Research Report

Local Sensation Changes and Altered Hip Muscle

Function Following Severe Ankle Sprain

Background and Purpose. Changes in sensory information have been shown Joanne E
to influence muscle function locally. Some clinicians, however, believe that the Bullock=Saxton
influem?may be m r e wemennve. To investigate this clinical concept, subjects with
severe ankle sprain were assesed for local sensation changes and p m ' m a l hipl
back muscle function. Subjects. Of a total of 361 potential subjects whose medi-
cal histolies were assessed, 20 men (age 18-35years) who had previously sus-
tained a severe unilateral ankle sprain and 11 matched "control" subjects with
no preuiaus lower-limb injury participated in the study. Metboth. Using this ex-
perimental model, tests of vibration sensation in the ankle (indicating sensation
changes) as well as surface electromyograpby of muscle recruitment patterns for
hip extension (indicating muscle function proximally) of the bicepsfemoris, glu-
teus m - i m u s ,and lumbar erector spinae muscles were made on both sides of the
unilaterally injured and matched control subjects. Results. Signflcant decreases
in vibrafionperception and signflcant delays in gluteus marimus muscle recruit-
ment during hip extension were found in the injured group. Conclusion and
Discusdon. The author concludes that both local sensory a n d p m ' m a l muscle
function changes are amciated with unilateral severe ankle sprain. [Bullock-
Saxton JE, Local sensation changes and altered hip muscle function following
severe ankle sprain. Pbys Ther. 1994;74:17-31.1

Key Words: Ankle; Electromyograpby; Hip; Muscle pe@ormance, lower extremity;

Sensation; Sprains and strains.

The existence of a complicated feed- input from the ankle joint and its gated the response to various articular
back system between muscles and influence on the postural control of stimuli of decerebrate and, in some
joints anti the central nervous system the muscles. instances, spinalized cats. They found
is well recognized. Interference with that increasing the articular pressure
sensory feedback may affect a per- This relationship between joint recep- in the knee joint, as well as pinching
son's ability to monitor movements o r tor information and muscle function its capsule, led to decreased quadri-
to make appropriate adaptations and has interested researchers for some ceps femoris muscle (ie, extensor)
adjustments to movement. For exam- years, and the relationships between activity, whereas pinching the knee
ple, a change in postural stability stimulation of joint afferents and mus- capsule elicited an increased response
when a person stands on one leg cle activity have been demonstrated in from the knee flexors (biceps femoris
following ankle sprain was attributed both animal and human studies.2-5 muscles). In their study of human
by Freerr1an1 to altered proprioceptive For example, Ekholm et alz investi- subjects, Stokes and Young3 consid-
ered that joint injury can decrease the
activity of muscles, leading to weak-
ness and wasting. They measured the
JE Bullock-Saxton, PhD, BPhty(Hons), is Lecturer, Department of Physiotherapy, The University of
Queenslanti, St Lucia, 4072, Queensland, Australia. rectified integrated electromyographic
(EMG) activity of both quadriceps
This study was approved by The University of Queensland Ethics Committee and the Australian
Armed Foxes Ethics Committee.
femoris muscles of patients who had
a meniscectomy or an arthrotomy and
Thls article was submitted August 12, 1992, and was accepted July 30, 1993.

Physical Therapy /Volume 74, Number ldanuary 1994

recorded large decreases (80%) in influence the muscle function about the motor regulation of body posture.
quadriceps femoris muscle activity on that joint. However, a more complex In their experimental study of pos-
the side of surgety. This effect per- relationship than this has also been tural stability following ankle sprain,
sisted for up to 15 days postopera- proposed. This proposed relationship Tropp et al" found a significant de-
tively (30%-40%), despite the lack of is that altered sensation in one joint crease in postural stability when com-
pain at that time. could lead to muscle function pared with noninjured subjects, thus
changes in another more proximal confirming Wykes'9 observations.
A possible mechanism for this de- joint. This concept has been the basis Recent experiments by Gauffin and
creased activity might be the excita- of teaching by Lewit7 and Jandas for colleagues12 have indicated that pa-
tion of joint afferents in the capsule several decades. Some experimental tients with unilateral anterior cruciate
because of pressure caused by joint data on cats d o exist demonstrating ligament injury demonstrated bilateral
infusion. Indeed, in 1965, De Andrade that the motor system has a tendency alterations in their postural control
et a14 showed that in healthy human to extend dysfunction into a larger when compared with uninjured sub-
subjects and in those with pathology, area.9 jects. They postulated that these alter-
infusion of saline into the knee joint ations may be due to "central adjust-
was responsible for decreased activity Although m a l t to extrapolate re- ments of motor control." The
of the quadriceps femoris muscles. sults from animal studies to human postulation that changes in sensory
Results of recent studies by Iles et al5 behavior, Wyke9 observed that in the input could cause alterations in the
have indicated that as the volume of cat, an injury of the joint capsule or function of muscles in a joint remote
saline infused into the human knee ligaments influences muscle activity from the injury seems to be well
joint is increased, the amplitude of not only in muscles that cross the justified, although no direct experi-
the H-reflex is decreased and that injured joint, but also in remote mus- mental evidence of this in humans
even apparently imperceptible vol- cles. Wyke stated that has yet been reported.
umes of saline could decrease quadri- . . . interruption of the flow of impulses
ceps femoris muscle activity. In these from the mechano-receptors in a joint This study was conducted, therefore,
studies, the decrease of extensor capsule into the central nervous system to investigate whether a localized
activity following afferent stimulation should result in clinically evident dis- lesion at a peripheral joint such as the
was highlighted. turbances of perception of joint posi- ankle influenced the sensation in that
tion and movement and of the reflexes area as well as the muscle function in
The relationship between ankle artic- concerned with posture and gak9W) more proximal regions such as the
ular mechanoreceptor function and hip and pelvis and, if so, whether
the reflex activity in the limb of the Thus, the arthrokinetic reflex might such changes were interrelated. The
cat was also investigated by Freeman be considered as a triggering factor sensory and muscle function in both
and Wyke.6 Establishing the normal that would initiate a whole chain of limbs of subjects who had previously
reflex muscular response of the tibia- adaptation reactions, eventually result- sustained a severe unilateral ankle
lis anterior and gastrocnemius mus- ing in a changed movement pattern. sprain was compared with that in
cles, these researchers decreased the The possibility that sensory deficits both limbs of noninjured ("control")
afferent information from the joint by associated with localized injury in one subjects.
local anesthesia and by electrocoagu- part of the body influence muscle
lation of the articular tissues. Both function in another and may ulti- For this study, appropriate tests of
procedures caused an abolition of the mately lead to pain has considerable sensory and muscle function needed
normal reflex muscular response to implications for the physical therapist, to be selected. Freeman et all3 theo-
movement, indicating the importance influencing both the preventive and rized that ankle instability following
of articular information to regulation therapeutic approaches to patient injury develops primarily due to le-
of muscle activity. Freeman and Wyke6 care. sions of mechanoreceptors in the
believe that muscle activity is regu- joint capsule and ligaments. This insta-
lated through the contribution of the Wyke9 also argued that articular sen- bility impairs both the static position
articular impulses to a facilitatory bias sory information is vital to normal and joint movement sense. This the-
to the gamma motoneurons of the postural reflexes. He cited observa- ory was not supported by Gross,l4
muscle spindles. If such an influence tions of impaired postural reflex activ- who compared active and passive
exists, then their early assertion that ity of muscles following severe ankle joint position sense in both injured
articular afferents influence local mus- sprain of humans and proposed that and uninjured subjects and found no
cle activity is correct and has been this might be a reflection of the im- significant difference between them.
supported by the later research of Iles paired proprioceptive information However, the method of testing used
et a1.5 from the damaged mechanoreceptors. for assessing joint position sense
As afferent impulses travel to cerebel- involved strapping the foot to a mov-
This experimental evidence supports lar and cortical centers,1° the im- able footplate with firm pressure. It is
the clinical observation that a joint paired afferent information from the possible that mechanoreceptors on
injury involving sensory receptors can ankle joint may be sufficient to impair both plantar and dorsal surfaces of

Physical Therapy/Volume 74, Number 1,January 1994

the foot, which were not compro- provide information on the integrity onstrated in the uninjured control
mised by the lateral ligament ankle of sensory receptors possibly dam- group.
injury, were able to provide sufficient aged due to ankle ligamentous injury.
cues for the subject to determine This assessment of sensation is capa- Method and Materials
ankle joint position. Barrack and col- ble of a high degree of control in
leagues's appear to have developed a comparison with current tests used Subjects
successful measurement procedure for assessing tactile, two-point dis-
for eliminating pressure cues during crimination or joint movement sense Two groups of subjects were studied:
testing of joint position sense of the around the ankle. For these reasons, an "injured" group, who had previ-
knee following anterior cruciate liga- vibration perception was chosen as ously sustained a severe unilateral
ment injury. These researchers found the variable for assessment of sensory ankle sprain, and a matched "control"
significant deficits in joint position function in this study. group, who had no previous lower-
sense of' the knee. limb injury. To control variables be-
Muscle function in each limb was tween the two groups of subjects, a
If damage to sensory receptors from investigated in terms of the temporal suitable population of sufficient size
severe ankle joint sprain is to be sequence of activation (as illustrated was sought. The armed forces pro-
accurately measured, a test that is in EMG signals) of the gluteus maxi- vided such a source. The Australian
sensitive to changes in sensory recep- mus, hamstring, and ipsilateral and Defence Force (Army) gave permis-
tor function is needed. Two factors contralateral lumbar erector spinae sion for their soldiers and officers to
were considered in this regard: (1) muscles during the movement of hip volunteer to be subjects for this study.
the influence of joint stress on dis- extension from a prone-lying position.
charge rates of mechanoreceptors and JandaI7has claimed that the determi- For the injured group, subjects were
(2) the effects of age on sensation. nation of the order of activation of included if they had previously sus-
Wyke? in his description of three muscles performing a simple move- tained a grade 11+ or 111 (severe or
types of articular nerves, outlined ment is important for the understand- unstablela) lateral ankle sprain that
how the frequently occurring group I1 ing of the methods used by patients was significant enough to have caused
nerve fibers terminated onto both to move their body and that this marked swelling at the time of injury
low-threshold,slowly adapting mecha- knowledge helps to reveal the area of and discomfort while walking. Treat-
noreceptors (type I) and low- impairment. Hip extension was se- ment must have included a period of
threshold, rapidly adapting mechan- lected for this study, not only because immobilization.The subjects' right
oreceptors (type II). The type I the studied muscles were separated side must also have been the pre-
mechanoreceptors, found in clusters from the site of injury but also be- ferred (or "skill") side. Subjects were
around the joint capsule, where the cause of its functional importance in excluded if they had had a significant
greatest degrees of stress during stance and locomotion. Due to the injury to any other lower-limb joint or
moveme:nt are likely to occur, are complexities of the gait process, it a significant injury to either leg. Of
sensitive to changes in joint pressure was considered advisable to isolate particular concern was the need to
and position. Their rate of discharge the hip extension motion rather than exclude subjects who may have had a
adapts ~ipidlyto the degree of joint to study muscle function during gait. history of incoordination or clumsi-
stress. It is likely that capsular tears, Much greater control could be im- ness (operationally defined as a his-
rupture of small nerve fibers, and posed experimentally by assessing tory of sensory and/or motor dysfunc-
joint edema following ankle sprain muscle activation during hip exten- tion not related to injury, in the
could cause alterations of discharge sion from a prone-lying position than absence of intellectual impairment). It
from these receptors, as indicated by would be possible during locomotion. was essential to ensure that an exist-
Freeman et ale13 ing neurological deficit was not a
In my study, the effects of ankle predisposing cause of the ankle
In persons without joint injuries, sprain (the independent variable) on sprain, because the finding of differ-
perception of some superficial and vibration perception at the ankle and ences in localized sensory function in
deep sensations decreases with age. the pattern of activation of specified the control group could then be said
These sensations include tactile, two- muscles around the hip and low back to be a cause, rather than an afteref-
point discrimination;vibration per- (the dependant variables) were mea- fect, of the injury.
ception; and joint movement sense.l3 sured. A matched control group was
Such an age-related decline suggests used for comparison. Only subjects I assessed the medical histories of 361
that these sensory modalities are who had sustained a unilateral ankle potential subjects; 80 subjects (22%)
vulnerable to change. Vibration per- sprain were included in the injured were found to have sustained an
ception requires information from group, so that side-to-side differences ankle sprain on both sides, and 233
both superficial and deep mechanore- between their injured and uninjured subjects (65%) either had injuries in
ceptors ;B well as a functional cortical sides could be compared with the other joints or were unavailable to
sensory association area.16Testing normal side-to-side differences dem- participate in the study. Sixty-four
vibration perception, therefore, would men (18%) underwent the detailed

Physical Therapy/Volume 74, Number

-
Table 1. Distribution of 4ge, Height, and Weight VariablesBetween Injured
(n=20)and Control (n=I I) Groups (Percentages)

Group
Age (Y)
18-20 21-25 25-30 31-35
the criteria were found. Table 1 illus-
trates the distribution of relevant
variables between the two groups.

Measurements

VlbmNon perception. Dyck et all9

have discussed the problems associ-
ated with the measurement of the
Injured 10 threshold of vibration perception and
Control 9 the inadequacy of current clinical
methods (such as the use of tuning
Helght (cm) forks) in providing reliable, repeat-
170-174 175-1 79 180-1 84 185-1 89 1@0+ able results. For my measurements of
vibration sensitivity, it was desirable
to ensure that frequency and ampli-
Injured 20 15 45 15 5 tude of vibration could be varied, a
Control 0 9 64 18 9 consistent pressure of application
could be maintained, and the subject
Welght (kg) could remain alert and cooperative.
70-74 75-79 80-64 85-89 90+
To meet the first of these criteria, a
mechanical oscillator* connected to a
Power oscillatort was used. This in-
Injured 5 15 20 30 30 strument allowed variation of both
Control 9 27 27 18 18 frequency and voltage and provided
measures of output voltage directly
Physlcal Actlvlty Durlng Occupatlon related to acceleration (force inten-
sity) of the oscillator head at each of
Sedentary Actlve
the chosen frequencies. To ensure a
Llght Fleld Heavy Fleld constant pressure of this device on
Drlver Admlnlstratlon Work* Workb the skin, the oscillator was suspended
from one end of a system of pulleys
Injured 15 30 and a mass of equal weight was sus-
Control 18 18 pended from the other end as a
counterbalance.
Physlcal Actlvlty Durlng Sports
The subject was positioned side lying
Dally Physlcal Welght with the leg to be tested uppermost
Therapy Football Runnlng Llttlng Cycling Swlmmlng and secured in a lower-leg rigid sup-
port (back slab) to control the degree
lnjured 90 of ankle dorsiflexion, and supported
Control 100 on a highdensity foam cushion (Fig.
1). The oscillator was positioned per-
"Includes signal operators, medical units, survey units. pendicular to and just touching a point
on the inferior fibular head, and the
b~ncludesinfantry, bombardiers.
weight on the counterbalance was
reduced by 50 g. The head of the
screening process, but only 20 (6%) weight, and lwel of physical activity
oscillator, therefore, made contact with
of these subjects met all inclusion (eg, during sports and work). Subjects
the fibular head with a gravity-applied
criteria. for the control group, who matched
force of 50 g. The voltmeter, which
subjects in the injured group in these
reflected the amplitude of oscillation
Variables measured were age, physical characteristics, were sought from the
of the vibrator, indicated both when
characteristics such as height and same army units. Eleven men fitting
vibration was occurring and when it
had ceased.

'Derriton p 2 ) Mechanical Oscillator, Derriton Electronics Ltd, Sedlescombe Rd N, St Leonards- Because there appears to have been
on-Sea, Sussex, United Kingdom. little research into the perception of
+Goodman'sPower Oscillator 09, Goodman Industries Ltd, Vibration Division, Axion Works, vibration at different frequencies, a
Wembley, Middlesex, United Kingdom. range of frequencies (ie, 100, 150,

26 120 Physical TherapyIVolume 74, Number loanuary 1994

 confidence interval limits of the
means (mean kt x standard error) and
their standard deviations of the
"within-subject-between-replication"
variation (derived from a suitable
analysis of variance) for each fre-
quency were comparable for each
group and are listed in Table 2. These
results indicated that for uninjured
subjects and for those with previous
ankle injury, the threshold of vibra-
tion perception at each frequency was
repeatable on the one day.

Muscle actlvatlon. Surface EMG

was used to provide information
regarding the activation of the speci-
fied muscles during hip extension,
utilizing bipolar surface electrodes
(silver-silver chloride) on each of the
ipsilateral and contralateral lumbar
erector spinae, gluteus maximus, and
hamstring muscles of both limbs.
Flgu~e1. Placement of the mechanovibrator over the lateral malleolus of the sup- With subjects positioned prone lying,
ported ankle. EMG sites were identified, and elec-
trodes were placed 10 mm (0.39 in)
200, ancl 250 Hz) was assessed. For sides was randomized. One re- apart on each prepared site. The elec-
each frequency, the oscillator ampli- searcher (JEB-S) tested all subjects. trodes were positioned parallel to the
tude (which could be termed "vibra- line of the muscle bellies and over
tion strength") was slowly increased To determine the consistency of sub- the area of greatest muscle bulk, de-
until the subject stated that he per- ject threshold to vibration perception, termined after a resisted contraction
ceived v.ibration. This voltage was a repeatability test was carried out of the specific muscle. Lumbar erector
recorded as the threshold. Two mea- prior to the major study. Ten repeti- spinae muscles were monitored adja-
suremerits at each frequency were tions at each frequency on one limb cent to the intervertebral level of L2-3,
taken on one limb to provide an were chosen randomly from a subset the upper fibers of the gluteus maxi-
mus muscle were monitored, and

-
estimate of measurement error, be- of the sample composed of five in-
fore commencing the series on the jured group subjects and five control electrodes were placed on the ham-
second limb. The order of presenta- group subjects. The repetitions of string muscles over the biceps femo-
tion of the series of frequencies to each frequency were taken within a ris muscle. For each hip extension
each subject and between subject 1-hour time span with a 30-second motion, signals from the four muscle
interval between repetitions. The groups were preamplfied using a

Table 2. Conjidence Interval and Standard Deviation for Repeatabilily Test of Vibration Perception (in Meters per Square
Second) r'n=20)

Frequency (Hz)

x9,.05=201.
'C =control group (n =5)
'I=injured group (n=5).

Physical Therapy /Volume 74, Number ldanuary 1994

Medelec PA65 preamplifiers before relation to the target line. By connect- subjects of the relative behavior of
passing to a Medelec AAM63 amplifi- ing the inclinometer to the chart muscles contributing to a group
er/filter.* The signals were sampled at recorder and computer, the position activity.
a rate of 2,500 Hz, were bandpass of the limb was recorded at rest and
filtered at a lower frequency of 0.8 Hz during movement. Because one o r more muscles might
and a higher frequency of 800 Hz, contract prior to the commencement
and were recorded on an eight- Speed of motion was controlled by of hip extension and the starting
channel ink jet chart recorders for the subject moving the limb through point of individual muscle contrac-
monitoring of the signal during data the 15degree range of motion at a tions relative to hip motion might
collection. The EMG signals were also rate equal to three beats of a metro- vary, a consistent reference point for
passed to an analogue-to-digital con- nome set at 72 beats per minute. That comparison purposes was needed.
verter in a computer and stored for is, the limb moved through a Therefore, the commencement of hip
analysis. 6degree arc of motion per second, movement (H) was taken as a refer-
which was considered to be approxi- ence. The temporal measure used to
The starting position of each leg was mately equal to a slow walking speed. recognize this was the time span (in
traced onto a sheet of paper placed seconds) between onset of individual
over the base of the test bed to en- Subjects were encouraged to relax muscle activity (0)and commence-
sure a consistency of position. A feed- prior to the hip extension, the chart ment of hip movement (H), as deter-
back system was devised to assist the recording indicating whether the mined by the inclinometer (ie, 0-H).
subject in controlling his own range muscles were at rest. Only then was
of motion. An inclinometed' provided the trial commenced, with EMG sig- Calculation of the time span between
a recording of the motion of the limb nals being recorded for a count of points of onset of the first and fourth
during hip extension. The inclinome- three beats of the metronome prior muscles provided a second quantita-
ter was connected to an oscilloscope to the request to extend the hip. This tive measure in relation to muscle
positioned below a face hole in the initial "at rest" recording not only activation, allowing a determination of
test bed to be monitored by the sub- provided a base line signal prior to whether injury influenced the time
ject. The inclinometer produced an hip extension, but also allowed the taken for activation of all four mus-
output in the form of a moving line EMG recording of any activity within cles. The second temporal measure
on the oscilloscope, and an initial the muscle as the subject prepared to used, therefore, was the time span (in
zero "base" line (representing the move into hip extension. An initial seconds) for the sequence of activa-
limb in neutral) and a "target" line training period ensured that the sub- tion of the first (MI) and fourth (MJ
(representing the designated 15" ject understood what was required of muscles ( 0 , MI-MJ. The incorpo-
range of hip extension) were marked. him. For each subject, a 10-second ration of a time reference into the
Thus, the subject had feedback for the recording of EMG and inclinometer sampling procedure and the com-
position of his limb as he moved the signals was made during each of the puter acquisition of EMG and limb-
limb through the range of motion. six tests on each side. A 10-minute position data allowed for a determina-
interval separated the tests on the two tion of these temporal measures.
The inclinometer, fixed to a curved sides to allow recovery from any
metal plate, was strapped to the lat- possible fatigue. The same researcher The EMG signals collected from the
eral side of the thigh on a line be- carried out all testing. four monitored muscles during hip
tween the greater trochanter and the extension were submitted to com-
lateral femoral condyle with the fe- For analysis of EMG data, the order of puter analysis to determine these
mur in a horizontal position, so lead- muscle activation represents the se- measures for each of the two limbs
ing to zero output of the inclinometer quence of each muscle's entry into a during the six trials for each subject.
(giving the base line). Passive limb coordinated muscle activity. Visual A special-purpose computer program
movement to the edge of a 15-degree observation is the method usually (language C) was written, in which
template allowed the l5degree target used by researchers for this deterrni- the 2,500 samples of data per second
line trace to be recorded on the oscil- nation. For this study, however, it was for each muscle for the 10-second
loscope. When the subject moved his important for statistical purposes to recording period could be analyzed.
limb, a third (moving) line provided find a quantitative measure that would Data used for this program related to
feedback of the limb's position in allow comparison between groups of the raw EMG signals, the EMG gain
used to acquire data, the period for
which the data were recorded, the
number of channels used, and the
* ~ e d e l e Ltd,
c Old Working Rd, Surrey, United Kingdom. data rate. The number of data points
#SimensAG Minograph Chart Recorder, ZW22, Postgach 101212, D-8000, Muchen 1, Federal Repub- was calculated and then read in bi-
lic of Germany. nary format. The data were stored
Il~chaevitz(A411-0001) Accelerometer, Applied Measurement, Baltec Systems, 26 Mayneview St, after multiplication by 100 to enable
Milton, 4064, Brisbane, Queensland, Australia. the program to use integer arithmetic

28 / 22 Physical Therapy /Volume 74, Number 1Januar-y 1994

Table 3. ConfidenceIntervals and Standard Deviations for Repeatabilig Test of Muscle Activation Relative to H@ Extensim (in
Seconds) (n =lo)
- --

Control Group (n=5)

Muscle Group
- -

Left Lumbar Erector Splnae Rlght Lumbar Erector Splnae Glulws Maximus Hamatring
SWlrSlc Left Right Left Right Left RbM La Rlght

injured Group (n=5)

Muscle ~ r o u p
Left Lumbar Erector Spinae Rlght Lumbar Erector Splnae Gluteus Maxlmur Hamstring
SWllc Uninjured Injured Unlnjured Injured Unlnjured Injured Uninjured Inlured

where possible in the analysis. The was activated in each case. From this, 24 values were rejected for this rea-
mean of' the first 500 points was cal- the 0, MI-M, calculation was made. It son, as reflected in the n values pre-
culated and subtracted from the raw also provided the incidence of each sented in Tables 3 and 4.
data to enable the data to be centered muscle occurring in each ranked
on zero. The data were then rectified position. To determine the repeatability of the
about this mean value and smoothed time of onset relative to hip motion, a
(four passes, 100-point bandwidth) to As a preliminary investigation of onset study was carried out using the analy-
remove the high-frequency compo- times, prior to adopting the use of the sis of six movement repetitions for
nents of the bursts yet still leave the computer analysis, an independent each subject. An analysis of variance
main burst shape. The filtering pro- "blind" visual observation of the EMG was applied to determine the mean
cess was carried out by using a filter signals on an IBM-compatible per- confidence interval and its standard
subroutine that used a rectangular sonal computer screen for a random deviation and the "within-
filter (the data were linearly averaged selection of subjects, trials, and mus- subject-between-replication" variabil-
over the bandwidth of the filter) and cles was performed to compare the ity for each of the four muscles. Re-
that allowed multiple passes over the accuracy of the computer analysis sults demonstrated an acceptable level
data. with human inspection. For the glu- of repeatability of the measurement
teus maximus and hamstring muscles, (Tab 3).
The me;m of the first 500 points was the computer analysis was found to
again calculated and subtracted from give comparable results to those ob- Data Anelysls
the data to ensure that the mean tained through manual inspection of
value of the initial region was zero. the computer-displayed signal, suit- Data acquired for muscle activation
The location of the peak data point ably magnified. Onset data for the during hip extension and for vibra-
was identified, and from this data lumbar erector spinae muscles were tion perception for both limbs were
point, the times at which the signal scrutinized visually to determine analyzed to investigate any differences
reached specified percentages of those trials in which the proximity of between the injured and control
maximum could be determined. For the heart beat signal to the signal of groups in muscle or sensory function
the purposes of this analysis, 5% of muscle activity made computer dis- in each limb. The general linear
the maximum weak value) EMG crimination of the onset of muscle model (GLM) of analysis of variance
signal was regarded as the onset of activity impossible. In such cases, the for unequal numbers was selected as
muscle activity. A second computer onset data were removed from the the most appropriate form of analysis
program was used to rank the order data set. Of the 372 possible values for these data. This model is used to
in which each of the four muscles for muscle onset in the entire sample, reveal the influence of any indepen-

Physical Therapy/Volume 74, Number loanuary 1994

 there were significant differences
between left and right sides at only
Table 4. Side-to-Side Comparisons of "~reshold"
Vibration Percephon (in Meters
quar
-per- re second)a
one of the four frequencies (200 Hz)
for the control group, there were
significant differences between in-
Control Group (n=ll) jured and uninjured sides at three of
Frequency Mean Difference the frequencies (150, 200, and 250
(Hz) Mean of Left Slde (L) Mean of Rlght Slde (R) L-R Hz) for the injured group (Tab. 4).

Comparison of the mean values for

each side of the injured group sub-
jects with those of each side of the
control group subjects showed that a
greater strength of vibration was nec-
essary to reach threshold perception
Inlured Group (n=20)
-- -
- -
on both the injured and uninjured
FrssUenW Mean DMbmnce sides of the injured group subjects
(Hz) Mean of Injured Slde (I) Mean of Uninjured Slds (U) I-U than on either the left or right side of
the control group subjects. To deter-
mine whether the threshold percep-
tion values for the uninjured side of
the injured group subjects contrib-
uted to these group differences, Stu-
dent's t tests were applied to the
uninjured side of the injured group
"Two repetitions for each side. subjects versus each side of the con-
' ~ o tsignificantly different. trol group subjects. Significant differ-
CSignificantlydiierent at P< ,001. ences were found to exist at all fre-
quencies (P< .05) (Tab. 5).
dent variable on the dependent vari- Secondly, to determine whether vibra-
able and demonstrates whether there tion perception or muscle activation Electmmyographlc Analysls
is any significant difference between in the injured or uninjured limb was
two groups of unequal numbers. The significantly different from that on Separate statistical analyses were per-
statistical package used was the SAS either side of the control group sub- formed on data for each temporal
for Personal computers.# jects, further analyses were under- variable (ie, 0-H; 0,MI-MJ. To deter-
taken comparing the side-to-side dif- mine whether there were sigmficant
Initially, to determine any group dif- ferences between groups. differences between the injured and
ferences in vibration perception, com- control groups, a GLM analysis was
parisons of the vibration strength Results performed.
required for subject perception at
each frequency were made between Vlbratlon Perception 0-H. Reflecting the preparatory acti-
the injured and control groups by vation of the muscles prior to the
consolidating the data for the two Analysis of data for vibration percep- limb motion in hip extension, the
limbs in each case. Accordingly, tion at each frequency demonstrated a onset times for each of the four mus-
"group" was included as an indepen- significant difference between the cles in almost all instances preceded
dent variable in the GLM analysis. injured and control groups (PC ,001). the time of commencement of the
Similar comparisons were made be- As Figure 2 demonstrates, vibration reference activity (ie, hip motion),
tween groups for each of the two strength needed to be greater for the giving a negative value for 0-H. The
EMG measures. These analyses of the injured group than for the control greater the negative time span, the
data do not reveal whether differ- group in order for the subjects to earlier the onset of activity of that
ences exist in one side or the other, perceive the stimulus. muscle prior to hip extension motion,
or in both, but only that overall some whereas the smaller the negative time
alterations in vibration perception or A comparison of side-to-side differ- span, the later the onset. The results
in muscle activation may be associ- ences between groups (uninjured of analyses of this variable need to be
ated with injury. versus injured) showed that whereas interpreted with this in mind. Figure
3 represents typical EMG recordings
of a control group subject and an
injured group subject.
*SAS Institute Inc, PO Box 8000, Cary, NC 27511.

30 / 24 Physical Therapy/Volume 74, Number l/January 1994

 The analysis of side-to-side differences
for the gluteus maximus and ham-
string muscles revealed that for the
control group, the time span (O-H)
was significantly greater on the left
A
5. (stance) side than on the right @re-
ferred or skill) side (P<.05), indicat-
Injured ing an earlier onset of gluteus maxi-
mus and hamstring muscle activity on
the left side for uninjured subjects
(Tab. 6). The side-to-side differences
(injured versus uninjured sides) in
the injured group did not reach sig-
nificance for either of these muscle
groups. The significantly later time of
Contml onset of gluteus maximus muscle
activity for the injured group com-
pared with that of the control group
A
- -- (ie, with data for two sides consolidat-

I& 150 2d0 2 h

ed), however, suggested a delay in
gluteus maximus muscle activation on
Frequency [Hz) both sides of the injured group sub-
jects. Examination of the data in Table
- 6 demonstrates that this was so. A
Flgum 2. Vibration perception threshold of injured versw conhwl group subjects. Student's t test applied to the gluteus
Note that the greater the strength of vibration, tbe poorer the subjects' perception of maximus muscle activity onset data
vibration. for the uninjured side of the injured
group versus each side of the control
With the: data for the two sides con- than for the control group. That is, group highlighted the dif-
solidated, the GLM analysis showed for the injured group subjects, an ference that existed (P<.0005). No
that for the gluteus maximus mus- overall delay in activation of the significant side-to-side differences
cle, highly significant differences gluteus maximus muscle was evi- were found to exist for either the left

-
existed between the injured group dent. For the hamstring and the left or right lumbar errnor spinae
(mean =:-0.092) and the control and right lumbar erector spinae muscles.
group (.mean= -0.349) (PC .001), muscles, results of the GLM analysis
the time of onset of gluteus maxi- showed that the group differences 0, MI-M,. Analyses of the consoli-
mus muscle activity being signifi- were not significant. dated data relating to the time span
cantly later for the injured group between the onset of activity of the

Table 5. Comparison of Mean Valuesfor Visual Perception (in Meters per Square Second.)for Injured and Uninjured Sides of the
Injured Group With the Left and Right Sides of the Control Group

Frequency Dliference Dliference Difference

(Hz) II. CLb II-CL II II-CR IUd CL IU-CL IU

--

"II=injurt:d group injured side.

bC~=controlgroup left side.
"CR=control group right side.
d ~ ~ = ~ n j u rgroup
e d uninjured side.
eP< .05.

Physical Therapy /Volume 74, Number l/Januax-y1994

 level and the timing of onset of glu-
teus maximus muscle activity relative
to hip extension (0-H). ~esultsdem-
a onstrated that a positive correlation
existed for the injured group between
threshold vibration perception and
gluteus maximus muscle activation for
the 250-Hz frequency (PC.05). That is,
inclinometer the less sensitive the subjects were to
vibration at 250 Hz, the longer the
Right Lumbar delay in recruitment of the gluteus
Erector Spinae
maximus muscle for hip extension.
3. Y

Lett Lumbar
Erector Spinae

Significant differences in the sensory

and muscle function of subjects with
severe ankle sprain were shown to
1. exist when compared with that of
Biceps Femorls uninjured subjects. The decreased
ability to perceive vibration appears to
confirm the views of Freeman1 and
Wyke9 that a ligarnentous/capsular
injury influences the integrity of local
sensory receptors on the side of in-
jury, presumably through direct
damage.

2. The significant delay in activation of

the gluteus maximus muscle in the
injured group subjects and the posi-
3. tive correlation between a poorer
perception of vibration at 250 Hz and
4. gluteal muscle delay suggests that
joint injury involving sensory recep-
1. tors could influence the function of
Biceps F e m o h muscles pmximal to and removed
from the injury side. Even though this
study could not determine cause and
Figure 3. Graphic illustration of an electromyographic recording of a typical re- effect, this association provides sup-
cruitment pattern of (a) a control group subject and @) an injured group subject. port for the idea of a reflex chain of
(Arrows indicate onset of activation; numbers indicate the sequence of activation.) events that occurs following injury, as
pmposed by Lewit7 and Janda?
first and fourth muscles (0, MI-MJ to gluteus maximus muscle was almost
be recruited revealed a highly signifi- always the fourth muscle to be acti- The normal activation behavior of the
cant difference (PC .001) between the vated. This delayed activation was hamstring and lumbar erector spinae
injured and control groups. As Table therefore responsible for the greater muscles in the injured group can be
7 demonstrates, although the mean 0,MI-M4 time span found in the viewed together with the delay in
time span for the control group was injured group. activation of gluteus maximus muscle.
0.306 second, it was 0.527 second for A change in activation of all muscles
the injured group, or 72% longer The delayed activation in the gluteus could have led to the assertion that all
than for the control group. maximus muscle was used as the subjects in the injured group had a
variable for a correlation analysis of motor regulation problem, as has
The GLM analysis showed that there muscle and sensory function. A Pear- been intimated by previous stud-
was no significant difference between son Product-Moment Correlation ies.llJ2 The finding of significant acti-
sides in the 0,MI-M4 time span for Coefficient analysis was applied to vation changes in the gluteus maxi-
either group. Examination of the data for both groups relating to vibra- mus muscle, however, only points to
ranking incidence indicated that the tion strength at threshold perception the possibility that such a change is
associated with the ankle injury. Be-

Physical Therapy /Volume 74, Number ldanuary 1994

 tempt was made to ensure that in-
D jured subjects included in this study
Table 6. Comparison of Time Spans Between Onset of Muscle Activity and had no history of incoordination. As
Commencement of Hip Movement for Groups (in Seconds)
has been shown by this study, the
injured group subjects did, however,
Control Grow In= 11l have a sensory deficit compared with
Slde Dlfference
the control group subjects. Although
Muscle Group (L) Rlght (R) L-R it has been assumed that any differ-
ences from normal in the injured
group occurred as a result of injury,
Gluteus rrlaximus -0.451 -0.249 -0.202a in a retrospective study the origin of
Hamstring -0.562 -0.448 -0.1 14a the differences cannot be determined.
Left lumbar erector spinae -0.530 -0.450 O.08Ob From the point of view of the man-
Right lumbar erector spinae -0.504 -0.464 -0.040" agement of patients following ankle
sprain, however, the origin of the
Injured Group (n=20)
deficit does not affect the need for the
physical therapist to pay due attention
Slde
Dlfference to the need to improve sensoty and
Muscle Group Unlnjured (U) Injured (I) U-l motor function.20
-

Gluteus maximus -0.086 -0.098 -0.012@ This study has a number of implica-
-0.486 -0.51 1 0.025~
tions for the physical therapist. In
Hamstring
view of the likelihood that a deficit in
Left lumbar erector spinae -0.486 -0.432 -0.054'
sensory function is associated with
Right lumbar erector spinae -0.440 -0.463 0.023b
decreased muscle activity around
other joints, a rehabilitation program
"PC.05. should include a focus on improving
not significant. sensory function. Because muscles
respond in different ways to periph-
cause cause and effect were not the implications for influences on the eral injury, the results of this study
focus of this study, further research is uninjured side. suggest that the effects need to be
warranted to help clarify these sought in areas remote from the site
interrelationships. These results suggest that as a result of injury. This study has examined
of injury to the ankle joint, the activity only some of the muscles around the
Differences in vibration perception of the hip extensors on both sides of hip. Further investigations could re-
and activation of the gluteus maximus the body is diminished. Whereas veal whether muscle function changes
muscle on the uninjured side as well Stokes and Young3 and Iles et a15 have also occur in other joints following
as the injured side of the injured demonstrated decreased extensor ankle injury (eg, in the knee o r verte
group subjects when compared with activity at the site of injury, the results bra1 joints), o r indeed, whether they

-
the control group subjects support of this study suggest that there could might occur as a result of the effects
the concept of central adjustment of be a direct relationship of decreased of gluteal muscle delay.
motor control following injury. This activity of the extensors of the lower
finding suggests that a reflex chain of limb, involving muscles not only re- The differences in sensory function
events is not limited to the side of mote from the site of injury but also and in the function of some muscles
injury, but that there could also b e o n the opposite side of the body. It is on the uninjured side are also impor-
also possible that even after pain tant in treatment. Whether such differ-
following the ankle injury had ceased, ences are due to dysregulation at the
the function of the gluteus maximus cortical level or at a spinal level has
muscle in extending the hip was still to be determined. Nevertheless,
Table 7. Comparison of Mean Time
Span Between Onset of Activity of First compromised due, perhaps, to an the existence of differences highlights
and Fourth Muscles alteration in gait pattern established the need to examine both sides of the
during the period of injury. Such body in assessment. These results
possibilities are the subject of further emphasize the importance of the
Group N (8) research. physical therapist paying attention to
motor control and to the function of
Control (r1=11) 120 0.306a The question could be asked whether muscles around joints separated from
Injured (r1=20) 229 0.527a subjects in the injured group had a the site of injury.
basic neurological deficit that led to
their initial ankle sprain. Every at-

Physical Therapy/Volume 74, Number ldanuary 1994

CO~C~US~O~ 4 De Andrade JR, Grant C, Dixon A. Joint dis- 1 3 Freeman MA, Dean MR, Hanham WF. The
tension and reflex inhibition in the knee. aetiology and prevention of functional instabil-
J Bone Joint Susg [Am]. 1965;47:313-322. ity of the foot.J Bone Joint Susg [Br]. 1965;46:
The this study have shown 5 Iles JF, Stokes M, Young A. Reflex actions of 678-685.
that both local sensory and proximal knee joint afferents during contractions of the 14 Gross MT. Effects of recurrent lateral ankle
muscle function changes are associ- human quadriceps. Clin physiol. 1990;10:489- sprains on active and passive judgments of
500. joint position. Phys Ther. 1987;67:1505-1509.
ated with unilateral severe ankle
6 Freeman MAR, Wyke B. Articular contdbu- 15 Barrack FU,Skinner HB, Buckley SL. Pro-
sprain and that when some aspects of tions to limb muscle reflexes. Br J Susg. 1966; prioception in the anterior cruciate deficient
sensory and motor function deficits 53:61-63. knee. Am J Sports Med. 1989;17:14.
are considered, there is a positive 7 Lewit K. Manipulative Therapy in Rehabilita- 16 Schmitz TJ. Sensory assessment. In: O'Sulli-
correlation between the two. If com- tion of the Motor System. London, England: van SB, Schmitz TJ. Physical Rehabilitation:
Buttemonh & Co (Publishers) Ltd; 1985. Assessment and Treatment. 2nd ed. Philadel-
prehensive and effective management 8 Janda V. Muscles, motor regulation and back phia, Pa: FA Davis Co; 1988: chap 6.
of injury is to be ensured, a holistic problems. In: Korr IM, ed. The Neurologic 17 Jan& V. Muscle Function Testing. Boston,
approach to assessment is essential. Mechanisms in Mantpulative Tberapy. New Mass: Buttemonh; 1983.
York, NK Plenum Publishing Corp; 1978:27- 18 Roy S, Irvin R. Sports Medicine: Prevention,
41. Evaluation, Manugement, and Treatment.
9 Wyke B. The neurology of joints. Ann R Coll New York, NY:Prentice-Hall Press; 1983.
Susg Engl. 1967;41:25-50. 19 Dyck PJ, Karnes J, O'Brien PC, Zimmerman
10 Chusid JG. Correlative Neuroanatomy and IR Detection Thresholds of Cutaneous Sensa-
1 Freeman MA. Instability of the foot after an- Functional Neurology. 16th ed. Los Altos, Cal- tion in Humam 2nd ed. In: Dyck PJ, Thomas
kle injuries to the lateral ligament of the ankle if: Lange Medical Publications; 1976. PK, Lambert EH, eds. Sydney, Australia: WB
J Bone Joint Susg [BY].1965;47:669-677. Saunders Co; 1984:1103-1 138.
11 Tropp H, Odenrick P, Gillquist J. Stabilom-
2 Ekholm J , Eklund G, Skoglund S. On the etry recordings in functional and mechanical 20 Bullock-Saxton JE, Jan& V, Bullock M. Re-
reflex effect from the knee joint of the cat. instability of the ankle joint. Int J Sports Med. flex activation of gluteal muscles in walking
Acta Pkysiol &and. 1960;50:167-174. 1985;6:180-182. with balance shoes: an approach to restoration
3 Stokes M, Young A. The contribution of re- 12 Gaufin H, Pettersson Y, Tegner Y, Tropp of muscle function for chronic low back pain
flex inhibition to anhrogenous muscle weak- H. Function testing in patients with old rup- patients. Spine. 1993;18:704-708.
ness. Clin Sci. 1984;67:7-14. ture of the anterior cruciate ligament. Int J
Sports Med. 1990;11:73-77.

Invited Commentary

Dr Bullock-Saxton's article examines Scientists, such as Goldberger,' Hul- has the same stylistic appearance,
theoretical concepts of neural adapta- liger,z and Pearson: have shown the regardless whether we write it bill-
tion and motor control changes fol- importance of aferent information to board or microchip size.
lowing soft tissue injury. Equally as motor control. Similar to Dr Bullock-
important, the author poses the clini- Saxton's findings in the lower extrem- The author of this report has taken on
cal question, Should motor planning ity, Smith and Brunolli4 have reported a complex task. It should not be sur-
learning rehabilitation be an integral proprioceptive deficits in the upper prising, therefore, that certain vulnera-
part of treatment protocols following extremity following soft tissue disrup- bilities exist regarding methodology
orthopedic-type injuries? These issues tion. In my laboratory, we have col- and data interpretation. Because ques-
are ripe for investigation, and, as a lected preliminary data indicating tions regarding methodology tend to
neuroscientist and a physical therapist, change in muscle afferent activity can be of interest mainly to other investi-
I was very pleased to be invited to actually alter neural responses to gators, I will only briefly mention a
provide this commentary. muscle fatigue.5 The possibility that few concerns. I would have found it
sensory changes in one joint may useful to know the average amount of
First and foremost, I would like to cause alterations at sites distant to it is time, and ranges, that had elapsed
commend the author for tackling a also consistent with recent computa- between subjects' ankle sprain and
technically, and theoretically difficult, tional models of human movement.617 testing. Given time, some injured
problem. The hypothesis that changes The nervous system appears to con- peripheral sensory fibers will regen-
within articular sensory receptors-as trol movement by comparing degrees erate. Relating time course of recov-
a result of soft tissue injury-alter of freedom at each joint. If joint A ery to vibratory sensory perception to
postural reflexes, has a strong theoret- moves x degrees, then other joints vibration would have been a correla-
ical rationale. The Bobaths and Signe will alter their movement trajectories tion of interest. The type of physical
Brunnstrom were among the first to accordingly in order to attain the therapy intervention and the length of
make clinical use of the importance desired end point of movement. Pre- time each subject was immobilized
of sensory input to motor control. sumably, this is why our handwriting following ankle sprain would also

Physical Therapy /Volume 74, Number loanuary 1994