the influence of horizontal velocity on interlimb symmetry in normal walking

Human Movement Science 22 (2003) 271–283

www.elsevier.com/locate/humov

The influence of horizontal velocity oninterlimb symmetry in normal walking

D.J. Goble a,b,*, G.W. Marino b, J.R. Potvin b

a Center for Human Motor Research, Division of Kinesiology, The University of Michigan,

401 Washtenaw Avenue, Ann Arbor, MI 48109-2214, USAb Faculty of Human Kinetics, Department of Kinesiology, University of Windsor, 401 Sunset Ave.,

Windsor, ON, N9B 3P4 Canada

Abstract

Changes in horizontal velocity (HV) are known to influence many biomechanical charac-

teristics of human locomotion. The purpose of the present study was to investigate this phe-

nomenon with respect to the interlimb symmetry of walking in a normal population. Peak and

temporal ground reaction force data from both feet of 20 able-bodied males were collected at

each of three relative velocity conditions (slow, normal and fast). These data were analyzed

using of a series of 2· 3 repeated measures ANOVAs, which revealed a high degree of inter-

limb (bilateral) symmetry across HV conditions despite significant intralimb (unilateral)

changes. In contrast to this primary finding were two significant interaction effects for the

stance time and peak vertical force at push-off measures respectively. These interactions indi-

cated greater asymmetries at the slow HV condition with a trend toward improved symmetry

at higher velocities. Although these results may provide some theoretical insight into the un-

derlying nature of symmetry in gait, their overall magnitude does not seem to invalidate the

current widespread use of symmetry assumptions in clinical and research settings today.

� 2003 Elsevier B.V. All rights reserved.

PsycINFO classification: 2330

Keywords: Walking; Velocity; Lateral dominance

* Corresponding author. Address: Center for Human Motor Research, Division of Kinesiology, The

University of Michigan, 401 Washtenaw Avenue, Ann Arbor, MI 48109-2214, USA. Tel.: +1-734-302-

1268; fax: +1-302-936-1925.

E-mail address: [email protected] (D.J. Goble).

0167-9457/$ - see front matter � 2003 Elsevier B.V. All rights reserved.

doi:10.1016/S0167-9457(03)00047-2

mail to: [email protected]

272 D.J. Goble et al. / Human Movement Science 22 (2003) 271–283

1. Introduction

Symmetry can be defined as an exact correspondence between opposite halves of a

figure or form. Conversely, asymmetry is any deviation from this �ideal� structure. Ingeneral, left/right symmetry is observable in humans when a mid-sagittal plane di-vides the body. Characteristic of this �bilateral equality� are opposing limbs, paired

sensory organs and a centralized spinal column. Despite this inherent symmetry in

the natural development of the human body, asymmetries also frequently exist.

Two common examples are leg length differentials (Kaufman, Miller, & Sutherland,

1996) and muscle imbalances induced by limb dominance (Chhibber & Singh, 1970).

In both these cases the body displays a systematic bias toward a particular limb and,

thus, a shift away from symmetry. The nature and extent of this shift have warranted

a fair measure of scientific investigation to date. For a comprehensive review of thiswork see Sadeghi, Allard, Prince, and Labelle (2000).

The presence of symmetry between the left and right lower limbs of a normal in-

dividual during walking is a common assumption in both clinical and research set-

tings. Examples from the clinic include comparisons between individuals with

dysfunctional gait and hypothetical �normal� individuals with perfect gait symmetry.

As well, examinations of symmetry between the lower limbs of impaired individuals

are often made pre- and post-rehabilitation as a method of measuring the efficacy of

a treatment paradigm (for example Becker, Rosenbaum, Kriese, Gerngrof, & Claes,1995). In the research domain, interlimb symmetry is a key consideration in many

biomechanical studies as the ability to assume symmetry allows a reduction in the

amount of data collected and, subsequently, reduces the complexity of the overall

data analysis.

The most common justification for the assumption of symmetry in human gait has

long been that it is a necessary means of maximizing energy efficiency. However, this

philosophy has recently been challenged based on studies of clinical populations with

neurological asymmetry (i.e. hemiparetic stroke, unilateral amputation, Hunting-ton�s disease, etc. . .). In these populations, affected individuals typically change their

preferred walking style to one of greater asymmetry in order to accommodate their

physiological limitations (i.e. having one side of the body that is more impaired than

the other). This new pattern of locomotion, although markedly asymmetric, usually

has greater metabolic efficiency than one where greater symmetry is imposed upon

the individual. This was shown clearly by Mattes, Martin, and Royer (2000) for

the energy efficiency of amputee gait. In this study it was found that patterning a

prosthetic leg to have more symmetrical weight and moment of inertia characteristicsto that of the unaffected limb actually decreased the energy efficiency of the individ-

ual�s gait. This result suggests that optimizing the efficiency of gait is not necessary

the key determinant of symmetry in an individual, but rather, symmetry is the con-

sequence of some other factor(s).

Since it is well established that changes in horizontal velocity (HV) have a substan-

tial effect on the biomechanical characteristics of walking (Andriacchi, Ogle, & Ga-

lante, 1977; Grieve, 1968; Jahnke, Hesse, Schreiner, & Maurity, 1995; Keller et al.,

1996; Kirtley, Whittle, & Jefferson, 1985), a logical hypothesis might be that HV is

D.J. Goble et al. / Human Movement Science 22 (2003) 271–283 273

in someway related to the degree of symmetry between the limbs in gait. This has been

tested in clinical populations of individuals with neurological asymmetry with results

indicating that walking speed does affect the amount of symmetry between lower

limbs (for example see Donker & Beek, 2002). However, it has yet to be determined

if a similar relationship might also exist in the gaits of able-bodied individuals.A common method of determining symmetry in gait is the assessment of ground

reaction force (GRF) data collected via a dynamometric force platform. Force plat-

forms provide objective measures of the foot–floor reaction forces created during

walking that are reliable, readily analyzed, and highly practical in their use as an as-

sessment tool of walking gait (Grabiner, 1993). Previous research using this instru-

mentation has demonstrated evidence for both gait symmetry (Giakas &

Baltzopoulos, 1997; Hamill, Bates, & Knutzen, 1984; Menard, McBride, Sanderson,

& Murray, 1992) and gait asymmetry (Herzog, Nigg, Read, & Olsson, 1989; White,Agouris, Selbie, & Kirkpatrick, 1999) in able-bodied subjects. However, these studies

did not adequately address the potential influence of HV on gait symmetry. This is

likely due to the difficulty associated with velocity testing, as it requires a great deal

of experimental rigor. In the present experiment we undertook this arduous process

in an attempt to formally assess the overall effect of varied horizontal velocities on

interlimb symmetry in normal gait.

2. Methods

2.1. Subjects

Ethical approval for the testing of human subjects was obtained, and informed

consent received, from a subject pool of 20 normal males (mean age 23.8 ± 2.2 yr;

mean height 178.8 ± 8.4 cm; mean weight 815.7 ± 96.3 N). All subjects underwent

a screening process to ensure the appropriateness of their participation in the study.The screening process consisted of a series of health-related questions concerning the

individual�s neuromuscular and musculoskeletal status, as well as a direct measure of

leg length inequality to ensure discrepancies of less than 0.02 m (Gross, 1978). Sub-

jects also had to demonstrate some degree of right lower limb dominance as deter-

mined by the leg with which subjects preferred to kick a ball (Coren & Porac, 1978).

2.2. Experimental setup

The experimental setup used for all subject testing is depicted in Fig. 1. This setup

included a rectangular gait runway with an AMTI (Advanced Mechanical Technolo-

gies Inc.) force platform embedded flush to the surface two-thirds of the distance from

the take-off area. The platform was interfaced with a microcomputer running Lab-

VIEW software for the collection of all GRF data. A Stalker ATS radar system

(Applied Concepts Inc.), consisting of a tripod-mounted radar gun (9� nominal band-

width, ±0.02 m/s accuracy) focused centrally along the subject�s estimated path of

travel, was used to monitor walking velocity. The radar gun was positioned to the

Fig. 1. Schematic of the experimental set-up used for data collection. A runway-embedded force platform

obtained GRF data from walking trials while a tripod-mounted radar gun monitored HV.


rear of the subject in amanner such that the subject walked away from it during the tri-

als. This provides themost accurate readings of the slower velocities associatedwith the

initiation of walking and limits any interference of subjects with the instrumentation.

For each subject the height of the radar gun was adjusted to a line of sight level

with that subject�s thorax. Such a position is important to minimize conflicting veloc-

ity readings from the swinging extremities during gait. The radar gun was manually

triggered during testing as the subject reached a distance of 4 m from the force plat-form. Continuous velocity recordings were made for the duration of the walking task

and subsequently recorded to a Dell Inspiron 4000 laptop computer via the Stalker

ATS collection program. Through a function of this program, a statistic was gener-

ated which outputted the subject�s instantaneous HV at 4 m, corresponding to the

point in time at which they were crossing the force platform. As well, an online re-

cording of a subject�s velocity profile throughout a trail was available to the re-

searcher, which ensured individuals crossed the platform at a relatively constant

velocity.

2.3. Experimental protocol

Prior to the GRF data collection subjects were instructed to mount the gait run-way from the takeoff area and to carry out a series of practice trials. The practice

trials allowed subjects to gain familiarity with the experimental setup and to adjust

to the constraints of the walking task. Although, there is evidence to suggest that tar-

geting a force plate has no significant effect on force values (Grabiner, Feuerbach,

Lundin, & Davis, 1995) there is still some argument as to whether or not this is a

valid concern. Therefore, for the purposes of this study, walking trials were only

deemed successful if the following criteria were met: (1) subject did not noticeably

target the force platform, (2) the subject contacted it entirely with the contact foot(CF) of interest and (3) the subject contacted the platform at the desired HV. As

well, subjects were required to make a straight approach to the platform from a dis-

tance of at least 4 m, and to take at least three steps post-contact with the platform.


Once accustomed to the experimental setup and walking task, specific walking ve-

locities at which the subject would be tested were determined. Overall, three HV

ranges were calculated for each subject based on a normal, self-selected speed of

walking. The normal speed was determined by averaging radar-measured HV values

for five consecutive individually self-paced trials. Ranges for each of the slow, nor-mal and fast HV conditions of walking were then determined from this normal HV

value. First, the normal range was calculated as being within ±2.5% of the normal

speed value. Next, based on the normal range, a fast velocity range of 10% greater

than normal, and a slow velocity range of 10% less than normal, were determined.

The value of 10% was chosen as it mimics a relatively normal range of velocities that

might be encountered during the performance of one�s daily activities.

Walking data for each subject was collected in six blocks of 15 successful gait tri-

als with one block for each combination of the HV (slow vs. normal vs. fast) and CF(left vs. right) conditions. To compensate for potential acquisition effects a random

block design for the order of conditions presented was used. Visual and computer-

based feedback aided the researchers in determining the acceptability of all trials.

Any trials that the researchers deemed to be unsuccessful were marked and elimi-

nated from the data set.

2.4. Data reduction and statistical analysis

GRF curves for the vertical (Fz) and anterior–posterior (Fy) directions were col-

lected for each subject trial at a sampling rate of 1000 Hz and subsequently reduced

to 11 representative parameters (Chao, Laughman, Schneider, & Stauffer, 1983). Fig.

2 illustrates 10 of the parameters on typical Fz and Fy curves with the 11th param-

eter used being a simple measure of the overall stance time. All peak force measures

were normalized to individual subject body weight and all corresponding time to

peak force measures were normalized to individual trial stance time. The individual

parameters are briefly defined as follows:

Fz1––Peak buildup of force at healstrike

Tz1––Time of occurrence of Fz1

Fz2––Peak unweighing of the body during knee flexion


Fz3––Peak force created during pushoff of the foot


Fy1––Peak braking force of the foot after contactTy1––Time of occurrence of Fy1

Fy2––Peak propulsive force of the foot prior to toe off

Ty2––Time of occurrence of Fy2

Stance time––Measure of the entire time the foot is in contact with the ground

The raw parameter data were initially collapsed into individual mean and stan-

dard deviation values for each experimental condition and used to calculate individ-

ual coefficient of variation (CV) values. Next, a series of 2 · 3 repeated measures

Fig. 2. Normal ground reaction force curves in the vertical (Fz) and anterior–posterior (Fy) directions

with the peak force (Fz1, Fz2, Fz3, Fy1 and Fy2) and time to peak force (Tz1, Tz2, Tz3, Ty1 and Ty2)

parameters selected for analysis.


ANOVAs were conducted generating F -ratios for the main effects of HV (fast vs.normal vs. slow) and CF (left vs. right) as well as the interaction effect between these

two independent variables (HV ·CF). Where necessary, post-hoc analyses were ac-

complished by means of paired t-tests. Statistical significance was set at a level of

p6 0:01 for all analyses to compensate for the effects of multiple comparisons.

3. Results

The mean (SD) velocity ranges calculated across all individuals in this study were

1.21 (0.16)–1.28 (0.17) m/s for the slow condition, 1.35 (0.18)–1.42 (0.19) m/s for the

normal condition and 1.48 (0.20)–1.56 (0.21) m/s for the fast condition. The overall

mean (SD) for each of the 11 gait parameters measured, collapsed across subjects,

are summarized in Table 1 along with their respective CV values. CV values calcu-

lated in this study were used as an indicator of the GRF data variability. Typically,

for work of this nature, a CV value of 12.5 or more is deemed an unacceptable

amount of variability (Munro, Visintainer, & Page, 1986). All variables measuredin this study were found to be within acceptable limits.

Table 1

Collapsed subject means (SD) for GRF parameters at each test condition with associated CV value

Parameter Slow Normal Fast CV

Right Left Right Left Right Left

Fz1 1.116 (0.035) 1.113 (0.037) 1.163 (0.040) 1.165 (0.037) 1.239 (0.042) 1.242 (0.040) 3.3

Fz2 0.804 (0.024) 0.809 (0.024) 0.751 (0.028) 0.745 (0.026) 0.682 (0.030) 0.676 (0.032) 3.8

Fz3 1.140 (0.025) 1.127 (0.022) 1.155 (0.026) 1.161 (0.026) 1.175 (0.033) 1.180 (0.033) 2.4

Fy1 )0.193 (0.018) )0.191 (0.020) )0.212 (0.019) )0.209 (0.018) )0.241 (0.021) )0.241 (0.020) 9.3

Fy2 0.197 (0.013) 0.197 (0.012) 0.220 (0.012) 0.222 (0.013) 0.244 (0.014) 0.243 (0.014) 6.0

Tz1 0.230 (0.016) 0.225 (0.020) 0.223 (0.014) 0.225 (0.011) 0.220 (0.009) 0.218 (0.010) 5.9

Tz2 0.457 (0.024) 0.446 (0.022) 0.467 (0.022) 0.468 (0.020) 0.478 (0.017) 0.475 (0.018) 4.4

Tz3 0.749 (0.011) 0.748 (0.012) 0.755 (0.011) 0.758 (0.011) 0.760 (0.009) 0.761 (0.011) 1.5

Ty1 0.147 (0.020) 0.145 (0.020) 0.151 (0.019) 0.152 (0.014) 0.154 (0.014) 0.154 (0.015) 10.4

Ty2 0.851 (0.010) 0.851 (0.009) 0.857 (0.009) 0.857 (0.010) 0.856 (0.009) 0.858 (0.009) 1.0

Stance time 0.725 (0.017) 0.735 (0.018) 0.690 (0.015) 0.686 (0.015) 0.648 (0.013) 0.647 (0.014) 2.3

D.J.Goble

etal./HumanMovem

entScien

ce22(2003)271–283

277

Table 2

F -ratios (2, 19) and significance values (p6 0:01 in bold) for the main effects of HV and CF as well as the

interaction between the two (HV·CF)

Parameter F -ratio (significance)

HV CF HV ·CF

Fz1 52.00 (<0.01) 0.02 (0.90) 0.36 (0.70)

Fz2 76.75 (<0.01) 0.13 (0.75) 1.84 (0.17)

Fz3 10.81 (<0.01) 0.03 (0.89) 5.43 (<0.01)

Fy1 32.51 (<0.01) 0.11 (0.75) 0.13 (0.73)

Fy2 64.05 (<0.01) 0.00 (0.99) 0.39 (0.68)

Tz1 1.68 (0.20) 0.31 (0.59) 1.29 (0.29)

Tz2 7.58 (<0.01) 1.03 (0.32) 3.04 (0.06)

Tz3 30.45 (<0.01) 0.30 (0.59) 1.90 (0.17)

Ty1 2.35 (0.11) 0.09 (0.76) 0.36 (0.56)

Ty2 7.88 (<0.01) 0.13 (0.72) 0.70 (0.50)

Stance time 114.94 (<0.01) 0.41 (0.53) 5.76 (<0.01)


F -ratios and the associated p-values for the 2 · 3 ANOVAs are detailed in Table 2.

From the analysis a significant main effect of HV was prevalent in all parameters

tested, except for the Tz1 and Ty1 variables. Post-hoc analyses among the levels

of HV produced varied results with a general trend such that higher HV significantly

increased the absolute values of Fz1, Fz3, Fy1, Fy2, Tz2, Tz3 and Ty2 while decreas-ing the Fz2 and stance time parameters. There was no main effect of CF for any of

the parameters tested.

Two significant interaction effects resulted from the ANOVAs. First, an interac-

tion effect was found for the stance time measure as depicted graphically in Fig. 3.

Fig. 3. Stance time as a function of HV. Significant differences ( ) were found in the mean difference

between the left and right feet for the slow HV condition as compared with the normal and fast HV con-

ditions.

*

Fig. 4. Fz3 as a function of HV. Significant differences ( ) were found in the mean difference between left

and right feet for the slow HV condition as compared to the normal and fast HV conditions.


Post-hoc analysis of the mean differences between feet at each HV for this interaction

revealed a significant effect of speed in determining symmetry. Significant differences

were found between right and left feet at the slow HV condition as compared to the

normal, tð19Þ ¼ �2:86, p < 0:01, and fast, tð19Þ ¼ �3:01, p < 0:01, conditions. Theasymmetry at the slow condition was characterized by larger stance times in the left

foot and shorter in the right.

The second interaction effect was for the Fz3 (peak vertical force at push-off) pa-

rameter as depicted in Fig. 4. Similar to the stance time measure, the post-hoc anal-

ysis for this interaction found significant differences in the mean left/right differences

between feet at the slow HV condition as compared to normal, tð19Þ ¼ 2:97,p < 0:01, and fast, tð19Þ ¼ 2:69, p < 0:01 conditions. However, in contrast to stance

time, the asymmetry at the slow HV was characterized by a tendency toward greatervalues in the right foot rather than the left foot.

4. Discussion

Not surprisingly changes in the overall magnitude of all GRF parameters were

witnessed with HV manipulation. This was evidenced by a strong main effect of

HV for the majority of the parameters tested and is attributable to the variedamounts of muscle force, and altered rates of loading, necessary for walking at dif-

ferent velocities. Since this effect has been discussed in detail by a variety of other

researchers (Andriacchi et al., 1977; Grieve, 1968; Jahnke et al., 1995; Keller et al.,

1996; Kirtley et al., 1985) it was not a primary focus of the present work. Rather, this

study set out to address the influence that walking at different HV�s within a normal


speed range might have on the interlimb symmetry of normal gait. Insight into this

question lies within the results for the main effect of CF and the interaction effect

(HV ·CF).A significant main effect of CF represents a systematic tendency toward the use of

one limb as opposed to the other regardless of the HV condition. For the data col-lected in the present study this effect was found to be non-significant for all 11 gait

parameters. This suggests that, on the whole, symmetry is largely maintained during

walking at different velocities and supports the work of previous studies that ascer-

tained walking is a fully symmetrical task. As well, this finding lends credence to clin-

ical and research practices that operate under the assumption of symmetry in normal

gait. The importance of this result should not be overlooked. In a large number of

clinical and experimental testing environments the ability to strictly regulate horizon-

tal walking velocity is logistically very difficult to accomplish. For example, in theclinic walking speeds of individuals vary not only in response to the testing proce-

dure but, also, over time with changes in physical functioning. In the lab, velocity

changes are prevalent for tasks that are either novel to a subject or involve a level

of complexity that dictates a slower pace of activity. The ability to confidently as-

sume symmetry in these situations has benefits in both the amount and in the time

of data collected and analyzed as they are significantly reduced when unilateral mea-

sures are made instead of bilateral measures.

Although the main effect of CF for this study indicates symmetry in gait, it doesso irrespective of speed. Since it was hypothesized that the degree of symmetry in gait

might be mediated by HV, the interaction effects (CF·HV) are also an important

means of assessing the degree of symmetry in gait. Overall, significant interactions

were found for both the stance time and Fz3 measures. Specifically, these measures

displayed increased symmetry at the normal and fast HV conditions as compared to

slow. This result is similar to that found in studies of symmetry for clinical popula-

tions with asymmetrical functioning but is seemingly the first of its kind for a sample

of able-bodied individuals. Perhaps the best explanation for this result is offeredwithin the domain of dynamical systems theory and motor control. In this domain

there are abundant examples of enhanced interlimb coupling with increases in veloc-

ity (for an overview of some of this research, see Beek, Peper, & Stegeman, 1995). In

this way, it is proposed that segments of the human body act as coupled oscillators

where symmetrical relationships (in-phase and anti-phase) are more easily main-

tained at higher velocities than other more complex phase relations. The significant

interactions from the present study follow a similar pattern of spatiotemporal con-

straint where at slow speeds the legs are afforded the opportunity to decouple andemploy different functional strategies, while at greater velocities more highly coupled

patterns of movement exist.

If the limbs in gait less coupled at lower velocities then is it possible that this is

done to accomplish some form of functional outcome? When interpreting the results

of the two interaction measures an interesting parallel with theories that purport

functional roles for the lower limbs in gait can be made. In these theories it is gen-

erally accepted that the left or non-dominant limb is used in a supportive manner

in gait while the right or dominant limb performs more of a propulsive function


(Hirokawa, 1989; Matsusaka, Fujitta, Hamamina, Norimatsu, & Suzuki, 1985;

Sadeghi, Allard, & Duhaime, 1997). This dissociation of roles between limbs is often

referred to as a �functional asymmetry� and is theorized to be a solution to the two

main goals of human locomotion: stability and speed.

The interaction for the stance time measure indicated that at the slow HV agreater asymmetry existed between left and right feet. This asymmetry was such that

the left foot spent a larger amount of time in the stance phase of walking than the

right. Stance time is a measure of the entire time the foot is in contact with the walk-

ing surface. Having a longer stance time during gait is often associated with an

attempt toward greater stability, as can be seen in amputee gait and in other abnor-

mal walking populations. Thus, at the slow HV it appears as though the subjects

tested used the left foot slightly more as a supportive mechanism than the right.

For the second interaction effect the Fz3 measure indicated a similar trend to thatof the stance time where at the slow HV condition greater asymmetry occurred be-

tween limbs. Specifically, this asymmetry indicated a tendency toward larger force

production by the right foot as opposed to the left. The Fz3 parameter is a measure

of the peak vertical force occurring during the push-off or propulsive phase of stance.

A tendency toward greater force production in this variable at the slow HV condi-

tion is suggestive of the right foot being used in more of a propulsive or mobilizing

function. This would confirm the use of the right or dominant limb for somewhat

more of a propulsive role during gait.Caution should be taken with regards to the interaction results for this study, as

the majority of evidence points toward symmetry in gait rather than asymmetry. This

is evident in the fact that similar interaction effects were not found for all the param-

eters tested especially those that are typically linked to the support or propulsive

phases of walking. For example, one would certainly expect to see a similar pattern

of response by the anterior–posterior propulsive peak force (Fy2) if one foot is being

used more specifically for propulsion. However, this result did not present itself in

the data collected and similar arguments can be made for any of the parameters thatfailed to show significance.

A second cause for consideration with respect to the results of the interactions is

their overall practicality. The few asymmetries found, although statistically signifi-

cant, are relatively small (<1%) when taken as a percentage of their respective

parameter�s average value. In the research setting, where statistical tests of signifi-

cance are made on a frequent basis, this small difference may or may not have a

direct influence on results based on assumptions of symmetry. As well, in the clinic,

it is unlikely that such a small difference would influence to any degree the rehabil-itation processes currently prescribed by clinician to patient.

Overall then it seems the results of this study have both theoretical and practical

merit. From a theoretical standpoint the possibility for velocity dependent symmetry

changes based on limb function seems plausible yet, warrants further investigation.

One particular aspect that should be looked at in greater depth is the effect that lar-

ger velocity manipulations than those used in this study may have on the symmetry

of gait. In this study the HV ranges chosen were approximately 10% different than

that of a normal self-selected pace. This choice had the advantage of keeping subjects


within a comfortable range of walking speeds and correlated to changes that might

occur naturally in a clinical or research situation where HV was not rigorously con-

trolled. However, with greater perturbations it is more likely that asymmetries could

be evoked (should they exist) and more learned about the underlying mechanisms of

the system as a whole. Finally, from a practical standpoint, the majority of evidencesupports the common practice of using symmetry assumptions in the assessment of

an individual�s gait. This is an important finding for both clinicians and researchers

who, for a variety of reasons, commonly make assumptions of symmetry in assess-

ments of human gait.

Acknowledgements

This work was funded in part by the Ontario Graduate Scholarship (OGS) fund.

A special thanks is offered to Diane Grondin for her selfless contributions to this

work.

References

Andriacchi, T. P., Ogle, J. A., & Galante, J. O. (1977). Walking speed as a basis for normal and abnormal

gait measurements. Journal of Biomechanics, 10, 261–268.

Becker, H. P., Rosenbaum, D., Kriese, T., Gerngrof, H., & Claes, L. (1995). Gait asymmetry following

successful surgical treatment of ankle fractures in young adults. Clinical Orthopaedics and Related

Research, 31(1), 262–269.

Beek, P. J., Peper, C. E., & Stegeman, D. F. (1995). Dynamical models of movement coordination. Human

Movement Science, 14, 573–608.

Chao, E. Y., Laughman, R. K., Schneider, E., & Stauffer, R. N. (1983). Normative data of knee joint

motion and ground reaction forces in adult level walking. Journal of Biomechanics, 16(3), 219–233.

Chhibber, S. R., & Singh, I. (1970). Asymmetry in muscle weight and one-sided dominance in the human

lower limbs. Journal of Anatomy, 106(3), 553–556.

Coren, S., & Porac, C. (1978). The validity and reliability of self-report items for the measurement of

lateral preference. Journal of Psychology, 69, 207–211.

Donker, S. F., & Beek, P. J. (2002). Interlimb coordination in prosthetic walking: Effects of asymmetry

and walking velocity. Acta Psychologica, 110, 265–288.

Giakas, G., & Baltzopoulos, V. (1997). Time and frequency domain analysis of ground reaction forces

during walking: An investigation of variability and symmetry. Gait and Posture, 5, 189–197.

Grabiner, M. D. (1993). Current issues in biomechanics. Champaign, IL: Human Kinetics Publishers.

Grabiner, M. D., Feuerbach, J. W., Lundin, T. M., & Davis, B. L. (1995). Visual guidance to force plates

does not influence ground reaction force variability. Journal of Biomechanics, 28(9), 1115–1117.

Grieve, D. W. (1968). Gait patterns and the speed of walking. Biomedical Engineering, 3, 119–122.

Gross, R. H. (1978). Leg length discrepancy: How much is too much? Orthopedics, 1(4), 307–310.

Hamill, J., Bates, B. T., & Knutzen, K. M. (1984). Ground reaction force symmetry during walking and

running. Research Quarterly, 55(3), 289–293.

Herzog, W., Nigg, B., Read, L., & Olsson, E. (1989). Asymmetries in ground reaction force patterns in

normal human gait. Medicine and Science in Sports and Exercise, 21(1), 110–114.

Hirokawa, S. (1989). Normal gait characteristics under temporal and distance constraints. Journal of

Biomedical Engineering, 11(6), 449–456.

Jahnke, M. T., Hesse, S., Schreiner, C., & Maurity, K.-H. (1995). Dependences of ground reaction force

parameters on habitual walking speed in hemiparetic subjects. Gait and Posture, 3(1), 3–12.


Kaufman, K. R., Miller, M. D., & Sutherland, D. H. (1996). Gait asymmetry in patients with limb-length

inequality. Journal of Pediatric Orthopaedics, 16, 144–150.

Keller, T. S., Weisberger, A. M., Ray, J. L., Hasan, S. S., Shiavi, R. G., & Spengler, D. M. (1996).

Relationship between vertical ground reaction force and speed during walking, slow jogging, and

running. Clinical Biomechanics, 11(5), 253–259.

Kirtley, C., Whittle, M. W., & Jefferson, R. J. (1985). Influence of walking speed on gait parameters.

Journal of Biomedical Engineering, 7, 282–288.

Matsusaka, N., Fujitta, M., Hamamina, A., Norimatsu, T., & Suzuki, R. (1985). Relationship between

right and left leg in human gait from a viewpoint of balance control. In D. A. Winter, R. Norman, R.

Wells, K. Hayes, & A. Patla (Eds.), Biomechanics IX-A. Champaign, IL, USA: Human Kinetics

Publishers.

Mattes, S. J., Martin, P. E., & Royer, T. D. (2000). Walking symmetry and energy cost in persons with

unilateral transtibial amputations: matching prosthetic and normal limb inertial properties. Archives of

Physical Medicine and Rehabilitation, 81, 561–586.

Menard, M. R., McBride, M. E., Sanderson, D. J., & Murray, D. D. (1992). Comparative biomechanical

analysis of energy-storing prosthetic feet. Archives of Physical Medicine and Rehabilitation, 73, 451–

458.

Munro, B. H., Visintainer, M. A., & Page, E. B. (1986). Statistical methods for health care research.

London: JB Lipincott Co.

Sadeghi, H., Allard, P., & Duhaime, M. (1997). Functional gait asymmetry in able-bodied subjects.

Human Movement Science, 16, 243–258.

Sadeghi, H., Allard, P., Prince, F., & Labelle, H. (2000). Symmetry and limb dominance in able-bodied

gait: a review. Gait and Posture, 12, 34–45.

White, R., Agouris, I., Selbie, R. D., & Kirkpatrick (1999). The variability of force platform data in

normal and cerebral palsy gait. Clinical Biomechanics, 14, 185–192.

the influence of horizontal velocity on interlimb symmetry in normal walking

Documents