Clinical Biomechanics 19 (2004) 71–77

www.elsevier.com/locate/clinbiomech

The role of selected extrinsic foot muscles during running

Kristian M. O’Connor *, Joseph Hamill
Biomechanics Laboratory, University of Massachusetts, Amherst, MA 01003, USA
Received 20 November 2002; accepted 4 September 2003

Abstract
Objective. To determine the kinematic, kinetic and EMG responses to perturbations of the foot by running in varus, neutral, and
valgus-wedged shoes.
Design. Within-subjects study comparing kinematics, kinetics and EMG while running in three different shoe conditions.
Background. Excessive pronation has been cited as a key contributor to many types of running injuries. However, the roles of the
extrinsic foot muscles (those that control motion of the foot) during the stance phase of running have not been adequately identified,
which is critical to determining the relationship between pronation and injury.
Methods. Ten males ran in varus, valgus, and neutral-wedged shoes while three-dimensional kinematic and kinetic data and
EMG data were collected. Surface EMG data were collected from the tibialis anterior, peroneus longus, medial and lateral gas-
trocnemius, and soleus. Indwelling EMG was obtained from the tibialis posterior. The net joint moment, power, and total positive
and negative work was calculated in the frontal plane. EMG onset, offset, and integrated values were reported.
Results. The maximum eversion angle, maximum inversion moment and total negative work done in the frontal plane were
greatest while running in the valgus shoe and least in the varus shoe. The greater joint moment was not accompanied by changes in
muscle activation patterns, although the tibialis posterior data were inconclusive in this respect.
Conclusions. Greater pronation leads to greater energy absorption in the foot invertor muscles and tendons. While not con-
clusive, the EMG data suggest that for these muscles there was not a neuromuscular adaptation to the perturbation.

Relevance
This study reinforces the hypothesized link between excessive pronation and injury and provides valuable insight into the
muscular responses (or lack thereof) when foot motion is altered. This information is critical in understanding the effects of shoe
design and orthotic interventions.
Ó 2003 Elsevier Ltd. All rights reserved.

Keywords: Pronation; Gait; Injuries; Ankle

1. Introduction proximal to the foot. These include the gastrocnem-

ius, soleus, tibialis posterior, tibialis anterior, peron-
Many overuse injuries in running are associated with eus muscles, and the flexor/extensor digitorum/hallucis
excessive pronation of the foot during stance (Hinter- muscles. Understanding the functional role of these
mann and Nigg, 1998; James et al., 1978), although the muscles is critical to understanding the etiology of many
term excessive has not been clinically defined (Nigg and running injuries.
Morlock, 1987). The etiology of leg injuries is uncertain The tibialis posterior is generally regarded as the pri-
and certainly multi-factorial but excessive pronation mary invertor of the foot based upon its moment arm
may impose stress on the extrinsic muscles of the foot about the subtalar joint (Perry, 1983). Hence, it is con-
that leads to injury. The extrinsic foot musculature in- sidered the muscle that primarily acts to control the
cludes all muscles that insert on the foot but originate amount of pronation (eversion) that occurs during the
stance phase of running. In support of this, McClay and
* Manal (1999) reported an inversion moment during
Corresponding author. Address: Department of Human Move-
ment Sciences, University of Wisconsin––Milwaukee, P.O. Box 413,
running stance that they postulated represented activity
Milwaukee, WI 53201, USA. in the tibialis posterior. Many of the other extrinsic foot
E-mail address: krisocon@uwm.edu (K.M. O’Connor). muscles, however, also exert an inversion moment about
0268-0033/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.clinbiomech.2003.09.001
72 K.M. O’Connor, J. Hamill / Clinical Biomechanics 19 (2004) 71–77

the subtalar joint. The gastrocnemius and soleus have a pronation (tibialis anterior, tibialis posterior, soleus,
small inversion moment arm, but these muscles also exert medial and lateral gastrocnemius) would be greatest in
large forces during running to provide the required the valgus shoes and least in the varus shoes.
plantar flexion moment (Perry, 1983). Reber et al. (1993)
reported that peak activity of the gastrocnemius, soleus,
and tibialis posterior occurred during early to mid-stance 2. Methods
which would support their role in resisting foot prona-
tion. It has also been demonstrated that the tibialis an- 2.1. Subjects and materials
terior is active during early stance, which is an invertor of
the foot. The peroneus muscles act as evertors of the Ten males, classified as rearfoot strikers, foot sizes 9–
foot, but they may be active during early stance in order 10, were recruited for this study. Their average height,
to increase overall joint stiffness. The toe flexors and mass and age were 1.72 (SD 0.07) m, 72.6 (SD 5.3) kg, 27
extensors also are capable of providing frontal plane (SD 5) years, respectively. All subjects were active
moments based on their lines of action, but based on recreationally and injury-free at the time of the study,
muscle size and moment arms they likely provide a small and none wore orthotics. All subjects signed an informed
contribution to the overall moment at this joint. consent in accordance with university regulations.
Perturbation studies can help to clarify the exact role Three pairs of shoes (size 9.5) were custom built for
these muscles play. A perturbation that either accentu- this study. All shoes utilized a midsole constructed of
ates or inhibits pronation may alter joint moments and ethylvinyl acetate (EVA) with a durometer of 45 (Shore
muscle activation patterns illustrating how these muscles A). The neutral shoes were constructed with a heel
act to control pronation of the foot. Mundermann et al. height of 2.5 cm. In order to attain an 8°-varus config-
(2003) reported that wearing a posted orthotic decreased uration, the medial aspect of the midsole at the heel was
the peak inversion moment during stance. While pro- 3-cm thick, and the lateral aspect was 2-cm thick. The
viding important information about the effects of orth- valgus shoes were built according to the same dimen-
oses on joint dynamics, the subject-specific nature of the sions but with the heights reversed. The midsoles ta-
orthotic intervention makes it difficult to determine a pered to a 1-cm thickness at approximately the
more general response to frontal plane perturbations metatarsals. Similar midsole designs have been used in
that could be useful for shoe midsole design. Also, these other studies (Milani et al., 1995; Perry and Lafortune,
results indicated that the kinetics of the foot could be 1993; van Woensel and Cavanagh, 1992). The upper
altered, but they do not establish what structures led to portion of the three shoes was modified such that there
the change in joint moments. was no heel counter. This design was employed in order
Systematic perturbations at the foot have been in- to directly track movement of the calcaneus. The shoe
duced by requiring subjects to run in shoes with varus- was fastened to the foot through the use of a cuff around
and valgus-wedged midsoles (Milani et al., 1995; Perry the ankle to which straps from the shoe were attached
and Lafortune, 1993; van Woensel and Cavanagh, (Fig. 1). In a pilot study, it was determined that subjects
1992). The varus-wedged shoes were designed with a were able to run comfortably in these shoes.
thicker midsole along the medial portion of the shoe,
and the valgus shoes were designed with a thicker mid- 2.2. Experimental setup
sole along the lateral portion of the shoe. These studies
reported systematic increases in pronation when running Kinematic, kinetic and EMG data were acquired
in the valgus shoe as compared to the varus shoe. These from the right lower limb of all subjects. Three-dimen-
authors did not, however, investigate joint kinetic pa-
rameters or muscle activation patterns. An under-
standing of how these parameters change with
perturbations to the foot is critical to understanding
possible mechanisms of injuries.
Therefore, the purpose of this study was to investi-
gate the role of selected extrinsic foot muscles during
running. This was accomplished by determining the ki-
nematic, kinetic, and EMG responses to running in
varus, neutral, and valgus-wedged shoes. It was hy-
pothesized that the varus shoe should decrease foot
pronation and the inversion moment while the valgus
shoe should increase pronation and the inversion mo-
ment. It was further hypothesized that the EMG acti- Fig. 1. Modified upper portion of the shoe designed to directly measure
vation levels of muscles involved in controlling calcaneus motion. Markers triad placement also demonstrated.
 K.M. O’Connor, J. Hamill / Clinical Biomechanics 19 (2004) 71–77 73

sional kinematic data were collected using a seven- Anatomical coordinate systems for the leg and foot were
camera Qualisys Pro-Reflex motion capture system derived from the marker locations collected in the
(East Windsor, CT, USA). Ground reaction force data standing calibration. The leg reference system was right-
were collected with a force platform (model BP6001200, handed and anatomically based, with the X -axis point-
A M T I , Inc., Watertown, MA, USA) mounted flush with ing laterally, the Y -axis point anterior, and the Z-axis
the floor. Ag–AgCl dual electrodes (2.0-cm inter-elec- pointing superiorly along the long axis of the segment.
trode distance) were placed at each recording site (model The foot coordinate system was aligned with the room
272, Noraxon, Inc., Scottsdale, AZ, USA). The in- coordinate system during the standing calibration.
dwelling electrode consisted of two 44-ga paired hook Three-dimensional angular data were calculated using
wires within a 25-ga cannula for insertion (Nicolet In- an XYZ Cardan rotation sequence (Cole et al., 1993).
strument Corporation, Madison, WI, USA). All EMG Kinematic joint parameters at the ankle were ex-
data were collected from a Telemyo-8 telemetered EMG tracted for the frontal plane. While the ankle joint
system (Noraxon, Inc., Scottsdale, AZ, USA). Speed complex contains two separate joints, the ankle and
was monitored by recording the time between two subtalar joints, these were treated as a single universal
photoelectric sensors placed at each end of the testing joint (Cole et al., 1993). The touchdown angle, maxi-
zone. Kinematic data were collected at 240 Hz while mum eversion angle, time to maximum eversion, range
ground reaction force and EMG data were collected of motion and peak eversion velocity were reported.
synchronously at 1920 Hz. Kinematic and force data were combined to calculate
joint kinetic data using an inverse dynamics Newton–
2.3. Protocol Euler procedure (Bresler and Frankel, 1950). Each seg-
ment was modeled as a frustra of a cone. Inertial
Four reflective markers on a rigid plate were attached parameters were derived from Dempster (1955). The
to the leg and a marker triad, also on a rigid plate, was ankle joint center was defined by the midpoint between
attached to the posterior aspect of the foot. Prior to the standing calibration markers placed on the medial
placement of the surface electrodes, the sites were and lateral malleoli. Moments and powers were calcu-
shaved, abraded, and cleaned with alcohol. Surface lated about the ankle joint and reported in the frontal
EMG data were collected from the tibialis anterior plane of the foot coordinate system. Maximum and
(TA), peroneus longus (Per), lateral gastrocnemius minimum joint moments and powers and the times to
(LG), medial gastrocnemius (MG), and the soleus (Sol). the moments were reported. Work was computed from
Electromyographic data for the tibialis posterior (TP) the time integral of the power time series. Negative and
were collected using a fine-wire electrode. The needle positive work were reported, as well as the total work
was inserted approximately one centimeter from the defined as the sum of the absolute values of the positive
medial edge of the tibia, one hand-width below the tibial and negative work (Eng and Winter, 1995).
tuberosity. Prior to the data collection, a standing cali- EMG data were high-pass filtered at 20 Hz and full-
bration trial was collected with the subject standing wave rectified prior to the calculation of variables. The
barefoot. For the standing calibration, additional re- high-pass filter was utilized in order to remove move-
flective markers were placed on the subject in order to ment artifact and the DC bias (Winter, 1990). Since full
define segment geometries and the segment coordinate stride data were unavailable due to data collection
systems. These markers were placed on the skin over the limitations, the period before foot contact equal in du-
medial and lateral femoral epicondyles, the medial and ration to the stance time was chosen as the beginning of
lateral malleoli and the heads of the first and fifth met- the time series, and toe off was chosen as the end.
atarsals and were subsequently removed prior to the Therefore, the data were reported from )100% to 100%
collection of the running trials. of stance with 0% representing foot contact. The inte-
Before collecting the overground running trials, sub- grated EMG and the mean of each period were calcu-
jects ran on a treadmill at 3.6 m/s for approximately 2 lated as well as the onset and offset of activity for each
min. Subjects then performed five acceptable running muscle. In order to determine the onset and offset of
trials in each shoe at 3.6 m/s ± 5% along a 30-m walkway muscle activity, the data were low-pass filtered at 24 Hz
across the force platform. Subjects were required to land and the magnitudes were normalized to the highest peak
on the force plate with their right foot while kinematic, from the five neutral condition trials of each subject.
kinetic and EMG data were recorded. The threshold for onset and offset was set at 10% of the
peak. Linear envelopes were also created by filtering the
2.4. Data analysis rectified signal with a 12 Hz low-pass filter for illustra-
tion purposes.
The three-dimensional coordinate data were filtered All data were time normalized to 100% of stance,
with a low-pass, fourth order, zero lag, Butterworth with the EMG data reported as described above. Ki-
filter with a 12-Hz cutoff frequency (Hamill et al., 1992). nematic, kinetic, and EMG parameters were extracted
74 K.M. O’Connor, J. Hamill / Clinical Biomechanics 19 (2004) 71–77

from each trial and mean curves were calculated for 15 30

each subject. 10 20

Angle (degrees)

Moment (Nm)
5 10
2.5. Statistical analysis
0 0

A one-way repeated measures A N O V A (P < 0:05) was -5 -10

performed on each kinematic, kinetic, and EMG pa- -10 -20

rameter to detect differences between shoe conditions. A (a) -15 (c) -30
Tukey’s post-hoc test was conducted where appropriate.
200 60

Angular Velocity (degrees/s)

150
100 40
3. Results 50

Power (Watts)
20
0
-50 0
There were significant differences in the frontal plane -100
-20
kinematic variables among conditions (Table 1 and Fig. -150
-200
2(a and b)). There were no differences in touchdown -250
-40

angles, but the varus shoe significantly decreased the -300 -60
range of motion. Maximum eversion was significantly 0 20 40 60 80 0 20 40 60 80

greater in the valgus shoe than the neutral and varus (b) Time (% stance) (d) Time (% stance)

shoes, but the times to maximum eversion were not Fig. 2. Mean (a) frontal plane angle, (b) angular velocity, (c) net joint
significantly different. moment, and (d) joint power during the stance phase of running for all
There was an inversion moment during the first subjects. The thin line represents the varus condition, the dashed line
portion of stance, followed by an eversion moment represents the neutral condition, and the thick line represents the
valgus condition Positive angle, velocity, and moment values represent
during the second portion of stance in all conditions
inversion. Positive power represents energy generation. The shaded
(Fig. 2c). There was a significantly greater inversion area represents ±1 SD (between subjects) from the neutral condition.
moment while running in the valgus shoe, and a signif- The large standard deviations are indicative of the differing individual
icantly smaller inversion moment while running in the movement patterns that were observed, although subjects responded
varus shoe as compared to the neutral condition (Table consistently to the three conditions.
1). The frontal plane power was characteristically tri-
phasic with energy absorption occurring during early
and late stance with energy generation only occurring
during a short period during the middle of stance (Fig.
Table 1 2d). There was significantly greater energy absorption
Group mean (SD) frontal plane kinematic and kinetic parameters while running in the valgus shoe as compared to the
Variable Condition other conditions.
Ensemble activation profiles were compiled for the
Varus Neutral Valgus
EMG of each muscle with the number of subjects shown
Touchdown angle (°) 4.1 (10.5) 5.0 (9.9) 4.2 (8.5)
Range of motion (°)
9.1 (2.9)a 11.1 (4.1)b 12.8 (5.5)b
for which acceptable data were collected (Fig. 3). Due to
Peak eversion angle (°)
)5.0 (8.2)a )6.1 (6.7)a )8.5 (7.0)b a variety of difficulties in collecting EMG data, data for
Time to peak eversion 40.2 (11.9) 40.0 (10.7) 36.4 (10.3) certain muscles were not obtained on given subjects. In
(% stance) particular, the tibialis posterior proved to be a difficult
Peak eversion velocity )253.8 )294.8 )335.4 signal to record from indwelling electrodes for an entire
(°/s)
(42.2)a (122.9)a;b (127.8)b
Peak inversion moment 12.5 (14.3)a 14.7 (14.5)a 21.0 (15.7)b
data collection session. The signal degraded in several
(N m)
subjects across the data collection period, leaving only
Time to inversion 43.0 (19.6) 37.4 (23.8) 33.2 (13.5) four subjects with full sets of data for this muscle. With
moment (% stance) only four subjects, statistical tests on this muscle were
Peak eversion moment )19.9 )16.3 )17.4 (14.4) not performed. There were no significant differences in
(N m) (15.8) (10.4)
Time to eversion 62.9 (12.2) 68.3 (15.6) 57.2 (13.0)
the integrated EMG, mean EMG, onset, or offset times
moment (% stance) between conditions for any of the other muscles re-
Minimum power (W)
)64.2 )63.9 )106.7 corded (Fig. 4).
(28.2)a (25.5)a (51.0)b
Maximum power (W) 67.6 (47.4) 54.3 (28.5) 66.6 (30.5)
Negative work (J) )2.8 (1.6) )2.6 (1.3) )4.1 (2.2)
Positive work (J) 2.4 (1.6) 1.9 (0.8) 2.4 (1.0) 4. Discussion
Total work (J) 5.2 (2.7) 4.5 (1.6) 6.5 (2.6)
*
Significant difference (P < 0:05). Like letters are not significantly The purpose of this study was to examine the role of
different. the extrinsic foot muscles during running by investigat-
 K.M. O’Connor, J. Hamill / Clinical Biomechanics 19 (2004) 71–77 75

(a) Tibialis Posterior (n=4) (d) Tibialis Anterior (n=8) 90

Integrated EMG (mv*s)

1 1
EMG Magnitude (a.u.)

60
Varus
Neutral
Valgus
30

0 0 0
(a) TP MG LG TA Per Sol
(b) Medial Gastrocnemius (n=9) (e) Peroneus Longus (n=8)
180
1 1
EMG Magnitude (a.u.)

150

Mean EMG (mv)

120

90

60

0 0 30

0
(c) Lateral Gastrocnemius (n=8) (f) Soleus (n=10) TP MG LG TA Per Sol

1 1
(b) M uscle
EMG Magnitude (a.u.)

Fig. 4. (a) Integrated and (b) mean EMG values for all subjects. The
mean and standard deviation across subjects are represented. There
were no significant differences between conditions.

0 0
-100 -50 0 50 -100 -50 0 50
directly on the calcaneus, the kinematic results of the
neutral shoe compare favorably to Reinschmidt et al.
Time (% stance) Time (% stance)
(1997) who reported three-dimensional ankle kinematics
Fig. 3. Group ensemble EMG activity profiles. The thin line represents derived from bone pins inserted into the tibia and cal-
the varus condition, the dashed line represents the neutral condition, caneus while running in shoes with a heel counter. The
and the thick line represents the valgus condition. The magnitudes are
joint displacement profiles in their study demonstrated
in arbitrary units. Time zero represents foot contact and )100% rep-
resents the instance in time before foot contact equal to stance time. similar patterns in all three planes and the joint excur-
The shaded area represents ±1 SD (between subjects) from the neutral sions were quite similar to the present study. The max-
condition. The number of subjects included in the ensemble average imum eversion angle of the bone pins was 8.6° while this
are shown in the parentheses. study reported 6.1°. In comparison, Reinschmidt et al.
(1997) also reported kinematics based on shoe-based
markers and reported a maximum eversion angle of 13°.
ing the kinematic, kinetic and muscle activation changes While not central to the purpose of this study, running
that occurred while running in varus, neutral and val- in the shoes without a heel counter appears to yield ki-
gus-wedged shoes. It was hypothesized that the varus nematic information quite similar to running in a shoe
shoe would decrease pronation during stance, reduce the with a heel counter.
net inversion joint moment and reduce activation levels With regard to the experimental manipulation, run-
in the invertor muscles (tibialis posterior, gastrocnemius, ning in shoes with a wedged midsole elicited the pre-
soleus). The valgus shoe was hypothesized to have the dicted kinematic response with the valgus-wedged shoe
opposite effect relative to the neutral condition. Antici- accentuating calcaneal eversion. Other studies (Milani
pated joint kinematic and kinetic changes were ob- et al., 1995; Perry and Lafortune, 1993; van Woensel
served, but there were no significant differences in and Cavanagh, 1992) that have employed these shoes
muscle activation profiles between shoes. have also reported significant differences in rearfoot
Given the design of the shoe upper chosen for this motion, although these studies based their results upon
study, the intent was not to directly compare kinematic shoe-based markers. van Woensel and Cavanagh (1992)
results of this study to traditional running shoe studies, and Perry and Lafortune (1993) both utilized shoes with
nor to infer the motion of the foot within a regular 10° wedges while Milani et al. (1995) utilized shoes with
running shoe. Rather, the purpose was to understand 8° wedges. In each case, the differences in rearfoot angle
the relationship between kinematic, kinetic and EMG between shoes were approximately equal to the inter-
variables when motion of the foot is perturbed in some vention. This could be predicted since they measured the
way during running. Although the markers were placed motion of the heel counter, which is rigidly attached to
76 K.M. O’Connor, J. Hamill / Clinical Biomechanics 19 (2004) 71–77

the midsole. The direct measurement of the foot in this formed. The tibialis posterior results, while only the
study, not using a heel counter, indicated that the 8°- product of four subjects, did not indicate trends that
wedge only altered the skeletal motion about 2°–3°. This might support the hypothesis that this muscle would
result seems consistent with other findings where sub- systematically alter its activation in response to the ex-
stantial shoe alterations were made that result in little perimental manipulation. While these results should
change in foot kinematics (Stacoff et al., 2001). Even so, certainly be viewed with caution, no clear differences
this magnitude of perturbation was sufficient to alter between shoe conditions emerged for this muscle. The
joint moment patterns. EMG data for all muscles generally did not indicate any
The frontal plane moments were highly variable systematic responses, which could suggest that acute
among subjects. McClay and Manal (1999), the only changes in foot frontal plane motion may not require an
study to report normative kinetic data for the secondary active response by the neuromuscular system.
planes of motion during running, also reported high The results of this study indicate that musculotend-
variability in the joint moment patterns in the frontal enous injuries may not be directly related to increased
plane. There was also high variability in the joint powers activity in muscles controlling pronation of the foot.
among subjects. Despite this variability between sub- This study cannot conclude, however, that perturbation
jects, the peak inversion moment and peak negative of the foot does not alter the forces in the muscles and
power and the peak abduction moment in these sec- impose excessive stress at the attachment sites. It has
ondary planes were significantly different among con- been shown that the passive properties of muscle, such
ditions. as the force–length and force–velocity characteristics,
The peak inversion moment was greatest for the modulate muscle force (van Soest and Bobbert, 1993)
valgus shoe which was designed to accentuate prona- and that the passive properties can regulate external
tion. There was 58% more energy absorbed in the forces (Herzog et al., 2000; Wright et al., 1998). At im-
frontal plane for the valgus shoe as compared to the pact, increased pronation and pronation velocity likely
neutral shoe (valgus ¼ )4.1 J; neutral ¼ )2.8 J). This increase the rate of stretch in the invertor muscles,
increased energy absorption occurred during early thereby increasing force in the muscle by increasing the
stance. While this amount of work is small compared eccentric velocity at a given activation level. Also, the
to the sagittal plane ()30 J), it may be sufficient to hard and soft tissue constraints within the joint may
cause injury if the additional load is primarily experi- contribute to a resistive moment if the subtalar joint
enced in a single structure such as the tibialis posterior approaches the end of its range of motion (Chen et al.,
muscle. Noyes (1977), for example, reported that the 1988) while wearing the valgus shoe. Therefore, while
anterior cruciate ligament of rhesus monkeys reached this study indicates that perturbation of the foot may
failure when the ligament absorbed 3.5 J of energy. not elicit adaptations in activation patterns, it is possible
While human tendons are likely thicker than the that passive properties may lead to greater tissue loads
monkey ACL, the physical properties are similar. about the ankle. Musculoskeletal modeling may lend
Therefore, an additional 1.6 J could be sufficient to insight into the load sharing and the role of passive
damage the tissue when repetitively stressed. While the properties of these muscles and ligaments.
mechanism of tendon-related injuries is unclear, ex- A possible limitation in this study is whether an ap-
perimental evidence indicates that increased energy propriate amount of time was given for subjects to adapt
absorption may contribute to injuries (Fisher, 2000). to each experimental condition. Little data exists on the
Therefore, increased energy absorption by the muscles exact time course of neuromuscular adaptations al-
caused by hyper-pronation may indeed be a mechanism though there is evidence that the nervous system can
for injury. quickly adapt to changes in the environment (Belanger
The muscle activity profiles in this study were not and Patla, 1984; Duysens et al., 1992; Ferris et al., 1999;
reported for a full stride. However, the stance and pre- White et al., 2002). While it is possible that activation
stance phases are similar to those in the literature for patterns may alter over time in response to the varus and
normal running (Reber et al., 1993). All of the muscles valgus shoes, the literature supports the assumption that
other than the tibialis anterior were generally quiet gait adaptations take place relatively quickly.
throughout swing. The pre-impact phase for the tibialis Another possible limitation may be the shoe design
anterior does appear to capture the prominent burst of utilized in this study. Although subjects reported being
activity in preparation for landing. Given this, the EMG able to run normally in these shoes and the kinematics
results as represented in this study likely still contain the were similar to Reinschmidt et al. (1997), there may
critical information about their activity during running have been changes to muscle activity attributed to the
for most muscles. heel-less shoe. Based on the subjective responses and
There were no significant differences in the integrated kinematics, the differences in muscular activity as com-
and mean EMG values or in the onset and offset times of pared to running in a typical shoe were assumed to be
the five muscles for which statistical tests were per- minimal.
 K.M. O’Connor, J. Hamill / Clinical Biomechanics 19 (2004) 71–77 77

5. Summary Ferris, D.P., Liang, K., Farley, C.T., 1999. Runners adjust leg stiffness
for their first step on a new running surface. J. Biomech. 32, 787–
794.
Running in the varus and valgus shoes altered the Fisher, K.J., 2000. Biological response to forces acting in the
joint kinematics and kinetics during running. Greater locomotor system. In: Nigg, B.M., MacIntosh, B.R., Mester, J.
negative work was performed in the frontal plane while (Eds.), Biomechanics and Biology of Movement. Human Kinetics,
running in the valgus shoes indicating that greater en- Champaign, IL, pp. 307–329.
ergy was absorbed in the structures that would con- Hamill, J., Bates, B.T., Holt, K.G., 1992. Timing of lower extremity
joint actions during treadmill running. Med. Sci. Sports Exerc. 24,
tribute to an inversion moment. The tibialis posterior, 807–813.
soleus, and gastrocnemius are likely to primarily absorb Herzog, W., Koh, T.J., Hasler, E., Leonard, T., 2000. Specificity and
this energy. There was not, however, a significant plasticity of mammalian skeletal muscle. J. Appl. Biomech. 16, 98–
change in the muscle activation levels of any of the 109.
muscles recorded. These results suggest that passive Hintermann, B., Nigg, B.M., 1998. Pronation in runners. Implications
for injuries. Sports Med. 26, 169–176.
properties may primarily account for the increased en- James, S.L., Bates, B.T., Osternig, L.R., 1978. Injuries to runners. Am.
ergy absorption when there is greater eversion of the J. Sports Med. 6, 40–50.
foot. This study was not, however, able to identify which McClay, I., Manal, K., 1999. Three-dimensional kinetic analysis of
muscle(s) accounted for the loading changes and there- running: significance of secondary planes of motion. Med. Sci.
fore would be more likely to be injured based upon Sports Exerc. 31, 1629–1637.
Milani, T.L., Schnabel, G., Hennig, E., 1995. Rearfoot motion and
EMG activity. pressure distribution patterns during running in shoes with varus
and valgus wedges. J. Appl. Biomech. 11, 177–187.
Mundermann, A., Nigg, B.M., Humble, R.N., Stefanyshyn, D., 2003.
Foot orthotics affect lower extremity kinematics and kinetics
Acknowledgements during running. Clin. Biomech. 18, 254–262.
Nigg, B.M., Morlock, M., 1987. The influence of lateral heel flare of
The authors would like to acknowledge New Balance, running shoes on pronation and impact forces. Med. Sci. Sports
Exerc. 19, 294–302.
Inc for their contribution of the shoes used in this study, Noyes, F.R., 1977. Functional properties of knee ligaments. Clin.
and to the International Society of Biomechanics Orthop. Rel. Res. 123, 210–242.
(Matching Dissertation Grant) and Life Fitness Acad- Perry, J.E., 1983. Anatomy and biomechanics of the hindfoot. Clin.
emy (Michael L. Pollock Memorial Grant) for their fi- Orthop. Rel. Res. 177, 9–15.
nancial support of this project. Perry, S.D., Lafortune, M.A., 1993. Effect of foot pronation on impact
loading. In: Bouisset, S., Metral, S., Monod, H. (Eds.), Bio-
mechanics XIV, II. International Society of Biomechanics, Paris,
pp. 1024–1025.
Reber, L., Perry, J., Pink, M., 1993. Muscular control of the ankle in
References running. Am. J. Sports Med. 21, 805–810.
Reinschmidt, C., van den Bogert, A.J., Murphy, N., Lundberg, A.,
Belanger, M., Patla, A.E., 1984. Corrective responses to perturbation Nigg, B.M., 1997. Tibiocalcaneal motion during running, measured
applied during walking in humans. Neurosci. Lett. 49, 291–295. with external and bone markers. Clin. Biomech. 12, 8–16.
Bresler, B., Frankel, J.P., 1950. The forces and moments in the leg Stacoff, A., Reinschmidt, C., Nigg, B.M., van den Bogert, A.J.,
during level walking. Trans. ASME 72, 27–36. Lundberg, A., Denoth, J., Stussi, E., 2001. Effects of shoe sole
Chen, J., Siegler, S., Schneck, C.D., 1988. The three-dimensional construction on skeletal motion during running. Med. Sci. Sports
kinematics and flexibility characteristics of the human ankle and Exerc. 33, 311–319.
subtalar joint––Part II: Flexibility characteristics. J. Biomech. Eng. van Soest, A.J., Bobbert, M.F., 1993. The contribution of muscle
110, 374–385. properties in the control of explosive movements. Biol. Cybern. 69,
Cole, G.K., Nigg, B.M., Ronsky, J.L., Yeadon, M.R., 1993. Appli- 195–204.
cation of the joint coordinate system to three-dimensional joint van Woensel, W.W., Cavanagh, P.R., 1992. A perturbation study of
attitude and movement representation: a standardization proposal. lower extremity motion during running. Int. J. Sport Biomech. 8,
J. Biomech. Eng. 115, 344–349. 30–47.
Dempster, W.T., 1955. Space requirements of the seated operator. White, S.C., Gilchrist, L.A., Christina, K.A., 2002. Within-day
Wright-Patterson Airforce Base, Ohio, pp. 55–159. accomodation effects on vertical reaction forces for treadmill
Duysens, J., Tax, A.A., Trippel, M., Dietz, V., 1992. Phase-dependent running. J. Appl. Biomech. 18, 74–82.
reversal of reflexly induced movements during human gait. Exp. Winter, D.A., 1990. Biomechanics and Motor Control of Human
Brain Res. 90, 404–414. Movement, second ed. John Wiley & Sons, Inc., New York.
Eng, J.J., Winter, D.A., 1995. Kinetic analysis of the lower limbs Wright, I.C., Neptune, R.R., van den Bogert, A.J., Nigg, B.M., 1998.
during walking: what information can be gained from a three- Passive regulation of impact forces in heel-toe running. Clin.
dimensional model? J. Biomech. 28, 753–758. Biomech. 13, 521–531.