the interpretation
■ The relative and absolute angular position of the foot, of clinical data, however, it is often advantageous to dis-
either computed from the motion capture data or mea- play each of the planar contributions to intersegmental
sured directly, e.g., electromagnetic sensors, power separately in order to better appreciate the mechan-
■ The location of the foot COM, the mass of the foot, mfoot , ical effect of an intersegmental moment about a particular
and the mass moment of inertia of the foot, Ifoot , typi- anatomical axis. Consequently, in the normative data pre-
cally estimated based on the anthropometric character- sented later in the chapter, sagittal and coronal interseg-
istics of the subject, and mental power will be described separately.
■ The linear acceleration of the foot COM, ag , and the an- Other investigators (5) have examined both rotational
gular acceleration of the foot, αfoot , calculated through and translational intersegmental power through the use
numerical differentiation of the location of the COM and of a six degree-of-freedom gait model (i.e., three degrees
the foot attitude, respectively (80–82). of rotation and three degrees of translation at each joint),
expressed as
The intersegmental forces and moments are determined
PJ = MJ · ω J + FJ · vJ
from the measurement of the external ground reaction
loads and estimates of segment weight and inertia. where FJ is the net intersegmental force and vJ is the trans-
Once the reactions at the ankle, FA and MA , have been lational velocity of one segment relative to the other at
determined, then the process can be repeated for an anal- joint J. Translational ankle intersegmental power values
ysis of the shank, or lower leg, to determine the knee inter- were found to be small relative to the associated rota-
segmental force and moment. In this step of the process, tional intersegmental power contributions during normal
the ankle intersegmental force, FA and intersegmental mo- gait, approximately an order of magnitude smaller. While
ment, MA , are treated as the known external loads to the these translational power contributions were statistically
distal shank in order to determine the unknown knee in- significant, their clinical significance and relevance is not
tersegmental force and moment, FK and MK , respectively. known.
Again, a number of measured values are required, e.g., the The computational process outlined above is referred to
locations of the ankle and knee joint centers, attitude of as “inverse dynamics,” where movement kinematics and
the shank. Estimated values are required, e.g., the loca- external loads are measured and internal reactions are
tion of the shank COM, the mass of the shank, and the computed (9). This process contrasts with a “forward dy-
mass moment of inertia of the shank. Intermediate com- namics” analysis or a “forward simulation” model, where
puted values are required as well, e.g., the linear accelera- internal forces, e.g., muscle forces, are numerically applied
tion of the shank COM and the angular acceleration of the and the resulting kinematics and ground reactions are
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

62 Chapter 4 / Kinetics of Normal Walking

predicted (83,85). The validity of each of these techniques force FGRF and the internal force Fmuscle , or
is predicated on the quality of the measured data and the
FJ1 = FGRF + Fpfm
appropriateness of the assumptions associated with the
specific biomechanical model. and the moment associated with the muscle force was ex-
The primary assumption in the inverse dynamics ap- pressed as
proach is that soft tissue movement relative to underlying
d2 × Fpfm = d1 × FGRF
bony structures is small. The body segments are assumed
to be “rigid” and not to deform when loaded. Moreover, For an inverse dynamics analysis (Figure 4-10B), the mus-
joint center locations are assumed to be well estimated cle force has been incorporated into the intersegmental
and to remain fixed relative to the associated segment co- force FA and the moment due to muscle force has been
ordinate system. An appreciation of the implications of incorporated into the intersegmental moment MA . Inverse
these assumptions is particularly important in the assess- dynamics again predicts an intersegmental moment MA
ment of clinical results (35). For example, excessive soft based on the external force and distance d1 , is
tissue movement of a patient with obesity can reduce the
MA = d1 × FGRF
quality of estimates of the joint center locations as well as
distort kinematic quantities such as angular velocities and Inverse dynamics, however, predicts an intersegmental
accelerations. force FA that is equal in magnitude to the external force
Two fundamental limitations associated with inverse FGRF ,
dynamics should be reiterated as well. The net interseg-
FA = FGRF
mental moment combines moments due to individual
muscle forces and other soft tissue structures that cross the and underestimates the joint contact force by an amount
joint. Moreover, the intersegmental force underestimates equal to Fmuscle .
the joint contact force produced between articulating sur- If agonist and antagonist muscle forces are introduced,
faces of the joint (77). Paul (50) reports peak hip force the dilemma becomes more challenging. From Figure
magnitudes of 3.9 times body weight compared to peak in- 4-10C, it can be seen that the joint contact force coun-
tersegmental force magnitudes of approximately 1.1 times ters both a plantar flexor muscle force Fpfm , a dorsiflexor
body weight (4). muscle force Fdfm as well as the external force FGRF , or
A reconsideration of the previous example illustrates
FJ2 = FGRF + Fpfm + Fdfm
this dilemma. Shown in Figure 4-10A, the static foot is
again loaded with a known external GRF FGRF at known The same free-body-diagram (Figure 4-10B) applies in
distance d1 anterior to the ankle. Other external loads this scenario. That is, both agonist and antagonist mus-
such as the weight of the foot are again neglected. Un- cle forces have been incorporated into the intersegmental
known internal muscle (Fpfm ) and joint (FJ1 ) forces are force FA and the moments due to both muscle forces have
also shown. From the previous discussion, it was appreci- been incorporated into the intersegmental moment MA .
ated that the joint contact force was the sum of the external Inverse dynamics continues to predict an intersegmental

FIGURE 4-10. Static equilibrium of the lower leg. (A) An external force FGRF and internal plantar
flexor muscle and joint forces, Fpfm and FJ1 , respectively, are shown. The weight of the foot is neglected.
(B) The two-dimensional “free-body-diagram” of the foot associated with the loading shown in (A)
and (C ). (C) The same schematic of the foot/ankle (A) with the addition of co-contraction, an internal
dorsiflexor muscle force Fdfm is depicted. Note that co-contraction does not change the value of the net
intersegmental moment MA and that the joint contact forces FJ1 and FJ2 are not equal, nor are they equal
to the intersegmental force FA .
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 63

force FA equal in magnitude to only the external force FGRF

and consequently, underestimates the joint contact force
by an amount equal to Fpfm + Fdfm . In Figure 4-10C, the
plantar flexor muscle moment must now balance the in-
ternal dorsiflexor muscle moment as well as the external
moment due to the force FGRF , or
d2 × Fpfm = d1 × FGRF + d3 × Fdfm
Alternatively, it can be seen that the combination of plantar
flexor and dorsiflexor moments balance the moment due
to the external force FGRF , or
d2 × Fpfm − d3 × Fdfm = d1 × FGRF
Intersegmental moment MA predicted by inverse dynamics
remains unchanged, or FIGURE 4-11. Schematic of a person just after right foot con-
tact. The external ground reaction force produces an external hip
MA = d1 × FGRF flexor moment that is resisted by an opposite and equal internal
hip extensor moment.
and has analytically combined the moment due to the plan-
tar flexor muscle force with the moment due to the dorsi-
flexor muscle force. Inverse dynamics underestimates the
joint contact force and cannot distinguish internal mo- moments are found in the literature, i.e., the display con-
ments produced by separate (and perhaps antagonistic) vention varies between journals. Also, the intersegmental
muscles. moment and power results are normalized with respect to
For additional details associated specifically with the body mass in this chapter. The literature is inconsistent
calculation of intersegmental moments and powers, read- with respect to this display convention as well.
ers are referred to Winter (77), Davis et al. (17), Kaufman The intersegmental moments described in this section
et al. (30), Õunpuu et al. (45), and Palladino and Davis (49). reflect the net moment applied across the joint, consistent
Readers are also encouraged to consider the classical con- with the discussion presented above. A net intersegmental
tributions of Braune and Fischer (republished 3), Elftman moment reflects the moments of both agonist and antag-
(20), Bresler and Frankel (4), and Paul (50). onist muscles as well any other soft tissue structures that
cross the joint. Consequently, for example, a net “extensor
moment” reflects a dominance of internal extensor forces
INTERSEGMENTAL MOMENTS AND over flexor forces. It does not imply the presence of only ex-
POWERS DURING WALKING tensor muscle forces or the absence of flexor muscle forces.

The intersegmental moments and powers described here

Ankle – Sagittal Plane of the
are associated with a group of 27 pediatric subjects
Talocrural Joint
(17 females, 10 males, mean age: 9.9 ± 3 years, mean
mass: 33 ± 12 kg, mean height: 137 ± 17 cm) evaluated at At and immediately following foot contact, the ground re-
Newington Children’s Hospital (Newington, Connecticut) action force is applied to the heel and its line of action
in 1988 and 1989. All of the pediatric subjects were without passes posterior to the ankle center. Consequently, it pro-
neuromuscular and orthopaedic impairment and walked duces an external moment that tends to plantar flex the
at a self-selected speed (119 ± 14 cm/sec) and cadence (130 ankle. This external plantar flex moment is countered by
± 13 steps/min). Readers are referred to Õunpuu et al. (46) an internal ankle dorsiflexor moment (Figure 4-12, Point
for a description of the protocols and methods used in the A). This dorsiflexor moment (associated with eccentric ac-
data collection and analysis. tivity of the ankle dorsiflexors) controls the lowering of
All of the moments presented in this section are the “in- the mid- and forefoot to the ground. The magnitude of
ternal” intersegmental moments that are produced by the this dorsiflexor moment is small, because the GRF passes
body to either modify or control the effects of “external” relatively near the ankle center and the magnitude of the
loads. Perry (52) describes this as an internal response to GRF is still small (relative to its peak value at the end of
an external demand. As illustrated in Figure 4-11, a GRF this double support interval); refer to Point A in Figure 4-5.
that passes anterior to the hip by distance d creates an ex- Note that the mass of the foot is small relative to the other
ternal hip flexor moment. The external hip flexor moment segments of the body and consequently the contribution
is balanced with an equal and opposite internal hip exten- of segment weight and segment inertia to the ankle inter-
sor moment. Presentations of both external and internal segmental moment can be considered negligible in stance.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

64 Chapter 4 / Kinetics of Normal Walking

FIGURE 4-12. Sagittal plane joint angles, moments and powers for the hip, knee and ankle associated
with a group of 27 pediatric subjects evaluated at Newington Children’s Hospital (Newington, Connecti-
cut) in 1988 and 1989, as described in (46). Shown are average values (solid line), one standard deviation
in average value (gray band), and average foot off (vertical gray line).

Once the foot is flat, the COP moves quickly forward un- mains small. Intersegmental power is being absorbed be-
der the plantigrade foot and the line of action of the GRF cause the plantar flexor moment is associated with a dor-
passes anterior to the ankle center. Consequently, the GRF siflexing motion.
produces an external dorsiflexor moment that increases The ankle begins to plantar flex after reaching a peak
rapidly in magnitude. This external dorsiflexor moment dorsiflexion angle at about 40% of the gait cycle. The ve-
is balanced by an internal plantar flexor moment (Figure locity of this plantar flexing motion increases as reflected
4-12, Point B). This internal plantar flexor moment (asso- by the increasing slope of the ankle angle plot (Figure
ciated with eccentric activity of the ankle plantar flexors) 4-12, Point D). Consequently, the internal plantar flexor
controls the forward advancement of the shank over the moment is combined with a significant plantar flexing an-
plantigrade foot and is commonly referred to as the “plan- gular velocity to produce a significant ankle power gener-
tar flexion-knee extension couple.” ation burst (Figure 4-12, Point E). This ankle power gener-
Note that ankle power magnitudes remain small dur- ation lifts the heel off the ground and accelerates the ankle
ing this interval (Figure 4-12, Point C), despite the rapid center upward and forward. This power generation is at-
increase in internal plantar flexor moment. Recall that in- tributed to concentric activity of the ankle plantar flexors
tersegmental power is the product of intersegmental mo- and has been termed “push-off” (75). However, the electri-
ment and joint angular velocity. The magnitude of joint cal (EMG) activity of the primary plantar flexor muscles
angular velocity is reflected by the slope (or steepness) of (gastrocnemius and soleus) begins to subside in the mid-
the joint angle plot. During this interval, the ankle is dor- dle of this power generation burst (at about 50% of the
siflexing, but at a slower and slower rate (as reflected in a gait cycle) (52). Perry (52) suggests that the plantar flexors
decreasing slope). Consequently, although the internal stiffen the ankle allowing the leg to rotate over a forefoot
plantar flexor moment is increasing, the ankle angular fulcrum, referring to this as “roll-off.” Results from in-
velocity is diminishing and the intersegmental power re- duced acceleration analysis (33) and forward simulation
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 65

(39), however, demonstrate a significant contribution to nificant motion of the knee in swing is not apparent from
forward progression of the body by this plantar flexor the knee moment and power.
“push-off” mechanism. During single support, the knee is stabilized by the plan-
tar flexion–knee extension couple described earlier. The in-
creasing internal ankle plantar flexor moment seen at this
Hip – Sagittal Plane of the Joint
time controls forward advancement of the shank and thus
In the first part of stance, the body COM is pulled forward helps stabilize knee position. The hip extensor moment in
and upward with a hip extensor moment (associated with stance also aids in stabilizing the knee by controlling the
concentric contraction of the hip extensors) (Figure 4-12, motion of the thigh. Sadeghi et al. (60) substantiates this
Point F). Hip power generation (Figure 4-12, Point G) aids high degree of coordination between the ankle and hip in
in elevating the body COM to its highest elevation in mid- providing support during stance.
stance while contributing to forward momentum. Later in Peak knee extension in stance occurs at approximately
single support, the body COM continues to translate for- 40% of the gait cycle, at which point the swing phase knee
ward and also falls downward under the influence of grav- flexion wave begins. Recall that at this same point in the
ity. The resulting hip flexor moment and power absorption gait cycle, the ankle is maximally dorsiflexed and ankle
is associated with the passive elongation of the hip flexor power generation begins to produce ankle plantar flexion,
musculature (Figure 4-12, Point H). thereby displacing the ankle center upward and forward.
The hip reaches full extension and begins to flex at 50% This action may also accelerate the knee center upward
of the gait cycle. The hip flexor moment produces the hip and forward as well, thereby initiating the swing phase
flexing motion as realized in the hip power generation knee flexion wave. A second source of power for the pro-
burst (Figure 4-12, Point I) that continues to increase in duction of knee flexion in swing is the late stance hip flexor
magnitude until foot off. This hip power burst (attributed moment with its associated hip power generation that be-
to concentric contraction of the hip flexors) helps propel gins at approximately 50% of the gait cycle (51,53,84). A
the lower extremity into swing. Note that the timing of hip flexor moment accelerates the thigh and the knee joint
the onset of this power generation (50% of the gait cycle) center forward. The proximal end of the shank translates
is closely aligned with peak ankle power generation. This forward as well, but segment inertia limits the simulta-
provides evidence that the ankle initially propels the stance neous rotation of the shank. This combination of thigh
limb toward swing in terminal stance and then the hip as- rotation and shank translation results in knee flexion.
sumes that role in double support and into swing phase. A knee flexor moment (Figure 4-12, Point M) is pro-
This observation is further supported by a principle com- duced in the second half of swing phase at the same time
ponent analysis of the hip and ankle moments (60). that the knee is rapidly extending. One might have more
The hip extensor moment in the second half of swing readily anticipated that this rapid knee extension would
(Figure 4-12, Point J) decelerates the lower extremity in have required a net knee extensor moment. This apparent
preparation for foot contact. It is observed that both the hip contradiction can be explained by again examining the si-
flexor moment/acceleration in late stance and early swing multaneous hip kinetics. Recall, that a hip extensor mo-
and the hip extensor moment/deceleration mechanisms af- ment is seen in the second half of swing. This hip extensor
fect knee motion, as will be described in the next section. moment decelerates the thigh and the knee joint center.
Consequently, the proximal end of the shank decelerates,
but the center of mass and distal end of the shank are
Knee – Sagittal Plane of the Joint
carried forward by momentum, causing the shank to ro-
A net internal knee extensor moment in early stance (Fig- tate forward. The forward rotation of the thigh is slowed
ure 4-12, Point K) supports and controls knee flexion dur- while the forward rotation of the shank is facilitated, thus
ing the first double support interval, also referred to as the knee is extended. The knee flexor moment controls the
loading response or weight acceptance. The correspond- rate of knee extension to avoid injury to posterior ligamen-
ing initial knee power absorption wave (Figure 4-12, Point tous structures. The hamstrings, in crossing both the hip
L) is consistent with a knee extensor moment and a si- and the knee, may aid in simultaneously producing a hip
multaneously flexing knee. This is followed by a smaller extensor moment that promotes knee extension and a knee
knee power generation wave that is also consistent with flexor moment that controls that motion.
the knee extensor moment and an extending knee motion.
This power absorption and generation may be associated
Hip – Coronal Plane of the Joint
with eccentric followed by concentric contraction of the
knee extensors. Aside from the early extensor moment, During ambulation, the line of action of the body weight
knee moments during stance are remarkably small given vector typically passes medial to the hip center (Figure
the 5 to 10 degrees of knee flexion found during single 4-13). Consequently, during single support, gravity pulls
support. Moreover, the mechanism that produces the sig- the body COM downward. An internal hip abductor
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

66 Chapter 4 / Kinetics of Normal Walking

moment (Figure 4-14A, Point A) counters this effect and

supports the pelvis and upper body during single support
(commonly associated with eccentric contraction of the
hip abductors). The magnitude of the coronal plane hip
moment during each double support interval is small, con-
sistent with the loading and unloading of the ground reac-
tions during this interval (assuming that the contributions
to the moment by segment weight and segment inertia are
also small). Early in stance, hip power absorption (Figure
4-14A, Point B) results from the combination of the hip ab-
ductor moment and hip adduction (caused by the drop of
the contralateral pelvis). Later in stance, power generation
(Figure 4-14A, Point C) is realized as the hip abductor mo-
ment aids in elevating the depressed contralateral pelvis.

Knee – Coronal Plane of the Joint

With respect to the knee, the line of action of the weight
vector also typically passes medial to the knee center (Fig-
ure 4-13). The distance between the force line of action and
FIGURE 4-13. Schematic of a person in single support. The the knee center is less than at the hip. Consequently, the
weight of the head, arms, and trunk would tend to cause hip pattern of the internal valgus knee moment (Figure 4-14B)
adduction and knee varus. The internal hip and knee moments is similar but reduced in magnitude.
counter this tendency.
For additional detail pertaining to intersegmental mo-
ments and powers during normal gait, readers are referred
to Davis and Õunpuu (16).

FIGURE 4-14. (A) Coronal plane

hip angle, moment and power, and
(B) coronal plane knee moment as-
sociated with a group of 27 pedi-
atric subjects evaluated at Newing-
ton Children’s Hospital (Newington,
Connecticut) in 1988 and 1989, as de-
scribed in (46). Shown are average
values (solid line), one standard de-
viation in average value (gray band),
and average foot off (vertical gray
line).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 67

FIGURE 4-15. Sagittal plane joint angles, moments and powers for the hip, knee and ankle for two ten-
year-old subjects, one with hemiplegic type cerebral palsy (thin line) and one with congenital proximal
myopathy (thick line). Shown also are one standard deviation in average value (gray band) and average
foot off (vertical gray line) associated with the normative data presented in Figure 4-12.

Example—Kinetics in Pathological Gait

A systematic review of this information by a multidisci-
With respect to clinical application, intersegmental mo- plinary clinical team affords an opportunity to distinguish
ment and power data should not be examined in isolation. primary deviations from compensatory mechanisms. Sub-
A comprehensive interpretation of a patient’s walking pat- tle abnormalities seen in the joint kinematics may be high-
tern for clinical decision-making should include their med- lighted by the intersegmental moment and power results.
ical history, observational gait findings documented by a To illustrate these points, the data for two ten-year-
video record, clinical examination results (e.g., lower ex- old females, one with right hemiplegic type cerebral palsy
tremity joint ranges of motion, muscle strength and tone, (CP) and the other with congenital proximal myopathy,
evidence of muscle spasticity and contracture, bony defor- are presented in Figure 4-15. While their diagnoses are
mity) as well as quantitative gait data: stride and temporal quite different, with one exhibiting primarily muscle spas-
measurements, segment and joint kinematics (i.e., angular ticity and hypertonicity (that overlays muscle weakness)
displacement), EMG tracings, and joint kinetics (15,18). and motor control deficits and the other presenting with
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

68 Chapter 4 / Kinetics of Normal Walking

proximal muscle weakness, their sagittal kinematic pat- A comprehensive surgical package was recommended
terns are quite similar. for the girl with CP, including right femoral derotational
Both girls are toe walkers with excessive ankle plantar rotational osteotomy, right medial hamstring lengthening,
flexion throughout the gait cycle (Figure 4-15, Point A) re- and right rectus femoris transfer. These recommendations
sulting from plantar flexor tightness and over-activity. This were based on her clinical examination, kinematics, and
results in an elevated plantar flexor moment in early stance EMG. The excessive plantar flexor moment and prema-
(Figure 4-15, Point B). Moreover, the peak plantar flexor ture ankle power generation in early stance, along with the
moment later in stance is reduced because the equinus (ex- excessive plantar flexion throughout stance and tightness
cessive plantar flexion) ankle position reduces the moment on clinical examination, supported a recommendation for
arm between the GRF and the ankle center. Premature an- right gastrocnemius recession.
kle power generation is seen with both subjects (Figure The referring physician for the girl with proximal my-
4-15, Point C) but for different reasons. For the girl with CP, opathy was considering bilateral gastrocnemius recession
a spastic stretch reflex ends an ankle dorsiflexing motion given her equinus gait pattern. Based on this girl’s limited
in early stance and produces a slight plantar flexing mo- hip power production (Figure 4-15, Point I) and her finely
tion. The girl with proximal myopathy produces the slight tuned control of the GRF to minimize hip extensor mo-
plantar flexing motion in midstance as a “vault” compen- ment requirements, the gait analysis team recommended
sation to aid in clearance for the contralateral swing limb. only continued clinical follow-up with no surgery. For ad-
The magnitude of ankle power generation in late stance ditional information related to the interpretation of patho-
(Figure 4-15, Point D) is reduced for both girls. For both logical gait patterns, readers are referred to Sutherland
girls, the excessively plantar flexed position of their ankles and Davids (66) and Õunpuu (43).
throughout stance predisposes them to poor plantar flex-
ion velocity in late stance. That is, the already-shortened
plantar flexor muscles may be poorly positioned to pro- WORK AND ENERGY
duce additional contractile force. Producing significantly
more plantar flexion in late stance is also counterproduc- Work is done when a force is displaced linearly some dis-
tive with respect to foot attitude, that is, a more vertically- tance along its line of action or a moment is displaced
oriented foot in early swing presents a greater challenge rotationally through some angular displacement. When a
with respect to clearance later in swing. book is put up on a shelf, work is done on the book (i.e.,
For both girls, their knees are excessively flexed at foot its weight is displaced upward). When a wrench tightens a
contact and throughout stance (Figure 4-15, Point E). The bolt, work is done by the wrench (i.e., the applied moment
knee intersegmental moment in stance is reasonably well is displaced in the direction of its application) and work
“compensated” in both cases (Figure 4-15, Point F). That is, is done on the bolt (i.e., the resisting moment offered by
one might anticipate greater knee extensor moment mag- the bolt is displaced opposite to its direction of resistance).
nitudes with 20 to 30 degrees of knee flexion in single sup- The rate at which work is done is defined as power. More
port. The excessive ankle plantar flexion-knee extension power is required to tighten the bolt quickly than slowly. In
couple restrains the forward motion of the shank and sup- mechanics, work and energy are intimately related. When
ports the knee during single support. For the girl with CP, work is done on a “system,” the energy level of the system
an anterior trunk lean (Figure 4-15, Point G) shifts her increases. Conversely, when work is done by a system, the
body COM forward and moves the GRF vector closer to energy level of the system decreases. System definition de-
the knee center. This also elevates her hip extensor mo- pends on the scope of the analysis. In placing the book on
ment in stance (Figure 4-15, Point H). The elevated hip the shelf, the system might be defined to include only the
extensor moment controls forward motion of the thigh, book or it might be expanded to include the book and the
also supporting the knee during single support. Her ad- person performing the task as well.
equate hip extensor strength (documented in the clinical The muscles that produce internal intersegmental mo-
examination) facilitates this compensation. With hip ex- ments during a sit-to-stand task do work. That is, the down-
tensor weakness associated with proximal myopathy, the ward weight (force) of the head, arms, and trunk (HAT) is
other girl is more limited with respect to using her trunk displaced vertically upward. Consequently, the lower ex-
posture to control the GRF. Alternatively, she walks with tremity muscles have done work on the HAT segment. Al-
a shorter step length (73% relative to an age-matched nor- ternatively, this work may be characterized as a change in
mal). In this way, the moments at both the hip and knee are the gravitational potential energy of the segment (defined
reduced as the excursion of the GRF will be reduced and as the product of the weight of the segment and its height
the elevated plantar flexor moment is sufficient to control above some reference level or datum), expressed mathe-
the knee. It is observed that her precise control of the GRF matically as
significantly reduces her need to produce an internal hip
extensor moment. EP = mgh
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 69

where mg is the segment weight (the product of mass and ponents of the angular velocity vector along with the asso-
gravitational constant) and h is the height of the segment ciated principal mass moment of inertia, i.e.,
COM above some datum. In this case, the gravitational po-
tential energy of the HAT segment is increased as the HAT EKR =1/2 Ixx ωx 2 +1/2 Iyy ωy 2 +1/2 Izz ωz 2
COM is displaced upward. In this same way, the gravita-
tional potential energy of the HAT segment increases and The interplay between potential and kinetic energy
decreases over the gait cycle (as described below). is quite interesting. Consider the simple pendulum
Another type of potential energy is elastic potential en- comprised of a small ball (concentrated mass) and a
ergy. The elastic potential energy of a mechanical spring lightweight string shown in Figure 4-16. As this pendu-
is increased when it is compressed (or stretched). Observe lum swings back and forth, kinetic and potential energy
that, again, this elastic potential energy is related to the lin- are exchanged. Immediately after the pendulum is released
ear displacement of the spring force, i.e., the work done on from its highest elevation (Figure 4-16, Point A), its poten-
the spring. In this same way, the elastic potential energies tial energy is maximum and its kinetic energy is zero (the
of the muscles and tendons of the lower extremity change pendulum is released from a stationary position). As the
during activities such as walking. Elastic potential energy pendulum swings down and the ball loses elevation and
cannot be directly assessed during gait, however, as this gains speed (Figure 4-16, Point B), the potential energy de-
would require measures of muscle force and displacement creases while the kinetic energy increases. When the pen-
as well as tendon force and displacement. This represents dulum reaches its lowest point and the ball is moving at its
the most significant limitation of this technique because fastest velocity (Figure 4-16, Point C), the potential energy
elastic potential energy stored in and returned from soft is minimized while the kinetic energy is maximized. Note
tissue structures may be significant in gait during inter- that in this example, if gravitational potential energy is ref-
vals with coactivity of agonists and antagonists muscles, erenced to the lowest elevation of the pendulum path, then
e.g., transitions from swing to stance. the gravitational potential energy would be equal to zero
Kinetic energy, the “energy of motion,” can, however, be at this point. As the pendulum swings away from vertical
computed. Translational kinetic energy, EKT , is a function and the ball gains elevation and loses speed, the potential
of the mass, m, and linear velocity, v, of the segment of energy increases while the kinetic energy decreases (Fig-
interest, expressed as ure 4-16, Point D). At its highest point when the ball stops
momentarily (Figure 4-16, Point E), all of the available ki-
EKT =1/2 mv2 netic energy has been transformed into potential energy,
which is maximized at this point. In this idealized example,
Similarly, rotational kinetic energy, EKR , is expressed as
energy losses to air resistance (friction) and strain in the
EKR =1/2 Iω2 string have been neglected and the transfer between kinetic
and potential energy is complete. Its total mechanical en-
where I is the centroidal mass moment of inertia about an ergy (the sum of kinetic and potential energy) remains con-
axis that coincides with the angular velocity, or a vector stant and illustrates the law of conservation of energy. This
component of the angular velocity, ω. For example, if the idealized pendulum would continue to swing in perpetuity.
motion of walking is assumed to lie in one plane, i.e., a In an extension of this example, agonist and antagonist
sagittal plane, then the axis associated with segmental an- muscles are added to the simple pendulum (Figure 4-17).
gular velocity and the centroidal mass moment of inertia Immediately after the pendulum is released, both muscles
would be perpendicular to that plane. In 3D analyses, the are inactive and the total energy level of the system (which
relationship is easily expanded to include the three com- now includes both muscles as well as the pendulum)

FIGURE 4-16. Energy changes due to mo-

tion of a simple pendulum. If air resistance
and strain in the string are neglected, there
is a complete transfer between potential and
kinetic energy during motion. Consequently,
the total energy of the pendulum remains
constant.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

70 Chapter 4 / Kinetics of Normal Walking

FIGURE 4-17. Agonist and antagonist muscles are added to

a simple pendulum. Concentric muscle contraction (m2 from
t1 to t2 ) increases the energy level of the pendulum, and eccentric
muscle contraction (m1 from t3 to t4 ) decreases the energy level
of the pendulum.
FIGURE 4-18. Potential energy, kinetic energy, and total energy
of the HAT for an adult female walking at 71 m/min.

is constant. It is assumed in this illustration that the en- The energy requirements of a lower extremity segment
ergy change associated with the passive elongation of the and each associated leg segment depicted in Figure 4-19
muscles is negligible. Muscle m2 concentrically contracts are consistent with other investigators (7). The change in
from time t1 to t2 . The muscle does work as its contractile the total energy for the leg segments over the gait cycle is
force results in muscle shortening (i.e., force and displace- due principally to the change in translational kinetic en-
ment are in the same direction). Consequently, energy is ergy in swing. However, for this subject, rotational kinetic
added to the system. From t2 to t3 both muscles are again energy did contribute up to 7% of the total energy of the
inactive and the system energy level remains constant. Ec- shank in swing (78). With each leg segment, the total en-
centric contraction of muscle m1 from t3 to t4 removes en- ergy is minimal in midstance when segment velocities and
ergy from the system. Work is done on the muscle since its kinetic energies are minimal as well. For gravitational po-
contractile force results in muscle lengthening (i.e., force tential energy, the vertical displacement of each segment is
and displacement are in opposite directions). measured relative to the ground. Energy is added to each
This transfer between kinetic and potential energy can leg segment in late stance and taken away from each leg
be seen in gait (7), as illustrated by the representative data segment in late swing.
presented in Figure 4-18 for the HAT. The gravitational po- The change in energy requirements for both legs is less
tential energy reaches a minimum twice in the gait cycle than the energy change for one leg during gait, 12 Joules
during each period of double support, when the COM of compared with 18 Joules for the individual’s data presented
the HAT is at its lowest point vertically. The kinetic en- in Figure 4-19. That is, the minimal energy of one (stance)
ergy (the sum of translational and rotational kinetic en- leg combines with maximal energy of the other (swing)
ergies) is greatest at these same points, coincident with limb to result in a biphasic pattern for both legs. A similar
the maximum velocities of the HAT. The gravitational po- biphasic pattern is seen for the total energy requirement
tential energy peaks at midstance, at about 30% and 80% of the entire body during gait, with change in energy seen
of the gait cycle in the data shown in Figure 4-18, when in late single support and during the second double sup-
the COM of the HAT has been lifted to its highest eleva- port interval for each side. An asymmetry in the energy
tion. Note that 80% in the gait cycle for the ipsilateral limb requirements between the two lower extremities can be
is coincident with midstance for the contralateral limb. appreciated for this individual with the peak body total en-
The velocity of the COM of the HAT has also slowed at ergy level in swing greater than the peak body total energy
midstance; consequently, kinetic energy is minimal at this level in stance (coinciding with swing for the contralateral
point. The change in total HAT energy over the gait cycle limb), i.e., 516 Joules as compared to 511 Joules.
(8 Joules) is less than either the change in gravitational po- The rate of change of these energy values reflects the
tential energy (11 Joules) or kinetic energy (14 Joules). The power requirements of the body during gait, i.e., power
exchange between kinetic and potential energy, although is the rate that energy is supplied or taken away. For the
incomplete, reduces the lower extremity muscular effort subject whose data are depicted in Figure 4-20, the whole
(and metabolic energy) required for walking (7). body in early and late stance requires power. During both
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 71

of these intervals, power is required to lift the body COM

from first double support to midstance and to propel the
stance limb into swing. Sources of energy include muscles
acting on the body segments, e.g., the lower extremities,
or the release of stored elastic energy. Power is also taken
away (energy is released or lost) from the whole body dur-
ing the second half of single support and in late swing.
During both of these intervals, the body COM falls from
its highest elevation at midstance and a swing leg is de-
celerated in preparation for foot contact. Work is done on
structures capable of absorbing energy, e.g., the passive
elongation of muscles shortened by contraction, such as
the hip flexors in single support and the hamstrings in late
swing.
These power “requirements” were derived mathemati-
cally from the segmental mechanical energy profiles. An al-
ternative power flow analysis during gait employs the rate
of displacement (i.e., linear and angular velocity) of the in-
tersegmental forces and moments (58). For example, the
power flow within the shank can be computed through a
consideration of the ankle and knee moments rotating with
the shank and the ankle and knee forces translating with
the ankle and knee centers, respectively. Robertson and
Winter (58) found good agreement between the rate of en-
ergy change and power flow in the lower extremity. Siegel
et al. (61) improved this correlation for the foot through a
full 3D analysis that also included the power contributions
FIGURE 4-19. Energy levels for the whole body, HAT, both lower of the ground reactions. The results of a power flow analy-
extremities, right lower extremity, thigh, shank, and foot for one sis for one adult subject are presented in Figure 4-21 along
adult female walking at 71 m/min. with her rate of mechanical energy change data for com-
parison. The correlation between the rate of energy change
and power flow over the gait cycle ranged from 0.796 for
the thigh to 0.997 for the foot, again, comparable to values
reported previously (58,61).
To appreciate the mechanisms associated with segmen-
tal power flow data depicted in Figure 4-21, the different
contributors, e.g., joint force and moment, are presented in
Figure 4-22. For the foot in stance, ankle moment power
generation is largely offset by ankle force power absorp-
tion. Ankle force power generation in the first half of swing
increases the translational kinetic energy of the foot (as
was seen in Figure 4-19). Conversely, ankle force power
absorption in the second half of swing decreases the trans-
lational kinetic energy of the foot. For the shank, a dom-
inant ankle force power generation (moderated by ankle
moment, knee moment, and knee force power absorption)
provides initial shank power generation in late stance.
Knee force power generation is the primary contributor to
shank power generation just before toe-off and into early
swing. This phasing of power generation is consistent with
an early contribution by the plantar flexor muscles fol-
lowed by a later contribution by the hip flexors. In late
FIGURE 4-20. Rate of energy change or power requirement for swing, ankle force power generation is offset by knee force
the whole body, HAT, both legs, and the right leg for one adult power absorption. Consequently, shank power absorption
female walking at 71 m/min. in the second half of swing is primarily due to knee
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

72 Chapter 4 / Kinetics of Normal Walking

FIGURE 4-21. Power flow of each right leg segment and the en-
tire leg (solid line) shown with the associated mechanical power
requirements for the same segments (dashed line) for one adult
female walking at 71 m/min.

moment power absorption, consistent with hamstring

power absorption. The sharp rise in knee moment power
generation at the end of the gait cycle for this particu-
lar individual is associated with an aberrant flexing of the FIGURE 4-22. Contributions to power flow of each right leg
knee in preparation for foot contact (a slight correction for segment for one adult female walking at 71 m/min.
knee hyperextension during terminal swing). A coincident
increase in knee force power absorption is seen as well this technique to estimate the role of the plantar flexors on
at this time. For the thigh, excessive hip moment power HAT energy. More recently, other investigators have used
absorption in stance for this individual results in over- forward dynamics techniques in conjunction with power
all thigh power absorption in stance. Hip moment power flow analysis to examine more directly the effect of inter-
generation at foot off (moderated by the other contribu- segmental moments on trunk propulsion (57) and the ef-
tors) results in thigh power generation, increasing the me- fect of intersegmental moments in one leg on the dynamics
chanical (translational kinetic) energy of the thigh. Hip of the contralateral leg as well as the trunk (62). These lat-
and knee forces contribute to thigh power absorption dur- ter approaches indicate that both positive and negative in-
ing midswing. Knee force power generation is balanced tersegmental power can contribute to the increase in the
by hip force power absorption in late swing. The sharp energy level of other segments. Inferences can be made
rise in thigh power generation in late stance is attributed from power flow analysis based on intersegmental mo-
to knee force power generation and results in a small in- ments computed through inverse dynamics, but this tech-
crease (2 Joules) in the energy level for the thigh (Figure nique cannot directly identify those muscles that are re-
4-19). Readers are referred to the work of Robertson and sponsible for propulsion and support during locomotion.
Winter (58,79) for additional details pertaining to these
energy transfer processes.
A power flow analysis provides a measure of the power SUMMARY
into and out of individual lower extremity segments. By
tracking the power flow between the segments in the leg Human locomotion is fundamentally related to the pro-
and assuming that segmental energy increases only with duction of muscle force that either creates or controls that
intersegmental power generation, Meinders et al. (37) used motion. Muscle forces act at some distance away from
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 73

the joints that they span. These moment arms provide 3. Braune W, Fischer O. The Human Gait (translation). Berlin: Springer,
the muscles with the leverage to generate intersegmental 1987.
4. Bresler B, Frankel JP. The forces and moments in the leg during level
moments that produce and control the rotations of those walking. Trans ASME 1950;72:27–36.
joints. It also allows the muscle forces to be transmitted 5. Buczek FL, Kepple TM, Siegel KL, Stanhope SJ. Translational and
to the joints so that these joint forces can accelerate and rotational joint power terms in a six degree-of-freedom model of the
normal ankle complex. J Biomech 1994;27:1447–1457.
move the HAT, referred to as the passenger unit by Perry 6. Cappozzo A, Berme N. Body segment inertial parameter determina-
(52). The ground reaction loads are an external reflection of tion. In: Berme N, Cappozzo A, eds. Biomechanics of Human Move-
these internal forces. Muscles provide “local” control over ment: Applications in Rehabilitation, Sports and Ergonomics. Wor-
thington, Ohio: Bertec Corporation; 1990;179–185.
the joints that they span, but as described by Zajak and his 7. Cappozzo A, Figura F, Marchetti M. The interplay of muscular and
colleagues (40,85), the primary function of muscles can be external forces in human ambulation. J Biomech 1976;9:35–43.
the redistribution of power and energy among the body 8. Chandler RF, Clauser CE, McConville JT, Reynolds HM, Young JW.
Investigation of inertial properties of the human body. AMRL-TR-74-
segments. 137, AD-A016-485. DOT-HS-801-430. Aerospace Medical Research
Inverse dynamics approaches provide some insight into Laboratories, Wright-Patterson Air Force Base, Ohio; 1975.
the mechanisms that move and control the lower extremity 9. Chao EYS. Determination of applied forces in linkage systems with
known displacements: with special application to biomechanics. PhD
joints. This information can be useful in assessing treat- Dissertation, University of Iowa, Iowa City; IA, 1971.
ment options or evaluating treatment approaches clin- 10. Chao EY, Rim K. Application of optimization principles in determin-
ically. Inverse dynamics is limited with respect to the ing the applied moments in human leg joints during gait. J Biomech
1973;6:497–510.
prediction of muscle forces during walking. Net internal 11. Clauser CE, McConville JT, Young JW. Weight, volume and center of
intersegmental moments combine contributions of both mass of segments of the human body. AMRL-TR-69-70. Aerospace
agonist and antagonist muscles. Intersegmental forces are Medical Research Laboratory, Wright-Patterson Air Force Base,
Ohio; 1969.
underestimated, i.e., do not reflect joint contact force mag- 12. Crowninshield RD. Use of optimization techniques to predict muscle
nitudes. Optimization techniques to allow the prediction forces. J Biomech Eng 1978;100:88–92.
of muscle force from computed intersegmental forces and 13. Davis J, Kaufman KR, Lieber RL. Correlation between active and
passive isometric stress and intramuscular pressure in the isolated
moments are improving (64), but are still limited with re- rabbit tibialis anterior muscle. J Biomech 2003;36:505–512.
spect to clinical application. The direct measurement of 14. Davis RB. The motion analysis laboratory. In: Gage JR, ed. The Treat-
muscle force with miniature intramuscular pressure trans- ment of Gait Problems in Cerebral Palsy. London: MacKeith Press;
2004;90–98.
ducers may ultimately resolve this empirical quandary 15. Davis RB, DeLuca PA, Õunpuu S. Analysis of gait. In: Schneck DJ,
(13,32). Bronzino JD, eds. Biomechanics: Principles and Applications Boca
Forward dynamics approaches show good promise in Raton: CRC Press; 2002;8:1–9.
16. Davis RB, Õunpuu S. Kinetics of normal gait. In: Gage JR, ed. The
elucidating the relationships between muscle force pro- Treatment of Gait Problems in Cerebral Palsy. London: MacKeith Press;
duction and gait kinematics as these techniques can pro- 2004;120–134.
vide estimates of muscle force. Sophisticated whole-body 17. Davis RB, Õunpuu S, Tyburski DJ, Gage JR. A gait analysis data
collection and reduction technique. Hum Mov Sci 1991;10:575–
forward simulations afford an excellent platform from 587.
which to investigate the role of different muscles and 18. DeLuca, PA. Gait analysis in the treatment of the ambulatory child
muscle groups in locomotion (1,2). The sensitivity of the with cerebral palsy. Clin Orthop Rel Res 1991;264:5–75.
19. Dempster WT. Space requirements of the seated operator: Ge-
conclusions of these simulations to uncertainty in the un- ometrical, kinematic, and mechanical aspects of the body with
derlying model and the model parameters needs to be fur- special reference to the limbs. WADC-55-159, AD-087-892. Wright
ther clarified (85). Current efforts to build models that Air Development Center. Wright-Patterson Air Force Base, Ohio;
1955.
incorporate subject-specific parameters related to muscle 20. Elftman H. Forces and energy changes in the leg during walking. Am
characteristics and bone geometry are expected to provide J Physiol 1939;125:339–356.
assistance in clinical decision-making (55,56). Significant 21. Eng JJ, Winter DA. Kinetic analysis of the lower limbs during walk-
ing: What information can be gained from a three-dimensional
challenges lie ahead, however. For example, forward simu- model? J Biomech 1995;28:753–758.
lations require optimization criteria to direct the computa- 22. Gage JR, Perry J, Hicks RR, Koop S, Werntz JR. Rectus femoris trans-
tional process. It is not clear that the optimization strategy fer to improve knee function of children with cerebral palsy. Dev Med
Child Neurol 1987;29:159–166.
that is appropriate in the investigation of able-body gait 23. Hanavan EP A mathematical model of the human body. AMRL-
will apply in pathological gait analysis. TR-64-102, AD-608-463. Aerospace Medical Research Laboratories,
Wright-Patterson Air Force Base, Ohio; 1964.
24. Harrington IJ. Static and dynamic loading patterns in knee joints
with deformities. J Bone Joint Surg 1983;65-A(2):247–259.
25. Hatze HA. A mathematical model for the human body. J Biomech
REFERENCES 1980;13:833–843.
26. Hullin MG, Robb JE, Loudon IR. Ankle-foot orthosis function in low-
1. Anderson FC, Goldberg SR, Pandy MG, Delp SL. Contributions of level myelomeningocele. J Pediatric Orthop 1992;12:518–521.
muscle forces and toe-off kinematics to peak knee flexion during the 27. Jenson RK. Body segment mass, radius and radius of gyration pro-
swing phase of normal gait: an induced position analysis. J Biomech portions of children. J Biomech 1986;19:359–368.
2004;37:731–737. 28. Johnson F, Leitl S, Waugh W. The distribution of load across the knee:
2. Anderson FC, Pandy MG. Dynamic optimization of human walking. a comparison of static and dynamic measurements. J Bone Joint Surg
J Biomech Eng 2001;123:381–390. 1980;62-B:346–349.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

74 Chapter 4 / Kinetics of Normal Walking

29. Judge J, Davis RB, Õunpuu S. Step length reductions in advanced 54. Prodromos CC, Andriacchi TP, Galante JO. A relationship between
age: the role of ankle and hip kinetics. J Gerontology: Med Sci gait and clinical changes following high tibial osteotomy. J Bone Joint
1996;51A:M303–M312. Surg 1985;67–A:1188–1194.
30. Kaufman KR, An KN, Chao EYS. A comparison of intersegmental 55. Riemer RF. Validated, subject-specific model for predicting tibio-
joint dynamics to isokinetic dynamometer measurements. J Biomech femoral knee force distribution from intersegmental forces. MS The-
1995;28:1243–1256. sis, Department of Mechanical Engineering, Houghton, Michigan:
31. Kaufman KR, An KN, Litchy WJ, Chao EYS. Physiological predic- Michigan Technological University; 2003.
tion of muscle forces: Part I - Theoretical formulation. Neurosci 56. Riemer RF, Haut Donahue T, Kaufman KR. Tibiofemoral load distri-
1991;40:781–792. bution during gait of normal subjects. Proceedings of American Soci-
32. Kaufman KR, Wavering T, Morrow D, Davis J, Lieber RL. Per- ety of Biomechanics, 2004.
formance characteristics of a pressure microsensor. J Biomech 57. Riley PO, Dela Croce U, Kerrigan DC. Effect of age on lower extremity
2003;36:283–287. joint moment contributions to gait speed. Gait Posture 2001;14:264–
33. Kepple TM, Siegel KL, Stanhope SJ. Relative contributions of the 270.
lower extremity joint moments to forward progression and support 58. Robertson DGE, Winter DA. Mechanical energy generation, absorp-
during gait. Gait Posture 1997;6:1–8. tion and transfer amongst segments during walking. J Biomech
34. Lai KA, Kuo KN, Andriacchi TP. Relationship between dynamic de- 1980;13:845–854.
formities and joint moments in children with cerebral palsy. J Pedi- 59. Rose SA, DeLuca PA, Davis RB, Õunpuu S, Gage JR. Kinematic
atric Orthop 1988;8:690–695. and kinetic evaluation of the ankle following lengthening of the gas-
35. Manal K, McClay I, Richards J, Galinat B, Stanhope S. Knee mo- trocnemius fascia in children with cerebral palsy. J Pediatric Orthop
ment profiles during walking: errors due to soft tissue movement of 1993;13:727–732.
the shank and the influence of the reference coordinate system. Gait 60. Sadeghi H, Somaye S, Prince F, Allard P, Labelle H, Vaughan CL.
Posture 2002;15:10–17. Functional roles of the ankle and hip sagittal muscle moments in
36. McConville JT, Churchill TD, Kaleps I, Clauser CE, Cuzzi J. able-bodied gait. Clin Biomech 2001;16:688–695.
Anthropometric relationships of body and body segment mo- 61. Siegel KL, Kepple TM, Caldwell GE. Improved agreement of foot
ments of inertia. Technical report AFAMRL-TR-80-119, Air Force segmental power and rate of energy change during gait: Inclusion of
Aerospace Medical Research Laboratory, Aerospace Medical Divi- distal power terms and use of three-dimensional models. J Biomech
sion, Air Force Systems Command, Wright-Patterson Air Force Base, 1996;6:823–827.
Ohio; Yr. 62. Siegel KL, Kepple TM, Stanhope SJ. Joint moment control of me-
37. Meinders M, Gitter A, Czerniecki JM. The role of plantar flexor work chanical energy flow during normal gait. Gait Posture 2004;19:
muscle work during walking. Scan J Rehab Med 1998;30:39–46. 69–75.
38. Meriam JL, Kraige LG. Engineering Mechanics—Dynamics. New 63. Spoor CW, van Leeuwen JL, Meskers CGM, Titulaer AF, Huson A.
York: John Wiley & Sons; 1997. Estimation of instantaneous moment arms of lower-leg muscles. J
39. Neptune RR, Kautz SA, Zajac FE. Contributions of the individual Biomech 1990;23:1247–1259.
ankle plantar flexors to support, forward progression and swing ini- 64. Stansfield BW, Nicol AC, Paul JP, Kelly IG, Graichen F, Bergmann
tiation. J Biomech 2001;34:1387–1398. G. Direct comparison of calculated hip joint contact forces with
40. Neptune RR, Zajac FE, Kautz SA. Muscle force redistributes seg- those measured using instrumented implants. An evaluation of a
mental power for body progression during walking. Gait Posture three-dimensional mathematical model of the lower limb. J Biomech
2004;19:194–205. 2003;36:929–936.
41. Olney SJ, Costigan PA, Hedden DM. Mechanical energy patterns 65. Sutherland DH, Cooper L, Daniel D. The role of the ankle plantar
in gait of cerebral palsied children with hemiplegia. Phys Ther flexors in normal walking. J Bone Joint Surg 1980;62-A:354–363.
1987;67:1348–1354. 66. Sutherland DH, Davids JR. Common gait abnormalities of the knee
42. Olney SJ, MacPhail HA, Hedden DM, Boyce WF. Work and power in in cerebral palsy. Clin Orthop 1993;288:139–147.
hemiplegic cerebral palsy gait. Phys Ther 1990;70:431–438. 67. Sutherland DH, Santi M, Abel MF. Treatment of stiff-knee gait
43. Õunpuu S. Patterns of gait pathology. In: Gage JR, ed. The Treat- in cerebral palsy: a comparison by gait analysis of distal rectus
ment of Gait Problems in Cerebral Palsy. London: MacKeith Press; femoris transfer versus proximal rectus release. J Pediatric Orthop
2004;217–237. 1990;10:433–441.
44. Õunpuu S, Bell KJ, Davis RB, DeLuca PA. An evaluation of the pos- 68. Thomson JD, Õunpuu S, Davis RB, DeLuca PA. The effects of the
terior leaf spring orthosis using joint kinematics and kinetics. J Pedi- ankle foot orthoses on knee and ankle motion in persons with
atric Orthop 1996;16:378–384. myelomeningocele: an evaluation using three-dimensional gait anal-
45. Õunpuu S, Davis RB, DeLuca PA. Joint kinetics: methods, interpreta- ysis. J Pediatric Orthop 1999;19:27–33.
tion and treatment decision-making in children with cerebral palsy 69. Wang J-W, Kuo KN, Andriacchi TP, Galante JO. The influence of walk-
and myelomeningocele. Gait & Posture 1996;4:62–78. ing mechanics and time on the results of proximal tibial osteotomy.
46. Õunpuu S, Gage JR, Davis RB. Three-dimensional lower extremity J Bone Joint Surg 1990;72-A:905–909.
joint kinetics in normal pediatric gait. J Pediatric Orthop 1991;11:341– 70. Wang H, Olney SJ. Relationships between alignment, kinematic and
349. kinetic measures of the knee of normal elderly subjects in level walk-
47. Õunpuu S, Muik E, Davis RB, Gage JR, DeLuca PA. Part I: the effect ing. Clin Biomech 1994;9:245–252.
of the rectus femoris transfer location on knee motion in children 71. Waters RL, Garland DE, Perry J, Habig T, Slabaugh P. Stiff-legged
with cerebral palsy. J Pediatric Orthop 1993;13:325–330. gait in hemiplegia: surgical correction. J Bone Joint Surg 1976;927–
48. Õunpuu S, Thompson JD, Davis RB, DeLuca PA. An examination 933.
of knee function during gait in children with myelomeningocele. J 72. Weinbach A. Contour maps, center of gravity, moments of inertia and
Pediatric Orthop 2000;20:629–635. surface area of the human body. Hum Bio 1938;10:356–371.
49. Palladino J, Davis RB. Biomechanics. In: Enderle J, Blanchard S, 73. Winter DA. Overall principle of lower limb support during stance
Bronzino J, eds. Introduction to Biomedical Engineering. Amslerdam: phase of gait. J Biomech 1980;13:923–927.
Elsevier Academic Press; 2005:127–210. 74. Winter DA. Use of kinetic analyses in the diagnostics of pathological
50. Paul JP. Forces transmitted by the joints in the human body. Proc Inst gait. Physiotherapy Canada 1981;33:209–214.
Mech Eng 1967;181:8–15. 75. Winter DA. Energy generation and absorption at the ankle and
51. Perry J. Distal rectus femoris transfer. Dev Med Child Neurol knee during fast, natural, and slow cadences. Clin Orthop Rel Res
1987;29:153–158. 1983;175:147–154.
52. Perry J. Gait Analysis: Normal and Pathological Function. Thorofare, 76. Winter DA. Kinematic and kinetic patterns in human gait: variability
NJ: Slack; 1992. and compensating effects. Hum Mov Sci 1984;3:51–76.
53. Piazza SJ, Delp SL. The influence of muscles on knee flexion during 77. Winter DA. Biomechanics and Motor Control of Human Movement.
the swing phase of gait. J Biomech 1996;29:723–733. New York: John Wiley and Sons; 2005;86–117.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

Chapter 4 / Kinetics of Normal Walking 75

78. Winter DA, Quanbury AO, Reimer GD. Analysis of instantaneous en- Force – The action of one object on another that tends to
ergy of normal gait. J Biomech 1976;9:253–257. change the state of rest or state of motion of the object to
79. Winter DA, Robertson DGE. Joint torque and energy patterns in nor-
mal gait. Biol Cybernet 1978;29:137–142. which it is applied; a vector quantity with characteristics
80. Winter DA, Sidwell HG, Hobson DA. Measurement and reduction of of magnitude, direction, and point of application (or line
noise in kinematics of locomotion. J Biomech 1974;7:157–159. of action). Unit of measure: Newtons (N).
81. Woltring HJ. On smoothing and derivative estimation from noisy
displacement data in biomechanics. Hum Mov Sci 1985;4:229– Concentrated force – The effect of a distributed force
245. acting at a point of application.
82. Wood GA. Data smoothing and differentiation procedures in biome- Distributed force – Forces acting between two bodies
chanics. Exerc Sport Sci Rev 1982;10:308–362.
83. Yamaguchi GT. Dynamic Modeling of Musculoskeletal Motion. that are dispersed over the contact area shared by the
Boston: Kluwer Academic Publishers; 2001;3–21. two bodies, e.g., pressure under the foot during gait.
84. Yamaguchi GT, Zajac FE. Restoring unassisted natural gait to para- Unit of measure: Newtons per meter squared (N/m2 )
plegics via functional neuromuscular stimulation: a computer simu-
lation study. IEEE Trans Biomed Eng 1991;37:886–902. or Newtons per centimeter squared (N/cm2 ). See also
85. Zajac FE, Neptune RR, Kaytz SA. Biomechanics and muscular co- pressure.
ordination of human walking: Part I: Introduction to concepts, External force – A force that is applied between the
power transfer, dynamics and simulations. Gait Posture 2002;16:215–
232. body and its environment. Examples include ground
reaction force, body (or body segment) weight, and
inertial force.
Ground reaction force – A concentrated force that rep-
APPENDIX resents the summation of the distributed forces (e.g.,
Glossary of Gait Kinetics Terminology ∗ pressure) applied to the plantar surface of the foot
at the center of pressure during activities such as
Acceleration – The rate of change of velocity with respect
gait.
to time; a vector quantity with characteristics of magni-
Inertial force – A force that is associated with the ac-
tude and direction.
celeration of an object. This force resists a change in
Angular acceleration – Acceleration associated with
motion and consequently, it is directed opposite to the
rotational displacement. Unit of measure: radians per
acceleration of the object.
second squared (rad/s2 ).
Internal force – A force that is produced within the
Linear acceleration – Acceleration associated with
body. Examples include muscle force, ligamentous
translational displacement. Unit of measure: meters
force, and joint contact force.
per second squared (m/s2 ).
Intersegmental force – A force acting across the joint
Body – In mechanics, a generic term that refers to the ob- that dynamically balances the body segment under
ject under analysis (e.g., the entire human body or just a investigation. The magnitude of the intersegmental
particular body segment such as the foot); in biomechan- force underestimates the joint contact force, i.e., it
ics, it generally refers to the entire human body. does not reflect the entire compressive load applied
Rigid body – An object whose deformation under the to the joint.
application of a set of forces may be considered negli- Joint contact force – The compressive force applied to
gible, i.e., the distance between any two points in the the joint, including all forces produced by muscles
object is assumed to remain constant. crossing the joint and those caused by gravity and any
Center of gravity – The point through which the concen- movement inertia.
trated gravitational force (weight) of the body or a body Normal force – A force applied perpendicular to a sur-
segment passes. In a uniform gravitational field, the cen- face, e.g., the vertical component of the ground reac-
ter of gravity and the center of mass are located at the tion force.
same point. Shear force – A force applied parallel to a surface, e.g.,
the anterior-posterior component of the ground reac-
Center of mass – The point where all of the mass of the
tion force.
body or a body segment could be concentrated and still
have the same mechanical effect. In a uniform gravita- Force couple – The torque produced by two equal and op-
tional field, the center of mass and the center of gravity posite forces that are noncollinear (not lying along the
are located at the same point. same line of action).
Center of pressure – The point of application of the Forward dynamics – A process by which movement kine-
ground reaction force. matics are predicted through the application of govern-
ing equations of motion that incorporate assumed (or
derived) muscle activation patterns and estimated (e.g.,
∗
Readers are also referred to the engineering mechanics text by segment inertial properties, muscle properties, muscu-
Meriam and Kraige (38). loskeletal geometry) quantities.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-04 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:11

76 Chapter 4 / Kinetics of Normal Walking

Inertia – A resistance to a change in motion, i.e., velocity. In two dimensional analyses, a moment is the product of
See also mass and mass moment of inertia. the magnitude of the force and the moment arm. Unit of
Inverse dynamics – A process by which intersegmental measure: Newton-meter (N·m).
forces and moments are calculated through the appli- External moment – The moment produced by an ex-
cation of governing equations (e.g., Newton’s equations ternal force.
of motion) that incorporate measured (e.g., kinematics, Internal Intersegmental moment – The moment act-
ground reactions) and estimated (e.g., segment inertial ing about the joint that dynamically balances the body
properties) quantities. segment under investigation; sometimes referred to as
“joint moment” or “muscle moment.”
Kinematics – The study of motion without reference to the
forces that cause that motion. See also kinetics. Moment arm – The shortest (perpendicular) distance be-
tween the line of action of a force and an axis of rotation.
Kinetics – The study of motion that relates the action of Synonymous with lever arm.
forces to motion. See also kinematics.
Momentum – The resistance of a moving object to accel-
Lever arm – Synonymous with moment arm. eration; defined mechanically as the product of the mass
Mass – A measure of the resistance of an object to a change of an object and its linear velocity or the mass moment
in translational motion, i.e., its translational inertia. Unit of inertia of an object and its angular velocity. See also
of measure: kilogram (kg). inertia.
Mass moment of inertia – A measure of the resistance of Pressure – Distributed normal forces. Unit of measure:
an object to a change in rotational motion, i.e., its rota- Newtons per meter squared (N/m2 ) or Newtons per cen-
tional inertia. The mass of the object and the distribu- timeter squared (N/cm2 ).
tion of that mass about particular axes passing through Power – The rate at which work is done on or by a system;
the object determine the magnitude of the mass moment the rate of change in mechanical energy in a system per
of inertia. Unit of measure: kilogram-meters squared unit time. Unit of measure: Watt (W).
(kg·m2 ). Intersegmental power – The rate at which work is
Mechanical energy – The capacity of an object or a done on or by an intersegmental moment.
system to do work; the sum of kinetic and potential Torque – The moment produced by a force couple. Note
energy. that a torque or force couple can only produce rotation
Kinetic energy – The energy associated with motion. of an object. Unit of measure: Newton-meter (N·m).
Kinetic energy combines either the mass of the object Ground reaction torque – The vertical torque pro-
and its translational velocity or the mass moment of duced by friction between the foot and ground.
inertia of the object and its rotational velocity. Unit of
Velocity – The rate of change of displacement with respect
measure: Joule (J).
to time; a vector quantity with characteristics of magni-
Potential energy (gravitational) – The energy associ-
tude and direction.
ated with work done against or by gravity. Gravita-
Angular velocity – Velocity associated with rotational
tional potential energy combines the weight of the
displacement. Unit of measure: radians per second
object and the height of its center of mass above
(rad/s).
some reference height or datum. Unit of measure:
Joule (J). Linear velocity – Velocity associated with translational
displacement. Unit of measure: meters per second
Potential energy (elastic) – The energy associated
(m/s).
with work done against or by the force of elastic el-
ements within the object, e.g., passive muscle tissue, Weight – The force on an object due to gravity represented
ligaments. Elastic potential energy combines the force as a concentrated force passing through the center of
of the elastic structure and the distance that the force gravity of the object. Weight is the product of the mass
is displaced along its line of action. Unit of measure: of an object and the acceleration due to gravity (approx-
Joule (J). imately 9.81 m/s2 on the surface of the Earth). Unit of
Moment – The tendency of a force to rotate an object about measure: Newton (N).
an axis; a vector quantity with characteristics of magni- Work – The change in energy associated with the linear
tude and direction. A moment is the vector cross product displacement of a force (along its line of action) or the
of the position vector from any point on the axis to any rotational displacement of a moment. Unit of measure:
point on the line of action of the force and the force vector. Joule (J).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:13

• C h a p t e r 5
•
◗ Energetics of Walking
Jessica Rose, Don W. Morgan, and James G. Gamble

People naturally walk in a manner that conserves energy. vast majority of chemical energy from food is released as
To accomplish this, limb and trunk motions are integrated heat (Fig. 5-1). Measuring the heat production of the body
to smooth the forward progression of the body and de- closely approximates the total amount of energy utilized
crease the displacement of the center of gravity. Walking in that state.
speeds are selected which minimize the energy expended During exercise, skeletal muscle uses the chemical en-
per distance walked. Deviation from the normal walking ergy from the hydrolysis of ATP to produce mechanical and
pattern increases energy expenditure and limits ambula- chemical work. In the process of this irreversible trans-
tion. Thus, for a person with a disability, energy expen- duction, movement occurs and energy is dissipated in the
diture is a primary consideration in decisions regarding form of heat. Part of this heat arises from friction of the
therapeutic treatment. mechanochemical coupling that produces physical work
(Fig. 5-1) (15,93). When performing physical work, human
energy efficiency approaches 30%; 70% of the energy uti-
METABOLIC ENERGY lized is ultimately released as heat (15). Humans are ∼24%
efficient while walking at a comfortable speed. At a slow
The human body utilizes metabolic energy in the form of walking speed, efficiency decreases to 14%. This is compa-
ATP to support physiologic processes, such as muscle con- rable with the internal combustion engine of an automo-
traction and relaxation, and for cellular reactions such as bile, which is only 10% to 20% efficient at converting the
active transport systems, hormone receptor signaling, and chemical energy of gasoline to heat and transducing heat
protein synthesis. The fuel for the body’s energy require- into the mechanical energy of torque at the drive shaft
ments comes from ingested nutrients or from stored glyco- (90).
gen and fat.
Metabolic Energy Storage
Energy Transductions in the Body and Utilization
Cells convert energy stored in the chemical bonds of food Cellular reactions, including skeletal muscle contraction,
sources to free energy that can be used for the molecular utilize chemical energy derived from ingested carbohy-
reactions of cellular metabolism. These physiologic pro- drates (glucose) and fat (lipids). Carbohydrates provide 4.2
cesses involve several forms of energy transduction, from kcal/g and lipids 9.5 kcal/g of energy. Heat of combustion
chemical energy of molecular bonds to thermal energy of is the same whether it occurs inside or outside the body
heat or to mechanical work of organ movements such as except for protein, which provides 26% more kilocalories
that which occurs in the heart and lungs. Mechanical en- when combusted outside the body (owing to the mecha-
ergy is used at a cellular level as well. For instance, a sin- nism of amino acid metabolism). While protein is not gen-
gle microtubule transport system within the cell generates erally considered a primary energy source, protein use can
2.6 × 10−7 dynes (3). All metabolic energy transductions increase during extreme conditions such as starvation and
obey the first law of thermodynamics, which states that en- during hard and protracted exercise (48). Ingested protein
ergy is conserved as it is interconverted. This means that provides essential amino acids that are required to synthe-
at rest, the chemical energy from blood glucose can be re- size enzymes and other cellular proteins.
leased as heat (thermal energy) or converted to the chem- The chemical energy of carbohydrates and fats is not
ical and mechanical energy of resting metabolism that is used directly for metabolic work, but is first transferred
ultimately released as heat (15). Only about 1% of the rest- to the high-energy phosphate bond of ATP. Energy in the
ing energy is converted directly to pressure-volume work terminal phosphate bond is available for molecular synthe-
on the atmosphere, such as exhalation. Thus, at rest, the sis, active transport, muscle contraction, and other cellular

77
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

78 Chapter 5 / Energetics of Walking

to drive the synthesis of ATP from ADP and inorganic phos-

phate (39).
For each pair of electrons transferred to oxygen in the
electron transport chain, sufficient free energy is released
to form approximately three ATP molecules. The complete
oxidation of a glucose molecule to carbon dioxide and wa-
ter yields a net total of 36 or 38 ATP molecules, depend-
ing on the location of ATP synthesis (48). This compares
quite favorably with the net gain of two ATP molecules per
glucose molecule resulting from anaerobic metabolism.
Oxidative (aerobic) ATP production provides most of the
body’s energy needs at rest and at submaximal workloads.
Energy for muscular contraction predominantly comes
from aerobic metabolism that involves the continuous ox-
idation of metabolites in the mitochondria. Only a small
amount of the total energy is provided by phosphagen
and glycolytic metabolism. In the absence of oxygen, ATP
production occurs exclusively in the cytosol of the cell.
These anaerobic sources of energy are primarily available
FIGURE 5-1. Energy transductions in the body.
for short-term, high-intensity work. At submaximal work-
loads, such as walking on a level surface, the ATP produced
reactions. The terminal phosphate bond of ATP has a high in the cytosol represents a very small amount of the total
free energy of hydrolysis, designated as a “high-energy” ATP produced (15).
phosphate bond. Current estimates of the free energy of In the absence of oxygen, pyruvate from glucose
hydrolysis per phosphate bond for physiologic conditions metabolism is reduced to lactic acid. At high work rates,
(such as those occurring in contracting muscle) are as high lactic acid is produced faster than it can be utilized, and
as 12–14 kcal/mole (112). lactate accumulation causes intracellular and extracellular
Muscle utilizes phosphate bond energy to slide the actin acid-base imbalance with unfavorable consequences such
and myosin protein filaments relative to each other, lead- as muscle fatigue and hyperventilation. At higher work
ing to shortening and increased tension of the muscle (20). rates, both aerobic and anaerobic ATP production share
In the presence of intracellular Ca2+ , one phosphate bond in energy generation. Anaerobic metabolism provides an
is cleaved by myosin ATPase per actin-myosin protein in- increasing proportion of the energy requirements at higher
teraction. Additional ATP is required to pump Ca2+ back work rates (15).
into the sarcoplasmic reticulum, disengaging the proteins
and allowing muscular relaxation. An average person at
rest requires 72 kcal per hour, utilizing six mol ATP every MEASUREMENT AND UNITS
hour. Level walking at 80 m per minute requires 240 kcal
per hour or 18 mol ATP. Historically, the measurement of heat production and the
changes in gas content of respired air were used to inves-
tigate metabolism. In the 1780s, Lavoisier studied the cen-
Oxidative and Glycolytic Metabolism
tral role of oxygen in metabolism (43). He compared the
Under normal conditions, ATP is produced by glycolytic combustion of food inside and outside the body and rec-
and oxidative reactions in the cytosol and mitochondria ognized that the same amount of heat was released during
of the cell. Carbohydrates are metabolized to pyruvate in the two processes. Lavoisier named the element that took
the cytosol by a series of glycolytic reactions. Glycolysis is up approximately one-fifth of the volume of air “oxygen,”
the source of anaerobic energy, yielding 2 mol ATP/mol glu- which means “acid former” because carbonic acid, sulfu-
cose. Fats are metabolized in the cytosol to free fatty acids. ric acid, or phosphoric acid were produced when organic
Acetyl coenzyme A, derived mostly from pyruvate and free substances were burned (43).
fatty acids, is the initial substrate for the Krebs Cycle. In the Measurement of the body’s heat production, also known
mitochondria, electrons and hydrogen ions released from as calorimetry, is an indication of the overall metabolic
metabolic substrates are transferred to nicotinamide and rate. Heat production can be measured directly (direct
flavin adenine dinucleotide, which act as hydrogen accep- calorimetry) or it can be estimated from the oxygen
tors and electron carriers, delivering the reducing equiv- consumed or from the carbon dioxide and water pro-
alents to the electron transport chain. The electron trans- duced (indirect calorimetry). The ultimate source of en-
port chain creates an electrochemical gradient across the ergy for metabolic and physical work is the combustion
inner mitochondrial membrane and this gradient is used of food, i.e., oxidation of nutrients. Measurement of the
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 79

oxygen consumed by the body provides a measure of the per minute.

energy expenditure of the body. The units for the mea-
surement of oxygen uptake can be expressed in absolute V02 /kg = 5.9 + 0.0011 v 2
(e.g., liters per minute) or relative (e.g., mL/kg per minute) Multiplying by the conversion factor 5 yields:
terms.
Oxygen uptake units can be transferred into heat units cal/kg per minute = 29.5 + 0.0055 v 2
such as the gram-calorie (cal) or the kilogram-calorie
(kcal). The gram-calorie is the amount of heat necessary to
raise the temperature of 1 g of water 1 degree C, whereas LYING, SITTING, AND STANDING
the kilocalorie is the amount of heat necessary to raise the
temperature of 1 kg of water 1 degree C. Expressing oxygen It is important, particularly when dealing with the effects
uptake in kilocalories allows comparison among nutrient of various types of disabilities, to compare the energy de-
energy, metabolic energy, and mechanical energy. mand of walking with that of an immobile (or relatively
In discussing metabolic energy expenditure, it is desir- immobile) state, such as lying, sitting, or quiet standing.
able to use the same units. However, this is not always fea-
sible. In calculations that involve measurements of speed
(m per minute) it is preferable to express energy expendi- Lying
ture in cal/kg per minute; however, kilocalories per hour, The “basal metabolic rate” is frequently referred to as
kilocalories per minute, and gram-calories per kilogram though, in the words of Durnin and Passmore, “it is a fun-
per minute body weight have also been used. damental biological property of the individual or species”
When dealing with mechanical energy levels and work, (29). It is preferable to refer to a resting rate, measured
joules (J) are the preferred units. For work rate, the units 3 to 4 hours after an ordinary meal. This measurement
are watts (W). Horse power (hp) is used occasionally in does not significantly differ from a basal rate, requires no
reference to work rate. Table 5-1 lists some commonly used special standardization of environmental and other condi-
conversion relationships. tions, and, most importantly, does not assume that a “fixed
In 1949, Weir (113) determined a method for expressing minimum metabolism on which life depends” is involved.
energy expenditure in units of kilocalories and calories. Resting metabolism has been shown to be indistinguish-
Using Weir’s mathematical analysis, Passmore and Draper able from basal metabolism (29). This is demonstrated in
(79) recommend the use of the following relation between Table 5-2.
energy expenditure (E) in kilocalories per minute and the Column 5 of Table 5-2 is the predicted value of the
oxygen percentage concentration of expired air (0e ): resting metabolic rate, according to tables from Durnin
and Passmore (29), for average male and female subjects.
4.92 V These authors divided subjects into four categories (thin,
E= (20.93 − 0e )
100 average, plump, and fat) corresponding to percent fat of
5, 10, 15, and 20% for males and 15, 20, 25, and 30% for
Where V is volume of expired air in liters per minute. females. These authors also show that the differences be-
This can be simplified to the following conversions: tween male and female subjects in the values of the basal
and resting metabolic rates are eliminated when lean body
5 kcal per minute = 1 liter oxygen per minute
mass is used instead of total body mass in the presentation
5 cal per minute = 1 mL oxygen per minute of metabolic values.
Column 6 of Table 5-2 is the predicted value of metabolic
As a somewhat more complex translation of oxygen val- rates as calculated from Kleiber’s (43) equation:
ues into calories, consider the following equation relating
oxygen consumption in milliliters per minute (V02 ) per metabolic rate (kcal/hr) = 3 (w)3/4
kilogram body weight to speed of walking (v) in meters
where w is gross body weight (kg). For female subjects,
the coefficient is scaled down to 2.775 from 3, which is
◗ TABLE 5-1 Common Conversion Relationships proportionally the same as the coefficients 37 and 40 for
the values given in column 5 of Table 5-2.
To Change To Multiply By
Column 7 of Table 5-2 is the value for the basal
cal J 4.186 metabolic rate as predicted by the Mayo Foundation Nor-
kcal J 4186 mal Standards (12) based on body surface area for young
kcal kg-m 427 males.
kcal/hr W 1.16
W cal/min 14.33 basic metabolic rate = 40 kcal/m2 /hr
hp W 746
The coefficient is 37 kcal/m2 /hr for females.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:13

80 Chapter 5 / Energetics of Walking

◗ TABLE 5-2 Predicted Resting Values Compared with Experimental Values for
Quiet Standing
Predicted Resting (kcal/hr)
Mean Resting Standing
Durnin and (kcal/hr) (kcal/hr) Standing:
Subject Sex Height (m) Weight (kg) Passmorea Kleiber b Mayo c (Predicted) d (Experimental) e Resting

1 M 1.74 69 71.8 71.9 73.0 72.2 86.9 1.20

2 M 1.72 64 67.7 68.0 70.2 68.6 87.0 1.27
3 M 1.80 64 67.7 68.0 72.5 69.4 96.3 1.39
4 M 1.70 67 70.1 70.5 71.0 70.5 89.8 1.27
5 M 1.70 66 69.2 69.8 70.6 69.9 79.0 1.13
6 M 1.73 63 66.2 67.3 70.1 67.9 82.4 1.21
7 M 1.96 87 85.4 f 85.5 87.8 86.2 126.1 1.46
8 M 1.72 68 70.9 71.0 71.9 71.3 96.2 1.35
9 M 1.89 83 82.6 f 82.5 83.8 83.0 97.5 1.17
10 M 1.82 79 79.7 79.5 79.9 79.7 115.2 1.45
11 M 1.68 62 66.2 66.3 68.0 66.8 89.8 1.34
12 M 1.79 80 80.4 80.2 79.3 80.0 101.8 1.27
13 M 1.80 84 83.3 f 83.2 81.3 82.6 99.6 1.21
14 M 1.79 71 73.4 73.4 75.4 74.1 92.8 1.25
15 M 1.79 79 79.7 79.5 78.9 79.4 94.2 1.19
16 M 1.87 69 71.8 71.8 76.9 73.5 92.7 1.26
17 M 1.84 75 76.8 76.5 78.8 77.4 101.1 1.31
Means 1.78 72 74.3 74.4 75.8 74.9 95.8 1.28
18 F 1.71 61 59.6 60.6 63.3 61.2 72.6 1.19
19 F 1.68 59 58.0 59.1 61.6 59.6 77.0 1.29
20 F 1.64 68 65.2 65.7 64.3 65.1 70.3 1.08
21 F 1.73 78 73.3 72.8 70.8 72.3 79.1 1.09
22 F 1.72 64 62.2 62.8 64.9 63.3 69.1 1.09
23 F 1.77 67 64.4 65.0 67.5 65.6 78.8 1.20
24 F 1.70 64 62.2 62.8 64.3 63.1 79.3 1.26
25 F 1.64 57 56.3 57.6 59.6 57.8 76.3 1.32
26 F 1.78 73 69.1 69.3 70.3 69.6 86.8 1.25
27 F 1.66 62 60.5 61.3 62.4 61.4 72.4 1.18
28 F 1.59 53 52.9 54.5 56.5 54.6g 49.4g 0.90g
29 F 1.62 56 55.4 56.8 58.7 57.0 79.9 1.40
30 F 1.70 59 58.0 59.1 62.1 59.7 66.4 1.11
31 F 1.65 70 66.6 66.9 65.2 66.2 78.3 1.18
Means 1.69 64 61.7 62.5 63.7 62.6 75.6 1.20
a Durnin and Passmore (1967), page 31. Values refer to “average” physique.
b Kleiber equation (1961). Coefficient 3 for males scaled down to 2.775 for females.
c Mayo Foundation standards (Boothby et al., 1936). Coefficient 40 for males and 37 for females.
d Mean of three preceding columns.
e Some experimental data from Molen and Rozendal (1967).
f Extrapolated. Other values are either direct or interpolated.
g Values are obviously incompatible and are unexplained. Not used in calculation of mean ratio.

It makes little difference in the results whether the val- Table 5-3 shows resting metabolic rates for 10 male sub-
ues for metabolic rate are calculated according to Durnin jects, measured in our laboratory, compared with values
and Passmore (29), Kleiber (43), or Boothby et al. (12). Not predicted by Durnin and Passmore (29). Except for sub-
only is the claim by Durnin and Passmore that basal and ject 1, the agreement is well within acceptable limits of
resting metabolic rates are virtually indistinguishable fully prediction. The discrepancy in the case of Subject 1 is due,
justified, but the Durnin and Passmore table also provides at least in part, to age.
a means of expressing metabolic rates in terms of lean body
mass, which is a notable improvement over earlier predic-
tive tables. Column 8 of Table 5-2 is the average value of
Sitting
columns 5 through 7. These values, recalculated as cal/kg Durnin and Passmore (29) measured the energy expendi-
per min, will be compared later in this chapter with the ture of seated men and women while they were engaged
values of sitting and standing. in activities such as reading and watching television. The
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:13

Chapter 5 / Energetics of Walking 81

◗ TABLE 5-3 Experimental Values of Resting ± SD for males is 22.1 ± 1.68 cal/kg per minute and for
Metabolism in 10 Male Subjects females 19.9± 2.60, with the ratio of means being 1.11.
Compared with Values Predicted by Molen and Rozendal (65) noted the lower energy de-
Durnin and Passmore (1968)a mand of standing in females and indicated that this in-
Age Weight Experimental Predicted fluenced the finding that the cost of walking was slightly
Subject (yr) (kg) (kcal/min) (kcal/min) Ratio less in females than in males. However, the difference is
so small that Molen and Rozendal (65) combined men and
1 53 70.9 0.89 1.17 1.31 women in formulating the regression equation relating en-
2 45 77.7 1.28 1.32 1.03 ergy expenditure to speed.
3 31 87.7 1.24 1.41a 1.14
Note that the difference in energy cost of standing for
4 26 58.6 1.04 1.07 1.03
5 23 72.7 1.21 1.25 1.03
males compared with females is eliminated if lean body
6 22 70.5 1.33 1.22 0.92 weight instead of gross body weight is used. The results in
7 22 64.3 1.25 1.13 0.90 this case using the values given by Durnin and Passmore
8 18 68.2 1.05 1.18 1.12 (29) for body fat content become 24.5 cal/kg per minute
9 17 62.3 1.19 1.10 0.92 for males and 24.8 cal/kg per minute for females.
10 18 80.9 1.38 1.35a 0.98 The ratio of standing to resting is 21.0:16.8 (i.e., 1.25)
Means 1.19 1.22 1.04 if males and females are combined and values expressed
a Extrapolated values. Other values in this column are direct or interpolated.
as calories/kilogram per minute, derived from Table 5-2,
columns 8 and 9. This ratio is similar to that for resting to
active sitting (1.22). The low cost of standing reflects low
measurements were made at various times of day and were activity of the postural muscles during standing, a finding
unrelated to food intake. Table 5-4 shows the average re- corroborated by electromyography.
sults recalculated to express energy expenditure.
Their average values, which range from 19.5 to 21.4
cal/kg/min, are substantially higher than the average val- ENERGY EXPENDITURE OF WALKING
ues for resting metabolism, derived from column 8 of
Table 5-2, that equal 17.2 cal/kg per minute for males A modest increase in energy expenditure, on the order of
and 16.4 cal/kg per minute for females. As will be seen 25%, occurs when a subject assumes a quiet standing po-
below, their sitting values are practically the same as sition compared with a supine position. As soon as an in-
the figures for quiet standing derived from column 9 of dividual beings to walk, a great increase in energy expen-
Table 5-2. This probably reflects the rather “active” sitting diture occurs, reflecting the metabolic cost to the muscles
engaged in by the subjects. If males and females are com- of moving the body against gravity and of accelerating and
bined, the ratio of active sitting to resting metabolism is decelerating the various body segments.
about 1.22.

Relation between Energy Expenditure

Standing and Speed in Level Walking
Molen and Rozendal (65) studied 10 males and 10 females Ralston (84) showed that a quadratic equation:
under conditions of quiet standing in the laboratory. Their
results were virtually identical with those obtained by Ral-
ston on seven males and four females and are combined, Ew = b + m v 2
therefore, with the data from Ralston for the 17 males and
14 females shown in Table 5-2, column 9. The average value adequately predicted the energy cost of walking for adults
at speeds up to ∼100 m per minute (Fig. 5-2). Ew is en-
ergy expenditure in cal/kg per minute, v is speed in m per
◗ TABLE 5-4 Sitting Energy Expenditure Based on minute, and b and m are constants. In deriving this equa-
Data by Durnin and Passmore (1967) tion, Ralston used data from various investigators. Since
that time, other investigators have found that an equa-
Age Range Mean ± SD
(years) Sex (cal/kg/min) SD/Mean tion of the same form provides an acceptable basis for
predicting energy-speed relations (11,22,24,90). Combin-
20–39 M 21.4 ± 3.85 0.18 ing their equations, using weighted averages of the con-
20–39 F 20.9 ± 5.09 0.24 stants according to number of subjects (86 total: 57 males,
40–64 M 21.1 ± 4.46 0.21 29 females), yields a grand average equation:
40–59 F 19.5 ± 3.45 0.18
65+ M 19.8 ± 3.85 0.19
60+ F 19.8 ± 5.64 0.28 Ew = 32 + 0.005 v 2 (1)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

82 Chapter 5 / Energetics of Walking

predicted values from Equation 1 and average values from

Durnin and Passmore. Because the quadratic Equation 1
implies that the energy cost is proportional to body weight,
Wyndham et al. (120) examined this relation in detail and
found that any error involved in such a relation was not
significant in the case of walking.
It is clear that over the range of 55 to 90 m per minute,
the predicted results in the last two columns of Table 5-5
are virtually identical. At 110 m per minute, the Durnin
and Passmore figures underestimate the average experi-
mental values (Ew = 93) by 11%, whereas Equation 1 is
still in good agreement with observed results (Fig. 5-2).
At still higher speeds, up to about 145 m/min, which is
about the top speed for natural walking, Equation 1 also
fails.
Zarrugh et al. (118) deduced a more general, hyperbolic
FIGURE 5-2. Energy expenditure during walking. Top curve,
equation for predicting energy expenditure during walk-
Em (cal/kg/m), calculated from Equation 4; middle curve, Ew ing, which considers step length and step rate:
(cal/kg per minute), calculated from Equation 3; bottom curve,
Ew (cal/kg per minute, calculated from Equation 1. (Adapted with Eo
permission from Zarrugh MY, Todd FN, Ralston HJ. Optimiza- Ew =


2
tion of energy expenditure during level walking. Eur J Appl Phys- S n2 (2)
1− 2 1− 2
iol 1974;33:293–306). Sµ nµ

where s is step length in meters, n is step rate in steps/

min, Eo is the value of Ew when s = n = 0 (resting), Su is
Note that for a 70-kg adult male, the increase in energy
the upper limit of s as Ew approaches infinity, and nu is
expenditure of walking at 80 m/min (Table 5-5) over resting
the upper limit of n as Ew approaches infinity.
values (derived from Table 5-2, column 4) is ∼ 3.7. This
In natural walking, where the subject adopts his own
increase is more than three times the ratio of active sitting
natural cadence for a particular speed, it can be shown
or quiet standing over resting values.
that Equation 2 reduces to
The previously mentioned studies included walking on
treadmill, floor, firm path, and grass surfaces. There was no
Eo
indication of systematic differences for the various walking Ew = (3)
surfaces. In such studies, however, close attention must be (1 − v/vu )2
paid to the nature of the footwear. A heavy boot or slippery where v is speed and vu is upper limit of v, equal to nu su
sole, for example, may alter the energy demand of walking. (Fig. 5-2).
Durnin and Passmore (29) provide a table for relating Eo has an average value of ∼28 cal/kg per minute and
weight of subject, speed, and energy expenditure over a Vu an average value of ∼240 m per minute. Thus, for a top
range of 55–110 m/min (2–4 miles/hr). Table 5-5 shows the natural speed of 145 m per minute,

28
Ew = = 179 cal/kg/min
◗ TABLE 5-5 Energy Expenditure during Level (1 − 145/240)2
Walking as Predicted by Equation 1
and by Durnin and Passmore (1967) which is in good agreement with average experimental val-
ues.
Durnin and Up to speeds of about 100 m per minute, Equations 1
Passmore
m/min km/hr mile/hr E w = 32 + 0.0050v 2
(1967) and 3 predict virtually identical values. Therefore Equation
1 may be used for most cases of natural walking.
25 1.50 1.0 35 Rather unexpectedly, Equation 1 predicts energy cost
40 2.40 1.5 40 within about 10% for competitive race walking, which is
55 3.30 2.0 47 46 quite unlike natural walking. Menier and Pugh (63) pro-
70 4.20 2.5 56 55 vide data on four male Olympic race walkers studied dur-
80 4.80 3.0 64 64 ing treadmill walking at an altitude of 1800 m. They state
90 5.40 3.5 73 74
that their findings agree with studies made at sea level by
110 6.60 4.0 93 85
other investigators.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 83

At a speed of 14 km/hr (233 m per minute) the walkers energy expenditure throughout the entire range of natural
clustered closely ∼ 60 mL oxygen/kg per minute, corre- walking speeds.
sponding to 300 cal/kg per minute. Equation 1 predicts:

Ew = 32 + 0.005 (233)2 = 303 cal/kg per minute Age

which is within 1% of the observed value. None of the previously mentioned studies found a system-
At a speed of about 135 m per minute, the walkers clus- atic effect of age on the energy expenditure of natural walk-
tered about a value of ∼ 25 mL oxygen/kg per minute, cor- ing, at least within an age range of 20 to 59 years, when
responding to 125 cal/kg per minute. Equation 1 predicts: expressed per unit body weight.
Waters et al. (108) compared the energy expenditure
Ew = 32 + 0.005 (135)2 = 123 cal/kg per minute of self-paced walking in young adults aged 20 to 59 years
and seniors aged 60 to 80 years. They found no signifi-
which is within about 2% of the observed value. cant difference in oxygen uptake (mL/kg per minute) with
Morgan et al. (71) recently developed a generalized age, although self-selected walking speed was slower in the
equation to predict oxygen uptake (mL/kg per minute) seniors.
while walking in children age 6 to 10 that takes into ac- Blessey (9) found that in adults, the energy expenditure
count both age and walking speed: of walking did not vary significantly with age. However,
VO2 = 24.9 + .0032(v)2 − .995(age) − .263(v) walking represented a higher percent of maximal capacity
in older adults aged 50 to 59 years (43%) compared with
where v is walking speed (m per minute) and age is in young adults aged 20 to 29 years (36%). This was because
years. In their longitudinal study, the children received of the decrease in maximal oxygen uptake per kilogram
extensive accommodation to treadmill walking on an an- that occurs with age in adults.
nual basis. Nearly all of the prediction errors associ- Martin et al. (61) examined the effects of age and phys-
ated with estimating V02 across a wide range of speeds ical activity on walking economy in young (18 to 28 years)
(40.2 to 107.2 m per minute) were less than 1 MET and older (66 to 86 years) men and women who were
(3.5 mL/kg per minute). Moreover, it was demonstrated sedentary or physically active. Results from their study
that use of adult or adolescent prediction equations to esti- indicated that older subjects showed an 8% higher mean
mate walking energy expenditure would result in an under- walking V02 (mL/kg/km) compared to younger subjects.
estimation of childrens’ V02 values. In adolescents, Walker Interestingly, physical activity levels did not have an influ-
and colleagues (107) developed and cross-validated gen- ence on walking economy. Although speculative, the au-
eralized equations for predicting relative V02 (mL/kg per thors suggested that the higher walking energy expendi-
minute) and caloric expenditure (kcal/kg per minute) dur- ture observed in the older adults may be linked to declines
ing horizontal walking and running. Their data indicated in the force-generating capacity of aging muscle that may
that in adolescents the association between walking speed require increased motor unit recruitment and the use of
and V02 was curvilinear for speeds varying from 67 to less-economical fast-twitch fibers.
134 m per minute. Malatesta et al. (55) compared self-paced walking in
adults aged 65 to adults aged 80. While they found no
significant differences in oxygen uptake at their preferred
Significance of Constant = 28 in
walking speed for the older group, the preferred walking
Equation 3
speed was significantly slower and walking at the preferred
At speed v = 0, Equation 3 yields Ew = 28 cal/ kg per speed required a higher fraction of the oxygen uptake
minute. This is much higher than the value for resting or corresponding to the ventilatory threshold. Malatesta
quiet standing. Ralston (84) found that the value for ex- et al. (54) compared oxygen uptake during walking on a
tremely slow walking, in five subjects, averaged 28.6 cal/kg treadmill at a range of walking speeds in adults aged 65
per minute. He concluded that the constant represented compared to adults aged 80. Across the different walking
the cost of maintaining the body in motion at a barely per- speeds, they found higher oxygen uptake values and
ceptible speed. greater gait instability in the older group. However, no
It is clear, therefore, that calculating the “net” cost of significant correlation was found between gait instability
walking, based on oxygen uptake, i.e., subtracting the rest- and higher energy expenditure.
ing value or the quietly standing value from the gross Waters et al. (109) compared energy expenditure of self
value, is neither theoretically nor practically justified. It paced walking in children and adolescents and found that
is of added interest that use of net values for energy ex- oxygen uptake (mL/kg per minute) was significantly higher
penditure during walking does not permit formulation of for children (15.3) than for adolescents (12.9), and the
Equation 2 and its derivative, Equation 3, which appear, walking speed was slightly but significantly slower for chil-
thus far, to be the only equations successfully predicting dren (70 vs. 73 m per minute).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

84 Chapter 5 / Energetics of Walking

Morgan et al. (70) examined age-related changes in en- than adults, with the magnitude of differences in econ-
ergy expenditure of walking in children. They studied a omy varying markedly among studies. His analysis also
range of walking speeds in children age 6 to 10 years and indicated that younger children are less economical than
found that for all walking speeds, oxygen uptake decreased older children and adolescents, with economy differences
yearly from age 6 to 8 years. When averaged across speeds, becoming larger as age-group comparisons become more
oxygen uptake was 27% higher for 6-year-olds compared pronounced. In this review, a number of potential factors
to 10-year-olds. In a more recent article, Morgan et al. (74) were identified that may help explain child-adult differ-
noted that in young children, changes in walking V02 with ences in relative walking energy costs. These include: (1)
age occurred in a fairly uniform manner. This finding sug- less efficient ventilation, (2) faster stride rates, (3) imma-
gests that in a group of typical healthy children, a trend to- ture gait patterns, (4) larger surface area to body mass
wards good, average, or below-average walking economy is ratios, (5) shorter stature, (6) an imbalance between body
manifested and generally maintained during a time period mass and leg muscle contraction speed, (7) more distal
characterized by marked increases in body size and varied distribution of mass in the lower extremities, (8) a greater
exposure to a wide spectrum of physical activity choices. reliance on fat as a metabolic substrate, and (9) a reduced
McCann and Adams (49) studied 184 children, ado- ability to use anaerobic energy sources.
lescents, adults, and seniors walking on a treadmill at a Taken together, these studies demonstrate that the en-
range of walking speeds. They observed that at each walk- ergy expended to walk per kilogram does not change signif-
ing speed, oxygen uptake per kilogram was higher in chil- icantly with age for adults until older age is attained. The
dren than in the older groups. They found that the higher energy expended to walk expressed per unit body mass,
mass-specific metabolic cost of children was explained by however, decreases with age for children and adolescents.
differences in standing metabolism and stature. Similarly,
DeJaeger et al. (27) reported that standing energy expendi-
Weight
ture was higher in children and was, on average, 43% lower
in young adults compared to children age 3 to 4 years. The aerobic demand of walking is typically expressed rel-
Skinner et al. (98) compared the energy expenditure of ative to total body mass, since this mass must be sup-
children and adolescents walking on a treadmill at speeds ported by the lower extremities and accelerated and de-
of 83 to 100 m per minute and inclines of 10% to 17.5%. celerated with each step. This ratio-scaling method of ad-
He found that oxygen uptake per kilogram decreased with justing for body mass, however, has been questioned by
age for a given workload. Higher oxygen uptake in children some investigators who have suggested that regression
may be explained by their increased resting metabolic rate, or allometrically-based scaling methods may be more ap-
small step length and higher step rate at a given workload; propriate for normalizing ambulatory energy demands
walking may represent a higher percent of maximum oxy- (89,114). In allometric scaling, V02 is expressed as a math-
gen uptake in children than in adolescents or adults. ematical function of a specific body dimension, such as
Astrand and Rodahl (4) compared children and adoles- body mass, according to the formula:
cents walking and running on a treadmill at various speeds.
y = a(x b )
He found that a typical 8-year-old running at 180 m per
minute was working at 90% of maximum oxygen uptake where y is the dependent variable (V02 ), x is the indepen-
and that a 16-year-old walking at the same speed worked at dent variable (body mass), b is the allometric coefficient
75% of maximum oxygen uptake. Astrand reported that an (or scaling factor), a is a proportionality coefficient, and
8-year-old can increase basal metabolic rate only 9.4 times a and b are derived by linear regression after logarithmic
in a maximal run of 5 minutes, but that a 17-year-old boy transformation of x and y variables to yield the following
can attain an aerobic power that is 13.5 times the basal equation:
metabolic rate.
log y = log a + b(log x)
Bar-Or (7) reported that children operate at a higher
percent of maximal oxygen uptake for a given workload While allometric scaling has been used to determine the
and that the higher energy cost in children cannot be ex- relationship between body size and aerobic power, the va-
plained merely by a difference in resting metabolism that lidity of this approach is reduced when small sample sizes
is only ∼1 to 2 mL/kg per minute, but rather by their rel- are studied. Moreover, the consistency of scaling factors
atively “wasteful” gait. Similarly, Astrand and Rodahl (4) derived using this technique has not been high (56) Al-
noted that small bodies are more metabolically active per though the question of how best to adjust for variation in
unit body weight than larger ones. However, they also indi- body size has yet to be resolved, ratio scaling continues to
cate that the child’s lower efficiency can partly be explained enjoy widespread support because it is a method of adjust-
by their high step rate that is energy-expensive per unit ing for variation in body size that is easy to understand and
time. interpret and involves simple mathematical computations.
In a comprehensive review of existing data, Morgan (68) For the measurement of resting and maximal metabolic
concluded that children are less metabolically economical rate, lean body mass (LBM) is more highly correlated to
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 85

oxygen uptake than total body weight (TBW) (4,29). Her- 20% weight gain over 16 weeks, and 4) return to normal
fenroeder and Schoene (35) studied the effect of body diet for 5 months with subject wearing backpack weight to
composition on maximum oxygen uptake in adolescents equal weight lost over 5 months. Oxygen uptake when ex-
and found the correlation between oxygen uptake and pressed in liters per minute increased with the total weight
TBW to be r = 0.71 and for oxygen uptake and LBM, r moved, but when expressed per unit weight moved, the
= 0.84. Katch (40) reported a correlation of 0.71 between value was relatively constant for a given work level.
TBW and oxygen uptake at basal and maximal metabolic These studies indicate that for walking, it is the total
rates. He notes that the use of the VO2 -to-kilogram ratio weight moved that correlates most highly with energy ex-
implies a correlation of r = 1.0 and direct proportional- penditure. However, the effects of loading the body depend
ity between VO2 and body weight and should be corrected on the nature of the loading (44). It has been shown that the
for when extrapolating energy expenditure from oxygen relative increase in the absolute cost of running per kg of
uptake in units per body weight. added mass is about 1% for added mass to the trunk, 3.5%
Although LBM is highly correlated with aerobic capac- for added mass to the thighs, and 7.0% for added mass
ity, most studies show that for submaximal activities such to the feet (60). Lloyd and Cooke (46) found that oxygen
as walking, TBW correlates more highly with oxygen up- uptake while walking was 5% lower using a backpack that
take than LBM (64,94). Similarly, Volpe and Bar-Or (106) allowed the load to be distributed between the back and
studied lean and obese adolescents matched for TBW and front of the trunk during uphill walking.
found that TBW, and not adiposity, was the main predictor Loads placed on the distal segments, especially the foot,
of walking energy cost. have a much greater effect than loads attached to the trunk,
Datta et al. (26) found that it was the total weight moved because of the large inertial effects associated with accel-
by subjects carrying loads of 0-50 kg on their head that eration and deceleration of the limb. Figure 5-3 shows the
correlated most highly with oxygen uptake, provided the effect of 2 kg on each foot on the metabolic cost of walking
load was not excessive or carried in an awkward position. at +2◦ , 0◦ , and –2◦ . At 0◦ , the value of Ew (cal/kg per minute)
Goldman and Iampietro (32) measured oxygen uptake in is increased by ∼30%, with similar results at +2◦ and −2◦ .
five subjects with various combinations of treadmill speeds Contrast this to the very small effect of 5 kg attached to the
and grades and backpack loads. They found that at a given trunk, where in seven subjects, the mean increase in Ew
speed and grade, the energy cost per unit weight (kcal/kg was only 4%.
per minute) was similar regardless of the distribution of A comparable experiment on lower leg loading is shown
total weight between body weight and load weight within in Figure 5-4. Unfortunately, this type of experiment is not
a range of 1-30 kg. Robertson et al. (88) also found this to
be true when oxygen uptake was expressed per kilogram
total weight moved for subjects walking and carrying loads
up to 15% of body weight.
Turrell et al. (103) and Bloom and Eidex (10) compared
the energy expenditure of walking for lean and obese sub-
jects. Although energy expenditure per unit time (l per
minute) was greater for the obese subjects, dividing by
TBW (mL/kg per minute) gave a similar value. Bloom and
Eidex (10) and Turrell et al. (103) compared lean subjects
carrying weight with obese subjects with a similar total
weight. They found that it was the total weight that was
most highly correlated with oxygen uptake. For example,
Bloom and Eidex (10) measured the energy cost of an obese
250-pound subject and a lean 150-pound subject carrying a
100-pound backpack. Dividing by total weight moved gave
essentially the same energy cost in all cases. Alternatively,
Myo-Thein et al. (77) found that predicting energy expen-
diture for trunk loads based on body weight led to overes-
timation of energy expenditure.
Most studies of the effects of TBW and/or obesity on the
energy cost of walking have been cross-sectional, i.e., com-
paring obese subjects with lean individuals at one point.
However, Hansen (34) studied the energy cost of treadmill
walking in male volunteers under four conditions: 1) con- FIGURE 5-3. Effects of loading foot during walking at several
trol on a balanced diet, 2) 14% to 16% weight gain via fat grades. Bottom, effect on energy expenditure; top, effect on ver-
supplements to a balanced diet for 11 weeks, 3) 18% to tical motion of body.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

86 Chapter 5 / Energetics of Walking

Cotes et al. (24) predicts lower values for women than for
men. The authors attributed this to a smaller step length
in women. Recent studies, however, indicate that a smaller
step length increases energy expenditure (4). Ralston (84)
did not find a significant gender difference in locomotor
energy cost.
Durnin and Passmore (29) stated that gender is not
a factor in walking energetics. Molen and Rozendal (65)
found a small difference in favor of the female when gross
energy expenditure was measured, but not when net en-
ergy expenditure (i.e., gross cost minus cost of quiet stand-
ing) was used. It has been shown that the cost of quiet
standing is slightly (but significantly) lower in the female
than in the male. It might be expected that at lower speeds,
this would explain a slight difference between males and
females. Corcoran and Brengelmann (22) and Ralston (84)
found a somewhat higher value for females than for males
during floor walking, but were in agreement with Molen
and Rozendal (65) that the difference was not important
enough to justify use of different equations for males and
females.
Blessey (9) found that the rate of energy expenditure
did not vary with gender and that male and female adults
walk at an equal percent of maximal V02 (39%). Similarly,
FIGURE 5-4. Effects of loading lower leg during walking at sev- Waters et al. (108) measured oxygen uptake per kilogram
eral grades. Bottom, effect on energy expenditure; top, effect on in young adults and seniors during free level walking and
vertical motion of body. found no difference between males and females in ei-
ther group. Walking speed was significantly slower in fe-
male compared with male young adults. No difference
entirely clearcut because, unexpectedly, the vertical mo- in walking speed was found between male and female
tion of the body is also increased by loads on the limbs seniors.
(as a result of greater step length), as shown in Figures 5-3 Waters et al. (109) also measured oxygen uptake per
and 5-4. As a consequence of this, both inertial and gravi- kilogram body weight in 114 children and adolescents dur-
tational effects are involved when the limbs are loaded. ing level walking. While they reported no difference be-
The metabolic cost of trunk loading in quiet standing is tween male and female adolescents, they did find a small
practically zero. As stated in “Lying, Sitting, and Standing,” but significant decrease in oxygen uptake in female chil-
the Ew for quiet standing in the young male adult is about dren compared to male children. Walking speed was not
22 cal/kg per minute. In an experiment on a young male different between males and females in either group.
subject weighing 64 kg, the Ew of quiet standing was not In related work, Morgan et al. (72) documented gen-
measurably altered by attaching 20 kg uniformly around der differences in running economy in 6-year-old boys and
the trunk. girls. Results from their investigation revealed that abso-
Laursen et al. (45) found that loads of more than 10% lute and mass-related values of gross and net V02 were
to 15% body weight resulted in higher energy expenditure significantly greater in boys compared to girls, but gross
and longer recovery from raised blood pressure. Similarly, V02 expressed relative to fat-free mass was not different
in children, Hong et al. (37) found that loads of more than between sexes. This finding suggests that the higher loco-
10% body weight caused significant increases in energy ex- motor V02 demand displayed by the boys was related to
penditure and longer recovery from raised blood pressure. the presence of a greater muscle mass.
Given the relatively heavy loads that are often carried in Leg length and body height are generally greater in
a typical school backpack, further investigation of these males compared with females, but may explain only small
physiologic effects is warranted. differences in energy expenditure. Increased leg length
may decrease the energy requirements of walking at a
given speed. Brockett et al. (14) demonstrated a small
Gender
but significantly higher correlation in the regression of
Booyens and Keatinge (13) found significantly lower val- energy cost on TBW and leg length (r = 0.77) than in
ues of energy expenditure for women than for men at the relationship between energy cost and TBW alone (r =
speeds of 91 and 107 m per minute, and the equation of 0.75). Van Der Walt and Wyndham (105) found a negligible
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 87

relationship between leg length and oxygen uptake (r metric transducers. Two cords were attached to the trunk,
= −0.073). Workman and Armstrong (119) found that one horizontal and one vertical, and drove transducers that
body height had a small but consistent effect on the pre- recorded motions of the HAT in both horizontal and verti-
dicted value of oxygen uptake while walking. According to cal directions.
their predictions, short people use more oxygen to walk This method ignores certain motions that are of rela-
than tall people of the same body weight at all walking tively small magnitude, such as rotational motions of the
speeds. HAT and limb segments and arm swing. Earlier investiga-
It may be concluded that most studies have found no tors similarly ignored such second-order effects (19). Ac-
consistent differences in oxygen uptake per kilogram while cording to Winter (117), who extended this type of study
walking in adults of different age, weight, and gender, to include rotational motion, only in the case of the shank
whereas for children and adolescents, some differences oc- does the rotational kinetic energy have significant value,
cur with age and possibly gender. When interpreting the contributing ∼ 10% of the lower leg energy. More recently,
effect of age, weight, or gender on energy expenditure, it computerized motion analysis has been used to obtain sim-
is important to distinguish whether oxygen uptake is ex- ilar measurements, including rotational motion.
pressed per unit time, per unit kilogram/time, or per lean Masses of the body segments were determined from the
body mass and whether the person is walking at a speed volumes of water displacement and from values of the spe-
associated with a percent of maximum workload (e.g., % cific gravities provided in the literature. From the motions
V02 max, % of maximal heart rate, or % maximal walking and masses, potential (gravitational) and kinetic (inertial)
speed), a self-selected speed or a set speed. In consider- energy levels of the body segments were calculated for in-
ing the latter two conditions, it is important to recognize tervals of 0.02 sec during the walking cycle.
that a U-shaped relationship exists between walking speed Figure 5-5 shows the changes in mechanical energy lev-
and V02 when V02 is expressed relative to distance traveled els of the various body segments, as labeled, of a young
(e.g., mL/kg/km or mL/kg/m). Self-selected walking speeds woman of 58.6 kg, 169 cm, walking at a speed of 73.2
are often used when evaluating therapeutic interventions m per minute. Limb motion during walking enables a
with an individual patient or within a group. However, be- smooth forward progression of the body with minimal
cause V02 measured under this condition is not indepen- displacement of the center of gravity. At the beginning
dent of speed, strong consideration should be given to hav- of the gait cycle, during initial stance, the HAT elevates
ing individuals walk at a set speed if group comparisons over the lower limb generated by forward kinetic energy.
in walking economy are being made or the impact of ther- In mid-stance as the HAT reaches maximal vertical eleva-
apy to reduce locomotor energy cost is being evaluated. tion, the forward kinetic energy is converted into potential
Clearly, the units in which oxygen uptake are expressed energy of HAT elevation. This potential energy is recon-
and the condition under which V02 is assessed can lead verted into forward kinetic energy in late stance as the HAT
researchers and clinicians to different interpretations and passes in front of the foot and enables energy transfer to
conclusions. the next step, maintaining approximately constant total
mechanical energy level for HAT during most of the gait
cycle.
WORK POWER AND EFFICIENCY The significant features of these data are: 1) the ap-
proximate mirror imaging of the HAT potential and HAT
Energy efficiency can be calculated using values of me- horizontal kinetic curves, 2) the flatness of the HAT to-
chanical energy levels (kinetic and potential) of the vari- tal curve during about two-thirds of each step, 3) the dis-
ous segments of the body during successive moments of turbance of the HAT total curve during transition from
the walking cycle along with measurements of energy ex- stance to swing phase, which, on the basis of electromyo-
penditure. Ralston and Lukin (86) originally used a direct graphic evidence, coincides with the major muscle ac-
method of measuring such energy levels that was suitable tivity during the walking cycle, 4) the large channel in
for certain motions of the body at moderate walking speeds energy level of the swinging leg, due almost entirely to ve-
up to ∼100 m per minute, corresponding to cadences up locity (kinetic energy) changes in the limb segments, and
to ∼130 steps per minute. 5) the large positive work peak in the total body after heel
The technique utilized cords that were attached to the contact.
principal segments of the body: head, arms, and trunk The positive work per step, measured by the increase
(HAT); thigh, shank, and foot. The attachments to thigh in the total body work level, averaged 29.85 J (7.13 cal).
and shank were at the center of mass of each segment, The subject walked at an average cadence of 100 steps per
placed according to the specifications of Dempster (28). minute, so the positive work per minute = 100 × 29.85 or
The attachment to the foot was at the heel, and the attach- 2985 J per minute (49.7 W, 713 cal per minute, 0.067 hp).
ment to the HAT at approximately the level of the second Such a value for mechanical work rate is in good agree-
sacral vertebra. The cords attached to the thigh, lower leg, ment with the results of Cavagna and Kaneko (18), who
and foot ran horizontally backward and drove potentio- used a cinematographic method in their studies.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

88 Chapter 5 / Energetics of Walking

FIGURE 5-5. Energy levels of

body segments and whole body
of a 19-year old woman walking
at 73.2 m per minute. (Adapted
with permission from Ralston HJ,
Lukin L. Energy levels of human
body segments during level walk-
ing. Ergonomics 1969;12:39–46.)

The burst of positive work in each step occurred during climbing a hill. Wilkie (115) referred to work in overcom-
an average interval of 0.19 sec. Hence, 29.85 J/0.19 = 157 W ing air resistance as external work; work associated with
(2250 cal per minute, 0.21 hp) was the maximal positive raising and lowering the center of mass and with changes
power output during each step. in kinetic energy of the limbs he called work that was “dis-
Wilkie (115) deduced that in single movements lasting sipated internally.” He evidently regards the work done in
<1 s, the usable “external power output” of the body is climbing a hill as external work.
limited to a value somewhat less than 6 hp. In brief bouts Cavagna et al. (19) referred to external work during lo-
of exercise of ∼6 seconds, he gives the value 2 hp, and for comotion as that associated with displacement of the cen-
long-term work lasting all day, 0.2 hp. Wilkie’s figures are ter of mass of the body and to internal work as work not
based on exercise by champion athletes, and he states that leading to a displacement of the center of mass. Ralston
ordinary healthy individuals can produce less than 70% to and Lukin (86) defined “total external positive work” as
80% as much power. Even so, it is evident that the power ex- work measured by increases in the sum total of the energy
penditure in the female subject (0.21 hp) walking at a nat- levels of the body segments.
ural rate of 73.2 m per minute is much below her maximal During walking, the reaction forces at the ground can
capability. It may be concluded that there is a large margin do no work because the points of application of the forces
of tolerance in the power demand of normal walking. are fixed (it is assumed that no slippage occurs such as
Statements regarding the “efficiency” of biological pro- when walking in sand). Thus, under conditions of level
cesses are notoriously confusing. As noted by Spanner walking on the treadmill, no external work is done. The
(101), “The notion of efficiency with which a process in- ground reaction forces can cause only acceleration of the
volving energy transformation is carried out is one which center of mass of the body. Conversely, the forces produced
comes up fairly frequently. However, the word is not al- by the muscles of the body can cause no acceleration of
ways employed in precisely the same sense, and this is apt the center of mass but are responsible for the changes in
to cause some confusion.” Spanner was referring to rel- energy level of the body, both kinetic and potential. Conse-
atively straightforward chemical processes. Human walk- quently, the internal work during walking is that done by
ing, on the other hand, is vastly more difficult. muscles, while there is no external work unless one con-
Clarification of the use of the terms “external work” and siders work done by the body on the environment (as in
“internal work” may prevent this confusion. Fenn (30) re- pushing molecules of air out of the way) or by the environ-
ferred to the elevation of the center of mass as “part of the ment on the body.
external work” of running. Muller (75) referred to work In describing the efficiency of human walking, it is
against external resistance as external work. Snellen (99) desirable to refer to work done on the environment as
stated that the external work in level walking is negligi- external work (as in pushing molecules of air out of the
ble and described external work as that work involved in way). Consequently, in the case of normal walking on
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 89

terrain (rather than simply level walking on the treadmill), As previously mentioned, many authors have used a net
the overall efficiency would be appropriately defined as the energy value in calculating efficiency. Such a net value is
sum of the external work plus the internal work divided obtained by subtracting a resting energy value from the
by the metabolic expenditure. This definition was used by energy expenditure or by subtracting an extrapolated value
Winter (118). from the energy expenditure. Such a procedure can lead
Under conditions of level walking on the treadmill, to very misleading conclusions. It has already been noted
when virtually no external work is done, we shall use a that use of a net value for mathematical formulation of the
definition of efficiency that is unambiguous and that we relation between walking speed and energy expenditure is
have found to be of considerable practical usefulness, es- unproductive.
pecially in the comparison of work done by normal and Wilkie (116), in speaking of human muscular exercise,
disabled human subjects during walking. said: “The usual procedure of subtracting the resting oxy-
Gross efficiency is defined as the total positive work gen consumption from the total does not correspond to any
per minute, determined from the total energy level of the clear hypothesis about what is being estimated. In order
body segments (Fig. 5-5) divided by the total metabolic to determine the efficiency of the working muscles them-
expenditure per minute. For the female subject in Figure selves one should also subtract the extra oxygen used by
5-5, the positive work, expressed in metabolic units, was heart, respiratory muscles, etc. . . . ”
713 cal per minute, and the metabolic energy expenditure
was 3010 cal per minute. The gross efficiency therefore is Energy of Expenditure Per Unit
713/3010 or 0.24. For the female subject walking slowly Distance Walked
at 48.8 m per minute, the total positive work, expressed
in metabolic units, was 350 cal per minute, and the total The measurement of energy efficiency is not feasible in
metabolic expenditure was 2544 cal per minute or 0.14. most research and clinical laboratories. The term effi-
Cavagna and Kaneko (18), who used net rather than ciency implies that the total amount of work performed
total metabolic expenditure in calculating efficiency, ob- by the body was calculated. The measurement of energy
tained similar results when total expenditure was used in expenditure per distance walked provides a quantitative
the calculation. This value is consistent with values of ef- measure of energy economy, much the same as fuel econ-
ficiency in the literature for various kinds of work, such omy is measured in the automobile.
as cycling on an ergometer. The data of Silverman et al. Many authors have discussed the energy cost of walk-
(97) for cycle ergometer exercise yield values of efficien- ing expressed per unit body weight per unit distance. The
cies very close to those found by Ralston and Lukin (86) curve of such energy cost, plotted against speed, is con-
for normal walking speeds, which commonly ranged from cave upward, and although fairly flat over a wide range
0.20 to 0.25. Paulsen and Asmussen (81) calculated net of speeds (65–100 m per minute), the curve still exhibits
efficiency from 23 experiments of arm vs. leg work on a minimal value (Fig. 5-6, top). Ralston (84) showed that
the ergometer and found arm work efficiency was 15.8% the mathematical form of such a curve could be deduced
compared with 21.9% for leg work. from the following equation:
Ralston and Lukin (86) compared normal subjects with Ew 32
patients who had motor disability. Two below-knee am- Em = = + 0.0050 v (4)
v v
putees, wearing conventional prostheses, were studied
with the techniques described previously. The subjects where Em is expressed in cal/kg/m and Ew is expressed in
could walk comfortably only at lower speeds of up to cal/kg per minute.
50 m per minute. As would be expected, the gross effi- Differentiating Em with respect to v (speed), and equat-
ciencies were relatively low, ranging from ∼ 0.08 to 0.14. ing to zero, yields a minimal Em of 0.80 cal/kg/m, corre-
However, 0.14 is about the same as that for the normal fe- sponding to an optimal speed of 80 m per minute.
male subject, described previously, when walking at 48.8 m Similarly, Zarrugh et al. (121) proposed the following
per minute. It may be concluded that one factor involved equation.
in the low value of the gross efficiency is simply the low Eo
speed, apart from the lack of normal muscle coordination Em = (5)
v(1 − v/vu )2
associated with the disability.
As is usual in cases of motor disability, the energy where Eo is value of Ew when step length and step rate
expenditure per step or per meter was high compared approaches zero and Vu is the upper limit of walking speed.
with normal subjects, whereas the energy expenditure per Differentiating Em with respect to v and equating to zero
minute was moderate, as a result of the low speeds used. yields an optimal speed,
It is the moderate energy cost per minute that is linked vopt = vu /3 = 240/3 = 80 m per minute (6)
to moderate changes in heart rate, blood pressure, and
pulmonary ventilation. as in the case of Equation 4.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

90 Chapter 5 / Energetics of Walking

FIGURE 5-6. Gross energy expenditure per me-

ter as function of walking speed v for typical sub-
ject walking naturally at different speeds. Lower
curves, as increasing amounts of energy (R) are
subtracted from gross energy expenditure per
minute, the optimal speed corresponding to mini-
mal Em becomes smaller. (From Ralston HJ. Ener-
getics of Human Walking. In: Herman RM, Grill-
ner S, Stein PSG, Stuart DG, eds. Neural Control
of Locomotion. New York: Plenum; 1976.)

As shown by Ralston (84), a person walking in a nat- ter is covered as efficiently, from an energy standpoint, as
ural self-selected manner tends to adopt a speed close to possible.
the optimal speed. This finding was confirmed by Corco- This principle is analogous to a biological conservation
ran and Brengelmann (22) in a study of 32 subjects during of energy law. The French physiologist E. J. Marey (1895),
floor walking. These authors found a natural average speed one of the pioneers in the study of animal locomotion,
of 83.4 m per minute, differing from the above theoreti- anticipated the statement of this principle over a century
cal value of 80 m per minute by only 4%. Rose et al. (91) ago (58). Commenting on the applications of measuring
found a similar curve based on oxygen uptake per meter for various gait parameters he stated, “The data afforded by
children and adolescents walking on a treadmill with an these measurements may be put to practical use, for they
optimal walking speed of 84 m per minute, ranging from indicate, according to the object in view, the best way of
64 to 91 m per minute. Similarly, Morgan and associates utilizing muscular force in walking or running: whether it
(73) reported no significant difference in the freely-chosen be to traverse the greatest distance with the least expendi-
walking speed (1.03 +/− 0.14 m per second) and the most ture of energy, or whether it be to cover a certain distance
economical walking speed (1.06 +/− 0.13 m per second) in the least possible time. There is then, for each pace, an
in six children with cerebral palsy. optimum rate of steps per minute, which corresponds to
Waters et al. (108,109) measured oxygen uptake per the point at which the velocity increases proportionately
kilogram per meter in children, teenagers, and adults walk- faster than energy is expended.”
ing at self-selected slow, comfortable, and fast walking Only the use of gross energy values leads to Equations 5
speeds and found a significant decrease in oxygen uptake and 6, which establish an optimum condition of economy
per meter with age. for free walking patterns and predict an experimentally ver-
Walking speed equals steps per minute multiplied by ifiable optimal speed. The top curve of Figure 5-6 shows the
step length. Thus, an energetically-optimal walking speed gross energy expenditure per meter as a function of walk-
must be based on a choice of step rate and step length ing speed v for a typical subject when walking naturally
that yields a minimal energy cost. This is an example at different speeds. The lower curves show that if increas-
of a fundamental feature of human motor behavior that ing amounts of energy (R) are subtracted from the gross
applies to many activities in addition to walking. In a energy expenditure per minute, the optimal speed corre-
freely chosen rate of activity, a rate is chosen that rep- sponding to minimal Em becomes smaller and smaller. It
resents minimal energy expenditure per unit task. In the becomes essentially zero when an amount equal to Eo is
case of natural walking, where the unit task is traversing subtracted. It is not very enlightening to find that the best
1 m of ground, a speed is adopted such that each me- way to avoid energy cost during walking is not to walk! It
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:13

Chapter 5 / Energetics of Walking 91

is suggested that in other types of exercise, the gross rather The near constancy in the value of s/n in walking has
than the net energy expenditures may be the fundamental physiologic significance. Not only does the value of s/n
energy parameter. characterize the walk of the male and female, but it is also
fundamental in determining the optimal speed of walking.
Energy Expenditure Per Step At any given speed, imposition of an unnatural step rate (or
step length) for that speed results in a higher than normal
Minimal energy expenditure per meter traversed per kilo- value of the energy expenditure per unit distance traversed
gram body weight (Em) occurs at a speed of 80 m per (5,75,96,121).
minute. The relations will now be considered between Shields (96) measured the time to reach 88% of age-
speed (v), step rate (n), step length (s), energy expenditure predicted maximum heart rate in adults walking at 90 m
per kilogram per minute (Ew ), and energy expenditure per per minute. She found that normal step length was 72%
kilogram per step (En ). of leg length and that this resulted in significantly better
The relevant data from Zarrugh et al. (121) are shown performance than a step length of 60% of leg length, but
in Table 5-6, based on studies of 10 normal males and 10 resulted in a significantly poorer performance than a step
normal females ranging 20 to 55 and 20 to 49 years, re- length of 80% of leg length.
spectively. Ew values are calculated from Equation 1. In the following mathematical treatment, the average
The difference in the value of s/n at the higher speeds in values of s/n = 0.007 and s/n = 0.0064 (Table 5-6) will be
males and females is highly significant. For example, at v = used for males and females, respectively.
73.2 m per minute, the odds against the difference being In a manner similar to that followed in the preceding
due to chance are better than 1000 to 1. Even at 48.8 m section, Equation 1 can be used to determine the step
per minute, the odds are 4 to 1. The smaller values of s rate corresponding to minimal energy expenditure per
and s/n in females reflect the shorter average leg lengths step:
in women compared with men, although this may not be Since v = sn, and s = 0.007n in males, v = 0.007n2 .
the only factor involved. Substituting in Equation 1
The values shown in Table 5-6 do not strictly correspond
to those of a natural walk, since the treadmill imposed
the speeds. However, 73.2 m per minute is fairly close to Ew = 32 + 0.0050v 2 = 32 + 0.0050 (0.007n2 )2
the optimal speed (80 m per minute) and therefore should Ew 32
En = = + 0.005 (0.007)2 n2
yield values of s/n close to those of natural walk. n n
Molen and Rozendal (66) studied 309 males and 32
= + 2.45 × 10−1 n2
224 females walking along pavement and path. At an av- n
erage speed of 83.4 m per minute, the average value of s/n
for males was 0.0072, in good agreement with the values of Differentiating En with respect to n and equating to
0.0070 and 0.0072 at speeds of 73.2 and 97.6 m per minute zero, En (minimum) occurs at n = 81.2 steps per minute.
in Table 5-6. At an average speed of 76.2 m per minute, the After interpolating from Table 5.6, this would corre-
average value of s/n for females was 0.0060, in excellent spond to a speed of about 46 m per minute, which is not the
agreement with the value of 0.0061 at 73.2 m per minute speed (80 m per minute) corresponding to minimal energy
in Table 5.6. expenditure per meter.

◗ TABLE 5-6 Average Values of v, n, s, s/n, Ew , and Em for 10 Male and

10 Female Subjectsa
v n s/n (m/steps/min) Ew En (cal/kg/step)
(m/min) (steps/min) s (m) (S.D.) (cal/kg/min) =E w/n

Males 24.4 59.5 ± 4.33 0.41 ± 0.025 0.0069 ± 0.00065 35.0 0.59
48.8 84.4 ± 6.48 0.59 ± 0.041 0.0070 ± 0.00072 43.9 0.52
73.2 102.2 ± 5.43 0.72 ± 0.056 0.0070 ± 0.00066 58.8 0.58
97.6 116.3 ± 3.13 0.84 ± 0.020 0.0072 ± 0.00026 79.6 0.68
Mean 0.0070 ± 0.00120
Females 24.4 60.0 ± 3.95 0.41 ± 0.029 0.0068 ± 0.00065 35.0 0.58
48.8 86.7 ± 7.26 0.57 ± 0.037 0.0066 ± 0.00070 43.9 0.51
73.2 109.0 ± 5.49 0.67 ± 0.037 0.0061 ± 0.00045 58.8 0.54
97.6 126.8 ± 8.98 0.77 ± 0.031 0.0061 ± 0.00050 79.6 0.63
Mean 0.0064 ± 0.00120
a Values are means ± SD. From Zarrugh MY, Todd FN, Ralston HJ. Optimization of energy expenditure during level
walking. Eur J Appl Physiol 1974;33:293–306.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

92 Chapter 5 / Energetics of Walking

By using the value s/n = 0.004 for females, En (op-

timum) occurs at n = 85 steps per minute. Interpolat-
ing from Table 5-6, this would correspond to a speed of
about 47 m per minute, again representing an undesirable
method of walking.
Clearly, the optimal step rate for minimal energy per
step does not represent a desirable method of walking, at
least in normal subjects.
The significance of this conclusion is addressed in
“Work, Power, and Efficiency” concerning disability, where
the inability to walk at higher speeds causes the energy ex-
penditure per step to greatly exceed normal values. In such
cases, the energy expenditure per minute is kept low be-
cause of low step rate.
FIGURE 5-7. Effect of grade on energy expenditure of a normal
young male while walking at three speeds.

Terrain
grades, the metabolic cost increases as a result of postural
Passmore and Durnin (80), in their review of energy expen-
changes and the use of muscles in “braking” action.
diture, state, “The type of surface may have a slight effect
Several investigators have developed equations relating
on the energy cost of walking. However, unless the surface
energy expenditure to speed and grade. Formulas for es-
is markedly rough, the effect will probably not exceed 10%
timating the energy requirements of walking at various
more than walking on a flat surface.” Haisman and Gold-
speeds and grades have been published by the American
man (33) found that energy expenditure of walking with
College of Sports Medicine (1). Montoye et al. (67) vali-
a 20-kg backpack was 10% higher when walking on grass-
dated these formulas and found that the formulas are ac-
land compared with blacktop at walking speeds of 53 and
curate for estimating the mean oxygen requirements in 6%
80 m per minute. Passmore and Durnin (80) reported an
to 18% grade walking in adult males. In horizontal walk-
increase of ∼ 35% in oxygen uptake for a subject walking
ing and walking at 3% grade, the formulas underestimate
at a speed of 90 m per minute on plowed field compared
oxygen uptake in all age groups.
with asphalt road. Strydom et al. (102), in a study of 11
In boys younger than 18 years, the formulas underesti-
young men, observed that the metabolic cost of walking at
mate the energy requirement in walking at all grade levels.
∼ 80 m per minute with loads of 23 kg was 80% greater on
Montoye et al. (67) concluded that when applied to adults,
loose sand then on a hard surface.
the formulas provide a reasonable estimate of the actual
In the latter two cases cited, the walking speed was fairly
oxygen requirement of steady-state treadmill walking.
brisk, and therefore the results might not be relevant to
Table 5-7 shows representative data, such as those
the effect of loose soil at lower speeds. However, is clear
provided by McDonald (51), based on experiments by
that the nature of the terrain must always be considered
Margaria (59) on three normal adult males. The calculated
in anticipating the metabolic demand of walking.
values of Ew (cal/kg per minute) and Ew (cal/kg/m) have
Terrain coefficients were developed to correct for the
been added, based on a body weight of 72 kg.
type of surface traversed (78). Single coefficients were
As has been noted earlier, during level walking, a person
found to fit all measured walking surfaces except for snow
tends to adopt a speed such that energy expenditure per
(76). The coefficient for soft snow was found to increase as
unit distance is minimized. From Table 5-7, it is evident
the snow footprint depth increased.
that minimal values of Em occur at slopes of + 10 to −20%
at speeds of 80 to 100 m per minute. Therefore, it may
be expected that persons will adopt such speeds at those
Slope Walking
grades. Above + 10%, and below −20%, Em minima do
Figure 5-7 presents the essential metabolic features of not occur within the range of speeds shown.
slope walking. A normal young adult male walked on the While the subject of Figure 5-7 exhibited minimal val-
treadmill at speeds of 48.8, 73.2, and 97.6 m per minute and ues of Ew at a slope of about −7% (−4◦ ), the minimal
at grades ranging from + 4◦ to −10◦ (+ 7% to −18%). The value of Ew in Table 5-7 occurs at a slope of about −10%.
metabolic cost rapidly increased with increasing grade. Similarly, Pivarnik and Sherman (83) found the minimal
Also noted is the more modest decrease in cost at grades in values of Ev occurred at grades of −5 to −10% at walk-
the range of 0 to −4◦ (0% to −7%). The experimental points ing speeds of 80 m per minute. Above these grades, Ew
are fitted by fourth-degree polynomial functions (smooth increased significantly with increasing grades of 0, 5, and
curves), all of which exhibit minima close to −4◦ . At lower 10%.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 93

◗ TABLE 5-7 Estimates of Ew (cal/kg/min) and Em (cal/kg/m) at Various Speeds

and Gradesa
Grade 40 m/min 60 m/min 80 m/min 100 m/min 120 m/min
% Degrees Ew Em Ew Em Ew Em Ew Em Ew Em

−40 −21.8 54 1.35 70 1.17 85 1.06 98 0.98 106 0.88

−35 −19.3 47 1.18 60 1.00 72 0.90 84 0.84 93 0.78
−30 −16.7 46 1.15 56 0.93 67 0.84 79 0.79 90 0.75
−25 −14.0 44 1.10 53 0.88 63 0.79 74 0.74 87 0.73
−20 −11.3 41 1.03 46 0.77 53 0.66 63 0.63 77 0.64
−15 −8.5 33 0.83 37 0.62 44 0.55 54 0.54 69 0.58
−10 −5.7 30 0.75 35 0.58 43 0.54 54 0.54 70 0.58
−5 −2.9 32 0.80 38 0.63 48 0.60 61 0.61 82 0.68
0 0 41 1.03 49 0.82 62 0.78 81 0.81 110 0.92
5 2.9 59 1.48 74 1.23 95 1.19 125 1.25 167 1.39
10 5.7 69 1.73 93 1.55 121 1.51 155 1.55 193 1.61
15 8.5 82 2.05 114 1.90 150 1.88 187 1.87
20 11.3 98 2.45 140 2.33 185 2.31
25 14.0 114 2.85 170 2.83
30 16.7 133 3.33 204 3.40
35 19.3 153 3.83
40 21.8 174 4.35
a (Modified from McDonald I. Statistical studies of recorded energy expenditure of man. Part II. Expenditure on walking
related to weight, sex, age, height, speed and gradient [abstract]. Nutr Abstr Rev 1961;31:739–762.)

Immobilization spine). The energy expenditure of three subjects walking

with both ankle braces locked and unlocked was compared
The effects of restricting motion at various joints such as
with normal values (Fig. 5-8). Normal male subjects of av-
the ankle, knee, hip, and the spine on the metabolic cost
erage physique tend to adopt a walking speed of ∼80 m
of walking were studied by Ralston (85) using immobi-
per minute. At this speed with both ankles immobilized,
lization with a brace (ankle) or plaster casts (knee, hip,
the energy expenditure was 9% above average. With one
ankle immobilized, the increase was 6%.
Knee restriction at 0, 15, and 30 degrees produced only
minimal increased energy needs (Fig. 5-9). At a walking
speed of 74 m per minute, locomotor energy demand was
significantly elevated at 45 degrees knee flexion, with en-
ergy expenditure being 19% higher than with a 0◦ knee flex-
ion angle and 37% higher than with the knee unrestricted.
Correction of leg length by use of a shoe lift did not sig-
nificantly alter this result. Hip restriction at an angle of
60 degrees was 28% higher than at 30 degrees and 39%
higher than with the hip unrestricted at a walking speed
of 57 m per minute (Fig. 5-10). Hip immobilization at 30
degrees consistently resulted in the lowest values of all the
angles studied, with 0 degrees being intermediate in re-
sults. Correction of leg length by use of a shoe lift did not
materially alter this result. Trunk restriction increased en-
ergy expenditure by 10% over a wide range of speeds (Fig.
5-11). Restriction of arm motion in seven subjects, over
FIGURE 5-8. Effects of immobilization of both ankle joints on a wide range of speeds, had no significant effect on the
average energy expenditure, as a function of speed, for three sub- energy cost of walking.
jects walking. Heavy line, average values for 12 normal subjects;
stippled area, scatter of normal values. (From Ralston HJ. Effects Similarly, Mattsson and Brostrom (62) found that in
of Immobilizations of Various Body Segments on the Energy Cost adults walking at self-selected comfortable speeds, immo-
of Human Locomotion. Ergonnomics 1965;(suppl):53.) bilization of the knee resulted in a 23% increase in oxygen
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

94 Chapter 5 / Energetics of Walking

FIGURE 5-11. Effect of restriction of trunk motion on aver-

FIGURE 5-9. Effect of angle of immobilization of knee on aver- age energy expenditure, as a function of speed, of seven subjects
age energy expenditures of eight subjects during walking. Broken with and without restriction of shoulder, spine, and pelvic mo-
lines, mean values for normal walking. (From Ralston HJ. Effects tion. Panel B: heavy line, average values for 12 normal subjects;
of Immobilizations of Various Body Segments on the Energy Cost stippled area, scatter of normal values. (Modified with permis-
of Human Locomotion. Ergonnomics 1965;(suppl):53.) sion, from Ralston HJ. Effects of immobilization of various body
segments on the energy cost of human locomotion. Ergonomics
1965;(suppl):53.)

uptake per meter walked (0.156 vs. 0.196 mL/kg per

minute). Walking speed decreased from 79 to 60 m per
minute. Waters and Mulroy (111) compared the effect
of restriction of the ankle, knee and hip on oxygen cost
(mL/kg/m) during walking and found a progressive rise in
the rate of energy cost caused by the loss of joint motion
progressing from the ankle, to the knee and to the hip.
Oxygen cost during walking increased slightly (13%) with
ankle fusion, to a greater degree (33%) with knee immobi-
lization with a cylinder cast and still higher with hip fusion
(46%). Waters and Lunsford (110) found that the energy
cost of walking with the knee flexed increased progres-
sively with the position of knee flexion. Immobilization of
the knee flexed 15 degrees resulted in increased oxygen
cost of 13% compared to an increase of 45% with the knee
flexed 45 degrees.

Disability
Measuring the energy expenditure of walking in persons
with a disability provides objective documentation of the
degree of disability and the effect of treatment such as
surgery, physical therapy, ambulatory aids, and prosthet-
ics. Bard and Ralston (6) measured oxygen uptake and
calculated calories/kilogram per minute in an amputee
walking with three different combinations of prosthet-
ics and/or crutches. They found that when the patient
FIGURE 5-10. Effect of angle of immobilization of hip on aver- walked at a comfortable walking speed (74 m per minute)
age energy expenditures of four subjects during walking. Broken
lines, mean values for normal walking. (From Ralston HJ. Effects with his usual suction-socket prosthesis with a SACH
of Immobilizations of Various Body Segments on the Energy Cost foot, the energy expenditure was only 20% above normal.
of Human Locomotion. Ergonnomics 1965;(suppl):53.) When using a pylon instead of his usual prosthesis, the
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 95

◗ TABLE 5-8 Average Values of Energy Expenditure of Normal and Hemiplegic

Subjects during Normal Stair Climbing Using Alternate Gaita
Energy Expenditure
Steps/min cal/kg/min cal/kg/step
Experimental
Conditions Normal Hemiplegic Normal Hemiplegic Normal Hemiplegic

Rest (sitting) 18.9 15.6

Low stairs 72 47 53.3 49.1 0.74 1.04
High stairs 70 44 70.3 61.5 1.00 1.40
a From Hirschberg GG, Ralston HJ. Energy cost of stair-climbing in normal and hemiplegic subjects. Am J Phys Med
1965;44:165.

energy expenditure increased sharply at speeds above Not shown in Table 5-8, but worthy of note, is the fact
60 m per minute. Energy expenditure was highest when that alternating gait required 28% to 30% less metabolic
using forearm crutches for all walking speeds. When ex- demand per step than unilateral leading gait in both nor-
pressed in calories/kilogram/meter, the optimal energy ex- mal and hemiplegic subjects. Also, high stairs elicited a
penditure with crutches was 30% higher than with the greater physiologic demand, per minute and per step, than
usual prosthesis, whereas the optimal walking speed was low stairs, in both normal and hemiplegic subjects.
20% less. McBeath et al. (47) measured oxygen uptake in nor-
In a comprehensive review of the energy cost of walking mal men walking with loftstrand and axillary crutches.
with disability, Waters and Mulroy (111) note that while They found that walking with crutches or canes resulted in
there are a substantial number of studies on the energy significantly higher energy expenditure and that the com-
expenditure of walking for persons with lower-extremity fortable walking speed of assisted ambulation was signif-
amputation, a direct comparison of the results of the dif- icantly slower than normal. With both loftstrand and ax-
ferent studies is difficult because young individuals with illary crutches, the oxygen uptake per meter walked was
amputations, which are usually traumatic, are not con- greater than normal for the three-point partial weight-
sistently distinguished from older persons with amputa- bearing and the two-point alternating gaits, but not as
tions, which are usually vascular. Yet, there are significant great as the costs for the three-point non-weight-bearing
differences between these two groups with respect to gait and the swing-through gaits.
performance. In addition, Waters and Mulroy (111) note Corcoran et al. (23) studied hemiplegic subjects walk-
that use of upper-extremity assistive devices, prosthetic ing with plastic and metal braces. They found that use of
fit and experience using the prosthesis influence the re- a brace significantly decreased oxygen uptake per meter,
sults but are not well-controlled for. They noted that re- but there were no differences between plastic versus metal
search has shown that there is a higher oxygen cost of braces. Rose et al. (92) measured oxygen uptake in chil-
walking for persons with lower limb amputations; how- dren with cerebral palsy walking at progressively increas-
ever, whether the increased cost resulted from a slower ing speeds on a treadmill. Oxygen uptake at rest and at a
walking speed or a higher rate of oxygen uptake depended maximal walking speed were not different from normal
on the level of amputation and physical fitness of the values. However, maximal walking speed averaged only
individual. 56 m per minute for cerebral palsy subjects compared
Hirschberg and Ralston (36) compared normal and with 122 m per minute for normal subjects. Oxygen up-
hemiplegic subjects climbing stairs with low (10 cm) and take per meter walked averaged 280% greater for children
high (19 cm) risers using alternating gait (Table 5-8). As with cerebral palsy compared with normal controls (92).
might be expected, the energy expenditure per step was Oxygen uptake per meter at the most economical walking
much greater, on the order of 40% higher, in the hemi- speed was 0.48 mL/kg/m compared with 0.17 mL/kg/m for
plegic subjects compared to the normal subjects. The normal subjects. Diplegic subjects had significantly higher
energy expenditure per minute is not very different in values than hemiplegic subjects.
the two types of subjects, because hemiplegic patients Unnithan and colleagues (104) reported that children
adopt a low step rate. Note that such physiologic vari- with cerebral palsy exhibited substantially higher tread-
ables as heart rate, blood pressure, and respiratory rate mill walking energy costs compared to a control group of
are linked to the metabolic rate per minute rather than able-bodied children. Similarly, Morgan et al. (73) demon-
to the metabolic rate per meter or per step. These phys- strated that children with spastic hemiplegic CP demon-
iologic variables, while slightly higher in the hemiplegic strated higher walking V02 values (mL/kg per minute)
than in the normal subjects, were still within acceptable while walking at 1.5, 2, and 2.5 mph compared to a
limits. matched group of able-bodied children. However, only V02
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

96 Chapter 5 / Energetics of Walking

comparisons at the two highest speeds were significantly Heart rate is also affected by climatic stress, dehydra-
different between groups. Campbell and Ball (17) mea- tion, fever, various diseases, and medications (4,7). At a
sured oxygen uptake (mL/kg per minute) in children with given workload, heart rate increases with the use of a
cerebral palsy walking at a single comfortable speed. They smaller active muscle mass such as with upper-extremity
found that oxygen uptake values varied widely but were work (6). Paulsen and Asmussen (81) compared heart rate
higher than normal for all children with cerebral palsy. and oxygen uptake and found the relationship to be linear,
In non-disabled children, oxygen uptake at the normal but the slope of the relationship was increased for upper-
walking speed decreases with age. The opposite trend oc- extremity work compared with lower-extremity work. The
curred in the children with cerebral palsy, in that oxy- increase in heart rate at a given level of oxygen uptake is in-
gen uptake values increased with age. This may reflect versely proportional to the size of muscle mass being used.
increased size and decreased relative strength with age Despite the aforementioned limitations, heart rate is a
in cerebral palsy, causing walking to become more diffi- clinically-useful estimate of energy expenditure. Because
cult because of increases in body mass and adiposity that of its marked sensitivity to any increase or decrease in con-
are not matched by commensurate gains in leg muscle ditioning, heart rate is a valuable gauge for determination
strength. From a practical standpoint, an increase in walk- of fitness and compliance to exercise programs. The slope
ing energy costs, when combined with lower values of max- of the relationship between heart rate and oxygen uptake
imal aerobic power (38,73,104), results in a higher relative is used to assess fitness level. With training, heart rate de-
exercise intensity at any given walking speed. This, in turn, creases for a given level of oxygen uptake or workload (4).
can lead to greater levels of fatigue, decreased commu- Saltin et al. (95) found that the slope of the relationship
nity walking activity, and increased reliance on wheelchair between heart rate and oxygen uptake (liters per minute)
use. increased with bed rest and decreased with training.
In order to accurately determine the impact of clinical During steady-state submaximal workloads, heart rate
treatments aimed at reducing locomotor energy cost in has been found to be a reliable estimate of energy expendi-
youth with cerebral palsy, it is important to consider ture. Astrand and Rodahl (4) found that in adults, oxygen
within- and between-day variability in walking V02 . uptake and heart rate increase linearly throughout a wide
Limited data (42,57) suggest that fairly stable ambula- range of submaximal workloads on the cycle ergometer.
tory oxygen uptake values can be achieved if testing is The relationship was unpredictable at maximum levels of
preceded by a short period (5 to 15 minutes) of treadmill exercise. Cooper et al. (21) found that a linear relation-
accommodation. ship existed between oxygen uptake (liters per minute) and
heart rate in 107 children performing on a cycle ergometer.
The mean correlation was 0.68, but the mode fell between
HEART RATE AS AN ESTIMATE OF 0.8 and 0.9, with the lowest correlation coefficients occur-
ENERGY EXPENDITURE ring in the youngest ages owing to a greater influence of
noise on a smaller range of oxygen uptake. The younger
The rate of oxygen uptake can be used to assess the en- subjects also started and stopped during the protocol, re-
ergy expended while walking, but the instrumentation is sulting in increased variation of heart rate. They found
cumbersome to wear and is unavailable in most clinical that the slope (0.33) of the relationship, when normalized
settings. Conversely, heart rate is an easily measured pa- for body weight, did not significantly change with age or
rameter and has been found to be an accurate and conve- weight, but the mean slope was significantly greater for
nient estimate of energy expenditure during steady-state boys (0.37) than for girls (0.29).
submaximal work in normal and disabled adults engaged Rose et al. (92) investigated the relationship between
in walking and cycling (4,50,81) and in normal and dis- heart rate and oxygen uptake for 18 children who were
abled children walking (92) and exercising on a cycle er- nondisabled and 13 children with cerebral palsy walk-
gometer (7,21). ing on a treadmill (Fig. 5-12). The relationship was lin-
Limitations exist for the use of heart rate as an esti- ear between heart rate (beats per minute) and oxygen up-
mate of energy expenditure. At very rapid rates of walk- take (mL/kg per minute) in the nondisabled children and
ing, the relationship is unpredictable between heart and children with cerebral palsy. The mean correlation within
oxygen uptake (4). Heart rate may be affected by factors an individual was very high (nondisabled, 0.98, cerebral
other than oxygen uptake, such as anxiety or anticipa- palsy, 0.99). Within the normal and cerebral palsy groups,
tion, particularly at rest and low levels of activity (4,7). the correlation was also high (nondisabled, 0.83, cere-
Ganguli and Datta (31) state that the effect of anxiety is bral palsy, 0.84). The linear relationship between heart
usually overcome quite readily by habituation after the first rate and mass-related oxygen uptake existed throughout
or second test in an investigation that constitutes a series a wide range of walking speeds (22-131 m per minute).
of tests. Furthermore, the importance of anxiety lessens as The slope of the relationship between heart rate and oxy-
the intensity of exercise is increased. gen uptake was not significantly different for the normal
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 97

capacity. Similarly, Blessey (9) found that in adults aged

20 to 29, walking represented 36% of maximal oxygen up-
take compared with 43% of maximal oxygen uptake for
adults aged 50 to 59.
Waters et al. (109) found that heart rate was higher in
children (114 beats per minute) than in adolescents (97
beats per minute) walking at a comfortable self-selected
speed. Walking speed was slightly, but significantly, higher
in the adolescents (73 m per minute) compared with the
children (70 m per minute). Similarly, Bar-Or (7) con-
cluded that submaximal heart rate declined with age in
children performing the same absolute exercise task. He
noted that some of this difference can be explained by vari-
ations in relative exercise intensity, i.e., given the level of
oxygen uptake or percent maximal oxygen uptake. As pre-
viously mentioned, Cooper et al. (21) and Rose et al. (92)
found that the slope of the relationship between heart rate
and oxygen uptake per kilogram did not differ significantly
with age for children and adolescents. It may be concluded
that the age-related decrease in heart rate in children and
adults while walking is related to the difference in relative
exercise intensity.

Gender

FIGURE 5-12. Scatterplot and linear regression line with 95%

Waters et al. (108) found that during self-paced walking,
confidence limits around the mean for heart rate vs. oxygen up- heart rate was slightly but significantly higher in young fe-
take for 18 non-disabled children (A) and 13 children with cere- male adults (104 beats per minute) and seniors (106 beats
bral palsy (B). (From Rose et al. J Pediatr Orthop 1989;9:276–279). per minute) compared with male young adults (94 beats
per minute) and seniors (97 beats per minute). Walk-
ing speed was significantly slower in female young adults
subjects (0.30) compared to children with cerebral palsy
(76 beats per minute) compared with male young adults
(0.32) and did not differ significantly with age.
(84 m per minute). Bassey et al. (8) found that women had
walking heart rates that are 4.4 beats per minute higher,
even when stature and weight are controlled for. Gender
Age
differences disappeared when percent fat and fat-free mass
Astrand and Rodahl (4) reported that adult heart rate were controlled.
does not change with age for a given workload (100 W). Waters et al. (109) found that heart rate during self-
Waters et al. (108) found no difference in heart rate paced walking was significantly higher in female chil-
between adults (20–59 years) and seniors (60–80 years) dren (118 beats per minute) and adolescents (103 beats
walking at a self-selected comfortable walking speed. How- per minute) compared with male children (11l beats per
ever, walking speed was significantly slower in the seniors minute) and adolescents (90 beats per minute).
(73 m per minute) compared with young adults (80 m per Bar-Or (7) also found that female children and adoles-
minute). Conversely, Bassey et al. (8) reported that during cents have a higher heart rate than males at a given work-
self-paced walking, heart rate was approximately 10 beats load. He suggested that the higher heart rates may be re-
per minute lower in subjects aged 20 to 29 compared with lated to decreased stroke volume or to lower levels of ha-
60 to 69 years when body fat was considered. bitual activity.
A given heart rate may represent a higher percent of Although the relative influence of body weight, age and
maximal capacity in older adults because maximum heart gender on heart have not been completely resolved, heart
rate decreases with age in adults (4). Astrand and Rodahl rate is an accurate and reliable estimate of energy expen-
(4) studied 350 subjects aged 25 to 60 working at 50% diture. More specifically, heart rate is an indication of the
V02 max. They reported that heart rate was 110 beats per cardiovascular demands of a given activity. Paulsen and
minute for a typical 65-year man compared with 130 beats Asmussen (81) stated: “While the caloric output estimated
per minute for a 25-year man. Thus, in the typical 65-year- from the pulse rate thus may be overestimated when com-
old, walking may represent a higher percent of maximal pared to the actual metabolism, it nevertheless gives an
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

98 Chapter 5 / Energetics of Walking

FIGURE 5-13. EEI based on oxygen uptake (A) and heart rate FIGURE 5-14. EEI based on oxygen uptake (A) and heart rate
(B) as a function of walking speed for 18 non-disabled children. (B) as a function of walking sped for 13 children with cerebral
(From Rose et al. Dev Med Child Neurol 1990;32:333–340.) palsy. (From Rose et al. Dev Med Child Neurol 1990;32:333–340.)

expression of the total load put on the organism during

the work.” The diplegic cerebral palsy subjects’ EEl values averaged
three times higher than the hemiplegic subjects.
Heart Rate Per Distance Walked MacGregor (52) had previously calculated this index for
73 normal adults walking at self selected walking speed and
Heart rate expressed per distance walked provides in-
found that 99% of the values for normal adults fell between
formation on energy economy, as does oxygen uptake
0.11 and 0.51 beats per minute. MacGregor referred to this
expressed per distance walked. Because resting heart rate
index as the physiologic cost index. MacGregor (52) stud-
is known to increase with decreased body size and level of
ied an adult polio patient and a nondisabled adult walking
fitness, the resting heart rate is subtracted from the walk-
on the floor at slow, fast, and self selected speeds. He found
ing heart rate, and this value is divided by the walking
that heart rate per meter walked was higher for the polio
speed. The following calculation of an energy expenditure
patient (0.5–0.9) than for the nondisabled adult (0.35). Fur-
index (EEI; beats per minute) can be made:
thermore, the rate increased for the polio patient when the
walking heart rate − resting heart rate brace was removed or unlatched.
EEI = Butler et al. (16) used this index to measure 72 children
walking speed
aged 3 to 12 years at a single self-selected speed and found
Rose et al. (90) compared oxygen uptake and heart rate the mean value to be 0.4 beats per minute. No change with
per meter walked in 12 children with cerebral palsy and 18 age was found. They also compared four children with
normal controls walking at progressively faster speeds on physical disability, two with cerebral palsy, and two with
a treadmill (Figs. 5-13 and 5-14). The curve generated by spina bifida, with and without their usual orthotic. They
the heart rate per meter walked was similar to the curve found that the four disabled children had higher values
generated by the oxygen uptake per meter walked for both compared with the nondisabled children, but use of the
cerebral palsy and normal subjects. orthotic decreased the index value. Children with spina bi-
Optimal walking speed averaged 75 m per minute for fida decreased from 6.9 and 5.8 beats per minute to 3.8
nondisabled subjects and 50 m per minute for subjects and 3.2 beats per minute and children with cerebral palsy
with cerebral palsy. EEl values for cerebral palsy subjects decreased from 1.5 and 1.6 beats per minute to 0.98 and
averaged 3.3 times higher than values for normal subjects. 0.7 beats per minute.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 99

value for the 102 subjects was 0.47 beats per minute
(Table 3-9). This value was similar to the most econom-
ical EEI (heart rate) value of 0.41 beats per minute for
18 normal children walking on the treadmill who were pre-
viously studied. These values for children averaged slightly
higher, but fell within the normal range of values for adults
reported by MacGregor (52).
The average self-selected comfortable floor walking EEI
decreased slightly with age, but the differences were not
statistically significant (Table 5.9). The comfortable walk-
ing speed was significantly slower for the younger age
group compared with the older three age groups, but there
was no significant difference in walking speed between
the three older age groups. The self-selected comfortable
EEl for all males (0.45 beats per minute) and females
(0.49 beats per minute) was not significantly different. In
the oldest age group, females displayed slight but signif-
icantly higher EEI values. Figure 5-16 shows data from
a 6-year-old boy with spastic hemiplegic cerebral palsy
who was tested before and after a bilateral split tibialis
posterior tendon transfer and tendoachilles lengthening.
He walked independently at self-selected slow, comfort-
FIGURE 5-15. EEI based on heart rate as a function of walking able, and fast walking speeds. The three self-selected walk-
speed. Shaded, EEI (mean ± 2 SD). Bar, range of comfortable ing speeds were slightly faster after surgery (32, 48, 65 m
walking speeds (mean ± 2 SD) for 103 non-disabled children per minute) compared with before surgery (26, 45, 58 m
and adolescents. (From Rose et al. J Pedatr Orthop 1991;11:571–
578.) Preoperative and postoperative EEI values for self-selected per minute). The comfortable and fast EEI values were
slow, comfortable, and fast values for a 6-year-old boy with spastic lower after surgery (0.78,0.42, and 0.59 beats per minute
diplegic cerebral palsy (CP) who underwent bilateral split tibialis compared with before surgery (0.74, 0.91, and 0.85 beats
posterior tendon transfer and tendoachilles lengthening. per minute).
While documentation of heart rate per distance walked
provides an easily-measured index of energy expenditure
Rose et al. (91) reported normal values for the EEI based and assessment of degree of walking impairment, data
on heart rate obtained from 103 children and adolescents from a recent study by Keefer et al. (41) of 13 children
aged 6 to 18 walking on the floor at self-selected slow, with hemiplegic CP revealed no association between EEI
comfortable, and fast speeds and walking on the tread- and gross V02 during treadmill walking at 0.67, 0.89, and
mill at progressively increasing speeds (Fig. 5-15). A plot 1.12 m per second. Similarly, a lack of relation between
of these values generates a curve showing that the lowest EEI and net V02 was evident at the two slowest walking
EEI value and optimal energy economy occurred at the speeds. Interestingly, analysis of individual data showed
self-selected comfortable walking speeds. EEI values for an unmatched response pattern between EEI and net V02
the self-selected fast walking speeds are significantly lower in a majority of subjects. Taken together, results from
than EEI values for the self- selected slow walking speeds. this investigation suggest that caution should be applied
The average self-selected comfortable floor walking EEI when using EEI to estimate walking energy expenditure in

◗ TABLE 5-9 Self Selected Walking Speeds and Corresponding EEI for Each
Age Groupa
Floor Walking Speed (m/min) EEI (beats/min)
Age (yr) Slow Comfortable Fast Slow Comfortable Fast

6–8 35 ± 9.9 65 ± 8.4 93 ± 13.1 0.75 ± 0.36 0.48 ± 0.15 0.60 ± 0.20
9–11 39 ± 11.0 70 ± 11.1 105 ± 12.0 0.69 ± 0.32 0.47 ± 0.11 0.61 ± 0.18
12–14 37 ± 11.0 76 ± 11.8 106 ± 11.6 0.69 ± 0.27 0.47 ± 0.11 0.58 ± 0.14
15–18 35 ± 11.2 75 ± 8.7 107 ± 11.2 0.70 ± 0.36 0.45 ± 0.14 0.57 ± 0.15
a Values are means ± SD. From Rose J, Gamble JG, Lee J, Lee R, Haskell WL. The energy expenditure index: A method to
quantitate and compare walking energy expenditure for children and adolescents. J Pediatr Orthop 1991; 11:571–578.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:13

100 Chapter 5 / Energetics of Walking

children with cerebral palsy. Because resting heart rate is 18. Cavagna GA, Kaneko M. Mechanical work and efficiency in level
used to determine EEI, and can be influenced by a variety walking and running. J Physiol (Lond) 1977;268:647–681.
19. Cavagna GA, Saibene FP, Margaria R. External work in walking.
of factors, including anxiety, it is possible that an adjusted J Appl Physiol 1963;18:1.
EEI measure that does not incorporate resting heart rate 20. Cooke R. Force generation in muscle. Curr Opin Cell Biol 1990;2:
may more directly reflect walking energy profiles in chil- 62–66.
21. Cooper DM, Weiler-Ravell D, Whipp BJ, Wasserman K. Growth-
dren with CP. Along these lines, preliminary analysis of related changes in oxygen uptake and heart rate during progressive
individual data from the Keefer et al. data set (41) has exercise in children. Pediatr Res 1984;18:845–851.
revealed a much more coherent response between gross 22. Corcoran PI, Brengelmann GL. Oxygen uptake in normal and hand-
icapped subjects, in relation to speed of walking beside velocity-
V02 and a modified EEI value (gross HR/meter). Given controlled cart. Arch Phys Med Rehabil 1970;51:78–87.
this promising initial finding, a reexamination of the as- 23. Corcoran PI, Jebsen RH, Brengelmann GL, Simons BC. Effects of
sociation between V02 and a modified expression of EEI plastic and metal leg braces on speed and energy cost of hemi-
paretic ambulation. Arch Phys Med Rehab 1970;51:69–77.
is warranted in children with CP in order to develop a 24. Cotes JE, Meade F. The energy expenditure and mechanical energy
protocol to assess exercise heart rate during free walk- demand in walking. Ergonomics 1960;3:97–119.
ing and treadmill walking. Such research could prove use- 25. Cotes JE, Meade F, Wise ME. Standardized test exercise. Fed Proc
1957;16:25–26.
ful in providing a simple-to-use field method of describ- 26. Datta SR, Chatteyee BB, Roy BN. The relationship between en-
ing locomotor energy profiles or quantifying changes in ergy expenditure and pulse rates with body weight and the load
walking energy cost following medical intervention and carried during load carrying on the level. Ergonomics 1973;16:
507–513.
rehabilitation. 27. DeJaeger D, Willems PA, Heglund NC. The energy cost of walking
in children. Eur J Physiol E-Pub 11/23/2000.
28. Dempster WT. Space requirements of the ’seated operator: Ge-
REFERENCES ometrical, kinematic, and mechanical aspects of the body with
special reference to the limbs. Wright Air Development Center.
1. American College of Sports Medicine. Guidelines for Graded Exercise Technical Report 55–159. Wright-Patterson Air Force Base, Ohio,
Testing and Exercise Prescription. 7th edition, Lippincott Williams & 1955.
Wilkins; 2006;286–299. 29. Durnin JVGA, Passmore R. Energy, Work, Leisure. London: Heine-
2. Armstrong N, Welsman JR. Assessment and interpretation of mann Educational Books, 1967.
aerobic fitness in children and adolescents. Exer Sport Sci Rev
30. Fenn WO. Work against gravity and work due to velocity changes in
1994;22:435–475.
running: Movements on the center of gravity within the body and
3. Ashkin A, Schutze K, Dziedzic JM, Euteneuer U, Schliwa M. Force
foot pressure on the ground. Am J Physiol 1930;93:42–43.
generation of organelle transport measured in vivo by an infrared
laser trap. Nature 1990;348:346–348. 31. Ganguli S, Datta SR. A new method for prediction of energy expen-
4. Astrand PO, Rodahl K. Textbook of Work Physiology. New York: diture from heart rate. Inst Engineer 1978;58:57–61.
McGraw and Hill, 1970:9, 59, 153, 166, 306–308, 330–338, 361,413. 32. Goldman RF, Iampietro PF. Energy cost of load carriage. J Appl
5. Atzler E, Herbst R. Arbeitsphysiologische Studien. Pfliigers Arch Physiol 1972;17:675–676.
1927;215:291–291. 33. Haisman MF, Goldman RF. The effect of terrain on the energy
6. Bard G, Ralston HJ. Measurement of energy expenditure during cost of walking with back loads and handcart loads. J Appl Phys-
ambulation, with special reference to evaluation of assistive devices. iol 1974;36:545–548.
Arch Phys Med Rehab 1959;40:415–420. 34. Hansen JS. Exercise responses following production of experimen-
7. Bar-Or O. Pediatric sports medicine for the practitioner: From tal obesity. J Appl Physiol 1973;35:587–591.
physiologic principles to clinical applications. New York: Springer- 35. Herfenroeder AC, Schoene RB. Predicting maximum oxygen up-
Verlag; 1983;4–10,22–23. take in adolescents. Am J Dis Child 1989;143:673–677.
8. Bassey EJ, MacDonald IA, Patrick JM. Factors affecting the heart 36. Hirschberg GG, Ralston HJ. Energy cost of stair-climbing in normal
rate during self paced walking. Eur J Appl Physiol 1982;48:105–115. and hemiplegic subjects. Am J Phys Med 1965;44:165.
9. Blessey R. Energy cost of normal walking. Ortho Clin North Am 37. Hong Y, Li JX, Wong ASK, Robinson PD. Effects of load carriage
1978;9:356–358. on heart rate, blood pressure and energy expenditure in children.
10. Bloom WI, Eidex MD. The comparison of energy expenditure in the Ergonomics 2000;43:717–727.
obese and lean. Metabolism 1967;16:685–692. 38. Hoofwijk M, Unnithan V, Bar-Or O. Maximal treadmill performance
11. Bobbert AC. Energy expenditure in level and grade walking. J Appl of children with cerebral palsy. Pediatr Exerc Sci 1995;7:305–313.
Physiol 1960;15:1015–1021. 39. Houston ME. Biochemistry Primer for Exercise Science, in Human
12. Boothby WM, Berkson J, Dunn HL. Studies of the energy Kinetics. 2nd edition. Champaign, IL 2001:65–70.
of metabolism of normal individuals: A standard for basal 40. Katch VL. Use of the oxygen/body weight ratio in correlational anal-
metabolism, with a nomogram for clinical application. Am J Physiol ysis: Spurious correlations and statistical considerations. Med Sci
1936;116:468–484. Sports 1973;5:253–257.
13. Booyens J, Keatinge WR. The expenditure of energy by men and 41. Keefer DJ, Tseh W, Caputo JL, et al. Comparison of direct and indi-
women walking. J Physiol (Lond) 1957;138:165–171. rect measures of walking energy expenditure in children with cere-
14. Brockett JE, Brophy EM, Konish F, et al. Influence of body size, body bral palsy. Dev Med Child Neurol 2004;46:320–324.
fat, nutrient intake and physical fitness on the energy expenditure of 42. Keefer DJ, Tseh W, Caputo JL, Apperson K, McGreal S Morgan DW.
walking. Army Med Nutr Lab Rep 177. Fitzsimmons Army Hospital, Within- and between-day stability of treadmill walking V02 in chil-
Denver, 1956. dren with hemiplegic cerebral palsy. Gait Posture, 2005;21:80–84.
15. Brooks GA, Fahey TD. Exercise Physiology: Human Bioenergetics and 43. Kleiber M. The fire of life. An Introduction to Animal Energetics.
its Applications. New York: John Wiley; 1984:39–54. Huntington NY: Robert E. Krieger; 1975:3–4,214.
16. Butler P, Engelbrecht M, Major RE, Tait JH, Stallard J, Patrick 44. Knapik JJ, Reynolds KL, Harman E. Soldier load carriage: his-
JH. Physiologic cost index of walking of normal children and its torical, physiologic, biomechanical and medical aspects. Mil Med
use as an indicator of physical handicap. Dev Med Child Neurol 2004;169:45–56.
1984;26:607–612. 45. Laursen B, Ekner K, Simonsen EB, Viogt M, Sjogaard G. Kinet-
17. Campbell J, Ball J. Energetics of walking in cerebral palsy. Orthop ics and energetics during uphill and downhill carrying of different
Clin North Am 1978;9:374–377. weights. Appl Ergon 2000;31:159–166.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:12

Chapter 5 / Energetics of Walking 101

46. Lloyd R, Cooke CB. The oxygen consumption associated with un- 70. Morgan DW, Tseh W, Caputo JL, et al. Longitudinal profiles of oxy-
loaded walking and load carriage using two different backpack de- gen uptake during treadmill walking in able-bodied children: the
signs. Eur J Appl Physiol 2000;81:486–492. locomotion energy and growth study. Gait Posture 2001;15:230–
47. McBeath AA, Bahrke M, Balke B. Efficiency of assisted ambula- 235.
tion determined by oxygen consumption measurement. J Bone Joint 71. Morgan DW, Tseh W, Caputo JL, et al. Prediction of the aerobic
Surg 1974;56A:994–1000. demand of walking in children. Med Sci Sports Exerc, 2002;34:2097–
48. McArdle, WD, Katch FI, Katch VL. Exercise Physiology. 5th ed. Bal- 2102.
timore: Lippincott Williams & Wilkins; 2001, p. 39,147. 72. Morgan DW, Tseh W, Caputo JL, et al. Sex differences in run-
49. McCann DJ, Adams WC. A dimensional paradigm for identifying ning economy of young children. Pediatr Exerc Sci 1999;11:
the size-independent cost of walking. Med Sci Sports Exerc 2002; 122–128.
34:1009–1017. 73. Morgan DW, Keefer DJ, Tseh W, et al. Walking energy use in children
50. McCardle WD, Magel JR, Lesmes GR, Pechar GS. Metabolic and with spastic hemiplegia. Pediatr Exerc Sci 2005;17:91–92.
cardiovascular adjustment to work in air and water at 18◦ , 25◦ , and 74. Morgan DW, Tseh W, Caputo JL, et al. Longitudinal stratification of
33◦ C. J Appl Physiol 1976;40:85–90. gait economy in young boys and girls: the locomotion energy and
51. McDonald I. Statistical studies of recorded energy expenditure of growth study. Eur J Appl Physiol 2004;91:30–34.
man. Part II. Expenditure on walking related to weight, sex, age, 75. Muller EA. Der Wirkungsgrad des Gehens. Arbeitsphysiologie
height, speed and gradient [abstract]. Nutr Abstr Rev 1961;31:739– 1950;14:236.
762. 76. Myer MJ, Steudel K. Effect of limb mass and its distribution on the
52. MacGregor J. The objective measurement of physical performance energetic cost of running. J Exp Biol 1985;116:363–373.
with long-term ambulatory physiologic surveillance equipment. 77. Myo-Thein , Lammert O, Garby L. Effect of trunkload on the energy
(L.A.P.S.E.). In: Scott FD, Rafferty EB, Goulding L, eds. Third Inter- expenditure on treadmill walking and ergometer cycling. Eur J Appl
national Symposium on Ambulatory Monitoring. London: Academic, Physiol 1985;54:122–124.
1980:29–38. 78. Pandolf KB, Haisman MF, Goldman RF. Metabolic energy expen-
53. MacGregor J. Rehabilitation ambulatory monitoring. In: Kenedi diture and terrain coefficients for walking on snow. Ergonomic
RM, Paul JP, Hughes J, eds. Disability. London: MacMillan; 1979; 1976;19:683–690.
159–172. 79. Passmore R, Draper MH. Energy Metabolism. In: Albanese A, ed.
Newer Methods of Nutritional Biochemistry. New York: Academic;
54. Malatesta D, Simar D, Dauvilliers Y, et al. Energy cost of walk-
1965.
ing and gait instability in healthy 65- and 80-yr-olds. J Appl Physiol
80. Passmore R, Durnin JVGA. Human energy expenditure. Physiol Rev
2003;95:2248–2256.
1955;35:801.
55. Malatesta D, Simar D, Dauvilliers Y, et al. Aerobic determinants of 81. Paulsen E, Asmussen E. Energy requirements of practical jobs from
the decline in preferred walking speed in healthy, active 65-and 80- pulse increase and ergometer test. Ergonomics 1963;5:33–36.
year olds. Pfluger Arch 2004;447:1915–21 (e-pub Dec. 10, 2003). 82. Perry J. The mechanics of walking in hemiplegia. Clin Orthop Rel
56. Malina RM, Bouchard C, Bar-Or O. Growth, Maturation and Phys- Res 1969 63:23–31.
ical Activity. In: Human Kinetics.2nd ed. Champaign, IL, 2004, 83. Pivarnik JM, Sherman NW. Responses of aerobically fit men and
p. 72. women to uphill versus downhill walking and slow jogging. Med Sci
57. Maltais D, Bar-Or O, Pierrynowski M, Galea V. Repeated treadmill Sports Exerc 1990;22:127–139.
walks affect physiologic responses in children with cerebral palsy. 84. Ralston HJ. Energy-speed relation and optimal speed during level
Med Sci Sports Exerc, 2003;35:1653–1661. walking. Int Z Angew Physiol 1958;17:277.
58. Marey EJ. Movement. Translated by E. Pritchard. New York: Apple- 85. Ralston HJ. Effects of immobilization of various body segments
ton, Century, Crofts, 1895. on the energy cost of human locomotion. Ergonomics 1965;
59. Margaria R. Sulla fisiologia e specialmente suI consumo energetico (suppl):53.
della marcia e della corsa a varie velocita ed inclinazioni del terreno. 86. Ralston HJ, Lukin L. Energy levels of human body segments during
vol 7 Atti Reale Accad. Naz Lincei Giovanni Bardi, Rome, 1938. level walking. Ergonomics 1969;12:39–46.
60. Martin PE. Mechanical and physiologic responses to lower ex- 87. Ralston HJ. Energetics of human walking. In: Herman RM, Grillner
tremity loading during running. Med Sci Sports Exerc 1985;17:427– S, Stein PSG, Stuart DG, eds. Neural Control of Locomotion, New
433. York: Plenum; 1976.
61. Martin PE, Rothstein DE, Larish DD. Effects of age and physical 88. Robertson RJ, Caspersen CJ, Allison TB, Skrinar GS, Abbott RA,
activity status on the speed-aerobic demand relationship of walking. Metz KF. Differentiated perceptions of exertion and energy cost of
J Appl Physiol 1992:73:200–206. young women while carrying loads. Eur J Appl Physiol Occup Physiol
62. Mattsson E, Brostrom LA. The increase in energy cost of walking 1982;49:69–78.
with an immobilized knee or an unstable ankle. Scand J Rehab Med 89. Rogers DM, Turley KR, Kujawa KI, Harper, KM, Wilmore JH. Allo-
1990;22:51–53. metric scaling factors of oxygen uptake during exercise in children.
63. Menier DR, Pugh LGCE. The relation of oxygen intake and ve- Pediatr Exerc Sci 1995;7:12–25.
locity of walking and running, in competition walkers. J Physiol 90. Rose J, Gamble JG, Burgos A, Medeiros J, Haskell WL. Energy ex-
1968;197:717–721. penditure index of activity for normal children and for children with
64. Miller AT, Blyth CS. Influence of body fat content on the metabolic cerebral palsy. Dev Med Child Neurol 1990;32:333–340.
cost of work. J Appl Physiol 1955;88:139–141. 91. Rose J, Gamble JG, Lee J, Lee R, Haskell WL. The energy expen-
65. Molen NH, Rozendal RH. Energy expenditure in normal test sub- diture index: A method to quantitate and compare walking en-
jects walking on a motor-driven treadmill. Proc Kon Ned Akad Wet ergy expenditure for children and adolescents. J Pediatr Orthop
Ser C 1967;75:192. 1991;11:571–578.
66. Molen NH, Rozendal RH. Fundamental characteristics of human 92. Rose J, Gamble JG, Medeiros JM, Burgos SA, Haskell WL. En-
gait in relation to sex and location. Proc Kon Ned Akad Wet Ser C ergy cost of walking in normal children and in those with cerebral
1972;75:215. palsy. Comparison of heart rate and oxygen uptake. J Pediatr Orthop
67. Montoye HS, Ayer T, Nagler F, Howley ET. The oxygen requirement 1989;9:276–279.
for horizontal and grade walking on a motor-driven treadmill. Med 93. Ross J, Schell M. Thermodynamic efficiency in nonlinear biochem-
Sci Sport Exer 1985;17:640–645. ical reactions. 1987;16:401–421.
68. Morgan DW. Economy of locomotion. In: Paediatric exercise science 94. Rush PJ, Pierson WR. The relationship of body surface area,
and medicine. Armstrong N, Mechelen WV, eds. Oxford: Oxford Uni- mass and indices of energy expenditure. Rev Can Biol 1962;21:
versity Press;. 2000:183–190. 1–6.
69. Morgan DW, Boblett K, Caputo J, et al. Comparison of freely-chosen 95. Saltin B, Bloomquist G, Mitchell JH, Johnson Jr RL, Wildenthakk,
and energetically-optimal walking speeds in children with cerebral Chapman CB. Response to submaximal and maximal exercise after
palsy. Pediatr Exerc Sci 1999;11:181–182. bed rest and training [abstract]. Circulation 1968;38(suppl 7).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-05 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:13

102 Chapter 5 / Energetics of Walking

96. Shields ST. The effect of varying lengths of stride on perfor- 108. Waters RL, Hislop HJ, Perry J, Thomas L, Campbell J. Compara-
mance during submaximal treadmill stress testing. J Sports Med tive cost of walking in young and old adults. J Orthop Res 1983;1:
1982;22:66–72. 73–76.
97. Silverman L, Lee G, Plotkin T, Sawyers LA, Yancey AR. Air flow 109. Waters RL, Hislop HJ, Thomas L, Campbell J. Energy cost of
measurements on human subjects with and without respiratory re- walking in normal children and teenagers. Dev Med Child Neurol
sistance at several work rates. Arch Industr Hyg Occup Med 1951; 1983;25:184–188.
3:461. 110. Waters RL, Lunsford BR. Energy expenditure of normal and patho-
98. Skinner JS, Bar-Or O, Bergesteinova V, et al. Comparison of con- logical gait: application to orthotic prescription. In: Atlas of Or-
tinuous and intermittent tests for determining maximal oxygen in- thotics. St Louis: C.V. Mosby Co; 1985.
take in children. Acta Paediatr Scand 1971;217(suppl):24–28. 111. Waters RL, Mulroy S. The energy expenditure of normal and patho-
99. Snellen JW. External work in level and grade walking on a motor- logic gait. Gait and Posture 1999;9:207–231.
driven treadmill. J Appl Physiol 1960;15:759. 112. Wassermann K, Hansen JE, Sue DY, Whipp BJ. Principles of Exercise
100. Soule RG, Goldman RF. Energy cost of loads carried on the head, Testing and Interpretation. Philadelphia: Lea and Febiger; 1986:4–5.
hands or feet. J Appl Physiol 1969;27:687–690. 113. Weir JB de V. New methods for calculating metabolic rate with spe-
101. Spanner DC. Introduction to Thermodynamics. New York: Aca- cial reference to protein metabolism. J Physiol (Lond.) 1949;109:1.
demic; 1965. 114. Welsman JR, Armstrong N, Nevill AM, Winter EM, Kirby BJ. Scal-
102. Strydom NB, Bredell GAG, Genade AJS, Morrison JG, Viljoen JH, ing peak V02 for differences in body size. Med Sci Sports Exerc
van Graan CH. The metabolic cost of marching at 3 m.p.h. over firm 1996;28:259–265.
and sandy surfaces. Int Z Angew Physiol 1966;23:166–171. 115. Wilkie DR. Man as a source of mechanical power. Ergonomics
103. Turrell DJ, Austin RC, Alexander JK. Cardiorespiratory responses 1960;3:1.
of very obese subjects to treadmill exercise. J Lab Clin Med 116. Wilkie DR. The efficiency of muscular contraction. J Mechanochem
1964;64:107–116. Cell Motil 1974;2:257–267.
104. Unnithan BB, Dowling JJ, Frost G, Bar-Or O. Role of cocontraction 117. Winter DA. Analysis of instantaneous energy of normal gait. J
in the 02 cost of walking in children with cerebral palsy. Med Sci Biomech 1976;9:253–257.
Sports Exerc 1996;28:1498–1504. 118. Winter DA. A new definition of mechanical work done in human
105. Van Der Walt WH, Wyndham CH. An equation for prediction of movement. J Appl Physiol 1979;46:79–83.
energy expenditure of walking and running. J Appl Physiol 1973; 119. Workman J, Armstrong BW. Metabolic cost of walking: Equation
34:559–563. and model. J Appl Physiol 1986;61:1369–1374.
106. Volpe AB, Bar-Or O. Energy cost of walking in boys who differ in 120. Wyndham CH, vander Walt WH, van Rensburg AJ, Rogers GG,
adiposity but are matched for body mass. Med Sci Sports Exerc 2003; Strydom NB. The influence of body weight on energy expenditure
35:669–674. during walking on a road and on a treadmill. Int Z Angew Physiol
107. Walker JL, Murray TD, Jackson AS, Morrow JR, Michaud TJ. The 1971;29:285–292.
energy cost of horizontal walking and running in adolescents. Med 121. Zarrugh MY, Todd FN, Ralston HJ. Optimization of energy expen-
Sci Sports Exerc 1999;31:311–322. diture during level walking. Eur J Appl Physiol 1974;33:293–306.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

• C h a p t e r 6
•
◗ Muscle Activity During Walking
Jennette L. Boakes and George T. Rab

Muscles provide the power for human locomotion. Their of the muscle fiber, surrounds each cell. A closer look at the
anatomic, molecular, and chemical structures provide a muscle fiber reveals that it is elegantly designed to gener-
biologically efficient source of power that is controlled by ate force beginning at the molecular level with a series of
an equally exquisite organization of central and peripheral chemical reactions.
nerves with diverse sensors and feedback loops. Muscles Although the gross appearance of individual mus-
and motor control are easy to take for granted since the cles varies widely, all skeletal muscles have a common
system works so effortlessly. Advances in the fields of microstructure composed of similar macromolecules.
molecular biology, ultrastructural anatomy, histology, and Viewed through a light microscope (Fig. 6-2), skeletal mus-
physiology in recent decades have allowed a much bet- cle has a regular pattern of lines or striations clearly visi-
ter understanding of this complex neuromuscular sys- ble (thus the term striated muscle). These striations corre-
tem. Muscles produce active movement by conversion of spond to the basic functional unit of all skeletal muscle, the
metabolic energy into muscle fiber contraction, using both sarcomere, which contains the contractile proteins actin
oxidative and glycolytic metabolism. In the human, mus- and myosin (Figs. 6-1, D–E). Sarcomeres join in a longitu-
cle structure varies with the functional demand: smooth dinal series to form myofibrils of 1 to 2 µm in diameter. At
muscle produces peristaltic contraction in the bowel and each end of the sarcomere is the Z disk, from which em-
sets vascular tone; cardiac muscle powers the contraction anates a group of parallel thin proteins called actin; the free
of the heart; and skeletal muscle produces motion of the ends of the actin face the center of the sarcomere. Here,
joints. Skeletal muscle will be considered in detail here. a second disk (which corresponds to the M line) anchors
parallel thick myosin molecules that interdigitate with the
actin groups (Fig. 6-3) (17,23,27). Myosin has globular pro-
jections that can cross link with the actin molecule; once
MUSCLE STRUCTURE this link has been formed, the globular chains undergo
stearic alteration and bend toward the M line, pulling the
The skeletal muscle system is the largest single organ of actin strands closer to the center of the sarcomere (14,15).
the body. It can be subdivided into muscle groups, which The process repeats as the cross-link is broken, the glob-
all perform similar actions, such as knee flexors or knee ular chain straightens, and a new cross-link is formed.
extensors. Muscle groups are further subdivided into spe- Thus, the two Z disks are pulled closer together by the
cific muscles. Each muscle is composed of individual cells, bending of multiple projections of the myosin, a process
called muscle fibers, which are grouped together into mus- that has been compared with the movement of oars. The
cle fascicles. Each fascicle has its own blood supply and its distance between the Z disks is defined as the sarcomere
own fibrous sheath called the perimysium (Fig. 6-1), and length.
the entire muscle is covered by the epimysium (23).
Skeletal muscle cells are roughly cylindrical with a di-
ameter of 10–100 µm and are up to 20 cm long. The mus-
cle cells contain elements common to every cell in the EXCITATION-CONTRACTION COUPLE
body, such as a cell membrane, mitochondria for oxidative
metabolism, and all of the machinery necessary for protein A group of reservoir-like membranes, the sarcoplasmic
synthesis and cell replication. Muscle cells are multinu- reticulum, and an excitable membrane, the sarcolemma,
cleated and densely packed with contractile proteins and surround the sarcomere. The sarcolemma extends into
energy stores. A complex network of supporting cells and the sarcomere with transverse extensions called T-tubules.
proteins, which are responsible for the growth and repair These membranes convert the electrical signals from the

103
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

104 Chapter 6 / Muscle Activity During Walking

FIGURE 6-1. Basic structural organization of

muscle tissue from macroscopic to molecular. A:
muscle; (B) muscle fiber; (C) myofibril; (D) sar-
comere; (E) molecular level. (From Nordin M,
Frankel VH. Basic Biomechanics of the Muscu-
loskeletal System, 2nd ed. Philadelphia: Lea &
Febiger; 1989.)

nervous system into a mechanical contraction of muscle pump, powered by ATP, returns calcium to the sarcoplas-
to generate force. This process is termed the excitation- mic reticulum.
contraction couple. In order for effective force production
to occur, contraction must be excited along the entire
length of the fiber simultaneously. Motor nerve axons from MOTOR UNIT
the brainstem or the spinal cord branch distally and form
motor end plates, which allow electrochemical transmis- The motor unit is the functional and anatomic grouping
sion between the nerve ending and muscle fibers (Figs. 6-2 of all the muscle fibers connected to a single motor axon.
and 6-4). The motor end plate contains microscopic vesi- It represents the smallest number of fibers that can con-
cles at the axon terminal that release the neurotransmitter tract in a muscle. Each individual axon can supply 3 to
acetylcholine. Acetylcholine diffuses and binds with a re- 200 muscle fibers, depending on the physiologic function
ceptor site on the muscle and initiates the action potential of the muscle. The number of muscle fibers per motor unit
in the sarcolemma. The signal is spread rapidly through is low in the extraocular muscles, where fine control is
the sarcolemma by electrical depolarization. This signals necessary; it is high in leg muscles, where power is more
the rapid release of calcium ions from the sarcoplasmic important. All fibers of a motor unit must necessarily con-
reticulum. The change in intracellular calcium concentra- tract together. The fibers, however, tend to be distributed
tion causes the molecular interaction of actin and myosin throughout a fascicle, so they may not physically touch,
described above. The contraction subsides as a calcium although they are neurologically connected. Motor unit
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

Chapter 6 / Muscle Activity During Walking 105

A B

FIGURE 6-2. (A) Skeletal muscle showing striations and the motor axon with associated end plates.
(From Cormack DH, ed. Ham’s Histology, 9th ed. Philadelphia: JB Lippincott, 1987:398). (B) A myofil-
ament shows several distinct bands, each of which has been given a special letter. The lightest (least
electron dense) band is known as the I band and consists mostly of actin. The wide, dark band, known
as the A band, is composed primarily of myosin. In the center of the I band is the Z-line. In the middle
of the A band is another dense line known as the M line. (Courtesy of R. Lieber.)

contraction is initiated by an action potential in the mo- bic (oxidative) metabolism. Other muscle fibers have a fast
tor axon. When a single action potential reaches a muscle twitch (10–50 msec) and generate a relatively high peak
fiber, there is a brief latent period followed by a contraction tension. Some fast-twitch fibers have few mitochondria
of that fiber. As the brain’s signal for contraction increases, and are unable to maintain high tensions for more than a
it recruits more motor units and increases the ‘firing fre- few contractions without rest; they rely on anaerobic (gly-
quency’ of the units already recruited. The duration of con- colytic) metabolism and are called type IIB, fast fatigable
traction and the tension generated is precisely controlled (FF) or fast glycolytic (FG) fibers. Some of the fibers with
by varying the number of motor units that are recruited faster contraction times are able to maintain force produc-
simultaneously. tion even after a large number of contractions. They tend
to have both oxidative and glycolytic enzymes and thus
are called fast oxidative-glycolytic (FOG) fibers or type IIA.
Types of Muscle Fibers—Physiologic
Because they differ in metabolic enzymes, slow-twitch and
and Histologic Properties
fast-twitch fibers can be distinguished from each other by
Four motor unit types have been described based on the various histochemical staining for ATPase and other enzy-
contractile properties of the motor units such as force, ve- matic activity (8) (Fig. 6-5). The characteristics of muscle
locity, and fatigability (5). Some muscle fibers have a slow fiber types are summarized in Table 6-1.
twitch (60–120 msec) and generate a relatively low peak All fibers in a single motor unit are of the same fiber
tension. Slow-twitch fibers (often called type I) have many type. The recruitment of motor units is determined by
capillaries and contain large quantities of myoglobin and the movement required. Most human muscles contain all
mitochondria; they are fatigue resistant and rely on aero- fiber types in varying proportions. Because slow-twitch
(type I) fibers are fatigue resistant, they are likely recruited
when skeletal muscles are performing activities where
great strength is not a requirement but activity for several
hours is. Fast-twitch (type II) fibers generate high forces,
and more of these are recruited when muscles are used for
short bursts of great strength.
Skeletal muscle is one of the most adaptable tissues in
the body. With training, muscle fibers will increase in size
and strength or increase their oxidative capacity or both
depending on the type of training. Chronic electrical stim-
FIGURE 6-3. Diagrammatic structure of sarcomere. ulation provides one of the purest experimental models of
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

106 Chapter 6 / Muscle Activity During Walking

FIGURE 6-4. Line drawing of

ultrastructure of a motor end
plate. (From Walton J. Disorders
of Voluntary Muscle. Edinburgh:
Churchill Livingstone; 1988:6.)

FIGURE 6-5. Variation in fiber types in normal

human muscle is demonstrated in this cross sec-
tion stained for ATPase activity (pH 4.6). Slow-
twitch fibers stain darker, and fast-twitch stain
lighter (×64). (From Filipe MI, Lake BD. Histo-
chemistry in Pathology. London: Churchill Living-
stone; 1983:93.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

Chapter 6 / Muscle Activity During Walking 107

◗ TABLE 6-1 Characteristics of Muscle Fiber Types

Myosin
Fiber Speed of Fatigue ATPase
Type Color Diameter Contraction Resistance Activity Vascularity Metabolism

I Red Small Slow More Lower Higher Oxidative

IIA White Medium Fast Less Higher Lower Oxidative glycolytic
IIB White Large Fast Less Higher Lower Glycolytic

muscle adaptation to ‘controlled exercise.’ When chronic, twitch of the muscle fibers in the motor unit. Skeletal mus-
low-frequency stimulation similar to the activity of a ‘slow’ cle contraction is maintained by asynchronous activity of
muscle is imposed on a predominantly ‘fast’ muscle, the many α-motoneurons that make up the motoneuron pool
muscle fibers undergo a predictable progression of trans- innervating each muscle. Increasing the force of muscle
forming their metabolic and contractile properties into a contraction is achieved by increasing the firing rate of α-
‘slow’ muscle (19). Conversely, the absence of normal elec- motoneurons and by recruitment (activation of more and
trical innervation, as in spinal cord injury, induces a trans- larger diameter motor units). A given amount of motoneu-
formation from slow-twitch to fast-twitch fiber type (18). ron activity produces a given level of tension; however, a
However, in humans exercise alone without a change in given tension results in different degrees of contraction de-
the neural input does not produce fiber type transforma- pending on the load on the muscle. Feedback is needed to
tion (1); rather, the proportion of each fiber type recruited assure that the intended contraction occurs. This feedback
changes depending on the type of training. is provided by highly specialized neural and muscle fibers
that provide information about muscle and tendon length,
tension, and contraction velocity to the central nervous
Motor Control
system.
A complex system of motor control allows voluntary and The muscle spindle and Golgi tendon organ are two re-
involuntary muscle activity. Motor signals originate in the ceptors particularly important to motor control. Muscle
cortex of the brain and pass through deeper cerebral cen- spindles are located throughout the muscle belly and lie
ters and the spinal cord before they reach muscle. Special- parallel to the skeletal muscle fibers; they provide feed-
ized sensor cells continuously feed information about joint back regulation of muscle length and static stretch. The
and muscle position and movement back to the brain and Golgi tendon organs are connected in series with the ten-
spinal cord, where the motor signal can be modified to pro- don at the musculotendinous junction and are sensitive to
duce smooth, coordinated muscle contractions necessary dynamic changes in length of the muscle (by responding to
for complex movement. active tension of contraction). These sensors combine with
Even the simplest voluntary movement of muscle is the joint position receptors as a “feedback loop” in a complex
result of exceptionally complex central nervous system ac- system of motor control that results in the normal motion
tivity. Cell networks in the cerebral cortex first conceive of of walking.
the movement and communicate with cells in the motor A simple schematic diagram of motor control (typical
cortex. There, under monitoring and direction of other cor- of many theoretical models) is shown in Figure 6-6.
tical and subcortical cell groups, an electrochemical mem-
brane potential is generated; this action potential travels in
the motor neuron axon through the brainstem and spinal WHOLE MUSCLE STRUCTURE
cord along the pyramidal tracts to synapse with a second
motor cell (the anterior horn cell) in the spinal cord. Action Not only is skeletal muscle highly organized microscopi-
potentials from the anterior horn cell travel through ax- cally, it is also highly organized at the macroscopic level.
ons in the peripheral nerve (α-motor neuron) to the motor Muscle architecture refers to the arrangement of muscle
end plate, and stimulate muscle contraction as described fascicles relative to the line of force generation. Muscles
above. At each level, feedback and modification (both in- with fibers oriented parallel to the line of force have paral-
hibitive and facilitative) of the cells can occur. Many extra lel architecture with muscle length roughly equal to fiber
neural cells (interneurons) are involved in the process and length. Muscles with fibers that are oriented at a single an-
assist in integrating and coordinating complex movement gle relative to the line of force are called unipennate. Most
(27). muscles, however, are multipennate with muscle fibers
Variations in muscle contraction depend on feedback oriented at differing angles relative to the line of force
from specialized sensory receptors within the muscle. A generation (Fig. 6-7). Muscle fibers of parallel architecture
single action potential in an α-motoneuron results in a can be much longer than pennate muscle fibers. Longer
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

108 Chapter 6 / Muscle Activity During Walking

fibers allow greater excursion and can contract much

faster. A pennate architecture allows more fascicles to be
packed next to each other in a small space compared with
a parallel architecture, increasing the physiological cross-
sectional area (PCSA) of the muscle. PCSA is the sum of
the cross-sectional areas of all the muscle fibers within the
muscle and is directly proportional to the amount of force
a muscle can generate. Greater pennation allows a muscle
to generate more force at the expense of decreasing overall
excursion. Older anatomy books almost made a “science”
of the different shapes of individual muscles, but neither
fiber length nor PCSA can be easily determined by the gross
appearance of the muscle (12).

Mechanical Properties of Muscle

The mechanical behavior of muscle has been studied at
the whole muscle, isolated muscle fiber, and myofibril lev-
FIGURE 6-6. Simplified diagram of basic functional neural re- els (4,10,12,17,27). The relationships among the muscle
lationships and feedback activities of motor control. length and generated tension, the speed of contraction, and
the balance between elastic and contractile structures at
each level influence muscle biomechanics during human
motion. If the muscle shortens while it is contracting, it is
termed a concentric contraction; eccentric contraction oc-
curs when muscle is lengthening while contracting. When
muscle generates tension but neither shortens, nor length-
ens, and no movement occurs, it is termed an isometric
contraction.

Length and Tension Relationships

An intact mammalian muscle contains contractile ele-
ments (sarcomeres) and elastic components (fascia and
tendons), both of which affect tension in the muscle. Gor-
don, Huxley, and Julian (10) defined the sarcomere length-
tension relationship that forms the basis for understand-
ing muscle force generation (Fig. 6-8). Muscle force varies
as a function of initial sarcomere length, based upon the
amount of myofilament overlap. On the ascending limb
of the length-tension curve, actin filaments from opposite
sides of the sarcomere touch or overlap (Fig. 6-8A). The
overlap interferes with the formation of cross-bridges with
A myosin, so that force generation is impaired. As sarcomere
length increases, so does force generation, until a plateau
is reached (Fig. 6-8B). The plateau occurs because the
myosin molecule contains a “bare region” where no further
cross-bridges with actin can form. Therefore, changes in
B sarcomere length do not increase force production on the
plateau. The plateau region is considered the “optimal sar-
FIGURE 6-7. Gross appearance of two different muscles demon- comere length” for force generation. Increasing sarcomere
strates variation associated with differing functions. A strap mus- length beyond the optimal length results in less myofila-
cle (A) provides high contractile velocity, whereas bipennate (B) ment overlap and forces production decreases (Fig. 6-8C).
muscles generate increased force at the expense of shortening ve-
locity. FL, average fiber length; CSA, average fiber cross-sectional The behavior of whole muscle can be predicted based
area. (From Lieber RL. Skeletal muscle adaptability. I. Review of on knowledge of the microscopic mechanics. Figure 6-9
basic properties. Dev Med Child Neurol 1986;28:390–397.) models a whole muscle exhibiting maximal isometric
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

Chapter 6 / Muscle Activity During Walking 109

FIGURE 6-8. The sarcomere length-tension curve for isolated FIGURE 6-10. Measured isometric length-tension relationship
frog skeletal muscle fibers. Thick line represents muscle max- in forearm flexor muscles of an amputee. Note similarity to theo-
imum tetanic tension and schematic sarcomeres represent the retical curve. T, total tension curve; P, passive tension curve; ,
relative overlap of actin and myosin filaments in the various por- contribution of contractile elements. See text. (From Ralston
tions of the length-tension curve: (A) ascending limb, (B) plateau, et al., Am J Physiol 1947;151:612.)
(C) descending limb. Thin line represents muscle passive tension.
(Modified from Gordon AM, Huxley AF, Julian FJ. The variation
in isometric tension with sarcomere length in vertebrate muscle Experimental laboratory measurements correspond
fibers. J Physiol 1966;184:170–192.) closely to the theoretical model. Figure 6-10 shows actual
length-tension relationships in an amputee who had a skin
tunnel created through the flexor muscles of the forearm.
contraction. (It is generating tension without a change
The active component of the contractile element () is de-
in total muscle length.) The force developed by the mus-
rived by subtracting the passive tension curve from the
cle during isometric contraction varies with its starting
total tension curve.
length. Note that there is an optimal length (Lo ) at which
In most muscles used in walking, there is an optimal
a muscle can generate maximal active contraction. Below
length corresponding to a position where that muscle must
Lo , the elastic components of muscle are slack, but as the
act with strength and efficiency. When posture deviation
muscle elongates passively there is a nonlinear increase
changes that length, the ability of the muscle to generate
in passive elastic tension. The sum of these active and
tension drops (Fig. 6-11), offering a practical illustration
passive tensile components is the overall length-tension
of the active length-tension relationship of muscle. Thus,
relationship.
mechanical behavior of the intact whole muscle is a direct
result of its microstructure.

Speed of Muscle Contraction

While the isometric length-tension relationship has an
anatomic basis, the force-velocity relationship does not.
Muscles produce motion by concentric or eccentric con-
tractions. The velocity of muscle contraction is dependent
on the force resisting the muscle. When a muscle is ac-
tivated to lift a load less than its maximum, the muscle
begins to shorten. A lighter load can be lifted more quickly
than a heavy load. As the load on the muscle increases be-
yond its maximum, the muscle can no longer shorten to
lift the load and begins to elongate while still maximally
contracting—an eccentric contraction. The absolute ten-
sion generated in the muscle during an eccentric contrac-
tion is very high in part because the elastic components
FIGURE 6-9. Theoretical isometric length-tension relationship in addition to the contractile components in muscle act to
of a whole muscle. Active tension (from contraction) and pas-
sive tension (from viscoelastic structures) combine to give total resist lengthening. It is important to understand the con-
tension (heavy line). At the optimal resting length of muscle, the cept of eccentric contraction, because much of the muscle
passive viscoelastic tension drops to zero. activity during gait occurs while it is actively lengthening.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

110 Chapter 6 / Muscle Activity During Walking

FIGURE 6-12. Schematic example of moment arm change with

joint angle. At small joint angles, the moment arm (dotted line) is
relatively small. At larger joint angles, moment arm increases
considerably. (From Lieber RL. Skeletal Muscle Structure and
Function. Williams & Wilkins, 1992:122.)

Muscles do not generate maximum forces throughout

the entire range of motion and may not even generate max-
imum force in the midrange of muscle motion. Studies of
FIGURE 6-11. Hip moment vs. hip abduction plotted for max-
imal strength of hip abduction in three above-knee amputees.
torque production in animals and humans have shown that
Note rapid decrease hip moment caused by muscle shortening the joint angle where the muscle generates maximal force
with progressive abduction. is not the same angle where the moment arm is maximum
(13,20,21). In fact, in many cases, muscle force is maxi-
mized at the extreme end of the range of motion.
Thus, during normal joint rotation, both moment arm
BIOMECHANICS OF MUSCLE and muscle force are constantly changing. The degree of
joint motion depends on both the excursion of the mus-
Most muscles originate proximally in a broad attachment cle and the distance of its lever arm from the joint center.
to either periosteum (the membrane on the surface of The excursion a muscle allows is proportional to its fiber
bone) or fascial sheets. Insertions at the distal end are length. If two muscles have the same fiber length, the one
thinner tendinous attachments to bone. By convention, we with the smaller lever arm will cause the joint to move
usually think of the origin as immobile and the insertion through a greater arc but will require more force to do so.
as moving. In the case of the brachialis muscle powering Muscles that span two joints (such as the hamstrings)
elbow flexion, this may be accurate; in other cases, such as have a more complex mechanical role since the lever arms
the gastrocnemius and soleus group raising a subject on may continually vary (in different ways) at each end of the
tiptoe, neither end of the muscle is truly immobile. muscle as it acts on two separate joints. Also, changes in
Muscles produce movement of a joint by generating lin- the lever arm at one end of a two joint muscle can change
ear force at a distance from the center of the joint. This its length, thus affecting length-tension properties at the
is known as a torque or moment. The moment is calcu- other end. Two joint muscles often allow transfer of energy
lated as the product of force and the distance from the between joints, and they are important to human walking.
muscle insertion to the center of joint movement. The far- An important concept in biomechanics of muscle is
ther the tendon of a muscle lies from the joint center, the work. Positive work is performed when a muscle is con-
greater the moment will be (23). The ability of a muscle centrically contracting—the biceps brachii performs posi-
to produce torque results from the interaction of muscle tive work when a weight is lifted while the elbow is flexing.
properties and joint biomechanics (13,20,21). As a joint Negative work is performed when a muscle is eccentrically
moves, the muscles crossing the joint change in length; this contracting; an example of this would be the activity of the
alters their ability to generate tension at each given joint hamstrings as they contract at the end of swing phase to
angle. The effective lever arm of the muscle also changes decelerate the leg before foot strike. These definitions of
(Fig. 6-12). In some joints, such as the knee, the center of work come from mechanics, where they are analogous to
rotation moves as the knee joint rotates; this characteris- what one would expect from a spring acting on a weight
tic contributes further to the variation in muscle moment (28). Note that these are not metabolic terms; contract-
during motion. ing muscles consume energy whether they are producing
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

Chapter 6 / Muscle Activity During Walking 111

◗ TABLE 6-2 Muscle Actions during Gait Cyclea

Phase of Gait Mechanical Goals Active Muscle Groups Examples

Stance phase
Initial contact Position foot, begin deceleration Ankle dorsiflexors, hip extensors, Anterior tibialis, gluteus maximus,
knee flexors hamstrings
Loading response Accept weight, stabilize pelvis, Knee extensors, hip abductors, Vasti, gluteus medius, gastrocnemius
decelerate mass ankle plantarflexors
Midstance Stabilize knee, preserve momentum Ankle plantarflexors (isometric) Gastrocnemius, soleus
Terminal stance Accelerate mass Ankle plantarflexors (concentric) Gastrocnemius, soleus
Preswing Prepare for swing Hip flexors lliopsoas, rectus femoris
Swing phase
Initial swing Clear foot, vary cadence Ankle dorsiflexors, hip flexors Anterior tibialis, iliopsoas, rectus femoris
Mid swing Clear foot Ankle dorsiflexors Anterior tibialis
Terminal swing Decelerate shank, decelerate leg, Knee flexors, hip extensors, ankle Hamstrings, gluteus maximus, anterior
position foot, prepare for contact dorsiflexors, knee extensors tibialis, vasti
a Modified and derived from Ref. 11.

positive or negative work. However, electromyographic form of contraction. Most muscles responsible for these
(EMG) activity is less in eccentric contraction than in con- contractions have a high proportion of fatigue-resistant
centric contraction and requires less motor unit activity. type I fibers. Shortening contractions, which use more en-
Normal human walking is an energy-efficient activity, ergy, are used only in brief bursts in normal walking.
and it is not surprising that much of the muscle activ- At initial foot contact (Fig. 6-13), the limb begins
ity during walking is isometric or eccentric. This negative to decelerate the body as it reaches the floor. This is
work allows the limbs to absorb energy while resisting the
pull of gravity, yet remains metabolically efficient. Posi-
tive work of muscles during walking allows acceleration
of limbs and powers such activities as push-off and exten-
sion of the hip after foot strike. Negative work holds us
upright, and positive work moves us forward!

PHASIC ACTION OF MUSCLES

IN HUMAN WALKING

Many muscles are active primarily in stance phase or pri-

marily in swing phase. This phasic contraction defines the
role of that muscle in producing normal walking activity.
Contraction that is prolonged or out of phase during walk-
ing may characterize abnormalities of motor control, such
as stroke or cerebral palsy. Out of phase muscle contrac-
tions can cause abnormal movements, or they can occur
while antagonist muscles are trying to produce normal
walking patterns, leading to co-contraction. This obviously
affects both the mechanical and metabolic efficiency of
walking.
In normal walking, muscles contract and relax in a pre-
cise, orchestrated fashion. EMG studies of muscles have
been coupled with accurate three-dimensional measure-
ment of gait movements to clarify the sequence of muscle
recruitment necessary for walking to occur (9,26). Level
walking at a comfortable pace is a natural place to study
phasic activity of muscle groups (Table 6-2). As the walk- FIGURE 6-13. Muscle activity and line of floor reaction force
ing cycle progresses, muscle action becomes primarily at initial foot contact. (From Gage JR. An overview of normal
isometric or eccentric, which is a more energy efficient walking. Instr Course Lect 1990;39:294.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

112 Chapter 6 / Muscle Activity During Walking

FIGURE 6-14. Muscle activity and line of floor reaction force

during loading response. (From Gage JR. An overview of normal
walking. Instr Course Lect 1990;39:297.)

FIGURE 6-15. Muscle activity and line of floor reaction force at

accomplished by simultaneous activity of the knee exten- midstance. (From Gage JR. An overview of normal walking. Instr
sor and flexor muscles to stabilize and position the knee in Course Lect 1990;39:294–297.)
space. Hip extensor contraction decelerates the thigh and
aids in knee extension and foot placement. At the same
time, the anterior tibialis begins a lengthening contraction allows the knee to remain extended without the need for
to gradually “ease down” the foot and keep it from slapping muscle action of the quadriceps.
down on the floor. At terminal stance (Fig. 6-16), the knee extension-ankle
Now the limb begins its loading response (Fig. 6-14) plantarflexion force couple continues to keep the knee pas-
in which it accepts the weight of the body by contract- sively extended, but now the ankle plantar flexors begin a
ing the knee extensors (vasti). The knee bends slightly and shortening contraction that accelerates the body forward.
begins extending under the shortening contraction of the This burst of energy is responsible for most power gener-
knee extensors, acting much like a spring. This knee ex- ation that keeps the body moving forward in normal gait
tension is aided by plantarflexion of the ankle (eccentric (29). There may be a small burst of iliopsoas activity to
contraction), which tends to move the contact point of the lead into the unloading response of preswing.
limb forward and shift the reaction force of the body an- Preswing (Fig. 6-17) begins after the opposite limb has
teriorly at the knee, inducing it to straighten. The gluteus reached the floor and begins to accept weight. In this
medius contracts isometrically and stabilizes the pelvis in phase, ankle plantar flexors are no longer active, and the
the frontal plane. hip flexors (iliopsoas and rectus femoris) begin to lift the
At midstance (Fig. 6-15), the center of gravity of the limb and swing it forward, generally by concentric con-
body has reached its highest point and is carried forward traction. Since the limb behaves like a passive pendu-
by momentum. As long as the knee remains extended, this lum for much of the swing phase, the energy consumed
can be accomplished with very little energy by keeping the in preswing muscle activity is efficiently brief, setting the
foot on the floor and allowing the body mass to “fall for- stage for the largely passive events of the following cycle
ward” in the line of walking progression. Lengthening con- of gait.
traction of the soleus muscle keeps the forefoot pressed Initial swing sees the end of iliopsoas and rectus femoris
against the floor, creating a force couple or linkage that activity. The exact duration of swing and length of stride
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

Chapter 6 / Muscle Activity During Walking 113

FIGURE 6-17. Muscle activity and line of floor reaction force at

preswing. (From Gage JR. An overview of normal walking. Instr
FIGURE 6-16. Muscle activity and line of floor reaction force at Course Lect 1990;39:297–299.)
terminal response stance. (From Gage JR. An overview of normal
walking. Instr Course Lect 1990;39:297–298.)

depend on the length of the freely swinging leg (pendulum)

and the dynamic friction of the knee joint and its associ-
ated tissues (thigh muscles, etc.). Prolongation of swing
phase activity of the hip or knee flexors, or premature con-
traction of the hamstrings, can change the geometry of the
freely swinging leg. For level walking, each person tends to
self-select the optimum speed for individual leg length and
energy efficiency. At the ankle (which was plantarflexed at
preswing), the ankle dorsiflexors begin a shortening con-
traction to allow the foot to clear the floor.
Midswing sees continuation of the passive pendulum
action of the leg. Foot clearance is continued by activity of
the anterior tibialis muscle.
At terminal swing (Fig. 6-18), the limb begins active de-
celeration by contracting the hamstrings eccentrically or
isometrically. This efficiently slows both hip flexion and
knee extension. The knee prepares to accept weight by
early quadriceps activity. The tibialis anterior continues its
activity but switches to an isometric or elongating contrac-
tion as it gently “floats” the foot over the floor immediately
before foot contact. As the foot touches the floor, the cycle
repeats itself. FIGURE 6-18. Muscle activity and line of floor reaction force
During level human walking, muscle contraction is during terminal swing. (From Gage JR. An overview of normal
controlled to produce maximum energy efficiency with walking. Instr Course Lect 1990;39:300.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

114 Chapter 6 / Muscle Activity During Walking

appropriate forward movement. Many muscles responsi- a muscle that is changing length; thus, some investigators
ble for walking contract isometrically, or while length- illustrate EMG activity as simple “on–off ” diagrams (Fig.
ening, to allow maintenance of upright posture against 6-19A). Others attempt to provide more information by
gravity or transfer and storage of energy between limb normalization of the EMG linear envelope (24,29), usually
segments. Brief bursts of more energy-expensive shorten- to the standard of EMG activity during maximal isomet-
ing contraction of muscle are added when needed to pro- ric contraction (Fig. 6-19B). Much variability in reported
vide power for forward motion. data is caused by sensitivity of muscle phasic contractile
Many investigators have studied the electrical activity patterns to walking velocity, but there is also evidence that
responsible for phasic contraction of muscles. The actual normal physiologic walking strategies include some vari-
length of phasic EMG activity of individual muscles will ation in muscle timing stride to stride (24). Depending on
depend on walking speed, age, body size, and the many the presence or absence of neuromuscular disease, and on
technical issues involved in EMG collection. For instance, the specific muscle, EMG data from five to ten gait cycles
cross talk surface electrodes but may be avoided with in- may need to be averaged to obtain a representative sample
dwelling wire electrodes. The magnitude of EMG signals (2). Examples of phasic EMG activity during walking re-
may not be directly proportional to the tension created in ported by different laboratories are shown in Figure 6-19.

FIGURE 6-19. Examples of normal EMG activity during gait. A: “On-off” diagrams of activity from the
University of California, San Francisco based on fine-wire EMG. (Modified from Ref. 22.) B: Curves are
normalized EMG envelopes (data from University of Waterloo). (Modified from Ref. 25.) C: Timing of
functional muscle activity with peak activity noted (data from Rancho Los Amigos Hospital.) (Modified
from Ref. 24.) Differences between these types of data reflect the measure variability in phasic activity
that has been observed during human walking (see text).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

Chapter 6 / Muscle Activity During Walking 115

FIGURE 6-19. (Continued )

P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

116 Chapter 6 / Muscle Activity During Walking

FIGURE 6-20. Phasic action of major muscle groups. Note that most muscles are active at the beginning
and end of swing phase. During midstance, there is minimal muscle activity. This suggests that the
main function of muscle is to accelerate and decelerate the limbs and that after weight acceptance, the
metabolic demands of muscle decrease as momentum allows body weight to advance forward.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

Chapter 6 / Muscle Activity During Walking 117

wave form whose shape depends on the type and location

of electrode, the number of motor unit action potentials
detected, the spatial geometry of the motor unit itself, and
filtering characteristics of muscle tissue. By processing the
raw EMG signal electrically and mathematically, informa-
tion about force generation, motor unit recruitment, and
muscle fatigue may be extracted (3).
The EMG signal is the sum of signals from multiple mo-
tor unit action potential “trains” (MUAPT). Each MUAPT
has a firing rate that depends on time and force generation.
The frequency distribution of the MUAPT firing impulses
can be described by the Fourier transform (frequency anal-
ysis) of the interpulse intervals. The total EMG signal,
FIGURE 6-21. Frequency spectrum of an EMG signal. Me- which is the sum of all MUAPTs, can similarly be ana-
dian and mean frequencies and bandwidth are useful parameters
for describing physiologic muscle function. (From Basmajian
lyzed by calculating the power density spectrum, which
JV, DeLuca CJ. Muscles Alive—Their Functions Revealed by Elec- depends on force, time, and the individual firing rates of
tromyography, 5th ed. Baltimore: Williams & Wilkins; 1985:99.) the MUAPTs (Fig. 6-21). From the power density spectrum,
the mean frequency (average frequency) and median fre-
quency (frequency that divides the spectrum into two re-
gions of equal power) can be calculated.
Estimates of phasic muscle activity and three- As a muscle contracts continuously, it exhibits fatigue.
dimensional motion of the limbs have been pictorially Fatigue eventually leads to a loss in force, pain, tremor, and
combined to give an overall view of gait and muscle ac- decline in work output (3). In addition, there are physiolog-
tion (Figure 6-20). ical and chemical changes in the muscle fiber itself, such
as the increase in lactate concentration. These changes,
known as “ localized muscle fatigue,” cause a decrease in
EMG INTERPRETATION the conduction velocity of the motor unit action potential
and therefore a decrease in the frequency of the EMG sig-
EMG of active muscle can give information about muscle nal (25). This shift of the mean or median frequency of
physiology and motor control beyond issues of timing dur- the power spectrum means that more low frequency of the
ing gait. Modern diagnostic EMG can identify neural in- power spectrum is transmitted to the EMG signal, and this
jury or compression, denervated muscle, or primary patho- power is detected as increased EMG amplitude since the
logic processes. The EMG signal itself is a highly complex muscle tissue acts as a low-pass filter (3). This increase in

A B

FIGURE 6-22. A: As a muscle exhibits isotonic contraction (force) over time, the EMG signal amplitude
(RMS) increases; eventually (at time b) the muscle fatigues and cannot sustain force. B: Power spectrum
of EMG undergoes a shift to lower frequency (from time a to time b) as muscle fatigue progresses
and isotonic contraction is no longer possible. (From Basmajian JV, DeLuca CJ. Muscles Alive—Their
Functions Revealed by Electromyography, 5th ed. Baltimore: Williams & Wilkins, 1985:205.)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-06 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:14

118 Chapter 6 / Muscle Activity During Walking

EMG amplitude and decrease in the mean (or median) fre- 12. Henneman E, Olson CB. Relations between structure and function
quency can be seen during sustained muscle contraction in the design of skeletal muscles. J Neurophysiol 1965;28:581–598.
13. Hoy MG, Zajac FE, Gordon ME. A musculoskeletal model of the
as fatigue occurs (Fig. 6-22). human lower extremity: the effect of muscle, tendon and moment
The ability to use EMG analysis to quantitate localized arm on the moment-angle relationship of musculotendon actuators
muscle fatigue has great importance for the study of ex- at the hip, knee, and ankle. J Biomechanics 1990;23:157–169.
14. Huxley AF. Muscle structure and theories of contraction. Prog Bio-
ercise physiology, athletic training, ergonomics, physical phys 1957;7:255–318.
therapy, and physical medicine (3). The noninvasive na- 15. Huxley HE, Hanson J. Chantes in the cross-striations of muscle
ture of this technique makes it particularly applicable to during contraction and stretch, and their structural interpretation.
Nature 1954;173:973–976.
studies of muscle physiology in animals and humans. 16. Kugelberg E, Edström L. Differential histochemical effect of
muscle contractions on phosphorylase and glycogen in various
types of fibers. Relation to fatigue. J Neurol Neurosurg Psychiatry
1968;31:415–423.
REFERENCES 17. Lieber RL. Skeletal muscle adaptability. I. Review of basic properties.
Dev Med Child Neurol 1986;28:390–397.
1. Always SE, MacDougall JD, Sale DG, Sutton JR, McComas AJ. Func- 18. Lieber RL. Skeletal muscle adaptability. II. Muscle properties follow-
tional and structural adaptations in skeletal muscle of trained ath- ing spinal-cord injury. Dev Med Child Neurol 1986;28:533–542.
letes. J Appl Physiol 1988; 64: 1114–1120. 19. Lieber RL. Skeletal muscle adaptability. III. Muscle properties follow-
2. Ambrosini D, VanEvery T, Kaufman KR, Sutherland DH. Averag- ing chronic electrical stimulation. Dev Med Child Neurol 1986;28:662–
ing of electromyographic data in gait analysis of cerebral palsy 670.
patients [abstract]. West Coast Gait Lab Conference, Los Angeles, 20. Lieber RL, Boakes JL. Sarcomere length and joint kinematics during
1991. torque production in the frog hindlimb. Am J Physiol (Cell Physiol)
3. Basmajian JV, DeLuca CJ. Muscles Alive—Their Functions Revealed 1988;23:C759–C768.
by Electromyography, 5th ed. Baltimore: Williams & Wilkins; 1985. 21. Lieber RL, Boakes JL. Muscle force and moment arm contributions
4. Bourne GH, ed. The Structure and Function of Muscle. New York: to torque production in the frog hindlimb. Am J Physiol (Cell Physiol)
Academic Press; 1960. 1988;23:C769–C772.
5. Burke RE, Levine DN, Tsairis P, Zajac FE III. Physiological types and 22. Lieber RL. Skeletal Muscle Structure and Function. (p. 122). Williams
histological profiles in motor units of the cat gastrocnemius. J Physiol & Wilkins; 1992.
1973:234:723–748. 23. Nordin M, Frankel VH. Basic biomechanics of the musculoskeletal
6. Cormack DH, ed. Ham’s Histology, 9th ed. Philadelphia: JB Lippin- system, 2nd ed. Philadelphia: Lea & Febiger; 1989.
cott; 1987. 24. Perry J. Gait Analysis-Normal and Pathological Function. Thorofare,
7. Edström L, Kugelberg E. Histochemical composition, distribution NJ: Slack; 1992.
of fibers and fatigability of single motor units. J Neurol Neurosurg 25. Stulen FB, DeLuca CJ. Frequency parameters of the myoelectric sig-
Psychiatry 1968;31:424–433. nal as a measure of muscle conduction velocity. IEEE Trans Biomed
8. Filipe MI, Lake BD. Histochemistry in Pathology. London: Churchill Eng 1981;28:515–523.
Livingstone, 1983. 26. Sutherland DH. Gait Disorders in Childhood and Adolescence. (p. 23–
9. Gage JR. An overview of normal walking. Course Lect 1990;39:291– 26). Baltimore: Williams & Wilkins; 1984.
303. 27. Vander AJ, Sherman JH, Lucuano DS. Human Physiology. New York:
10. Gordon AM, Huxley AF, Julian FJ. The variation in isometric ten- McGraw-Hill; 1990.
sion with sarcomere length in vertebrate muscle fibers. J Physiol 28. Winter DA. Biomechanics of Human Movement. New York: John
1966;184:170–192. Wiley & Sons; 1979.
11. Hill AV. The mechanics of active muscle. Proc Soc Lond [Biol] 29. Winter DA. The Biomechanics and Motor Control of Human Gait.
1953;141:104–117. Waterloo: University of Waterloo Press; 1987.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

• C h a p t e r 7
•
◗ Development of Gait
Rosanne Kermoian, M. Elise Johanson, Erin E. Butler,
and Stephen Skinner

Walking is the motor milestone most eagerly awaited by is interspersed with bouts of crawling (5,83). During the
parents and is the hallmark of young children’s emerg- first 3 to 6 months of walking, many gait characteristics
ing independence. Although gait is a complex motor skill, such as velocity, stride length and joint motions change
it begins in infancy and develops rapidly, attaining many rapidly and are only loosely related (18,38). During the
adult characteristics by 7 to 8 years of age. This chapter following months and up to 7 to 8 years of age, the rate
summarizes the changes that occur in children’s gait with of change slows, and the relation between the characteris-
age, addresses the search for mechanisms underlying the tics of gait become more predictable (17,18,83). Bril and
development of gait and the functional consequences of Breniere (18) interpret this pattern of gait acquisition as
children’s mobility. reflecting a two-step process of development. During the
early integration phase, children are learning to combine
the independent elements that contribute to gait for phys-
FIRST FUNCTIONAL CHALLENGES ical growth and a wide range of environmental and task
demands.
Learning to walk is a complex motor problem that is mag- Three particularly notable features of early walking are
nified by factors unique to childhood. The acquisition of (1) the intensity with which gait skills are practiced, (2) en
postural balance during gait is made more difficult by bloc movement patterns (i.e., moving separate parts of the
infants’ body dimensions and their rapid rate of growth body as a single mechanical unit), and (3) marked variabil-
(3,4,88). Relative to adults and older children, infants are ity in a given child’s gait from step to step, trial to trial, and
top-heavy, with large heads and trunks and short legs that day to day.
make them less stable during movement (88). Their short Adolph and colleagues (3,4) have likened the prac-
stature results in faster body sway, requiring more rapid tice schedules of newly walking infants to that of elite
corrections to prevent falls and their physical growth re- athletes and musicians. Converging evidence acquired
sults in the need for ongoing adaptation to changes. Chil- from foot switches placed in the infant’s shoes, daily
dren’s lack of experience also contributes to the difficulty of checklists and telephone diaries of observations by moth-
the motor challenge. Prior to taking the their first steps, for ers who were trained to record their children’s move-
example, children rarely practice the dynamic bipedal and ments show that newly walking infants are upright more
unipedal stance control necessary for independent ambu- than 6 hours per day, averaging between 500 and 1,500
lation (18). walking steps per hour. By the end of each day in-
Despite the magnitude of the motor problem, children fants may have taken 9,000 walking steps and traveled
typically begin to pull-to-stand and take steps while hold- a distance equivalent to the length of 29 football fields
ing onto things around 8 to 10 months of age. Supported (3,4).
walking or “cruising” continues for 2 months, on average, During the first months of walking, practice is dis-
with crawling continuing as a means of getting from place tributed throughout the day, a pattern that may be critical
to place. During supported walking, children require ex- to maintaining the infants’ motivation to practice their new
ternal support to maintain balance and their movements gait skills and to consolidating their motor learning (3).
are characterized by erratic joint motions, variable step Walking occurs in different locations and over indoor and
rates and step lengths, and inconsistent timing of muscle outdoor surfaces that vary in height, rigidity, and texture
actions (36,81,87). (3,4). Walking occurs for a variety of purposes as diverse as
The onset of independent walking typically begins attaining a desired object to following a sibling from place
around the first birthday and, as with supported walking, to place (49).

119
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

120 Chapter 7 / Development of Gait

Early walking is characterized by en bloc movement slower rate than between 1 and 3 years of age (67,83,89,95).
patterns that can be seen in the stiff coupling of body The principal reduction in cadence occurs between the
parts (e.g., trunk and arms) and co-activation of mus- ages of 1 and 2 years, and thereafter declines gradually
cle groups. These movement patterns can be observed until adulthood. Although cadence decreases between 3.5
during motor learning throughout life, as in the stiff and 7 years of age, it remains approximately 26% higher
legged posture of novice skiers. En bloc movement pat- than the normal adult mean (83).
terns are thought to simplify the motor problem by Velocity, stride length and cadence are only loosely re-
decreasing the number of biomechanical degrees of free- lated in the first few months of walking, which leads to
dom that must simultaneously be controlled, thereby lim- instability and limited speed (38). By 4 to 5 years of age,
iting the number of components of a new skill that must velocity, stride length, and cadence become closely related
be learned at the same time (61). It has been hypothesized and exhibit a more adultlike pattern, one in which both ca-
that en bloc movement patterns in early walking make it dence and stride length are increased as a means to walk
possible for infants to focus on the critical skills of main- faster.
taining an upright posture while moving forward (18).
Measurements of early walking in children are charac-
Kinematics
terized by marked levels of variability that decrease with
age and walking experience (18,83). This variability in the Children begin walking around 12 months of age with rel-
gait of newly walking children has been attributed to a atively stiff, slightly flexed hips and knees, a flat foot-strike,
number different of factors including poor motor control and a wide base of support. The arms are held away from
and interlimb coordination, inconsistent walking speed, the body in a symmetrical position for balance. They lack
active exploration of different movement patterns, and the reciprocal arm swing (83). With age, the flexed posture of
challenge of obtaining accurate gait measurements from the hips and knees normalizes (6,34,81) and a well-defined
young children (18,27,32). heel strike at initial contact occurs. The base of support
(step width) decreases and there is a lowered arm posture
(34,53,83). The arm and leg movements progress from an
CHARACTERIZING CHILDREN’S GAIT
ipsilaterally synchronous pattern to a reciprocal pattern
(34,87).
Measurements of time and distance (temporal-spatial) pa-
Head: Head motion becomes more predictable with
rameters, joint motions, muscle activation, energy cost,
age, a process that may continue past 12 years of age
and ground reaction forces during walking have proved
(7,45). During the first 10 to 15 weeks of walking, lateral
to be sensitive indicators of changes in gait patterns that
and anterior-posterior movement of the head and trunk de-
occur as children increase in age and gain walking expe-
creases, a change that parallels decreases in step width and
rience. Understanding the timing of the changes in these
double limb support and increases in velocity and stride
measurements and how they are linked to other neurolog-
length (54). During the following 4 months of walking,
ical and behavioral events provides insight into the mech-
head motion is minimized by stiff coupling of the head to
anisms underlying children’s development. To date, the
the trunk so that they move as a single unit. Consequently,
most comprehensive characterization of children’s gait is
as the head is translated upward during initial swing, the
that of Sutherland, Olshen, Biden and Wyatt (83), whose
orbital region of the eye and vestibular apparatus of the
rich description of walking patterns in children ages one
inner ear also move upward; as the head translates down-
to seven years has provided a solid foundation for the char-
ward during terminal stance the position of the eyes and
acterization of children’s gait.
ears move downward.
After walking 18 months, children show some evidence
Temporal-Spatial Parameters
of the mature pattern of head motion (18,55). In mature
Changes in gait that are associated with age and walking gait, there is greater disassociation of the head and trunk.
experience are described by measurements of temporal- As the head necessarily translates upward during initial
spatial parameters: velocity, stride length, cadence, and swing, it tilts downward approximately the same number
time spent in single or double limb support. Following the of degrees as the upward translation so that the position
first few months of walking, gait velocity, stride length, of the eyes and ears remain level. As the head translates
and single limb stance begin to increase while cadence downward during terminal stance, the head tilts upward.
and double limb support decrease (83). As leg length and Compensatory head tilt in response to head translation
height continue to increase, stride length and velocity also during gait thus serves to keep the orbital region of the
increase. With age, there are further developments in the eye and vestibular apparatus of the inner ear at a constant
temporal-spatial parameters of gait (9,83) that are associ- height relative to the ground throughout the gait cycle.
ated with changes in the size of body segments. However, This pattern of head movement contributes to postural
after 3 years of age the increase occurs at a significantly control during gait by providing a stable platform for the
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

Chapter 7 / Development of Gait 121

visual and vestibular sensory systems in the same way that

keeping the head oriented toward a specific spot helps
ballet dancers maintain their balance during a pirorette
(44,66).
Trunk and upper extremities: A wide base of support
and lateral motion of the trunk characterize early walking.
By approximately 2.5 years of age, the base of support re-
duces to within the width of the pelvis and movement of
the trunk is minimized (19).
During early walking, children’s arms are typically held
stiffly in a high-guard position (i.e., both shoulders raised
in abduction and external rotation with elbow flexion). As
step width decreases, the arms lower to a middle- and low-
guard position (34,53). The raised and symmetrical po-
sitioning of the arms during early walking is thought to
play a role in maintaining the body in an upright posture
while moving forward, the arms being lowered as pos-
tural balance improves (53). Reciprocal arm swing, the FIGURE 7-1. Ankle motion during gait at one year of age (n =
mature form of arm movement during gait, in which the 45) and seven years of age (n = 46). Note the lack of dorsiflexion
leg and contralateral arm move forward in synchrony, is at initial contact (0% gait cycle) and in terminal swing (80% to
seen in 50% of children by 1.5 years of age and 100% by 90% gait cycle) for the one-year-olds. Dorsiflexion is increased
in stance phase (35% to 50% gait cycle) for the seven-year-olds.
4 years of age (83). Arm swing and free moments tend to (Adapted from Sutherland D et al. The Development of Mature
reinforce each other and act to balance the trunk torques Walking, Philadelphia: Lippincott 1988; 93, 97.)
induced by the lower limb (56).
Pelvis: During early walking the pelvis is more anteri- Ankle: In early walkers, 1 to 1.5 years of age, a combina-
orly tilted than in mature walking. In 1- to 2-year-old chil- tion of ankle plantar flexion and knee flexion result in a flat
dren, in the frontal plane, swing phase is accompanied by foot at initial contact. A consistent heel strike is achieved at
an ipsilateral pelvic hike above neutral. By age 2.5 years, an average of 22 weeks after independent walking (19) or
the pelvis drops below neutral in swing. Changes in pelvic 18 months of age (83). Once heel contact is established, the
obliquity can be explained, in part, by the progressive nar- ankle plantar flexes during the loading response. With age,
rowing of the wide base of support (83). Pelvic rotation as the body is moved over the stance limb in single limb
in the transverse plane emerges approximately 4 weeks af- support, the degree of ankle dorsiflexion increases. This
ter independent walking (19), and by age 3 years there is a change has been attributed to the developing strength of
trend toward a lower dynamic range of pelvic rotation (83). the calf muscles that control tibial advancement over the
Hip: Early walking is characterized by increased supporting limb (47). In general, swing phase dorsiflexion
flexion and external rotation of the hip throughout stance is present in early walking but increases with age until nor-
(83). Full hip extension during single limb support in ter- mal adult motions are achieved. The described changes in
minal stance is normally seen by age 2 years; hip rotation ankle motion are illustrated in Figure 7-1.
throughout the gait cycle stabilizes in mature form by
age 3 years (83). Hip abduction and adduction are closely
Kinetics
linked to pelvic obliquity, with progressively less stance
phase adduction observed until approximately 2.5 years of In a study of children ranging from 11 months to 14 years
age when pelvic obliquity stabilizes in mature form (83). of age, Beck et al. (9) found a trend toward decreasing
Knee: The knee is flexed throughout the gait cycle when force amplitudes until 5 years of age, after which the verti-
children begin to walk. This differs from the mature gait cal, fore-aft and medial-lateral ground reaction forces re-
pattern in which the knee is extended throughout most of mained constant at all three walking speeds studied (slow,
stance phase, with the exception of loading response and comfortable, and fast). The initial decrease in the relative
preswing. By age 2 years, there is an increase in knee flex- magnitude of the ground reaction force was hypothesized
ion during the loading response (81,83). However, a ma- to indicate the development of a more efficient gait, based
ture knee pattern from initial contact to terminal stance on the assumption that a reduction in the floor reaction
is not consistently observed until age 4 years. The sec- forces reflects a reduction in muscular effort (9). Other
ond phase of knee flexion during the gait cycle occurs just investigators, reporting age differences in ground reac-
prior to swing (preswing) to advance the limb and is in- tion force, have found that children show changes in the
distinguishable from the adult pattern by 2 years of age amplitude and timing of kinetic data until 9 years of age
(83). (29,62,84). The difficulty of collecting force data in young
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

122 Chapter 7 / Development of Gait

FIGURE 7-2. Average times of onset and cessa-

tion of EMG activity as measured by surface elec-
trodes over the tibialis anterior in each of 10 age
groups expressed as percentages of the gait cycle.
For the one-year-olds, the activation is prolonged
during stance, lasting until nearly toe off. This ac-
tivity is co-contraction with the calf muscles that
are also active in stance. (From Sutherland et al.
The Development of Mature Walking 1988;155.)

children may contribute to the discrepancy in age-related in stance. With age, the swing phase activity of the calf
changes. Walking speed may also account for the discrep- disappears. However, approximately 25% of children from
ancy. Stansfield et al. (78) demonstrated that normalized age 2 to 7 years have premature firing of the ankle plantar
speed, and not age, characterized ground reaction force flexors in swing and early stance phase (83).
patterns in 5- to 12-year-old children when speed, step The observed co-contraction in newly walking children
length, cadence, and force were normalized to unitless and emergence of mature patterns of muscle activation
quantities using the method described by Hof and Zijlstra has been attributed to a number of different factors. One
(43). explanation is that co-contraction maintains the body in
equilibrium in an upright position and that the integration
of the stretch-reflex activity into the preprogrammed leg
Muscle Timing
muscle EMG with age corresponds to an increase in gas-
New walkers is characterized by prolonged activity in most trocnemius activity stabilizing the body during gait (11).
muscle groups, resulting in co-contraction of antagonist Forssberg (36) and others (e.g., 58) attribute the emerging
leg muscles during the gait cycle. With age, the duration pattern of muscle activation to innate central pattern gen-
of muscle activity shortens and a reciprocal pattern is ob- erators (CPGs) in the spinal cord that are hypothesized to
served between agonist and antagonist muscles. This pat- produce both infant stepping and generate the basic loco-
tern is most notable in the quadriceps and hamstrings, motor rhythm in adults. Further work is needed to deter-
and calf and pretibial muscle groups (83,87). In new walk- mine the role of the stretch reflex, supraspinal influences,
ers, there is nearly continuous activity in the quadriceps and other factors in shaping and fine-tuning muscle activ-
throughout the gait cycle, which is thought to be essential ity during gait (35).
to the maintenance of an upright posture in the presence
of excessive knee flexion in stance that is characteristic of
Effort and Energy Cost
early walking (83). Hamstring muscle activity is also pro-
longed and has similar timing to that of the quadriceps The amount of energy required for walking is directly
muscles, beginning in mid swing and ending at the end of proportional to walking speed at all ages. Oxygen con-
single limb stance. By the age of 2 years, the hamstrings sumption measured in children walking at slow, comfort-
and quadriceps achieve nearly normal timing (83). able, and fast speeds demonstrates that small children
At the ankle, the pretibial muscles are active at initial have a higher expenditure of energy during walking than
contact and loading response, secondary to the flat foot teenagers or adults (92). A linear relationship between age
contact, with prolonged activity in stance. From age 2 to and oxygen consumption per kilogram of body weight per
7 years, the pretibial muscles terminate activity after the minute and oxygen consumption per kilogram per me-
beginning of single limb stance and resume their activity at ter walked has been observed, with the youngest subjects
initial swing, similar to that of mature walking. Figure 7-2 (6 years of age) having the highest values of oxygen con-
demonstrates the prolonged activation of the tibialis an- sumption. In a study of the energy cost of walking as a func-
terior during stance at younger ages. For most new walk- tion of walking speed in children from 3 to 12 years, the net
ers, the calf group is active in terminal swing as well as cost of transport of the 3-to-4-year olds and 5-to-6-year olds
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

Chapter 7 / Development of Gait 123

was found to be 70% and 40% greater, respectively, than in ten reported as a percentage of walking speed for non-
adults (31). impaired children of the same age. Measurements of self-
Measurement of heart rate can also provide an estimate selected walking speed are related to children’s age and
of the effort required for walking; at least at submaximal height, increase with walking experience and are asso-
levels, oxygen uptake and heart rate are linearly related. ciated with children’s lowest energy expenditure. Thus
Rose et al. (70), using treadmill ambulation, determined they provide a useful baseline characterization of gait in
that oxygen consumption and heart rate in children were clinical settings, where the primary goal of surgical or
linearly related over a wide range of walking speeds (22 to rehabilitative therapy is often to improve children’s self-
130 m per minute). selected walking speed.
In contrast, gait measurements collected while children
are walking at their fastest speed reveal their ability to
CURRENT DIRECTIONS IN
change speed and the strategy by which they walk faster.
CHARACTERIZING CHILDREN’S GAIT
The normal strategy observed in mature gait is to increase
both the stride length and cadence linearly. In immature
Adaptation to Environmental and
gait, factors which contribute to an increase in stride
Task Demands
length such as heel strike, hip extension in late midstance,
Flexible adaptation of gait to current conditions is the and full knee extension in terminal swing may not yet be
hallmark of mature walking, contributing to such critical developed. In gait pathology, stride length can be restricted
functions as initiating gait, changing speed or direction, by joint contracture, spasticity, or musculoskeletal defor-
stepping over obstacles without interrupting ongoing gait, mity leaving increased cadence as the only strategy for
walking through narrow openings, and stopping before walking faster. Recent data collected by Abel and Damiano
reaching an intended target of locomotion (59). Character- (1) have shown that the strategy children use to change
izing age-related changes in the way children adapt their their walking speed differentiates between the gait of chil-
gait to environmental and task demands is a relatively new dren with cerebral palsy and their nonimpaired peers. Chil-
and promising area of investigation. dren with cerebral palsy, although able to increase their
Anticipatory control: Children’s ability to flexibly walking speed, rely more on increasing cadence than stride
adapt their gait in anticipation of future motor and sen- length to change speed.
sory events improves with age. Data to date suggest that The advantages of measuring speed, stride length, and
anticipatory control during gait develops more slowly and cadence at both children’s self-selected free and fast speed
is more challenging than skills tapped by more traditional in clinical populations is illustrated in Figure 7-3. The
measures of gait. The ability to respond to external threats plot illustrates the relation between speed, stride length,
to balance during the gait initiation process, for example, and cadence for a child instructed to walk at both self-
does not emerge until 2 to 3 years of age, and children do selected and fast walking speeds. The darkened reference
not react efficiently to perturbation during gait initiation line indicates the expected values for a non-impaired child
until after 4 to 5 years of age (99). Consistently anticipat- of the same height walking at the same speed. The data
ing a change in the direction of walking by turning the demonstrate the compromised stride length and dramatic
head prior to pivoting does not occur before 5.5 years of increase in cadence in the barefoot condition when a
age and even at 8.5 years of age has not attained the adult 5-year-old child with spina bifida is asked to walk at a fast
pattern of turning the head a full second prior to pivot- speed. When the child wears ankle-foot orthoses (braced
ing the body (37). Difficulty with anticipatory control may condition) he is able to increase his walking speed by
play an important role in mild gait problems, such as those increasing both stride length and cadence.
seen in children who have been identified as clumsy, and
thus be more sensitive to the effects of premature birth
Minimizing Variance in
and subtle neurologic difficulties than other measures of
Gait Measurements
gait (40,94).
Mechanics of changing walking speed: Todd et al. A serious limitation of standard gait measurements for
(89) were among the first to propose that children’s abil- children is that they yield a wide range of “normal” values
ity to vary their walking speed and the strategy by which due to unspecified factors. Much of the variability observed
they do so are sensitive measures of gait function. Their in the data of children who are the same age is thought to
data demonstrate that children modify their speed in a be caused by factors that are specific to childhood, such as
predictable manner, using characteristic combinations of individual differences in the rate of growth (e.g., height and
cadence and stride lengths. limb length) and individual differences in the rate of motor
Time and distance variables (speed, stride length, ca- skill acquisition. The extent to which children’s ages are not
dence) are typically recorded while children are instructed perfectly correlated with their growth or motor skill will
to walk at a self-selected comfortable speed. For children necessarily result in unexplained variance in measurement
with disabilities, their self-selected walking speed is of- (91). Differences in children’s self-selected walking speed
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

124 Chapter 7 / Development of Gait

FIGURE 7-3. Speed, stride length, and ca-

dence data from a five-year-old boy, 105 cm
in height, with spina bifida can be visualized
on this graph. The child was instructed to
walk at his self-selected and fast speed in two
conditions: braced (with orthoses) and bare-
foot. For each condition the self-selected and
fast walking data points (circled) are con-
nected by a line. The expected relationship
between speed, stride length, and cadence
for a typically-developing child of the same
height is indicated by the solid line drawn
along the dashed reference lines for height
in centimeters. In the barefoot condition, he
walks slower compared to the braced con-
dition even when he is attempting to walk
at his fastest speed, and he is unable to in-
crease his stride length. In order to increase
his walking speed, the child increases his ca-
dence from 122 steps per minute to 180 steps
per minute without an increase in stride
length. With the braces on, his self-selected
and his fastest speed are greater than his
barefoot speed. The additional stability pro-
vided by the braces enabled him to increase
both his stride length and cadence propor-
tionally to achieve the faster walking speed.
(From Todd FN, Lamoreux LW, Skinner SR,
et al. J Bone Joint Surg Am 1989;71-A:196–
204.)

may also increase the variability in gait data by differen- et al. (91) have used this technique to eliminate the effects
tially affecting temporal-spatial parameters, joint kinemat- of height to study the changes in a number of other vari-
ics and kinetics, and muscle timing (9,78,79,90). Currently, ables including walking speed, step length, and cadence.
there are a number of techniques being used to minimize Other techniques that have been used to eliminate vari-
the variability in children’s gait data so as to better under- ability due to height include statistical detrending with re-
stand the mechanisms contributing to the development of spect to two or more variables. This normalization method
gait, and to increase the comparability of data from differ- involves iteratively removing the trends between two or
ent children or individual children followed over time. more variables, such as height and stride length. O’Malley
To minimize variability in gait data owing to height, (63), using Sutherland’s data to compare several statistical
Hof and Zijlstra (43) proposed a nondimensional scaling methods for removing the effects of height and age, recom-
technique in which the effect of height can be accounted mended simultaneous normalization when two or more
for in a mechanically consistent way using a pendulum variables are involved. This method has the additional ben-
model of gait. The technique uses geometrical scaling in efit of retaining the original units of measurement.
which leg length and acceleration caused by gravity are To minimize variability that results from changes in
the only factors: The data are unitless. The assumption walking speed as well as height, Todd et al. (89) used data
is made that any variability remaining in the data after from 324 non-impaired children to develop a set of equa-
these factors are accounted for must be caused by factors tions that predict the appropriate stride length (in cm) for
other than height. Stansfield et al. (78,79) and Vaughn a child of a specific height walking at his or her chosen
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

Chapter 7 / Development of Gait 125

walking speed. Placing the emphasis on stride length studies in which Thelen and colleagues experimentally ma-
rather than walking speed reduces the variability that is in- nipulated infants’ leg weight and strength. If young in-
troduced by children voluntarily varying their self-selected fants who were still producing neonatal stepping move-
speed. Normalizing stride length for height and walking ments had their legs weighted to simulate the leg weight
speed assures that children with disabilities are not penal- gained over the first 2 months of life, they could no longer
ized merely because they are smaller in stature or walk at step (88). Yet, if young infants who were no longer spon-
a slower speed than their peers. taneously producing neonatal stepping movements had
The equations that were used to define the normal ref- their legs partially submerged in a tank of water to min-
erence curves in Figure 7-3 describe the expected relation imize the effect of gravity, they were once again able to
between speed, stride length, and cadence for children who step (88). Furthermore, when infants were held over a mo-
are between 80 and 170 cm in height. The equations are torized treadmill to augment their leg strength, they pro-
presented in the Appendix along with two sets of constants, duced stepping movements throughout the first year of life,
one for boys and one for girls. Measurements of a child’s well past the time the “stepping reflex” had disappeared
height and free walking speed are used to calculate the and prior to the time they were spontaneously produc-
predicted stride length. Measurement of the child’s actual ing independent stepping movements (86). Taken together,
stride length can be expressed as a percentage of the pre- these studies show how the peripheral factors of leg weight
dicted stride length. This method can be used to express and muscle strength can reliably alter the developmental
the degree of walking disability or changes in functional course of one component of independent walking, that of
status over time. stepping.
The search for peripheral factors contributing to gait
has resulted in research on the role of experience in the de-
THE SEARCH FOR MECHANISMS velopment of gait. Such investigations are surprisingly di-
UNDERLYING THE DEVELOPMENT verse, ranging from the effects of prenatal movement expe-
OF GAIT rience (e.g., breech positioning during the third trimester
results in increased likelihood of hip flexion during gait
Over the past decade, there has been a distinct shift in at 12 to 18 months of age [76]), to the effect of seasonal
the study of gait development from characterizing gait in and environmental influences on motor development (e.g.,
children of different ages to searching for the mechanisms bundling of Inuit children in the fall and winter results in
that bring about developmental change. Why do infants delayed walking onset relative to spring and summer, un-
around the world typically begin to walk between 12 and published Kaplan-Estrin, et al).
18 months of age and look strikingly similar from the time Gait is determined by interplay among factors: Al-
that they begin to negotiate their first steps? though peripheral factors may independently account for
Historically, the answers to these questions were some changes in gait development, it is likely that nei-
straightforward. The emergence of the major motor mile- ther peripheral nor central factors function in isolation.
stones, such as walking, and the subsequent refinement Recent studies demonstrate how multiple factors, each
of these skills was attributed to maturation of the central operating independently, can interact in a way that
nervous system (CNS). Although elegant in its simplicity, either facilitates or limits the emergence of walking. In
the brain-to-behavior causal link has been seriously chal- these studies, neuromaturation is typically treated as one
lenged over the past decade by theory and empirical evi- factor among many. Work by Woollacott et al. (100) and
dence, suggesting that the CNS may not always have pri- Sienko-Thomas et al. (74), for example, provide an ele-
mary status in the development of gait (85). gant demonstration of how central and peripheral factors
Peripheral factors contribute to the development of jointly considered elucidate our understanding of mech-
gait: Among the most powerful tests of neuromaturational anisms constraining the development of gait in children
explanations of gait are studies demonstrating how periph- with cerebral palsy. To test the effects of abnormal lower ex-
eral factors (non-CNS factors), such as the weight of the tremity postures observed in children with cerebral palsy
lower extremities, can determine the emergence of new on muscular responses in standing and walking, non-
motor patterns. Thelen’s studies of infant stepping were impaired children were positioned in a crouching posture
among the first to demonstrate that the development of and activation patterns in the lower extremity musculature
a behavior that had been attributed to the CNS could be were recorded. Standing in the abnormal posture elicited
determined by change in peripheral factors (85). Thelen muscular responses in non-impaired children that resem-
hypothesized that neonatal stepping movements (the bled that of children with cerebral palsy, including activa-
“stepping reflex”) would seemingly disappear as young in- tion of muscles in a proximal to distal fashion and exces-
fants gained leg fat and subsequently reappear toward the sive co-contraction. These studies demonstrate one means
end of the first year of life when their leg muscles became by which CNS damage, as evidenced by muscle and joint
strong enough to lift their limbs against gravity. Among contracture and crouched position, may affect the devel-
the most convincing data supporting this explanation are opment of standing and gait.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

126 Chapter 7 / Development of Gait

FUNCTIONAL CONSEQUENCES cross-sectional and longitudinal data demonstrate that

OF MOBILITY the onset of walking is associated with increased sensi-
tivity of the otolith but not the canal system, regardless
Physiological Change of the age at which the infants begin to walk (96,97).
These findings demonstrate that development of the otolith
Bone mineral density: In children, as in adults, the act system occurs with the onset of walking, although fur-
of walking contributes to health by improving aerobic ther work is needed to determine whether walking expe-
capacity, resting metabolic rate, body composition, and rience brings about changes in the otolith system or vice
fat metabolism. Many studies on the effect of mobility versa.
on physiology address the issue of bone mineral accrual. Mobility improves visual proprioception. Adults control
Mechanical loading associated with physical activity in- their posture and perceive their direction of self-motion
creases bone mineral density in children (57). Intermittent from the pattern of visual motion (optical flow) generated
weight-bearing, such as occurs during walking, is thought as they move (6). This coupling between posture and mo-
to be more effective for bone accrual than static weight- tion in large areas of the visual field has been demon-
bearing (13,52,71). Not surprisingly, children with mobil- strated even in neonates, who make directionally appro-
ity deficits, such as those with cerebral palsy, have been priate compensatory movements of their heads to optical
found to be at higher risk for fractures and osteoporo- flow (46). Only after mobility onset, however, does the to-
sis than can be predicted by their nutritional deficiencies tal amount of visual stimulation needed to control posture
and abnormal vitamin D metabolism from anticonvulsant decrease significantly: Optical flow in the periphery of the
medication (25,41,73). Using a number of different mea- visual field, as compared to movement of the entire visual
sures of bone mineral accrual (e.g., bone mineral den- field, is sufficient to elicit directionally appropriate postu-
sity, bone mineral composition and broadband ultrasound ral adjustments in infants who are independently mobile
attenuation, a measure of disruption of the microstruc- but not in pre-mobile infants of the same age (12,42). Only
tural parameters), findings across studies show that skele- mobile infants respond to the same illusion of movement
tal fragility in children is directly related to the extent that adults experience when seated in a stationary train
of their mobility (26,77,98). Children who walk indepen- that is passed by a moving train.
dently have the highest bone density, followed by children
who walk with a mobility aid, children who are non-mobile
but who regularly weight-bear in a standing frame, and Neural Change
children who are non-mobile and do not weight-bear. Fur- Just as the CNS plays a role in the process of learning
ther work is needed to determine whether there are sensi- to walk, learning to walk may in turn affect brain de-
tive periods during development when ambulation is most velopment. Neural change that is associated with the ac-
critical to bone mineral accrual. However, recent data sug- quisition of complex motor skills has been attributed to
gest that intermittent weight-bearing may be particularly neuroplasticity, the active reciprocal process of specific
important during the two-year window of the pubertal experience shaping neural connectivity (22). This process
growth spurt when human growth hormone and sex hor- occurs during both development and normal learning (82).
mones surge and approximately 25% of the adult skeleton Recent data demonstrate that complex motor skill learn-
develops (65). ing in adults results in selective anatomical changes in the
Increased sensitivity of the vestibular and visual cortex (e.g., learning to juggle increases gray matter in cor-
sensory systems: The onset of mobility is associated tical areas that are associated with the processing and stor-
with increased responsiveness of the vestibular and vi- age of complex visual motion [33]) and selective functional
sual sensory systems. Both systems, although functional changes in the cortex (e.g., training in a complex finger
from a very young age, undergo prolonged postnatal de- movement task results in remodeling of the somatosen-
velopment and become more finely tuned with mobility sory cortex [60]). In children, the effect of neuroplasticity
experience. is thought to be even more powerful than in adults be-
Although the vestibular system is anatomically differ- cause there are more uncommitted resources in the brain
entiated at birth, the capacity for the CNS to utilize head and less competition for these resources among developing
movement information transmitted by the otolith and functional systems (51). Thus, it is reasonable to assume
canal system of the inner ear does not reach adult-level that the act of learning to walk has a positive effect on the
functioning until approximately 12 years of age (64). Re- structure and function of the brain.
search on delays in postural and locomotor development Although this issue has not been investigated directly,
associated with hypoactive or absent vestibular function data on pre-walking mobility are consistent with the hy-
suggests that the otolith system, in particular, facilitates pothesis that the demands associated with learning a
walking because of its role as a gravitational and iner- complex mobility skill bring about changes in cortical
tial force detector (64). Consistent with that prediction, organization. Bell and Fox (10) found a higher degree of
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

Chapter 7 / Development of Gait 127

connectivity and synaptic elaboration, as measured by to non-mobile children of the same age are more likely
resting EEG coherence, between intrahemispheric frontal, to generalize information to new settings. For example,
parietal, and occipital sites in children who had been they are more likely to activate a new toy if it works in
crawling 1 to 4 weeks than in pre-mobile children or chil- the same way as one they saw demonstrated 24 hours
dren who had been crawling for more than 4 weeks. All before in a different location (39). This important skill
children were 8 months of age. This pattern of findings underlies children’s development in domains as seem-
points to the possibility that learning how to crawl results ingly unrelated to mobility as language learning (72).
in an initial overproduction of cortico-cortical connections ■ The need and opportunity for children to make
that is followed by subsequent pruning of unused synapses independent decisions occurs constantly during mo-
as the motor skill of crawling becomes more routine. A bility, teaching children that they have their own unique
more efficient pattern of interconnectivity may also follow wishes and desires, a critical step in socioemotional
the onset of walking. development. The consequences of increased autonomy
can be seen in the “testing of wills” between newly
walking children and their parents. Children who are
Behavioral Change mobile are more likely to persist in their activities
The emergence of independent mobility has widespread and express anger if their activities are thwarted than
and enduring functional consequences for children’s cog- same-age peers (48). Parents, in turn, begin using the
nitive, perceptual, and socioemotional development (e.g., word “no” with their mobile children and expecting
16,20,49). Before children are mobile, they spend relatively them to comply with their demands (24).
long periods of time in one place and their play is limited
to objects that are within arms’ reach. Following mobility Converging evidence from studies using different re-
onset, they have many new opportunities. Learning how search designs demonstrate that the dramatic changes in
to navigate through the environment is particularly chal- behavior are not merely a correlate of independent mo-
lenging. A few specific demands associated with safe and bility but a consequence of it. In general, children of the
successful navigation and selected findings from studies same age who are crawling on hand-and-knees perform at
on the effects of mobility on children’s development are higher levels than children who are not yet mobile (48,49).
listed below. Children who cannot move independently, but who can
move from place to place with the aid of a baby walker, be-
■ Visual attention to the environment is critical in have similarly to infants who are crawling. Case studies of
order to monitor the route, and avoid collisions and infants with mobility delays caused by heavy casting that
falls. Wariness of heights, a skill that is essential to safe prevented crawling or spina bifida (L4-5 or below) perform
mobility, develops after mobility onset even though poorly until after they begin to move, at which time they
children can perceive heights long before they move show a spurt in performance.
independently (23). Essential features of mobility: Behavioral changes
■ Working memory is needed in order to keep a destina- associated with mobility emerge slowly over a period of
tion in mind, regardless of the time it takes to get there weeks or months and improve concomitantly with gains
and distractions encountered along the way. Object in motor skill. This pattern of acquisition suggests that
permanence, the ability to remember where an object movement parameters such as speed, distance traveled and
has been hidden, is the hallmark of infant intelligence efficiency of movement, either separately or jointly con-
and a measure of working memory. Object permanence sidered, may be essential for behavioral change. A direct
improves after mobility onset (8,10,50) and continues link between movement parameters and outcome is sup-
to improve the longer children have been independently ported by the fact that mobility in a baby walker, a fast
mobile (50). and efficient form of movement, is associated with behav-
■ Strategies for remembering the route are necessary ioral change; whereas, belly crawling, a slow and effortful
so as to avoid getting lost. Following mobility onset form of movement, is not (8,50). Walking, as compared
infants are significantly more likely to use landmarks to pre-walking mobility, results in higher movement ve-
and environmental cues to locate objects, even when locity, efficiency, greater distance traveled, and total time
they are confined to a seated position (28,48,49). spent moving (3,4). Studies to date suggest that walking
■ The ability to remember and transfer information magnifies the changes that have been documented with
acquired in one setting to similar problems en- creeping on hands-and-knees (e.g., 2,28). The gains in
countered in another setting is essential to successful biomechanical efficiency that come with upright mobility
navigation in new locations. One of the many things chil- may be particularly important to behavioral and nueral
dren need to learn is that doorways, regardless of their change, making it possible for children to pay less at-
size and shape, are the place one enters or leaves a room. tention to the act of moving and more attention to the
Findings indicate that mobile children when compared environment.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

128 Chapter 7 / Development of Gait

Clinical populations: Taken together, these stud- 2. Adolph KE. Learning in the development of infant locomotion.
ies demonstrate important functional consequences of Monogr Soc Res Child Dev 1997;62:1–158
3. Adolph KE. Learning to keep balance. Adv Child Dev Behav 2002;30:
mobility. Under typical circumstances, mobility can facil- 1–40.
itate other areas of children’s development. However, it is 4. Adolph KE, Avolio AM. Walking infants adapt locomotion to chang-
unlikely that it is the only means by which children can ing body dimensions. J Exp Psychol Hum Percept and Perform
2000;26:1148–1166.
learn (69). Unfortunately, few studies have examined the 5. Adolph KE, Vereijken B, Shrout PE. What changes in infant walking
effect of mobility delays or disability on infants and chil- and why. Child Dev 2003;74:475–497.
dren with orthopaedic or neuromuscular disorders. Inter- 6. Andersen GJ. Perception of self-motion: psychophysical and com-
putational approaches. Psychol Bull 1986;99:52–65.
pretation of the findings from studies to date is limited 7. Assaiante C, Amblard, B. Ontogenesis of head stabilisation in space
by the possibility that the children’s underlying medical during locomotion in children: influence of visual cues. Exp Brain
condition is responsible for both learning and mobility. Res 1993;93:499–515.
8. Bai DL, Bertenthal BI. Locomotor status and the development of
In addition, the studies do not capitalize on the power spatial search skills. Child Dev 1992;63:215–226.
of experimental or longitudinal designs. The sample sizes 9. Beck RJ, Andriacchi TP, Kuo KN, et al. Changes in the gait pat-
are small and heterogeneous. Keeping these limitations in terns of growing children. J Bone Joint Surg Am 1981;63:1452–
1457.
mind, the most conservative interpretation is that mobility, 10. Bell MA, Fox, NA. Individual differences in object permanence per-
although not necessary for the acquisition of rudimentary formance at 8 months: Locomotor experience and brain electrical
skills, commonly affects the rate of acquisition and final activity. Dev Psychobiol 1997;31:287–297.
11. Berger W, Altenmueller E, Dietz V. Normal and impaired develop-
level of skill attainment of a number of spatial and so- ment of children’s gait. Hum Neurobiol 1984;3:163–170.
cial skills (68,75,80,93). When powered mobility aids are 12. Bertenthal BI, Bai DL. Infants’ sensitivity to optic flow for control-
provided, mobility has beneficial and predictable conse- ling posture. Dev Psychol 1989;25:936–945.
13. Biewener AA, Bertram JE. Structural response of growing bone to
quences, even for children with severe motor and cognitive exercise and disuse. J Appl Physiol 1994;76:946–955.
deficits (15,21,30). 14. Biringen Z, Emde, RN, Campos JJ, Appelbaum MI. Affective reorga-
nization in the infant, the mother, and the dyad: the role of upright
locomotion and its timing. Child Dev 1995;66:499–514.
15. Bottos M, Bolcati C, Sciuto L, Ruggeri C, Feliciangeli A. Powered
SUMMARY wheelchairs and independence in young children with tetraplegia.
Dev Med Child Neurol 2001;43:769–777.
The study of children’s gait over the past decade has led to a 16. Bremner, JG. Motor abilities as causal agents in infant cognitive de-
velopment. In: Salvesberg G, ed. The Development of Co-ordination
more complete characterization of how gait changes with in Infancy. Amsterdam: Elsevier, 1993:347–377.
age and walking experience. To address the wide range 17. Bril B, Breniere Y. Posture and independent locomotion in early
of “normal values” found in children’s gait measurements childhood: learning to walk or learning dynamic postural control?
In: Salvesberg G, ed. The Development of Co-ordination in Infancy.
due to unexplained factors, there is an increased focus Amsterdam: Elsevier, 1993:337–358.
on the factors that systematically contribute to measure- 18. Bril B, Ledebt A. Head coordination as a means to assist sensory
ment variability. One technique for normalizing gait data integration in learning to walk. Neurosci Biobehav Rev 1998;22:
555–563.
to eliminate variability due to height and walking speed is 19. Burnett CN, Johnson EW. Development of gait in childhood. II. Dev
presented in this chapter. To better address issues of gait Med Child Neurol 1971;13:207–215.
function, research characterizing gait now encompasses 20. Bushnell EW, Boudreau JP. Motor development and the mind: The
potential role of motor abilities as a determinant of aspects of per-
studies of children’s ability to adapt their gait to environ- ceptual development. Child Dev 1993;64:1005–1021.
mental and task demands. These skills (e.g., stepping over 21. Butler C. Effects of powered mobility on self-initiated behaviors of
an obstacle without stopping) are notably poor in imma- very young children with locomotor disability. Dev Med Child Neurol
1986;28:325–332.
ture and pathological gait. 22. Byl NN, Merzenich M. Principles of neuroplasticity: implications
In addition to advances in characterizing gait, gait re- for neurorehabilitation and learning. In: Gonzalez EG Myers SJ,
search has broadened to include studies of the mech- Edelstein JE, Lieberman JS, Downey JA eds. Downey and Darling’s
Physiological Basis of Rehabilitation Medicine, 3rd ed. Boston: But-
anisms underlying age-related changes in gait and the terworth Heinemann, 2001:609–628.
functional consequences of mobility for other domains 23. Campos JJ, Bertenthal B, Kermoian R. Early experience and emo-
of children’s development. Investigations of mechanisms tional development: The emergence of wariness of heights. Psychol
Sci 1992;3:61–64.
underlying the development of gait have highlighted the 24. Campos JJ, Kermoian R, Zumbahlen MR. Socioemotional transfor-
importance of biomechanical and experiential factors in mation in the family system following infant crawling onset. New
promoting development. Studies of the functional conse- Dir Child Dev 1992;55:25–40.
25. Chad KE, McKay HA, Zello GA, Bailey A, Faulkner A, Snyder RE.
quences of mobility have revealed surprisingly widespread Body composition in nutritionally adequate ambulatory and non-
and enduring effects of mobility for children’s physiologi- ambulatory children with cerebral palsy and a healthy reference
cal, neural, and behavioral development. group. Dev Med Child Neurol 2000;42:334–339.
26. Chad EK, Bailey DA, McKay HA, Zello GA, Snyder RE. The effect of
a weight-bearing physical activity program on bone mineral content
and estimated volumetric density in children with spastic cerebral
REFERENCES palsy. J Pediatr 1999;135:115–117.
27. Clark JE, Whitall J, Phillips SJ. Human interlimb coordination: the
1. Abel MF, Damiano DL. Strategies for increasing walking speed in first 6 months of independent walking. Dev Psychobiol 1998;21:445–
diplegic cerebral palsy. J Pediatr Ortho 1996;16:753–758. 456.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

Chapter 7 / Development of Gait 129

28. Clearfield MW. The role of crawling and walking experience in infant tion during the first months of independent walking. Neuroreport
spatial memory. J Exp Child Psycho 2004;89:214–241. 1995;6:1737–1740.
29. Cupp T, Oeffinger D, Tylkowski C, et al. Age-related kinetic changes 55. Ledebt A, Wiener-Vacher S. Head co-ordination in the sagittal plane
in normal pediatrics. J Pediatr Orthop 1999;19:475–478. in toddlers during walking: preliminary results. Brain Research Bul-
30. Deitz J, Swinth Y, White O. Powered mobility and preschoolers letin 1996;40:371–373.
with complex developmental delays. Am J Occup Ther 2002;56: 56. Li Y, Wang W, Crompton RH, Gunther MM. Free vertical movements
86–96. and transverse forces in human walking and their role in relation
31. De Jaeger D, Willems PA, Heglund NC. The energy cost of walking to arm-swing. J Exp Biol 2001;204:47–58.
in children. Pflugers Arch 2001;441:538–543. 57. MacKelvie KJ, Khan KM, McKay HA. Is there a critical period for
32. Diop M, Rahmani A, Calmels P, et al. Influence of speed variation bone response to weight bearing exercise in children and adoles-
and age on the intrasubject variability of ground reaction forces and cents? A systematic review. Br J Sports Med 2002;36:250–257.
spatiotemporal parameters of children’s normal gait. Ann Readapt 58. Marder E, Bucher D. Central pattern generators and the control of
Med Phys 2004;47:72–80. rhythmic movements. Curr Biology 2001;11:R986–R996.
33. Draganski B, Baser C, Busch V, Schuierer G, Bogdahn U, May 59. Mattiello D, Woollacott M. Posture control in children: development
A. Changes in grey matter induced by training. Nature 2004;427: in typical populations and in children with cerebral palsy and down
311–312. syndrome. In: Connolly KJ, Forssberg H, eds. Neurophysiology and
34. Farmer SE. Key factors in the development of lower limb coordi- Neuropsychology of Motor Development: Clinics in Developmental
nation: implications for the acquisition of walking in children with Medicine No. 143/144. London: Mac Keith Press; 1977;54–77.
cerebral palsy. Disabil Rehabil 2003;25:807–816. 60. Merzenich M, Recanzone M, Jenkins W, Grajski KA. Adaptive mech-
35. Forssberg H. Neural control of human motor development. Curr anisms in cortical networks underlying cortical contributions to
Opin Neurobiol 1999;9:676–682. learning and nondeclarative memory. Cold Spring Harbor Symposia
36. Forssberg H. Ontogeny of human locomotor control. I. Infant on Quantitative Biology, LV, 1990:873–887.
stepping, supported locomotion and transition to independent 61. Newell K. Change in movement and skill: learning, retention and
locomotion. Exp Brain Res 1985;57:480–493. transfer. In: Latash M, Turvey M, eds. Dexterity and Its Development.
37. Grasso R, Assaiante C, Prevost P, Berthoz A. Development of an- Mahwah: LEA; 1996:393–430.
ticipatory orienting strategies during locomotor tasks in children. 62. Noguchi M. (Development of gait in children–characteristics of gait
Neurosci Biobehav Rev 1998;22:533–539. in children from the viewpoint of ground reaction force) Nippon
38. Grieve DW, Gear RJ. The relationship between length of stride, step Seikeigeka Gakkai Zasshi 1986;60:787–799.
frequency, time of swing and speed of walking for children and 63. O’Malley MJ. Normalization of temporal-distance parameters in pe-
adults. Ergonomics 1966;9:379–399. diatric gait. J Biomech 1996;29:619–625.
39. Herbert J, Gross J, Hayne H. The effect of age and locomotor expe- 64. Ornitz EM. Normal and pathological maturation of vestibular func-
rience on deferred imitation in human infants. Dev Psych. (In Press) tion in the human child. In Romand R, ed. Development of Au-
40. Hempel MS. Neurological development during toddling age in nor- ditory and Vestibular Systems. New York: Academic Press; 1983:
mal children and children at risk of developmental disorder. Early 479–535.
Hum Dev 1992;34:47–57. 65. Petit MA, McKay HA, MacKelvie KJ, Heinonen A, Khan KM, Beck
41. Henderson RC. Bone density and other possible predictors of frac- TJ. A randomized school-based jumping intervention confers site
ture risk in children and adolescents with spastic cerebral palsy. and maturity-specific benefits on bone structural properties in girls:
J Bone Joint Surg Am 1995;77:1617–1618. hip structural analysis study. J Bone Miner Res 2002;17:363–372.
42. Higgins CI, Campos JJ, Kermoian R. Effect of self-produced loco- 66. Pozzo T , Levik Y, Berthoz A. Head and trunk movements in the
motion on infant postural compensation to optic flow. Dev Psychol frontal plane during complex dynamic equilibrium tasks in humans.
1996;32:836–841. Exp Brain Res 1995;106:327–338.
43. Hof AL, Zijlstra W. Comment on “Normalization of temporal- 67. Preis S, Klemms A, Muller K. Gait analysis by measuring ground
distance parameters in pediatric gait.” J Biomech 1997;30:299,301– reaction forces in children: changes to an adaptive gait pattern
302. between the ages of one and five years. Dev Med Child Neurol
44. Holands MA, Sorensen KL, Patla AE. Effects of head immobilization 1997;39:228–233.
on the coordination and control of head and body reorientation and 68. Rendeli C, Salvaggio E, Sciascia-Cannizzaro G, Bianchi E, Cal-
translation during steering. Exp Brain Res 2001;140:223–233. darelli M, Guzzetta F. Does locomotion improve the cognitive pro-
45. Holt KG, Ratcliffe R, Jeng SF. Head stability in walking in children file of children with meningomyelocele? Child’s Nerv Syst 2002;
with cerebral palsy and in children and adults without neurological 18:231–234.
impairment. Phys Ther 1999;79:1153–1162. 69. Riviere J, Lecuyer R. Spatial cognition in young children with spinal
46. Jouen F. Visual-proprioceptive control of posture in newborn in- muscular atrophy. Dev Neuropsychol 2002;21:273–283.
fants. In: Amblard B, Berthoz A, Clarac F, eds. Posture and Gait: 70. Rose J, Gamble JG, Medeiros J, Buygos A, Haskell WL. Energy cost
Development, Adaptation and Modulation. Elsevier Science Publish- of walking in normal children and in those with cerebral palsy: com-
ers B.V.; 1988:59–65. parison of heart rate and oxygen uptake. J Pediatr Orthop 1989;
47. Katoh M, Mochizuki T, Moriyama A. Changes of sagittal-plane an- 9:276–279.
kle motion and ground reaction force (fore-aft shear) in normal 71. Rubin CT, Lanyon LE. Regulation of bone formation by applied
children aged 4 to 10 years. Dev Med Child Neurol 1993;35:417–423. loads. J Bone Joint Surg Am 1984;66:397–402.
48. Kermoian R. Locomotor experience and psychological development 72. Schafer G. Infants can learn decontextualized words before their
in infancy. In: Furumasu J, ed. Pediatric Powered Mobility: Develop- first birthday. Child Dev 2005;76:87–96.
mental Perspectives, Technical Issues, Clinical Approaches. Washing- 73. Shaw NJ, Weite CP, Fraser WD, Rosenbloom L. Osteopenia in cere-
ton, DC: RESNA Press; 1997;7–21. bral palsy. Arch Phys Med Rehab 1994;7:235–238.
49. Kermoian R. Locomotor experience facilitates psychological func- 74. Sienko-Thomas S, Moore, CA, Kelp-Lenane C, Norris C. Simulated
tioning. In: Gray D, Quatrano L, Lieberman M, eds. Designing and gait patterns: the resulting effects on gait parameters, dynamic, elec-
Using Assistive Technology: The Human Perspective. Baltimore, MD: tromyography, joint moments, and physiological cost index. Gait
Paul Brookes Publishing Company; 1997:249–266. Posture 1996;4:100–107.
50. Kermoian R, Campos JJ. Locomotor experience: a facilitator of spa- 75. Simms B. The route learning ability of young people with Spina
tial cognitive development. Child Dev 1988;18:299–307. Bifida and hydrocephalus and their able-bodied peers. Z Kinderchir
51. Knudsen EI. Sensitive Periods in the Development of the Brain and 1987;42(suppl 1):53–56.
Behavior. J Cogn Neurosci 2004;16:1412–1425. 76. Sival DA, Prechtl HF, Sonder GH, Touwen BC. The effect of intra-
52. Lanyon LE, Rubin CT. Static vs dynamic loads as an influence on uterine breech position on postnatal motor functions of the lower
bone remodelling. J Biomech 1984;17:897–905. limbs. Early Hum Dev 1993;32:161–176.
53. Ledebt A, Bril B. Acquisition of upper body stability during walking 77. Stallings VA, Cronk CE, Zemel BS, Charney EB. Body composi-
in toddlers. Dev Psychobiol 2000;36:311–324. tion in children with spastic quadriplegic cerebral palsy. J Pediatr
54. Ledebt A, Bril B, Wiener-Vacher S. Trunk and head stabiliza- 1995;126:833–839.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-07 Rose- 2252G GRBT092-Rose-v3.cls November 10, 2005 0:51

130 Chapter 7 / Development of Gait

78. Stansfield BW, Hillman SJ, Hazlewood ME, et al. Normalized 91. Vaughn CL, Langerak NG, O’Malley MJ. Neuromaturation of hu-
speed, not age, characterizes ground reaction force patterns in 5- to man locomotion revealed by non-dimensional scaling. Exp Brain
12-year-old children walking at self-selected speeds. J Pediatr Orthop Res 2003;153:123–127.
2001;21:395–402. 92. Waters RL, Hislop HJ, Thomas L, et al. Energy cost of walking in
79. Stansfield BW, Hillman SJ, Hazlewood ME, et al. Sagittal joint kine- normal children and teenagers. Dev Med Child Neurol 1983;25:184–
matics, moments, and powers are predominantly characterized by 188.
speed of progression, not age, in normal children. J Pediatr Orthop 93. Wedell K, Newman CV, Reid P, Bradbury IR. An exploratory study
2001;21:403–411. of the relationship between size constancy and experience of mo-
80. Stanton D, Wilson PN, Foreman N. Effects of early mobility bility in cerebral palsied children. Dev Med Child Neurol 1972;14:
on shortcut performance in a simulated maze. Beh Brain Res 615–620.
2002;136(1):61–66. 94. Weisglas-Kuperus N, Baerts W, Fetter WP, et al. Minor neurological
81. Statham L, Murray MP. Early walking patterns of normal children. dysfunction and quality of movement in relation to neonatal cere-
Clin Orthop 1971;70:8–24. bral damage and subsequent development. Dev Med Child Neurol
82. Stiles J. Neural plasticity and cognitive development. Dev Neuropsy- 1994;36:727–735.
chol 2000;18:237–272. 95. Wheelwright EF, Minns RA, Law HT, et al. Temporal and spatial
83. Sutherland D, Olshen R, Biden E, et al. The Development of Mature parameters of gait in children. I: Normal control data. Dev Med
Walking. Philadelphia: Lippincott; 1988. Child Neurol 1993;35:102–113.
84. Takegami Y. Wave pattern of ground reaction force of growing 96. Wiener-Vacher S, Ledebt A, Bril B. Changes in otolithic vestibulo-
children. J Pediatr Orthop 1992;12:522–526. ocular reflex of off vertical axis rotation in infants learning to
85. Thelen E. Motor development: a new synthesis. Am Psychol walk: preliminary results of a longitudinal study. Ann N Y Acad Sci
1995;50:79–94. 1996;781:709–712.
86. Thelan E. Treadmill-elicited stepping in seven-month-old infants. 97. Wiener-Vacher SR, Toupet F, Narcy P. Canal and otolith vestibulo-
Child Dev 1986;57:1498–1506. ocular reflexes to vertical and off vertical axis rotations in chil-
87. Thelen E, Cooke DW. Relationship between newborn stepping dren learning to walk. Act Otolaryngol (Stockh) 1996;116:657–
and later walking: a new interpretation. Dev Med Child Neurol 665.
1987;29:380–393. 98. Wilmshurst S, Ward K, Adams JE, Langton CM, Mughal MZ. Mo-
88. Thelen E, Fisher DM, Ridley-Johnson R. The relationship be- bility status and bone density in cerebral palsy. Arch Dis Childhood
tween physical growth and a newborn reflex. Infant Behavior Dev 1996;7:5,164–165.
1984;7:479–493. 99. Woollacott M, Assaiante C. Developmental changes in compen-
89. Todd FN, Lamoreux LW, Skinner SR, et al. Variations in the gait satory reponses to unexpected resistance of leg lift during gait ini-
of normal children. A graph applicable to the documentation of tiation. Exp Brain Res 2002;144:385–396.
abnormalities. J Bone Joint Surg Am 1989;71:196–204. 100. Woollacott M, Burtner P, Jensen J, Jasiewicz J, Roncesvalles N,
90. van der Linden ML, Kerr AM, Hazlewood ME, et al. Kinematic and Sveistrup H. Development of postural responses during standing
kinetic gait characteristics of normal children walking at a range of in healthy children and children with spastic diplegia. Neuroscience
clinically relevant speeds. J Pediatr Orthop 2002;22:800–806. Biobehav Rev 1998;22:583–589.

APPENDIX
The set of equations described below make it possible to For example, using the equations to calculate the
determine if a child achieves an appropriate stride length predicted stride length for a 6-year-old girl, 112 cm in
given their body height and chosen walking speed. The height, who is walking 53.3 m per minute, yields the
graph shown in Figure 7-3 illustrates the mathematical re- following:
lationship between stride length, cadence, and body height Intermediate value A = 9.8; intermediate value B = 11.
for normal children walking over a range of speeds. In or- Thus, her predicted stride length in cm. = 82.5 cm. If her
der to relate mathematically the gait variables to height, actual measured stride length is 76.5 cm, her stride length
a three-dimensional surface was fitted to experimental normalized to height and walking speed is 93% of her non-
data from 324 children (Todd et al., 1989). Data points impaired peers.
were plotted against three perpendicular scales represent-
ing walking speed, stride length, and body height.
Two equations define the normal reference curves in ACKNOWLEDGMENTS
Figure 7-3, one for boys, and one for girls. Each equation
depends on six constants (C1-C6), which, taken together This project was supported in part by a Distinguished
with a child’s height, enable computation of a second or- Research Fellowship from the National Institute of Dis-
der curve relating stride length and walking speed. In these ability and Rehabilitation Research (NIDRR) awarded
computations, “H” represents the child’s height in cen- to Rosanne Kermoian. M. Elise Johanson’s contribution
timeters, and “C1” through “C6” are constants, defined in was made possible in part by the Department of Veteran
the table below, which were derived from normal data. Two Affairs, Rehabilitation Research and Development Service,
intermediate values (A and B) are calculated as follows: Project #B2785R.
√
Stride Length = A ∗ Speed + B
∗ ∗
A = C1 H 2+ C2 H + C3 B = C4 ∗ H 2 + C5 ∗ H + C6
C1 C2 C3 C4 C5 C6
BOYS -0.0002802 0.1544256 -3.5589657 -0.0010762 0.2803360 -10.3155200
GIRLS -0.0005145 0.1896516 -5.0245599 -0.0002684 0.2465649 -13.2757848
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

• C h a p t e r 8
•
◗ Gait Adaptations in Adulthood:
Pregnancy, Aging, and Alcoholism
Erin E. Butler, Maurice Druzin, and Edith V. Sullivan

In this chapter we describe the changes to the normal Physiological Changes and Posture
adult gait pattern that occur with pregnancy, aging, and
The majority of women experience musculoskeletal pain
alcoholism. We present an overview of the physiologi-
during the course of pregnancy (85). On average, the to-
cal processes enabling healthy gait and balance and how
tal weight gain is approximately 9 to 14 kg (2,197). Most
changes with pregnancy and the normal aging process
of the weight gained is due to the enlarging uterus, fe-
contribute to changes in the normal adult gait pattern.
tus, and breasts. The lower trunk has significantly greater
We also identify some factors that can predict when
rates of change than all other body segments during the
a person will fall and provide information about prac-
second and third trimester of pregnancy (97). This weight
tices that can reduce the risk of falling and improve gait
gain leads to increased forces on the articular cartilage at
and overall health. Finally, we discuss the changes that
the joints. As the fetus grows, the position of the mother’s
occur with acute alcohol use and chronic alcoholism,
center of gravity moves superiorly and anteriorly (68). The
where the effects of alcohol disrupt normal balance and
developing fetal load places an increased demand on the
gait.
lumbar spine and abdominal muscles (92). The changing
From the time gait matures in childhood (28,209), it
shape and inertia of the lower trunk requires postural ad-
remains stable through midlife. For example, velocity, one
justments that can cause back pain.
of the most basic parameters of gait, remains relatively
Increased ligamentous laxity during pregnancy allows
unchanged between the ages of 10 to 59 years (25,156). A
for greater joint motion at the pelvis (3) and the peripheral
decline in the comfortable walking speed does not occur
joints (32,52,188). Beginning in the tenth to twelfth week
until after 60 years of age. Many gait variables are highly
of pregnancy, the sacroiliac synchondroses and symph-
correlated with walking speed (118), suggesting that gait
ysis pubis begin to widen. The hormone relaxin is thought
kinematics also remain stable until the age of 60 years.
to be a major contributor to increases in joint laxity
Similarly, postural equilibrium is most stable between the
(123). The dramatic decrease in the strength of the ab-
ages of 16 to 60 years and declines thereafter (91). Thus,
dominal muscles is related to excessive lengthening and
there are few natural changes in gait or postural stability
overstretching of the muscles to accommodate the fetus
until later in life, other than the transient changes that
(57,71).
occur during pregnancy.
A qualitative assessment of standing posture during
pregnancy reveals an elevation of the head, hyperexten-
sion of the cervical spine, and extension of the knee and
PREGNANCY ankle joints (68). Similarly, a quantitative analysis of stand-
ing posture reveals a more posterior head position and
Throughout the 280 days of human gestation, numerous an increase in lumbar lordosis and anterior pelvic tilt
physical and hormonal changes take place in a woman’s (65).
body. Many of these changes, including weight gain, posi- These alterations in joint laxity, weight, and alignment
tion of the body’s center of gravity, increased joint laxity, of the spine are related to decreased postural stability and
and alterations in skeletal alignment, lead to an altered the development of physical discomfort. Indeed, the most
posture and gait. These changes influence postural stabil- common areas of pain during pregnancy are the low-back,
ity and may cause musculoskeletal pain and discomfort. pelvis, hips, knees, and calves (180).

131
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

132 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

FIGURE 8-1. Comparison of pelvic and hip kinematic measures of gait during the first trimester (solid
line) and the third trimester (dashed line) for one pregnant subject. Note the increased maximum anterior
pelvic tilt, increased maximum hip flexion, and increased stance-phase hip adduction. Motion & Gait
Analysis Laboratory, Lucile Packard Children’s Hospital at Stanford.

Gait and Balance during Pregnancy

Gait during pregnancy has been described by the tem- nificant increase in double-support time and a decrease in
porospatial characteristics of velocity, stride length, steps single-support time. In a similar analysis performed at our
taken per minute (cadence), and time spent in single or laboratory, we found a trend toward increased maximum
double support, i.e., with one or both feet on the ground. anterior pelvic tilt, increased maximum hip flexion, and
Gait can be further analyzed in terms of joint kinematics increased stance-phase hip adduction between the first and
and kinetics. However, to date, only a few studies have ad- third trimester of pregnancy (Figure 8-1).
dressed these parameters during pregnancy. Base of support during walking has been found to in-
A cinematographic analysis of two young pregnant crease significantly in pregnant women (22). Although step
subjects during treadmill walking revealed no differ- width increases, pelvic width also increases; thus, the ra-
ences in temporal measures or sagittal plane kinemat- tio of base of support to pelvic width remains constant
ics of the lower limbs throughout the term of preg- (64). An investigation of plantar foot pressures in preg-
nancy (210). The authors speculate that the gradual ad- nant women at 38 weeks gestational age revealed increased
vancement of pregnancy allows the body time to adjust loads on the lateral side of the foot and the hindfoot com-
to the new loads. However, this study failed to address pared to nonpregnant controls (155). Most physiologic and
the impact of pregnancy on kinetics and on movement morphologic changes of pregnancy persist for 4 to 6 weeks
in the frontal and transverse planes. Golomer et al. (76) postpartum (6).
studied ten women between the third and eighth months Static postural sway has been found to increase
of pregnancy. Although the subjects’ velocity remained throughout pregnancy (68). Good static postural control
constant during that time, there was a trend indicat- is characterized by a small sway path, and by a small area
ing an increased cadence and decreased step length. covered by the center of foot pressure movements (217).
The authors proposed that the maintenance of velocity While it might be suspected that this decreased postural
throughout pregnancy reflects the body’s ability to adapt stability is related to an increase in the height of the center
to the change in posture and economize total energy of gravity and the development of the fetal load, much of
expenditure. this sway remains present at six weeks postpartum (68).
A three-dimensional kinematic and kinetic gait anal- The authors speculated that the decrease in postural sta-
ysis performed on 15 women between 35 and 40 weeks bility may be related to laxity of the pelvic ligaments, since
of pregnancy revealed statistically significant increases in laxity does not readily return to pregravida conditions, un-
maximum anterior pelvic tilt, maximum hip flexion, and like other mechanical factors.
stance-phase hip adduction during walking, compared to Approximately 25% of all women employed during
values at 1 year postpartum (64). Increases were found in pregnancy experience a fall, a rate comparable to elderly
the maximum hip extensor moment, maximum hip abduc- persons aged 65 years and older (53). The leading causes of
tor moments, maximum hip power generation in the sagit- falls at work for pregnant women include slippery floors,
tal and frontal planes, the maximum ankle plantar flexor moving at a fast pace, and carrying an object or a child. The
moment, and maximum ankle plantar flexion power ab- aforementioned physiological changes and the increased
sorption. However, these changes were determined to be levels of postural instability during pregnancy may lead
secondary to substantial increases in body mass and an to this high incidence of falls in the (employed) pregnant
anterior shift of the center of gravity. There was also a sig- population.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 133

Exercise during Pregnancy prior to hip fracture become functionally dependent for
walking 6 to 12 months after injury (99,135,139).
The American College of Obstetricians and Gynecologists
Considerable financial costs are associated with the
states, “in the absence of medical or obstetric complica-
functional dependence that occurs after falls in the el-
tions, 30 minutes or more of moderate exercise a day on
derly. Functionally dependent elderly incur $10,000 more
most, if not all, days of the week is recommended for
in healthcare expenditures over 2 years than similarly aged
pregnant women” (6). This recommendation is based on
independent persons (67). In 1995, $26. 1 billion was spent
research which shows that women who exercise more dur-
on healthcare for the 1.4 million persons making the tran-
ing pregnancy have decreased lengths of hospital stays, re-
sition from an independent to a dependent living status
duced incidence of cesarean section, higher Apgar scores
(128). Although less than 12% of the population is older
of newborn infants (81), and a 24% reduced risk of
than the age of 65 years (221), this group sustains 25% to
preeclampsia (199) than women who do not exercise dur-
28% of all fatal injuries and accounts for 33% to 38% of
ing pregnancy. Additionally, regular exercise during preg-
the total US trauma care costs (43,61,148,160). Therefore,
nancy is related to reduced discomforts of swelling, leg
it is of general public concern to understand the impact
cramps, fatigue, and shortness of breath (87,202). Exer-
of the aging process on falling and investigate methods to
cise may also be beneficial in the prevention of gestational
combat the risks of falling in the elderly.
diabetes, especially in morbidly obese women (6).
Walking is one of the most beneficial forms of exercise
during pregnancy, and is the preferred exercise by the Normal Age-Related Changes that
majority of pregnant women (87,149,154). In a study of Affect Gait and Balance
nearly 10,000 pregnant women, 43% of women reported
Changes associated with aging that negatively affect bal-
walking as their leading activity, followed by swimming
ance and gait include declining strength, muscle mass, and
(12%), and aerobics (12%) (234). However, pregnant
bone density; redistributed body mass; impaired respira-
women should be aware of the decreased levels of oxygen
tory capacity; selective atrophy of central nervous system
available for aerobic exercise and modify the intensity of
components controlling balance and gait; and deteriora-
their workouts accordingly (6).
tion in peripheral sensory functions. Additionally, the in-
creased use of medications in the elderly may also have a
In summary, pregnancy results in physical and hor-
negative impact on balance and gait.
monal changes that affect postural balance and gait. Re-
Strength, Muscle Mass and Bone Density, Respira-
search to date indicates that the most prominent changes
tory Capacity. Strength, muscle mass, and bone density
in gait are increased maximum anterior pelvic tilt, in-
decline with age. The age-related change in muscle mass
creased maximum hip flexion, and increased stance-phase
and function is called ‘sarcopenia,’ sarx meaning flesh and
hip adduction. The majority of gait and postural stability
penia meaning loss in Greek (183). Consequences of sar-
measures return to baseline after 6 weeks postpartum.
copenia include decreased strength, metabolic rate, and
maximal oxygen consumption (138). Sarcopenia is mul-
tifactorial and affects mostly type II (fast-twitch) skele-
AGING tal muscle fiber number (129) and cross-sectional area
(8,129). Lack of regular physical activity, change in pro-
Over the last 40 years, the elderly population in the United tein metabolism (i.e., a deficit between protein synthesis
States, age 65 years and older, has increased by 110% (220). and degradation), alterations in the endocrine system, in-
The elderly population at age 85 and older is projected to be cluding decreases in growth hormone and testosterone and
the fastest growing age group, doubling in size from 1995 an increase in cortisol and cytokines, loss of neuromuscu-
to the year 2030, and increasing fivefold by the year 2050. lar function, altered gene expression, and cell apoptosis
This extension in lifespan is unfortunately often associated all underlie the age-related change in muscle mass (138).
with diminished walking ability and decreased postural Cross-sectional and longitudinal studies have found a sig-
stability that results in frequent falls and injury. It has been nificant relation between muscle strength and gait velocity
documented that a large proportion (45%) of falls occur (9,19,59,93).
during walking (157,170,226). Associated with a decline in muscle mass are changes in
Falls account for nearly half of all geriatric trauma cases body composition. Body fat increases while muscle mass
(77,177), representing a leading cause of injury death in decreases (8,89,90,198). This shift in body composition
persons aged 65 to 79 years, second only to motor vehicle with age is often masked by relative stability in overall
traffic accidents (150). For persons over 80 years, falls are body weight and occurs even in physically active older
the leading cause of death (150). Upwards of 90% of hip adults who exercise (138). With advancing age, the mass
fractures are directly related to falling (146) and approxi- of the upper body increases at the expense of the lower
mately 75% of patients who were independent ambulators body, thereby elevating the body’s center of mass (158).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

134 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

FIGURE 8-2. The data presented are based on 61 healthy men, spanning the normal adult age range and
are taken from Sullivan et al. (204). Left panel: Number of seconds an individual was able to stand on one
leg with eyes closed. Note the marked decline in ability with advancing age. Middle panel: Midsagittal
slice of a high-resolution MRI, displaying manual outlining of regions of the cerebellar vermis (44). The
anterior superior vermis declines in volume with age (right panel), and its volume is related to balance
ability (204,205).

Additionally, a change in body posture as a result of skele- decreases in dendritic connections and receptor sites do
tal changes may lead to the stooped posture observed in occur (62,108,163) and likely contribute to compromise of
the elderly, causing an anterior shift in the body’s center neurotransmitter activity, especially of the dopamine sys-
of mass and an increased demand on the posterior mus- tem affected in Parkinson’s disease and other movement
cles, further stressing the balance control system (29,158). disorders (for reviews, see [78,79]). A notable exception is
In addition, muscle strength and age have been shown to the substantia nigra, which loses 5% to 10% of its cells
be independent predictors of loss of balance (102). Thus, per decade (179). White matter also shows disruptive age-
with increasing age and decreasing muscle strength, loss related degradation (110) of myelin and microtubules and
of balance becomes more probable. axon deletion (1,145). These white matter changes may
In a multisite study of over 54,000 healthy, non- contribute to motor slowing and other features of gait
smoking men and women, aged 30 to 70 years, mea- that characterize normal aging, including stooped posture,
surements showed a loss of 0.36% per year for the shortened stride, and loss of motion fluidity.
musculoskeletal system and a loss of 0.84% per year for Magnetic resonance imaging (MRI) studies have
the respiratory system (193). These losses may lead to a documented volume shrinkage of tissue in regional
reduced resistance to fatigue and an inability to sustain cerebral and cerebellar cortex in normal aging
long periods of walking. Reduction in maximal energy ca- (23,24,41,79,98,166,172,175,204,207) that can contribute
pacity (lower VO2 maximum) results in a proportionally to postural instability. Degradation of basal ganglia motor
higher cost of energy during walking (158). Zeleznik (233) and attentional systems (173), cerebellar-pontine circuitry
reported a qualitative emphysematous change in lung his- (208), and frontal lobes (167,172) associated with aging
tology and lung-thorax mechanics in the elderly. Although can also exert untoward forces on gait and balance.
these changes, along with altered lung volumes, affect For example, declines in the gray matter of the anterior
oxygenation and oxygen consumption levels, there is no superior vermis may underlie imbalance in static pos-
evidence that the changes in the respiratory system with ture, measured quantitatively with the Fregley-Graybiel
aging have a negative impact on the daily activities of older Walk-a-Line Test battery (66) (Figure 8-2).
adults. Only when physiologic demands reach the limits of A general relation has been noted between brain white
supply do they become evident. matter abnormalities and falls in the elderly. Early studies
Central Nervous System. Age-related changes of the using computed tomography (CT) reported correlations
central nervous system may negatively affect gait and bal- between the numbers of hypodensities read on CT films
ance. These changes include shrinkage of neuronal soma and impaired gait and disequilibrium (140). Later stud-
and processes of the cerebral cortex (110), especially in ies using MRI (98,189) noted greater frequency and extent
the frontal lobe, and similar degradation of cell struc- of white matter signal hyperintensities, typically occurring
ture in the vermis of the cerebellum (161,162). Although in deep white matter and periventricularly, as predictive of
neuronal loss does not necessarily occur with aging (50), impaired gait or high incidence of falls (16,18,31,35,211).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 135

times as much contrast as younger persons for the detec-

tion of objects in the environment (213). Unfortunately,
while visual acuity is decreasing, there is an increased re-
liance on the visual system to guide locomotion in the el-
derly (158). When elderly and young subjects were asked
to walk a straight path while wearing liquid crystal glasses
that provided a view of the terrain only when they pressed
a handheld switch, the elderly subjects had longer viewing
durations of the terrain than the younger subjects. In ad-
dition, the elderly had a 50% success rate changing their
step length when cued 1 to 2 steps before, versus an 80%
success rate for the younger group. Thus, the increased re-
liance on vision for guidance during walking may serve to
compensate for a decreased ability to alter step lengths in
avoiding an unstable or dangerous surface. In short, the
elderly need to pay more attention to the terrain in order
to allow more time to plan required avoidance and com-
pensation strategies.
Medications. The elderly take more medication than
younger individuals do, and these prescription medica-
tions by themselves or in combination can negatively
FIGURE 8-3. Left two axial images are from high-resolution affect the neuromuscular and vestibular systems (213).
MRI at the level of the lateral ventricles. Right two axial images Benzodiazepines, tricyclic antidepressants, neuroleptics,
are from diffusion tensor imaging, where the brighter the white, and psychotropic medications are the most common
the higher the white matter anisotropy, a measure of fiber tract problem sources related to declines in gait and balance
integrity. Note the expanded ventricles and sulci in the 71 year (51,131,187).
old (bottom left) and less intense anisotropy (bottom right), es-
pecially in the frontal regions of the brain. (Images reproduced
from reference 168).
CHANGES IN GAIT WITH AGE

The volume of white matter on MRI has been shown to be Temporal and Spatial Parameters. Older persons gener-
significantly correlated with the development of gait and ally adopt a more conservative gait pattern than young peo-
balance impairments in the elderly over a 10-year period ple; that is, older persons walk more slowly, take shorter
(17). Neuropathologic correlation with antemortem MRI steps, and are more variable in step timing than younger
indicate that the white matter signals abnormalities that people (147). Walking on an irregular surface increases
are associated with degradation of myelin, thinning of ven- the characteristics of this type of conservative gait in the
tricular membrane, and increased interstitial spaces (16). elderly. The age-related decrease in overall strength means
Recent studies using diffusion tensor imaging, a measure that older people need to use a greater percentage of mus-
of white matter microstructural fiber coherence, provide cle power during gait. This can lead to the use of alter-
further evidence that widespread degradation of brain native muscles for propulsive power (158) that can alter
white matter in normal aging is predictive of declines in temporospatial parameters, such as stride length and ve-
static balance (203) (Figure 8-3). locity. As mentioned previously, there is a direct relation
Peripheral Sensory and Motor Loss. In addition to between muscle strength and gait velocity (9,19,59,93).
age-related changes in the central nervous system, older Decreased gait velocity in the elderly is a func-
people experience a number of peripheral sensory deficits tion of greater time spent in double support and de-
that significantly affect their ability to maintain balance creased step length, rather than decreased cadence
and avoid falls. The peripheral nervous system suffers age- (25,100,156,187,227). The increase in double support
related compromise that may affect gait and balance, in- times may reflect impaired motor control of the body dur-
cluding cell dysmorphology of motor neurons of the spinal ing single support (104). Table 8-1 shows the average re-
cord and decreased density and number of peripheral sults for these temporospatial parameters for a typical
nerve fibers (187). healthy young adult, a healthy elderly adult, and an elderly
Older adults typically have significantly poorer visual adult with pathologic gait tested in our laboratory.
acuity (both high- and low-contrast), contrast sensitivity, A study investigating the relationship between step
and depth perception than younger adults (147). Conse- length and joint kinetics in older subjects (average age
quently, healthy older persons require approximately three 79 years) and younger subjects (average age 26 years)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

136 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

◗ TABLE 8-1 Comparison of Temporospatial Parameters Between Healthy and

Impaired Elderly Adults and Young Adults
Temporal-Spatial Parameters Healthy 30-Year-Old Healthy 75-Year-Old Impaired 80-Year-Old

Velocity (cm/s) 140.9 ± 2.77 88.3 ± 7.47 71.7 ± 6.29

Cadence (steps/min) 110.6 ± 2.13 95.3 ± 4.21 99.7 ± 5.35
Stride Length (cm) 145.1 ± 3.06 117.8 ± 4.75 87.2 ± 6.37
Double Support (% cycle) 22.0 ± 1.70 25.3 ± 2.50 33.4 ± 7.50
Step Width (cm) 12.5 ± 1.36 13.0 ± 0.93 15.8 ± 0.88
Notice the trend toward decreased velocity and stride length, and increased double support time and step width with
increasing age and pathology. Motion & Gait Analysis Laboratory, Lucile Packard Children’s Hospital at Stanford.

found that older persons had a 12% shorter step length tension and compensatory increase in anterior pelvic tilt
during normal walking, when corrected for leg length, than during both comfortable and fast speeds indicate the
younger persons (104). Elble et al. (54) found a 17% to 20% prevalence of hip flexion contractures in the elderly popu-
reduction in gait velocity and stride length in an elderly lation (115).
group compared with young adults. Stride width variabil- Maximum gait velocity and leg power decline with ad-
ity has also been found to increase with age (26,78). vancing age and are positively correlated (121). In a group
A diminished ability to maintain high levels of speed of elderly women, weaker subjects demonstrated lower
with advancing age occur even in elite athletes. Sprint per- concentric ankle and eccentric knee mechanical energy
formance, for example, declines with age in elite runners expenditures and higher eccentric low-back mechanical
(120). This decline becomes most evident around 65 to energy expenditures than stronger subjects (144). As pre-
70 years of age. Master runners of both sexes (men, viously noted, a decrease in overall strength requires a
40 to 88 years, women, 35 to 87 years) were recorded greater percentage of muscle activation during gait, which
with high-speed cameras at the European Veterans Ath- may lead to the use of alternative muscles for propulsive
letic Championships. The velocity during a 100-m sprint power, such as hip and low-back muscles, and also suggests
decreased 5% to 6% per decade in men and 5% to 7% a more proximal mechanism of motor control during gait
per decade in women. Stride length showed clear re- with advancing age.
ductions with increasing age, whereas cadence remained At self-selected walking speeds, elderly adults generate
relatively unchanged. Ground contact time was signifi- less joint torque and power in the lower extremities than
cantly longer, and flight time was shorter in older age young adults. Several studies have reported a decrease in
groups. ankle plantar flexor power during the late stance phase
Kinematics and Kinetics. Increased joint stiffness and of gait in the elderly, when adjusted for speed (45). The
reduced functional range of motion with age may limit the decline in ankle plantar flexor power across the age and
ability of otherwise healthy muscles to generate power at health status is illustrated in Figure 8-4. DeVita and Hor-
the variable speeds required to navigate different terrains tobagyi (45) report increases in hip extensor moment and
(158). Supporting this possibility are the observations of power in healthy elderly subjects, with simultaneous re-
decreased push-off power at the ankle and a more flat- ductions in ankle and knee power. Figure 8-5 illustrates
footed landing in the elderly compared with young adults the increase in hip extensor moments in stance with age,
(227). The joint motions most likely to be limited by mus- especially notable between the healthy 30-year-old and
culotendinous tightness or articular disease include peak the healthy 75-year-old. Judge et al. (100) report an in-
ankle dorsiflexion, peak knee extension, and peak hip ex- crease hip flexor power in older persons to compensate for
tension (105). Figures 8-4 and 8-5 illustrate some of the reduced plantar flexor power, and McGibbon and Krebs
ways in which typical young, healthy elderly, and impaired (143) report increases in knee power absorption in late
elderly persons differ in kinematic and kinetic parameters stance and hip extensor power in early stance. While not
at the ankle and hip. Elderly persons may also exhibit re- all of these studies agree in the specific aspects of change,
ductions in maximum toe-floor clearance, arm swing, and they do suggest that neuromuscular changes that occur
rotations of the hips and knees; however, these differences with aging cause changes in gait characteristics.
may be due to reductions in velocity and stride length,
rather than age alone (54).
Changes in Static Balance with Age
Healthy older people exhibit increased anterior pelvic
tilt, reduced peak hip extension in stance, and decreased Stability during walking is dependent upon normal
peak ankle plantar flexion in swing, independent of speed postural control (104). Postural equilibrium is most
(100,113,115). The consistent reduction in peak hip ex- stable between the ages of 16 to 60 years (91). Function of
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 137

Healthy 30 y.o. Healthy 75 y.o. Impaired 80 y.o.

Ankle Dors/Plantar Ankle Dors/Plantar Ankle Dors/Plantar
40 40 40
Joint Motion (degrees)

20 20 20
Pla - Dor

Pla - Dor

Pla - Dor
0 0 0

−20 −20 −20

−40 −40 −40

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
RHS

RHS

RHS

RHS

RHS

RHS
LTO

LHS

RTO

LTO

LHS

RTO

LTO

LHS

RTO
Ankle Dors/Plantar Ankle Dors/Plantar Ankle Dors/Plantar
2 2 2
Joint Moments (Nm/kg)

1 1 1
Dor - Pla

Dor - Pla

Dor - Pla
0 0 0

−1 −1 −1

−2 −2 −2
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
FIGURE 8-4. Sagittal plane an-
RHS

RHS

RHS

RHS

RHS

RHS
LTO

LHS

RTO

LTO

LHS

RTO

LTO

LHS

RTO
kle motion, moments, and power Ankle Dors/Plantar Ankle Dors/Plantar Ankle Dors/Plantar
for a healthy 30-year old, a healthy 3 3 3
Joint Powers (watts/kg)

75-year old, and an 80-year old 2 2 2

with pathological gait. Note the re-

Abs - Gen

Abs - Gen

Abs - Gen
1 1 1

duced ankle dorsiflexion and di- 0 0 0

minished ankle power generation −1 −1 −1
in late stance in both of the elderly −2 −2 −2
subjects. Motion & Gait Analysis −3 −3 −3
Laboratory, Lucile Packard Chil- 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
RHS

RHS

RHS

RHS

RHS

RHS
LTO

LHS

RTO

LTO

LHS

RTO

LTO

LHS

RTO
dren’s Hospital at Stanford.
Joint Motion (degrees)
Joint Moments (Nm/kg)

FIGURE 8-5. Sagittal plane hip

motion, moments, and power for
a healthy 30-year-old, a healthy
75-year old, and an 80-year old
with pathologic gait. A reduction in
peak hip extension and an increase
in hip power during late stance
can be observed, particularly in
the healthy elderly subject, com-
pared to the young subject. It has
Joint Powers (watts/kg)

been suggested that this increase

in hip power in late stance serves
to compensate for diminished plan-
tar flexor power in late stance. Mo-
tion & Gait Analysis Laboratory,
Lucile Packard Children’s Hospital
at Stanford.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

138 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

the sensory system begins to degrade after the age of (45 to 55 years) or young subjects (15 to 25 years) in stance
65 years, adversely affecting balance and contributing and stance-related tasks, such as tandem walking (70). In-
to falls. A reduction in somatosensation and vestibular creased trunk angular sway and angular velocity indicate
function has been strongly linked to postural instability an increase in balance instability and may be useful in de-
in elderly adults (133,137,165). After age 60 years, sway tecting balance disorders in individuals prone to falling
velocity during quiet standing increases, with the largest (70).
values found in persons older than 76 years of age (91). Ad- Amiridis et al. (7) measured center of pressure varia-
vancing age correlates significantly with longer sway path tions, ankle and hip electromyographic activity (tibialis
lengths and greater postural sway velocity during quiet anterior, medial gastrocnemius, rectus femoris, and semi-
standing (12,84,194) and occurs regardless of visual condi- tendinosus) and hip and ankle kinematics in older adults
tions (eyes open, eyes open with visual feedback, and eyes (mean age 70.1 ± 4.3 years) and younger adults (20.1
closed) (13,37,56,80,119,182). ± 2.4 years). The elderly group had significantly greater
The increased reliance on the visual system to guide lo- and more variable center of pressure excursions than the
comotion in the elderly also applies to postural stability. younger adults in normal quiet standing, tandem Romberg
An investigation of women, aged 20 to 80 years, found an stance (nondominant heel in front of the dominant toe,
age-related increased reliance on vision for postural stabil- arms on the hips), and one-legged stance (standing on the
ity, beginning at the age of 40 years (37). Sway velocities dominant leg with the nonsupporting leg flexed and stabi-
increase with the eyes closed in all age groups, but improve lized on the standing leg). In addition, the elderly showed
with eyes open, even in the elderly (91). Indeed, reduced an increased reliance on the hip muscles for maintaining
visual input increases the odds of a loss of balance fivefold upright balance during tandem Romberg stance and one-
(102). legged stance tasks. No significant difference was noted in
The use of multiple sensorimotor (tactile, visual, and the ankle muscles. This overreliance on the hip muscles
stance) cues can also improve postural sway (94,218). Fig- has also been noted in elderly gait (100).
ure 8-6 demonstrates the anterior-posterior and medial- Physiological control mechanisms considered to con-
lateral postural sway paths for a young healthy woman tribute to static posture involve two basic component pro-
and an elderly healthy woman during quiet standing. No- cesses: a short-term and long-term component. The short-
tice how the sensorimotor cues improve postural sway in term component behaves like an open-loop control system
both the young and elderly subjects. mostly devoid of feedback, whereas the long-term com-
The range of trunk angular sway and angular velocity in ponent behaves like a closed-loop system with feedback
the medial-lateral and anterior-posterior planes is greater based on afferent input (145,146). The open-loop compo-
in elderly subjects (ages 65 to 75 years) than middle-aged nent is affected little by attentional and other cognitive

FIGURE 8-6. Postural sway for

a young healthy woman and an
elderly healthy woman without
(left panels) and with (right pan-
els) the aid of sensorimotor cues.
The plots are stabilograms, taken
from center of pressure calcula-
tions; sampling rate was 50 Hz
and the resulting sway paths are
averages of three 30-second trials.
Note the significant reduction in
sway path length for the younger
(top two plots) and the older (bot-
tom two plots) when cues are
available. Without cues, the sway
path of the older woman is sub-
stantially longer than that of the
younger woman.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 139

processes and is temporally brief (40,49,96). The closed- port phase than elderly nonfallers (109), although these
loop component can be affected by internal and exter- gait changes may be fear-related adaptations, rather than
nal perceptual information. Several experiments demon- risk factors that increase the likelihood of falling. In a kine-
strated that the sway paths of young healthy adults (176), matic analysis of elderly gait, fallers had a reduced range
elderly adults (195), sober alcoholics (205), and individuals of motion at the ankle and a delay in the maximum stance
with fetal alcohol exposure (181) can increase under con- phase ankle dorsiflexion, which the authors speculate may
ditions of sensory challenge that influence the open-loop be predictive of falls (109). An isolated and consistent re-
control component of balance. duction in hip extension during walking occurs with aging,
Sway can increase while engaging in a cognitive task but is exaggerated in elderly fallers compared with elderly
that influences closed–loop control, indicating that static nonfallers (113). Elderly fallers relative to nonfallers have
balance is not solely under automatic control. A study also been shown to exhibit a substantially smaller first step
quantifying cognitive demands of secondary tasks demon- at gait initiation, and the first step length variability of el-
strated a linear relationship between increased sway and derly fallers is more than twice that observed for nonfall-
task difficulty (159). Sway can also be affected by the na- ers (142). An increased stride-to-stride variability in stride
ture of a secondary task. Tasks requiring articulation cause length, velocity, and double support time significantly in-
more interference to stability than do mental tasks (42) creases the odds that an elderly person will experience fu-
and may be related to timing factors intrinsic to speech ture falls, regardless of fear of falling (136).
prosody (157). The difficulty levels of the primary balance Kerrigan et al. (114) found an increase in peak external
and the secondary cognitive tasks may increase indepen- hip flexion moment in stance and reductions in the peak
dently without interacting (112,122) to permit separate hip extension moment, knee flexion moment in preswing,
assessment of each task (232). However, the presence of and knee power absorption in preswing in elderly fallers
pathology associated with aging or neurologic conditions compared with nonfallers, at both comfortable and fast
may introduce an interaction between tasks as a function speeds. Similarly, McGibbon et al. (144) found that the dis-
of difficulty. abled elderly expend less ankle energy in late stance and
more low-back energy in midstance than the healthy el-
derly. When controlling for walking speed, the difference
Slips and Falls
in ankle mechanical energy expenditure disappeared, but
A fall is defined as “an event which results in a person midstance hip mechanical energy expenditure remained
coming to rest inadvertently on the ground or other lower significantly higher in the disabled group. The authors sug-
level, and other than a consequence of the following: sus- gest that increased energy transfer to the low-back and
taining a violent blow, loss of consciousness, sudden onset pelvis may be a strategy used to assist in advancing the
of paralysis, as in a stroke, or an epileptic seizure” (69). swing leg. However, increased trunk energy may also com-
For the elderly, falls may be an outcome of the deficits in promise dynamic stability and increase the risk of falling.
gait and static balance described above. Balance Predictors of Falls. Measures of balance also
Independent risk factors for serious falls are older age; correlate with the number of falls a person sustains. Dise-
white race; decreased bone and mineral density; low body quilibrium is often associated with frequent falls and con-
mass index; cognitive impairment; abnormal neuromus- cerns about falling (111).
cular findings, such as decreased reaction time or balance To determine the physiological basis of age-related
disturbance; poor visual acuity; previous history of falls changes in postural control, center-of-pressure displace-
and fall injuries; and specific chronic illnesses (216). ments and electromyographic data were collected from the
In a study of the biomechanics of slips and falls, kine- tibialis anterior, soleus, vastus lateralis, and biceps femoris
matic and kinetic measurements of young and elderly during quiet standing in elderly fallers, elderly nonfallers,
participants were obtained on slippery and nonslippery and healthy young subjects (125). Elderly fallers demon-
walking surfaces. Older subjects were found to have a sig- strated significantly greater sway in the anterior-posterior
nificantly faster horizontal heel contact velocity, shorter direction and greater muscle activity during quiet stand-
step length, and slower transitional acceleration of the ing than the young subjects. Elderly nonfallers had signif-
whole body center of mass during walking than younger icantly greater muscle activation and coactivation com-
subjects (130). Older participants were also found to slip pared with younger subjects. Short-term postural sway
longer and faster and fall more often than younger partic- was significantly correlated with muscle activity in each
ipants. group. This study suggests that high levels of muscle activ-
Gait Predictors of Falls. A number of kinematic and ity are characteristic of age-related declines in postural sta-
kinetic gait variables distinguish elderly fallers from non- bility and that such activity is correlated with short-term
fallers. In general, the age-related effects on walking that increases in postural sway.
occur in healthy nonfallers are exaggerated in elderly fall- Functional base of support is the anterior-posterior pro-
ers. Fallers have a slower velocity and longer double sup- portion of foot length used in maximal sustained forward
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

140 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

and backward leaning. Base of support declines with age, muscles may also contribute to maintenance of step length
and the extent of the decline may be a predictor of falls in in advanced age (100). A combination of resistance and
older persons (117). balance training has been found to increase gait velocity
Fear of Falling. Falling begets falling. Fear of falling of- by 13% in older persons (106), improve postural control
ten leads to a reduction in activity levels and decondition- (103,134) and reduce the incidence of falls (33,34,186). A
ing, which leads to declines in the quality of life and phys- hip flexor-stretching program for the elderly results in im-
ical well-being, and further increases the danger of falling provements in gait, with an increase in dynamic hip ex-
(86,88,222). Indeed, fear of falling, which many older indi- tension during both comfortable and fast walking speeds
viduals express, is a liability because it can alter the mag- (116). The improved hip extension also leads to significant
nitude of postural adjustment to maintain erect posture, improvements in peak ankle plantar flexion and a tendency
resulting in overcorrection (4). Sway velocity on static and toward improved ankle power generation.
dynamic posturography is greater in elderly persons who Home-based exercise programs have recently gained
report a fear of falling compared with those who do not considerable attention for their ability to improve gait and
(13,15). functional performance. In one study (152), 70 community
dwelling, elderly men and women completed a 6-month
clinical trial. Half of the participants were given an exercise
EXERCISE TO IMPROVE GAIT program focusing on strength and balance training and
AND AVOID FALLS encouraged to increase overall physical activity; the other
half of the participants were given 6 months of nutrition
Although the physiological effects of aging are widely education. The exercise group was asked to perform body
accepted as inevitable, physical activity and exercise weight exercises, lower body strength training with ankle
throughout aging can diminish or even negate age’s un- weights, upper body strength training with dumbbells, and
toward effects (36). There are many benefits to exercise, balance training exercises 3 times per week. Exercises were
including improved strength, flexibility, reaction time, gait, demonstrated by an exercise physiologist and explained in
and postural control (192). Despite age-related decline a detailed booklet outlining the program. At the end of the
in musculoskeletal and respiratory systems, the underly- six months, the exercise group saw a significant improve-
ing plasticity of the muscles, autonomic nervous system, ment in functional performance and balance and coordi-
bones, and joints remains intact even in very elderly per- nation, compared to the group who only received nutrition
sons and is amenable to the effects of conditioning ex- education. The authors concluded that a home-based, mul-
ercises (21,58–60). In one study of nine frail, institution- tidimensional exercise program in community-dwelling el-
alized volunteers, average age 90 years, who underwent ders is feasible and can be effective in improving functional
8 weeks of high-intensity resistance training, strength in- performance, despite limited supervision.
creased 174% and mean tandem gait speed improved 48% Additionally, Tai Chi has been shown to improve bal-
(59). ance in the elderly (101,219,230). Tai Chi was originally
Any exercise program designed for older individuals developed as a martial art form but has been used by
should incorporate the following four goals: increase con- elderly Chinese people as an exercise form for the past
ditioning, especially endurance; improve muscle strength, 300 years (231). Tai Chi focuses on slow sequential move-
particularly of the lower extremities; minimize risk of in- ments, providing a smooth, continuous and low intensity
jury; and promote enjoyment without causing excessive fa- activity. After only 4 weeks of intensive Tai Chi training,
tigue (124). The two main components of a good exercise community-dwelling elderly subjects (mean age 69 years)
program are dynamic aerobic exercise and strength train- significantly improved their ability to use somatosensory,
ing. Dynamic aerobic exercise includes walking, swim- visual, and vestibular information to control body sway
ming, cycling, and jogging. In fact, brisk walking has been during quiet standing, and an improved ability to volun-
deemed one of the most ideal forms of exercise readily tarily weight shift to various spatial positions within their
available to the elderly population (55) and has been shown base of support, compared to the control group (219).
to improve velocity, standing balance, speed of muscular Furthermore, the improved balance performance at week
contraction, and lead to an increase in the maximal rate of 4 of the elderly subjects was comparable to that of ex-
oxygen consumption and VO2 maximum (30). perienced Tai Chi practitioners. One investigator points
Gait velocity is directly related to muscle strength out that with Tai Chi training there is a counterintu-
(9,19,59,93). Progressive resistive training, including hip itive reduction in anticipatory postural adjustments of
and knee flexion/extension, ankle dorsi/plantar flexion, standing posture while stability of standing is improved
and hip ab/adduction exercises have been shown to (63). These findings suggest that practicing Tai Chi may
significantly increase muscle strength in older persons lead to a greater use of the elasticity of the peripheral
(107,171,186,196). Strength training of the plantar flexor structures involving muscles, ligaments, and tendons and
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 141

a decrease in the participation of the central neutral struc- edgment of alcoholism, its potential effects on postural sta-
tures in postural equilibrium, thereby improving postural bility, its mechanisms of action, and strategies for mitigat-
stability. ing imbalance should be pursued. In a recent survey of
Elderly adults who participate in sports and a high level alcoholic men and women who volunteered for research
of total physical activity, walking, or household activity ex- studies, proportionately more alcoholics (60%) than con-
perience a smaller decline in mobility over time than older trols (30%), regardless of sex, reported a history of at least
adults who are inactive. In a 3-year prospective study of one fall in the past year (184). Contributing special lia-
over 2,000 men and women aged 55 to 85 years, two mo- bility to falling is alcoholism and alcoholism’s interaction
bility tests (timed 6 meter walk and repeated chair stands) with aging (47,72,204,223). The chronic effects of alcohol
were performed at baseline and again 3 years later. Con- intoxication on balance and gait can linger in alcoholics
tinued physical activity was associated with the smaller who have withdrawn from alcoholism and remain sober
decline in mobility (225). Older persons who exercise reg- (224).
ularly perform better on tests of strength, flexibility, re- Gait ataxia in alcoholics has traditionally been at-
action time, walking, and balance maneuvers than older tributed to damage to the anterior superior vermis of
persons who do not exercise on a regular basis (20,132). the cerebellum, as determined with postmortem study
Similarly, older adults who participate in 20 to 30 minutes (11,82,224). More recently, in vivo neuroimaging stud-
of moderate-intensity exercise on most days of the week ies confirm these speculations and have reported signif-
(1,000 kcal per week) have better physical function, e.g., icant correlations between alcoholism-related ataxia and
endurance, lower extremity strength, gait speed, and bal- low regional glucose metabolism quantified with positron
ance, than older persons who are active throughout the emission tomography (73) and anterior superior vermian
day but do not exercise, or who are inactive (27). shrinkage quantified with MRI (204).
Even if exercise has not been a lifelong habit, research Balance platform studies of recovering alcoholics pro-
shows that adults who participate in sports and physical vide evidence for enduring instability measurable with
activities in old age have better postural ability than adults static (46,126,223) or dynamic posturography, that is,
who only exercised at an early age. In a study of 65 adults stance during platform perturbation (126). Although de-
over the age of 60 years, those subjects who only started gree of impaired postural control has been related to the
participating in sports and physical activities after age 60 amount of alcohol drunk in the 6 months prior to exam-
had postural performances close to that of the subjects ination (228,229), persistent deficits in balance diminish
who had always exercised, and had significantly better bal- with protracted sobriety in recovering alcoholics but do
ance than those subjects who only exercised at an early age not necessarily fully resolve (185,206). Although periph-
or never exercised (164). eral neuropathy, which can be a concomitant of chronic
alcoholism, may exacerbate imbalance, neuropathy does
not necessarily account for imbalance (191). Another likely
ALCOHOL AND ALCOHOLISM factor mitigating against full recovery is the presence of
cerebellar pathology, noted in chronic alcoholics and also
Falling is one of the leading causes of morbidity and mor- in children with prenatal alcohol exposure (200). Lesions
tality in otherwise healthy individuals and its liability is in this vermian region characteristically result in postu-
exacerbated with alcohol (214,215). In an investigation of ral tremor (∼3 Hz) (74,151), which is prominent in the
ground-level falls, ethanol was present in nearly 50% of all anterior-posterior direction (141,223) and detectable with
cases tested (83). Acute alcohol ingestion reduces the func- spectral analysis of sway velocity (14).
tion of the vestibular system and leads to balance and gait A study from our laboratory (205) assessed sway during
disturbances. The acute effects of alcohol use include posi- quiet standing in alcoholic men who had been sober for
tional nystagmus and gaze nystagmus (127), reductions in several months and examined whether postural instabil-
the gain of the vestibulo-ocular reflex (169,190,212), and ity, measured in terms of sway path length and direction,
increases in body sway during static and dynamic postur- could be ameliorated by sensorimotor visual, tactual, or
ography (75,127,153). A comparison of the sway pattern stance cues, which are known to exert stabilizing forces
of subjects after acute alcohol ingestion closely resembles in normal, healthy adults (95).The alcoholics were signifi-
that of patients with chronic lesions of the cerebellar an- cantly less stable than the controls when maintaining erect
terior lobe, i.e., the spinocerebellum (48,153). posture in the absence of visual, tactile, or stance cues.
The incidence of alcohol abuse and dependence is high, Although longer sobriety was predictive of shorter sway
estimated at upwards of 15% of the US population and 30% paths, residual imbalance without stabilizing cues was still
to 35% of patients treated at university medical centers, measurable months after cessation of drinking. In the pres-
making it a significant public health concern, although it is ence of cues, however, the sway paths of the alcoholics were
often overlooked in clinical interview (10). Thus, acknowl- indistinguishable from those of the controls, indicating
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

142 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

FIGURE 8-7. Left panels: Force plate stabilograms collected with and without sensorimotor cues in a
61-year-old control man (top) and a 56-year-old alcoholic man (bottom). The MRI images on the right are
midsagittal MRI (proton density weighted on the left and T2-weighted on the right), showing a volume
deficit in the anterior superior vermis of the alcoholic relative to the control (205).

that the alcoholics had functionally adequate sensorimo- on Aging (AG17919) and the National Institute on Alcohol
tor integration skills despite the likely cerebellar basis for Abuse and Alcoholism (AA10723) granted to E.V.S.
imbalance (Figure 8-7). The frequency characteristics of
sway velocity in the 2 to 7 Hz frequencies was also greater
in alcoholics than controls, but postural tremor at all of REFERENCES
these frequencies was ameliorated with sensorimotor cues.
1. Aboitiz F, Rodriguez E, Olivares R, Zaidel E. Age-related changes in
These results provide support for a cerebellar vermian ba- fibre composition of the human corpus callosum: sex differences.
sis of alcoholism-related postural instability and indicate Neuroreport 1996;7:1761–1764.
that alcoholics can successfully engage in sensorimotor in- 2. Abrams B, Parker JD. Maternal weight gain in women with good
pregnancy outcome. Obstet Gynecol 1990;76:1–7.
tegration to override CNS-related deficits in static balance 3. Abramson D, Roberts SB, Wilson P. Relaxation of the pelvic joints
(205). in pregnancy. Surg Gynecol Obstet 1934;58:595–613.
Stabilization that availed with sensorimotor cues has 4. Adkin AL, Frank JS, Carpenter MG, Peysar GW. Fear of falling
modifies anticipatory postural control. Exp Brain Res 2002;143:
implications for rehabilitation strategies for individuals 160–170.
recovering from chronic alcoholism and probably also for 5. Ahmad S, Rohrbaugh JW, Anokhin AP, Sirevaag EJ, Goebel JA.
nonalcoholic patients with similar postural instability. The Effects of lifetime ethanol consumption on postural control: a com-
puterized dynamic posturography study. J Vestibul Res 2002;12:
progressive improvement in balance and decrease in sway 53–64.
with increasing numbers of cues suggests that that when 6. American College of Obstetricians and Gynecologists. Exercise dur-
external cues are present, impaired individuals can adapt ing pregnancy and the postpartum period. Clin Obstet Gynecol
2003;46:496–499.
mechanisms to compensate for underlying cerebellar in- 7. Amiridis IG, Hatzitaki V, Arabatzi F. Age-induced modifications of
sult and such cues may even reduce fear of falling. static postural control in humans. Neurosci Lett 2003;350:137–140.
8. Aniansson A, Hedberg M, Henning GB, Grimby G. Muscle mor-
phology, enzymatic activity, and muscle strength in elderly men: a
follow-up study. Muscle Nerve 1986;9:585–591.
ACKNOWLEDGMENTS 9. Aniansson A, Rundgren A, Sperling L. Evaluation of functional ca-
pacity in activities of daily living in 70-year-old men and women.
Scand J Rehab Med 1980;12:145–154.
The authors thank Margaret J. Rosenbloom, M.A. for help 10. Arciniegas DB, Beresford TP. Neuropsychiatry: An Introductory Ap-
in manuscript editing and preparation, Adolf Pfefferbaum, proach. Cambridge: Cambridge University Press; 2001.
M.D. for the balance platform and MRI figures, and Jessica 11. Baker K, Harding A, Halliday G, Kril J, Harper C. Neuronal
loss in functional zones of the cerebellum of chronic alcoholics
Rose, Ph.D. for the gait analysis studies. Partial support with and without Wernicke’s encephalopathy. Neuroscience 1999;91:
for this project was obtained from the National Institute 429–438.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 143

12. Baloh RW, Corona S, Jacobson KM, Enrietto JA, Bell T. A prospec- genetic influences to measures of lower-extremity function in older
tive study of posturography in normal older people. J Am Geriatr male twins. J Gerontol A Biol Sci Med Sci 2000;55:B49–53.
Soc 1998;46:438–443. 36. Chakravarthy MV, Joyner MJ, Booth FW. An obligation for primary
13. Baloh RW, Fife TD, Zwerling L, et al. Comparison of static and dy- care physicians to prescribe physical activity to sedentary patients
namic posturography in young and older normal people. J Am Geri- to reduce the risk of chronic health conditions. Mayo Clin Proc
atr Soc 1994;42:405–412. 2002;77:165–173.
14. Baloh RW, Jacobson KM, Beykirch K, Honrubia V. Static and dy- 37. Choy NL, Brauer S, Nitz J. Changes in postural stability in women
namic posturography in patients with vestibular and cerebellar le- aged 20 to 80 years. J Gerontol A Biol Sci Med Sci 2003;58:525–
sions. Arch Neurol 1998;55:649–654. 530.
15. Baloh RW, Spain S, Socotch TM, Jacobson KM, Bell T. Postur- 38. Collins JJ, De Luca CJ, Burrows A, Lipsitz LA. Age-related changes
ography and balance problems in older people. J Am Geriatr Soc in open-loop and closed-loop postural control mechanisms. Exp
1995;43:638–644. Brain Res 1995;104:480–492.
16. Baloh RW, Vinters HV. White matter lesions and disequilibrium 39. Collins JJ, De Luca CJ. Open-loop and closed-loop control of pos-
in older people: 2. clinicopathologic correlation. Arch Neurol ture: A random-walk analysis of center-of-pressure trajectories. Exp
1995;52:975–981. Brain Res 1993;95:308–318.
17. Baloh RW, Ying SH, Jacobson KM. A longitudinal study of gait 40. Collins JJ, De Luca CJ. The effects of visual input on open-loop and
and balance dysfunction in normal older people. Arch Neurol closed-loop postural control mechanisms. Exp Brain Res 1995;103:
2003;60:835–839. 151–163.
18. Baloh RW, Yue Q, Socotch TM, Jacobson KM. White matter le- 41. Courchesne E, Chisum HJ, Townsend J, et al. Normal brain devel-
sions and disequilibrium in older people: 1. case-control compar- opment and aging: quantitative analysis at in vivo MR imaging in
ison. Arch Neurol 1995;52:970–974. healthy volunteers. Radiology 2000;216:672–682.
19. Bassey EJ, Bendall MJ, Pearson M. Muscle strength in the tri- 42. Dault MC, Yardley L, Frank JS. Does articulation contribute to
ceps surae and objectively measured customary walking activity modifications of postural control during dual-task paradigms? Cogn
in men and women over 65 years of age. Clin Sci (Lond)1988;74: Brain Res 2003;16:434–440.
85–89. 43. DeMaria EJ, Merriam MA, Casanova LA, Gann DS, Kenney PR. Do
20. Baylor AM, Spirduso WW. Systematic aerobic exercise and compo- DRG payments adequately reimburse the costs of trauma care in
nents of reaction time in older women. J Gerontol 1988;43:P121– geriatric patients? J Trauma 1988;28:1244–1249.
126. 44. Deshmukh A, Desmond JE, Sullivan EV, et al. Quantification of cere-
21. Beere PA, Russell SD, Morey MC, Kitzman DW, Higginbotham bellar structures with MRI. Psychiatr Res Neuroimaging 1997;75:
MB. Aerobic exercise training can reverse age-related peripheral 159–172.
circulatory changes in healthy older men. Circulation 1999;100: 45. DeVita P, Hortobagyi T. Age causes a redistribution of joint torques
1085–1094. and powers during gait. J Appl Physiol 2000;88:1804–1811.
22. Bird AR, Menz HB, Hyde CC. The effect of pregnancy on footprint 46. Diener H-C, Dichgans J, Bacher M, Guschlbauer B. Characteris-
parameters. A prospective investigation. J Am Podiatr Med Assoc tic alteration of long-loop “reflexes” in patients with Friedreich’s
1999;89:405–409. disease and late atrophy of the cerebellar anterior lobe. J Neurol,
23. Blatter DD, Bigler ED, Gale SD, et al. Quantitative volumetric anal- Neurosurg, Psychiatry 1984;47:679-685.
ysis of brain MRI: Normative database spanning five decades of life. 47. Diener H-C, Dichgans J, Bacher M, Guschlbauer. Improvement in
Am J Neuroradiol 1995;16:241–245. ataxia in alcoholic cerebellar atrophy through alcohol abstinence.
24. Blatter DD, Bigler ED, Gale SD, et al. Quantitative volumetric anal- J Neurol 1984;231:258–262.
ysis of brain MR: normative database spanning 5 decades of life. 48. Diener HC, Dichgans J, Bacher M, Hulser J, Liebich H. Mechanisms
Am J Neuroradiol 1995;16:241–251. of postural ataxia after intake of alcohol. Z Rechtsmed 1983;90:159–
25. Bohannon RW. Comfortable and maximum walking speed of adults 165.
aged 20–79 years: reference values and determinants. Age Ageing 49. Diener HC, Dichgans J, Guschlbauer B, Bacher M, Langenbach
1997;26:15–19. P. Disturbances of motor preparation in basal ganglia and cere-
26. Brach JS, Berthold R, Craik R, VanSwearingen JM, Newman AB. bellar disorders. In: Allun JHH, Hulliger M, eds. Progress in Brain
Gait variability in community-dwelling older adults. J Am Geriatr Research, Vol 80, Afferent Control of Posture and Locomotion. Ams-
Soc 2001;49:1646–1650. terdam: Elsevier Scientific Publishers B.V.; 1989:481–488.
27. Brach JS, Simonsick EM, Kritchevsky S, Yaffe K, Newman AB. The 50. Dlugos CA, Pentney RJ. Morphometric evidence that the total num-
association between physical function and lifestyle activity and ex- ber of synapses on Purkinje neurons of old F344 rats is reduced after
ercise in the health, aging and body composition study. J Am Geriatr long-term ethanol treatment and restored to control levels after re-
Soc 2004;52:502–509. covery. Alcohol Alcohol 1997;32:161–172.
28. Bril B, Breniere Y. Postural requirements and progression velocity 51. Dominguez RO, Bronstein AM. Assessment of unexplained falls and
in young walkers. J Motor Behav 1992;24:105–116. gait unsteadiness: the impact of age. Otolaryngol Clin North Am
29. Brocklehurst JC, Robertson D, James-Groom P. Skeletal deformities 2000;33:637–657.
in the elderly and their effect on postural sway. J Am Geriatr Soc 52. Dumas GA, Reid JG. Laxity of knee cruciate ligaments during preg-
1982;30:534–538. nancy. J Orthop Sports Phys Ther 1997;26:2–6.
30. Brown M, Holloszy JO. Effects of walking, jogging and cycling on 53. Dunning K, LeMasters G, Levin L, Bhattacharya A, Alterman T,
strength, flexibility, speed and balance in 60- to 72-year olds. Aging Lordo K. Falls in workers during pregnancy: risk factors, job
1993;5:427–434. hazards, and high risk occupations. Am J Ind Med 2003;44:664–
31. Cahn DA, Malloy PF, Salloway S, et al. Subcortical hyperintensi- 672.
ties on MRI and activities of daily living in geriatric depression. 54. Elble RJ, Thomas SS, Higgins C, Colliver J. Stride-dependent
J Neuropsychiatry Clin Neurosci 1996;8:404–411. changes in gait of older people. J Neurol 1991;238:1–5.
32. Calguneri M, Bird HA, Wright V. Changes in joint laxity occurring 55. Elward K, Larson EB. Benefits of exercise for older adults. A review
during pregnancy. Ann Rheum Dis 1982;41:126–128. of existing evidence and current recommendations for the general
33. Campbell AJ, Robertson MC, Gardner MM, Norton RN, Buch- population. Clin Geriatr Med 1992;8:35–50.
ner DM. Falls prevention over 2 years: a randomized controlled 56. Era P, Heikkinen E, Gause-Nilsson I, Schroll M. Postural balance in
trial in women 80 years and older. Age Ageing 1999;28:513– elderly people: changes over a five-year follow-up and its predictive
518. value for survival. Aging Clin Exp Res 2002;14:37–46.
34. Campbell AJ, Robertson MC, Gardner MM, Norton RN, Tilyard MW, 57. Fast A, Weiss L, Ducommun EJ, Medina E, Butler JG. Low-back pain
Buchner DM. Randomised controlled trial of a general practice pro- in pregnancy. Abdominal muscles, sit-up performance, and back
gramme of home based exercise to prevent falls in elderly women. pain. Spine 1990;15:28–30.
Br Med J 1997;315:1065–1069. 58. Fiatarone MA, Evans WJ. The etiology and reversibility of muscle
35. Carmelli D, Kelly-Hayes M, Wolf PA, et al. The contribution of dysfunction in the aged. J Gerontol 1993;48:77–83.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

144 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

59. Fiatarone MA, Marks EC, Ryan ND, Meredith CN, Lipsitz LA, Evans 85. Heckman JD, Sassard R. Musculoskeletal considerations in preg-
WJ. High-intensity strength training in nonagenarians. Effects on nancy. J Bone Joint Surg Am 1994;76:1720–1730.
skeletal muscle. JAMA 1990;263:3029–3034. 86. Hindmarsh JJ, Estes EH, Jr. Falls in older persons. Causes and in-
60. Fiatarone MA, O’Neill EF, Ryan ND, et al. Exercise training and nu- terventions. Arch Intern Med 1989;149:2217–2222.
tritional supplementation for physical frailty in very elderly people. 87. Horns PN, Ratcliffe LP, Leggett JC, Swanson MS. Pregnancy out-
N Engl J Med 1994;330:1769–1775. comes among active and sedentary primiparous women. J Obstet
61. Finelli FC, Jonsson J, Champion HR, Morelli S, Fouty WJ. A case Gynecol Neonatal Nurs 1996;25:49–54.
control study for major trauma in geriatric patients. J Trauma 1989; 88. Howland J, Lachman ME, Peterson EW, Cote J, Kasten L, Jette
29:541–548. A. Covariates of fear of falling and associated activity curtailment.
62. Flood DG. Critical issues in the analysis of dendritic extent in aging Gerontologist 1998;38:549–555.
humans, primates, and rodents. Neurobiol Aging 1993;14:649–654. 89. Hughes VA, Frontera WR, Roubenoff R, Evans WJ, Singh MA. Lon-
63. Forrest WR. Anticipatory postural adjustment and T’ai Chi Ch’uan. gitudinal changes in body composition in older men and women:
Biomed Sci Instrum 1997;33:65–70. role of body weight change and physical activity. Am J Clin Nutr
64. Foti T, Davids JR, Bagley A. A biomechanical analysis of gait during 2002;76:473–481.
pregnancy. J Bone Joint Surg Am 2000;82:625–632. 90. Hughes VA, Frontera WR, Wood M, et al. Longitudinal muscle
65. Franklin ME, Conner-Kerr T. An analysis of posture and back pain strength changes in older adults: influence of muscle mass, physical
in the first and third trimesters of pregnancy. J Orthop Sports Phys activity, and health. J Gerontol A Biol Sci Med Sci 2001;56:B209–217.
Ther 1998;28:133–138. 91. Hytonen M, Pyykko I, Aalto H, Starck J. Postural control and age.
66. Fregly AR, Graybiel A, Smith MS. Walk on floor eyes closed Acta Otolaryngol 1993;113:119–122.
(WOFEC): A new addition to an ataxia test battery. Aerosp Med 92. Ireland ML, Ott SM. The effects of pregnancy on the musculoskele-
1972;43:395–399. tal system. Clin Orthop 2000; 169–179.
67. Fried TR, Bradley EH, Williams CS, Tinetti ME. Functional disabil- 93. Ito H, Nagasaki H, Maruyama H, Hashizume K, Nakamura R.
ity and health care expenditures for older persons. Arch Intern Med [Age related changes in the walking cycle during fastest walking in
2001;161:2602–2607. healthy male subjects]. Nippon Ronen Igakkai Zasshi 1989;26:347–
68. Fries EC, Hellebrandt FA. The influence of pregnancy on the loca- 352.
tion of the center of gravity, postural stability, and body alignment. 94. Jeka JJ, Lackner JR. Fingertip contact influences human postural
Am J Obstetr Gynecol 1943;46:374–380. control. Exp Brain Res 1994;100:495–502.
69. Gibson MJ. The prevention of falls in later life. Dan Med Bull 1987; 95. Jeka JJ, Lackner JR. Fingertip contact influences human postural
34(suppl 4):1–24. control. Exp Brain Res 1994;100:495–502.
70. Gill J, Allum JH, Carpenter MG, et al. Trunk sway measures of pos- 96. Jeka JJ, Schoner G, Dikjstra T, Ribeiro P, Lackner JR. Coupling of
tural stability during clinical balance tests: effects of age. J Gerontol fingertip somatosensory information to head and body sway. Exp
A Biol Sci Med Sci 2001;56:M438–447. Brain Res 1997;113:475–483.
71. Gilleard WL, Brown JM. Structure and function of the abdominal 97. Jensen RK, Doucet S, Treitz T. Changes in segment mass and mass
muscles in primigravid subjects during pregnancy and the immedi- distribution during pregnancy. J Biomech 1996;29:251–256.
ate postbirth period. Phys Ther 1996;76:750–762. 98. Jernigan TL, Archibald SL, Berhow MT, Sowell ER, Foster DS,
72. Gilman S, Adams K, Koeppe RA, et al. Cerebellar and frontal Hesselink JR. Cerebral structure on MRI .1. Localization of age-
hypometabolism in alcoholic cerebellar degeneration studied related changes. Biol Psychiatry 1991;29:55–67.
with positron emission tomography. Ann Neurol 1990;28:775– 99. Jette AM, Harris BA, Cleary PD, Campion EW. Functional recovery
785. after hip fracture. Arch Phys Med Rehabil 1987;68:735–740.
73. Gilman S, Adams K, Koeppe RA, et al. Cerebellar and frontal hy- 100. Judge JO, Davis RB, 3rd, Ounpuu S. Step length reductions in ad-
pometabolism in alcoholic cerebellar degeneration studied with vanced age: the role of ankle and hip kinetics. J Gerontol A Biol Sci
Positron Emission Tomography. Ann Neurol 1990;28:775–785. Med Sci 1996;51:M303–312.
74. Gilman S, Bloedel JR, Lechtenberg R. Disorders of the Cerebellum. 101. Judge JO, Kenny AM, Kraemer WJ. Exercise in older adults. Conn
Philadelphia: F.A. Davis Co; 1981. Med 2003;67:461–464.
75. Goebel JA, Dunham DN, Rohrbaugh JW, Fischel D, Stewart PA. 102. Judge JO, King MB, Whipple R, Clive J, Wolfson LI. Dynamic bal-
Dose-related effects of alcohol on dynamic posturography and ocu- ance in older persons: effects of reduced visual and proprioceptive
lomotor measures. Acta Otolaryngol Suppl 1995;520 Pt 1:212–215. input. J Gerontol A Biol Sci Med Sci 1995;50:M263–270.
76. Golomer E, Ducher D, Arfi GS, Sud R. [Simple locomotion and 103. Judge JO, Lindsey C, Underwood M, Winsemius D. Balance im-
during load carrying in pregnant women]. J Gynecol Obstet Biol provements in older women: effects of exercise training. Phys Ther
Reprod (Paris) 1991;20:406–412. 1993;73:254–262; discussion 263–255.
77. Gomberg BF, Gruen GS, Smith WR, Spott M. Outcomes in acute 104. Judge JO, Ounpuu S, Davis RB, 3rd. Effects of age on the
orthopaedic trauma: a review of 130,506 patients by age. Injury biomechanics and physiology of gait. Clinical Geriatric Medicine
1999;30:431–437. 1996;12:659–678.
78. Grabiner PC, Biswas ST, Grabiner MD. Age-related changes in spa- 105. Judge JO, Ounpuu S, Davis RB, 3rd. Effects of age on the biome-
tial and temporal gait variables. Arch Phys Med Rehabil 2001;82:31– chanics and physiology of gait. Clin Geriatr Med 1996;12:659–
35. 678.
79. Gur RC, Gunning-Dixon FM, Turetsky BI, Bilker WB, Gur RE. Brain 106. Judge JO, Underwood M, Gennosa T. Exercise to improve gait ve-
region and sex differences in age association with brain volume: locity in older persons. Arch Phys Med Rehabil 1993;74:400–406.
a quantitative MRI study of healthy young adults. Am J Geriatr 107. Judge JO, Whipple RH, Wolfson LI. Effects of resistive and balance
Psychiatr 2002;10:72–780. exercises on isokinetic strength in older persons. J Am Geriatr Soc
80. Hageman PA, Leibowitz JM, Blanke D. Age and gender effects on 1994;42:937–946.
postural control measures. Arch Phys Med Rehabil 1995;76:961–965. 108. Kaasinen V, Aalto S, K NA, Hietala J, Sonninen P, Rinne JO. Extra-
81. Hall DC, Kaufmann DA. Effects of aerobic and strength condition- striatal dopamine D(2) receptors in Parkinson’s disease: a longitu-
ing on pregnancy outcomes. Am J Obstet Gynecol 1987;157:1199– dinal study. J Neural Transm 2003;110:591–601.
1203. 109. Kemoun G, Thoumie P, Boisson D, Guieu JD. Ankle dorsiflexion
82. Harper CG, Kril JJ. Neuropathological changes in alcoholics. In: delay can predict falls in the elderly. J Rehabil Med 2002;34:278–
Hunt WA, Nixon SJ eds. Alcohol Induced Brain Damage: NIAAA 283.
Research Monograph No 22. Rockville, MD: National Institute of 110. Kemper TL. Neuroanatomical and neuropathological changes dur-
Health; 1993:39–69. ing aging and dementia. In: Albert ML, Knoefel JE, eds. Clin Neurol
83. Hartshorne NJ, Harruff RC, Alvord EC, Jr. Fatal head injuries in Aging New York: Oxford University Press; 1994:3–67.
ground-level falls. Am J Forensic Med Pathol 1997;18:258–264. 111. Kerber KA, Enrietto JA, Jacobson KM, Baloh RW. Disequilib-
84. Hasselkus BR, Shambes GM. Aging and postural sway in women. rium in older people: a prospective study. Neurology 1998;51:574–
J Gerontol 1975;30:661–667. 580.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 145

112. Kerr B. Processing demands during mental operations. Mem Cognit 137. Manchester D, Woollacott M, Zederbauer-Hylton N, Marin O. Vi-
1973;1:401–412. sual, vestibular and somatosensory contributions to balance control
113. Kerrigan DC, Lee LW, Collins JJ, Riley PO, Lipsitz LA. Reduced hip in the older adult. J Gerontol 1989;44:M118–127.
extension during walking: healthy elderly and fallers versus young 138. Marcell TJ. Sarcopenia: causes, consequences, and preventions.
adults. Arch Phys Med Rehabil 2001;82:26–30. J Gerontol A Biol Sci Med Sci 2003;58:M911–916.
114. Kerrigan DC, Lee LW, Nieto TJ, Markman JD, Collins JJ, Riley PO. 139. Marottoli RA, Berkman LF, Cooney LM, Jr. Decline in physical func-
Kinetic alterations independent of walking speed in elderly fallers. tion following hip fracture. J Am Geriatr Soc 1992;40:861–866.
Arch Phys Med Rehabil 2000;81:730–735. 140. Masdeu JC, Wolfson L, Lantos G, et al. Brain white matter changes
115. Kerrigan DC, Todd MK, Della Croce U, Lipsitz LA, Collins JJ. Biome- in the elderly prone to falling. Arch Neurol 1989;46:1292–1296.
chanical gait alterations independent of speed in the healthy elderly: 141. Mauritz KH, Dichgans J, Hufschmidt A. Quantitative analysis of
evidence for specific limiting impairments. Arch Phys Med Rehabil stance in late cortical cerebellar atrophy of the anterior lobe and
1998;79:317–322. other forms of cerebellar ataxia. Brain 1979;102:461–482.
116. Kerrigan DC, Xenopoulos-Oddsson A, Sullivan MJ, Lelas JJ, Riley 142. Mbourou GA, Lajoie Y, Teasdale N. Step length variability at gait
PO. Effect of a hip flexor-stretching program on gait in the elderly. initiation in elderly fallers and non-fallers, and young adults. Geron-
Arch Phys Med Rehabil 2003;84:1–6. tology 2003;49:21–26.
117. King MB, Judge JO, Wolfson L. Functional base of support de- 143. McGibbon CA, Krebs DE. Discriminating age and disability ef-
creases with age. J Gerontol 1994;49:M258–263. fects in locomotion: neuromuscular adaptations in musculoskeletal
118. Kirtley C, Whittle MW, Jefferson RJ. Influence of walking speed on pathology. J Appl Physiol 2004;96:149–160.
gait parameters. J Biomed Eng 1985;4:282–288. 144. McGibbon CA, Puniello MS, Krebs DE. Mechanical energy transfer
119. Kollegger H, Baumgartner C, Wober C, Oder W, Deecke L. Spon- during gait in relation to strength impairment and pathology in
taneous body sway as a function of sex, age, and vision: posturo- elderly women. Clin Biomech (Bristol, Avon) 2001;16:324–333.
graphic study in 30 healthy adults. Eur Neurol 1992;32:253–259. 145. Meier-Ruge W, Ulrich J, Bruhlmann M, Meier E. Age-related white
120. Korhonen MT, Mero A, Suominen H. Age-related differences in 100- matter atrophy in the human brain. Ann N Y Acad Sci 1992;673:260–
m sprint performance in male and female master runners. Med Sci 269.
Sports Exerc 2003;35:1419–1428. 146. Melton LJ. Epidemiology of Fractures. In: Riggs BL, Melton LK,
121. Kozakai R, Tsuzuku S, Yabe K, Ando F, Niino N, Shimokata eds. Osteoporosis: Etiology, Diagnosis, and Management. New York:
H. Age-related changes in gait velocity and leg extension power Raven Press; 1988:111–131.
in middle-aged and elderly people. J Epidemiol 2000;10:S77– 147. Menz HB, Lord SR, Fitzpatrick RC. Age-related differences in walk-
81. ing stability. Age Ageing 2003;32:137–142.
122. Lajoie Y, Teasdale N, Bard C, Fleury M. Attentional demands for 148. Morris JA, Jr., MacKenzie EJ, Damiano AM, Bass SM. Mortality in
static and dynamic equilibrium. Exp Brain Res 1993;97:139–144. trauma patients: the interaction between host factors and severity.
123. Laros Jr RK. Physiology of normal pregnancy. In: Willson JR, Car- J Trauma 1990;30:1476–1482.
rington ER, eds. Obstetrics and Gynecology. St. Louis, MO: Mosby 149. Mottola MF, Campbell MK. Activity patterns during pregnancy. Can
Year Book; 1991:242. J Appl Physiol 2003;28:642–653.
124. Larson EB, Bruce RA. Exercise. In: Cassell CK, Leipqiz RM, Cohen 150. National Center for Injury Prevention and Control. 10 Leading
HJ, Larson EB, Meier DE, Capello C, eds. Geriatric Medicine: An Causes of Injury Deaths, United States. http://wwwcdcgov/ncipc/
Evidence-Based Approach. New York: Springer-Verlag; 2003: 1023– 2004.
1029. 151. Neiman J, Lang AE, Fornazzari L, Carlen PL. Movement disorders
125. Laughton CA, Slavin M, Katdare K, et al. Aging, muscle activity, in alcoholism: a review. Neurology 1990;40:741–746.
and balance control: physiologic changes associated with balance 152. Nelson ME, Layne JE, Bernstein MJ, et al. The effects of multidi-
impairment. Gait Posture 2003;18:101–108. mensional home-based exercise on functional performance in el-
126. Ledin T, Odkvist L. Abstinent chronic alcoholics investigated by dy- derly people. J Gerontol A Biol Sci Med Sci 2004;59:154–160.
namic posturography, ocular smooth pursuit and visual suppres- 153. Nieschalk M, Ortmann C, West A, Schmal F, Stoll W, Fechner G.
sion. Acta Otolaryngol 1991;111:646–655. Effects of alcohol on body-sway patterns in human subjects. Int J
127. Ledin T, Odkvist LM. Effect of alcohol measured by dynamic pos- Legal Med 1999;112:253–260.
turography. Acta Otolaryngol Suppl 1991;481:576–581. 154. Ning Y, Williams MA, Dempsey JC, Sorensen TK, Frederick IO,
128. Levit KR, Olin GL, Letsch SW. Americans’ health insurance cover- Luthy DA. Correlates of recreational physical activity in early preg-
age, 1980–91. Health Care Financ Rev 1992;14:31–57. nancy. J Matern Fetal Neonatal Med 2003;13:385–393.
129. Lexell J, Taylor CC, Sjostrom M. What is the cause of the ageing 155. Nyska M, Sofer D, Porat A, Howard CB, Levi A, Meizner I. Planter
atrophy? Total number, size and proportion of different fiber types foot pressures in pregnant women. Isr J Med Sci 1997;33:139–
studied in whole vastus lateralis muscle from 15- to 83-year-old men. 146.
J Neurol Sci 1988;84:275–294. 156. Oberg T, Karsznia A, Oberg K. Basic gait parameters: reference
130. Lockhart TE, Woldstad JC, Smith JL. Effects of age-related gait data for normal subjects, 10–79 years of age. J Rehabil Res Dev
changes on the biomechanics of slips and falls. Ergonomics 1993;30:210–223.
2003;46:1136–1160. 157. Overstall PW, Exton-Smith AN, Imms FJ, Johnson AL. Falls in the
131. Lord SR, Anstey KJ, Williams P, Ward JA. Psychoactive medication elderly related to postural imbalance. Br Med J 1977;1:261–264.
use, sensori-motor function and falls in older women. Br J Clin Phar- 158. Patla AE. A framework for understanding mobility problems in the
macol 1995;39:227–234. elderly. In: Craik RL, Oatis CA, eds. Gait Analysis: Theory and Appli-
132. Lord SR, Caplan GA, Ward JA. Balance, reaction time, and muscle cation. St. Louis, MO: Mosby-Year Book Inc; 1995:436–449.
strength in exercising and nonexercising older women: a pilot study. 159. Pellecchia GL. Postural sway increases with attentional demands of
Arch Phys Med Rehabil 1993;74:837–839. concurrent cognitive task. Gait Posture 2003;18:29–34.
133. Lord SR, Clark RD, Webster IW. Physiological factors associated 160. Pellicane JV, Byrne K, DeMaria EJ. Preventable complications and
with falls in an elderly population. J Am Geriatr Soc 1991;39:1194– death from multiple organ failure among geriatric trauma victims.
1200. J Trauma 1992;33:440–444.
134. Lord SR, Ward JA, Williams P. Exercise effect on dynamic stability in 161. Pentney R, Quackenbush LJ, O’Neill M. Length changes in dendritic
older women: a randomized controlled trial. Arch Phys Med Rehabil networks of cerebellar Purkinje cells of old rats after chronic ethanol
1996;77:232–236. treatment. Alc Clin Exp Research 1989;13:413–419.
135. Magaziner J, Simonsick EM, Kashner TM, Hebel JR, Kenzora JE. 162. Pentney RJ, Dlugos CA. Cerebellar Purkinje neurons with altered
Predictors of functional recovery one year following hospital dis- terminal dendritic segments are present in all lobules of the cere-
charge for hip fracture: a prospective study. J Gerontol 1990;45: bellar vermis of ageing, ethanol-treated F344 rats. Alcohol Alcohol
M101–107. 2000;35:35–43.
136. Maki BE. Gait changes in older adults: predictors of falls or indica- 163. Pentney RJ. Remodeling of neuronal dendritic networks with aging
tors of fear. J Am Geriatr Soc 1997;45:313–320. and alcohol. Alcohol Alcohol 1991;(suppl 1):393–397.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

146 Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism

164. Perrin PP, Gauchard GC, Perrot C, Jeandel C. Effects of physical and 189. Scheltens P, Barkhof F, Leys D, Wolters EC, Ravid R, Kamphorst
sporting activities on balance control in elderly people. Br J Sports W. Histopathologic correlates of white matter changes on MRI
Med 1999;33:121–126. in Alzheimer’s disease and normal aging. Neurology 1995;45:883–
165. Peterka RJ, Black FO. Age-related changes in human posture con- 888.
trol: sensory organization tests. J Vestib Res 1990;1:73–85. 190. Schmal F, Thiede O, Stoll W. Effect of ethanol on visual-vestibular
166. Pfefferbaum A, Mathalon DH, Sullivan EV, Rawles JM, Zipursky interactions during vertical linear body acceleration. Alcohol Clin
RB, Lim KO. A quantitative magnetic resonance imaging study of Exp Res 2003;27:1520–1526.
changes in brain morphology from infancy to late adulthood. Arch 191. Scholz E, Diener H, Dichgans J, Langohr H, Schied W, Schupmann
Neurol 1994;51:874–887. A. Incidence of peripheral neuropathy and cerebellar ataxia in
167. Pfefferbaum A, Sullivan EV, Rosenbloom MJ, Mathalon DH, Lim chronic alcoholics. J Neurol 1986;233:212–217.
KO. A controlled study of cortical gray matter and ventricular 192. Seguin R, Nelson ME. The benefits of strength training for older
changes in alcoholic men over a five year interval. Arch Gen Psy- adults. Am J Prev Med 2003;25:141–149.
chiatry 1998;55:905–912. 193. Sehl ME, Yates FE. Kinetics of human aging: I. Rates of senescence
168. Pfefferbaum A, Sullivan EV. Diffusion MR imaging in psychiatry and between ages 30 and 70 years in healthy people. J Gerontol A Biol
ageing. In: Gillard J, Waldman A, Barker P eds. Physiological Mag- Sci Med Sci 2001;56:B198–208.
netic Resonance in Clinical Neuroscience. Cambridge: Cambridge 194. Sheldon JH. The effect of age on the control of sway. Gerontol Clin
University Press; 2005, pp. 558–578. (Basel) 1963;5:129–138.
169. Post RB, Lott LA, Beede JI, Maddock RJ. The effect of alcohol on the 195. Shumway-Cook A, Woollacott M, Kerns KA, Baldwin M. The ef-
vestibulo-ocular reflex and apparent concomitant motion. J Vestib fects of two types of cognitive tasks on postural stability in older
Res 1994;4:181–187. adults with and without a history of falls. J Gerontol 1997;52:M232–
170. Prudham D, Evans JG. Factors associated with falls in the elderly: 240.
a community study. Age Ageing 1981;10:141–146. 196. Skelton DA, Young A, Greig CA, Malbut KE. Effects of resis-
171. Pyka G, Lindenberger E, Charette S, Marcus R. Muscle strength tance training on strength, power, and selected functional abili-
and fiber adaptations to a year-long resistance training program in ties of women aged 75 and older. J Am Geriatr Soc 1995;43:1081–
elderly men and women. J Gerontol 1994;49:M22–27. 1087.
172. Raz N, Gunning FM, Head D, et al. Selective aging of the human 197. Soltani H, Fraser RB. A longitudinal study of maternal an-
cerebral cortex observed in vivo: Differential vulnerability of the thropometric changes in normal weight, overweight and obese
prefrontal gray matter. Cereb Cortex 1997;7:268–282. women during pregnancy and postpartum. Br J Nutr 2000;84:95–
173. Raz N, Rodrigue KM, Kennedy KM, Head D, Gunning-Dixon F, 101.
Acker JD. Differential aging of the human striatum: longitudinal 198. Song MY, Ruts E, Kim J, Janumala I, Heymsfield S, Gallagher
evidence. Am J Neuroradiol 2003;24:1849–1856. D. Sarcopenia and increased adipose tissue infiltration of muscle
174. Reeves S, Bench C, Howard R. Ageing and the nigrostriatal in elderly African American women. Am J Clin Nutr 2004;79:874–
dopaminergic system. Int J Geriatr Psychiatry 2002;17:359–370. 880.
175. Resnick SM, Goldszal AF, Davatzikos C, et al. One-year changes in 199. Sorensen TK, Williams MA, Lee IM, Dashow EE, Thompson ML,
MRI brain volumes in older adults. Cereb Cortex 2000;10:464–472. Luthy DA. Recreational physical activity during pregnancy and risk
176. Riley MA, Clark S. Recurrence analysis of human postural sway of preeclampsia. Hypertension 2003;41:1273–1280.
during the sensory organization test. Neurosci Lett 2003;342:45–48. 200. Sowell ER, Jernigan TL, Mattson SN, Riley EP, Sobel DF, Jones KL.
177. Riley R. Accidental falls and injuries among seniors. Health Rep Abnormal development of the cerebellar vermis in children prena-
1992;4:341–354. tally exposed to alcohol: size reduction in lobules I-V. Alc Clin Exp
178. Rinne JO, Lonnberg P, Marjamaki P. Age-dependent decline in hu- Res 1996;20:31–34.
man brain dopamine-D1 and dopamine-D2 receptors. Brain Res 201. Spencer RM, Zelaznik HN, Diedrichsen J, Ivry RB. Disrupted tim-
1990;508:349–352. ing of discontinuous but not continuous movements by cerebellar
179. Rinne JO, Rummukainen J, Paljarvi L, Sako E, Molsa P, Rinne UK. lesions. Science 2003;300:1437–1439.
Neuronal Loss in the Substantia Nigra in Patients with Alzheimer’s 202. Sternfeld B, Quesenberry CP, Jr., Eskenazi B, Newman LA. Exercise
Disease and Parkinson’s Disease in Relation to Extrapyramidal during pregnancy and pregnancy outcome. Med Sci Sports Exerc
Symptoms and Dementia. Alzheimer’s Disease and Related Disorders 1995;27:634–640.
1989;317:325–332. 203. Sullivan EV, Adalsteinsson E, Hedehus M, et al. Equivalent disrup-
180. Ritchie JR. Orthopedic considerations during pregnancy. Clin Ob- tion of regional white matter microstructure in aging healthy men
stet Gynecol 2003;46:456–466. and women. Neuroreport 2001;12:99.
181. Roebuck RM, Simmons RW, Mattson SN, Riley EP. Prenatal expo- 204. Sullivan EV, Deshmukh A, Desmond JE, Lim KO, Pfefferbaum
sure to alcohol affects the ability to maintain postural balance. Alc A. Cerebellar volume decline in normal aging, alcoholism, and
Clin Exp Res 1998;22:252–258. Korsakoff’s syndrome: Relation to ataxia. Neuropsychology 2000;
182. Rogind H, Lykkegaard JJ, Bliddal H, Danneskiold-Samsoe B. Pos- 14:341–352.
tural sway in normal subjects aged 20–70 years. Clin Physiol Funct 205. Sullivan EV, Rose J, Pfefferbaum A. Effect of vision, touch, and
Imaging 2003;23:171–176. stance on cerebellar vermian-related sway and tremor: A quantita-
183. Rosenberg IH. Sarcopenia: origins and clinical relevance. J Nutr tive MRI and physiological study. Cereb Cortex 2005;in press.
1997;127:990S–991S. 206. Sullivan EV, Rosenbloom MJ, Lim KO, Pfefferbaum A. Longitudinal
184. Rosenbloom MJ, O’Reilly A, Sassoon SA, Sullivan EV, Pfefferbaum changes in cognition, gait, and balance in abstinent and relapsed
A. Persistent cognitive deficits in community treated alcoholic men alcoholic men: Relationships to changes in brain structure. Neu-
and women volunteering for research: Limited contribution from ropsychology 2000;14:178–188.
psychiatric comorbidity. J Stud Alcohol 2005;66:254–265. 207. Sullivan EV, Rosenbloom MJ, Serventi KL, Pfefferbaum A. Effects of
185. Rosenbloom MJ, Pfefferbaum A, Sullivan EV. Recovery of short- age and sex on volumes of the thalamus, pons, and cortex. Neurobiol
term memory and psychomotor speed but not postural stabil- Aging 2004;25:185–192.
ity with long-term sobriety in alcoholic women. Neuropsychology 208. Sullivan EV. Compromised pontocerebellar and cerebellothalamo-
2004;18:589–597. cortical systems: speculations on their contributions to cognitive
186. Rubenstein LZ, Josephson KR, Trueblood PR, et al. Effects of a and motor impairment in nonamnesic alcoholism. Alc Clin Exp
group exercise program on strength, mobility, and falls among fall- Research 2003;27:1409–1419.
prone elderly men. J Gerontol A Biol Sci Med Sci 2000;55:M317–321. 209. Sutherland DH, Olshen RA, Biden EN, Wyatt MP. The Development
187. Rubino FA. Gait disorders. Neurologist 2002;8:254–262. of Mature Walking. Clinics in Developmental Medicine No. 104–105.
188. Schauberger CW, Rooney BL, Goldsmith L, Shenton D, Silva London: MacKeith Press; 1988.
PD, Schaper A. Peripheral joint laxity increases in pregnancy but 210. Taves C, Charteris J, Wall JC. The kinematics of treadmill walking
does not correlate with serum relaxin levels. Am J Obstet Gynecol during pregnancy. Physiotherapy Canada 1982;34:321–324.
1996;174:667–671. 211. Tell GS, Lefkowitz DS, Diehr P, Elster AD. Relationship between
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

Chapter 8 / Gait Adaptations in Adulthood: Pregnancy, Aging, and Alcoholism 147

balance and abnormalities in cerebral magnetic resonance imaging generation occurring in alcoholic patients. Arch Neurol 1959;1:577–
in older adults. Arch Neurol 1998;55:73–79. 688.
212. Tianwu H, Watanabe Y, Asai M, Shimizu K, Takada S, Mizukoshi 224. Victor M, Adams RD, Collins GH. The Wernicke-Korsakoff Syndrome
K. Effects of alcohol ingestion on vestibular function in postural and Related Neurologic Disorders Due to Alcoholism and Malnutri-
control. Acta Otolaryngol Suppl 1995;519:127–131. tion, 2nd ed. Philadelphia: F.A. Davis Co; 1989.
213. Tideiksaar R. Falling in Old Age. New York: Springer Publishing 225. Visser M, Pluijm SM, Stel VS, Bosscher RJ, Deeg DJ. Physical ac-
Company, Inc; 1997. tivity as a determinant of change in mobility performance: the Lon-
214. Tinetti M, Doucette J, Claus E. The contribution of predisposing gitudinal Aging Study Amsterdam. J Am Geriatr Soc 2002;50:1774–
and situational risk factors to serious fall injuries. J Am Geriatr Soc 1781.
1995;43:1207–1213. 226. Wild D, Nayak US, Isaacs B. Description, classification and preven-
215. Tinetti M, Williams C. The effect of falls and fall injuries on func- tion of falls in old people at home. Rheumatol Rehabil 1981;20:153–
tioning in community-dwelling older persons. J Gerontol 1998;53A: 159.
M112–M119. 227. Winter DA, Patla AE, Frank JS, Walt SE. Biomechanical walking
216. Tinetti ME, Doucette JT, Claus EB. The contribution of predisposing pattern changes in the fit and healthy elderly. Phys Ther 1990;70:340–
and situational risk factors to serious fall injuries. J Am Geriatr Soc 347.
1995;43:1207–1213. 228. Wober C, Wober-Bingol C, Karwautz A, Nimmerrichte RA, Deecke
217. Toupet M, Gagey PM, Heuschen S. Vestibular patients and aging L, Lesch OM. Postural control and lifetime alcohol consump-
subjects lose use of visual input and expend more energy in static tion in alcohol-dependent patients. Acta Neurol Scand 1999;99:
postural control. In: Vellas BJ, Toupet M, Rubenstein L, et al, eds. 48–53.
Balance and Gait Disorders in the Elderly. Paris: Elsevier; 1992: 229. Wober C, Wober-Bingol C, Karwautz A, Nimmerrichter A, Walter H,
183–198. Deecke L. Ataxia of stance in different types of alcohol dependence-a
218. Tremblay F, Mireault AC, Dessureault L, Manning H, Sveistrup H. posturographic study. Alcohol 1998;33:393–402.
Postural stabilization from fingertip contact: I. Variations in sway 230. Wolf SL, Barnhart HX, Kutner NG, McNeely E, Coogler C, Xu T. Re-
attenuation, perceived stability and contact forces with aging. Exp ducing frailty and falls in older persons: an investigation of Tai Chi
Brain Res 2004;157:275–285. and computerized balance training. Atlanta FICSIT Group. Frailty
219. Tsang WW, Hui-Chan CW. Effect of 4- and 8-wk intensive Tai Chi and Injuries: Cooperative Studies of Intervention Techniques. J Am
Training on balance control in the elderly. Med Sci Sports Exerc 2004; Geriatr Soc 1996;44:489–497.
36:648–657. 231. Wolf SL, Coogler C, Xu T. Exploring the basis for Tai Chi Chuan
220. U.S. Census Bureau. Aging in the United States: Past, Present, and as a therapeutic exercise approach. Arch Phys Med Rehabil 1997;78:
Future. http://wwwcensusgov/ipc/prod/97agewcpdf 2004. 886–892.
221. U.S. Census Bureau. Census 2000 PHC-T-9. Population by Age, Sex, 232. Woollacott M, Shumway-Cook A. Attention and the control of pos-
Race, and Hispanic or Latino Origin for the United States: 2000. ture and gait: a review of an emerging area of research. Gait Posture
http://wwwcensusgov/population/cen2000/phc-t9/tab01pdf 2000. 2002;16:1–14.
222. Vellas BJ, Wayne SJ, Romero LJ, Baumgartner RN, Garry PJ. Fear 233. Zeleznik J. Normative aging of the respiratory system. Clin Geriatr
of falling and restriction of mobility in elderly fallers. Age Ageing Med 2003;19:1–18.
1997;26:189–193. 234. Zhang J, Savitz DA. Exercise during pregnancy among US women.
223. Victor M, Adam RD, Mancell EL. A restricted form of cerebellar de- Ann Epidemiol 1996;6:53–59.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-08 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:16

148
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

• C h a p t e r 9
•
◗ Walking for Health
William L. Haskell and Leslie Torburn

Evolutionary theory and data have demonstrated that hu- increase in obesity and type II diabetes mellitus in many
mans carrying genetic alleles favoring such motor skills as technologically advanced cultures (46). The present in-
stamina, endurance, strength, speed, and agility at relevant crease in obesity, type II diabetes, and cardiovascular dis-
genes were more likely to experience better reproductive ease indicate that walking and other physical activity needs
fitness than their counterparts who were not so endowed to be included in the daily life of both children and adults.
(7). Good mobility was critical for survival, and walking One way to reverse this decline and effectively combat
became an increasingly important feature of daily life in this “hypokinetic state” is to increase substantially the
the evolving human. Without other means of transporta- amount of brisk or vigorous walking performed by a large
tion, people who could walk considerable distances during segment of the sedentary population. This chapter pri-
a day or on consecutive days were more likely to survive. marily will review the evidence that supports the health
This increase in physical activity activated gene expression and performance benefits of walking, how much walking
that further enhanced their motor fitness or capacity (29). the population currently performs, and some issues sur-
For more than 99.5% of the last 100,000 years, modern rounding attempts to improve health through increased
Homo sapiens have primarily relied on walking for moving walking.
about and muscular work to perform most daily chores. It
“I find in the domestic duck that bones of the wing weigh
is only within the past 200 years that man has been able
less and the bones of the leg more, in proportion to the
to reduce systematically his required daily physical activ- whole skeleton, than do the same bones in the wild-duck;
ity and especially walking through a variety of technologi- and I presume this change may be safely attributed to the
cal developments. Thus, modern man has inherited a body domestic duck flying much less and walking more, than
exquisitely designed for a wide range of physical activities, its wild parent” Charles Darwin – 1859 (9)
and it functions best when activities, such as walking, are
a significant part of daily life.
WALKING IN THE USA
Throughout much of recorded western history, physi-
cians, philosophers, educators, and scientists have pro-
Data from Surveys
moted the idea that being physically active contributes
to improved health, better physical functioning, and in- Based on various physical activity surveys of representa-
creased longevity. Taking frequent walks or similar activity tive samples of the United States adult population con-
to help prevent or treat various chronic diseases and pro- ducted over the past 30 years, walking is the most fre-
mote “successful aging” has been a frequent recommenda- quently reported physical activity. Most surveys have only
tion by a number of early major thought leaders (“Before assessed the walking performed during leisure time or for
supper take a little walk, after supper do the same” – Eras- transportation and have not included walking involved
mus [1514] or “Of all exercises walking is the best” – Thomas during occupational work or household chores. Data col-
Jefferson [1791]). lected in 2000 using the Behavioral Risk Factor Surveil-
While these personal observations or clinical impres- lance System by the Centers for Disease Control and Pre-
sions were valuable in promoting a healthier lifestyle, it vention indicated that about 30% of men and 47% of
was not until the mid-1900s that data collected and ana- women reported walking as a leisure-time physical activ-
lyzed scientifically were published describing some of the ity (47). Both men and women who reported walking for
health benefits of walking and other moderate or vigor- leisure time walked an average of 2.87 times per week, with
ous intensity activities. The fact that walking and similar an average duration of 34.4 minutes for men and 29.9 min-
activities need to be included in the lives of many chil- utes for women. The prevalence for walking during leisure
dren as well as adults is highlighted by the recent rapid time was highest for men and women age 45 to 54 years

149
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

150 Chapter 9 / Walking for Health

(53.5%) and lowest for those age 18 to 24 years (38.1%), per day was strongly associated with degree of adiposity:
whereas 52% of college graduates reported walking for normal weight (BMI <25) = 7,029, overweight (BMI
leisure time but only 35.6% of participants reporting less 25–29.9) = 5,813, and obese (BMI ≥30) = 4,618 steps
than a high school education reported such walking. Non- per day.
Hispanics reported the highest rate of walking (49.4%), The idea that a good walking goal for promoting health
while Hispanic men and women reported the lowest preva- is 10,000 steps per day was introduced in Japan in the
lence (39.9%). These data are quite similar to the results 1960s (17). This 10,000-step level as a physical activity goal
of a random sample survey of 1,818 United States men for adults was exported to many countries in the late 1990s.
and women collected during 1999 to 2000, where 34% of Since then, a number of investigators have evaluated how
the population were identified as regular walkers (5 times this amount of walking relates to current public health
per week for ≥30 minutes), 45.6% as occasional walkers recommendations, such as the CDC/ACSM public health
(<5 times per week or <30 minutes per session) and 20.7% guidelines for physical activity that recommends at least
as never walkers (13). 30 minutes of moderate intensity activity on most, prefer-
ably all, days of the week (41). The results of these evalu-
ations generally indicate that when most adults regularly
Walking Based on Data
achieve 10,000 steps per day, they meet the current recom-
from Pedometers
mendations for physical activity. For example, in healthy
Pedometers and motion sensors (usually accelerometers) adults who reported at least 30 minutes of moderate inten-
have been used to evaluate the physical activity habits, sity activity per day on a seven-day physical activity recall,
especially walking, of various groups of individuals. Ac- 73% recorded at least 10,000 steps per day on a pedome-
celerometers have been used to study habitual physical ac- ter (58). Practitioners have rapidly adopted the idea of a
tivity or changes in activity and have been preferred over 10,000-step goal for the promotion of physical activity in a
pedometers because they can be used to evaluate a variety variety of clinical and community settings. However, little
of different types of activity, not just ambulation, and they systematically collected data have been published on how
have greater accuracy in determining an overall activity successful this approach is in maintaining an increase in
profile (27). However, the substantially greater costs and daily walking as compared to other approaches.
logistical challenges when using accelerometers as com-
pared to pedometers to collect data over extended peri-
ods of time in a large number of participants has made WALKING AND MAINTAINING GOOD
pedometers a popular instrument for recording walking. HEALTH THROUGHOUT THE LIFESPAN
Systematic evaluations of pedometers have demonstrated
that a number of well-made pedometers can provide ac- Over the past 50 years, scientific data have continued to
curate data on the amount of walking performed by most accumulate supporting the association between increased
adults (45). Pedometers may be less accurate in children habitual moderate intensity physical activity, including
and older adults who have limited mobility and an altered brisk walking, with a significantly lower prevalence of vari-
(shuffle) gait. ous chronic diseases, greater physical independence in the
The results of several studies have been reported where elderly and a better overall health status and quality of life
pedometers were used to assess the amount of walking among many adults. Men and women who include bouts
performed throughout the day in representative samples of brisk walking on most days have lower occurrences of
of the adult population. These data indicate that for adults fatal and nonfatal heart attack and stroke, hypertension,
the number of steps per day typically ranges between 2,500 type II diabetes mellitus, site-specific cancers (especially of
and 15,000, with very sedentary people achieving less than the colon and breast), osteoporosis, and depression. Fre-
4,000 steps per day. Sedentary people being in the range of quent walking by older persons helps to maintain and/or
4,000 to 6,000, moderately active in the range of 6,000 to increase cardiorespiratory and muscle endurance, as well
9,000, active (achieving public health recommendations as muscle strength, balance and gait speed, all important
for physical activity) 9,000 to 12,000, and the very active factors for retaining the capacity to perform a wide range
at more than 12,000 steps per day (51). Pedometer data of tasks of daily living and maintain independent living.
collected over a seven-day period on a representative
sample of adults living in South Carolina indicated that
Cardiovascular Disease
the average number of steps taken per day was 5,931 (sd
± 3,634) with men recording 7,192 ± 3,596 and women Men and women who accumulate 120 minutes or more
5,210 ± 3,518 steps per day (52). Adults at the age of 35 to per week of brisk walking during most weeks generally
44 years accumulated more steps per day than either have lower rates of cardiovascular disease (CVD) and clin-
younger or older adults did, with those 65+ years record- ical events caused primarily by the atherothrombotic pro-
ing only 3,766 ± 2,805 steps per day. The number of steps cess. The initial support for this relationship using modern
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

Chapter 9 / Walking for Health 151

scientific methods was published by Morris and colleagues outside the home for a duration of ≥10 minutes without
in 1953 (34). They reported an inverse association between stopping, the average duration of each walk, and the walk-
on-the-job activity and future CVD clinical events in two ing pace. Using energy expenditure data provided in the
occupations, double-decker bus conductors and postal car- Physical Activity Compendium (2), the estimated amount
riers as compared to their sedentary counterparts, bus of energy expended while walking was calculated as MET-
drivers, and civil service workers who sat at desks most hours per week. The average length of follow-up was
of the day. The rate of first clinical episodes of coronary 3.7 years with 232,971 person-years of exposure and 1,551
artery disease (CAD) in the conductors was about 30% CVD events.
lower than in the drivers. While actual walking time or In Figure 9-1 the relation between amount of energy
distance was not measured, the major activity of the bus spent per week walking and development of CVD during
conductors was walking up and down the stairs collect- follow-up is provided for women ages 50 to 59, 60 to 69,
ing fares on the double-decker buses and all of the postal and 70 to 79 years. Women were categorized into quintiles
carriers completed most of their routes by walking. according to MET-hours per week while walking with the
Over the past 50 years, since this initial report by Morris values for low to high quintiles being 0, 0.1 to 2.5, 2.6 to
and colleagues, there have been a number of well-designed, 5.0, 5.1 to 10.0, >10.0 MET-hours per week. There was a
prospective observational studies that have linked in- highly significant association ( p <0.001) between higher
creased amounts of walking with lower CVD risk for men amounts of walking and decreased risk of having a CVD
and women. In most cases the walking intensity has been clinical event over the next two to six years. The relation-
defined as “usual pace” or “brisk” and the amount in min- ship was similar for women in all three age groups, with
utes per day or hours per week. In the data analysis of a somewhat less strong association in the oldest women.
several of these reports, the amount of walking has been The association between pace of walking and CVD clini-
separated from other activities, especially more vigorous cal events was also analyzed in this study. Women were
activities. This approach allows for a more refined analy- categorized by usual self-reported walking pace into five
sis of the association between walking amount and clinical categories: 1) rarely or never walk, 2) <2 mph, 3) 2 to 3
CVD events independent of other physical activity. A sum- mph, 4) 3 to 4 mph, and 5) >4 mph. As can be seen in Fig-
mary of the results of such a study is presented in Figures ure 9-2, there was a highly significant trend ( p for trend
9-1 and 9-2. These data are from the Observational Study = 0.002) with women who report walking faster having
of the Women’s Health Initiative (30). Participants were less CVD during follow-up. The overall favorable associa-
73,743 healthy women (without cardiovascular disease or tion between walking and better cardiovascular health was
cancer) 50 to 79 years of age at study entry who were eval- seen in women representing different ages, ethnicities, and
uated in 42 clinics and represented a broad spectrum of body size as determined by body mass index.
the US population. At a baseline visit, a variety of health A number of other studies reporting on the relation of
and lifestyle measurements were recorded, including a de- routine walking amount and risk of CVD are quite consis-
tailed assessment of recreational physical activities clas- tent with the overall results of the Women’s Health Initia-
sified as mild, moderate, or vigorous intensity. Questions tive. For example, results from the US Nurses Health Study
were asked regarding the frequency of walks performed demonstrated that the incidence of CHD among 72,488

FIGURE 9-1. Age-adjusted relative risks for car-

diovascular disease according to quintile of walk-
ing amount (MET–hours/week) in women age 50 to
79 years (Women’s Health Initiative Study).
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

152 Chapter 9 / Walking for Health

FIGURE 9-2. Age and walking amount adjusted

relative risks for cardiovascular disease according
to average speed of walking in women age 50 to
79 years (Women’s Health Initiative Study).

women age 40 to 65 years during 8 years of follow-up in most of these studies, the design of the physical
was much lower in women who accumulated ≥10 MET- activity assessments did not allow for just an analysis
hours per week while walking than those who accumulated of walking independent of other activity. In the Nurses
≤0.5 MET-hours per week (R2 adjusted for age = 0.46, Health Study involving 72,488 women aged 40 to 65 years,
p <0.001) (31). The authors concluded from the results of where walking was analyzed independent of other activi-
this study “brisk walking and vigorous exercise are asso- ties, women reporting walking ≥3.3 miles per week versus
ciated with substantial and similar reductions in the in- those walking <0.2 miles per week experienced a signif-
cidence of coronary heart disease among women.” In the icantly lower incidence of stroke (R2 = 0.66, p <0.01)
Honolulu Heart Program, men age 71 to 93 years who re- during an average follow-up period of eight years (19).
ported walking at least 1.5 miles per day had nearly 50%
lower CHD incidence over a 2- to 4-year follow-up period
compared to men reporting walking less than 0.25 miles
per day (age adjusted, p = 0.002) (16). Also, in a random Peripheral Arterial Disease
sample of generally healthy men and women enrollees in a In patients with peripheral arterial disease (PAD), steno-
health maintenance organization age ≥65 years, walking sis of the arteries in the legs reduces blood flow caus-
more than 4 hours per week was associated with a signifi- ing hypoxia to the contracting muscles during exercise,
cantly lower rate of hospitalizations due to CVD compared especially during walking. This condition, frequently re-
to their contemporaries who reported walking <1 hour per ferred to as intermittent claudication, is caused by the
week (R2 = 0.69, p <0.01) (24). Total mortality rate was atherothrombotic process that develops most frequently
lower in those who walked more than 4 hours per week, in the intermediate-size arteries of the lower legs. When
especially in women (R2 = 0.45, p <0.01). the reduced capacity for blood flow in such arteries can-
not meet the increased demands for flow caused by ex-
ercise, tissue ischemia and local muscle pain occurs. As
Stroke
the exercise continues in duration and/or intensity, the is-
A lower stroke rate in more physically active or physically chemia increases until the patient stops the exercise due
fit persons compared to the least active or fit in the study to leg pain. Major risk factors for developing PAD include
population has been reported in a number of studies. cigarette smoking, diabetes mellitus, hypertension, hyper-
In a meta-analysis of data available in 2002, Lee and cholesterolemia, a diet high in animal fats and low in plant
colleagues concluded that moderate and high amounts products, and a sedentary lifestyle (8). Limited observa-
of physical activity are associated with a reduced risk tional data support the idea that frequent walking helps to
of total, ischemic, and hemorrhagic strokes in men and prevent the development of PAD in older men and women,
women (26). This report points to the need for more but compared to the data linking walking to the reduced
data on physical activity amount and intensity and stroke risk of CHD, the data on walking and PAD prevention is
incidence, especially for hemorrhagic stroke. However, still quite sparse (38).
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

Chapter 9 / Walking for Health 153

In patients with PAD, frequent walking is considered fibrinolysis, a decrease in myocardial oxygen demand at
to be one of the cornerstones of therapy. In conjunction rest and work along with an increase in aerobic capacity
with pharmacologic treatment and possibly surgery, fre- (maximal oxygen uptake), and an increase in endothelium-
quent walking is prescribed to maintain physical func- mediated coronary artery vasodilatation capacity and an-
tioning and independence. In a systematic review of the giogenesis. While not all of the changes have been shown
effects of exercise training in patients with PAD, Gardner to be produced specifically by an increase in walking, most
and Poehlman concluded that exercise, especially walking, appear to improve to some degree by moderately intense
was effective in reducing the clinical symptoms and in- physical training. For example, substantial data exist that
creasing walking performance (15). The claudication pain increased walking will help prevent the development of hy-
end point used during exercise rehabilitation (i.e., inter- pertension and reduce arterial blood pressure. In a study
mittent exercise to near-maximal pain) was the most im- of 6,017 generally healthy Japanese men age 35 to 60 years
portant program component for producing improvements who had a baseline blood pressure <140/90 mm Hg, it was
in distance walked, as it explained 55% and 40% of the observed that men who walked more had a lower rate of
variance in the increases in the distances to onset and developing hypertension than men who walked little (50).
to maximal claudication pain, respectively. Exercise pro- During 59,574 person-years of follow-up, the multivariate
grams that had patients walk to near-maximal claudication relative risk for developing hypertension was 29% lower
pain (high pain end point) demonstrated greater improve- ( p for trend = 0.002) for men who reported walking >20
ments in claudication symptoms than programs that had minutes to work each day compared to men reporting a
patients stop walking at the onset of claudication pain (low walk to work of ≤10 minutes per day. In this population,
pain end point). These data support the notion that greater for every 26 men who walked more than 20 minutes to
amounts of ischemia induced within the claudicating mus- work daily, one case of hypertension was prevented.
culature may produce greater improvements in pain symp- In people who are quite inactive and have a low aerobic
toms, possibly due to greater hemodynamic and metabolic capacity, brisk walking puts a sufficient stress (overload)
adaptations. on a number of the body’s tissues or systems to produce an
increase in their efficiency or capacity (42). That increased
exercise intensity, within the normal range of walking,
Physiological Changes Contributing
produces step-wise increases in cardiorespiratory func-
to Less CVD
tion (aerobic capacity) in inactive women was effectively
A number of biological measures causally linked to the demonstrated by Duncan and colleagues (10). Displayed
development of CVD and/or the precipitation of clinical in Figure 9-3 are changes in maximal oxygen uptake mea-
events are known to be altered favorably with moderately sured during walking on a treadmill and HDL-cholesterol
intense physical activity. These include the decrease in ar- concentrations for 102 sedentary, healthy premenopausal
terial blood pressure, alterations in the lipoprotein profile, women randomly assigned to control (no change in phys-
especially the reduction in triglyceride and the increase in ical activity), “strollers” who walked at 3 mph, “walkers”
high-density cholesterol concentrations, enhancement of who walked at 4.0 mph, and “brisk walkers” who walked
insulin-mediated glucose uptake (glucose tolerance), less at 5.0 mph. All the women assigned to walking walked
of a tendency for red blood cell clotting and/or increased 3 miles per day, five times per week for 24 weeks. All

FIGURE 9-3. Percent changes after 24 weeks in

maximal oxygen uptake (VO2 max) and high-density
lipoprotein cholesterol (HDL-C) in response to walk-
ing at different speeds but the same distance.
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

154 Chapter 9 / Walking for Health

three exercise groups achieved a significant increase in effect where an improvement in insulin sensitivity is de-
maximal oxygen uptake, but those who walked faster (but tected only after weeks or months of increased exercise, the
still just 3 miles per day) had greater increases than the improvement is seen after only a few sessions of moderate
women who walked slower. These data support the idea intensity walking (acute response). Rogers and colleagues
that by increasing walking speed while holding the walk- reported that after only seven sessions of brisk walking and
ing distance constant, a person can gain a greater increase cycling (50 to 60 minutes per session on seven consecutive
in cardiorespiratory capacity. However, this does not ap- days) by 10 men with poor glucose tolerance, they had a
pear to be true for changes in HDL-cholesterol as all three significant reduction in both glucose and insulin concen-
exercise groups had significant but similar increases in trations over three hours following an oral glucose load of
HDL-cholesterol concentration over the 24 weeks. Thus, 100 g (Figure 9-4) (44).
for HDL-cholesterol, it appears here and in other studies Men and women who are habitually more physically
the total amount of walking performed is more important active, including more walking, are significantly less likely
than the speed of walking or intensity of the exercise. For to develop type II diabetes in the future than less active
a number of health related measures, such as blood lipids, persons. In the Nurses Health Study, women in the upper
insulin-mediated glucose uptake and obesity prevention, quintile of reported walking had a relative risk for devel-
the total amount of exercise performed is more important oping type II diabetes over an 8-year follow-up period of
than performing it at a very high intensity (22,48). 0.58 ( p = 0.001 for trend) and 0.74 when adjusted for
BMI ( p = 0.01 for trend) compared to women in the
lowest quintile of walking. A faster pace of walking was
Type II Diabetes
Since the 1980s there has been a steady and rapid increase
in the prevalence of type II diabetes in the populations of
most technologically developed countries. In addition to
the increase in prevalence, the age of onset of type II dia-
betes has continued to decrease with an increasing num-
ber of cases being reported in boys and girls in their teens
(11). This increase in prevalence and earlier age of onset
is beginning to cause major havoc with health care costs
due to the long-term and expensive care to deal not only
with the medical management of the diabetes but also the
treatment of the blindness, leg amputations, end state re-
nal disease, and coronary heart disease caused by the di-
abetes. While there is genetic variation among individuals
in their susceptibility to type II diabetes, a major cause of
this disease is a chronic exposure to inactivity, overeating,
positive energy balance and the resulting obesity (21). The
process of type II diabetes is initiated when there is an
increase in the resistance to insulin-mediated glucose up-
take (insulin resistance) and the pancreas has to produce
more and more insulin in order to move glucose from the
blood into tissues. Eventually, the pancreas is not able to
produce sufficient insulin to overcome this resistance and
blood sugar rises to a level where diabetes is diagnosed.
The major tissue involved in this “insulin resistance”
is the skeletal muscle. Substantial experimental data col-
lected in animals and humans have demonstrated that fre-
quent muscle contractions, such as occur while walking,
play a critical role in maintaining or increasing insulin
sensitivity or decreasing insulin resistance in the skeletal
muscle (43). These increases in insulin-mediated glucose
uptake produced by exercise can significantly reduce blood
glucose levels following an oral glucose challenge (glu- FIGURE 9-4. Effect of 7 days of exercise (primarily walking)
on plasma glucose and insulin response to a 100 gram glucose
cose tolerance test) (23). The amount of insulin needed tolerance test (OGTT) in 10 men with abnormal glucose toler-
to produce this improved glucose uptake is significantly ance. OGTT was performed 18 hours after last exercise bout.
reduced. Rather than being mainly an “exercise training” NE = non-exercise control group; E = exercise group.
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

Chapter 9 / Walking for Health 155

independently associated with a lower risk of type II di- naires, pedometers, and a log for seven consecutive days.
abetes (18). British men age 40 to 59 years who reported The average number of steps per day for men was 18,425
moderate intensity activity (including a lot of walking) had and 14,196 for women and the men reported walking
a relative risk of 0.42 ( p <0.001 for trend across activity 12.0 hours per week and the women 5.7 hours per week.
levels) for developing type II diabetes during eight years of None of Amish men and only 9% of the women were obese
follow-up compared to men considered inactive (56). The (BMI ≥30), while 25% of the Amish men and 26% of the
results of these prospective observational studies showing women were overweight or obese (BMI >25). By contrast,
a reduced risk of developing type II diabetes in more ac- in the general population, the majority of adults do not
tive men are supported by experimental studies where in- walk more than 2 to 3 hours per week, accumulate less
active persons at high risk of developing type II diabetes than 5,000 steps per day, and approximately 56% have BMI
are randomized to an exercise group or a sedentary con- >25 and 20% are obese (BMI ≥30) (33).
trol group. Data from the Da Qing IGT and Diabetes Study The contribution of increased walking to weight loss
in China support the idea that an increase in physical ac- in overweight adults has been investigated repeatedly over
tivity by men and women with elevated blood glucose lev- the past 30 years with mixed results. Programs that have
els significantly reduces the rate of development of type II included an increase in walking ≥12 miles per week for
diabetes as compared to an inactive control group (39). more than 12 weeks frequently have reported significant
Over a 6-year period of intervention, there were 15.7 new weight loss, while programs that included less exercise typ-
cases per 100 person years in the control and 8.3 new cases ically have not. Pollock and colleagues (42) had sedentary
per 100 person years in the exercise group, (46% decrease, middle-aged men walk 40 minutes per session, 4 days per
p <0.001). The exercise program for patients included a week for 20 weeks, with walking distance averaging 12 to
mixture of mild to vigorous activities, with slow to fast 13 miles per week over the last 12 weeks. Compared with
speed walking being one of the major forms of exercise. sedentary controls, these men lost an average of 1.3 kg
The exercise benefits occurred in subjects who were either body weight ( p <0.05) and reduced their percentage of
lean or overweight at baseline. body fat 1.1% ( p <0.01). In a 12-month weight loss
study involving 201 premenopausal, overweight sedentary
women, Jakicic and colleagues demonstrated that the mag-
Walking and the Obesity Epidemic
nitude of weight loss was related to the amount of exercise
The relation between physical activity and obesity in the performed (20). As displayed in Figure 9-5, women who
general public is complex because body-weight gain or loss walked <150 minutes per week had about a 6% weight
depends on the energy intake from food as well as the en- loss at 6 and 12 months, while those women who walked
ergy expenditure through physical activity. When one looks >200 minutes per week had a 12% to 14% weight loss dur-
at the extreme ends of the physical activity continuum in ing the same period. In some studies where an increase
a population, those at the low end generally have a higher in physical activity is provided during structured sessions,
prevalence of being overweight (BMI = 25.0 to 29.9) or no or little weight loss was observed because participants
obese (BMI ≥30) than those at the high end. However, reduced activity performed during other times of the day.
in the broad and relatively homogenous middle range of
energy expenditure in the adult population, the reported
Osteoporosis, Osteoarthritis,
level of physical activity amount usually does not corre-
and Bone Health
late well with measures of adiposity. This is particularly
true more recently as “daily required physical activity” has Most current recommendations for improving general
decreased for a greater and greater proportion of the popu- bone health, especially for increasing or maintaining bone
lation because of technological advances. People now have mineral density and preventing fractures related to os-
a reduced need to move in the home and on the job. Fur- teoporosis include the frequent performance of moderate
thermore, economic policies and incentives favor sitting at or vigorous intensity weight-bearing physical activity, in-
a desk for long hours, as the work environment constrains cluding walking. Osteoporosis is a disease characterized
a person’s opportunities for daily activity (25). by decreased bone mass and leading to increased bone
Bassett and colleagues documented that high levels of fragility and susceptibility to fracture. Compression and
daily physical activity, especially substantial walking, are bending forces placed on the skeleton during muscle con-
associated with a healthy body weight in Old Order Amish tractions and the action of gravity during exercise stim-
men and women (4). These Old Order Amish living in ulate bone formation and strength. Generally, vigorous
Ontario, Canada, refrain from driving automobiles, using intensity or high-impact exercise produces the greatest
electrical appliances, and employing other modern conve- stimulus for increased bone strength, but most weight-
niences, and manual labor farming was the primary occu- bearing activities provide benefit. Walking briskly (see
pation. The amount of physical activity performed daily by section on power walking) as well as up and down hills or
98 men and women was assessed using standard question- stairs provides sufficient stress on the bones of the hip, legs,
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

156 Chapter 9 / Walking for Health

FIGURE 9-5. Percent changes in body weight based

on exercise duration (walking on treadmill in home)
in 196 premenopausal women participating in a
comprehensive weight loss program. Group assign-
ment based on average amount of exercise at 6 and
12 months. Groups that have the same symbol (as-
trisk or dagger) are significantly different from one
another.

and feet to stimulate increased bone growth in children similar activities prevent the development of osteoarthri-
and young adults and decrease the rate of bone density tis, but there are data from observational and experimen-
loss in postmenopausal women and older men (55). As tal studies showing that moderate-intensity resistance and
with many of the health benefits of walking, other good aerobic exercise, including walking, can help reduce joint
health habits such as not smoking and adequate calcium pain and swelling and increase functional capacity in pa-
intake are important in maximizing bone health. tients with osteoarthritis (53). Ettinger and colleagues con-
In addition to evidence that frequent exercise helps to ducted an 18-month randomized clinical trial comparing
maintain bone strength, it has been shown that leisure- walking or resistance exercise to participation in a health
time physical activity, including walking, reduces the risk education class in 365 men and women age ≥60 years with
of hip fractures in postmenopausal women (14). A total osteoarthritis of the knee (12). The walking and resistance
of 61,200 generally healthy women aged 40 to 77 years exercise sessions lasted 60 minutes and were performed
were followed for a period of 12 years during which time three times per week. Overall compliance with the exer-
incidence of hip fracture was assessed. After controlling for cise sessions was 68% in the walking group and 70% in
age, body mass index, use of postmenopausal hormones, the resistance-training group. When compared with par-
smoking, and dietary intake, the most active women had ticipants in the health education class (exercise controls)
a 55% lower incidence of hip fracture than the least active patients in both exercise groups showed significant im-
women (R2 = 0.45, 95% CI = 0.32–0.63). Among women provements in self-reported pain and disability scores, six-
who did no other exercise, walking at least 4 hours per minute walk time, stair climb test, and time to lift and
week was associated with 41% lower risk of hip fracture carry 10 pounds. These data support the use of supervised
(R2 = 0.59, 95% CI = 0.37–0.94) compared to women who walking in the comprehensive medical management of
walked <1 hour per week. osteoarthritis.
Osteoarthritis is a chronic degenerative joint disease
characterized primarily by progressive loss of articular car-
Cancer
tilage, which leads to the narrowing of the joint space, pain,
restriction in motion, crepitus, and deformity. Chronic, Over the past two decades data have continued to accu-
repetitive muscular loading of the joint, especially dur- mulate supporting an inverse relationship between level
ing occupational work and vigorous athletic training and of habitual physical activity and overall cancer incidence
competition can lead to an increase in the risk of os- as well as selected site-specific cancers, especially of the
teoarthritis (40,54). However, joints tolerate slowly applied colon and breast. For many other cancers there are inade-
loads much better than sudden impact or torsional load- quate data to determine if a significant relationship exists,
ing. There is no good evidence that increased walking or but favorable trends have been reported for lung and some
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

Chapter 9 / Walking for Health 157

reproductive organ cancers. However, due to the complex and 24% of those women in the highest to lowest quartiles
nature of the etiology of most cancers, including the long of blocks walked ( p <0.001 for trend). After adjustment
delay between risk exposure and the clinical display of the for age, educational level, comorbid conditions, smoking
tumor, the relatively low-prevalence of most site-specific status, estrogen use, and functional limitations, women in
cancers in the general population and our lack of under- the highest quartile of walking remained 34% less likely
standing about biological mechanisms by which exercise than women in the lowest quartile to develop cognitive
might protect against a cancer, it has not yet been possi- decline. Walking has been reported to protect against cog-
ble to establish a causal link between physical activity and nitive decline in older men as well as women (3).
cancer morbidity or mortality. There is no consistent evi- Distance walked per day was assessed by questionnaire
dence that an increase in physical activity, including walk- in 2,257 physically capable men aged 71 to 93 years as part
ing, decreases a person’s risk of acquiring cancer. In the of the Honolulu-Asia Aging Study (1). During the course
numerous observational studies that have investigated ei- of a 6- to 8-year follow-up, men who walked the least
ther occupational or leisure-time physical activity and can- (<0.25 miles per day) had a 1.8-fold excess risk of devel-
cer risk, it appears that moderate intensity activities (4–5 oping dementia compared to those who walked more than
METS or higher) are more strongly associated with lower 2 miles per day (17.8 versus 10.3/1,000 person-years).
cancer risk than activities that require less than 4 METS Somewhat similar results have been reported from the
(49). Thus, for walking to meet these criteria for cancer Canadian Study of Health and Aging in which a nation-
prevention, walking briskly (≥4 mph), walking up stairs wide population-based sample of 4,615 men and women
or hills or hiking with a pack all should be considered. aged ≥65 years free of dementia at baseline were evaluated
Recently, some cancer treatment centers have been pro- 5 years later (28). One of the stronger predictors of de-
viding supervised aerobic and resistance exercise train- veloping Alzheimer disease was baseline level of physical
ing programs for patients following surgery, radiation, activity. Subjects reporting regular physical activity (a sub-
or chemotherapy for their cancer. These programs are stantial amount being walking) had a 41% lower risk of de-
designed to improve physical performance and indepen- veloping Alzheimer disease compared to persons reporting
dence, increase lean body mass and reduce adiposity, no regular physical activity. A biological basis for such an
and enhance health-related quality of life. The effects effect is not known but could be related to less vascular
of a home-based walking program on fatigue, physical disease and a lower rate of multi-infarct dementia in the
functioning, and emotional distress were evaluated in more physically active older persons.
46 women during a 6-week program of radiation ther-
apy for breast cancer (32). Women randomly assigned
to the walking program experienced significantly greater POWER WALKING FOR FITNESS
improvements in physical functioning, fatigue, emotional AND HEALTH
distress, and sleeping compared to women assigned to
usual care. In order to achieve the beneficial effects of exercise, adults
must participate in moderate-intensity physical activity on
most days of the week, for at least 30 minutes (41). In
Cognitive Function and Dementia
terms of energy expenditure, moderate-intensity activity
Moderate intensity exercise when performed on a regu- is defined as 3 to 6 METS (work metabolic rate/resting
lar basis appears to have favorable influences on a vari- metabolic rate). Brisk walking at a pace of 4 mph meets
ety of mental and psychological functions. It has been re- this criterion for most adults (2). The actual energy expen-
ported that frequent walking contributes to a higher level diture for an individual will vary depending on walking
of cognitive function and to lower occurrence of demen- speed, surface (smooth or uneven, soft or hard, etc.), and
tia in older women and men. Among 18,766 women aged gradient (horizontal, up, down) (2). Personal factors such
70 to 81 years who reported no vigorous physical activity, as mechanical efficiency (e.g., inefficient gait due to hip or
those women in the upper two quartiles of walking dura- leg deformity) may also affect energy expenditure in some
tion (1.5 to 2.8 hours per week and >2.8 hours per week) individuals.
had significantly higher cognitive scores on all measures Comfortable walking speed is typically about 3 mph,
as compared to women in the lowest quartile of walking or 80.5 m/min (36,57). To achieve brisk walking speed
(<38 minutes per week) (59). The average speed of walking of 4 mph (107 m/min), or a vigorous exercise pace of
was estimated at 21 to 30 minutes per mile or 2 to 3 mph. 5 mph (134 m/min), the mechanics of walking must
The relationship of walking to the development of cogni- be altered from that of regular comfortable walking.
tive impairment was studied in 5,925 community-dwelling Murray et al. studied walking at slow, free, and fast
women aged ≥65 years without baseline cognitive impair- speeds. When comparing free walking (85 m/min) to
ment or physical limitations who were followed for 6 to fast walking (115 m/min), they found fast walking re-
8 years (60). Cognitive decline occurred in 17%, 18%, 22% sulted in an increased arc of hip motion due to increased
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

158 Chapter 9 / Walking for Health

◗ TABLE 9-1 Velocity, Stride Length, and Cadence of Free Versus Power Walking
Subject Walking Pace Velocity (m/min) Cadence (steps/min) Stride Length (cm)

Subject 1 Free 91.9 123 150.0

Subject 1 Power 148.8 146 200.0
Subject 2 Free 91.5 126 145.7
Subject 2 Power 139.1 169 165.6

flexion in terminal swing and increased extension in length and cadence beyond that of normal controls during
terminal stance; increased knee flexion in early stance; in- fast walking (35).
creased shoulder and elbow arc of motion; and an increase Many people participate in walking for fitness and
in the vertical excursion of the head. Fast walking also re- health. Brisk walking (4 mph, or 107 m/min) is sufficient
sulted in increased stride length and increased cadence to achieve the physiological benefits of exercise (37). How-
(36). Race walkers use different joint motions than that ever, to achieve a higher level of fitness than that afforded
observed for fast walking, and have an increased stride by moderate intensity of brisk walking, people may choose

FIGURE 9-6. Pelvic kinematics for (a) Subject 1 and (b) Subject 2. Solid line is free walking; dashed line
is power walking; light gray lines represent normal comfortable speed walking ± 2 standard deviations.
LHS = left heel strike; RTO = right toe off; RHS = right heel strike; LTO = left toe off. X-axis is percent
of gait cycle. Y-axis is degrees of motion. LHS to LTO represents stance phase. LTO to LHS represents
swing phase. LHS to RTO is loading response (initial double limb support) and RHS to LTO is preswing
(terminal double limb support).
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

Chapter 9 / Walking for Health 159

to participate in vigorous exercise (greater than 6 METS) length. During power walking, Subject 1 increased her
(41). Fitness, or power, walking (5.0 mph, 134 m/min, or velocity by 62% over free walking primarily with an in-
greater) meets the criteria of vigorous exercise (2). creased stride length (33% increase) and a small increase
There is a paucity of information in the literature de- (19%) in cadence. Subject 2 achieved a 52% increase in
scribing the kinematic and kinetic data looking specifically velocity primarily through an increase in her cadence
at the recreational power walker. We studied two women (34%) with only a slight increase (14%) in stride length
(Subjects 1 and 2) who participate in power walking as (Table 9-1).
part of their regular fitness routine. Both achieved walk-
ing speeds sufficient to generate an exercise effect as de-
Kinematics
scribed in the literature (2,41). Both subjects had a free
walking pace faster than that reported by Murray et al. of During power walking, both subjects had the elbows flexed
85.0 m/min, as shown in Table 9-1 (36). from 75 to 100 degrees and had an increased arc of
An increased speed, or velocity, of walking, can be shoulder motion during reciprocal arm swing. During free
obtained by increased cadence and/or increased stride walking, the elbows have an arc of motion from about 20 to

FIGURE 9-7. Hip, knee, and ankle

kinematics for (a) Subject 1 and (b)
Subject 2. Solid line is free walking;
dashed line is power walking; light
gray lines represent normal com-
fortable speed walking ±2 standard
deviations. LHS = left heel strike;
RTO = right toe off; RHS = right
heel strike; LTO = left toe off. X-
axis is percent of gait cycle. Y-axis
is degrees of motion. LHS to LTO
represents stance phase. LTO to
LHS represents swing phase. LHS
to RTO is loading response (initial
double limb support) and RHS to
LTO is preswing (terminal double
limb support).
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

160 Chapter 9 / Walking for Health

FIGURE 9-8. Knee kinetics (1)

and power (2) for (a) Subject 1
and (b) Subject 2. Solid line is
free walking; dashed line is power
walking; light gray lines repre-
sent normal comfortable speed
walking ± 2 standard deviations.
LHS = left heel strike; RTO =
right toe off; RHS = right heel
strike; LTO = left toe off. X-axis is
percent of gait cycle. (1) Kinetics:
Y-axis is the joint moment
(Nm/kg). (2) Power: Y-axis is in
Watts/kg, expressed as positive
for power generation, or negative
for power absorption. LHS to
LTO represents stance phase.
LTO to LHS represents swing
phase. LHS to RTO is loading
response (initial double limb
support) and RHS to LTO is
preswing (terminal double limb
support).

40 degrees of flexion. Subject 1, who used primarily an trunk and pelvis. If one has mechanical pathology in the
increased stride length to achieve power walking velocity, lumbar spine that may be aggravated by increased trunk
had an increase in pelvic motion in all three planes (Fig- and pelvis rotation, it may be best to use increased ca-
ure 9-6a), and increased hip flexion at initial contact, with dence, rather than increased stride length, to achieve walk-
less change in sagittal plane ankle motion than Subject ing speeds of vigorous exercise.
2 (Figure 9-7a). Subject 1 also demonstrated a signifi-
cant increase in trunk rotation that accompanied the in-
crease in pelvic rotation (Figure 9-6a): for the left limb
Kinetics
at initial contact, the pelvis rotated to the left while the
trunk rotated to the right and in terminal stance, the In loading response, both subjects had similar increases
pelvis rotated to the right while the trunk rotated to in knee flexor and hip extensor moments. In late stance
the left. When using cadence as the primary means of (preswing) there were similar increases in knee extensor
achieving an increased velocity for power walking, Sub- and hip flexor moments for both women (Figures 9-8 and
ject 2’s trunk and lower extremity motion only varied by an 9-9). At the ankle, there was an increased generation of
increased anterior pelvic tilt and increased ankle dorsiflex- power in late stance during power walking compared to
ion in swing and early stance (Figures 9-6b and 9-7b). Her comfortable speed walking (Figure 9-10).
peak ankle dorsiflexion position occurred at initial contact These two subjects demonstrated power walking at
with a gradual progression toward plantar flexion during speeds adequate to generate a vigorous exercise effect.
stance (Figure 9-7b). Their techniques of increasing walking speed were differ-
The kinematic data from these two subjects demon- ent in that one used primarily an increased cadence and
strate two different strategies used to achieve increased one used primarily an increase in stride length. The result-
speed of walking. Both subjects had increased elbow flex- ing joint kinematics, kinetics, and power varied depending
ion and increased shoulder range of motion during power on the chosen technique. However, there were similarities
walking. However, when increased cadence was chosen as in some of these variables as well. As participation in power
the primary mode to achieve faster walking speed, there walking grows, especially in the older population, it is im-
were minimal changes in the kinematics. When faster portant to continue to expand our understanding of the
walking speed was achieved by increasing the stride length, mechanics of high speed walking through research to pro-
there were significant changes in the kinematics of the mote safe techniques.
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

Chapter 9 / Walking for Health 161

FIGURE 9-9. Hip kinetics (1)

and power (2) for (a) Subject 1
and (b) Subject 2. Solid line is
free walking; dashed line is power
walking; light gray lines represent
normal comfortable speed walk-
ing ± 2 standard deviations. LHS
= left heel strike; RTO = right toe
off; RHS = right heel strike; LTO
= left toe off. X-axis is percent
of gait cycle. (1) Kinetics: Y-axis
is the joint moment (Nm/kg). (2)
Power: Y-axis is in Watts/kg, ex-
pressed as positive for power gen-
eration, or negative for power ab-
sorption. LHS to LTO represents
stance phase. LTO to LHS repre-
sents swing phase. LHS to RTO
is loading response (initial dou-
ble limb support) and RHS to LTO
is preswing (terminal double limb
support).

FIGURE 9-10. Ankle kinetics (1)

and power (2) for (a) Subject 1
and (b) Subject 2. Solid line is
free walking; dashed line is power
walking; light gray lines represent
normal comfortable speed walk-
ing ± 2 standard deviations. LHS
= left heel strike; RTO = right toe
off; RHS = right heel strike; LTO
= left toe off. X-axis is percent
of gait cycle. (1) Kinetics: Y-axis
is the joint moment (Nm/kg). (2)
Power: Y-axis is in Watts/kg, ex-
pressed as positive for power gen-
eration, or negative for power ab-
sorption. LHS to LTO represents
stance phase. LTO to LHS repre-
sents swing phase. LHS to RTO
is loading response (initial dou-
ble limb support) and RHS to LTO
is preswing (terminal double limb
support).
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

162 Chapter 9 / Walking for Health

Conclusion program in older adults with knee osteoarthritis: The Fitness Arthri-
tis and Seniors Trial (FAST). JAMA 1997;277:25–31.
That frequent walking has the potential to provide a num- 13. Eyler AA, Brownson RC, Bacak SJ, Housemann RA. The epidemi-
ber of significant health benefits throughout a person’s ology of walking for physical activity in the United States. Med Sci
Sports Ex 2003:35:1529–1536.
lifespan has been scientifically documented over the past 14. Feskanic D, Willett W, Colditz G. Walking and leisure-time physical
50 years. These benefits include the reduction in the in- activity and risk of hip fracture in postmenopausal women. JAMA
cidence of coronary heart disease, hypertension, stroke, 2002;288:2300–2306.
15. Gardner AW, Poehlman ET. Exercise rehabilitation programs for the
obesity, type II diabetes, osteoporosis, and several site- treatment of claudication pain: a meta-analysis. JAMA 1995;274:975–
specific cancers. Walking is also effective in the treatment 980.
or rehabilitation process for many of these and other dis- 16. Hakim AA, Curb D, Petrovich H, et al. Effects of walking on coronary
heart disease in elderly men: The Honolulu Heart Program. Circula-
orders, including osteoarthritis. In older persons, walking tion 1999;100:9–13.
contributes to enhanced physical and cognitive function- 17. Hatano Y. Use of the pedometer for promoting daily walking activity.
ing and a lower incidence of dementia. Many of the stud- Jap C Health Physical Ed Rec 1993;29:4–8.
18. Hu FB, Sigal RJ, Rich-Edwards JW, et al. Walking compared with
ies report significant health benefits from brisk walking vigorous physical activity and risk of type 2 diabetes in women. JAMA
(3+ mph) for two or more hours per week, with greater 1999;282:1433–1439.
benefits resulting from faster walking or longer durations. 19. Hu FB, Stampfer MJ, Colditz GA, et al. Physical activity and risk of
stroke in women. JAMA 2000;283:2961–2967.
There are numerous reasons for promoting walking to im- 20. Jakicic JM, Marcus BH, Gallagher KJ, Napolitano M, Lang W.
prove the health of the population over a number of other Effect of exercise duration and intensity on weight loss in over-
activities. Compared to jogging or running, walking pro- weight sedentary women: a randomized trial. JAMA 2003;290:1323–
1330.
duces substantially fewer musculoskeletal injuries (5) and 21. Koh-Banerjee P, Wang Y, Hu FB, Spiegelman D, Willett WC, Rimm
by building walking into each person’s lifestyle the public EB. Changes in body weight and body fat distribution as risk factors
health goals for physical activity can be achieved (6). Walk- for clinical diabetes in US men. Am J Epidemiol 2004;159:1150–1159.
22. Kraus WE, Houmard JA, Duscha BD, et al. Effects of the amount
ing also provides an environmental friendly form of trans- and intensity of exercise on plasma lipoproteins. N Engl J Med
portation and can be done alone or as a socially oriented 2002;347:1483–1492.
group hike or trek. Most people possess the skill and ability 23. Krotkiewski M, Lonnroth P, Mandroukas K, et al. The effects of phys-
ical training on insulin secretion and effectiveness and on glucose
to walk daily and special equipment requirements, other metabolism in obesity and Type-2 (noninsulin dependent) diabetes
than good walking shoes and possibly a walking stick, are mellitus. Diabetologia 1985;28:881–890.
minimal. 24. LaCroix AZ, Levielle SG, Hecht JA, Grothaus LC, Wagner EH.
Does walking decrease the risk of cardiovascular disease hospital-
izations and death in older adults? J Am Geriatr Soc 1996;44:113–
120.
REFERENCES 25. Lakdawalla D, Philipson T. The growth of obesity and technologi-
cal change: a theoretical and empirical examination. NBER Working
1. Abbott RD, White LR, Ross GW, Masaki KH, Curb JD, Petro- Paper No 8946. National Bureau of Economic Research, Cambridge,
vitch H. Walking and dementia in physically capable men. JAMA MA; 2002.
2004;292:1447–1453. 26. Lee CD, Folsom AR, Blair SN. Physical activity and stroke risk: A
2. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical meta-analysis. Stroke 2003;34:2475–2482.
activities: an update of activity codes and MET intensities. Med Sci 27. Leenders NYJM, Sherman WM, Nagaraja HN, Kien CL. Evaluation
Sports Exerc 2000;32:S498–S516. of methods to assess physical activity in free-living conditions. Med
3. Albert MS, Jones K, Savage CR, et al. Predictors of cognitive change Sci Sports Ex 2001;33:1233–1240.
in older persons. MacArthur Studies of Successful Aging. Psychol 28. Lindsay J, Laurin D, Verreault R, Helliwell B, Hill GB, McDowell I.
Aging 1995;10:578–589. Risk factors for Alzheimer’s disease: a prospective analysis from the
4. Bassett DR, Schneider PL, Hunington GE. Physical activity in an Old Canadian Study of Health and Aging. Am J Epidemiol 2002;156:445–
Order Amish community. Med Sci Ex Sports 2004;36:79–85. 453.
5. Colbert LH, Hootman JM, Macera CA. Physical activity-related in- 29. Malina RM. Darwinian fitness, physical fitness and physical activity.
juries in walkers and runners in the Aerobics Center Longitudinal In: Mascie-Taylor CGN, Lasker GW, eds. Applications of Biological An-
Study. Clin J Sports Med 2000;10:259–263. thropology. Cambridge: Cambridge University Press; 1991:143–184.
6. Coleman KJ, Raynor HR, Mueller DM, Cerny FJ, Dorn JM, Epstein 30. Manson JE, Greenland P, LaCroix AZ, et al. Walking compared to vig-
LH. Providing sedentary adults with choices for meeting their walk- orous exercise for the prevention of cardiovascular events in women.
ing goals. Prevent Med 1999;28:510–519. N Engl J Med 2002;347:716–725.
7. Cordain L, Gotshall RW, Eaton SB, Eaton SB 3rd . Physical activity, 31. Manson JE, Hu FB, Rich-Edwards JW, et al. A prospective study
energy expenditure, and fitness: an evolutionary perspective. Int J of walking as compared with vigorous exercise in the prevention
Sports Med 1998;19:328–335. of coronary heart disease in women. N Engl J Med 1999;341:650–
8. Criqui MH. Peripheral arterial disease—epidemiological aspects. 658.
Vasc Med 2001;6:S3–S7. 32. Mock V, Dow KH, Mears CJ, et al. Effects of exercise on fatigue, phys-
9. Darwin C. The Origin of Species by Means of Natural Selection, or the ical functioning, and emotional distress during radiation therapy for
Preservation of Favored Races in the Struggle for Life. Watts and Co., breast cancer. Oncol Nurs Forum 1997;24:991–2000.
London; 1859:546. 33. Mokdad AH, Bowman B, Ford E, Vinicor F, Marks JS, Koplan JP.
10. Duncan JJ, Gordon NF, Scott CB. Women walking for health and The continuing epidemic of obesity and diabetes in the United States.
fitness: how much is enough? JAMA 1991;266:295–299. JAMA 2001;286:1195–1200.
11. Engelgau MM, Gelss LS, Saaddine JB, et al. The evolving dia- 34. Morris JN, Heady JA, Raffle PAB, et al. Coronary heart disease and
betes burden in the United States. Ann Intern Med 2004;140:945– physical activity of work. Lancet 1953;6795:1053–1057.
950. 35. Murray MP, Guten GN, Mollinger LA, Gardner GM. Kinematic and
12. Ettinger WH, Burns R, Messier SP, et al. A randomized trial compar- electromyography patterns of Olympic race walkers. Am J Sport Med
ing aerobic exercise and resistance exercise with a health education 1983;11:68–74.
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

Chapter 9 / Walking for Health 163

36. Murray MP, Mollinger LA, Gardner GM, Sepic SB. Kinematic and exercise on body weight, body composition, and measures of central
EMG patterns during slow, free, and fast walking. J Orthop Res 1984; obesity. Arch Intern Med 2004;164:31–39.
2:272–280. 49. Thune I, Furberg A-S. Physical activity and cancer risk: dose re-
37. Murtagh EM, Boreham CAG, Murphy MH. Speed and exercise inten- sponse and cancer, all sites and site-specific. Med Sci Sports Exerc
sity of recreational walkers. Prevent Med 2002;35:397–400. 2001;33:S530–S550.
38. Newman AB. Peripheral arterial disease: insights from population 50. Tomoshige H, Tsumura K, Suematsu C, Okada K, Fujii S, Endo G.
studies of older adults. J Am Geriatr Soc 2000;48:1157–1162. Walking to work and the risk of hypertension in men: The Osaka
39. Pan X-R, Lei G-W, Hu Y-H, et al. Effects of diet and exercise Health Survey. Ann Intern Med 1999;130:21–26.
in preventing NIDDM in people with impaired glucose tolerance. 51. Tudor-Locke C, Bassett DR. How many steps/day are enough: prelim-
The Da Quing IGT and Diabetes Study. Diabetes Care 1997;20:537– inary pedometer indices for public health. Sports Med 2004;34:1–5.
544. 52. Tudor-Locke C, Ham SA, Macera C, Ainsworth BE, Kirtland KA, Kin-
40. Panush RS, Inzinna JD. Recreational activities and degenerative joint sey CD. Descriptive epidemiology of pedometer-determined physical
disease. Sports Med 1994;17:1–5. activity. Med Sci Sports Ex 2004;36:1567–1573.
41. Pate RR, Pratt M, Blair SN, Haskell WL, et al. Physical activity and 53. Van Barr ME, Assendelft WJJ, Dekker J, Oostendorp RAB, Bulsma
public health: a recommendation from the Centers for Disease Con- JWJ. Effectiveness of exercise therapy in patients with osteoarthritis
trol and Prevention and the American College of Sports Medicine. of the hip or knee. A systematic review of randomized clinical trials.
JAMA 1995;273:402–407. Arthritis Review 1999;42:1361–1369.
42. Pollock ML, Miller HS, Janeway R, Linnerud AC, Robertson B, 54. Vingard E, Alfredsson L, Goldie I, Hogstedt C. Occupation and os-
Valentino R. Effects of walking on body composition and cardio- teoarthrosis of the hip and knee: a registry-based cohort study. Int J
vascular function in middle-aged men. J Appl Physiol 1971;30:126– Epidemiol 1991;20:1025–1031.
130. 55. Vuori I. Dose-response of physical activity and low back pain, os-
43. Rodnick KJ, Haskell WL, Swislocki ALM, Foley JE, Reaven GM. Im- teoarthritis and osteoporosis. Med Sci Sports Ex 2001;33:S551–S586.
proved insulin action in muscle, liver and adipose tissue in physically 56. Wannamethee SG, Shaper AG, Alberti GMM. Physical activity and
trained humans. Am J Physiol 1987;253:E489–E495. metabolic factors, and the incidence of coronary heart disease and
44. Rogers MA, Yamamoto C, King DS, Hagberg JM, Ehansi AA, Holloszy type 2 diabetes. Arch Intern Med 2000;160:2108–2116.
JO. Improvement in glucose tolerance after one-week of exercise in 57. Waters RL, Lunsford BR, Perry J, Byrd R. Energy-speed relationship
patients with mild NIDDM. Diabetes Care 1988;11:613–618. of walking: standard tables. J Orthop Res 1988;6:215–222.
45. Schneider PL, Crouter SE, Lukajic O, Bassett JR. Accuracy and reli- 58. Welk GJ, Differding JA, Thompson RW, et al. The utility of the Digi-
ability of 10 pedometers for measuring steps over a 400-meter walk. walker step counter to assess daily physical activity patterns. Med Sci
Med Sci Sports Ex 2003;35:1779–1784. Sports Ex 2000;32:S481–S488.
46. Seidell JC. Obesity, insulin resistance and diabetes—a worldwide epi- 59. Weuve J, Kang JH, Manson JE, Breteler MM, Ware JH, Grodstein F.
demic. BMJ 2000;83:S5–S8. Physical activity, including walking, and cognitive function in older
47. Simpson ME, Serdula M, Galuska DA, et al. Walking trends among women. JAMA 2004; 292:1454–1461.
US adults: the Behavioral Risk Factor Surveillance System, 1987– 60. Yaffe K, Barnes D, Nevitt M, Lui L-Y, Covinsky K. A prospective study
2000. Am J Prev Med 2003;25:95–100. of physical activity and cognitive decline in elderly women. Arch In-
48. Slentz CA, Duscha BD, Johnson JL, et al. Effects of the amount of tern Med 2001;161:1703–1708.
P1: NDL/OVY P2: NDL/OVY QC: NDL/OVY T1: NDL
GRBT092-09 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 11:19

164
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

• C h a p t e r 10
•
◗ Gait Analysis: Clinical Decision Making
Janet M. Adams and Jacquelin Perry

THE DEVELOPMENT OF GAIT another. Disruption of normal function by disease, trauma,

ANALYSIS degeneration, fatigue, or pain produces functional limita-
tions. Potential for normal function varies with the severity
Clinical use of gait analysis began as an observational skill of the primary impairment and the ability to utilize resid-
when physicians, therapists, and prosthetists defined nor- ual function for substitute action. Compensatory strategies
mal and abnormal limb motion and common deviations depend on residual muscle strength, motor control, joint
displayed by patients. Because the hip, knee, and ankle mobility, and sensory capabilities. Intervention depends
move at different rates and through asynchronous arcs, a on the ability of the clinician to identify and treat the im-
systematic approach was devised corresponding to the pat- pairment contributing to the functional loss, not the more
tern of limb motion. Critical events (i.e., heel strike) were conspicuous compensatory action.
selected to differentiate intervals and define phases. Clini-
cians from the Pathokinesiology and Physical Therapy De-
partments at Rancho Los Amigos National Rehabilitation Gait Cycle Organization
Center developed generic terminology for both normal and & Interpretation
abnormal gait. The functional subdivisions of the gait cycle
were expanded to eight phases, each differentiating a ma- A stride is a complete cycle of limb motion that is divided
jor kinematic event. A “Full Body: Observational Gait Anal- into two periods: stance and swing. Stance, the period of
ysis Form” was developed which identified 35 commonly foot-floor contact, is subdivided into 5 phases: initial con-
observed gait deviations and their phasic occurrence (15). tact, loading response, mid-stance, terminal stance, and
Using this form as a guide, clinicians established a proto- preswing. Stance comprises 62% of the gait cycle (GC).
col for analysis that included observing the patient walk, The remaining 38% of the gait cycle constitutes the swing
identifying excessive or inadequate motion, and when it period, initiated as the foot leaves the support surface. It is
occurs in the gait cycle. Reliability and validity of this tool equally divided into three phases: initial swing, mid-swing,
has been affirmed by two separate studies (4,7). The obser- and terminal swing.
vational analysis is correlated with clinical assessments of Sequential combinations of these phases accomplish
strength, range of motion (ROM), spasticity, motor con- three tasks: weight acceptance, single limb support, and
trol, and sensation to determine the probable cause(s) of limb advancement. Initial contact and loading response
the deviation and its contribution to functional deficits. constitute the period of weight acceptance (0% to 12%
Combined with an observational analysis, instru- GC), an interval of double limb support when both feet
mented analysis provides a more precise and quantitative are in contact with the floor for the transfer of body
means of identifying deviations and probable cause. Or- weight from one limb to the other. Mid-stance (13% to
thopedists, treating children and adults with upper motor 30% gait cycle) and terminal stance (31% to 49% gait cy-
neuron lesions increasingly rely on instrumented analysis cle) are involved in the task of single limb support when
for presurgical decision making (6,14,17,22,24,25). the weight of the body is fully supported by the stance
limb. The task of swing limb advancement is initiated dur-
ing preswing. Preswing is the period of terminal double
FUNCTIONAL PERSPECTIVE limb support when weight is again transferred to the con-
tralateral limb. Swing limb advancement includes preswing
Walking requires antigravity muscular support, joint mo- (50% to 62% GC), initial swing (63% to 74% GC), midswing
bility to allow smooth progression, and adequate motor (75% to 87% GC), and terminal swing (88% to 100% GC)
control for the transition of body weight from one limb to (17,18).

165
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

166 Chapter 10 / Gait Analysis: Clinical Decision Making

Task 1: Weight Acceptance rocker to preserve momentum and advance the limb over
a stationary foot. Controlled ankle dorsiflexion is critical
The purpose of the two phases involved in weight accep-
to progression. The larger soleus muscle is the predom-
tance is to protect the joints from impact injury and pre-
inant decelerating force, eccentrically contracting along
serve progression. The critical functions are establishing a
with the gastrocnemius and perimalleolar (posterior tib-
heel rocker and shock absorption.
ialis [PT], flexor digitorum longus [FDL], flexor hallucis
Phase 1: Initial contact is a brief event that initiates
longus [FHL], peroneus longus [PL], and peroneus brevis
stance phase. Foot position and contact point determine
[PB]) muscles (17,21,23). By the end of mid-stance, the
the availability of the heel, or first, rocker. Heel contact cre-
ankle achieves 5 degrees of dorsiflexion and the center of
ates a moment that initiates plantar flexion and subtalar
pressure is moving toward the forefoot while the hip and
eversion resisted by eccentric action of the pretibial (ante-
knee are passively extending.
rior tibialis [AT], extensor digitorum longus [EDL], exten-
Errors in motor control that cause premature, excessive
sor hallucis longus [EHL]) muscles and posterior tibialis
or spastic plantar flexor muscle action and contractures
(17,21,23).
frequently produce premature lifting of the heel, relocat-
Abnormal foot contact patterns may be caused by
ing the pivotal axis to the forefoot to preserve momentum
pretibial muscle weakness, inadequate knee extension or
(14). When patients lack momentum to progress onto the
plantar flexion contractures. The normal event of heel
forefoot, knee hyperextension or genu recurvatum results.
strike may be replaced by a low or abbreviated heel strike,
Plantar flexor weakness (<4/5) causes unrestrained
foot-flat, or forefoot contact. All three reduce the pivotal
dorsiflexion. The tibia accelerates forward, inducing knee
action that the heel rocker provides, resulting in decreased
flexion that increases the demand on the quadriceps. Even
momentum, reduced stride length, and decreased velocity
when the quadriceps are graded good or normal in strength
(17).
(4 or 5), they no longer have a stable base over which
Phase 2. Loading response is characterized by an in-
to exert force and full knee extension cannot occur. The
crease in knee flexion as weight is transferred onto the
dorsiflexion moment increases and knee flexion persists,
limb. The eccentric action of the quadriceps limits knee
requiring continuous quadriceps action. Once maximum
flexion to 15 degrees. The knee flexes to provide shock
available dorsiflexion range is reached, the heel lifts off the
absorption to protect proximal joints. Deviations may be
support surface (premature heel rise). Forefoot contact re-
either excessive or absent knee motion. Weak quadriceps
sults but is coupled with excess dorsiflexion and knee flex-
(<4/5 in manual muscle testing [MMT]) result in the avoid-
ion. This pattern is often mistakenly referred to as equinus
ance of knee flexion; the patient compensates by position-
since the heel is lifted, however the ankle is dorsiflexed in-
ing body weight anterior to the knee joint axis creating
stead of plantar flexed.
an extensor moment, thereby eliminating the demand for
Just as primary ankle impairments cause knee devia-
quadriceps action. Plantar flexion contractures, spasticity,
tions, knee impairments produce changes in ankle kine-
or premature activation of the gastrocnemius may prevent
matics that alter the foot contact pattern. For example,
knee flexion during loading (13,14,20). This phase also re-
when a patient presents with a knee flexion contracture
quires peak hip extensor and abductor muscle action to in-
and strong calf muscles, momentum will be preserved
crease weight-bearing stability of the limbs and pelvis (9).
by voluntarily plantar flexing, assuming a forefoot con-
tact pattern. This increases the relative length of the in-
volved limb and enables progression over a stationary
Task 2: Single Limb Support
foot. Similarly, hip flexion contractures induce secondary
Contralateral toe-off places weight exclusively on the knee flexion, requiring modification of the ankle posi-
stance limb, defining the third and fourth phases of tion and foot-floor contact pattern to preserve momen-
the gait cycle, mid and terminal stance. During this task, tum.
the body advances over a stationary foot while maintaining Phase 4: Terminal stance progression relies on hip ex-
limb stability. Critical functions include stabilization of the tension mobility and controlled dorsiflexion to progress
hip in the sagittal and coronal planes to prevent excessive the body over a stationary foot. The forefoot becomes the
pelvic & trunk motion, as well as unrestrained ankle dor- axis about which the weight-bearing limb pivots; it is called
siflexion to allow progression over a stationary foot. The the forefoot, or third, rocker. Heel rise occurs, reducing the
second (ankle) and third (forefoot) rockers preserve mo- base of support to the metatarsal heads. Eccentric action
mentum. The rate of tibial advancement is restrained by of the soleus and gastrocnemius peak at 45% GC and 40%
eccentric action of the plantar flexors (17,21,23). The criti- GC, respectively, coupled with posterior tibialis and per-
cal factors include adequate strength of the plantar flexors oneals activation which stabilizes the subtalar and trans-
and mobility of the ankle, knee, and hip. verse tarsal joints for forefoot weight bearing (17,21,23).
Phase 3: Mid-stance is a period of flat foot support The ankle dorsiflexes 10 degrees while the center of pres-
and relies on the pivotal mobility of the ankle, or second, sure advances over the contour of the metatarsal heads.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

Chapter 10 / Gait Analysis: Clinical Decision Making 167

Inadequate plantar flexion strength allows unrestrained ual muscle test [MMT] <3/5) causes “foot drop” requir-
tibial advancement resulting in a flexed knee during stance ing increased hip and knee flexion to achieve foot clear-
and a possible loss of heel rise. If quadriceps strength is ance (steppage gait). Weak contralateral abductors, spas-
inadequate to support a flexed knee (<4/5), knee exten- tic quadriceps, or weak dorsiflexors all produce a relative
sion persists. Excessive plantar flexion from contracture, lengthening of the swing limb requiring compensatory ac-
spasticity or disordered motor control accentuates the ex- tion for clearance. In terminal swing, step length may be
tensor moment causing hyperextension at the knee. The compromised by inadequate knee extension, tight or spas-
center of pressure cannot progress to the forefoot, which tic hamstrings or disordered motor control obstructing
requires premature contralateral initial contact and leads knee extension with hip flexion.
to a reduced step length.
Temporal Gait Characteristics
Task 3: Swing Limb Advancement Walking speed for 20 to 60 year old adults at the normal
self selected pace averages 80 m per minute (men 82 m
This task includes the final phase of stance (preswing) and per minute and women 78 m per minute) +/− 10 m per
all three swing phases. The initial requirement is rapid minute (1 SD) (26). At this speed, the duration of stance
transformation of a “rigid” extended limb into a flexible is 62% of the gait cycle and swing is 38%. Double limb
mobile segment. The critical events are limb advancement support averages 24% GC with the remaining 38% as sin-
and foot clearance. gle limb support. Wide variability in stride characteristics
Phase 5: Pre-swing is the final phase of stance (50% to exists within the normal population depending on speed,
62% gait cycle). Contralateral foot contact initiates double age, gender, and geographic location (2,5,11,26).
limb support phase. The rapid transfer of weight abruptly Stance as a percentage of the gait cycle increases as
unloads the stance limb and two events occur: 20 degrees walking speed becomes slower (5). Among persons with a
of plantar flexion and 40 degrees of knee flexion. The en- disability, a correlation between severity and self-selected
ergy for this action is the result of the abrupt release, velocity has been identified. Normal 20 to 60 year old indi-
in terminal stance, of the soleus and gastrocnemius ec- viduals walking at slow speeds average 43 m per minute;
centric action (12). At 54% of the gait cycle, there is a consequently, disabled individuals with slower velocities
prominent spike of plantar flexor power that generates must exert maximal efforts to accompany peers (26).
the dynamic elastic response characterized as “push-off”
(17,21,23). The adductor longus contributes to the arc of
CLINICAL TESTING TO IDENTIFY
hip flexion for limb advancement.
PRIMARY IMPAIRMENTS
Spasticity of the vasti or rectus femoris may obstruct
knee flexion requiring compensation at the hip and/or
Strength Assessment
pelvis to achieve foot clearance (8,20). Reduced knee flex-
ion may also result from joint effusion, pain, or the sec- For patients with musculoskeletal pathology, the standard
ondary effect of reduced velocity. manual muscle testing (MMT) procedures are appropriate
Phases 6, 7 & 8: Initial, Mid & Terminal Swing Dur- as they have the ability to isolate movement of one joint
ing initial swing the hip flexes 20 degrees, the knee flexes from that of adjacent joints. Clinical interpretation of the
another 20 degrees (total arc of 60◦ ), and the ankle be- MMT grades 0 to 5 represent absent to normal strength;
gins dorsiflexing to accomplish foot clearance. Momentum however, it often overestimates the patient’s capabilities.
from hip flexion is the primary determinant of knee flexion Correlation of MMT grades and torque measurements of
during this interval. Mid-swing continues the task of limb individuals with polio and age-matched controls defined
advancement and foot clearance. Hip flexion (25◦ ) and a grade 4/5 (“good”) was equivalent to 40% of true nor-
ankle dorsiflexion to neutral are the critical events while mal strength, grade 3/5 (“fair”) was 20% of normal and
advancing the tibia past its vertical alignment. Terminal grade 2/5 (“poor”) was 10% of normal (1). Thus, a grade
swing completes limb advancement with full knee exten- 4/5 (good) actually represents a significant loss of strength
sion, while keeping the hip at 25 degrees of flexion and the (as much as 60%) despite the ability of the examiner to
ankle dorsiflexed to neutral. Hip flexion is terminated by offer substantial resistance. The weakness may introduce
the hamstring muscles while momentum and quadriceps subtle deviations as a result of muscle fatigue that do not
action extend the knee to neutral. In preparation for ini- appear during the standard gait analysis (6–10 m of walk-
tial contact, the gluteus maximus and adductor magnus ing) but limit community mobility.
activate (9).
The major deviation during swing is toe drag. Limb
Mobility Assessment
clearance is impaired if there is limited hip flexion, knee
flexion, or dorsiflexion. The primary cause may be inade- Contractures restrict passive soft tissue mobility, requiring
quate knee flexion in pre-swing. Weak dorsiflexors (man- postural adaptations to preserve progression. “Rigid” and
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

168 Chapter 10 / Gait Analysis: Clinical Decision Making

“elastic” or “yielding” contractures should be differenti- cient during midswing to lift the foot against gravity for
ated as they have different functional implications. Elastic foot clearance. Thus, the patient will exhibit their most
contractures yield under the force of body weight but re- conspicuous gait deviation during loading where a grade
sist full range during a manual stretch using minimal force of 4/5 is needed to eccentrically control the rate of plantar
(2 fingers) (17). In contrast, rigid contractures do not flexion following heel contact. A fair grade (3/5) introduces
stretch significantly regardless of the force exerted by body a “foot-slap” (the foot was sufficiently dorsiflexed in swing
weight. This differentiation is especially critical at the an- for heel contact but experiences uncontrolled rapid plan-
kle. Yielding contractures can be a valuable supplement tar flexion during loading). An alternate strategy is initial
for weak plantar flexors, acting as an internal orthosis, contact with a “low” or “abbreviated” heel contact (the an-
providing tibial stability in the absence of sufficient mus- gle to the floor ≤5◦ ). This effectively eliminates the heel
cle strength. In contrast, rigid contractures are usually ob- rocker, reducing the knee flexion moment during loading.
structive and require serial casting or surgical lengthening. A severe foot drop, 30 degrees to 50 degrees during swing,
Differentiation between the two types of contractures, es- requires conspicuous lifting of the limb in swing for clear-
pecially at the ankle, is critical for deciding on an optimal ance. As the limb is lowered for stance, initial contact is
intervention strategy. made with the forefoot and the limb rapidly drops to create
foot-flat. Paralysis of anterior tibialis with 4 or 5 grades of
the extensor digitorum longus provides sufficient strength
The Significance of Pathology:
for foot clearance in swing; however, the medial rays of the
Musculoskeletal versus Neuromuscular
foot drop resulting in dorsiflexion with eversion.
An essential first step in the clinical application of gait anal- WEAK CALF MUSCLES: The soleus and gastrocnemius ac-
ysis is identification of the patient’s medical diagnosis as count for 91% of the force required to eccentrically con-
a means of classifying the primary impairments. Impair- trol the dorsiflexing ankle during mid and terminal stance.
ments affecting gait primarily fall into two broad practice The perimalleolar group (FDL, FHL, PT, PL, PB) pro-
patterns, neuromuscular and musculoskeletal. Patients vides the remaining 9% as they control the subtalar and
with neuromuscular impairments primarily have disor- metatarsal joints (17,21,23). The severity of gastrocsoleus
dered motor control resulting from upper motor neuron muscle weakness is evident in the timing of heel rise and
lesions (3). This significantly limits their ability to substi- knee kinematics. Grade 3/5 or “fair” (20% of normal) plan-
tute for impaired function. Musculoskeletal impairments tar flexion strength is required to accomplish one heel rise
limit functional performance caused by pathological alter- but it is not adequate to restrain the tibia’s forward transla-
ation in the peripheral structures. These patients generally tion, resulting in knee instability (10). Calf strength graded
have excellent substitutive capability. 3+/5 is sufficient to restrain the tibia in mid-stance but
heel rise is delayed. Grade 4/5 will provide single stance
heel rise for short distances, while unlimited community
Primary Impairments in Patients with
ambulation requires a 5/5.
Musculoskeletal Pathology
Uncontrolled advancement of tibia during mid and ter-
Patients with gait abnormalities often exhibit asymmetry minal stance causes excessive knee flexion. This increases
with one limb more affected than the other. If the impair- the demand on the quadriceps at a time when they are
ment is moderate, that limb will be the focus of disabil- normally inactive. Quadriceps demand can be reduced by
ity. If, however, the impairment is severe the “strong” limb maintaining the center of pressure close to the calcaneus.
may be the symptomatic site of overuse. Thus, at least a Knee flexion can therefore be avoided by reducing the an-
preliminary assessment of both limbs should precede in- kle dorsiflexion moment. Tibial control may be restored
strumented gait analysis. Even a MMT grade 2/5 can be by an articulated ankle foot orthosis (AFO) with a dorsi-
beneficial for functional tasks despite the absence of anti- flexion stop. The orthosis should allow free plantar flexion
gravity capabilities. The initial gait modification is a re- to protect the quadriceps from an exaggerated heel rocker
duced gait velocity. during loading which would increase the flexor moment
WEAK DORSIFLEXORS: Strength requirements of the dor- at the knee. If plantar flexion is blocked the patient must
siflexors (AT, EDL and EHL) vary throughout the gait cycle. have sufficient quadriceps strength to meet the demand.
When graded <3/5, drop foot is most evident in midswing. WEAK QUADRICEPS. The normal pattern and rate of walk-
Compensatory hip and knee flexion provide foot clearance ing requires quadriceps with grade 4+ to 5 strength. Grade
(steppage gait). The heel (first) rocker at initial contact and 4/5 (good) is the minimum strength required to resist the
loading is eliminated resulting in forefoot or flatfoot con- knee flexion moment created by the heel rocker. Patients
tact. As only foot weight need be supported, a posterior with quadriceps graded less than good (<4/5) avoid knee
leaf-spring dorsiflexion assist orthosis is sufficient to sup- flexion during loading but have an otherwise normal gait
port the foot during swing and to preserve the heel rocker. as long as the hip extensors and plantar flexors are at least
In contrast, grade 3/5 strength (20% of normal) is suffi- a grade 4/5. In patients graded <3/5, knee stability may
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

Chapter 10 / Gait Analysis: Clinical Decision Making 169

be maintained by increased activity of the hip extensors to Primary Impairments in Patients

control the femur and premature activation of the ankle with Neuromuscular Dysfunction
plantar flexors to retract the tibia. At initial contact, slight
Upper motor neuron lesions resulting from cerebrovascu-
(10◦ ) plantar flexion avoids the heel rocker reaction as the
lar accidents, cerebral palsy, or traumatic brain injuries or
limb is loaded. The knee extends to its passive limit of at
trauma can lead to disruptions in neural control mecha-
least 5◦ hyperextension for stability in the absence of an
nisms interfering with the patients’ ability to walk (3). Im-
orthotic restricting recurvatum. If the hip extensor or calf
pairments may include spasticity, impaired or inadequate
muscles are less than 3/5 and knee hyperextension range is
motor control, the emergence of primitive locomotor pat-
inadequate, a knee-ankle foot orthosis (KAFO) is optimal.
terns, muscle weakness, restricted joint mobility, impaired
A locked knee joint is customary but a freely mobile, pos-
balance and sensation, and may include cognitive deficits.
teriorly off-set orthotic joint is indicated for patients with
Muscle weakness may result from loss of selective con-
severe recurvatum (>20◦ ). The off-set joint is designed to
trol, dependence on primitive control patterns, or lack of
enhance the extensor moment and limit recurvatum while
normal somatosensory input. Joint mobility may be re-
providing flexion for swing limb clearance.
stricted by antagonistic muscle activation and spasticity, as
WEAK HIP EXTENSORS; Four muscles are involved in
well as contracture. Cognitive and somatosensory deficits
hip extension; gluteus maximus and adductor magnus are
further limit the patients substitutive capabilities. Signifi-
the primary extensors, and the larger hamstring muscles
cant functional loss is common; patients typically exhibit
(semimembranosus and biceps femoris, long head) serve
varying degrees of impaired or inadequate selective control
as a supplementary force. When the extensors are <3/5,
with difficulty modifying the rate, direction, duration, and
a backward trunk lean preserves the standing posture by
intensity of muscle activity. Within each disability, there
positioning the line of gravity posterior to the hip joint
is a significant range of deficits, from mild to severe. All
axis. Patients immediately assume this posture after initial
but the smallest lesions result in poor substitutive actions
contact. Quadriceps demand is correspondingly increased
since the capacity to voluntarily modify the motions or
by the alignment of the ground reaction force vector pos-
postures of the paretic limb is dependent on selective mo-
terior to the knee axis. The hamstrings may accentuate
tor control. Substitutive mechanisms are further limited
this demand. In many patients, the posterior lean is cou-
when proprioception is impaired. The feedback needed to
pled with excessive lordosis, as a hip flexion contracture is
define limb position and balance to guide the compen-
common.
satory action may be limited or absent. This limits the
WEAK HIP ABDUCTORS: Four muscles contribute to hip
repertoire of movement strategies available to effectively
abduction: gluteus medius and minimus, upper gluteus
compensate.
maximus and tensor fasciae latae (9). Lateral trunk lean
over the stance limb is the sign of inadequate hip abductor
strength (<4/5). The abductors counter the large adduc- GAIT ANALYSIS: CLINICAL DECISION
tor moment created by lifting the contralateral limb for MAKING CASE STUDIES
swing. Weakness deprives the patient of the force needed
to hold the pelvis level. Compensation is initiated in load- The following case studies illustrate the use of instru-
ing by shifting the trunk toward the stance limb to reduce mented gait analysis in clinical decision making. Recom-
the opposing adduction moment. With complete abductor mendations include surgical decisions (by a medical doc-
paralysis, the trunk’s center of mass will lie directly over tor [MD]), as well as exercise prescription and orthotic
the hip joint center to neutralize the abductor demand. prescription (by a physical therapist [PT])
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

170 Chapter 10 / Gait Analysis: Clinical Decision Making

CASE 1
Diagnosis (Dx)-Anterior Compartment Syndrome with Drop Foot

This 17-year old male was diagnosed with anterior com-

partment syndrome (left) following open reduction inter-
nal fixation of a tibial fracture 9 months ago at the site of
a benign tumor excised 3 months previously.
Chief Complaint: Difficulty clearing the foot during swing
resulting in loss of balance and fear of tripping limiting
community ambulation, participation in sports, and recre-
ational activities.

Clinical Examination:
PROM: Dorsiflexion with knee flexed: 10 degrees
Dorsiflexion with knee extended: −5 degrees

Strength tests MMT Torque Strength tests MMT Torque

Anterior Tibialis 2+ Gastrocnemius 2
EDL 3+ Soleus 2
EHL 1 PT, FHL, PL, PB 4
Dorsiflexion 10.2% N Plantar flexors 22.4% N

Palpation: Anterior tibialis muscle contraction palpable FIGURE 10-1. Initial contact in 9 degrees of plantar flexion with
a “low” heel strike minimizing the flexion moment at the knee.
without tendon tension Case 1: Drop Foot.
Gait Instrumentation: 3-D motion, stride characteristics,
and kinesiologic electromyography (EMG) during self se-
lected paced walking.

Results
Foot Floor Contact: Low heel strike at initial contact (Fig.
10-1). Loading: Foot flat at 3% GC. Normal heel off in ter-
minal stance.
% Normal
Stride: Velocity: 56.9 m/min 70
Stride Length: 1.135 m 75
Cadence: 100.3 steps/min 93
Single limb support: left 82

Motion:
Ankle: Initial Contact (IC) in 10 degrees of plantar flexion
(PF). Terminal Stance: 5 degrees of dorsiflexion (DF) in
terminal stance (yielding contracture reduces with body
weight).
Swing Phase: 10 degrees of PF throughout swing phase
(Figs. 10-2 and 10-3)
Knee: IC in 5 degrees of flexion increasing to 10 degrees
in loading, achieving neutral in terminal stance.
Hip: IC in 15 degrees of flexion, followed by immediate FIGURE 10-2. Mid-swing ankle dorsiflexion is inadequate; note
extension to neutral, reaching 10 degrees of extension in the slight compensatory increase in hip and knee flexion with
terminal stance. minimal clearance.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

Chapter 10 / Gait Analysis: Clinical Decision Making 171

ANKLE Patient Interpretation:

Dorsiflexion Despite moderate anterior tibialis EMG activity during
40 Normal
manual muscle testing (2+) and torque testing (10%
20 normal), there was no functionally significant anterior tib-
ialis EMG activity during walking. Failure to lift foot may
Degrees

0
be a result of tendon entrapment in the operative scar or
-20 partial denervation. In contrast, the EDL demonstrated
Plantar Flexion 38% MVC throughout swing with a grade of 3+, but lacked
-40 the leverage to effectively dorsiflex the foot to neutral for
foot clearance.
Extensor
Digitorum
Longus Recommendations:
■ Surgical:
■ Transfer the EDL to the dorsum of the foot with the
tendon removed from the retinaculum to improve its
Anterior
Tibialis dorsiflexion leverage. Selective lengthening of the gas-
trocnemius only to 10 degrees of dorsiflexion with
knee extended. The soleus should be left intact to pre-
serve plantar flexion strength.
% Gait Cycle 0 12 50 62 100 ■ Physical Therapy
IC TO IC ■ Strengthen the AT, EDL & EHL to counter disuse at-

FIGURE 10-3. No functionally significant EMG activity of the

rophy, Functional electrical stimulation to partially
anterior tibialis with 38% MVC of the EDL. Despite limitations denervated muscles
in dorsiflexion PROM the ankle reaches 5 degrees of DF in ter- ■ Strengthen the plantar flexors, quadriceps & hip ex-
minal stance (yielding contracture), yet remains plantar flexed tensors
throughout swing. ■ Gait & balance training
■ Orthotic
EMG: (Fig. 10-3) ■ Articulated AFO with DF-stop and a DF-assist un-

Anterior Tibialis: No functionally significant EMG til EDL recovers adequate strength to dorsiflex foot.
activity Post surgically, if active dorsiflexion to neutral is not
EDL: 38% maximum voluntary contraction achieved (>6 mos) a posterior leaf spring is recom-
(MVC) with premature onset mended.
& cessation
Soleus: 41% MVC, normal phasing
Gastrocnemius: 41% MVC, normal phasing
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

172 Chapter 10 / Gait Analysis: Clinical Decision Making

CASE 2
Dx-Post Polio Syndrome with Excess Dorsiflexion in Stance Caused by
Inadequate Plantar Flexor Strength

This 49-year-old woman was diagnosed with post polio

syndrome 6 months ago.
Chief Complaint: Pain, weakness and fatigue in the right
lower extremity limiting community mobility. Complaints
of knee collapse in late stance inducing fear of falling with
reports of instability on uneven terrain limiting commu-
nity ambulation and recreation.
Clinical Examination:
PROM: DF (knee extended) −5◦
Subtalar eversion 3◦
Knee hyperextension 5◦
Straight Leg Raise (SLR) 70◦

Strength Tests MMT Torque (% N)

Plantar flexion 2 6
Knee extension 3− 21
Hip Abductors 4+ 8
Hip Extensors 4+ 28
FIGURE 10-4. Loading response: avoidance of knee flexion
during loading response caused by weakness and fatigue in the
Gait Instrumentation: 3-D motion, stride characteris- quadriceps (torque testing 21% N) leading to tibial instability.
Case 2: Post Polio Syndrome.
tics, and kinesiologic EMG performed during self selected
paced walking.

Results

Foot Floor Contact: IC with the heel, premature contact

of 5th metatarsal (5% GC), heel off delayed until double
limb support (preswing).
Stride: % Normal
Velocity: 30.3 m/min 39
Stride Length: 0.639 m 48
Cadence: 94.7 steps/min 80
Single limb
support: Right 48.3

Motion:
Ankle: IC in 10 degrees of plantar flexion (Fig. 10-4), re-
versing to 20 degrees of dorsiflexion by terminal stance
(Fig. 10-5), Swing phase plantar flexion 20-5 degrees.
Knee: Hyperextension of 5 degrees at initial contact (Fig.
10-4) with increasing flexion of 15 degrees in terminal
stance.
FIGURE 10-5. Terminal stance: excess dorsiflexion caused by
Hip: IC in 5 degrees of flexion and 30 degrees of medial weak calf muscles (torque testing 6% N), lack of heel rise and
rotation continuing throughout stance. persistent knee flexion increasing quadriceps demand.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

Chapter 10 / Gait Analysis: Clinical Decision Making 173

KNEE Patient EMG: All recordings sparse with high amplitude activity
80 Flexion (Figs. 10-6 and 10-7)
Normal
60
Soleus: (16% MVC) slightly prolonged
40 Vastus Lateralis: (26% MVC) premature and
Degrees

prolonged
20
Semimembranosus: (64% MVC) premature and
0 prolonged
Extension Biceps Femoris: (8% MVC) out of phase, active
-20
in terminal stance
Vastus Gluteus Maximus: (8% MVC) low amplitude
Lateralis (premature and prolonged)

Interpretation:
Sparse EMG activity indicates a reduced number of
Semi- motor units available for activation. High amplitude
Membranosus
EMG is characteristic of recovery through motor unit
enlargement as a result of reinnervation by sprouting. A
mild yielding PF contracture assists with tibial control
% Gait Cycle 0 12 50 62 100 in stance, however dorsiflexion is excessive (20◦ ) increas-
IC TO IC
ing the quadriceps demand at the knee. With 21% N
quadriceps strength and 6% N plantar flexion strength,
FIGURE 10-6. Knee hyperextension (5◦ ) in loading with persis-
tent knee flexion throughout stance. Sparse high amplitude EMG knee instability is likely. Small arcs of joint motion and
in the vastus lateralis indicative of large motor units (reinnerva- slow velocity indicate voluntary protection of the few
tion) with decreased activation in terminal stance contributing remaining large motor units. Motor unit enlargement
to knee collapse. indicates more muscle fibers are activated simultaneously
which may contribute to overuse and fatigue.

Recommendations:
■ No surgical recommendations
ANKLE Patient ■ Physical Therapy
Normal Client education and guidance in modification of func-
Dorsiflexion tional activities to avoid overuse and fatigue. (J Perry
40
recommendations for patients diagnosed with Post-
20 Polio) (16,19)
Degrees

0 ■ Avoid activities causing fatigue that persists more

than 10 minutes
-20 ■ Interrupt active day by two rest periods
Plantar Flexion
-40 ■ Stop an activity before pain develops
■ No resistive exercise causing muscle or joint pain
■ Balance training
Soleus ■ Gait training with SPC (single point cane) to protect
and unload overused muscles
■ Orthotics

% Gait Cycle 0 ■ Articulated AFO (lightweight) modified as follows:

12 50 62 100
IC TO IC A dorsiflexion stop at 5 degrees to assist and protect a weak
calf and reduce the flexor moment at the knee throughout
FIGURE 10-7. Ankle progressively dorsiflexes from initial con-
tact in 10 degree of plantar flexion to 20 degrees of dorsiflexion
stance. Allow free plantar flexion to protect the weak quads
in terminal stance and a delay reversal to plantar flexion in late during IC and loading by attenuating the effect of the heel
preswing. In mid-swing the ankle briefly dorsiflexed to neutral rocker at initial contact. The goal is to prevent fatigue and
but drops into plantar flexion in terminal swing. overuse of quadriceps and hip extensors.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

174 Chapter 10 / Gait Analysis: Clinical Decision Making

CASE 3
Dx-Post CVA with Equinovarus Foot

This 36-year-old male had a 2-year history of right hemi-

paresis following a CVA

Chief Complaint: Pain in the forefoot and toes during

weight bearing. Difficulty with foot clearance and impaired
balance with the use of a quad cane limiting household and
community ambulation.

Clinical Examination:
PROM: DF (knee flexed) −10◦
DF (knee extended) −30◦

Strength Testing (in Synergy-no selective control)

Anterior Tibialis Weak
Posterior Tibialis Weak
Soleus & Gastrocnemius Weak

Spasticity
Muscle EMG response
AT Absent
Gastrocnemius Prolonged (>5 sec of clonus)
Soleus Prolonged (>5 sec of clonus)
FDL & FHL Prolonged (>5 sec of clonus)
PL Prolonged (>5 sec) sparse EMG
PB Normal response

Gait Instrumentation: 3-D motion, stride characteris-

tics, and kinesiologic EMG performed during self-selected
paced walking.

Results
Foot Floor Contact: IC with 5th metatarsal (MT)
only, absent 1st MT or heel contact throughout stance FIGURE 10-8. Mid-stance with equinovarus foot position with-
(Fig. 10-8). out heel, 1st MT or toe contact. Case 3: CVA with equinovarus.

Stride: Velocity 13% N, Stride Length 25% N, Cadence

53% N, Single limb support 14% N
EMG: Figures 10-12 and 10-13
Motion: Anterior Tibialis: Sparse activity throughout initial &
Ankle: IC in 20 degrees of PF reducing to 12 degrees of PF mid-swing
in terminal stance (Fig. 10-9). Inversion 25 degrees. Soleus, Gastrocnemius, FHL & FDL: all exhibit prema-
ture onset in terminal swing that continues into terminal
Knee: IC: 12 degrees of flexion decreasing to 5 degrees in
stance.
mid-stance with a peak of 30 degrees in initial swing (Fig.
Posterior Tibialis: delayed onset with sparse & curtailed
10-11).
intensity
Hip: IC: flexed 25 degrees, reducing to 15 degrees in mid- Peroneus Longus: no activity throughout stance
stance increasing to 30 degrees in initial swing. Peroneus Brevis: sparse throughout stance
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

Chapter 10 / Gait Analysis: Clinical Decision Making 175

ANKLE Patient
Normal
Posterior Tibialis
Dorsiflexion
40
20
Degrees

0 Flexor Hallucis
Longus
-20
Plantar Flexion
-40

% Gait Cycle 0 Flexor Digitorum

12 50 62 100 Longus
IC TO IC

FIGURE 10-9. Ankle Motion: Initial contact in 20◦ of plantar

flexion reducing to 12 degrees of PF in terminal stance dropping
to 20◦ throughout swing. Peroneus
Longus
Patient
FOOT
Inversion Normal
40 % Gait Cycle 0 12 50 62 100
IC TO IC
20
Degrees

FIGURE 10-12. Premature EMG activity of the FHL and FDL

0
without significant posterior tibialis activity suggests the toe flex-
-20 ors significantly contribute to forefoot varus.
Eversion
-40

% Gait Cycle 0 12 50 62 100

IC TO IC

FIGURE 10-10. Forefoot Motion: Inversion of 28◦ at initial con-

tact increasing to 40 degrees during mid-stance reducing to 20◦
throughout swing. Peroneus Brevis

Soleus

Patient
KNEE Normal
80 Flexion
Gastrocnemius
60

40
Degrees

20
Anterior Tibialis
0
Extension
-20
% Gait Cycle 0 12 50 62 100 % Gait Cycle 0 12 50 62 100
IC TO IC IC TO IC

FIGURE 10-11. Stiff legged gait pattern with peak knee flexion FIGURE 10-13. Premature action of the soleus and gastrocne-
reduced to 30 degrees in initial swing. Coupled with the equino- mius prevent heel contact throughout stance. Lack of peroneals
varus foot position, swing limb clearance is compromised. activity contributes to inversion throughout stance.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

176 Chapter 10 / Gait Analysis: Clinical Decision Making

Interpretation: ■ FHL transfer to the posterior-lateral calcaneus with

Primitive synergistic patterns dominate with no se- the ankle in DF.
lective motor control. Insufficient anterior tibialis activity, ■ FDL release PL transfer to the lateral forefoot
a PF contracture (−30◦ DF PROM reducing to −20◦ dur- (cuboid) to supplement eversion. PT lengthened only
ing weight bearing) and highly spastic gastroc & soleus if foot remains in fixed inversion after TAL & toe flexor
(>5 sec of clonus) all contribute to equinovarus. Sparse release.
EMG levels of posterior tibialis suggest it is not a primary
Additional Recommendation: Kinesiologic EMG to de-
varus contributor. Equinus is likely caused by the sus-
termine cause of stiff legged gait with possible rectus
tained and highly spastic soleus, gastrocnemius, and FHL.
femoris distal transfer to enhance knee flexion in swing
Low level continuous EMG indicates a combination of
for foot clearance.
primitive motor control activation and contracture. The
relatively high intensity toe flexor EMG suggests their ■ Physical Therapy
activation and length is a major obstruction to forward ■ Post operative functional strengthening coupled with
progression. The marked subtalar and forefoot inversion talocrural and subtalar mobilization
may represent control error, but the EMG intensity of AT, ■ Balance and gait training with quad cane progressing
PT and soleus was significantly less than the toe flexors. to SPC to decrease risk for falls to improve community
Reduced everter activity may also contribute to forefoot ambulation
inversion. Inversion may also be the voluntary avoidance ■ Functional activities to promote independent house-
of weight bearing on the forefoot caused by toe clawing. hold and community mobility
Pain may also limit progression of the COP onto the ■ Orthotics
forefoot. ■ Articulated AFO set in neutral for 6 months postoper-
atively to protect transfers.
Recommendations:
■ Surgical:
■ Achilles tendon lengthening (with a medial emphasis)
to neutral.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

Chapter 10 / Gait Analysis: Clinical Decision Making 177

CASE 4
Dx-Cerebral Palsy-Crouched Gait with Forefoot Contact Only (No Heel Contact)

This 4 year 9-month-old child diagnosed with spastic diple-

gia cerebral palsy is independently ambulatory with bilat-
eral solid polypropylene AFOs and “tone inhibiting” foot-
plate inserts. No previous surgeries were reported. During
walking she exhibited an anterior pelvic tilt with excess
pelvic rotation; flexion, adduction and medial rotation of
the thigh; and knee flexion with persistent heel off during
stance.
Chief Complaint: Decreased functional ambulation (dis-
tance and speed) at home & school.
Referring MD request: Evaluate for a possible heel
cord lengthening, split anterior tibialis tendon transfer
(SPLATT), posterior tibialis lengthening or transfer, and
a hip adductor release.
Clinical Examination:
PROM: DF (knee flexed) 20◦
DF (knee extended) 10◦
Eversion 30◦
Knee hyperextension 20◦
Hip extension −20◦
SLR 70◦

Strength Testing (In synergy) Spasticity

(EMG response)
Anterior Tibialis-Weak None
Posterior Tibialis-Weak + (0.5 sec)
Soleus-Weak + (1.5 sec)
Gastrocnemius-Weak + (2.0 sec) FIGURE 10-14. Foot Floor Contact: IC by forefoot (1st and 5th
Adductor Longus-Weak + (>3.5 sec) MT with toes), no heel contact in stance. Anterior rotation of
the pelvis causes “apparent” medial hip rotation and medial foot
Gait Instrumentation: 3-D motion, stride, and kinesio- progression angle. No heel contact throughout stance is caused
logic EMG analyses during self selected paced walking. by hip and knee flexion as the ankle is in a neutral position. Case
4: Cerebral Palsy: Crouched gait with “apparent” equinus.
Results
Foot Floor Contact: IC by forefoot (1st and 5th MT with
toes), no heel contact throughout stance (Fig. 10-14)
Knee: Flexed 40 degrees at initial contact decreasing to
Stride: % Normal 20 degrees in mid-stance, 40 degrees in terminal stance
Velocity: 44.5 m/min 69 reaching a peak of 60 degrees in mid-swing.
Stride Length: 0.666 m 79 Thigh Rotation: 40 degrees of medial rotation at initial
Cadence: 133 86 contact decreasing to 25 degrees in loading, 35 degrees in
mid-stance, decreasing to 10 degrees of external rotation in
Motion: preswing. Thigh progressively medially rotates throughout
swing to reach 40 degrees in terminal swing (Fig. 10-16).
Ankle: Neutral at initial contact progressing to 5 degrees
of plantar flexion in mid-stance increasing to 15 degrees of Pelvic rotation: 20 degrees of medial rotation at initial
PF degrees in terminal stance rapidly reaching 50 degrees contact increasing to 30 degrees during loading, reaching
in preswing (Fig. 10-15). 8 degrees of lateral rotation in preswing (Fig. 10-17).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

178 Chapter 10 / Gait Analysis: Clinical Decision Making

ANKLE
Patient PELVIS Patient
40 Normal Internal Rotation
Dorsiflexion 40 Normal

20 20

Degrees
Degrees

0 0
-20 -20
-40 External Rotation
Plantar Flexion -40
-60
% Gait Cycle 0 12 50 62
% Gait Cycle 0 12 50 62 100
100
IC TO IC
IC TO IC

FIGURE 10-17. Medial rotation of the pelvis causes the “appar-

FIGURE 10-15. Ankle motion: Loading reduces the ankle to ent” medial rotation of the thigh suggesting the lower extremity
neutral that is maintained until late mid-stance with 50 degrees of rotation is caused by excess anterior pelvic rotation.
plantar flexion in preswing decreasing to 10 degrees in terminal
swing. Forefoot only contact is caused by excess knee and hip
flexion rather than plantar flexion.
position, while hamstring tightness is exaggerated by ex-
cessive hip flexion. The extreme ankle plantar flexion seen
EMG: in preswing most likely represents a rebound response as
Anterior Tibialis: (67.8% MVC) premature onset unloading the foot releases energy in the stretched tissues.
& cessation It signifies some gastrocsoleus contracture but correcting
Soleus: (66.5% MVC) premature onset this by lengthening the entire Achilles tendon would result
& normal cessation in excess dorsiflexion, as loading the flexed limb on the
Gastrocnemius: (28.3% MVC) premature onset forefoot would continue to stretch the tendon. Knee flexion
& cessation can be reduced by correcting the hip flexion contracture to
Posterior Tibialis: (36.8% MVC) premature onset make full hamstring length available. The “apparent” ad-
& cessation duction and internal rotation of the limb and the medial
Adductor Longus: (6% MVC) normal onset foot progression angle are the result of the pelvic forward
& cessation rotation rather than medial rotation of the hip caused by
adductor activity. EMG testing of the medial hamstrings is
Interpretation: The persistent heel off posture implies recommended.
equinus of the foot but the ankle is dorsiflexed to neutral
at initial contact through loading response. Subsequent
plantar flexion develops in late stance as advancement of Recommendations:
the limb over the forefoot allows the knee to partially ex- ■ Surgical:

tend. The ankle position is caused by weight bearing on ■ Iliopsoas: Aponeurotic lengthening

a markedly flexed knee that holds the tibia in a trailing ■ Possible medial hamstring lengthening (EMG recom-
mended)
■ Soleus: Conservative lengthening of the aponeurosis
to neutral. Preserve the gastrocnemius, and take cau-
THIGH tion to avoid overlengthening.
Internal Rotation
40 ■ No posterior tibialis lengthening or transfer
■ No adductor surgery indicated
20
■ Physical Therapy
Degrees

0 ■ Functional strengthening of the hip extensors, abduc-

tors, knee extensors, plantar flexor and dorsiflexors
-20 ■ Optimize motor control strategies to reduce pelvic ro-
External Rotation
-40 tation
■ Gait and balance training
■ Functional training for improved community ambu-
% Gait Cycle 0 12 50 62 100
IC TO IC
lation and recreation
■ Orthotic

FIGURE 10-16. The “apparent” medial thigh rotation reflects ■ Continue with bilateral solid polypropylene AFOs and
pelvic rotation. “tone inhibiting” footplate inserts.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

Chapter 10 / Gait Analysis: Clinical Decision Making 179

CASE 5
Dx-Post Polio Syndrome with Genu Recurvatum

This 48-year-old male with a diagnosis of post polio syn- % Normal

drome was a functional community ambulatory until he Stride: Velocity: 54.5 m/min 67
fractured his distal femur in a fall resulting in a femoral Stride Length: 1.331 m 88
deformity requiring an osteotomy. Cadence: 82 steps/min 76
Single limb
Chief Complaint: Pain in medial and posterolateral knee;
support: Right 90.1
weakness and fatigue in right limb with knee instability
limiting community ambulation. He is currently wearing
a KAFO with an articulated knee (0–145◦ ) joint and ankle Motion:
allowing free dorsiflexion with plantar flexion restricted to
neutral. Ankle: Excess plantar flexion in loading through mid-
stance reversing to excess dorsiflexion in terminal stance
Gait Instrumentation: Motion, stride, and kinesiologic causing knee instability.
EMG analyses during self-selected walking-Barefoot.
Knee: Genu recurvatum throughout loading and mid-
Strength MMT Torque (% Normal) stance with rapid knee flexion in terminal stance (Fig.
Soleus/ 2− Plantarflexion 0 10-19).
Gastrocnemius
Vastus Lateralis 3 Knee extension 8
Patient
Gluteus Medius 4 Abduction 45 KNEE
Gluteus Maximus 3+ Hip Extension 30 70 Flexion Normal
Semimembranosus 50
& BFLH 3+
30
Degrees

10
Foot floor contact: Initial contact with a flat foot avoiding
the heel rocker thereby eliminating the flexor moment at -10 Extension
the knee (Fig. 10-18). Prolonged heel contact throughout
-30
terminal stance into preswing.
Gluteus
Maximus

Biceps
Femoris,
Long Head

Vastus
Lateralis

Soleus

% Gait Cycle 0 12 50 62 100

FIGURE 10-18. Initial contact with a flat foot in 10 degrees of IC TO IC
plantar flexion that effectively eliminates the heel rocker reduc-
ing the knee flexion moment during loading. Patient has weak FIGURE 10-19. Knee hyperextension in stance for weight bear-
quadriceps (3/5) and uses available hyperextension as a passive ing stability with no EMG activity of the soleus or vastus lateralis.
extensor moment to stabilize the knee throughout stance. Case Gluteus maximus is active to control the knee in the presence of
5: Dx Post Polio Syndrome with genu recurvatum. weak quadriceps.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

180 Chapter 10 / Gait Analysis: Clinical Decision Making

EMG: ■ Avoid activities causing fatigue that persists more

Upper and Lower Gluteus Maximus: Large motor units- than 10 minutes
onset normal with prolonged activity into mid-stance (Fig. ■ Interrupt active day by two rest periods
10-19) ■ Stop an activity before pain develops
Vastus Lateralis: No EMG activity ■ No resistive exercise causing muscle or joint pain
Soleus: No EMG activity ■ Gait & balance training for independent functional
community ambulation
Interpretation: The primary deviation was knee hyperex-
■ Orthotic:
tension in early stance with knee collapse in late stance.
■ KAFO: Lightweight
The symptoms were aggravated by a knee extensor thrust
Ankle: A dorsiflexion stop at 5 degrees to assist
into valgus and hyperextension. The mechanism for knee
and protect a weak calf while reducing the flexor
control was dynamic hip extension (intense gluteus max-
moment at the knee throughout stance. Free plan-
imus EMG) for stance stability caused by quadriceps weak-
tar flexion (to protect weak quads during IC and
ness (3/5). Motor unit enlargement indicates more muscle
loading) to attenuate the effect of the heel rocker
fibers are activated simultaneously contributing to overuse
at initial contact. The goal is to prevent fatigue and
and fatigue. With the patient dependent on the gluteus
overuse of quadriceps and hip extensors.
maximus for stability, orthotic restraint at the knee and
ankle is essential to prevent joint deformity and pain. Knee: freely mobile in flexion, restricted to 5 to 10
degrees of hyperextension with a posteriorly offset
Recommendations:
knee joint axis for added stability
■ No surgical recommendations.
■ Physical Therapy
Client education and guidance in modification of func-
tional activities to avoid overuse. (J Perry recommenda-
tions for post polio syndrome [16,19]).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls November 9, 2005 22:17

Chapter 10 / Gait Analysis: Clinical Decision Making 181

CASE 6
Dx-Congenital Hip Dysplasia with Hip Pain During Jogging

This 17-year-old female was diagnosed with right congen-

ital hip dysplasia
Chief complaint: Significant hip and low back pain when
jogging which she performs for recreation and fitness.
Referring MD request: Evaluate for a possible total hip
replacement
Past Medical History: Chiari osteotomy on the right hip
2.5 years ago to stabilize the hip joint
Clinical Examination:
True leg length discrepancy: 1 inch shorter on the right.
Strength Testing
Muscle Group Grade MMT Torque (% Normal)
TFL 3− Abduction 58.5
Gluteus Medius 3−
Adductor Longus 4
Gluteus Maximus 5 Extension 79.8
Hamstrings 5
DF, PF, Knee 5
flexors & ext

Gait Instrumentation: Motion, stride, and kinesiologic

EMG analyses during self-selected and fast walking.

Results
Slight ipsilateral trunk lean and a contralateral pelvic drop FIGURE 10-20. Stance phase on the right LE results in con-
throughout stance (Fig. 10-20) tralateral pelvic drop coupled with a slight ipsilateral trunk lean.
One inch lift on the right equalizes SLS time. Case 6: Dx Congen-
Stride Characteristics with shoes & 1 inch lift ital hip dysplasia.

Self selected Fast

Velocity: 83% N 112% N
Stride length: 88% N 104% N
Cadence: 94% N 109% N Hip Moment
Single limb support: 101% N 116% N
Single limb stance: Rt 100% N L 115% N 1.5 Adduction Barefoot
Shoes
Normal
Nm/kg_m

Motion: Contralateral pelvic drop and a slight ipsilateral

trunk lean in stance phase reducing the hip adduction mo- 0.5
ment (Fig. 10-21).
0
EMG:
Abduction
Gluteus Medius: (31% MVC) premature onset and pro- -0.5
longed activity into terminal stance (Fig. 10-22) % Gait Cycle 0 12 50 62 100
Tensor Fascia Lata: (97.5% MVC) with premature onset
IC TO IC
and prolonged activity into initial swing (Fig. 10-22)
Upper & Lower Gluteus Maximus: no functionally sig- FIGURE 10-21. Right hip adductor moment reduced with an
nificant activity despite high intensity EMG during torque ipsilateral trunk lean decreasing the demand on the abductors.
testing. Joint reaction force is decreased which may reduce pain.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

182 Chapter 10 / Gait Analysis: Clinical Decision Making

nificantly alter her ability to weight bear on the right limb.

Tensor A valgus moment at the knee and a reduced adduction
Fasciae moment at the hip was consistent with a slight ipsilateral
Latae
lateral trunk lean. There was also a contralateral pelvic
drop in stance indicative of hip abductor weakness. Glu-
teus medius and tensor fascia lata both displayed vigorous
Gluteus
Medius muscle activity and adequate strength to provide support
if a total hip arthroplasty is elected at a later date, if pain
significantly limits function. Since the pain is experienced
only during elective jogging and the Chiari procedure
% Gait Cycle 0 12 50 62 100 is functioning adequately, total hip replacement can be
IC TO IC delayed.
FIGURE 10-22. Gluteus medius and tensor fascia latae both
displayed vigorous muscle activity and adequate strength to pro- Recommendations:
vide support if a total hip arthroplasty is elected at a later date if
pain significantly limits function. ■ Surgical
■ Joint replacement surgery is not indicated at this time
■ Right total hip arthroplasty if pain exacerbates and
limits function
Interpretation: Both free and fast walking (with shoes and ■ Physical Therapy
1 inch lift) and muscle strength testing indicated a high ■ Continue with 1 inch lift on right shoe
level of function despite modest hip abduction instability ■ Strengthen right hip abductors (TFL & G medius–
and pain. Self selected velocity was 83% of normal with 58.5% normal) and hip extensors (80% normal)
minimal asymmetry in single limb stance time (Rt-100% ■ Cardiovascular training for health promotion and fit-
N versus L-115% N) signifying that hip pain did not sig- ness (UE ergometer and swimming)
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

Chapter 10 / Gait Analysis: Clinical Decision Making 183

ACKNOWLEDGMENTS 10. Mulroy S, Perry J, Gronley J. A comparison of clinical tests for ankle
plantar flexion strength. Trans Orthop Res Soc. 1991;16:667.
11. Murray MP, Mollinger LA, Gardner GM, et al. Kinematic and
The authors wish to thank Judith Burnfield PT, PhD for EMG patterns during slow, free and fast walking. J Orthop Res.
the preparation of the figures and Alina Kaufman for her 1984;2(3):272–280.
assistance with the references. 12. Olney S, Winter D. Predictions of knee and ankle moments of force
in walking from EMG and kinematic data. J Biomech. 1985;18:9–20.
13. Olney S. Hemiparetic gait following stroke. Part I: characteristics.
Gait Posture. 1996;4:136–148.
REFERENCES 14. Ounpuu S. Joint kinematics: methods, interpretation and treatment
decision-making in children with cerebral palsy and myelomeningo-
1. Beasley WC. Quantitative muscle testing: principles and application cele. Gait Posture. 1996;4:62–78.
to research and clinical services. Arch Phys Med Rehabil. 1961;42:398– 15. Pathokinesiology Department, Physical Therapy Department, Ran-
425. cho Los Amigos Medical Center. Observational Gait Analysis Hand-
2. Brown L, Adams JM, Roller P, et al. Functional distances and veloc- book. Downey, CA: LAREI; 1989.
ities required of community ambulators in Ventura County. Masters 16. Perry J, Young S, Barnes G. Strengthening exercise for post-polio
Thesis. Northridge: California State University; June 2001. sequelae. Arch Phys Med Rehab. 1987;68:660.
3. Duncan PW, Badke MB. Determinants of abnormal motor control. 17. Perry J. Gait Analysis: Normal and Pathological Function. Thorofare,
In: Duncan PW, Badke MB, eds. Stroke Rehabilitation, The Recovery NJ: Slack; 1992.
of Motor Control. Chicago: Year Book Medical Publishers; 1987:135– 18. Perry J. Integrated function of the lower extremity including gait
159. analysis. In: Cruess RL, Rennie WR, eds. Adult Orthopedics. New
4. Everett, D, Adams, J, Wolfe G, et al. The criterion-related validity of York: Churchill Livingston; 1984:1161–1207.
the Rancho Los Amigos observational gait analysis form in persons 19. Perry J. Poliomyelitis. In: Nickel VL, Botte MJ, eds. Orthopaedic Re-
with cerebral vascular accident. Masters Thesis. Northridge: Califor- habilitation. New York: Churchill Livingstone; 1992:493–520.
nia State University; 2000. 20. Richards CL. Hemiparetic gait following stroke. Part II: recovery and
5. Finley RF, Cody KA. Locomotive characteristics of urban pedestrians. physical therapy. Gait Posture. 1996;4:149–162.
Arch Phys Med Rehabil. 1970;51:423–426. 21. Sutherland D. An electromyographic study of the plantar flexors
6. Gage JR. Gait analysis for decision-making in cerebral palsy. Bull of the ankle in normal walking on the level. J Bone Joint Surg.
Hosp Joint Dis Orthop Inst. 1983;43:147–163. 1966;48A(1):66–71.
7. Greenberg M, Gronley J, Perry J, et al. Reliability and concur- 22. Sutherland D. Gait Disorders in Childhood and Adolescence. Balti-
rent validity of Rancho Los Amigos Medical Center’s observa- more: Williams & Wilkins, 1984.
tional gait analysis. California Chapter Conference Proceedings; Nov. 23. Sutherland D. The development of mature gait. Gait Posture.
1996. 1997;6:163–170.
8. Kerrigan DC, Gronley J, Perry J. Stiff-legged gait in spastic paralysis: 24. Sutherland DH, Cooper L. The pathomechanics of progressive crutch
a study of quadriceps and hamstring activity. Am J Phys Med Rehabil. gait in spastic diplegia. Orthop Clin North Am. 1978;9(1):143–154.
1991;70(6):294–300. 25. Waters RL, Garland DE, Perry J, et al. Stiff-legged gait in hemiplegia:
9. Lyons K, Perry J, Gronley J, et al. Timing and relative intensity of hip surgical correction. J Bone Joint Surg. 1979;61A:927–934.
extensor and abductor muscle action during level and stair ambula- 26. Waters RL, Lunsford BR, Perry J, et al. Energy-speed relationship of
tion: an EMG study. Phys Ther. 1983;63:1597–1605. walking: standard tables. J Orthop Res. 1988;6(2):215–222.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-10 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 10:0

184
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 10:0

• C h a p t e r 11
•
◗ Lower Limb Prostheses: Implications
and Applications
John W. Michael

In 1945, engineers and scientists at the University of Cal- gen per distance traveled per kilogram of body weight than
ifornia Biomechanics Laboratory were the first to objec- one permitting unrestricted knee flexion during the swing
tively analyze the gait of amputees with lower limb pros- phase (22). Thus, the primary goal of the clinician is to cre-
theses. Their work provided the first comprehensive look ate prostheses and orthoses facilitating movements that
at the biomechanical attributes of normal and amputee approximate the kinematics and energetics of normal
gait. Their concepts were illustrated with free body dia- walking.
grams and theoretical descriptions of the forces thought When interpreting amputee gait studies, it is impor-
to be involved in controlling a passive artificial limb using tant to note that no prosthetic components generate active
the remnant musculoskeletal system. Figure 11-1 is a re- propulsive motion at any of the lower limb joints. Due to
production of one such illustration (18). These constructs the large torques developed during weight bearing, current
have proven invaluable in the decades that followed, shap- electromechanical technology is unable to produce suffi-
ing how prosthetists worldwide conceptualize their goals cient power in a lightweight, compact package to actively
and practice their craft. plantarflex the ankle or to actively extend the knee under
In more recent decades, as instrumented gait analy- load. This means that all prosthetic and orthotic compo-
sis laboratories became more widely available, a growing nents studied to date basically permit joint motion, or im-
body of knowledge has increased our understanding of the pede it, by limiting the range of motion or by applying
fundamental characteristics of ambulation with an artifi- resistance to decelerate movement.
cial limb. Available scientific evidence supports the ma- From a biomechanical perspective, a free motion pros-
jority of the concepts proposed by Inman and colleagues thetic component is analogous to the joint in an insen-
more than half a century ago, with only a few exceptions sate biological limb with flaccid paralysis. The individual
that will be noted later in this chapter. who walks with a hip disarticulation prosthesis having free
motion hip, knee, and ankle components develops a gait
pattern similar to that of the patient with a flail leg, mak-
ing the study of amputee gait particularly relevant to the
FUNDAMENTAL ASSUMPTION rehabilitation of individuals having limb paralysis. How-
ever, two primary differences between a paralyzed leg and
The fundamental assumption shared by rehabilitation ex- a prosthetic limb must be kept in mind. First, the weight
perts is that grossly abnormal gait patterns are cosmet- of a modern prosthesis is less than one-third that of a nor-
ically and mechanically undesirable. An ample body of mal leg. Secondly, because each artificial joint is individu-
evidence supports this broad statement. For example, ally aligned and adjusted during walking trials, the pros-
Waters et al. have shown that applying a plaster cast to thetist can optimize the biomechanical alignment of the
block ankle movement in a normal subject decreases gait artificial limb. Because the prosthetist can configure the
efficiency, while immobilizing the knee results in even artificial limb in an ideal manner, and because individuals
more inefficient ambulation (37). Furthermore, eliminat- with paralysis frequently have associated bony and soft
ing motion at both the knee and ankle results in a gait that tissue deformities, ambulation with an orthosis applied to
requires more energy than if only one joint or the other a paralyzed leg may not be as effective in restoring nor-
was restricted. Kaufman et al. have shown that an orthosis mal walking as would a prosthesis. These factors may ex-
that locks the knee in extension throughout the entire gait plain the observation that people with paralysis who wear
cycle required a post-polio subject to consume more oxy- an ankle-foot orthosis rarely compete in sports events that

185
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls November 9, 2005 22:20

186 Chapter 11 / Lower Limb Prostheses: Implications and Applications

FIGURE 11-1. Anteroposterior

forces in below-knee amputee
wearing PTB prosthesis. Left, at heel
contact. Middle, during midstance
phase. Right, during push-off. W,
body weight; S, support force;
K, force on patellar ligament; P,
posterior force; A, anterior force; R,
floor reaction.

require running, while a number of amputees with modern lying deficiency still remains. For example, because the
transtibial prostheses do so and some learn to run faster transfemoral residual limb no longer has functional knee
than most nondisabled individuals. extensors, people with this amputation who walk with a
More technologically advanced prosthetic components basic prosthetic knee mechanism quickly develop a char-
use various electromechanical methods to selectively add acteristic “stiff knee” gait pattern throughout stance phase.
resistance to artificial joints to dampen undesired motion. The remaining hip extensor muscles are used to alter the
Perhaps the best example is the use of a hydraulic cylin- ground reaction forces to create an external knee extension
der linked to the knee joint that controls knee movement moment throughout stance, thus ensuring that the passive
during swing phase. From a biomechanical perspective, prosthetic knee will not buckle or collapse during weight
prosthetic components that resist motion partially simu- bearing.
late muscles that are contracting eccentrically. However, This gait deviation is a valuable compensation for the
there are two important distinctions from normal muscle. loss of active knee extension control, and without it the
Current prosthetic components are not actively controlled amputee could not walk using a basic artificial limb. There-
by the amputee’s neuromuscular system. The present state- fore, efforts to train the amputee to allow the knee to flex
of-the-art requires that the control be inherent in the mech- during early stance or to produce a more normal hip ex-
anism itself, so even the most advanced microprocessor- tensor muscle pattern while using such a simple prosthetic
controlled knee applies preestablished algorithms to knee component would be inappropriate. This illustrates
simulate normal swing phase control. Equally important, the concept that the optimal gait pattern for a particular
prosthetic components that dampen movement should amputee will not necessarily be fully normal and serves
also absorb energy. The amputee must generate sufficient as a reminder that when evaluating amputee mobility, it
motion by using remaining hip flexor muscles to produce is important not only to identify any gait abnormalities
adequate swing phase function. The resistance within a present but also to determine if they serve a useful clinical
prosthetic knee converts some of this kinetic energy into purpose.
heat to limit excessive movement. This explains why am-
bulation with sophisticated prosthetic components may be
more energy efficient than walking with more primitive AMPUTEE GAIT
devices, but no artificial limb is as efficient as an intact
neuromuscular system. Studies of amputee walking have consistently demon-
One very important caveat is that some gait deviations strated characteristic disruptions in temporospatial, kine-
are recognized clinically as necessary compensations for matic, kinetic, and metabolic gait parameters compared
the musculoskeletal abnormalities inherent in limb am- to normal walking. A key finding is that amputees typi-
putation. Eliminating these aberrations may not result cally spend less time bearing weight on the artificial limb
in a more effective gait pattern so long as the under- and proportionately more time on the intact leg. This
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 10:0

Chapter 11 / Lower Limb Prostheses: Implications and Applications 187

results in stride asymmetry with a shortened stance phase sected and allowed to retract. To illustrate, Waters et al.
and more prolonged swing phase on the prosthesis (40). have demonstrated that Syme ankle disarticulation am-
The typical amputee gait pattern is often characterized by putees consume less energy than subjects with transtibial
more rapid unloading of the prosthetic limb in terminal or higher amputations (39) and similar results have been
stance combined with earlier loading of the contralateral shown for knee disarticulation compared to transfemoral
limb (29). Amputees consistently demonstrate a slower amputation and for hip disarticulation amputation com-
than normal self-selected walking speed and transfemoral pared to transpelvic levels.
amputees often have a greater stride width, suggestive of The biomechanical value of preserving remnants of the
a wider base of support to increase stability (21). Knee foot has recently been challenged and may prove to be an
flexion during stance phase is markedly reduced on the exception to the general rule that added bony length trans-
prosthesis side (31) while hip extension, knee flexion, and lates into more efficient gait. Several authors have sug-
ankle dorsiflexion of the opposite limb are increased from gested that the critical factor in partial foot amputation is
normal (3). Unilateral amputees demonstrate increased the preservation of the metatarsal heads by disarticulation
muscle activity on both the involved and contralateral at the metatarsal-phalangeal junction. Recent gait studies
side, with the pattern varying depending on whether the have shown that once the metatarsal heads are disrupted,
amputation is at the transtibial (19) or transfemoral gait mechanics become grossly abnormal because the
level (20). This general pattern of compensating with the ability to actively push off is lost (30). Prostheses and or-
contralateral leg and with proximal muscles on the in- thoses studied to date have been unable to restore this
volved limb is believed to be one of the major causes ability for those with transmetatarsal and more proximal
for the increased effort required to walk with an artificial partial foot ablations (27).
limb.
Cause of Amputation
Level of Amputation
The cause of an amputation affects the quality of walk-
One of the principal findings reported in the literature is ing with a prosthesis. As a group, individuals with trau-
that more proximal amputations disrupt gait mechanics matic amputations have been shown to walk at a faster
more extensively than distal amputations and increase the self-selected velocity, to have a lower net energy cost dur-
net energy cost to ambulate with a prosthesis over a given ing ambulation, and to have a gait pattern that more closely
distance (9). These observations are the foundation for the approximates normal than do the group of individuals who
surgical dictum to “preserve the functioning knee joint and have lost a limb as a result of vascular insufficiency (with or
spare all functional limb length.” without associated diabetes mellitus). Amputation as a re-
It seems intuitive that as more biological joints are lost sult of vascular disease tends to occur in older individuals,
to amputation, the ability to actively control motion is fur- and many dysvascular amputees have associated comor-
ther compromised, and available data consistently support bidities. These factors have been speculated to account, at
this notion. Some reports in the literature also support the least in part, for the difference in performance compared
concept that the shorter the bony remnant remaining, the to traumatic amputees. Analyzing gait data according to
more abnormal and inefficient the amputee’s gait becomes. the cause of the subject’s limb loss helps to clarify differ-
For example, Gonzales et al. demonstrated that subjects ences between these two groups. Waters and Mulroy have
with long transtibial amputations had a more energy effi- published an excellent review of the energy costs of normal
cient gait than individuals with short transtibial residual and pathologic gait discussing many of these factors (38).
limbs (14). This suggests that a longer skeletal segment Figure 11-2 illustrates two general trends from the litera-
may provide the amputee with more leverage to control ture: individuals with more proximal amputations expend
the prosthesis, which presumably leads to a more normal more energy to cover a given distance, as well as walk at
gait pattern. progressively slower velocities than those with longer
Disarticulations are believed to offer characteristic residual limbs.
biomechanical advantages over more proximal diaphyseal
ablations, particularly when distal end weight bearing is
possible. In addition to offering the longest possible bony PROSTHETIC COMPONENTS
lever arm for control of the artificial limb, disarticulation
generally results in preservation of the full mass of all A growing body of evidence has demonstrated that the
proximal muscle bellies, and the muscles are usually an- functional components within the prosthesis can have a
chored distally under normal tension. Preserving all of the significant effect on the kinematics and kinetics of gait.
muscle’s fibers and reattaching them at the normal resting The classic study by Doane and Holt was the first to con-
length should allow the amputee to generate more force clude that the primary biomechanical distinction of the
for prosthesis control than if the muscles had been tran- articulated, single axis foot was that it achieved foot flat
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 10:0

188 Chapter 11 / Lower Limb Prostheses: Implications and Applications

cians had not anticipated that prescription of a specific

prosthetic component might reduce stresses on the surviv-
ing foot, which is known to be at risk for lesions in the case
of the dysvascular amputee.
Research to date has not demonstrated any energy ef-
ficiency advantage for dynamic response feet when the
amputee is walking on level surfaces at a relatively slow
self-selected pace (32). However, during uphill walking
or at more rapid cadences, dynamic response feet have
been shown to reduce the net oxygen costs of ambulation
compared to walking with less sophisticated mechanisms
(24,36). However, it is important to recognize that ambu-
lation with even the most advanced prosthetic ankle-foot
devices still requires significantly more effort than normal
walking. Gait studies have shown that amputee gait is char-
FIGURE 11-2. This overlay of two graphs, modified from Waters
acterized by a reduction in the magnitude of the second
and Mulroy (1999), summarizes two important general trends vertical ground reaction force peak (2), sometimes termed
from the literature on walking with a lower limb prosthesis. In- the A2 power burst, which occurs in the latter half of stance
dividuals with higher amputation levels expend more net energy phase (4), as depicted in Figure 11-3. This illustrates that
to walk a given distance than those having more distal amputa- passive, spring keel prosthetic feet are unable to effectively
tions. In addition, their self-selected walking speed is progres-
sively slower as the level of amputation becomes more prox-
restore this aspect of gait kinetics, which may contribute
imal. TP = transpelvic amputation; HD = hip disarticulation to the higher energy costs of amputee walking. Studies of
amputation; TF = transfemoral amputation; KD = knee disar- normal subjects have established that individuals tend to
ticulation amputation; TT = transtibial amputation; normal = choose a walking speed that is most energy efficient. It is
nonamputee. possible that some of the increased effort required for am-
putees to ambulate a given distance is directly related to
the slower cadence adopted by this population.
more rapidly than other tested alternatives (8), a conclu- Childress’ group at Northwestern University has done
sion borne out by subsequent researchers (7). To illustrate many of the more recent investigations of lower limb
the importance of such findings, the primary indication prosthetic components. One intriguing proposal from this
today for the prescription of single axis ankle-foot mech- laboratory has been that prosthetists dynamically align
anisms is to enhance prosthetic knee stability (25). This the ankle-foot mechanism to optimize its functional per-
occurs because the net vertical ground reaction force vec- formance in view of each transtibial amputee’s physical
tor quickly moves anteriorly at foot flat, from the back edge condition and individual gait pattern. It appears that the
of the heel to the ankle mortise, thus creating an external subjective changes the prosthetist makes to the angulation
moment on the prosthetic knee that stabilizes it in exten- and linear position of the foot with respect to the prosthetic
sion. The more rapidly this happens, the more stable the socket gives the prosthesis a “rollover characteristic” that
prosthetic limb. approximates that of a normal ankle and foot (15).
A number of researchers have looked at prosthetic feet Although prosthetic ankle-foot components have been
incorporating a plastic spring element within the forefoot, studied extensively, less research has been conducted on
commonly referred to as “dynamic response” components. the effect of prosthetic knee mechanisms on amputee
One consistent trend is that because they have a more flex- gait. The growing clinical utilization of microprocessor-
ible forefoot than earlier single axis or solid ankle cush- controlled devices should stimulate future research in
ion heel (SACH) feet, dynamic response feet allow a more this area. Basic prosthetic knee mechanisms with friction
normal range of passive ankle dorsiflexion in late stance swing phase controls can only be optimized for a single
(23),although the total range of motion is still less than cadence. Hicks et al. corroborated the clinical impression
normal. Particularly for those components whose design that such single axis constant friction prosthetic knees only
incorporates a spring that extends fully to the phalangeal swing at a fixed rate which is largely determined by the pen-
region of the prosthetic foot, this seems to result in a dular effect of the shank (17). Their results showed that
longer stride length on the prosthetic side. Interestingly, increasing or decreasing the amount of mechanical fric-
dynamic response feet have also been shown to result in tion on the knee changed the magnitude of knee flexion in
less vertical impact on the contralateral heel during load- early swing but had very little effect on overall swing phase
ing response (1). This illustrates one of the primary advan- timing.
tages of instrumented gait analysis: it sometimes reveals One important implication of this information is that
effects that are unexpected and counterintuitive. Clini- studies of ambulation with a prosthesis should report on
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 10:0

Chapter 11 / Lower Limb Prostheses: Implications and Applications 189

FIGURE 11-3. This graph of the vertical ground

reaction forces (GRF) generated during stance
phase illustrates typical differences between am-
putee gait, shown as the solid line, and normal
gait, shown as the dotted line. During the first
half of stance phase, the prosthetic limb usu-
ally exhibits a somewhat lower maximum verti-
cal GRF than normal, suggesting that the am-
putee is placing less load onto the artificial limb.
During the latter half of stance phase, the second
peak of the vertical GRF on the prosthetic limb
is significantly lower than normal. This indicates
that the prosthetic ankle-foot mechanism gener-
ates less propulsive force than the normal leg. Dy-
namic response prosthetic feet, which have fore-
foot springs, generate a higher second peak than
less sophisticated components but do not exceed
100 percent of body weight; the intact limb gen-
erates about 120 percent of body weight during
normal walking.

the specific components being used. For example, if gait that the range of knee flexion with these passive “stance
analysis shows that a group of transfemoral amputee sub- flexion” components is less than normal and begins later
jects walked more slowly than a comparable transtibial in the stance phase, but these components clearly offer
group, it would be important to report that all of the TF the potential for the amputee to walk without the stiff-
subjects used a constant friction (one speed) knee. Similar legged gait that characterizes basic single axis devices (11).
results in a group that included TF subjects whose pros- This emphasizes the point that gait data from subjects fit-
theses contained fluid-controlled knee mechanisms would ted with more technologically advanced knee components
be of particular interest, since hydraulic and pneumatic should be expected to differ significantly from earlier stud-
knees and particularly those that are microprocessor- ies based on single axis, constant friction knee mecha-
controlled have been shown to permit amputees to walk nisms. Some of the kinematic abnormalities identified in
safely throughout a range of cadences (17). Unfortunately, early gait studies, such as full knee extension throughout
many of the gait studies of earlier decades did not report during early stance, were necessary compensations for the
any information about the components within the pros- biomechanical limitations of the available components of
thesis and therefore their results must be interpreted with that era.
caution. Telescoping pylons, often termed “vertical shock sys-
tems” have become increasingly well accepted clinically,
and preliminary research suggests that these components
SHOCK ABSORPTION approximate the overall stiffness of the intact human leg.
Gait analysis has shown that shock pylons can significantly
Inman and colleagues proposed that knee flexion during reduce the loading on the residual limb during loading re-
stance minimized the vertical displacement of the center sponse (6) but may create a prosthesis that is function-
of mass and therefore helped contribute to efficient am- ally short under weight bearing (13). Further investigation
bulation. More recent studies have cast doubt on this as- is needed to better document the biomechanical impli-
sumption, suggesting that the primary benefit of knee flex- cations of using these nonanatomic structures to provide
ion during loading response is to provide shock absorption shock absorption.
(28). Current data also suggest that pelvic list, the second
element in Inman’s original six determinants of gait also
contributes more to shock absorption than was originally
believed (12). This increased appreciation for the role of ADDITIONAL FACTORS
shock absorption during normal walking has led to the
development of prosthetic components that are intended In the first edition of this text, Radcliffe and colleagues dis-
to provide this function during amputee gait. cussed a rationale for alignment of prosthetic knee devices
Within the past decade, several innovative prosthetic based on the external moments created about the mechan-
knee designs that allow controlled knee flexion during ical joint axis. Experimental evidence is now available to
loading response have been developed. Gait analysis shows support this hypothesis, and at least one researcher has
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls November 9, 2005 22:20

190 Chapter 11 / Lower Limb Prostheses: Implications and Applications

shown that a prosthesis that was subjectively considered the United States occur in individuals who are retirement
“optimally aligned” by an experienced prosthetist and am- age and older, and many have advanced vascular disease
putee subjects minimized the internal hip flexion and ex- and associated cardiopulmonary restrictions, a substan-
tension control moments generated on the affected side tial percentage of patients who use a prosthesis cannot
(26). complete the multiple trials necessary to participate in a
Moving the prosthetic knee one centimeter anterior scientific study. Many of those who do have the physical
or posterior to the subjectively determined “optimal” lo- ability to walk for longer periods of time require balance
cation in the sagittal plane increased the overall inter- aids such as canes, crutches, or a walker, which alter their
nal hip moments, suggesting that the amputee would gait mechanics.
exert more muscle force to control a knee that was As a practical matter, most published studies have a
malaligned. Moving the knee posteriorly resulted in an in- selection bias in that the subjects who participated were
crease in the hip flexion moment, supporting Radcliffe’s younger, more physically active individuals who required
argument that more stable alignment would require in- amputation due to trauma or tumor but who were oth-
creased force to “break” the knee to permit flexion dur- erwise healthy. Results from this population are not nec-
ing swing phase. Moving the knee anteriorly, which Rad- essarily valid for the more typical elderly dysvascular
cliffe (33) proposed should make the knee less stable, amputee.
required more hip extension force during loading re- Finally, the technological sophistication of the com-
sponse (5), presumably to prevent the prosthetic knee from ponents incorporated in the prosthesis has been shown
collapsing. to have a significant potential to influence the objective
Socket design is widely believed to be one of the most parameters measured during instrumented gait analysis.
critical aspects of amputee rehabilitation, but very few gait Most recently, microprocessor-controlled hydraulic and
studies have investigated the impact of this aspect of the pneumatic prosthetic knee mechanisms have been shown
artificial limb. Some authors have suggested that modern to increase the efficiency of gait in tested subjects more
“ischial containment” transfemoral sockets result in a than noncomputerized equivalents (10,35).
more energy-efficient gait than the earlier “quadrilateral” These limitations mean that current data on gait with
type (33), but the small number of subjects investigated a prosthesis, and the summary statements made in this
limits the conclusions that can be drawn from such stud- chapter, are best viewed as suggestive evidence rather than
ies. Suspension mechanisms are also believed to be sig- verified fact. While it is likely that many of these findings
nificant factors that influence use of the prosthesis, but will be corroborated by larger and better controlled future
very little data is available on the influence of suspension work, it is also quite possible that some of our beliefs will
alternatives on prosthetic gait. These areas remain fertile be shown to have significant limitations.
ground for future investigations.

CONCLUSION
LIMITATIONS IN CURRENT STUDIES
Free motion prosthetic joints have biomechanical charac-
Perhaps the greatest limitation to our current knowledge, teristics that are quite similar to a limb with flail paral-
present in almost all investigations of amputee gait pub- ysis, except that the alignment and adjustments of the
lished to date, is the small number of subjects studied. Ma- prosthetic components can be optimized for each indi-
jor leg amputations are, fortunately, not a common prob- vidual patient. Altering the alignment or adjustment of
lem and higher levels of loss occur less frequently than the mechanisms and testing the results with observational
those below the knee joint. For example, hip disarticula- or instrumented gait analysis can provide useful infor-
tion prostheses are only one or two percent of all artificial mation. Adding resistance to prosthetic joints is similar
limbs fitted, so it is virtually impossible to recruit more to walking by using only eccentric muscle contractions.
than a handful of local subjects who walk with these de- The passive nature of currently available prosthetic com-
vices for gait analysis. Transtibial amputation is the most ponents provides a mechanically simple mechanism for
common major lower limb loss and therefore this level walking that can seem surprisingly effective clinically but
has the largest number of subjects enrolled, as well as the has significant limitations that instrumented gait analy-
greatest number of studies published. For most other lev- sis makes apparent. Insights from the growing body of
els, less than a dozen subjects have been studied at any one evidence provided by modern gait analysis have spurred
time. designers and clinicians to develop more sophisticated
Another factor making these studies challenging is that prosthetic components and advanced clinical methods in-
a certain level of physical fitness is required to walk for an tended to reduce the gap between normal and prosthetic
extended period of time. Since most new amputations in walking.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 10:0

Chapter 11 / Lower Limb Prostheses: Implications and Applications 191

REFERENCES 20. Jaegers SMHJ, Arendzen JH, De Jongh HJ. An electromyographic

study of the hip muscles of transfemoral amputees in walking. Clin
1. Barth DG, Schumacher L, Thomas SS. Gait analysis and energy cost Orthop 1996;328:119–128.
of below-knee amputees wearing six different prosthetic feet. J Pros- 21. James U, Oberg K. Prosthetic gait pattern in unilateral above-knee
thet Orthot 1992;4:63–75. amputees. Scand J Rehabil Med 1973;5:35–50.
2. Barth DG, Schumacher L, Thomas SS. Gait analysis and energy cost 22. Kaufman KR, Irby SE, Mathewson JW. Energy-efficient-knee-ankle-
of below-knee amputees wearing six different prosthetic feet. J Pros- foot orthosis: a case study. J Prosthet Orthot 1996;8:79–85.
thet Orthot 1992;4:63–75. 23. MacFarlane PA, Nielsen DH, Shurr DG. Mechanical gait analysis of
3. Bateni H, Olney SJ. Kinematic and kinetic variations of below-knee transfemoral amputees: SACH foot versus the Flex-Foot. J Prosthet
amputee gait. Prosthet Orthot 2002;14:2–10. Orthot 1997;9:144–151.
4. Bateni H, Olney SJ. Kinematic and kinetic variations of below-knee 24. MacFarlane PA, Nielsen DH, Shurr DG. Transfemoral amputee phys-
amputee gait. J Prosthet Orthot 2002;14:2–10. iological requirements: comparison between SACH foot walking and
5. Blumentritt S, Scherer HW, Michael JW. Transfemoral amputees Flex-Foot walking. J Prosthet Orthot 1997;9:138–143.
walking on a rotary hydraulic prosthetic knee mechanism: a pre- 25. Michael J. Modern prosthetic knee mechanisms. Clin Orthop
liminary report. J Prosthet Orthot 1998;10:61–70. 1999;361:39–47.
6. Blumentritt S, Scherer HW, Wellershaus U. Design principles, biome- 26. Michaud SB, Gard SA, Childress DS. A preliminary investigation of
chanical data and clinical experience with a polycentric knee offering pelvic obliquity patterns during gait in transtibial and transfemoral
controlled stance phase knee flexion: a preliminary report. J Prosthet amputees. J Rehabil Res Dev 2000;37:1–10.
Orthot 1997;9:18–24. 27. Murphy AT, Burgess-Limerick. Influence of partial foot amputa-
7. Culham EG, Peat M, Newell E. Below-knee amputation: a compari- tion on sagittal gait mechanics and locomotor function. Abstract
son of the effect of the SACH foot and single axis foot on electromyo- from 7th Annual Gait and Clinical Movement Analysis Society Confer-
graphic patterns during locomotion. Prosthet Orthot Int 1986;10:15– ence, April 16-29, 2002 [available online at www.motionanalysis.com/
22. pdf/gcmas2002/murphy.pdf]
8. Doane NE, Holt LE. A comparison of the SACH and single axis foot 28. Murray MP, Mollinger LA, Sepic SB, Gardner GM, Linder MT. Gait
in the gait of unilateral below-knee amputees. Prosthet Orthot Int patterns in above-knee amputee patients: hydraulic swing control
1983;7:33–36. versus constant-friction knee components. Arch Phys Med Rehabil
9. Fisher GV, Gullickson G. Energy cost of ambulation in health and 1983;64:339–345.
disability: a literature review. Arch Phys Med Rehabil 1978;59:124– 29. Nissan M. The initiation of gait in lower limb amputees: some related
133. data. J Rehabil Res Dev 1991;28:1–12.
10. Gailey RS, Lawrence D, Burditt C. The CAT-CAM socket and quadri- 30. Pinzur MS, Wolf B, Havel RM. Walking pattern of midfoot and ankle
lateral socket: a comparison of energy cost during ambulation. Pros- disarticulation amputees. Foot Ankle Int 1997;18: 635–638.
thet Orthot Int 1993;17:95–100. 31. Powers CM, Rao S, Perry J. Knee kinetics in trans-tibial amputee
11. Gard SA, Childress DS. The effect of pelvic list on the vertical gait. Gait & Posture 1998;8:1–7.
displacement of the trunk during normal walking. Gait Posture 32. Powers CM, Torburn L, Perry J. Influence of prosthetic foot design
1997;5:233–238. on sound limb loading in adults with unilateral below-knee amputa-
12. Gard SA, Childress DS. What determines the vertical displacement tions. Arch Phys Med Rehabil 1994;75:825–829.
of the body during normal walking? J Prosthet Orthot 2001;13: 33. Radcliffe CW. Functional considerations in the fitting of above-knee
64–67. prostheses. Artificial Limbs 1955;2:35–60.
13. Gard SA, Konz RJ. The effect of a shock absorbing pylon on the gait 34. Schmalz T, Blumentritt S, Jarasch R. Energy expenditure and biome-
of persons with unilateral transtibial amputation. J Rehabil Res Dev chanical characteristics of lower limb amputee gait: the influence of
2003;40:109–124. prosthetic alignment and different prosthetic components. Gait Pos-
14. Gonzales EG, Corcoran PJ, Reyes RL. Energy expenditure in below- ture 2002:16:255–263.
knee amputees: correlation with stump length. Arch Phys Med Rehabil 35. Taylor MB, Clarke E, Offord EA. A comparison of energy expenditure
1974;55:111–118. by a high level trans-femoral amputee using the Intelligent Prosthe-
15. Hafner BJ, Sanders JE, Czerniecki JM, Fergason J. Transtibial sis and conventionally damped prosthetic limbs. Prosthet Orthot Int
energy-storage-and-return prosthetic devices: a review of energy con- 1996;2:116–121.
cepts and a proposed nomenclature. J Rehabil Res Dev 2002;39: 36. Torburn L, Perry J, Ayyappa E. Below-knee amputee gait with dy-
1–11. namic elastic response prosthetic feet: a pilot study. J Rehabil Res
16. Hansen AH, Childress DS, Knox EH. Prosthetic foot roll-over shapes Dev 1990;27:369–384.
with implications for alignment of trans-tibial prostheses. Prosthet 37. Waters RL, Campbell J, Thomas L, Hugos L, Davis P. Energy costs
Orthot Int 2000;24:205–215. of walking in lower extremity prosthetic casts. J Bone Joint Surg Am
17. Hicks R, Tashman S, Cary JM. Swing phase control with knee friction 1982;64: 896–899.
in juvenile amputees. J Orthop Res 1985;3:198–201. 38. Waters RL, Mulroy S. The energy expenditure of normal and patho-
18. Inman VT, Ralston HJ, Todd F. Human Walking. Baltimore/London: logical gait. Gait and Posture 1999;9:207–231.
Williams & Wilkins 1982;140. 39. Waters RL, Perry J, Antonelli D. Energy cost of walking of amputees:
19. Isakov E, Bruger H, Krajnik J. Knee muscle activity during am- influence of level of amputation. J Bone Joint Surg 1976;58A:42–46.
bulation of trans-tibial amputees. J Rehabil Med 2001;33:196– 40. Zuniga EN, Leavitt LA, Calvert JC. Gait patterns in above-knee am-
199. putees. Arch Phys Med Rehabil 1972;53:373–382.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-11 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 10:0

192
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

• C h a p t e r 12
•
◗ Simulation of Walking
Frank C. Anderson, Allison S. Arnold, Marcus G. Pandy,
Saryn R. Goldberg, and Scott L. Delp

Many elements of the neuromusculoskeletal system in- A theoretical framework is needed, in combination with
teract to enable walking. Scientists fascinated by human experiments, to advance our understanding of neuromus-
movement have performed an extensive range of stud- culoskeletal function during walking, e.g., to uncover the
ies to describe these elements, e.g., to identify the pro- principles that govern the coordination of muscles dur-
cesses involved in neuromuscular activation, characterize ing normal gait, to determine how neuromuscular impair-
the mechanics of muscle contraction, describe the geomet- ments contribute to abnormal gait, and to predict the func-
ric relationships between muscles and bones, and quan- tional consequences of treatments. It is imperative that this
tify the motions of joints. Clinicians who treat walking framework reveals the cause-effect relationships between
abnormalities in individuals with cerebral palsy, stroke, neuromuscular excitation patterns, muscle forces, ground
and other neuromusculoskeletal disorders have exam- reaction forces, and motions of the body. A dynamic sim-
ined the electromyographic (EMG) patterns, gait kinemat- ulation of walking that integrates facts about the anatomy
ics, and ground reaction forces of literally thousands of and physiology of the neuromusculoskeletal system and
patients, both before and after treatment interventions. the mechanics of multi-joint movement provides such a
However, synthesizing detailed descriptions of the neu- framework.
romusculoskeletal system with gait measurements to cre- What is a dynamic simulation of movement, and how
ate an integrated understanding of normal gait, to iden- can it complement experimental studies of walking? A dy-
tify the sources of pathologic gait, and to establish a namic simulation is, in essence, a solution to a set of equa-
scientific basis for treatment planning remains a major tions that describe how the forces acting on a system cause
challenge. motions of the system over time, as governed by the laws of
Using experiments alone to meet this challenge has physics. A “muscle-driven” dynamic simulation of walking,
two fundamental limitations. First, important variables, therefore, describes how the forces produced by muscles
including the forces generated by muscles, are not read- (and other sources of force, such as gravity) contribute to
ily accessible in experiments. Second, even when variables motions of the body segments during the gait cycle.
can be measured accurately, it is often difficult to establish The process for developing, testing, and analyzing
cause-effect relationships. As a result, elucidating the func- a muscle-driven simulation of movement involves four
tions of muscles from experiments is not straightforward. stages (Figure 12-1). Stage 1 is to create a computer model
For example, ground reaction forces (see Chapter 4) can that characterizes the dynamic behavior of the neuromus-
be measured and used to estimate the accelerations of the culoskeletal system with sufficient accuracy to answer spe-
body’s center of mass. However, force plate measurements cific research questions. The models that we, and others
alone offer little insight into how muscles contribute to have developed typically include detailed descriptions of
these accelerations, and therefore to the critical tasks of musculoskeletal geometry and equations that describe the
supporting and propelling the body forward. EMG record- activation and force production of muscles and the multi-
ings (see Chapter 6) can indicate when a muscle is active, joint dynamics of the body. Stage 2 is to find a set of muscle
but examination of EMG recordings does not allow one to excitations which, when applied to the model, generate
determine which motions of the body arise from a mus- a simulation that reproduces the movement of interest.
cle’s activity. Indeed, determining how individual muscles Stage 3 is to verify that the simulation is indeed repre-
contribute to observed motions is not necessarily intuitive, sentative of the movement of interest by comparing the
as explained below, because a muscle can accelerate joints results of the simulation to experimental data. Stage 4 is
that it does not span and body segments that it does not to analyze the simulation to answer the research or clinical
touch (105). questions posed.

193
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

194 Chapter 12 / Simulation of Walking

FIGURE 12-1. Stages involved in creating and analyzing a simulation of walking. A rigorous assessment
of muscle function during walking requires (1) a model of neuromusculoskeletal dynamics, typically
formulated as a set of equations that relate accelerations of the limb segments to the forces generated
by muscles and other external forces, such as the force of gravity, (2) a simulation of the gait cycle,
obtained by applying a set of muscle excitations to the model and integrating the equations of motion
forward in time, (3) tests to verify that the simulation is sufficiently accurate to answer specific research
questions, generally performed by comparing aspects of the model and simulation to experimental data,
and (4) analyses of the simulation to determine how individual muscles generate forces and contribute
to motions of the body.

A muscle-driven dynamic simulation provides unique celerated (Figure 12-2). The acceleration of the segment is
capabilities that complement experimental approaches. resisted by the inertia of adjoining segments, giving rise to
Simulations provide estimates of important variables, intersegmental forces at the joints. These intersegmental
such as muscle and joint forces, which are difficult to mea- forces are transmitted from one segment to another due to
sure experimentally. Simulations also enable cause-effect the “coupled” multi-articular nature of the body. The trans-
relationships to be explained. For instance, the contribu- mission of these intersegmental forces, often referred to as
tion that a muscle makes to the ground reaction force can dynamic coupling, means that a force applied to one body
be calculated. A simulation also allows “what if?” studies segment accelerates all body segments, not just the seg-
to be performed in which, for example, the excitation pat- ment to which the force is applied. The relative magnitudes
tern of a muscle can be changed and the resulting motion of the intersegmental forces at the joints depend, in part,
can be observed. These capabilities provide new ways to on the mass and inertial properties of the segments be-
characterize the functions of muscles during walking and ing accelerated. For example, intersegmental forces at the
other tasks. hip that arise from accelerations of the thigh are generally
Why are simulations needed to characterize the func- greater than intersegmental forces at the metatarsopha-
tions of muscles during walking? Simulations are needed langeal joints that arise from accelerations of the toes be-
because experimental approaches to infer a muscle’s ac- cause the thigh is more massive than the toes. The mag-
tions, based on the muscle’s attachments, EMG activity, nitudes and directions of the intersegmental forces also
and measured motions of the body, do not explain how the depend on the applied muscle force and the configuration
forces produced by the muscle accelerate the body seg- of the body.
ments and contribute to motions of the joints. When a How significant are the intersegmental forces induced
muscle applies a force to a segment, that segment is ac- by muscles during walking? In particular, are these forces
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

Chapter 12 / Simulation of Walking 195

FIGURE 12-2. Dynamic actions of soleus during single-limb stance (knee flexion angle is ex-
aggerated for the purpose of illustration). (A) The force applied by soleus, a uniarticular muscle
spanning the ankle, not only generates an ankle plantarflexion moment (circular black arrow),
but also induces intersegmental forces throughout the body. The magnitudes and directions of
these intersegmental forces depend on the force applied by the muscle, the moment arms of the
muscle, the inertial properties of the segments, and the configuration of the body. In this example,
the force applied by soleus produces a counter-clockwise angular acceleration of the shank. This
acceleration requires the location of the knee joint to accelerate to the left and upward. The iner-
tia of the thigh resists this acceleration, resulting in an intersegmental force at the knee (straight
black arrow). The intersegmental force at the knee accelerates the thigh, which in turn induces
an intersegmental force at the hip (straight black arrow), and so on. (B) As a consequence of the
intersegmental forces induced by soleus, soleus accelerates not only the ankle, but all the joints of
the body. At the body position shown, soleus accelerates the ankle toward plantarflexion, the knee
toward extension, the hip toward extension, and the trunk upward. Over time, these accelerations
give rise to changes in position. Thus, due to dynamic coupling, soleus does not function solely as
an “ankle plantarflexor”—in many situations, it likely does much more. In similar fashion, other
muscles induce intersegmental forces and accelerate joints that they do not span.

large enough to influence our interpretation of the mus- the hip toward extension (9,45,58) and provides vertical
cles’ actions? The answer to this question, in many cases, support for the trunk (6,56,106). Based on this work and
is yes. Due to dynamic coupling, a muscular moment gen- other examples, we believe that the effects of dynamic
erated at one joint induces accelerations of that joint and coupling must be considered when attempting to deter-
other joints— including joints that the muscle does not mine the functional roles of muscles during gait. Using
span. These “muscle-induced” accelerations are generally a muscle-driven dynamic simulation, the intersegmental
small at joints far removed from the muscle; however, the forces that result from a muscle’s force can be calculated,
induced accelerations of nearby joints can be substantial. and the resulting motions of the body segments can be
For example, soleus exerts only an ankle plantarflexion quantified.
moment, yet Zajac and Gordon (105) have demon- The remainder of this chapter summarizes our experi-
strated that soleus can accelerate the knee into exten- ences with the development and analysis of muscle-driven
sion more than it accelerates the ankle into plantarflexion. simulations of human walking. There are many plausi-
During the stance phase of normal gait (Figure 12-2), ble approaches, and this chapter is not intended to be a
soleus accelerates both the knee and the ankle toward ex- comprehensive review. Rather, the chapter describes the
tension (9,45,56), an action commonly termed the plan- development (Stages 1 and 2) and testing (Stage 3) of a
tarflexion – knee extension couple (33). The plantarflexion particular simulation—a three-dimensional dynamic sim-
moment exerted by soleus during stance also accelerates ulation of normal gait (5). Three studies are reviewed in
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

196 Chapter 12 / Simulation of Walking

which this simulation was analyzed (Stage 4) to gain new lated by motor unit action potentials, can be modeled by
insights into the functions of individual muscles during relating the time rate of change of muscle activation (ȧ) to
walking. The chapter concludes by summarizing some of muscle activation (a) and excitation (u):
the limitations of current modeling, simulation, and analy-
sis techniques. Consideration of these limitations suggests ȧ = (u2 − ua)/ · τact + (u − a)/τdeact , (1)
future research. where τact and τdeact are the time constants for activation
and deactivation, respectively (62). Excitation and activa-
tion levels in Eq. 1 are allowed to vary continuously be-
STAGE 1: CREATING A DYNAMIC tween zero (no excitation and activation) and one (full ex-
MODEL OF THE MUSCULOSKELETAL citation and activation). Activation levels serve as inputs
SYSTEM to the equations for musculotendon contraction dynamics
that estimate muscle forces.
In a muscle-driven dynamic simulation, elements of the Musculotendon contraction dynamics, the time course
neuromusculoskeletal system are modeled by sets of dif- of muscle force generation as determined by the energet-
ferential equations that describe muscle activation dynam- ics of cross bridge formation, the sliding of actin filaments,
ics, musculotendon contraction dynamics, musculoskele- and the dynamics of tendon, can be modeled by relating
tal geometry, and skeletal dynamics (Figure 12-3). These the time rate of change of muscle force ( f˙M ) to muscu-
equations characterize the time-dependent behavior of the lotendon length (l M T ), musculotendon shortening velocity
musculoskeletal system in response to neuromuscular ex- (l˙M T ), and muscle activation (a):
citation. Formulating these equations is the necessary first f˙M = f (l M T , l˙M T , a), (2)
stage in generating a simulation (see also Pandy (67) for a
review). where the function f (l M T , l˙M T , a) characterizes the force-
Muscles do not generate forces instantaneously in re- length-velocity properties of muscle and the force-length
sponse to neuromuscular excitation (see also Chapter 6). properties of tendon (66,107). The contraction dynamics
Muscle activation dynamics, the time course of Ca++ – of a particular muscle can be estimated by scaling the
mediated activation of the contractile apparatus as modu- function by five parameters: maximum isometric muscle

FIGURE 12-3. Elements of a muscle-driven simulation of walking. When a dynamic musculoskeletal


model (shaded region) is driven by muscle excitations (u ), the resulting motions of the body, (q¨ , q˙ , q)
can be calculated. The equation for muscle activation dynamics relates the muscle excitations to muscle
activations (a). The  equations for musculotendon contraction dynamics relate the muscle activations
to muscle forces ( f ); the muscle forces depend on musculotendon lengths and shortening velocities
  ˙ 
( l , l ). The muscle forces are transmitted to the skeleton resulting in joint torques (), which accelerate
the limb segments according to the equations of motion that govern the multi-joint dynamics of the
body. These accelerations (q¨ ) depend on the mass and inertial characteristics of the limb segments,
Coriolis and centrifugal forces, gravity, muscle forces, and any external forces produced when the body
interacts with the environment. Equations 1–3 are numerically integrated forward in time to produce a
simulation.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

Chapter 12 / Simulation of Walking 197

force and the corresponding muscle fiber length, max-

imum shortening velocity, pennation angle, and tendon
slack length (i.e., the length at which tendon begins to
transmit force when stretched). When applied to the skele-
ton, musculotendon forces act to accelerate the body
segments.
The accelerations of the body segments in response to
muscle forces and other loads can be computed from a
description of the musculoskeletal geometry and the equa-
tions of motion that govern the multi-joint dynamics of the
body:
¨ ↔−1 
     
˙ ↔   
˙
 
q = I (q) · {G(q) + C (q, q) + R(q) · f M + F E (q, q)}.
(3)
˙
  ¨
In this set of equations, q, q, and q are the generalized co-
ordinates (e.g., joint angles), velocities, and accelerations
↔−1
of the model, respectively. I is the inverse of the system
mass matrix, which specifiesthe mass and inertial prop-
erties of the body

segments. G is a vector of forces arising
from gravity. C is a vector of forces arising from Corio-
lis and centrifugal forces, which are generated from the
rotations of the body segments (e.g., like the forces en-
countered when swinging
↔
an object around in circles at FIGURE 12-4. Skeletal degrees of freedom of the model. The
the end of a rope). R is a matrix of muscle moment arms, head, arms, and torso were represented as a single rigid segment

and fM is a vector of muscle forces. The muscle moment that articulated with the pelvis via a ball-and-socket joint located
at approximately the third lumbar vertebra. Each hip was mod-
arms, or lever arms, determine the joint moments

that re- eled as a ball-and-socket joint, each knee as a hinge joint, each
sult from application of the muscle forces. FE is a vector ankle-subtalar joint as a universal joint, and each metatarsal joint
of external forces that characterize the interactions with as a hinge joint. Figure adapted from Anderson and Pandy (7).
the environment.
We have used this general approach to create a model
to simulate walking (4,5). The skeleton was represented as time constants of 22 and 200 msec, respectively (66,107).
a 10 segment, 23 degree-of-freedom linkage (Figure 12-4). Musculotendon contraction dynamics were described us-
The pelvis was modeled as a rigid segment that was free ing Eq. 2 with musculotendon parameters based on data
to translate and rotate with respect to the ground. The re- reported in the literature (18,25,31,97,107). Values of max-
maining nine segments branched in an open chain from imum isometric force and tendon slack length were ad-
the pelvis. Segment masses, centers of masses, and mo- justed for each actuator so that the isometric torque-angle
ments of inertia were derived based on the regression equa- curves for each joint in the model approximated the av-
tions reported by McConville et al (52) and anthropomet- erage torque-angle curves measured for five subjects (4).
ric measures obtained from five healthy adult males. The The model was completed when Eqs. 1–3 were formulated
equations of motion (Eq. 3) for the model were generated for the 54 musculotendon actuators and the 10 segment,
using a software package called SD/Fast (PTC, Needham, 23 degree-of-freedom linkage.
Massachusetts). Interactions between the feet and the
ground were characterized using a series of spring-damper

units distributed under the sole of each foot (i.e., FE in STAGE 2: GENERATING A
Eq. 3). Torques representing the actions of ligaments were MUSCLE-DRIVEN SIMULATION
applied to the joints to prevent the joints from hyperex- OF WALKING
tending during a simulation.
The model was driven by 54 musculotendon actuators. To generate a simulation of movement, the differential
The path (i.e., line of action) of each actuator was speci- equations of a dynamic model are integrated forward in
fied based on anatomical landmarks (25). When an actu- time (Figure 12-3). If the simulation is driven by muscles,
ator wrapped around a bone or other muscles, via points a set of muscle excitations (i.e., u in Eq. 1) must be ap-
and/or via cylinders were introduced to represent the path plied. A set of initial state variables (i.e., joint angles, joint
more accurately (26,35). Muscle activation dynamics were angular velocities, muscle forces, and muscle activations
described using Eq. 1 with activation and deactivation at the initial time step) must also be specified. Integration
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

198 Chapter 12 / Simulation of Walking

of the equations yields the time histories of all state vari- of the model (16). To enforce repeatability of the gait cycle,
ables in the model, including the muscle activations, mus- we specified a number of terminal constraints. Specifically,
culotendon forces, and joint angles. the values of the joint angular displacements, joint angu-
Finding a set of muscle excitations that produces a co- lar velocities, muscle forces, and muscle activations at the
ordinated movement can be challenging. This is especially end of the simulation were required to be the same as the
true for a movement as complex as walking. Not only values at the beginning. The values of the state variables
must many degrees of freedom be controlled (e.g., 23 in at the beginning of the simulation were based on averaged
the case of the musculoskeletal model described above), experimental data.
but also the time-dependent, nonlinear force-generating Thus, the dynamic optimization problem that we solved
properties of muscle must be taken into account. To was to find time histories of the muscle excitations that
meet this challenge, dynamic optimization can be used minimized J (Eq. 4) and met the constraints imposed
(e.g., 4,5,23,44,55,56,65,103). to enforce repeatability. To solve this problem, we imple-
Dynamic optimization is a mathematical approach for mented a parameter optimization algorithm on parallel
finding a set of control values (e.g., the time histories of supercomputers (8,62). We applied the resulting muscle
muscle excitations) for a dynamic system (e.g., a dynamic excitations to the dynamic model, integrated the equations
musculoskeletal model) that minimizes or maximizes a of motion forward in time, and generated a simulation of
time-dependent performance criterion, possibly subject to walking. The locomotor pattern predicted by the optimal
constraints. To simulate walking or other multi-joint move- solution (Figure 12-5) successfully reproduced the salient
ments, a variety of performance criteria can be formulated. features of normal gait.
One approach is to solve an optimal tracking problem (e.g.,
23,56). In this approach, the performance criterion is spec-
ified based on the difference between simulated and experi- STAGE 3: TESTING THE ACCURACY
mentally determined quantities, such as joint angles, joint OF DYNAMIC SIMULATIONS
powers, and ground reaction forces. By minimizing the
performance criterion over the period of the simulation, Before a simulation of movement can be analyzed, the sim-
the model is driven explicitly to reproduce experimental ulation and the underlying model should be tested. In par-
data. In other formulations, a set of muscle excitations ticular, it is important to verify that the dynamic behavior
might be found that minimizes or maximizes the model’s of the neuromusculoskeletal system is represented with
performance of some hypothesized motor goal. sufficient fidelity to answer the research questions posed.
We have used this goal-based approach, in combination For example, if the actions of individual muscles are of in-
with the musculoskeletal model described above, to gener- terest, the model should accurately characterize the mus-
ate a muscle-driven simulation of walking (5). We hypoth- cle moment arms for the ranges of body positions assumed
esized that the locomotor patterns of healthy individuals by the model during the simulation. Confidence in the mus-
result from minimizing metabolic energy expenditure per cle moment arms can be gained by comparing the mo-
unit distance traveled. This hypothesis is supported by ment arms predicted by the model to the moment arms
measurements of metabolic energy consumption and ob- determined experimentally from image data (e.g., 43) or
servations of preferred walking speeds (76). The perfor- from cadaveric specimens (e.g., 13,24). Confidence in the
mance criterion (J ) was therefore formulated as follows: moment-generating capacities of muscles can be gained by
comparing the maximum isometric joint moments gener-

tf ated by the model to the moments generated by human
Ė M
0
total subjects (e.g., 4).
J = + penalty terms, (4) A simulation of movement can be tested by comparing
X cm (tf ) − X cm (ti )
quantities predicted by the simulation to quantities deter-
M
where Ėtotal is the rate at which total metabolic energy is mined experimentally in the laboratory. For instance, the
consumed in the model and X cm (ti ) and X cm (tf ) denote muscle excitation patterns, joint angles, joint moments,
the position of the model’s center of mass at the initial and and ground reaction forces from a simulation of walking
final times of the simulated gait cycle, respectively. The can be compared to EMG, kinematic, and kinetic data ob-
penalty terms were appended to increase the value of the tained from gait analysis. The similarity between predicted
performance criterion if any of the joints hyperextended and measured values of metabolic energy consumption
M
during a simulation. Ėtotal was computed by adding the can also be assessed.
basal metabolic heat rate of the whole body to the acti- We have made detailed comparisons of the walking sim-
vation heat rate, maintenance heat rate, shortening heat ulation, described above, to experimental data from five
rate, and the mechanical work rate of each muscle in the subjects (5). In most cases, the joint angles predicted by
model. These rate terms were computed from the muscle the simulation were within one standard deviation of the
activations, forces, shortening velocities, and other states joint angles measured for the subjects. The simulated and
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

Chapter 12 / Simulation of Walking 199

FIGURE 12-5. Snapshots of the muscle-driven simulation of walking obtained by solving a dynamic
optimization problem.

measured ground reaction forces and muscle excitations influence the motions of the body segments over time. We
also compared favorably (e.g., Figure 12-6). These tests used this method to determine which muscles have the
provide some confidence that our simulation of walking greatest potential to diminish knee flexion velocity prior
reproduces the dynamics of normal gait (5). to toe-off, a possible cause of stiff-knee gait. All three ex-
amples were generated using the dynamic optimization
solution for walking described above.
STAGE 4: ANALYZING A SIMULATION
OF WALKING Example 1: Decomposition of the Ground Reaction
Force to Quantify Contributions to Support
Once a muscle-driven simulation of movement is gen- The vertical ground reaction force is measured rou-
erated and tested, the simulation can be analyzed in tinely, and its characteristic shape is well known for normal
several ways. Qualitatively, the motions produced from walking (Figure 12-6A). Because achieving adequate verti-
a particular set of muscle excitations can be visualized. cal support is one of the basic requirements of walking
Quantitatively, the contributions of individual muscles to (42), quantifying how muscles contribute to the vertical
the joint moments, joint angular accelerations, ground re- ground reaction force, and therefore to the vertical accel-
action forces, segmental energies, and other variables of eration of the center of mass, is an important part of our
interest can be determined. It is through the rigorous anal- basic understanding of gait mechanics.
ysis and interpretation of such data that the value of a sim- Several studies have examined how the shape of the
ulation can be realized. ground reaction force is influenced by muscle activity
The following examples demonstrate how a dynamic (48,53,56,63,64,68,72,88,100). There has been broad con-
simulation can be analyzed to extract information about sensus that the second maximum observed in the vertical
the actions of individual muscles during movement. Ex- ground reaction force is due largely to the forces exerted
ample 1 describes a method for decomposing the ground by the plantarflexors during late stance. However, an expla-
reaction force. We used this method to determine which nation for the shape of the ground reaction force during
muscles contribute to the vertical ground reaction force, early stance and midstance has been less definitive. Mus-
and therefore to the vertical support of the body, during cles that potentially contribute have been inferred from
walking. Example 2 describes a technique for calculating experiments based on similarities between net joint mo-
the instantaneous angular accelerations of the joints in- ments and the shape of the ground reaction force (100),
duced by individual muscles during movement. We used changes in the ground reaction force after the administra-
this technique to determine which muscles are responsi- tion of nerve blocks (88), and the timing of muscle activ-
ble for generating knee extension during the single-limb ity (72). However, without a theoretical framework for at-
stance phase. Example 3 describes a method for quantify- tributing portions of the ground reaction force to the forces
ing how individual muscles (or other elements of a model) produced by individual muscles, determining the relative
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

200 Chapter 12 / Simulation of Walking

FIGURE 12-6. Tests of the muscle-driven simulation of walking. (A) The vertical ground reaction force
generated during the simulation (black line) is representative of the ground reaction forces measured
from five subjects during normal gait (gray lines). (B) The muscle excitation histories predicted by
the dynamic optimization solution (black lines) compare favorably with EMG data recorded from one
subject (gray lines) and with EMG data reported in the literature (horizontal gray bars) (75). The marked
kinematic events are contralateral toe-off (CTO), contralateral heel-strike (CHS), toe-off (TO), and heel-
strike (HS). (Figure adapted from Anderson and Pandy [5].)

contributions of muscles to the shape of the ground reac- ity, ligament, and muscle forces that are applied:
tion force is difficult.
A number of studies (48,53,56) have used models of gait fE = fEC + fEG + fEL + fEM , (5)
dynamics to estimate how muscles and other sources of
force, such as gravity, contribute to the ground reaction where fEC , fEG , fEL , and fEM are the contributions made to
force. This example summarizes the analyses we have per- the ground reaction force by the Coriolis and centrifugal,
formed to identify which sources of force make the largest gravity, ligament, and muscle forces, respectively. During
contributions to the vertical ground reaction force, and walking, deformations of the foot and ground are small
therefore to support of the body during walking (6). (i.e., typically much less than 1 cm), and fEC , fEG, fEL, and
When interpreting experimental data obtained from fEM can be computed by assuming rigid contact between
gait analysis, it is often appealing to think of the ground each foot and the ground. That is, when a foot contacts
reaction force as existing of its own accord, as a force to be the ground, the component of the ground reaction force
balanced by net joint moments. Physically, however, the re- caused by a particular action force can be estimated by
verse is the case. “Reaction” forces, including the ground applying that action force to the model in isolation and cal-
reaction force, arise as a consequence of “action” forces culating the reaction force needed to prevent the portion
that act on and within the body. When muscles or other of the foot in contact with the ground from accelerating.
sources of force act on the body, they accelerate each foot The reaction force induced by a muscle, fEMi, for example,
with respect to the ground. Over time, these accelerations can be computed as the force necessary to prevent the foot
result in motion. When this motion brings a foot into con- from accelerating when muscle Mi alone acts on the body.
tact with the ground, the foot and the ground deform (e.g., We have used this method, in conjunction with our walk-
the compression of the sole of a shoe, or the depression of ing simulation, to generate a decomposition of the vertical
a wooden floor). Reaction forces are generated as a result ground reaction force throughout the stance phase (6).
of these deformations. Muscles made the largest contribution to vertical sup-
A muscle-driven dynamic simulation enables the cause- port in our simulation, accounting for 50% to 95% of the
effect relationships between action forces, such as the vertical ground reaction force generated in stance (Figure
forces produced by muscles, and reaction forces, such as 12-7A, Muscle+Ligaments). The hip and knee extensors
the ground reaction force, to be established. Specifically, (gluteus maximus, posterior portion of gluteus medius,
the ground reaction force generated during a simulation, and vasti) were the main contributors to support in early
fE , can be attributed to the Coriolis and centrifugal, grav- stance, although prior to foot-flat the ankle dorsiflexors
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

Chapter 12 / Simulation of Walking 201

made important contributions as well (Figure 12-7B). The

hip abductors (anterior and posterior portions of gluteus
medius and minimus) were the main contributors to sup-
port in midstance (Figure 12-7C). The ankle plantarflex-
ors (primarily soleus and gastrocnemius) were the main
contributors to support in late stance (Figure 12-7D). The
passive transmission of force through the joints and bones
in resistance to gravity accounted for 20% to 50% of the
ground reaction force when the foot was flat on the ground,
but made much smaller contributions before foot-flat and
A
after heel-off (Figure 12-7A, Gravity). Coriolis and centrifu-
gal forces acted to reduce the ground reaction force, but
these contributions were far less than the contributions
made by muscles or gravity (Figure 12-7A, Coriolis and
Centrifugal).
Thus, by analyzing the cause-effect relationships be-
tween muscle forces and ground reaction forces during
walking, a quantitative picture of how muscles contribute
to the shape of the vertical ground reaction force has
emerged. The first maximum in the ground reaction force B
results from the actions of the hip and knee extensors;
the second maximum from the ankle plantarflexors. These
muscles, along with the hip abductors, are the predomi-
nant contributors to vertical support during normal gait.
The skeleton offers some support in resistance to gravity,
but this support is less than 50% of body weight.
The methodology described in this example has numer-
ous applications. By analyzing a simulation of a patient’s
abnormal gait, for instance, it may be possible to link devi-
C
ations in the ground reaction force to the actions of partic-
ular muscles, perhaps identifying the cause of the patient’s
gait abnormality. In addition, as described in the next ex-
ample, performing a decomposition of the ground reaction
force is necessary for calculating the accelerations of the
joints induced by individual muscles.

Example 2: Induced Acceleration Analysis to Quan-

tify Muscle Actions in Single-Limb Stance
Children with cerebral palsy frequently walk with ex-
cessive knee flexion during the stance phase. This move- D

ment abnormality, called crouch gait, is problematic be-

cause it increases patellofemoral force (69), impedes toe
FIGURE 12-7. Decomposition of the ground reaction force dur-
clearance, and dramatically increases the energy require- ing normal gait. (A) Contributions made to the ground reaction
ments of walking (19,81,87). The persistent knee flexion force by muscle and ligament forces (Muscles + Ligaments),
typically worsens if left uncorrected (32,89) and can lead Coriolis and centrifugal forces (Coriolis and Centrifugal), and
to altered patellofemoral joint mechanics and chronic knee resistance to gravity provided by the bones and joints of the skele-
pain (17,50,82,90). ton (Gravity). Contributions made to the ground reaction force
(B) during early stance by the dorsiflexors (DF), vasti (VAS), glu-
Unfortunately, the biomechanical causes of crouch gait teus maximus (GMAX), the posterior portion of gluteus medius
are often unclear, making it challenging to determine the and minimus (GMEDP), (C) during mid-stance by the posterior
most appropriate treatment. In some cases, abnormally and anterior portions of gluteus medius and minimus (GMEDA
“short” or “spastic” hamstrings are presumed to limit knee and GMEDP, respectively), and (D) during late stance by soleus
extension, and surgical lengthening of the hamstrings is (SOL), gastrocnemius (GAS), and the ligaments (Ligaments). The
marked kinematic events are heel-strike (HS), foot-flat (FF), con-
performed (1,17,29,70). In other cases, diminished plan- tralateral toe-off (CTO), heel-off (HO), contralateral heel-strike
tarflexion strength is thought to be a factor, and ankle-foot (CHS), metatarsal-off (MO), and toe-off (TO). (Figure adapted
orthoses are prescribed (34,79). Other hypothesized causes from Anderson and Pandy [(6]).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

202 Chapter 12 / Simulation of Walking

of crouch gait include malrotation of the femur, tibia, and

foot (34,85), tight hip flexors (17,60,77,80), weak hip ex-
tensors (98), weak knee extensors (15,22,34), and poor bal-
ance (32).
Successful treatment of crouch gait is difficult, in part,
because the factors that contribute to knee extension dur-
ing normal gait are not well understood, and because the
potential of individual muscles to produce knee flexion or
extension during the stance phase is unknown. Knee mo-
tions are influenced not only by muscles that cross the
knee, but also by muscular moments that are generated at
other joints. To understand how knee extension is achieved
during normal gait, and to elucidate the factors that po-
tentially contribute to crouch gait, the effects of dynamic
coupling must be considered. That is, the intersegmen-
tal forces that arise from a muscle’s force must be com- FIGURE 12-8. Contributions of vasti (VAS), gluteus maximus
puted, and the resulting multi-joint accelerations of the (GMAX), soleus (SOL), and gravity to the angular accelerations of
body must be evaluated. the knee during the single-limb stance phase of normal gait. Grav-
We have used the simulation of walking, described ity, together with its contribution to the ground reaction force,
above, to quantify the angular accelerations of the knee in- accelerated the knee toward flexion. The effects of gravity were
resisted by muscles crossing the hip, knee, and ankle.
duced by the gluteus maximus, hamstrings, vasti, soleus,
and other muscles during single-limb stance (9). This ex-
ample illustrates how a simulation can be analyzed to evaluated the relative potential of a muscle to generate (or
reveal the dynamic actions of individual muscles during limit) knee extension if, for example, the muscle was acti-
movement. vated inappropriately or was producing excessive passive
We determined the muscle-induced accelerations of the force.
knee using the equations of motion for the model (Eq. 3) Examination of the muscle-induced accelerations
and the decomposition of the ground reaction force de- showed that the gluteus maximus, vasti, soleus, and poste-
scribed in Example 1. At each time step in the simulation, rior portion of the gluteus medius made substantial contri-
a muscle’s contributions to the instantaneous accelerations butions to knee extension during normal gait (Figure 12-8).
of the generalized coordinates (q¨ Mi ) were computed by ap- Per unit force, the gluteus maximus had greater potential
plying that muscle’s force ( fMi ), as generated during the than the vasti to accelerate the knee toward extension (Fig-
simulation, and the corresponding portion of the ground ure 12-9). These data suggest that diminished force in
reaction force induced by that muscle ( fEMi ): the hip extensors, knee extensors, or ankle plantarflexors
↔
↔ ↔

q ) · fMi + E(
q )−1 · R(
q¨ Mi = I( q ) · fEMi (6)

↔
All other forces in the model were set to zero. E is a ma-
trix that converts the foot-ground spring forces into gener-
alized forces (46).
Two descriptions of the muscle actions during single-
limb stance were examined. First, the angular accelera-
tions of the knee induced by individual muscles were quan-
tified to determine which of the muscles enabled knee
extension in the simulation of normal gait. Second, the
muscle-induced accelerations of the knee per unit force
were calculated to assess the “dynamic potential” of each
muscle to accelerate the knee toward flexion or extension.
This measure of a muscle’s actions (obtained by setting
fMi = 1N, computing the corresponding fEMi , and substi- FIGURE 12-9. Angular accelerations of the knee per unit force,
tuting these quantities into Eq. 6) does not depend on the averaged over the single-limb stance phase (17%–50% of the gait
cycle), induced by gluteus maximus (GMAX), vasti (VAS), soleus
muscle excitations or forces applied during the simulation; (SOL), biceps femoris short head (BFSH), iliopsoas (ILPS), sar-
rather, it reflects the influence of a muscle’s moment arms torius (SAR), rectus femoris (RF), hamstrings (HAMS), and gas-
and the inertial properties of the body. Hence, this analysis trocnemius (GAS).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

Chapter 12 / Simulation of Walking 203

may contribute to crouch gait, and strengthening these erations of the knee during double support were known,
muscles, particularly the gluteus maximus, may help to then perhaps treatments to correct stiff-knee gait could be
improve both hip and knee extension. Abnormal forces designed more effectively.
generated by spasticity or contracture of the iliopsoas may Comparisons of EMG recordings and measured gait
also cause crouch gait in some cases, since these muscles kinematics have suggested that gastrocnemius, popliteus,
have a large potential to accelerate the knee toward flexion and occasionally gracilis contribute to knee flexion during
(Figure 12-9). the late stance phase (72). Rectus femoris (72) and, in some
The potential of the biarticular hamstrings, rectus cases, vasti (101) are also active during this period and are
femoris, and gastrocnemius muscles to induce angu- thought to limit knee flexion. However, the potential of
lar accelerations of the knee during single-limb stance these and other muscles to produce knee flexion or exten-
was small relative to other muscles (Figure 12-9). This sion during walking cannot be deduced from kinesiologic
was caused by dynamic coupling. Each of these mus- observations alone. We have used the muscle-driven simu-
cles generated a moment about the knee and a moment lation of walking, described above, to identify the muscles
about an adjacent joint, and these moments induced that influence knee flexion velocity during double support
opposing accelerations of the knee. For example, the and to determine which muscles have the greatest poten-
knee flexion moment generated by hamstrings acted to tial to alter this velocity (37).
accelerate the knee toward flexion, but the hip exten- In a muscle-driven simulation, joints are accelerated
sion moment generated by hamstrings acted to acceler- because of muscle forces. If a muscle’s force is altered,
ate the knee toward extension. During the stance phase, or “perturbed,” by a small amount during a simulation,
in fact, the hamstrings had the potential to weakly acceler- the resulting changes in the motions of the joints can be
ate the knee toward extension in our model. This occurred quantified (i.e., by reintegrating the equations of motion
because the hamstrings’ hip extension moment accelerated forward in time). This technique, called perturbation anal-
the knee toward extension more than the hamstrings’ knee ysis, is useful for investigating how individual muscles or
flexion moment accelerated the knee toward flexion. This other elements of a model influence the angular displace-
unexpected result suggests that abnormally short or spas- ments and velocities of the joints. We used this technique
tic hamstrings, a reputed cause of crouch gait, may not be in this example.
the direct source of excessive knee flexion in some patients. We quantified the actions of individual muscles in
Our analysis of the muscle-induced accelerations of the our simulation by systematically perturbing each mus-
knee, as described in this example, has clarified some of the cle’s force in double support and calculating the resulting
actions of muscles during walking and has identified fac- changes in peak knee flexion velocity (Figure 12-10). We
tors that are likely to contribute to excessive knee flexion altered the muscle forces in two ways. First, we increased
in persons with cerebral palsy. This work emphasizes the the force in each muscle by a percentage of the muscle’s
need to consider how muscular forces contribute to multi- unperturbed force. The resulting change in knee flexion
joint movement when attempting to identify the causes of velocity depended on the muscle’s unperturbed force and
a patient’s abnormal gait. Another method for characteriz- characterized how much that muscle’s force influenced
ing the actions of muscles during movement is described peak knee flexion velocity during the simulation. Second,
in the next example. we increased the force in each muscle by a fixed amount
in Newtons. Using this approach, the resulting change in
Example 3: Perturbation Analysis to Quantify Muscle knee flexion velocity per unit force was independent of the
Actions in Double Support muscle’s unperturbed force, and characterized the poten-
Knee flexion velocity at toe-off is an important factor tial of the muscle to influence knee flexion velocity based
in generating swing-phase knee flexion during normal gait on the muscle’s moment arms and the inertial properties
(53,73). Low knee flexion velocity at toe-off is a potential of the body.
contributor to stiff-knee gait, a movement abnormality as- Analysis of the simulation revealed that iliopsoas and
sociated with stroke and cerebral palsy in which swing- gastrocnemius were the largest contributors to peak knee
phase knee flexion is diminished. Stiff-knee gait is com- flexion velocity during double support (Figure 12-11A).
monly attributed to excessive activity of the rectus femoris, Each of these muscles exerted relatively large forces in the
which is thought to limit knee flexion by producing an simulation, and each had a large potential to increase knee
excessive knee extension moment during swing (71,91). flexion velocity (Figure 12-11B). The forces generated by
However, we have shown that many individuals with stiff- vasti, soleus, and rectus femoris, by contrast, decreased
knee gait do not exhibit excessive knee extension moments knee flexion velocity (Figure 12-11A). Vasti decelerated
during swing phase, but instead walk with a low knee flex- knee flexion the most. This is because vasti had the largest
ion velocity at toe-off (38). During normal gait, just prior to potential to decrease knee flexion velocity (Figure 12-
toe-off, knee flexion velocity increases dramatically during 11B), and because these muscles developed passive forces
double support. If the muscles that produce angular accel- during double support. Soleus also exerted large forces
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

204 Chapter 12 / Simulation of Walking

FIGURE 12-10. Steps to assess muscle function in a perturbation analysis. (A) A perturbation in an
individual muscle force was introduced during the period of double support. (B) A dynamic simulation
was performed with this altered muscle force and the resulting change in peak knee flexion velocity
during double support was observed. (C) After repeating steps A) and B) for a range of perturbation
sizes, the change in peak knee flexion velocity was plotted vs. the size of the force perturbation. The
resulting slope was computed for different muscles to assess the relative influence of these muscles on
knee flexion velocity.

during double support, but had a small potential relative tus femoris could limit knee flexion velocity during double
to vasti to decrease knee flexion velocity (Figure 12-11B). support, and cause swing-phase knee flexion to be dimin-
The muscles with the greatest potential to increase knee ished. The stance-phase actions of these and other muscles,
flexion velocity during double support, per unit force, were such as gracilis, should be considered before performing
sartorius and gracilis (Figure 12-11B). When activated, muscle-tendon surgery to treat stiff-knee gait.
these muscles generate hip flexion moments and knee flex-
ion moments, both of which promote knee flexion. Biceps
femoris short head, a uniarticular knee flexor, also had a CHALLENGES AND FUTURE
relatively large potential to increase knee flexion velocity. DIRECTIONS
These analyses have advanced our understanding of
muscle function during normal gait, and have helped to The first muscle-driven simulation of walking was devel-
identify several possible causes of stiff-knee gait. In par- oped more than three decades ago by Chow and Jacobson
ticular, our results suggest that insufficient force in iliop- (21). Their dynamic model consisted of a single leg
soas or gastrocnemius, or excessive force in vasti or rec- confined to the sagittal plane that was driven by four

FIGURE 12-11. The muscles with the most influence and the most potential to influence peak knee
flexion velocity. (A) The influence of selected muscles on the peak knee flexion velocity during double
support. The influence was calculated as the slope of the change in peak knee flexion velocity (V) vs.
perturbation size as a percentage of unperturbed muscle force (F) throughout the period of double
support. (B) The potential influence of selected muscles on the peak knee flexion velocity during double
support. These values characterize the change in peak knee flexion velocity (V) due to a 1 N change in
muscle force (F) throughout the period of double support. (Figure adapted from Goldberg, et al.) (37).
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

Chapter 12 / Simulation of Walking 205

idealized muscles. Nearly two decades passed before Davy develop subject-specific models from magnetic resonance
and Audu (23) and then Yamaguchi and Zajac (104) pub- images or ultrasonography scans (e.g., 43,51,86,99). How-
lished significantly advanced muscle-driven simulations of ever, this approach may not always be practical. An alterna-
walking. Over the last few years, there has been a dramatic tive approach that combines medical images with generic
increase in the number and complexity of walking simu- musculoskeletal models, we believe, offers a promising,
lations, enabled, in large part, by new modeling software tractable way to construct models that are representative
and increases in computer speed. To date, simulations of of patients. For instance, it may be possible to transform
normal gait have been developed and analyzed to estimate a generic model to represent a range of individuals with
the forces produced by muscles (e.g., 7), to determine how cerebral palsy using multi-dimensional scaling techniques,
individual muscles support the body (56,58), accelerate algorithms for deforming bones, and a few subject-specific
the joints (44,45), and distribute energy among the limb parameters derived from image data or experimental mea-
segments (58,59), and to evaluate theories of neuromo- surements (11,20). We have begun to develop and evaluate
tor control (36,40,61,92,106). Muscle-driven simulations such models (10,11), and we believe that additional efforts
have also been created and used to evaluate exercise pro- are warranted.
tocols for persons with spinal cord injury (84) and patients The equations for musculotendon dynamics that we
with patellofemoral pain (57), to examine the influence of have used in simulations must be further tested. While ex-
foot positioning and joint compliance on the occurrence isting models capture many features of muscle force gen-
of ankle sprains (102), to assess computational prototypes eration in unimpaired subjects, they do not account for
of knee implants (74), and to investigate causes of stiff- adaptations that can occur in persons with neuromuscu-
knee gait (3,37,73,78). These studies, and the examples lar disorders or alterations that might occur after surgery.
presented in this chapter, demonstrate the utility of We have developed models that attempt to account for de-
muscle-driven simulations for elucidating the functions creases in the muscle fiber lengths that may occur with
of muscles during movement and, potentially, improving contracture (28). However, muscle-tendon models that ac-
the outcomes of treatments for persons with neuromuscu- count for structural changes in the extracellular matrix
loskeletal impairments. that may occur with chronic spasticity (49), or alterations
Although models of the musculoskeletal system have in force transmission due to scar tissue (14) are not yet
become more sophisticated and novel approaches for generally available. Muscle-tendon models that character-
analyzing simulations have been developed, rigorous ize the effects of pathology, surgery, and other treatment
dynamics-based techniques for determining which impair- modalities on the time course of muscle force generation
ments contribute to the abnormal gait patterns of persons are needed to assess the impact of these effects on move-
with neuromusculoskeletal disorders do not exist. We be- ment and neuromotor control.
lieve that the limitations of current models and analyses
must be addressed before simulations can be widely used
Simulation Challenges
to guide treatment decisions for patients. Some of the im-
portant issues to be resolved in future studies are outlined Using dynamic optimization to determine the muscle
below. excitation patterns needed to simulate complex three-
dimensional movements, such as walking, incurs great
computational expense. The dynamic optimization solu-
Modeling Challenges
tion for normal walking presented in this chapter required
Models that more accurately and efficiently characterize over 5000 computer processor hours to compute, and re-
the musculoskeletal geometry and the joint kinematics of lied heavily on the use of parallel supercomputers (5).
individual subjects need to be developed. This is impera- Although approaches for solving dynamic optimization
tive because the results of simulations are often sensitive problems are improving (41,47,54,96), this process is still
to the accuracy with which the lengths and moment arms very slow— at best, a solution for pathological gait might
of muscles can be estimated. Studies of muscle function be obtained in a few days or a week. If dynamic simula-
during walking have typically relied on “generic” models tions are to guide treatment decisions, then efficient com-
of adult subjects with normal musculoskeletal geometry. putational algorithms for generating subject-specific sim-
We have modified generic models to represent bone defor- ulations must be developed.
mities (10-12), osteotomies (30,83), and tendon transfer Fortunately, alternatives to dynamic optimization are
surgeries (27). However, more work is needed to under- emerging. Adaptations of traditional robotics control tech-
stand how variations in musculoskeletal geometry due to niques appear to be particularly promising (e.g., 93,95). A
size, age, deformity, or surgery might influence the predic- technique called computed muscle control, for example,
tions of a model, and to determine when, and under what has been used to generate a simulation of bicycle ped-
conditions, simulations based on generic models are ap- aling approximately 100 times faster than conventional
plicable to individual patients. One approach might be to dynamic optimization approaches (93). We believe that
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

206 Chapter 12 / Simulation of Walking

this technique will soon enable subject-specific simula- cially if a direct comparison with experimental data is not
tions of walking to be generated in a few minutes. feasible. Ultimately, controlled clinical studies are required
Perhaps the most profound limitation of the simula- to determine if the insights gained from simulations can
tion of walking described in this chapter is its exclusion of indeed improve treatment outcomes.
the nervous system. The simulation was performed open
loop; that is, the muscle excitations were not modulated
Experiments and Theory
by reflexes. This is true of most dynamic simulations of
movement generated to date, particularly those involving We believe that a comprehensive explanation for how mus-
the lower extremity. However, simulations that are driven cles are coordinated to produce gait is emerging. This be-
by neural networks and/or central pattern generators, that lief is supported by the convergence of findings between
include simplified models of the nervous system, are being some experimental and theoretical studies. For instance,
developed (e.g., 39,61,92). The incorporation of accurate it was no surprise in Example 1 that the plantarflexors are
representations of sensory-motor control (e.g., 94) into dy- largely responsible for the second maximum in the verti-
namic simulations of abnormal movements is one of the cal ground reaction force (Figure 12-7D); this was inferred
most critical challenges to be overcome if models are to be previously from several experimental studies, and a num-
developed that can predict the outcomes of treatments. ber of simulations have confirmed a cause-effect relation-
ship. Other results from our simulation-based analyses of
walking are more surprising. For example, our finding that
Analysis Challenges
hamstrings weakly accelerate the knee toward extension
Characterizing the functions of muscles and extracting the during stance (Example 2, Figure 12-9) was unexpected.
principles that govern muscle coordination from dynamic It is the unexpected findings from simulations, and the
simulations is nontrivial. Three approaches for quantify- reconciliation of these findings with what we believe we
ing the actions of muscles were illustrated in this chapter, already know from experiments, that have the potential to
but each of these analyses has limitations. Our method for deepen our understanding of walking.
decomposing the ground reaction force, as performed in With the development, analysis, and testing of muscle-
Example 1, assumes rigid contact between the foot and driven dynamic simulations, we are now in a position to es-
the ground. To the extent that contact is not rigid, a por- tablish quantitative, cause-effect relationships between the
tion (usually a small portion) of the reaction force cannot neuromuscular excitation patterns, muscle forces, ground
be explained (2). Calculations of muscle-induced acceler- reaction forces, and motions of the body that are observed
ations, as performed in Example 2, rely on an accurate in the laboratory. Coupled with high quality experimental
decomposition of the ground reaction force, and depend measurements, dynamic simulations will be used to elu-
on the degrees of freedom that are included (or not in- cidate how the elements of the neuromusculoskeletal sys-
cluded) in the model. Perturbation analysis, as performed tem interact to produce movement and, we hope, improve
in Example 3, provides an intuitive method for assessing the outcomes of treatments for persons with movement
how a muscle influences movement (i.e., not just accelera- disorders.
tions, but also velocities and positions) by altering muscle
force and quantifying the changes in movement over time.
However, the results of perturbation analyses can be mis- REFERENCES
leading if the perturbations in muscle force are too large or 1. Abel MF, Damiano DL, Pannunzio M, Bush J. Muscle-tendon
if the time windows over which perturbations are applied surgery in diplegic cerebral palsy: Functional and mechanical
are too long. We believe that improved methods for ana- changes. J Pediatr Orthop 1999;19:366–375.
2. Anderson FC, Arnold AS, Delp SL. Reaction forces induced by mus-
lyzing simulations will evolve by applying these and other cles in the presence of compliant contact. Proceedings from the 4th
techniques to a range of research problems and critically World Congress of Biomechanics, 2002.
evaluating the results. 3. Anderson FC, Goldberg SR, Pandy MG, Delp SL. Contributions of
muscle forces and toe-off kinematics to peak knee flexion during the
Experimental observations and a theoretical frame- swing phase of normal gait: An induced position analysis. J Biomech
work are needed to establish a scientific basis for the 2004;37:731–737.
treatment of gait abnormalities. Before a simulation of 4. Anderson FC, Pandy MG. A dynamic optimization solution for verti-
cal jumping in three dimensions. Comput Methods Biomech Biomed
walking can be used to make treatment decisions, the Engin 1999;2:201–231.
simulation and the underlying model must be tested. If 5. Anderson FC, Pandy MG. Dynamic optimization of human walking.
possible, simulation results should be compared with ex- J Biomech Eng 2001;123:381–390.
6. Anderson FC, Pandy MG. Individual muscle contributions to sup-
perimental data to verify that a particular simulation is port in normal walking. Gait Posture 2003;17:159–169.
of sufficient fidelity to answer the clinical question being 7. Anderson FC, Pandy MG. Static and dynamic optimization solu-
posed. Sensitivity studies must be performed to determine tions for gait are practically equivalent. J Biomech 2001;34:153–161.
8. Anderson FC, Ziegler JM, Pandy MG, Whalen RT. Application of
whether the conclusions drawn from analysis of a simula- high-performance computing to numerical simulation of human
tion are sensitive to variations in model parameters, espe- movement. J Biomech Eng 1995;117:155–157.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

Chapter 12 / Simulation of Walking 207

9. Arnold AS, Anderson FC, Pandy MG, Delp SL. Muscular contribu- 34. Gage JR. Treatment principles for crouch gait. In: Gage JR, ed. The
tions to hip and knee extension during the stance phase of normal Treatment of Gait Problems in Cerebral Palsy. London: Mac Keith
gait: A framework for investigating the causes of crouch gait. In Press, 2004;382–397.
press. 35. Garner BA, Pandy MG. The obstacle-set method for representing
10. Arnold AS, Blemker SS, Delp SL. Evaluation of a deformable mus- muscle paths in musculoskeletal models. Comput Methods Biomech
culoskeletal model for estimating muscle-tendon lengths during Biomed Engin 2000;3:1–30.
crouch gait. Ann Biomed Eng 2001;29:263–274. 36. Gerritsen KG, van den Bogert AJ, Hulliger M, Zernicke RF. Intrinsic
11. Arnold AS, Delp SL. Rotational moment arms of the medial ham- muscle properties facilitate locomotor control: A computer simula-
strings and adductors vary with femoral geometry and limb po- tion study. Motor Control 1998;2:206–220.
sition: Implications for the treatment of internally-rotated gait. J 37. Goldberg SR, Anderson FC, Pandy MG, Delp SL. Muscles that in-
Biomech 2001;34:437–447. fluence knee flexion velocity in double support: Implications for
12. Arnold AS, Komattu AV, Delp SL. Internal rotation gait: A compen- stiff-knee gait. J Biomech 2004;37:1189–1196.
satory mechanism to restore abduction capacity decreased by bone 38. Goldberg SR, Ounpuu S, Delp SL. The importance of swing
deformity? Dev Med Child Neurol 1997;39:40–44. phase initial conditions in stiff-knee gait. J Biomech 2003;36:1111–
13. Arnold AS, Salinas S, Asakawa DJ, Delp SL. Accuracy of muscle 1116.
moment arms estimated from MRI-based musculoskeletal models 39. Hase K, Miyashita K, Ok S, Arakawa Y. Human gait simulation with
of the lower extremity. Comput Aided Surg 2000;5:108–119. a neuromusculoskeletal model and evolutionary computation. J Vis
14. Asakawa DJ, Blemker SS, Rab G, Bagley A, Delp SL. Three dimen- Comput Animat 2003;14:73–92.
sional muscle tendon geometry after rectus femoris transfer. J Bone 40. Hase K, Yamazaki N. Computer simulation study of human lo-
Joint Surg Am 2004;86–A:348–354. comotion with a three-dimensional entire-body neuro-musculo-
15. Beals RK. Treatment of knee contracture in cerebral palsy by ham- skeletal model (I. Acquisition of normal walking). JSME Int J Ser
string lengthening, posterior capsulotomy, and quadriceps mecha- C 2002;45:1040–1072.
nism shortening. Dev Med Child Neurol 2001;43:802–805. 41. Higginson JS, Neptune RR, Anderson FC. Simulated parallel an-
16. Bhargava LJ, Pandy MG, Anderson FC. A phenomenological model nealing within a neighborhood for optimization of biomechanical
for estimating metabolic energy consumption in muscle contrac- systems. J Biomech 2005;38:1938–1942.
tion. J Biomech 2004;37:81–88. 42. Inman VT, Ralston HJ, Todd F. Human Walking. Baltimore: Williams
17. Bleck EE. Orthopaedic Management in Cerebral Palsy. London: Mac & Wilkins; 1981.
Keith Press, 1987. 43. Ito M, Akima H, Fukunaga T. In vivo moment arm determination
18. Brand RA, Pedersen DR, Friederich JA. The sensitivity of muscle using b-mode ultrasonography. J Biomech 2000;33:215–218.
force predictions to changes in physiologic cross-sectional area. 44. Jonkers I, Stewart C, Spaepen A. The complementary role of the
J Biomech 1986;19:589–596. plantarflexors, hamstrings and gluteus maximus in the control
19. Campbell J, Ball J. Energetics: Application to the study and man- of stance limb stability during gait. Gait Posture 2003;17:264–
agement of locomotor disabilities. Energetics of walking in cerebral 272.
palsy. Orthop Clin North Am 1978;9:374–377. 45. Jonkers I, Stewart C, Spaepen A. The study of muscle action during
20. Chao EYS, Lynch JD, Vanderploeg MJ. Simulation and anima- single support and swing phase of gait: Clinical relevance of forward
tion of musculoskeletal joint system. J Biomech Eng 1993;115:562– simulation techniques. Gait Posture 2003;17:97–105.
568. 46. Kane T, Levinson D. Dynamics: Theory and application. New York:
21. Chow C, Jacobson D. Studies of human locomotion via optimal McGraw-Hill; 1985.
programming. Math Biosci 1971;10:239–306. 47. Kaplan ML, Heegaard JH. Predictive algorithms for neuromuscular
22. Damiano DL, Vaughan CL, Abel MF. Muscle response to heavy re- control of human locomotion. J Biomech 2001;34:1077–1083.
sistance exercise in children with spastic cerebral palsy. Dev Med 48. Kepple TM, Siegel KL, Stanhope SJ. Relative contributions of the
Child Neurol 1995;37:731–739. lower extremity joint moments to forward progression and support
23. Davy DT, Audu ML. A dynamic optimization technique for predict- during gait. Gait Posture 1997;6:1–8.
ing muscle forces in the swing phase of gait. J Biomech 1987;20:187– 49. Lieber RL, Steinman S, Barash IA, Chambers H. Structural and
201. functional changes in spastic skeletal muscle. Muscle Nerve 2004;29:
24. Delp SL, Hess WE, Hungerford DS, Jones LC. Variation of rotation 615–627.
moment arms with hip flexion. J Biomech 1999;32:493–501. 50. Lloyd-Roberts GC, Jackson AM, Albert JS. Avulsion of the distal pole
25. Delp SL, Loan JP, Zajac FE, Topp EL, Rosen JM. An interactive of the patella in cerebral palsy. J Bone Joint Surg Br 1985;67B:252–
graphics-based model of the lower extremity to study orthopaedic 254.
surgical procedures. IEEE Trans Biomed Eng 1990;37:757– 51. Maganaris CN. In vivo measurement-based estimations of the mo-
767. ment arm in the human tibialis anterior muscle-tendon unit. J
26. Delp SL, Loan JP. A computational framework for simulation and Biomech 2000;33:375–379.
analysis of human and animal movement. IEEE Comput Sci Eng 52. McConville JT, Clauser CE, Churchill TD, Cuzzi J, Kaleps I. An-
2000;2:46–55. thropometric Relationships of Body and Body Segment Moments of
27. Delp SL, Ringwelski DA, Carroll NC. Transfer of the rectus femoris: Inertia. Report AFAMRL-TR-80-119. Air Force Aerospace Medical
Effects of transfer site on moment arms about the knee and hip. J Research Laboratory, Wright-Patterson AFB, Ohio, 1980.
Biomech 1994;27:1201–1211. 53. Mochon S, McMahon TA. Ballistic walking: An improved model.
28. Delp SL, Statler K, Carroll NC. Preserving plantar flexion strength Math Biosci 1980;52:241–260.
after surgical treatment for contracture of the triceps surae: A com- 54. Neptune R. Optimization algorithm performance in determining
puter simulation study. J Orthop Res 1995;13:96–104. optimal controls in human movement analyses. J Biomech Eng
29. DeLuca PA, Ounpuu S, Davis RB, Walsh JH. Effect of hamstring 1999;121:249–252.
and psoas lengthening on pelvic tilt in patients with spastic diplegic 55. Neptune RR, Hull ML. Evaluation of performance criteria for simu-
cerebral palsy. J Pediatr Orthop 1998;18:712–718. lation of submaximal steady-state cycling using a forward dynamic
30. Free SA, Delp SL. Trochanteric transfer in total hip replacement: model. J Biomech Eng 1998;120:334–341.
Effects on the moment arms and force-generating capacities of the 56. Neptune RR, Kautz SA, Zajac FE. Contributions of the individual
hip abductors. J Orthop Res 1996;14:245–250. ankle plantar flexors to support, forward progression and swing
31. Friederich JA, Brand RA. Muscle fiber architecture in the human initiation during walking. J Biomech 2001;34:1387–1398.
lower limb. J Biomech 1990;23:91–95. 57. Neptune RR, Kautz SA. Knee joint loading in forward versus
32. Gage JR, Schwartz M. Pathological gait and lever-arm dysfunction. backward pedaling: Implications for rehabilitation strategies. Clin
In: Gage JR, ed. The Treatment of Gait Problems in Cerebral Palsy. Biomech 2000;15:528–535.
London: Mac Keith Press, 2004;180–204. 58. Neptune RR, Zajac FE, Kautz SA. Muscle force redistributes seg-
33. Gage JR. Gait Analysis in Cerebral Palsy. London: Mac Keith Press, mental power for body progression during walking. Gait Posture
1991. 2004;19:194–205.
P1: NDK/OVY P2: NDK/OVY QC: NDK/OVY T1: NDK
GRBT092-12 Rose- 2252G GRBT092-Rose-v2.cls October 28, 2005 9:58

208 Chapter 12 / Simulation of Walking

59. Neptune RR, Zajac FE, Kautz SA. Muscle mechanical work require- 84. Schutte LM, Rodgers MM, Zajac FE. Improving the efficacy of
ments during normal walking: The energetic cost of raising the electrical stimulation-induced leg cycle ergometry: An analysis
body’s center-of-mass is significant. J Biomech 2004;37:817–825. based on a dynamic musculoskeletal model. IEEE Trans Rehab Eng
60. Novacheck T, Trost J, Schwartz M. Intramuscular psoas lengthen- 1993;1:109–125.
ing improves dynamic hip function in children with cerebral palsy. 85. Schwartz M, Lakin G. The effect of tibial torsion on the dy-
J Pediatr Orthop 2002;22:158–164. namic function of the soleus during gait. Gait Posture 2003;17:113–
61. Ogihara N, Yamazaki N. Generation of human bipedal locomo- 118.
tion by a bio-mimetic neuro-musculo-skeletal model. Biol Cybern 86. Sheehan FT, Zajac FE, Drace JE. Using cine phase contrast magnetic
2001;84:1–11. resonance imaging to non-invasively study in vivo knee dynamics.
62. Pandy MG, Anderson FC, Hull DG. A parameter optimization ap- J Biomech 1998;31:21–26.
proach for the optimal control of large-scale musculoskeletal sys- 87. Stout JL, Koop S. Energy expenditure in cerebral palsy. In: Gage JR,
tems. J Biomech Eng 1992;114:450–460. ed. The Treatment of Gait Problems in Cerebral Palsy. London: Mac
63. Pandy MG, Berme N. Quantitative assessment of gait determinants Keith Press, 2004;146–164.
during single stance via a three-dimensional model: Part 1. Normal 88. Sutherland DH, Cooper L, Daniel D. The Role of the Ankle Plantar
gait. J Biomech 1989;22:717–724. Flexors in Normal Walking. J Bone Joint Surg Am 1980;62-A:354–
64. Pandy MG, Berme N. Quantitative assessment of gait determinants 363.
during single stance via a three-dimensional model: Part 2. Patho- 89. Sutherland DH, Cooper L. The pathomechanics of progressive
logical gait. J Biomech 1989;22:725–733. crouch gait in spastic diplegia. Orthop Clin North Am 1978;9:143–
65. Pandy MG, Garner BA, Anderson FC. Optimal control of non- 154.
ballistic muscular movements: A constraint-based performance cri- 90. Sutherland DH, Davids JR. Common gait abnormalities of the knee
terion for rising from a chair. J Biomech Eng 1995;117:15–26. in cerebral palsy. Clin Orthop 1993;288:139–147.
66. Pandy MG, Zajac FE, Sim E, Levine WS. An optimal control model 91. Sutherland DH, Santi M, Abel MF. Treatment of stiff-knee gait in
for maximum-height human jumping. J Biomech 1990;23:1185– cerebral palsy: A comparison by gait analysis of distal rectus femoris
1198. transfer versus proximal rectus release. J Pediatr Orthop 1990; 10:
67. Pandy MG. Computer modeling and simulation of human move- 433–441.
ment. Ann Rev Biomed Eng 2001;3:245–273. 92. Taga G. A model of the neuro-musculo-skeletal system for anticipa-
68. Pandy MG. Simple and complex models for studying muscle func- tory adjustment of human locomotion during obstacle avoidance.
tion in walking. Philos Trans R Soc Lond B Biol Sci 2003;358:1501– Biol Cybern 1998;78:9–17.
1509. 93. Thelen DG, Anderson FC, Delp SL. Generating forward dy-
69. Perry J, Antonelli D, Ford W. Analysis of knee-joint forces during namic simulations of movement using computed muscle control.
flexed-knee stance. J Bone Joint Surg Am 1975;57A:961–967. J Biomech 2003;36:321–328.
70. Perry J, Newsam C. Function of the hamstrings in cerebral palsy. 94. van der Helm FCT, Rozendaall LA. Musculoskeletal systems with
In: Sussman MD, ed. The Diplegic Child: Evaluation and Manage- intrinsic and proprioceptive feedback. In: Winters JM, Crago PE,
ment. Rosemont, IL: American Academy of Orthopaedic Surgeons, eds. Biomechanics and Neural Control of Posture and Movement.
1992;299–307. New York: Springer-Verlag, Inc., 2000;164–174.
71. Perry J. Distal rectus femoris transfer. Dev Med Child Neurol 95. van der Helm FCT, van Soest AJ. Planning of human motions: How
1987;29:153–158. simple must it be? In: Winters JM, Crago PE, eds. Biomechanics and
72. Perry J. Gait analysis: Normal and pathological function. Thorofare, Neural Control of Posture and Movement. New York: Springer-Verlag,
NJ: Slack Incorporated, 1992. Inc., 2000;373–381.
73. Piazza SJ, Delp SL. The influence of muscles on knee flexion during 96. van Soest AJK, Casius LJR. The merits of a parallel genetic algo-
the swing phase of gait. J Biomech 1996;29:723–733. rithm in solving hard optimization problems. J Biomech Eng 2003;
74. Piazza SJ, Delp SL. Three-dimensional dynamic simulation of to- 125:141–146.
tal knee replacement motion during a step-up task. J Biomech Eng 97. Wickiewicz TL, Roy RR, Powell PL, Edgerton VR. Muscle architec-
2001;123:599–606. ture of the human lower limb. Clin Orthop 1983;275–283.
75. Rab GT, Muscle. In: Rose J, Gamble JG, eds. Human Walking. 2nd 98. Wiley ME, Damiano DL. Lower-extremity strength profiles in spastic
ed. Baltimore: Williams and Wilkins, 1994;101–121. cerebral palsy. Dev Med Child Neurol 1998;40:100–107.
76. Ralston HJ. Energetics of human walking. In: Herman RM, Grillner 99. Wilson DL, Zhu Q, Duerk JL, Mansour JM, Kilgore K, Crago PE. Es-
S, Stein PSG, Stuart DG, eds. Neural Control of Locomotion. New timation of tendon moment arms from three-dimensional magnetic
York: Plenum Press, 1976;77–98. resonance images. Ann Biomed Eng 1999;27:247–256.
77. Reimers J. Static and dynamic problems in spastic cerebral palsy. J 100. Winter DA. Overall principle of lower limb support during stance
Bone Joint Surg Br 1973;55-B:822–827. phase of gait. J Biomech 1980;13:923–927.
78. Riley PO, Kerrigan DC. Torque action of two-joint muscles in the 101. Winter DA. The Biomechanics and Motor Control of Human Gait:
swing period of stiff-legged gait: A forward dynamic model analysis. Normal, Elderly, and Pathological. Waterloo, Ontario, Canada: Wa-
J Biomech 1998;31:835–840. terloo Biomechanics, 1991.
79. Rodda J, Graham HK. Classification of gait patterns in spastic hemi- 102. Wright IC, Neptune RR, van den Bogert AJ, Nigg BM. The influence
plegia and spastic diplegia: A basis for a management algorithm. of foot positioning on ankle sprains. J Biomech 2000;33:513–519.
Eur J Neurol 2001;8(suppl 5):98–108. 103. Yamaguchi GT, Zajac FE. A planar model of the knee joint to char-
80. Roosth HP. Flexion deformity of the hip and knee in spastic cerebral acterize the knee extensor mechanism. J Biomech 1989;22:1–10.
palsy: Treatment by early release of spastic hip-flexor muscles. J 104. Yamaguchi GT, Zajac FE. Restoring unassisted natural gait to para-
Bone Joint Surg Am 1971;53-A:1489–1510. plegics via functional neuromuscular stimulation: A computer sim-
81. Rose J, Gamble J, Medeiros J, Burgos A, Haskell W. Energy expen- ulation study. IEEE Trans Biomed Eng 1990;37:886–902.
diture index of walking for normal children and for children with 105. Zajac FE, Gordon ME. Determining muscle’s force and action in
cerebral palsy. Dev Med Child Neurol 1990;32:333–340. multi-articular movement. Exerc Sport Sci Rev 1989;17:187–230.
82. Rosenthal RK, Levine DB. Fragmentation of the distal pole of the 106. Zajac FE, Neptune RR, Kautz SA. Biomechanics and muscle coor-
patella in spastic cerebral palsy. J Bone Joint Surg Am 1977;59A:934– dination of human walking part II: Lessons from dynamical simu-
939. lations and clinical implications. Gait Posture 2003;17:1–17.
83. Schmidt DJ, Arnold AS, Carroll NC, Delp SL. Length changes of the 107. Zajac FE. Muscle and tendon: Properties, models, scaling, and ap-
hamstrings and adductors resulting from derotational osteotomies plication to biomechanics and motor control. Crit Rev Biomed Eng
of the femur. J Orthop Res 1999;17:279–285. 1989;17:359–411.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

• C h a p t e r 13
•
◗ The Next Step: Restoring Walking
After Paralysis
Ronald J. Triolo and Rudi Kobetic

Loss of the ability to walk as a result of paralysis follow- inputs, absence of feedback from natural or body-mounted
ing spinal cord injury (SCI) can compromise the ability to sensors, inability to isolate and fully activate target mus-
work, engage in social or leisure activities, pursue an ed- cles, reverse order of motor unit recruitment with electrical
ucation or participate in other activities associated with stimulation, muscle deconditioning and rapid fatigue, as
an independent and productive lifestyle. Long periods of well as critical issues related to stimulus timing and mod-
immobility after SCI can cause degenerative changes of al- ification during the gait cycle. Although all major muscles
most every major organ system including the bones, joints, of the trunk, hips, knees, and ankles can be activated with
heart, lungs, and skin. However, if the damage to the ner- electrical stimulation, walking with FES is not normal and
vous system is confined to the upper motor neurons, then requires various compensatory mechanisms such as brac-
the peripheral nerves (alpha motor neurons) can often ing and walking aids.
be excited with small electric currents to cause the mus- The state of the art in standing and walking with FES via
cles they innervate to contract. Functional electrical stim- transcutaneous, intramuscular end implanted electrode
ulation (FES) refers to a developing assistive technology technologies are reviewed in this chapter. Current research
that can generate purposeful, useful movements of the ex- directions addressing the many challenges to achieving
tremities through such electrical activation of paralyzed or natural and energy efficient lower extremity function after
spastic muscles. In selected individuals with SCI, standing paralysis, through more efficient methods for interfacing
and walking can be achieved by coordinating the actions with the nervous system and controlling stimulation are
of weak, paralyzed, or uncontrollable muscles with each identified.
other and with remaining voluntary movement.
This chapter describes the status of motor system
MUSCULAR RESPONSES TO
neural prostheses (technologies that facilitate or restore
ELECTRICAL STIMULATION
movement by replacing or augmenting the damaged or
dysfunctional nervous system) for standing and walking
Contractile Properties
for individuals paralyzed by low cervical (tetraplegia) or
thoracic (paraplegia) SCI. In these systems, FES can sup- The contractile properties of the electrically activated mus-
port the body against gravity to prevent collapse and pro- cles are a large determinant of the quality of standing and
vide forward progression through the actions of the electri- walking possible with FES. In general, muscles that are
cally activated musculature. Balance is provided through paralyzed after spinal cord injury are atrophied and com-
voluntary interactions with walkers, crutches, or other posed mostly of fast-twitch fibers (29). When electrically
walking aids. Such neural prostheses have the potential activated, such muscles produce less than normal force
to postpone or prevent medical complications secondary and fatigue quickly.
to paralysis and to improve the independence of people The force of a muscle contraction is related to the net
with SCI by providing a means to exercise, stand, step, charge delivered to the nerve. The net force output and
and negotiate physical barriers. speed of the contraction increases with increasing stim-
For an individual with complete absence of motor and ulus current amplitude, pulse duration, and frequency.
sensory function below the level of injury, smooth, en- Muscle fibers in the intact neuromuscular system are re-
ergy efficient, and cosmetically appealing walking with cruited in an asynchronous fashion, generally beginning
FES presents many technical challenges. These challenges with slow-twitch, fatigue resistant fibers and progressing
include the lack of higher level descending command toward the fast-twitch fibers when additional forces are

209
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

210 Chapter 13 / The Next Step: Restoring Walking After Paralysis

required. This recruitment order is reversed with FES, frequency of 10 Hz was effective in increasing endurance
which elicits contractions in the larger fast-twitch motor (60). With an exercise protocol of 30 minutes of stimula-
units first, because of their lower threshold to depolariza- tion per day at a frequency varying from 16 to 30 Hz, we
tion by an external electric field. Electrical stimulation also observed a significant increase in the strength and fatigue
produces a synchronous contraction of all supra-threshold resistance of paralyzed muscle within 3 months of train-
motor units. Furthermore, muscle composition after SCI ing. Electrical exercise was a mix of isometric, eccentric,
tends to shift from slow to predominantly fast fiber types. isokinetic, and isotonic contractions that more closely ap-
These phenomena combine to produce nonlinearities (de- proximates muscle activity in an individual with an intact
lays, thresholds, strength, and fatigability) in the recruit- nervous system.
ment properties of stimulated muscle that complicate con- The well-conditioned electrically stimulated muscle will
trol and coordination. retain a stable fiber type composition with predictable fa-
Muscle contraction and relaxation properties are criti- tigue properties (60). At a constant stimulus frequency, fa-
cal in the synthesis of walking with electrical stimulation tigue causes a shift in recruitment properties manifesting
because event times and relative phasing between bursts in an elevated threshold of pulse duration required to elicit
of muscle activities during the gait cycle are on the order a contraction as well as a decreased maximum output. This
of tenths of a second. Muscle contraction rise time is de- shift in the recruitment curve is an ongoing process dur-
pendent on the frequency of stimulation (15) and on the ing stimulation until it reaches some steady state position.
muscle fiber type composition. Typical rise times, from the During the rest period, this process is reversed. Similarly,
first stimulus to 90% of maximal moment developed at a at constant stimulus pulse duration the muscle undergoes
joint, are on the order of 100 to 300 msec. Therefore, stim- frequency dependent fatigue and recovery (23). A negligi-
ulation must be initiated before a desired joint moment is ble increase in the knee moment has been observed during
required to make the stimulation effective. Computer sim- isometric contraction of the quadriceps muscles at stimu-
ulations of gait in paraplegics using electrical stimulation lation frequencies above 50 Hz both in normal and para-
show great sensitivity to on/off timing of muscle activity lyzed human muscle (15,22).
(81). This has been confirmed by our observations where Relevant measures of fatigue and/or endurance include
a delay in muscle activation as small as 20 msec produced the number of repetitions above a target joint moment and
a visually noticeable effect on walking. the amount of moment remaining during continuous stim-
Half relaxation time, the time from last stimulus to a ulation. During cyclic stimulation at various on/off stim-
50% decrease in the joint moment, was found to corre- ulation times the peak isokinetic moment at the knee set-
late with the percentage of fast-twitch type II fibers in tles to a steady state value within the first 10 minutes and
the muscle. Estimated values of half relaxation times are remains constant for more than 1 hour. The normalized
240 msec for slow-twitch fibers and 57 msec for fast-twitch steady state moment is strongly related to the amount of
fibers (56). An electrically stimulated paralyzed muscle can rest time between the cycles. Thus, the shorter the off-
have a half relaxation time in this range of values depend- time the lower the amount of usable moment available
ing on conditioning; the time is longer when the muscle is to produce a desired function at the joint. During continu-
fatigued. Relaxation time has been found to increase both ous stimulation, there is a marked reduction in isometric
with increases in the stimulation frequency (15) and with moment within the first 2 minutes of stimulation. The re-
the onset of muscle fatigue (21). maining steady state moment is dependent on stimulation
frequency. The available moment at the knee during con-
tinuous quadriceps stimulation is a good indicator of the
Strength and Endurance
subject’s ability to stand.
Reconditioning of the paralyzed muscle is required before The moment produced by electrical stimulation of a
initiating walking with electrical stimulation to reverse the paralyzed muscle about a particular joint varies greatly
effects of disuse and improve stimulated strength and fa- among different subjects. The strength of an electrically
tigue resistance to the levels required to generate safe and activated contraction is a function of many factors such
effective standing and stepping motions. Although no op- as electrode position and its ability to recruit all muscle
timal program to maximize both strength and endurance fibers at a safe maximal level of stimulation. Other factors
of paralyzed muscle has been established, some general include the frequency of stimulation, which has negligible
trends have emerged. An electrical exercise program of effect above 50 Hz in a well-conditioned electrically acti-
isotonic muscle contraction of 30 minutes per day, with vated muscle (15), and antagonistic or synergistic muscle
a 1:1 duty cycle (4 sec on and off times) at a stimulation activity. Caution must be used when measuring maximal
frequency of 20 Hz produced a strength increase in most isometric or eccentric contraction in paraplegic individu-
individuals with SCI. However, only a few individuals ex- als, especially in those with considerable osteoporosis (10).
perienced an increase in endurance with this protocol (7). Ultimately, the strength of contraction depends primarily
Other studies demonstrated that exercise at a stimulation on the state of muscle atrophy and the position of the
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

Chapter 13 / The Next Step: Restoring Walking After Paralysis 211

◗ TABLE 13-1 Isometric Joint Moments* duced by maintaining activation to the quadriceps of the
stance limb while initiating a flexion withdrawal in the
Moment Angle
Action (N-m) (Degree) contralateral limb. Depression of the crutch- or walker-
mounted switch on the swing limb side stimulates the af-
Trunk extension 70 0 ferent sensory fibers and triggers a spinal reflex arc that
Hip flexion 60 0 causes hip, knee, and ankle flexion in response to the
Hip extension 70 45 electrical stimulus. To complete the advancement phase
Hip abduction 40 0 of the stride, activation of the knee extensors on the swing
Hip adduction 30 0 limb is initiated while the reflex is still active and flexing the
Knee extension 80 45
hip. The stimulus producing the flexion reflex is then re-
Knee flexion 15 90
Ankle plantar flexion 55 15 (dorsi)
moved, leaving the user in double-limb support once again
Ankle dorsiflexion 15 15 (plantar) with bilateral quadriceps stimulation. It has been reported
that some paralyzed subjects walk at speeds approaching
∗ Average values of joint moments at specified angles obtained from six
one quarter of normal, and can ascend a curb or step with
conditioned volunteers with paraplegia with electrical stimulation at
maximum pulse width of 150 µsec and maximum frequency of 50 Hz. surface stimulation.
Complicating issues with this system include active flex-
electrode. Well-conditioned muscles produce from 30% to ion at the hip generated by the rectus femoris muscle that
60% of normal joint moment. Table 13-1 presents aver- compromises erect standing posture and results in an an-
age values of stimulated joint moments. Except for ankle terior pelvic tilt with compensatory lordosis, or excessive
plantar flexion and hip abduction, measured moments weight on the arms to remain upright. In addition, not
were sufficient to sustain independent normal speed walk- all patients will exhibit a flexion withdrawal reflex that is
ing as predicted by computer simulation (81). strong or repeatable enough for stepping. The reflex can
habituate with repeated activation, limiting the number of
steps that can be taken.
ELECTRICAL STIMULATION
TECHNOLOGIES Intramuscular Stimulation
Walking with FES after complete paraplegia requires the Intramuscular electrodes with percutaneous leads that exit
application of depolarizing electric fields to the nerves that the skin are thin, helically coiled, and insulated multi-
innervate the muscles required for the desired movement. stranded stainless steel wires similar to kinesiologic EMG
These fields are applied either transcutaneously through recording electrodes. These devices can be introduced with
electrodes placed on the skin surface, or via implanted elec- a hypodermic needle. The uninsulated stimulating tip is
trodes in closer proximity to the targeted neural structures. placed near the target nerve, thus bypassing cutaneous sen-
sory fibers and providing improved selectivity and access
to deep muscles (51). The configuration of a stainless steel
Transcutaneous Stimulation
helix around a polypropylene core provides strain relief
Transcutaneous (surface) stimulation achieves strong con- to reduce fatigue fracture of the wire and allows move-
tractions of large muscle groups served by superficial ment with surrounding tissue (66). Leads from multiple
nerves, but lacks selectivity and is unable to access deeper electrodes distributed about the lower extremities, pelvis,
muscles. Furthermore, surface electrodes need to be reap- and trunk can be routed subcutaneously to a common
plied each time such a system is to be used, compromising exit site for convenient connection to an external stimu-
day-to-day repeatability of the stimulated responses, and lator. Monopolar cathodic stimulation is commonly em-
the management of external cables rapidly becomes dif- ployed with the anode placed on the skin surface over a
ficult as the number of stimulus channels increases. As bony prominence. Once implanted, such electrodes can re-
few as two surface stimulation channels per leg can pro- main in the body as long as they produce strong, isolated
duce standing and reciprocal stepping motions through a contractions. The integrity of an electrode is monitored
combination of direct activation of the quadriceps mus- by the strength, selectivity, and recruitment properties of
cles via a pair of surface electrodes on the anterior thigh the muscle contraction it elicits as well as its electrical
and the triggering of a flexion withdrawal reflex via a impedance. The most frequent reason for electrode failure
second pair of electrodes over the dermatomes of the per- in the first few weeks after implantation is spontaneous
oneal, sural or saphenous sensory nerves (6,25,28). Stand- movement away from the motor point, which alters the
ing is achieved by simultaneously activating the quadri- contractile response. Another cause is breakage beneath
ceps bilaterally in response to a command input, such the skin with loss of the uninsulated section. This results
as the simultaneous depression of switches on the han- in high electrical impedance since the surface area for de-
dles of a rolling walker or crutches. A stride is pro- livery of charge is reduced (55). An average survival rate
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

212 Chapter 13 / The Next Step: Restoring Walking After Paralysis

of the intramuscular electrode is 70% at 1 year, and many codified the process of generating stimulus profiles in such
have remained functional for more than 5 years. a way as to be repeatable by other clinicians or researchers.
Chronically indwelling helically coiled fine wire intra- Next, experiments with subjects with multichannel percu-
muscular electrodes with percutaneous leads have allowed taneous systems defined the minimal muscle set required
researchers to synthesize complex lower extremity mo- to achieve stable standing and repeatable stepping (47).
tions (45) by activating numerous muscle groups with up This provided a set of primary target muscles for implan-
to 48 separate channels of stimulation under the control tation. Finally, the locations and stimulated responses of
of a programmable microprocessor-based external stimu- intramuscular electrodes with percutaneous leads were
lator (12). Users select one of a series of movement pat- used to guide the establishment of the surgical approaches
terns by scrolling through a menu of options presented on required to access the motor points of the target muscles
a liquid-crystal display via switches on a command ring (69). Cadaver and intraoperative tests confirmed the in-
worn on the index finger. Preprogrammed patterns of stim- sight provided by the experience with percutaneous elec-
ulation to the appropriate muscles are synchronized with trodes. To date, these approaches have been used to install
the gait cycle by successive switch depressions for each surgically implanted lower extremity systems for standing,
step, or by insole mounted pressure sensors that detect transfers, and short distance mobility in the vicinity of the
contact and loading. The quality of the motions produced wheelchair to 19 volunteers with complete thoracic or in-
with this system depends on the availability, strength, complete low-cervical injuries (46,77).
and endurance of the paralyzed muscles; the ability The implanted components of these systems include
of the therapist or engineer to specify patterns of stimu- an eight-channel receiver-stimulator, illustrated in Figure
lation for ambulation; and the subject’s experience with 13-1, and epimysial (3) and surgically-implanted intramus-
the device. Some well-trained subjects are able to walk cular (53) electrodes. The implant is a passive pacemaker-
300 meters repeatedly at 30 m per minute with this system like device that receives power and command signals from
(44). Although such systems are best suited for temporary a wearable external control unit (70). To achieve standing,
therapeutic applications, or to simulate the actions of a epimysial electrodes are installed bilaterally in the vastus
completely implanted system, with the proper care and lateralis (to achieve knee extension without hip flexion),
maintenance FES systems for walking based on intramus- the semimembranosus (and alternately the posterior por-
cular electrodes with percutaneous leads can remain op- tion of the adductor magnus), and the gluteus maximus,
erational and safe for functional use for many years (2). while intramuscular electrodes are inserted at the lumbar
spinal roots to activate the erector spinae muscles. The
entire system can be implanted in a single surgical pro-
Implanted Systems
cedure and allows recipients to rise from a seated posi-
Implanted systems utilizing electrodes placed on or tion, perform one-handed reaching tasks to retrieve objects
around the target nerve (epineural), sutured or inserted from wheelchair inaccessible shelves and achieve swing-
to the muscle near the nerve entry point (epimysial and through gait with a walker (16,17).
intramuscular), and connected to surgically implanted Standing and stepping with completely implanted sys-
pulse generators provide many major advantages over sur- tems can be achieved without braces for persons with com-
face and percutaneous stimulation. Advantages of this plete paraplegia with 16 channels of stimulation (the four
pacemaker-like approach include improved convenience, listed above for standing, plus the tibialis anterior, tensor
cosmetics, reliability, maintenance, and repeatability (42). fascia lata, sartorius and hamstrings, bilaterally). These
Multichannel implanted systems for walking after paraple- systems rely on two eight-channel devices as depicted in
gia provide standing and swing-through gait (13,34). Ex- Figure 13-2. Users of the standing systems with complete
ercise and standing functions have been reported with a thoracic level injuries below the level of T4 can have the
cochlear implant modified to deliver 22 channels of stimu- second implant installed to activate the additional muscles
lation to the lower extremities (18), and a 12-channel sys- required for walking in a second surgical procedure. Figure
tem for intradural stimulation of the L2-S2 motor roots 13-3 shows reciprocal gait with a rolling walker using the
(64). 16-channel dual-implant system. System recipients trig-
Because of their ability to activate muscles inaccessible ger each step via successive depressions of a ring-mounted
to surface stimulation, intramuscular electrodes with per- thumb switch. Walking speeds of up to 10 m per minute
cutaneous leads have been valuable tools to simulate the at cadences of 26 steps per minute were achieved with this
action of completely implanted FES systems. Researchers approach (46).
at the Cleveland FES Center utilized these devices to com-
plete three studies essential to proceeding with clinical
Hybrid Systems
trials of implanted FES technology. First, multichannel
percutaneous systems were employed to devise a standard- Hybrid systems combine bracing with FES in an at-
ized procedure and set of rules for specifying and adjusting tempt to overcome the disadvantages of stimulation or
patterns of stimulation for reciprocal walking (45). This orthoses alone for providing ambulation after SCI. The
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

Chapter 13 / The Next Step: Restoring Walking After Paralysis 213

FIGURE 13-1. The eight-channel implanted receiver-stimulator (IRS-8) developed at Case Western
Reserve University and Cleveland Department of Veterans Affairs Medical Center. The electronics shown
at the top are hermetically sealed in a titanium package that serves as a common system anode. Eight
leads provide in-line connections to stimulating electrodes implanted in or on the target muscle or
around the target nerve.

advantages of orthoses lie primarily in their ability to con- 14 weeks of training with 3 hours of walking per week,
strain the motions of the joints, reduce the degrees of free- significant reduction in spasticity, total cholesterol and
dom of movement and provide mechanical stability. For low density lipids, hydroxyproline to creatinine ratio, and
static activities such as quiet standing, individuals with increased knee extensor torque were evident (73). The ad-
paraplegia can assume a stable posture with little or no dition of FES to the gluteals during stance when using the
muscular exertion by locking the knees of a brace and hy- long leg braces resulted in a 36% reduction in the crutch
perextending the hips, thus avoiding the fatigue associ- force (52), a 30% reduction in PCI (74), and provided for-
ated with continuous stimulation. FES is quite effective ward propulsion by driving the stance leg into extension.
at introducing large impulsive forces to move the body By incorporating joint locks or brakes, standing and
forward through activation of large lower extremity mus- stance-limb stability against collapse can be accomplished
cles, which would reduce the upper extremity exertion re- with minimal muscle stimulation. Numerous examples of
quired for walking in conventional braces. Combining FES such systems have been prototyped and tested, yet few
and bracing in a hybrid orthosis offers an opportunity to have proven clinically viable (30,39,41,59). A hybrid or-
take advantage of the positive aspects of each technology thosis that utilized surface FES and incorporated closed-
and minimize the potential shortcomings. To date, hybrid loop, computer-controlled magnetic particle brakes at the
systems remain primarily exercise devices with limited hip and knee is reported to have been successful in reg-
functionality (4,5,38,61,62,67). ulating desired position and velocity of the joints during
One such system is an FES-powered reciprocating gait gait (19,26). The muscle fatigue was significantly reduced
orthosis (RGO) (32,48). Adding FES to a standard RGO when compared with FES-only gait (27). Major drawbacks
improved walking distance from 100 to 800 m, reduced of this system were size, weight, ease of donning and doff-
energy expenditure by 15%–30%, improved balance on in- ing, and cosmetics of the brace. With fully locked joints,
clines, and provided unassisted rising (72). There was a sig- the power requirements were prohibitive for practical use
nificant reduction in heart rate (8) and in physiologic cost out of the laboratory.
index (PCI) when walking by adding FES, but no change A controlled knee joint that provides stability during
in cadence, step length or velocity was noted (37). After stance phase and allows freedom during swing has been
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

214 Chapter 13 / The Next Step: Restoring Walking After Paralysis

FIGURE 13-2. A schematic presentation of the FES walking system is shown on the left. Stimulus
patterns are created on a personal computer and downloaded to a wearable external control unit that
transmits power and stimulus information to the implanted components. On the right is a radiograph
of 16-channel implanted walking system consisting of two receiver-stimulators, 14 epimysial and two
intramuscular electrodes.

designed based on a wrap-spring clutch (36). This system cles at the joint level can often be simplified by recruiting
provided significant reduction in oxygen consumption dur- multiple muscles with a single electrode. For example, one
ing walking when compared to a locked knee brace (35). precisely located electrode can produce balanced dorsi-
The hybrid combination of a brace with an isocentric re- flexion by stimulating the deep peroneal nerve to activate
ciprocator and lockable knee joints resulted in gait that the tibialis anterior and peroneus tertius while minimiz-
was slower, but with less forward trunk lean than walking ing the extensor digitorum longus and extensor hallucis
with FES alone (43). longus. Similarly, one judiciously placed electrode can re-
cruit the soleus and both heads of the gastrocnemius to
maximize the plantarflexion moment. Tibialis posterior re-
SYNTHESIS OF GAIT WITH
cruitment is usually avoided because it causes foot inver-
ELECTRICAL STIMULATION
sion. The gluteus minimus is usually activated with the
tensor fascia lata, although it is possible to implant the
Muscle Functions and Interactions
tensor without the gluteus minimus. Muscles with multiple
The 48-channel percutaneous stimulation system em- innervations are difficult to recruit with a single peripher-
ployed at our center has provided an excellent opportunity ally placed electrode, although they can be accessed with
to study human functional musculoskeletal anatomy and electrodes located at the spinal roots above the lumbar-
develop control strategies that mimic the human central sacral plexus. For example, second and third lumbar motor
nervous system. Muscles are generally targeted for implan- root stimulation recruits the iliopsoas but may also include
tation for their primary actions, although they often have undesirable stimulation of the adductors and quadriceps,
secondary and tertiary motor functions that need to be which are to be avoided.
counterbalanced with additional channels of stimulation. Muscles not only accelerate the joint or joints that they
The process of balancing the actions of the stimulated mus- span, but can also induce accelerations at other joints (82).
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

Chapter 13 / The Next Step: Restoring Walking After Paralysis 215

FIGURE 13-3. A subject stepping with a 16-channel implanted FES walking system using a hand switch
to initiate each step and a walker for balance. The external control unit is attached to his belt and ankle
foot orthoses protect against ankle injury.

Therefore, the action of a stimulated muscle depends on be customized for the unique responses from individual
the posture and motion of the entire musculoskeletal sys- electrodes. Multichannel computer-controlled stimulators
tem at the time it is activated, rather than solely by its have been developed that can control pulse duration, fre-
direct anatomical connections. For example, gluteus max- quency and current amplitude on a pulse by pulse basis
imus acts as an extensor of the hip and an external rotator with a resolution of 1 msec. Conditional jumps to special-
of the femur. During stance, gluteus maximus rotates the ized patterns of stimulation, or other control actions such
femur posteriorly, thereby extending the knee. Thus, dur- as delays and wait states, have been implemented based
ing walking the concept of a total extension moment at on volitional switch closures, insole pressure sensors or
the hip, knee, and ankle may be more useful than con- information from other body-mounted sensors.
sidering the moment produced by individual muscles at Figure 13-4 illustrates a typical temporal pattern of
their respective joints (80,82). Biarticulate muscles may stimulation for generating stepping motions in individu-
produce flexion at one joint and extension at another. For als with complete SCI. The shaded areas in the figure show
example, semimembranosus produces flexion at the knee when the stimulation is delivered to the muscles. The am-
and extension at the hip; rectus femoris produces flexion plitude of the shaded bars corresponds to the durations of
at the hip and extension at the knee. These dual functions individual pulses with a maximum set at 150 µsec. Ramp-
at two joints are difficult to separate when using electrical ing up and down of the bars indicates increasing and de-
stimulation. Other muscles have dual (antagonistic) func- creasing the width of each successive pulse. The pattern
tions at the same joint depending on position. For example, is divided into “tics” representing a generalized time base
gracilis and adductor longus act as hip flexors to about 30 over which stimulus pulse widths and interpulse intervals
degrees of hip flexion, but at greater angles they become are varied for each electrode throughout the gait cycle. In
hip extensors (75). this example, the gait cycle is divided into 100 tics. Con-
trol actions occur at “breakpoints” (BPs) that divide the
stimulation pattern into segments and allow the defini-
Stimulation Patterns
tion of how command inputs are to be interpreted. For
General templates for patterns of stimulation for step- example, progression through the pattern can continue un-
ping after complete paralysis have been developed through interrupted, and can be delayed at a BP in order to await
a process of trial and error based on the well-studied trigger input signals from switches or insole-mounted
EMG activity during normal gait. These patterns can foot-floor contact sensors before initiating the next step.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

216 Chapter 13 / The Next Step: Restoring Walking After Paralysis

FIGURE 13-4. A graphical pre-

sentation of an 18-channel stim-
ulation pattern for the walking
illustrated in Figure 13-5. Pulse
duration versus time through one
gait cycle is indicated by the mag-
nitude of the shaded areas. Stim-
ulus pulse duration can be modu-
lated from 0 to 150 microseconds
and frequency can be modulated
between 16 and 30 pulses/second
continuosly for a given pulse
amplitude.

Walking speed can also be varied by scaling the amount of and plantar flexors were turned off and right swing was
time between BPs. initiated. The stimulation was provided to tensor fascia
Breakpoint 1 (BP1) in the stimulation pattern repre- lata, iliopsoas, and gracilis to produce hip flexion, and
sents right heel strike (RHS). At this point, stimulation is tibialis anterior to dorsiflex the ankle for the foot-floor
provided to the right and left quadriceps to keep the knees clearance during swing. At the end of right swing, the
extended and to the left hip and trunk extensors to main- flexors were turned off and knee extensors were activated
tain posture. The processor can remain at BP1 until the in preparation for heel strike. The pattern then looped
subject presses the finger switch indicating readiness for from BP5 back to BP1 to repeat the entire gait cycle.
the next step or can continue uninterrupted in well trained The timing of the stimulation depends on the speed of
users, as was the case in this example. The period between walking and on the type of control used. When each step
BP1 and BP2 corresponds to weight shifting to the right was initiated by a hand switch, the left hip extensors were
leg and renewal of momentum for progression accom- activated before BP3 to maintain hip stability and posture
plished by activation of the right hip and trunk extensors while the user prepared for the next step. Similarly, the
and left plantar flexors. Following BP2, stimulation to the right hip extensors were activated before the BP5.
left knee and hip extensors was turned off, and the left
hip flexors and ankle dorsiflexors were activated. Left toe
off (LTO) occurred shortly thereafter. Just before BP3, the MECHANICS OF WALKING WITH FES
left hip flexors were turned off, and at the same time knee
extensors were activated in preparation for the left heel The kinematics of electrically stimulated walking vary de-
strike (LHS). Left ankle dorsiflexors remained active to pending on the availability, strength, and conditioning
prevent slapping of the foot at heel strike. A second burst of paralyzed muscles, the therapist’s ability to synthesize
of stimulation was provided to the right hip and trunk gait, and the subject’s experience. Although maximal forces
extensors to maintain posture, but then turned off shortly are not required during normal gait (24), reduced muscle
during the latter part of stance phase to delay the onset of strength in the major muscle groups of the trunk and lower
muscle fatigue. Left hamstrings may be activated earlier at extremities will reduce speed (11), symmetry (58), and sta-
BP3 to reduce extension moment on the knee and prevent bility of gait and increase the energy cost (65). Muscles
hyperextension. At BP3, the right step was initiated with- acting at the hip and ankle joints are particularly sensitive
out a delay. Pulse width and frequency of stimulation to to strength reductions, and the forces at these joints have
the left hamstrings, posterior portion of adductor magnus a significant correlation to walking speed (11). In stimu-
and gluteus maximus, and the right plantar flexors were lated gait, elimination of hip or trunk extensors or plantar
increased to produce a strong and rapid contraction flexors reduces walking speed. Hamstring elimination re-
to help transfer weight to the left leg and to move the duces speed by up to half. Removal of the posterior adduc-
body forward. At BP4, the right hip and trunk extensors tor causes a wide stance with side sway, and removal of the
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

Chapter 13 / The Next Step: Restoring Walking After Paralysis 217

FIGURE 13-5. A paraplegic subject (T–9, ASIA A) walking with a percutaneous FES system connected
to a belt-worn stimulator cycling through the activation pattern shown in Figure 13-4. Different phases
of the gait cycle are illustrated.

erector spinae reduces speed and requires increased arm Midstance: Single limb support is initiated in mid-
support. The salient mechanical features of stepping with stance with the foot in total contact with the floor. A sub-
FES illustrated in Figure 13-5 are summarized below to stantial amount of weight is accepted on the supporting
illustrate the similarities and differences with normal gait: leg in this phase. The GRV remains in front of the ankle,
Initial Contact: The stance phase begins with initial knee, and hip throughout, indicating that the ankle is pas-
foot contact. In FES walking, the heel strikes the floor after sively forced into dorsiflexion throughout midstance with-
an exaggerated swing phase with the foot in slight plantar out pretibial stimulation. As the heel rises the ankle dorsi-
flexion. The knee is fully extended, and the hip is flexed up flexes up to ∼ 10 degrees. Midstance knee extension causes
to 50 degrees. Neutral ankle position is difficult to achieve excessive vertical fluctuation of the center of mass requir-
in some individuals because of heel cord tightness, even ing extra hip and trunk extensor force to bring the body
with maximal activity of the pretibial muscles. The knee forward over the stance leg. As a result, much metabolic
is extended by the activity of the quadriceps in late swing. energy is wasted. The knee remains locked in extension as
Extensors of the knee move the leg into excessive hip flex- the GRV passes in front of the knee. Quadriceps are active
ion before heel contact, but the hip angle at initial contact during midstance, and the hip extensors are relaxed be-
is somewhat controlled by hamstring, gluteus maximus, fore their second burst of activity during terminal stance.
and posterior adductor activity. When hip extensor stimulation is stopped, a slight knee
Weight Acceptance: In this period of double limb sup- flexion occurs because of the flexion moment at the knee.
port, the leading leg progresses from initial heel strike to The GRV remains behind the knee for the rest of stance.
total foot contact with the floor. The ankle moves into fur- In subjects with weak hip and trunk extensors, the hip
ther plantar flexion even though restrained by maximum flexes in midstance rather than extending, interfering with
pretibial muscle stimulation. In some individuals with SCI, the body’s forward momentum. Increased hip flexion effec-
maximum pretibial stimulation is undesirable since it may tively lengthens the swing leg and interferes with toe clear-
cause a withdrawal reflex resulting in knee and hip flexion. ance, shortening step length and making weight trans-
During the loading phase, the knee tends to slightly hyper- fer difficult. Much of the body weight must be absorbed
extend instead of flex. In the early part of this phase, the through the arms to prevent falling. The GRV passes lateral
ground reaction vector (GRV) passes behind the ankle joint to the subtalar joint and medial to the knee and hip joint in
and in front of the knee and hip, generating moments that the frontal plane. This indicates that an eversion moment
plantar flex the foot and extend the knee. Knee extension is is occurring at the ankle and an adduction moment at the
partly caused by the hip extensors that produce posterior hip. In this phase, stimulation of the posterior portion of
rotation of the thigh. Weight acceptance on the leading the adductor magnus produces hip extension, and erector
leg is not completed until halfway into midstance. With spinae and gluteus medius produce hip abduction.
the body weight still behind the stance foot and excessive Terminal Stance: This phase begins with heel rise and
forward lean on the trunk, the hip flexes, resulting in loss terminates with double limb support. Activity of the plan-
of forward momentum and hampering weight transfer of tar flexors is initiated to generate the push-off force and to
the leading leg. At contralateral toe-off, the GRV passes in prevent excessive dorsiflexion because of the passive mo-
front of the ankle and remains in front of the knee and ment created by the GRV positioned in front of the an-
hip. This produces a dorsiflexion moment at the ankle and kle. This is a critical point during FES-induced gait. If the
a flexion moment at the hip. The knee remains in forced plantar flexors act too soon, they tend to raise the body
extension. Weight transfer to the stance leg is delayed, and instead of propelling it. This effectively shortens the swing
brief forces of up to 22% of body weight are exerted on leg and prevents proper swing foot contact. In addition,
each arm through the walking aid. activity of the swing leg hip extensors will rotate the leg
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

218 Chapter 13 / The Next Step: Restoring Walking After Paralysis

backward, reducing its effective step length and interfer- but a benefit is that a vigorous leg swing provides needed
ing with weight transfer. A push-off delayed into double progressional momentum during terminal stance of the
stance is less effective in weight transfer. With the body opposite leg.
weight far posterior and with loss of momentum, the hip Terminal Swing: An overly flexed hip is extended by
of the weight-accepting leg will tend to flex, and much of gravitational forces of the leg and activity of the hamstrings
the body weight must be supported by the arms. A sec- to control step length in preparation for foot contact.
ond burst of stimulation is delivered to the hip and trunk
extensors. It helps to stabilize the hip and trunk, especially Energy Expenditure
at the opposite heel strike. The GRV is now behind the hip, The metabolic energy currently required for ambulation
helping the active hip extensors. The knee remains in slight with FES is still too high to make it a practical alternative
flexion with the GRV behind it. Knee collapse is prevented to the wheelchair. The goal is to bring the energy require-
by active knee and hip extensors. ment to below 50% of the individual’s maximum aerobic
Preswing: Regulation of the knee and ankle is crit- capacity. At that level of energy consumption, walking can
ical for weight transfer to the other leg and for ini- be sustained for hours (79).
tiation of swing. A stiff knee of the stance leg can Average maximum aerobic capacity of untrained female
effectively shorten the swing leg, making transfer more dif- and male individuals with paraplegia was determined to be
ficult. Similarly, if the body weight has not progressed far 16 and 25 mL O2 /kg per minute, respectively (14). Trained
enough forward, the stance knee will collapse and inter- males with paraplegia averaged 28 mL O2 /kg per minute
rupt weight transfer. Proper control of knee flexion is dif- (14), whereas individuals with thoracic SCI using lower
ficult with pre-programmed stimulation and depends on extremity stimulation plus an arm ergometer while seated
the relaxation properties of the extensor muscles and the reached a maximal aerobic capacity of 36 mL O2 /kg per
contraction properties of the knee flexors. Knee flexion is minute (20). At this fitness level, prolonged walking with
achieved by stimulation of the sartorius and gracilis. Knee FES should require no more than 18 mL O2 /kg per minute.
flexion magnitude is slightly less than normal and varies However, walking with FES at 30 m per minute currently
between 30 and 40 degrees. Even though the GRV is far averages 28 mL O2 /kg per minute. This far exceeds the goal
behind the knee, the residual knee extension moment dur- of 50% maximum aerobic capacity, as well as the nominal
ing quadriceps relaxation resists flexion. However, an ob- value of 8 mL O2 /kg per minute for normal able-bodied
servable improvement in forward progression is achieved gait (54).
with even 5 to 10 degrees of knee flexion compared with a A subject lying supine on a mat using a pattern of
stiff knee. The ankle continues to plantar flex and reaches stimulation designed to produce walking motions with-
10 degrees before toe-off. Plantar flexor stimulation is dis- out weight bearing or ground contact consumes 20 mL
continued halfway into this phase, but the relaxation mo- O2 /kg per min (54). Therefore, 8 mL of O2 /kg per minute
ment continues to create further plantar flexion for the are typically used by the voluntarily controlled arm and
next 100 msec. Early deactivation of hip extensors results trunk muscles for support, balance, and propulsion via
in internal leg rotation and foot inversion. Hip flexion is a walker or crutches. Progressive addition of the muscle
initiated by the sartorius, gracilis, tensor fascia lata, and groups to the walking stimulation pattern in the supine
iliopsoas. Neutral is reached before toe-off. position results in a lower energy requirement for the full
Initial Swing: This phase is initiated at toe-off. Knee set of muscles than if all muscle groups were initiated at
flexion is produced by contraction of the sartorius and once. This suggests that a warm up prior to walking might
gracilis and is often initially inadequate, resulting in toe reduce the energy demand. Finer control of stimulation to
drag and poor swing phase mechanics. The stance leg reduce reliance on the upper extremities and more efficient
must be cleared as the leg swings through. Excessive in- generation of stepping motions are needed to reduce en-
toeing of the swing leg can catch easily. Because the gra- ergy consumption to reasonable levels.
cilis also strongly adducts, balance from the tensor fascia
lata and the gluteus minimus or medius is needed to clear
the stance leg. The tensor fascia lata medial thigh rotation RESEARCH DIRECTIONS AND
also balances the sartorius external rotation. Stimulation FUTURE DEVELOPMENTS
of the contralateral abductors will give additional toe and
leg clearance. The quest for efficient and effective methods of reanimat-
Midswing: Maximum knee flexion in FES-generated ing the paralyzed musculature to restore standing and
gait is nearly normal. At peak knee flexion, hip flexion is walking function to individuals with spinal cord injuries
normal and continues to overshoot to 60 degrees. This ex- continues in research laboratories and clinics around the
aggerated flexion is caused by acceleration of the leg result- world (57). Research and development efforts are fo-
ing from stimulation of the quadriceps while hip flexors cused on expanding the capabilities of implanted sys-
are deactivated. Energy is wasted because muscle action tems through additional channels of stimulation, new elec-
is used to overcome gravity for an unnecessary motion, trode configurations, more natural command and control
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

Chapter 13 / The Next Step: Restoring Walking After Paralysis 219

sources, and advanced techniques for automatic control all available motor units by virtue of their location in the
and regulation of movement (1,78). periphery (typically at a single nerve entry point to the tar-
A new family of implanted stimulation devices is be- geted muscle). On the other hand, nerve-based electrodes
ing developed and readied for clinical testing at our cen- can achieve maximal recruitment because of their intimate
ter as the common platform for the next-generation of contact with the nerve above the point where it branches
neural prostheses (71). The most important improvements into multiple motor points, usually at the expense of se-
offered by this new technology is the capability to provide lectivity. Since all distal muscles are represented by dis-
up to 16 independent channels of nerve or muscle-based tinct fascicles in the proximal nerve trunk, an effective
stimulation, and the capability to telemeter information nerve electrode must also be able to selectively recruit only
from implanted sensors to an external controller that can the desired fascicles within a multifascicular nerve. Selec-
utilize it to modulate stimulation in a closed-loop fash- tive activation of subpopulations of axons in a synergistic
ion (76). Such devices have been successfully implemented nerve trunk can delay the effects of fatigue by switching
clinically to process information from implanted joint an- between different muscles or portions of a muscle while
gle transducers consisting of a Hall-effect sensor array and maintaining a constant net moment about the joint. This
magnet packaged in threaded titanium capsules that are intermittent “sequential” or “cyclical” stimulation reduces
inserted into the bones on either side of the joint (9,40), and the effective duty cycle of each muscle fiber, allows some
to modulate stimulation based on feedback from multiple recovery from fatigue, and has been shown to improve en-
channels of EMG data acquired from muscles under voli- durance for locking the knee in extension (63).
tional control as sensed by implanted recording electrodes. Further advantages of selective stimulation are realized
This ability to derive “afferent” information about body by implanting a cuff electrode proximally where the nerve
position or voluntary muscle activity and exploit it to alter contains axons to multiple muscles that produce different
the “efferent” stimulation delivered to a large number of functions. A single multicontact electrode that can selec-
paralyzed muscles offers the potential for unprecedented tively activate individual fascicles within a nerve trunk is
improvements in the quality of standing and walking mo- highly desirable as it could simplify implant surgery while
tions and more interactive, intimate, and intuitive control simultaneously providing access to an increased number
of neural prostheses (71). of muscles. This approach has the potential to profoundly
Preliminary success has also been reported in the use improve neural prosthesis implementation by streamlin-
of “natural sensors” to control stimulation, such as can be ing the implant procedure, obviating the need to distribute
accomplished by recording the electroneurogram (ENG) and manage multiple lead wires to various points in the pe-
from an afferent nerve serving the mechanoreceptors in riphery, eliminating the time-consuming process of expos-
the plantar surface of the foot or the joint afferents ing and deploying an individual electrode on each of many
of the knee (76). A recording cuff electrode placed around target muscles, and activating muscles that are too small
the appropriate sensory nerve can sense the biopotentials or inaccessible for muscle-based electrodes. While such a
related to the mechanical events of foot-floor contact, system can enhance the performance of standing neural
weight acceptance, or loading response. Appropriately prostheses, it will be even more important for achieving
processed ENG signals can then be used to trigger, ter- the advanced functions of stepping and stair climbing in
minate, or modulate stimulation to the muscles required the future.
for the corresponding phase of the gait cycle. Such tech-
niques have been applied clinically to produce active CONCLUSION
dorsiflexion during swing and control plantarflexion dur-
ing initial stance to correct footdrop, improve swing limb Walking with FES for individuals with motor and sen-
clearance, and prevent foot slap with initial contact in indi- sory complete SCI requires the cooperative effort of peo-
viduals with hemiparesis (9). Similar methods can be ap- ple from diverse disciplines including surgery, engineer-
plied to individuals with partial paralysis resulting from ing, and therapy. Although the feasibility of ambulation
incomplete SCI to seamlessly coordinate the actions of with FES has been established experimentally and through
stimulated muscles with intact volitional movements. computer simulation, much work remains to be done to
Other improvements to our ability to synthesize more efficiently interface with the nervous system. New
smooth, cosmetically acceptable, and energy efficient gait and deeper understanding of the response of paralyzed
after paralysis will result from the development and ap- muscle to electrical stimulation are required to advance
plication of new nerve-based electrode technologies and the performance of lower extremity neural prostheses.
stimulation techniques that can completely and selectively More effective means of controlling the motion of individ-
activate all of the motor units in a targeted muscle or ual joints and multisegmental body structures are needed
block unwanted and counterproductive activity because to make FES-generated gait energy efficient and clinically
of spasticity in antagonist muscles. The muscle-based elec- practical. Improved activation and deactivation of muscles
trodes currently employed in neural prostheses for walk- should reduce unnecessary stimulation, improve balance,
ing achieve selectivity at the expense of full activation of and reduce energy requirements.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

220 Chapter 13 / The Next Step: Restoring Walking After Paralysis

Functional electrical stimulation provides an opportu- 8. Beillot J, Carre F, Le Claire G, et al. Energy consumption of para-
nity to activate and utilize muscles no longer under direct plegic locomotion using reciprocal gait orthosis. Eur J Appl Physiol
1996;73:376–381.
volitional control (40). In complete spinal cord injury, FES 9. Bhadra N, Peckham PH, Keith MW, et al. Implementation of an im-
allows short distance walking, stair climbing, maneuver- plantable joint-angle transducer. J Rehabil Res Dev 2002;39(3):411–
ability in small spaces, and other functional activities. This 422.
10. Biering-Sorensen F, Bohr H. Bone mineral content of the lumbar
technology has the potential to provide patients with par- spine and lower extremities years after spinal cord lesion. Paraplegia
tial spinal cord injuries with enhanced function and less 1988;26:293–301.
need for conventional bracing. 11. Bohannon RW. Relevance of muscle strength to gait performance
in patients with neurologic disability. J Neuro Rehabil 1989;3:97–
In addition to their applications after complete thoracic 100.
SCI, implanted stimulation systems will soon be clinically 12. Borges G, Ferguson K, Kobetic R. Development and operation of
available to restore or improve walking or transfers in peo- portable and laboratory electrical stimulation systems for walk-
ing in paraplegic subjects. IEEE Trans Biomed Eng 1989;36:798–
ple with partial paralysis or spasticity (68) from spinal cord 800.
injury, stroke, head injury, cerebral palsy, multiple sclero- 13. Brindley GS, Polkey CE, Rushton DN. Electrical splinting of the knee
sis, and Parkinson’s disease. More complex multichannel in paraplegia. Paraplegia 1979;16:248.
14. Burkett LN, Chisum J, Stone W, Fernhall B. Exercise capac-
systems incorporating sensory feedback for closed loop ity of untrained spinal cord injured individuals and the relation-
control will require much more development and testing ship of peak oxygen uptake to level of injury. Paraplegia 1990;28:
before they can be readied for widespread clinicalk use. 512–521.
15. Carroll SG, Triolo RJ, Chizeck HJ, Kobetic R, Marsolais EB.
Tetanic response of electrically stimulated paralyzed muscle at
varying interpulse intervals. IEEE Trans Biomed Eng 1989;36:644–
653.
ACKNOWLEDGMENTS 16. Davis JA, Triolo RJ, Uhlir JP, et al. Surgical technique for installing
an 8-channel neuroprosthesis for standing. Clin Orthop Rel Res
The authors acknowledge the contribution of the Cleve- 2001;2001:237–252.
17. Davis JA, Triolo RJ, Uhlir JP, Bieri C, Rohde L, Lissy D. Preliminary
land FES Center, a consortium consisting of the Louis performance of a surgically implanted neuroprosthesis for standing
Stokes Cleveland Department of Veteran Affairs Medical and transfer. J Rehabil Res Dev 2001;38:609–617.
Center, Case Western Reserve University and MetroHealth 18. Davis R, Eckhouse R, Patrick JF, Delehanty A. Computer-controlled
22-channel stimulator for limb movement. Acta Neurochirurgica
Medical Center. Some of the work described was supported Suppl 1987;39:117–120.
by the Rehabilitation R&D service of the US Department 19. Durfee W, Goldfarb M. Controlled-brake orthosis, U.S. Patent 5 476
of Veterans Affairs, the Office of Orphan Product Develop- 441,1995.
20. Edwards BG, Marsolais EB. Metabolic responses to arm ergome-
ment of the United States Food and Drug Administration try and functional neuromuscular stimulation. J Rehabil Res Dev
and the National Institutes of Health (NINDS and NIBIB). 1990;27:107–114.
Further recognition is due to Dr. E. Byron Marsolais, the 21. Edwards RHT, Hill DK, Jones DA. Heat production and chemical
changes during isometric contractions of the human quadriceps
General Clinical Research Center (GCRC) at Metro Health muscle. J Physiol (Lond) 1975;251:303–315.
Medical Center, and all the research participants, engi- 22. Edwards RHT, Young A, Hosking GP, Jones DA. Human skeletal mus-
neers, therapists, physicians, and technicians who have cle function: Description of tests and normal values. Clin Sci Mol Med
1977;52:283–290.
worked on this project over the past 30 years for their in- 23. Edwards RHT. Human muscle function and fatigue. In: Human mus-
valuable contributions. cle fatigue: Physiological mechanisms. Ciba Found Symp 1981;82:
1–18.
24. Ericson MO, Nisell R, Ekholm J. Quantified electromyography of
lower-limb muscles during level walking. Scand J Rehabil Med 1986;
REFERENCES 18:159–163.
25. Gallien P, Brissot R, Eyssette M, Tell L, Barat M, Wiart L, Petit H.
1. Abbas JJ, Riener R. Using mathematical models and advanced con- Restoration of gait by functional electrical stimulation for spinal cord
trol systems techniques to enhance neuroprosthesis function. Neu- injured patients. Paraplegia 1995;33:660–664.
romodulation 2001;4:187–195. 26. Goldfarb M, Durfee WK. Design of a controlled-brake orthosis for
2. Agarwal S, Kobetic R, Nandurkar S, Marsolais EB. Functional elec- FES-aided gait. IEEE Trans Rehabil Eng 1996;1:13–24.
trical stimulation for walking in paraplegia: 17 year follow-up of 27. Goldfarb M, Korkowski K, Harrold B, Durfee WK. Preliminary eval-
2 cases. J Spinal Cord Med 2003;26:86–91. uation of a controlled-brake orthosis for FES-aided gait. IEEE Trans
3. Akers JM, Peckham PH, Keith MW, Merritt K. Tissue response to Rehabil Eng 2003;11:241–248.
chronically stimulated implanted epimysial and intramuscular elec- 28. Graupe D, Kohn K. Functional Electrical Stimulation for Ambulation
trodes. IEEE Trans Rehabil Eng 1996;5:207–220. by Paraplegics. Malabar FL: Krieger Publishing Co; 1994;1–26.
4. Andrews BJ, Baxendale RH, Barnett R, Phillips GF, Paul JP, Freeman 29. Grimbv G, Broberg C, Krotkiewska 1, Krotkiewski M. Muscle fiber
PA. A hybrid orthosis for paraplegics incorporating feedback control. composition in patients with traumatic cord lesion. Scand J Rebabil
Proc 9th Symp ECHE, 1997;297–311. Med 1976;8:37–42.
5. Andrews BJ, Baxendale RH, Barnett R, et al. Hybrid FES orthosis 30. Harrison R, Lemaire E, Jeffreys Y, Goudreau L. Design and pilot test-
incorporating closed loop control and sensory feedback. Biomed Eng ing of an orthotic stance-phase control knee joint. Orthopadie Tech-
1988;10:189–195. nik; 2001;2–4.
6. Bajd T, Kralj A, Turk R. Standing up of a healthy subject and a para- 31. Haugland MK, Sinkjær T. Cutaneous whole nerve recordings used for
plegic patient. J Biomech 1982;15:1–10. correction of footdrop in hemiplegic man. IEEE Trans Rehabil Eng
7. Bajd T, Kraij A, Turk R, Benko H, Sega J. The use of a four-channel 1995;3:307–317.
electrical stimulator as an ambulatory aid for paraplegic patients. 32. Hirokawa S, Grimm M, Thanh LE, et al. Energy consumption in para-
Phys Ther 1983;63:1116–1120. plegia ambulation using the reciprocating gait orthosis and electrical
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

Chapter 13 / The Next Step: Restoring Walking After Paralysis 221

stimulation of the thigh muscles. Arch Phys Med Rehabil 1990;71:687– 56. Moulds RFW, Young A, Jones DA, Edwards RHT. A study of the
694. contractility, biochemistry and morphology of an isolated prepa-
33. Hoffer JA, Stein RB, Haugland MK, et al. Neural signals for com- ration of human skeletal muscle. Clin Sci Mol Med 1977;52:291–
mand control and feedback in functional neuromuscular stimulation 297.
[review]. J Rehabil Res Dev 1996;33:145–157. 57. Munih M, Ichie M. Current status and future prospects for upper
34. Holle J, Frey M, Gruber H, Kern H, Stohr H, Thoma H. Func- and lower extremity motor system neuroprostheses. Neuromodula-
tional electrostimulation of paraplegics. experimental investigations tion 2001;4:176–186.
and first clinical experience with an implantable stimulation device. 58. Murray MP, Cuten GN, Sepic SB, Gardner GM, Baldwin JM. Function
Orthopaedics 1984;7:1145–1160. of triceps surae during gait. J Bone Joint Surg 1978;60- A:473–475.
35. Irby SE, Kaufman KR, Mathewson JW, Sutherland DH. Automatic 59. Myers WN, Shadoan MD, Frobes JC, Baker KJ, Darron CR. Selec-
control design for a dynamic knee-brace system. IEEE Trans Rehabil tively lockable knee brace, U.S. Patent 5 490 831,1996.
Eng 1999a;7:135–139. 60. Peckham PH, Mortimer JT, Marsolais EB. Alteration in the force and
36. Irby SE, Kaufman KR, Wirta RW, Sutherland DH. Optimization and fatigability of skeletal muscle in quadriplegic humans following exer-
application of a wrap-spring clutch to a dynamic knee-ankle-foot or- cise induced by chronic electrical stimulation. Clin Orthop Relat Res
thosis. IEEE Trans Rehabil Eng 1999;7:130–134. 1976;114:326–334.
37. Isakov E, Douglas R, Berns P. Ambulation using the reciprocat- 61. Petrofsky JS, Phillips CA, Douglas R, Larson P. A computer-controlled
ing gait orthosis and functional electrical stimulation. Paraplegia walking system: The combination of an orthosis with functional elec-
1992;30:239–245. trical stimulation. J Clin Eng 1986;11:121–133.
38. Jaeger RT, Yarkony GM, Roth EG. Standing by a combined or- 62. Popovic D, Schwirtlich L. Hybrid powered orthoses. In: Popovic D,
thotic/electrical stimulation system in thoracic paraplegic. Proc IC- ed. Advances in External Control of Human Extremities IX. Yugoslav
CAART, Montreal, 1988:336–337. Committee for ETAN, Yugoslavia; 1987:95–104.
39. Jaspors P, Van Petegem W, Van der Perre G, Peeraer L. Optimisation 63. Pournezam M, Andrews BJ, Baxendale RH, et al. Reduction of muscle
of a combined ARGO-FES system: adaptation of the knee mechanism fatigue in man by cyclical stimulation. J Biomed Eng 1988;10:196–
to allow flexion of the knee during the swing phase. IEEE/EMBS 200.
Conference, Montreal, 1995. 64. Rushton DN, Perkins TA, Donaldson N, et al. LARSI: How to obtain
40. Johnson MW, Peckham PH, Bhadra N,et al. An implantable trans- favorable muscle contractions. Proceedings of the Second Annual
ducer for two-degree-of-freedom joint angle sensing. IEEE Trans IFESS Conference (IFESS ‘97) and Neural Prosthesis: Motor Sys-
Biomed Eng 1999;7:349–359. tems 5 (NP ‘97), Burnaby, British Columbia, Canada, 16–21 Aug.,
41. Kagaya H, Shimada Y, Sato K, Satom, Iizuka K, Obinata G. An electri- 1997:163–164.
cal knee lock system for functional electrical stimulation. Arch Phys 65. Saunders JB, Inman VT, Eberhart HD. The major determinants
Med Rehabil 1996;77:870–873. in normal and pathologic gait. J Bone Joint Surg 1953;35-A:543–
42. Kilgore KL, Peckham PH, Keith MW, et al. An implanted upper ex- 558.
tremity neuroprosthesis: A five patient follow-up, J Bone Joint Surg 66. Scheiner A, Polando G, Marsolais EB. Design and clinical application
Am 1997;79A:533–541. of a double helix electrode for functional electrical stimulation. IEEE
43. Kobetic R, Marsolais EB, Davy D, Gaudio R, Triolo R. Development of Trans Biomed Eng 1994;41:425–431.
a hybrid gait orthosis: a case report. J Spinal Cord Med 2003;26:254– 67. Schwirtlich L, Popovic D. Hibrid orthosis for deficient locomotion.
258. In: Advances in External Control of Human Extremities VIII edited by
44. Kobetic R, Marsolais EB, Samame P, Borges G. The next step: arti- Dejan Popovic, Yugoslav Committee for Electronics and Automation,
ficial walking. In: Rose J, Gamble JG, eds. Human Walking. 2nd ed. Belgrade. 1984:23–32.
Baltimore: Williams & Wilkins; 1994:225–252. 68. Scott TR, Haugland M. Command and control interfaces for ad-
45. Kobetic R, Marsolais EB. Synthesis of paraplegic gait with multi- vanced neuroprosthetic applications. Neuromodulation 2001;4:165–
channel functional neuromuscular stimulation. IEEE Trans Rehabil 175.
Eng 1994;2:66–79. 69. Sharma M, Marsolais EB, Polando G, et al. Implantation of a 16-
46. Kobetic R, Triolo RJ, Uhlir J, et al. Implanted functional electrical channel functional electrical stimulation walking system. Clin Ortho
stimulation system for mobility in paraplegia: A follow-up case re- 1998;347:236–242.
port. IEEE Trans Rehabil Eng 1999;7:390–398. 70. Smith B, Peckham PH, Keith MW, Roscoe DD. An externally
47. Kobetic R, Triolo RJ. Muscle selection and walking performance of powered, multichannel implantable stimulator for versatile con-
multichannel FES systems for ambulation in paraplegia. IEEE Trans trol of paralyzed muscle. IEEE Trans Biomed Eng 1987;34:499–
Rehabil Eng 1997;5:23–29. 508.
48. Larson PF, Douglas RD, Petrofsky JS, Phillips CA. Walking assistance 71. Smith B, Tang Z, Johnson M, et al. An externally powered, multi-
system, U.S. Patent 4 697 808,1987. channel, implantable stimulator-telemeter for control of paralyzed
49. Marsolais EB, Kobetic R, Barnicle K, Jacobs J. FNS application for muscle. IEEE Trans Rehabil Eng 1998;45:463–475.
restoring function in stroke and head-injury patients. J Clin Eng 72. Solomonow M, Baratta R, Hirokawa S, et al. The RGO Generation II:
1990;15:489–496. Muscle stimulation powered orthosis as a practical walking system
50. Marsolais EB, Kobetic R. Functional electrical stimulation for walk- for thoracic paraplegics. Orthopedics 1989;12:1309–1315.
ing in paraplegia. J Bone Joint Surg 1987;69-A:728–733. 73. Solomonow M, Reisin E, Aguilar E, Baratta RV, Best R, D’Ambrosia
51. Marsolais EB, Kobetic R. Implantation techniques and experience RD. Reciprocal gait orthosis powered with electrical muscle stimula-
with percutaneous intramuscular electrodes in the lower extremities. tion (RGO II). Part II: medical evaluation of 70 paraplegic patients.
J Rehabil R & D 1986;23:1–8. Orthopedics 1997a;20:411–418.
52. McClelland M, Andrews BJ, Patrick JH, Freeman PA, El Masri WS. 74. Stallard J, Major RE. The influence of orthosis stiffness on para-
Augmentation of the Oswestry parawalker orthosis by means of sur- plegic ambulation and its implication for functional electrical stim-
face electrical stimulation: gait analysis of three patients. Paraplegia ulation (FES) walking systems. Prosthet Orthot Int 1995;19:108–
1987;25:32–38. 114.
53. Memberg WD, Peckham PH, Keith MW. A surgically-implanted in- 75. Steindier A. Kinesiology of the Human Body. 5th ed. Springfield:
tramuscular electrode for an implantable neuromuscular stimulation Charles C Thomas; 1977:316.
system. IEEE Trans Biomed Eng 1994;2:80–91. 76. Tang Z, Smith B, Schild JH, Peckham PH. Data transmission
54. Miller PC, Kobetic R, Lew RD. Energy cost of walking and stand- from an implantable biotelemeter by load-shift keying using cir-
ing using functional electrical stimulation. Rehabilitation Engineer- cuit configuration modulator. IEEE Trans Biomed Eng 1995;42:525–
ing Society of North America, 13th Annual Conference on Rehabilita- 528.
tion Technology, Washington DC, 1990:155–156. 77. Triolo RJ, Bieri C, Uhlir J, Kobetic R, Scheiner A, Marsolais EB.
55. Mortimer JT, Kaufman D, Roessmann U. Intramuscular electri- Implanted FNS systems for assisted standing and transfers for indi-
cal stimulation: Tissue damage. Ann Biomed Eng 1980;8:235– viduals with cervical spinal cord injuries: Clinical case reports. Arch
244. Phys Med & Rehabil 1996;77:1119–1128.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-13 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:57

222 Chapter 13 / The Next Step: Restoring Walking After Paralysis

78. Troyk PR, Donaldson N. Implantable FES stimulation systems: what 81. Yamaguchi GT, Zajac FE. Restoring unassisted natural gait to para-
is needed? Neuromodulation 2001;4:196–204. plegics via functional neuromuscular stimulation: A computer sim-
79. Waters RL, Lunsford BR. Energy expenditure of normal and patho- ulation study. IEEE Trans Biomed Eng 1990;37:886–902.
logic gait: Application to orthotic prescription. In: Atlas of Orthotics. 82. Zajac FE, Gordon ME. Determining muscle’s forces and action
2nd ed. St. Louis: Mosby; 1985:151–159. in multi-articular movement. In: Pandolf K, ed. Exercise and
80. Winter DA. Biomechanics and Motor Control of Human Movement. sports science reviews. Baltimore: Williams & Wilkins; 1989;17:187–
2nd ed. New York: Wiley; 1990:85. 230.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-14 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:55

• C h a p t e r 14
•
◗ Human Walking: Six Take-Home Lessons
James G. Gamble and Jessica Rose

LESSON 1: BIPEDALISM, THE ABILITY carry objects such as meat, nuts, fruit, and, perhaps most
TO WALK UPRIGHT, WAS THE FIRST importantly, infants and children. It must have been a great
ANATOMIC AND BEHAVIORAL advantage to be able to scoop up the baby and run in a
CHARACTERISTIC TO DISTINGUISH time of crisis. From a biomechanical perspective, bipedal-
OUR PRIMAL ANCESTORS FROM APES ism decreases the energy expenditure required during slow
walking speed and permits tool and weapon use, attributes
As Weaver and Kline discuss in Chapter 2, the ability to that would have been important for early hominoids dur-
walk upright on two legs, bipedalism, appeared before a ing both hunting and gathering. Bipedalism requires the
dexterous hand with an opposable thumb and before a lower limbs to support all the body weight, and it requires
large brain with the ability to use language and design tools balance against gravity.
and weapons. Bipedalism is at the root of what it means to An important question facing paleoanthropologists to-
be human. Our closest genetic ancestors, chimpanzees, do day is, “why did our ancestors first begin to walk upright in
not have a bipedal gait. They use a characteristic gait called the first place?” Of course, any discussion of “why bipedal-
“knuckle-walking.” Chimps do not fully extend their knees, ism” must consider natural selection. This is a term first
and they must use more muscle power than humans to used by Charles Darwin in 1859 and relates to the fitness
support the body in midstance. Chimps have to rock their of any species to its environment. Darwin envisioned that
bodies from side to side to keep their center of gravity over in any given environment, some individuals are better able
the weight-bearing leg. to find food, find a mate, care for their offspring, and es-
Early hominoid fossils show biomechanical evidence cape predators due to their genetic make-up. What were
of bipedal ambulation. A computerized tomography of the the advantages that bipedalism gave the early hominins?
internal structure of fossil femora from the six million- What were the environmental pressures favoring bipedal-
year-old Orronin tugenensis matches closely that of hu- ism over other forms of locomotion? Previous speculation
mans and is distinct from those of gorillas and chimps held that our ancestors became bipedal as they migrated
(1). Four million-year-old fossils of Australopithecus ana- out of the forest and into the savannah. Indeed the environ-
mensi have pelvic and lower extremity anatomical fea- ment was changing with increasing expanses of savannah
tures permitting habitual bipedalism. The iliac wings face emerging among shrinking areas of forest. Presumably,
laterally, but the pelvis is narrow and short bringing the standing and walking upright in the savannah permitted
hips closer to the sacrum. Strong abductor muscles pre- observations above the tall grasses and provided an ener-
vented rotation of the body on the femoral head and de- getic advantage when searching for food or avoiding preda-
creased body sway in midstance. In our bipedal ancestors, tors. The problem with this scenario is that recent research
the femoral neck-shaft angle medialized the knee, bring- at early hominoid sites indicates that the environment was
ing the feet closer to the center of gravity. The bicondylar not a savannah but was a lightly to densely wooded area,
angle of their knees was larger (∼10◦ ) than that of chim- suggesting that early hominins may have ventured out to
panzees (0◦ ), and their feet had well developed longitudi- the savannah in search of food but returned to the forest
nal arches with the great toe aligned parallel to the lesser to live. While more research is necessary to understand
toes. our history of bipedalism, it is evident that bipedalism
From an evolutionary standpoint, it is obvious that was the characteristic that first distinguished humans from
bipedalism would free up the hands to use weapons and apes.

223
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-14 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:55

224 Chapter 14 / Human Walking: Six Take-Home Lessons

LESSON 2: THE GAIT CYCLE IS THE Initially we can divide the gait cycle into two phases,
BASIS FOR UNDERSTANDING stance phase when the foot is in contact with the ground,
NORMAL AND PATHOLOGICAL and swing phase when the foot is in the air. Stance phase
HUMAN WALKING accounts for about 60% of the cycle, and swing accounts
for the remaining 40%. We can subdivide stance phase
A cycle is a recurrent series of events. We can understand into five periods known as: 1) initial contact, 2) load-
many complex phenomena by thinking in terms of a cycle. ing response, 3) midstance, 4) terminal stance, and 5)
Ancient astronomers achieved a better understanding of preswing. In addition, we can divide swing into three pe-
the seasons by observing the celestial cycles and the lunar riods: 1) initial swing, 2) midswing, and 3) terminal swing
cycle. Biologists proposed the cell cycle, and biochemists (Fig. 14-1).
devised the Krebs and the Calvin cycles. The gait cycle is First we will consider the five periods of stance phase.
to the biomechanics of walking as the Krebs cycle is to the The first period, initial contact, begins when the foot
biochemistry of intermediary metabolism. We speak of the touches the ground. The second period, loading response,
gait cycle as applying separately to each lower extremity, is a time of double limb support when both feet are on
and we define the gait cycle as the events that occur from the ground (usually 10% to 12% of stance). During this
one heel strike to the next. period, the limb accepts the weight of the body. Loading

FIGURE 14-1. The two phases of the gait cycle: stance phase when the foot is on the floor, and swing
phase when the foot is in the air. The gait cycle is further subdivided into tasks as well as periods
known as (1) initial contact, (2) loading response, (3) midstance, (4) terminal stance, and (5) preswing,
(6) initial swing, (7) midswing, and (8) terminal swing.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-14 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:55

Chapter 14 / Human Walking: Six Take-Home Lessons 225

response ends when the opposite foot leaves the ground. LESSON 3: KINEMATICS DESCRIBES
Midstance is the period when the center of gravity moves MOVEMENT OF THE BODY
over the foot, and the limb fully supports the weight of SEGMENTS, AND KINETICS
the body. The fourth period, terminal stance, begins with DESCRIBES THE FORCES CAUSING
the center of gravity directly over the foot and ends as the AND RESULTING FROM MOVEMENT
heel rises from the ground and the opposite foot contacts OF THE JOINTS
the ground. The last period of stance is preswing, another
time of double limb support, when the foot is about to be- Kinematic data provides information about joint move-
come airborne and the opposite limb progressively accepts ment and is expressed as degrees of displacement in any of
more weight. three planes of movement, i.e., sagittal, coronal, or trans-
The swing phase occupies less time of the cycle. Initial verse planes. Kinematically, we can describe the position,
swing begins when the foot leaves the ground and contin- velocity, and acceleration of body segments during walk-
ues as the knee flexes. Midswing begins with the knee in ing. From a motion analysis perspective, the body seg-
maximum flexion and ends when the leg is perpendicular ments include the HAT (head, arms, trunk), the pelvis, the
to the ground. The last period, terminal swing, begins with thigh, the shank (leg), and the foot.
the leg perpendicular to the ground and ends when the foot Kinetic data provides information about the forces
contacts the ground again. that act across the joints. Figure 14-2 shows a graphic

FIGURE 14-2. A graphic representation of the hip,

knee, and ankle kinematics and kinetics in the sagit-
tal plane during normal walking.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-14 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:55

226 Chapter 14 / Human Walking: Six Take-Home Lessons

Certain Temporal Spatial Definitions Help in logical task involving gross or fine body movement, mus-
Understanding Gait cles must convert the chemical energy in carbohydrates
Step: Advancement of the foot from and fats to the mechanical energy of actomyosin contrac-
toe off to heel strike tion. This conversion permits us to accomplish the various
Step length: Longitudinal distance between tasks of gait such as weight acceptance, single limb sup-
the two feet when both are on port, and limb advancement.
the ground Specific muscles are active during each period of the
Cadence: Number of steps per minute gait cycle (Fig. 14-3). In general, the muscles that are ac-
Double The time when both feet are on tive during stance prevent the stance limb from collaps-
support: the ground ing as the limb supports the body weight while the center
Float: A time when neither foot is on the of gravity advances. The major stance phase muscles in-
ground, as occurs in running clude the gluteus maximus, medius and minimus, tensor
Stride: One step by each foot, or one fascia lata, adductors, quadriceps, and the gastrocsoleus.
complete turn of the gait cycle
Stride length: The distance covered during one
turn of the gait cycle
Velocity: Stride length per cycle time,
measured in meters per minute

representation of the hip, knee, and ankle kinematics

and kinetics in the sagittal plane during normal walking.
The slope of each graph represents velocity of movement,
and degrees of displacement represents the position of
the joint relative to the neutral position. Note that hip
and knee flexion result in extension moments, and an-
kle dorsiflexion causes a plantar flexion moment at the
ankle.
The kinematic data show that at initial floor contact, the
hip is flexed, the knee is extended, and the ankle is close
to neutral. During early stance, the hip begins to extend,
the knee flexes slightly and then extends, and the ankle
dorsiflexes as the center of gravity passes over the foot.
In preswing, the hip is in extension, the knee begins to
flex, and the ankle is in plantarflexion. In swing, the hip
and knee flex, and the ankle dorsiflexes to clear the foot.
The knee rapidly extends in terminal swing after peak knee
flexion in preparation for floor contact. The kinetic data
illustrates that the ankle plantarflexion moment generated
in terminal stance is the largest moment acting across any
joint during any point in the gait cycle and is important in
advancing the limb.

LESSON 4: SPECIFIC MUSCLES ARE

ACTIVE DURING THE INDIVIDUAL
PHASES AND PERIODS OF THE GAIT
CYCLE, AND THE MAGNITUDE,
DURATION, AND TIMING OF
MUSCULAR ACTIVITY DETERMINES
THE QUALITY AND EFFICIENCY OF GAIT

Skeletal muscles hold the trunk upright against gravity, FIGURE 14-3. The normal timing of the lower extremity mus-
stabilize the supporting limb during stance, and move the cles measured by electromyography (EMG) recorded as a percent
advancing limb during swing. To accomplish any physio- of the gait cycle.
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-14 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:55

Chapter 14 / Human Walking: Six Take-Home Lessons 227

At initial contact and during loading response, active mus- 2. Pelvic obliquity: The pelvis drops prior to heel strike to
cles include the hamstrings, gluteus maximus, medius, increase the length of the leading leg and drops again
minimus, and the tensor fascia lata to stabilize the hip. with the trailing leg at toe-off.
The quadriceps stabilizes the knee. During the third and 3. Stance phase knee flexion: The knee flexes slightly dur-
fourth period, midstance and terminal stance, the gastroc- ing loading response, presumably to reduce the length
soleus complex stabilizes the ankle as the center of gravity of the extremity at midstance and decrease the vertical
moves over the foot. movement of the CoM.
The major swing phase muscles responsible for advanc- 4. Ankle rockers: Ankle dorsiflexion at heel-strike and
ing the limb include the iliopsoas and rectus femoris, active plantar flexion at toe-off increase the functional length
in early swing to flex the hip, followed in terminal swing by of the extremity. The heel rise from the foot flat position
the quadriceps to extend the leg, and finally the hamstrings raises the CoM when it is at its lowest.
to restrain knee extension at terminal swing. Tibialis an- 5. Transverse rotation: The pelvis internally rotates with
terior and other pretibial muscles are active throughout the leading leg and externally rotates with the trailing
swing to hold the foot in a position such that it clears the leg to functionally increase the leg length.
floor. 6. Genu valgus: Valgus at the knee permits us to walk with
Figure 14-3 shows the normal timing of the lower ex- a narrow base, resulting in less lateral shift of the CoM
tremity muscles measured by electromyography (EMG) during stance.
recorded as a percent of the gait cycle. Any abnormality
in the timing or magnitude of muscular activity or in the Whereas Inman proposed that stance phase knee flex-
motion of the joints during the gait cycle produces devia- ion decreased the vertical displacement of the trunk, Chil-
tions in motion. dress and Gard propose that this determinant, as well as
pelvic obliquity, have more of an effect as shock absorbers.
Nonetheless, in theory, these six determinants work to limit
the up-and-down and the side-to-side displacement of the
LESSON 5: SIX DETERMINANTS OF
center of mass to what is seen experimentally, which is
GAIT PROVIDE A MODEL FOR
about 2.5 cm in the adult.
BIOMECHANICAL MECHANISMS USED
TO MAXIMIZE WALKING EFFICIENCY

A half-century ago, Saunders, Inman, and Eberhart at the LESSON 6: THE RIDDLE OF THE
Biomechanical Laboratory of the University of California SPHINX: WALKING CHANGES WITH
at Berkeley described what they called the six determinants AGE
of gait as a logical way to understand how the body mini-
mizes the displacement of the center of mass (CoM) during According to Greek mythology, the Sphinx was a monster
walking. The determinants are discussed more critically that took up residence outside the gates of Thebes, asking
and in more detail in Chapter 1 of this book. Saunders, a riddle of all passers-by. If the passer-by gave an incor-
Inman, and Eberhart reasoned that if our legs worked as rect answer, the Sphinx would eat them. If the correct an-
the arms of a compass, our CoM would bounce up and swer was given, the Sphinx would kill himself. The riddle
down and jolt from side to side during each step, result- was: What goes on four legs in the morning, two legs at
ing in a cumbersome and inefficient gait. They reasoned noon, and three legs in the evening? For quite some time
that the six determinants worked to convert these harsh the Sphinx enjoyed many meals. Needless to say, the daily
jolts and bounces into a smooth sinusoidal movement, in- commerce of Thebes was a mess when Oedipus happened
creasing the efficiency as well as the cosmetics of walk- along. Oedipus thought about the question and gave the
ing. As Childress and Gard point out in the Commentary correct answer, MAN, whereupon the Sphinx proceeded to
for Chapter 1, researchers are currently investigating the die. The people of Thebes were so joyous that they made
biomechanical validity of these determinants. However, Oedipus their King. Oedipus realized that morning, noon,
minimizing the displacements of the CoM while walking and evening were metaphors for the stages of a person’s
is important, even though the mechanisms of action of the life. People crawl on all fours as a babies, walk on two legs
determinants offered half-century ago are undergoing as adults, and use a cane in old age. The myth is a great
scrutiny and revision. example of how even the ancient Greeks recognized that
The original six determinants of gait are: human gait adapts and changes during life (Chapters 7
and 8).
1. Pelvic tilt: The pelvis tilts anteriorly with the leading Infants usually can sit by six months, crawl at nine
leg at the beginning of stance and posteriorly with the months, walk with support (cruise) at eleven months, and
trailing leg at the end of stance, prior to the beginning walk without support (toddle) at a year. The age range for
of swing. toddling can be from eight to eighteen months with an
P1: KWW/KKL P2: KWW/KFR QC: KWW/PBR T1: KWW
GRBT092-14 Rose- 2252G GRBT092-Rose-v3.cls October 28, 2005 9:55

228 Chapter 14 / Human Walking: Six Take-Home Lessons

average around 11.5 months. We recognize the toddler’s The most obvious change in gait with advancing age
“tossing gait” as a result of the large vertical oscillations is that people walk more slowly. Velocity slows and step
of the body. Toddlers walk with an arched back (lumbar length decreases 15% to 20% by the seventh decade. The
lordosis), with legs far apart (wide base of support), and stance phase increases and swing decreases, thus increas-
the arms are elevated in a “high guard” position. It is ing the time of double limb support from 18 percent in
normal for the hips, the knees, and even the ankles to a 20-year-old to 26% in a 70-year-old. This pattern of gait
be flexed and relatively stiff. Toddlers take short choppy is a more conservative motion, and presumably protects
steps and contact the floor with a flat foot if they are not against falling. Finally, as Oedipus realized, an older per-
up on toes. A consistent heel strike appears by around son may use a cane while walking, resulting in three con-
1.5 years, and at that same time, about half the children tact points with the floor, which can supplement balance
will develop reciprocal arm swing. Toddlers walk an and provide a mechanical advantage to the abductors in
astounding amount each day. New walkers can cover the stance, thus decreasing the weight bearing forces across
distance of over 20 football fields in a single day! By the the hip.
age of two years 90 percent of children have reciprocal In conclusion, these six take-home lessons are meant to
arm swing, and the fixed hip flexion of infancy is gone. help students organize their thinking about human walk-
The knees show what Sutherland called a “knee flexion ing. Many concepts make up the intellectual understand-
wave” during loading. By four years the time and distance ing of gait, and facts are rapidly being added to this super-
parameters have stabilized, and by seven years a child has structure. However, we believe that these six take-home
a mature gait pattern. lessons (bipedalism, the gait cycle, kinematics and kinet-
The kinematics and kinetics of gait remain stable ics, muscle timing, walking efficiency, and adaptations and
throughout much of an adult’s life but can adapt and changes with age) can provide a blueprint for understand-
change with any temporary or permanent physiologi- ing the intellectual superstructure and for helping to or-
cal change. For instance, pregnancy temporarily changes ganize all the fascinating new facts that will continue to
kinematics and kinetics, as is obvious from the lumbering come out of motion analysis laboratories.
gait of a woman in late term. Obesity also changes both
the kinetics and the kinematics of gait, predisposing to
degenerative joint disease of the hips and knees. Trauma REFERENCES
resulting in joint or skeletal damage changes gait. Chronic
1. Galik K, Senut B, Pickford M, Gommery D, Treil J, Kuperavage AJ,
disease can influence gait such as occurs with people who Eckhardt RB. External and internal morphology of the BAR 1002’00
have Parkinson’s disease or after a person suffers a stroke. Orrorin tugenensis femur. Science 2004;305:1450 –1453.
P1: OSO/OVY P2: OSO/OVY QC: OSO/OVY T1: OSO
GRBT092-IND Rose- 2252G GRBT092-Rose-v4.cls October 28, 2005 21:5

• I N D E X
•
Note: Page numbers followed by f indicate figures; page numbers followed by t indicate tables.

Acceleration, 75 Bicycling, 33 peripheral factors contributing to,

Adenosine triphosphate (ATP), 77–78 Biomechanics, 34. See also Kinematics of 125
Adulthood walking physiological change in, 126
aging influencing gait in, 133–140, Biostereometrics, 60 remembering route/landmarks in,
134f, 135f, 136t, 137f, 138f Bipedalism 127
alcohol/alcoholism with, 141–142 evolution of, 23–24, 30 stepping reflex in, 125
exercise to improve gait with, 140–141 primal ancestors/apes distinguished by, stride length determining equations
falls with, 139–141 223 for, 130
gait adaptations with, 131–142, 132f, Body, 75 tuning phase in, 119
134f, 135f, 136t, 137f, 138f, 142f Bone density, 133–134 vestibular/visual system sensitivity
Age Bone health, 155–156 in, 126
energy expenditure with, 83–84 Bone mineral density, 126 visual attention in, 127
gait with, 42, 43f working memory in, 127
heart rate with walking for, 97 CAD. See Coronary artery disease effort/energy cost with, 122–123
Aging, gait influenced by, 133–140, 134f, Cadence, 226 flexible adaptation measurements of,
135f, 136t, 137f, 138f, 227–228 Calf muscles, 168 123–124, 124f
bone density with, 133–134 Calories, 79, 79t kinematics of, 120–121, 121f
central nervous system with, 134–135, Cancer, 156–157 ankle in, 121, 121f
134f, 135f Cardiovascular disease (CVD) head in, 120–121
exercise to improve, 140–141 equinovarus foot following, 174–176, hip in, 121
falls with, 139–141 174f, 175f knee in, 121
kinematics/kinetics with, 136, 137f physiological changes from walking pelvis in, 121
medications with, 135 lessening, 152f, 153–154 trunk in, 121
motor loss with, 135 walking influencing, 150–152, 151f, upper extremities in, 121
MRI showing, 134–135, 134f, 135f 152f kinetics of, 121–122
muscle mass with, 133–134 Center of gravity, 75 mechanics of changing walking speed
normal changes of, 133–135, 134f, Center of mass (COM) with, 123, 124f
135f ankle influencing, 9, 10f minimizing variance in measurements
postural sway with, 138–139, 138f defined, 75 of, 123–125
respiratory capacity with, 133–134 foot influencing, 7–9, 9f, 11f muscle timing with, 122, 122f
Romberg stance with, 138 ground reactions with, 55, 56f temporal-spatial parameters of, 120
sarcopenia with, 133 heel strike with, 7, 9, 9f Chimpanzee, 24–25, 25f
sensory loss with, 135 sinusoidal curve in displacements of, 3, Cognitive function, walking influencing,
slips with, 139–140 3f, 11f 157
static balance with, 136–139, 138f six determinants of gait with, 19, 21–22 COM. See Center of mass
strength with, 133–134 three foot rocker mechanisms with, 20, Concentrated force, 75
Tai Chi to improve, 140–141 20f Congenital hip dysplasia, 181–182, 181f,
temporal/spatial parameters with, Center of pressure (COP), 20, 22 182f
135–136, 136t defined, 75 COP. See Center of pressure
Alcohol/alcoholism, 141–142 ground reactions with, 56 Coronal plane, walking kinematics of,
Amputee gait, 186–188, 188f, 189f Central nervous system, aging influencing 46–47, 47f
cause of amputation in, 187, 188f, 189f gait with, 134–135, 134f, 135f Coronary artery disease (CAD), 151
ground reaction forces in, 188, 189f Cerebral palsy (CP) CP. See Cerebral palsy
level of amputation in, 187 crouched gait with, 177–178, 177f, 178f Crouched gait, 177–178, 177f, 178f
Angular acceleration, 75 energy efficiency of walking with, CVD. See Cardiovascular disease
Angular velocity, 76 95–96
Ankle heart rate in walking with, 96–97, 97f Decay stage, 33, 34f
children’s gait kinematics for, 121, 121f intersegmental moments with, 67–68, Degrees-of-freedom, 36
knee flexion influenced by, 9, 9f 67f joint motion with six, 38–39, 39f
plantar flexion/dorsiflexion with, 46, Chaslés’ Theorem, 38 Dementia, 157
46f Children’s gait, 119–128, 121f, 122f, 124f, Development stage, 33, 34f
power walking kinematics with, 159f, 130 Diabetes type II, 154–155, 154f
160 anticipatory control with, 123, 124f Disability, 94–96, 95t
power walking kinetics with, 160, 161f current directions with, 123–124, 124f Displacement, 36
rotations in, 13–18, 14f–16f development of, 119–128, 121f, 122f, Displacements of body during walking,
sagittal plane intersegmental moments 124f, 130 2–19, 3f–19f
with, 63–65, 64f, 67f behavioral change in, 127–128 alternative views of, 20, 20f
Anterior compartment syndrome, bone mineral density in, 126 ankle influencing, 9, 10f, 13–18,
170–171, 170f, 171f en bloc movement patterns in, 119 14f–16f
ATP. See Adenosine triphosphate factors interplaying in, 125 center of mass in, 3, 3f
Australopithecus afarensis, 29 first functional challenges with, foot influencing, 7–9, 9f, 13–18,
Australopithecus africanus, 23, 24f, 27f 119–120 14f–16f
Australopithecus anamensis, 28–29, 223 independent decision making in, 127 heel strike with, 7, 9, 9f
information transfer in, 127 interrupted light studies on, 10f
Basal metabolic rate, 79–80, 80f integration phase in, 119 knee flexion in, 4–9, 8f–11f
Behavioral change, 127–128 mechanisms underlying, 125 lateral, 9–11, 12f
Behavioral reconstruction, 28 neural change in, 126–127 pelvis list in, 4, 7f

229
P1: OSO/OVY P2: OSO/OVY QC: OSO/OVY T1: OSO
GRBT092-IND Rose- 2252G GRBT092-Rose-v4.cls October 28, 2005 21:5

230 Index

Displacements of body during walking oxidative metabolism in, 78 external, 75

(cont.) storage/utilization of, 77–78 ground reactions, 75
pelvis rotation in, 4, 6f transductions in body of, 77, 78f inertial, 75
rocker-based inverted pendulum model resting metabolism with, 79–81, 80f, 81f internal, 75
with, 20, 20f sitting with, 80–81, 80f, 81f intersegmental, 75
rotations in ankle with, 13–18, 14f–16f standing metabolism with, 80f, 81 joint contact, 75
rotations in foot with, 13–18, 14f–16f work power in, 87–96, 88f, 90f, 91t, normal, 75
rotations in leg with, 12–13, 13f 92f–94f, 93t, 95t shear, 75
rotations in shoulders with, 12 Energy expenditure index (EEI), 98–100, Force couple, 75
rotations in thigh with, 12–13, 13f 98f–100f, 100t Forward dynamics, 75
rotations in thorax with, 12 Equinovarus foot, 174–176, 174f, 175f Free-body-diagram, intersegmental
rotations in transverse plane with, 11 Eulerian angle description, 37–38, 38f forces with, 58, 58f, 59f
simple model with, 3–4, 5f Evolution of walking, 23–31 Free speed-walking, 33
six determinants of gait with, 19–20, Australopithecus afarensis in, 29 Free walking, 157–158, 158t
21–22 Australopithecus africanus in, 23, 24f, FS. See Foot strike
summary of, 20–21 27f Functional electrical stimulation (FES),
three foot rocker mechanisms with, 20, Australopithecus anamensis in, 28–29 209–219, 213f–216f, 217f
20f, 22 behavioral reconstruction with, 28 energy expenditure with, 218
Distance walked bipedalism in, 23–24, 30 future developments for, 218–219
energy efficiency of walking with, chimpanzee in, 24–26, 25f mechanics of walking with, 216–218,
88–91, 90f hominin phylogeny in, 24f 217f
heart rate with, 98–100, 98f–100f, 100t Homo erectus in, 26–27 initial contact in, 217
Distributed force, 75 Homo ergaster in, 26–27, 27f, 29–30 initial swing in, 218
Dorsiflexion, 13, 16f, 46, 46f, 227 Homo genus in, 23–25, 24f, 25f midstance in, 217
post polio syndrome with, 172–173, Homo habilis in, 26 midswing in, 218
172f, 173f Homo neanderthalensis in, 27 preswing in, 218
weakness in, 168 Homo rudolfensis in, 26 terminal stance in, 217–218
Double support, 226 Homo sapiens in, 27 terminal swing in, 218
human walking fossil evidence with, weight acceptance in, 217
Early swing, 41 28–30 muscular response to, 209–211
EEI. See Energy expenditure index Orrorin tugenensis in, 25 contractile properties in, 209–210
Electromyography (EMG) overview of, 25–28, 27f endurance with, 210–211, 211t
MUAPT with, 117 summary of, 30 strength with, 210–211, 211t
muscle interpretation with, 117–118, Exercise, 140–141. See also Power walking research directions for, 218–219
117f External force, 75 stimulation patterns for, 215–216,
phasic muscle action in, 113f, 114, 117f External hip flexor moment, 63, 63f 216f
En bloc movement patterns, 119 External moment, 76 synthesis of gait with, 214–216, 216f
Energetics of walking, 77–100 technologies for, 211–214, 213f–215f
children’s gait in, 122–123 Falls hybrid systems as, 212–214
efficiency in, 87–96, 88f, 90f, 91t, aging influencing gait with, 139–141 implanted systems as, 212,
92f–94f, 93t, 95t balance predictors of, 139–140 213f–215f
body segments with, 87–88, 88f exercise to avoid, 140–141 intramuscular electrodes with
disability with, 94–96, 95t fear of, 140 percutaneous leads as, 211–212
energy expenditure per distance gait predictors of, 139 physiologic cost index with, 213
walked with, 88–91, 90f Tai Chi helping avoid, 140–141 reciprocating gait orthosis with,
energy expenditure per step with, Fast fatigable fibers, 105 213
91–92, 91t Fast glycolytic fibers, 105 surface stimulation as, 211
immobilization with, 93–94, 93f, 94f Fast oxidative-glycolytic fibers (FOG transcutaneous stimulation as, 211
slope walking with, 91–92, 91f, 92t fibers), 105
terrain with, 91 Fast-twitch muscle fiber, 105–107, 106f, Gait. See also Gait analysis; Locomotion
energy expenditure in, 81–87, 82f, 82t, 107t adaptations in adulthood for, 131–142,
86f FES. See Functional electrical stimulation 132f, 134f, 135f, 136t, 137f, 138f,
age with, 83–84 Float, 226 142f
constant of body maintenance with, FO. See Foot-off aging influencing, 133–140, 134f, 135f,
83 FOG fibers. See Fast oxidative-glycolytic 136t, 137f, 138f
gender with, 86–87 fibers bone density with, 133–134
speed of walking with, 81–83, 82f, 82t Foot central nervous system with,
weight with, 84–86, 85f, 86f dorsiflexion with, 13, 16f 134–135, 134f, 135f
energy expenditure with FES in, 218 Hicks’ windlass action with, 14 exercise to improve, 140–141
heart rate energy expenditure with, knee flexion influenced by, 7–9, 9f falls with, 139–141
96–100, 97f–100f, 100t metatarsal heads’ angle with, 17f kinematics/kinetics with, 136, 137f
age in, 97 plantar flexion with, 13, 16f medications with, 135
distance walked in, 98–100, 98f–100f, rising of toes with, 17f motor loss with, 135
100t rocker mechanisms with, 20, 20f, 22, MRI showing, 134–135, 134f, 135f
energy expenditure index for, 227 muscle mass with, 133–134
98–100, 98f–100f, 100t rotations in, 13–18, 14f–16f normal changes of, 133–135, 134f,
gender in, 97–98 shoes of, 17f 135f
lying metabolism with, 79–80, 80f, Foot-off (FO), 39–40, 40f, 40t postural sway with, 138–139, 138f
81f Foot progression angle, 49–50, 49f respiratory capacity with, 133–134
measurement/units for, 78–79, 79t Foot strike (FS), 39–40, 40f, 40t Romberg stance with, 138
metabolic energy with, 77–78, 78f Force, 75 sarcopenia with, 133
ATP for, 77–78 concentrated, 75 sensory loss with, 135
glycolytic metabolism in, 78 distributed, 75 slips with, 139–140
P1: OSO/OVY P2: OSO/OVY QC: OSO/OVY T1: OSO
GRBT092-IND Rose- 2252G GRBT092-Rose-v4.cls October 28, 2005 21:5

Index 231

static balance with, 136–139, 138f mid swing in, 40–41, 40t, 41t Ground reaction force (GRF), 53–58, 54f,
strength with, 133–134 opposite foot-off in, 39–40, 40f, 40t 56f–58f
Tai Chi to improve, 140–141 opposite foot strike in, 39–40, 40f, 40t amputee gait with, 188, 189f
temporal/spatial parameters with, periods in, 40–41, 40t, 41t body leaning with, 53–54, 54f
135–136, 136t phases in, 40, 40f, 40t center of mass with, 55, 56f
alcohol/alcoholism with, 141–142 second double limb support in, center of pressure with, 56
amputee, 186–188, 188f, 189f 40–41, 40t, 41t defined, 75
cause of amputation in, 187, 188f, single limb support in, 40–41, 40t, 41t gait cycle with, 55–57, 56f, 57f
189f stance phase in, 40, 40f, 40t simulation analysis with, 199–201,
ground reaction forces in, 188, 189f step length in, 41f, 42 200f
level of amputation in, 187 stride length in, 41f, 42, 42t vertical jump with, 55, 56f
cerebral palsy with, 67–68, 67f swing phase in, 40, 40f, 40t, 41 Ground reaction torque, 76
children’s, 120–125, 121f, 122f, 124f terminal swing in, 40–41, 40t, 41t
anticipatory control with, 123, 124f walking speed in, 42 HAT. See Head, arms, and trunk
current directions with, 123–124, maturation, 42, 43f Head, children’s gait kinematics for,
124f pathological intersegmental moments 120–121
effort/energy cost with, 122–123 with, 67–68, 67f Head, arms, and trunk (HAT), 34
flexible adaptation measurements of, pregnancy with, 131–133, 132f, 134f, energy efficiency with, 87–88, 88f
123–124, 124f 135f kinetic energy changes in, 68, 70, 70f,
kinematics of, 120–121, 121f balance during, 132, 132f 71f
kinetics of, 121–122 cinematographic analysis of, 132 Health, walking for, 149–162, 151f–154f,
mechanics of changing walking exercise during, 133 156f, 158f–161f, 158t
speed with, 123, 124f hip during, 132, 132f bone health with, 155–156
minimizing variance in kinematic/kinetic gait analysis of, cancer with, 156–157
measurements of, 123–125 132 cardiovascular disease with, 150–152,
muscle timing with, 122, 122f pelvis during, 132, 132f 151f, 152f
temporal-spatial parameters of, 120 physiological changes of, 131 cardiovascular physiological changes
cycle posture of, 131 lessening, 152f, 153–154
ankle in, 46f six determinants of, 19–20, 21–22, 227 cognitive function with, 157
foot progression angle in, 49f Gait analysis, 165–182 coronary artery disease with, 151
ground reactions with, 55–57, 56f, anterior compartment syndrome in, dementia with, 157
57f 170–171, 170f, 171f diabetes type II with, 154–155, 154f
hip in, 45f, 47f, 48f case studies in, 170–182, 170f–175f, lifespan with, 150–157, 151f–154f
knee in, 45f, 49f 177f–179f, 181f, 182f obesity with, 155, 156f
muscle action during, 111, 111t, cerebral palsy-crouched gait in, osteoarthritis with, 156
226–227, 226f 177–178, 177f, 178f osteoporosis with, 155–156
normal/pathological walking congenital hip dysplasia in, 181–182, peripheral arterial disease with,
understood through, 224–225, 224f 181f, 182f 152–153
pelvis in, 44f, 47f, 48f development of, 165 power walking for fitness with,
development of, 119–128, 121f, 122f, equinovarus foot in, 174–176, 174f, 157–162, 158f–161f, 158t
124f, 130 175f stroke with, 152
behavioral change in, 127–128 functional perspective of, 165–167 USA pedometer data on, 150
bone mineral density in, 126 gait cycle organization/interpretation USA survey data on, 149–150
en bloc movement patterns in, 119 with, 165–167 Heart rate with walking, 96–100,
factors interplaying in, 125 single limb support in, 166–167 97f–100f, 100t
first functional challenges with, swing limb advancement in, 167 age in, 97
119–120 weight acceptance in, 166 distance walked in, 98–100, 98f–100f,
independent decision making in, 127 gait deviation characteristics in, 100t
information transfer in, 127 168–169 energy expenditure index for, 98–100,
integration phase in, 119 mobility assessment in, 167–168 98f–100f, 100t
mechanisms underlying, 125 musculoskeletal pathology in, 168–169 gender in, 97–98
neural change in, 126–127 weak calf muscles in, 168 Heel strike, center of mass with, 7, 9, 9f
peripheral factors contributing to, weak dorsiflexors in, 168 Hicks’ windlass action, 14
125 weak hip abductors in, 169 Hip
physiological change in, 126 weak hip extensors in, 169 abduction/adduction, 47–48, 47f
remembering route/landmarks in, weak quadriceps in, 168–169 children’s gait kinematics for, 121
127 pathology’s significance with, 168 congenital hip dysplasia in, 181–182,
stepping reflex in, 125 post CVA with equinovarus foot in, 181f, 182f
stride length determining equations 174–176, 174f, 175f coronal plane intersegmental moments
for, 130 post polio syndrome in, 172–173, 172f, with, 65–66, 66f
tuning phase in, 119 173f, 179–180, 179f flexion/extension, 44–45, 45f
vestibular/visual system sensitivity dorsiflexion/plantar strength with, pain during jogging with, 181–182,
in, 126 172–173, 172f, 173f 181f, 182f
visual attention in, 127 genu recurvatum with, 179–180, power walking kinematics with, 159f,
working memory in, 127 179f 160
events, 39–42, 40f, 40t, 41f, 41t, 42t strength assessment in, 167 power walking kinetics with, 160, 161f
early swing in, 41 temporal gait characteristics in, 167 rotation, 48–49, 48f
foot-off in, 39–40, 40f, 40t testing for impairments in, 167–168 sagittal plane intersegmental moments
foot strike in, 39–40, 40f, 40t Gender with, 64f, 65, 67f
initial double limb support in, 40–41, energy expenditure with, 86–87 weak abductors with, 169
40t, 41t heart rate with walking for, 97–98 weak extensors with, 169
initial swing in, 40–41, 40t, 41t Genu recurvatum, 179–180, 179f Homo erectus, 26–27
late swing in, 41 Gimbal lock situation, 38 Homo ergaster, 26–27, 27f, 29–30
P1: OSO/OVY P2: OSO/OVY QC: OSO/OVY T1: OSO
GRBT092-IND Rose- 2252G GRBT092-Rose-v4.cls October 28, 2005 21:5

232 Index

Homo genus, 23–25, 24f, 25f trunk with, 121 forces with joint movement described
Homo habilis, 26 upper extremities with, 121 by, 225–226, 225f
Homo neanderthalensis, 27 coronal plane in, 46–47, 47f kinematics v., 225–226
Homo rudolfensis, 26 description of joint movement for, Kinetics of walking, 53–73
Homo sapiens, 27 36–39, 36f–39f aging influencing gait with, 136, 137f
Horse power, 79, 79t Chaslés’ Theorem in, 38 children’s gait in, 121–122
Eulerian angle description in, 37–38, energy with, 68–72, 69f–72f
ICR. See Instantaneous center of rotation 38f ground reactions in, 53–58, 54f, 56f–58f
Immobilization, energy efficiency of gimbal lock situation with, 38 body leaning with, 53–54, 54f
walking with, 93–94, 93f, 94f instantaneous center of rotation in, center of mass with, 55, 56f
Induced acceleration analysis, 201–203, 36f, 37, 37f center of pressure with, 56
202f planar motion in, 36–37, 36f, 37f gait cycle with, 55–57, 56f, 57f
Inertia Reuleaux’s method in, 37, 37f vertical jump with, 55, 56f
defined, 76 rotational three-dimensional motion intersegmental forces with, 58–63, 58f,
intersegmental forces with, 60 in, 37–38, 38f 59f, 62f
Inertial force, 75 screw displacement axis in, 38–39, biostereometrics for, 60
Initial double limb support, 40–41, 40t, 39f free-body-diagram of foot for, 58, 58f,
41t six degree-of-freedom in, 38–39, 39f 59f
Initial foot contact gait events with, 39–42, 40f, 40t, 41f, inertia in, 60
mechanics of walking with FES in, 217 41t, 42t inverse dynamics for, 61–62
phasic muscle action of, 111–112, 111f, early swing in, 41 stereophotometrics for, 60
111t foot-off in, 39–40, 40f, 40t intersegmental moments/powers with,
Initial swing, 40–41, 40t, 41t foot strike in, 39–40, 40f, 40t 58–68, 58f, 59f, 62f–64f, 66f, 67f
mechanics of walking with FES in, initial double limb support in, 40–41, ankle/sagittal plane in, 63–65, 64f, 67f
218 40t, 41t cerebral palsy example of, 67–68,
phasic muscle action of, 111t, 112–113 initial swing in, 40–41, 40t, 41t 67f
Instantaneous center of rotation (ICR), late swing in, 41 external hip flexor moment in, 63, 63f
36f, 37, 37f mid swing in, 40–41, 40t, 41t hip/coronal plane in, 65–66, 66f
Integration phase, gait development with, opposite foot-off in, 39–40, 40f, 40t hip/sagittal plane in, 64f, 65, 67f
119 opposite foot strike in, 39–40, 40f, 40t internal hip flexor moment in, 63,
Internal force, 75 periods in, 40–41, 40t, 41t 63f
Internal hip flexor moment, 63, 63f phases in, 40, 40f, 40t knee/coronal plane in, 66, 66f
Internal intersegmental moment, 76 second double limb support in, knee/sagittal plane in, 64f, 65, 67f
Intersegmental force, 75 40–41, 40t, 41t pathological gait example of, 67–68,
Intersegmental power, 76 single limb support in, 40–41, 40t, 41t 67f
Inverse dynamics, 61–62, 76 stance phase in, 40, 40f, 40t push-off in, 64–65
step length in, 41f, 42 roll-off in, 64
Jogging, hip pain during, 181–182, 181f, stride length in, 41f, 42, 42t talocrural joint/sagittal plane in,
182f swing phase in, 40, 40f, 40t, 41 63–65, 64f, 67f
Joint contact force, 75 terminal swing in, 40–41, 40t, 41t power walking in, 160–161, 160f, 161f
Joint movement walking speed in, 42 static equilibrium with, 54f, 58f, 62f
description of, 36–39, 36f–39f gait maturation with, 42, 43f summary of, 72–73
Chaslés’ Theorem in, 38 hip abduction/adduction in, 47–48, 47f work with, 68–72, 69f–72f
Eulerian angle in, 37–38, 38f motion curves for, 42–43 Knee
gimbal lock situation with, 38 motion measurement principles for, children’s gait kinematics for, 121
instantaneous center of rotation in, 33–36, 34f, 35f coronal plane intersegmental moments
36f, 37, 37f degrees-of-freedom in, 36 with, 66, 66f
planar motion in, 36–37, 36f, 37f displacement in, 36 flexion/extension of, 45–46, 45f
Reuleaux’s method in, 37, 37f external coordinate system in, 34–35, power walking kinematics with, 159f
kinetics forces with, 225–226, 225f 35f power walking kinetics with, 160, 160f
rotational three-dimensional, 37–38, markers in, 34–35, 35f rotation, 49, 49f
38f scalars in, 34 sagittal plane intersegmental moments
screw displacement axis in, 38–39, vectors in, 34 with, 64f, 65, 67f
39f overview of, 33 Knee flexion
six degree-of-freedom in, 38–39, 39f power walking in, 158f, 159–160, 159f ankle influencing, 9, 10f
talocrural joint/sagittal plane in, 63–65, sagittal plane in, 43–46, 44f–46f foot influencing, 7–9, 9f
64f, 67f ankle plantar flexion/dorsiflexion heel strike with, 7, 9, 9f
Joules, 79, 79t with, 46, 46f walking displacements with, 4–9, 8f–11f
anterior pelvic tilt with, 43–44, 44f Knuckle-walking, 26, 223
Kinematics hip flexion/extension with, 44–45, 45f
body segment movement described by, knee flexion/extension with, 45–46, Late swing, 41
225–226 45f LBM. See Lean body mass
defined, 76 summary of, 50 Lean body mass (LBM), 84–85
kinetics v., 225–226 three stages of, 33, 34f Leg rotation, 12–13, 13f
Kinematics of walking, 33–50 transverse plane in, 48–50, 48f, 49f Lever arm, 76
aging influencing gait with, 136, foot progression angle with, 49–50, Lifespan
137f 49f bone health with, 155–156
children’s gait in, 120–121, 121f hip rotation with, 48–49, 48f cancer with, 156–157
ankle with, 121, 121f knee rotation with, 49, 49f cardiovascular disease with, 150–152,
head with, 120–121 pelvic rotation with, 48, 48f 151f, 152f
hip with, 121 Kinetic energy, 76 physiological changes lessening,
knee with, 121 Kinetics 152f, 153–154
pelvis with, 121 defined, 76 cognitive function with, 157
P1: OSO/OVY P2: OSO/OVY QC: OSO/OVY T1: OSO
GRBT092-IND Rose- 2252G GRBT092-Rose-v4.cls October 28, 2005 21:5

Index 233

coronary artery disease with, 151 Midswing Opposite foot-off (OFO), 39–40, 40f, 40t
dementia with, 157 mechanics of walking with FES in, 218 Opposite foot strike (OFS), 39–40, 40f, 40t
diabetes type II with, 154–155, 154f phasic muscle action of, 111t, 113 Osteoarthritis, 156
obesity with, 155, 156f Mobility, gait assessment of, 167–168 Osteoporosis, 155–156
osteoarthritis with, 156 Moment, 76
osteoporosis with, 155–156 Moment arm, 76 PAD. See Peripheral arterial disease
peripheral arterial disease with, Momentum, 76 Parallel muscle fiber, 107
152–153 Motor loss, aging influencing gait with, Paralysis, restoring walking after,
stroke with, 152 135 209–219, 213f–216f, 217f
walking influencing, 150–157, MUAPT. See Multiple motor unit action conclusion to, 219–220
151f–154f potential trains electrical stimulation technologies for,
Linear acceleration, 75 Multipennate muscle fiber, 107 211–214, 213f–215f
Linear velocity, 76 Multiple motor unit action potential hybrid systems as, 212–214
Loading response, phasic muscle action trains (MUAPT), 117 implanted systems as, 212,
of, 111t, 112, 112f Muscle, 103–118 213f–215f
Locomotion, 1–22 biomechanics of, 110–111, 110f intramuscular electrodes with
displacements of body during, 2–19, children’s gait with, 122, 122f percutaneous leads as, 211–212
3f–19f contraction speed with, 109 physiologic cost index with, 213
alternative views of, 20, 20f, 22 EMG interpretation for, 117–118, 117f, reciprocating gait orthosis with, 213
ankle influencing, 9, 10f 226f, 227 surface stimulation as, 211
center of mass in, 3, 3f excitation-contraction couple with, transcutaneous stimulation as, 211
foot influencing, 7–9, 9f 103–104, 105f, 106f energy expenditure with FES for, 218
heel strike with, 7, 9, 9f fiber types, 105–107, 106f, 107t FES stimulation patterns for, 215–216,
interrupted light studies on, 10f fast-twitch, 105–107, 106f, 107t 216f
knee flexion in, 4–9, 8f–11f slow-twitch, 105 future developments for, 218–219
lateral, 9–11, 12f type I, 105 mechanics of walking with FES for,
pelvis list in, 4, 7f type II, 105–107, 106f, 107t 216–218, 217f
pelvis rotation in, 4, 6f gait cycle with, 111, 111t, 226–227, initial contact in, 217
rocker-based inverted pendulum 226f initial swing in, 218
model with, 20, 20f length relationships with, 108–109, midstance in, 217
rotations in ankle with, 13–18, 109f, 110f midswing in, 218
14f–16f mechanical properties of, 108 preswing in, 218
rotations in foot with, 13–18, 14f–16f motor control with, 107, 108f terminal stance in, 217–218
rotations in leg with, 12–13, 13f motor unit of, 104–107, 106f, 107t, 108f terminal swing in, 218
rotations in shoulders with, 12 phasic action with walking of, 111–117, weight acceptance in, 217
rotations in thigh with, 12–13, 13f 111f–116f, 111t muscular response to FES for, 209–211
rotations in thorax with, 12 EMG with, 114, 114f, 117f contractile properties in, 209–210
rotations in transverse plane with, 11 initial foot contact in, 111–112, 111f, endurance with, 210–211, 211t
simple model with, 3–4, 5f 111t strength with, 210–211, 211t
six determinants of gait with, 19–20, initial swing in, 111t, 112–113 research directions for, 218–219
21–22 loading response in, 111t, 112, synthesis of gait with FES for, 214–216,
summary of, 20–21 112f 216f
three foot rocker mechanisms with, midstance in, 111t, 112, 112f Paraplegia, 209. See also Paralysis,
20, 20f, 22 midswing in, 111t, 113 restoring walking after
model of bipedal, 3–4, 5f pre-swing in, 111t, 112, 113f PCI. See Physiologic cost index
process of walking in, 2 terminal stance in, 111t, 112, 113f Pelvis
Lying, metabolism of, 79–80, 80f, 81f terminal swing in, 111t, 113, 113f anterior tilt of, 43–44, 44f
response to FES of, 209–211 children’s gait kinematics for, 121
Magnetic resonance imaging (MRI), contractile properties in, 209–210 displacements of, 3–7, 3f–6f
aging influencing gait shown in, endurance with, 210–211, 211t knee flexion with, 4–9
134–135, 134f, 135f strength with, 210–211, 211t list in, 4, 7f
Mass, 76 structure, 103, 104f, 105f rotation in, 4, 6f
Mass moment of inertia, 76 tension relationships with, 108–109, rotations in leg with, 12–13, 13f
Measurement conversions, 79t 109f, 110f simple model with, 3–4, 5f
Mechanical energy, 76 whole muscle structure with, 107–110, transverse rotation of, 14f, 227
Mechanics of walking with FES in, 217f 108f–110f walking speed variation in, 3f
Medications, 135 multipennate architecture in, 107 power walking kinematics with, 158f,
Metabolism, 77–78, 78f parallel architecture in, 107 160
ATP for, 77–78 unipennate architecture in, 107 rotation, 48, 48f, 227
basal metabolic rate with, 79–80, 80f Muscle mass, 133–134 Periods, 40t, 41t
glycolytic, 78 Musculoskeletal pathology, 168–169 Peripheral arterial disease (PAD),
lying, 79–80, 80f, 81f weak calf muscles in, 168 152–153
oxidative, 78 weak dorsiflexors in, 168 Perturbation analysis, 203–204, 204f
resting, 79–81, 80f, 81f weak hip abductors in, 169 Phases, 40f, 40t
sitting, 80–81, 80f, 81f weak hip extensors in, 169 Physiologic cost index (PCI), 213
standing, 80f, 81 weak quadriceps in, 168–169 Physiological change, development of
storage/utilization in, 77–78 gait with, 126
transductions in body of, 77, 78f Neural change, 126–127 Planar motion, 36–37, 36f, 37f
Metatarsals, angle of head with, 17f Normal force, 75 Plantar flexion, 13, 16f, 46, 46f
Mid swing, 40–41, 40t, 41t push-off in, 64–65
Midstance Obesity, 155, 156f roll-off in, 64
mechanics of walking with FES in, 217 OFO. See Opposite foot-off Plantar strength, polio syndrome with,
phasic muscle action of, 111t, 112, 112f OFS. See Opposite foot strike 172–173, 172f, 173f
P1: OSO/OVY P2: OSO/OVY QC: OSO/OVY T1: OSO
GRBT092-IND Rose- 2252G GRBT092-Rose-v4.cls October 28, 2005 21:5

234 Index

Post polio syndrome, 172–173, 172f, 173f, Screw displacement axis (SDA), 38–39, Stroke, 152
179–180, 179f 39f Swing limb advancement, gait cycle with,
dorsiflexion/plantar strength with, Second double limb support, 40–41, 40t, 167
172–173, 172f, 173f 41t Swing phase, 40, 40f, 40t, 41
genu recurvatum with, 179–180, 179f Sensory loss, aging influencing gait with,
Postural sway, aging influencing gait 135 Tai Chi, 140–141
with, 138–139, 138f Shear force, 75 Talocrural joint, 63–65, 64f, 67f
Potential energy, 76 Shoes, 17f TBW. See Total body weight
Power, 76 Shoulder rotation, 12 Terminal stance
Power walking, 157–162, 158f–161f, 158t Simulation of walking, 193–206, mechanics of walking with FES in,
conclusion to, 161 194f–197f, 199f–202f, 204f 217–218
free walking v., 157–158, 158t analyzing, 199–204, 200f–202f, 204f phasic muscle action of, 111t, 112, 113f
kinematics of, 158f, 159–160, 159f ground reaction force decomposition Terminal swing, 40–41, 40t, 41t
kinetics of, 160–161, 160f, 161f in, 199–201, 200f mechanics of walking with FES in, 218
Pre-swing induced acceleration analysis in, phasic muscle action of, 111t, 113, 113f
mechanics of walking with FES in, 218 201–203, 202f Terrain, 91
phasic muscle action of, 111t, 112, 113f perturbation analysis in, 203–204, Thigh rotations, 12–13, 13f
Pregnancy, 131–133, 132f, 134f, 135f 204f Thorax rotations, 12
balance during, 132, 132f quantifying contributions to support Three foot rocker mechanism, 20, 20f, 22
cinematographic analysis of, 132 in, 199–201, 200f Torque, 76
exercise during, 133 quantifying muscle action of double Total body weight (TBW), 84–85
hip during, 132, 132f support in, 203–204, 204f Transverse plane
kinematic/kinetic gait analysis of, 132 quantifying muscle actions of foot progression angle with, 49–50, 49f
pelvis during, 132, 132f single-limb stance in, 201–203, hip rotation with, 48–49, 48f
physiological changes of, 131 202f knee rotation with, 49, 49f
posture of, 131 challenges for, 204–206 pelvic rotation with, 48, 48f
Pressure, 76 analysis, 206 rotations in, 11
Prostheses, lower limb, 185–190, 186f, modeling, 205 walking kinematics of, 48–50, 48f, 49f
188f, 189f simulation, 205–206 Trunk, children’s gait kinematics for, 121
alignment of, 189–190 examples of, 199–204, 200f–202f, Tuning phase, gait development with,
amputee gait with, 186–188, 188f, 189f 204f 119
cause of amputation in, 187, 188f, experiments with, 206 Type I muscle fiber, 105
189f future directions in, 204–206 Type II muscle fiber, 105–107, 106f, 107t
ground reaction forces in, 188, 189f muscle-driven simulation in, 196f,
level of amputation in, 187 197–198, 199f Unipennate muscle fiber, 107
components of, 187–189, 189f musculoskeletal system for, 196–197, Upper extremities, children’s gait
SACH feet as, 188 196f, 197f kinematics for, 121
conclusions to, 190 stage 1 in, 194f, 196–197, 196f, 197f
fundamental assumption with, 185–186 stage 2 in, 194f, 196f, 197–198, 199f Vectors, 34
limitations in current studies on, 190 stage 3 in, 194f, 198–199, 200f Velocity, 76, 226
shock absorption with, 189 stage 4 in, 194f, 199–204, 200f–202f, Vertical jump, 55, 56f
socket design for, 190 204f Vestibular system sensitivity, 126
Push-off, 64–65 testing of, 198–199, 200f Visual system sensitivity, development of
theory with, 206 gait with, 126
Quadriceps, 168–169 Single limb support, 40–41, 40t, 41t
gait cycle with, 166–167 Walking, 2
Reciprocating gait orthosis (RGO), 213 Sitting, metabolism of, 80–81, 80f, displacements of body during, 2–19,
Respiratory capacity, aging influencing 81f 3f–19f
gait with, 133–134 Slips, 139–140. See also Falls energetics of, 77–100
Resting metabolism, 79–81, 80f, 81f Slope walking, 91–92, 91f, 92t evolution of, 23–31
Reuleaux’s method, 37, 37f Slow-twitch muscle fiber, 105 for health, 149–162, 151f–154f, 156f,
RGO. See Reciprocating gait orthosis Solid ankle cushion heal (SACH), 188 158f–161f, 158t
Rhythmic stage, 33, 34f Sphinx, riddle of, 227–228 kinematics of, 33–50
Rigid body, 75 Spinal cord injury (SCI), 209. See also kinetics of, 53–73
Rocker-based inverted pendulum model, Paralysis, restoring walking after after paralysis, 209–219, 213f–216f,
20, 20f Stance phase, 40, 40f, 40t 217f
Roll-off, 64 Standing, metabolism of, 80f, 81 power, 157–162, 158f–161f, 158t
Romberg stance, 138 Static balance, 136–139, 138f process of, 2
Rotational three-dimensional motion, Step simulation of, 193–206, 194f–197f,
37–38, 38f defined, 226 199f–202f, 204f
energy efficiency of walking with, two requisites for, 2
SACH. See Solid ankle cushion heal 91–92, 91t Walking speed, 42
Sagittal plane Step length, 41f, 42, 226 children’s gait with, 123, 124f
ankle plantar flexion/dorsiflexion with, Stepping reflex, 125 energy expenditure index for, 98–100,
46, 46f Stereophotogrammetry, 60 98f–100f, 100t
anterior pelvic tilt with, 43–44, 44f Stereophotometrics, 60 energy expenditure with, 81–83, 82f, 82t
hip flexion/extension with, 44–45, 45f Strength mechanics of changing, 123, 124f
knee flexion/extension with, 45–46, aging influencing gait with, 133–134 Weight, 76
45f gait assessment of, 167 energy expenditure with, 84–86, 85f, 86f
walking kinematics of, 43–46, 44f–46f Stride, 226 Weight acceptance
Sarcopenia, 133 Stride length, 41f, 42, 42t gait cycle with, 166
Scalars, 34 defined, 226 mechanics of walking with FES in, 217
SCI. See Spinal cord injury equations for, 130 Work, 76