| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
M Martin, PT, MS, is Staff Physical Therapist, 64th Combat Support Hospital, US Army, Wurzburg, Germany
M Shinberg, PT, MSPT, is Physical Therapist, Shriner's Hospital for Children, Philadelphia, Pa
M Kuchibhatla, PhD, is Assistant Research Professor, Department of Biostatistics and Bioinformatics, Center for the Study of Aging and Human Development, Duke University, Durham, NC
L Ray, PT, MS, is Pediatric Physical Therapist, Duke University Health System, Durham, NC
JJ Carollo, PhD, PE, is Director, Center for Gait and Movement Analysis, The Childrens Hospital, Denver, Colo, and University of Colorado Health Sciences Center, Denver, Colo
ML Schenkman, PT, PhD, is Professor, Physical Therapy Program, University of Colorado Health Sciences Center, C-244, 4200 E Ninth Ave, Denver, CO 80262-0244 (USA) (margaret.schenkman{at}UCHSC.edu).
Address all correspondence to Dr Schenkman
Submitted October 31, 2000;
Accepted December 6, 2001
 | Abstract |
|---|
Key Words: Gait analysis • Gait initiation • Kinesiology/biomechanics • Neuromuscular disorders • Parkinson disease
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussion and ConclusionsReferences |
|---|
Initiation of gait from a stationary position requires a transition from being relatively stationary to moving. Gait initiation is a task that challenges the balance control system by forcing an individual from a state of stable balance to a continuously unstable posture during walking.8 A better understanding of gait initiation may be helpful for designing interventions for people in early stages of PD and for monitoring the benefits of such interventions.
One approach to understanding initiation of gait in people in relatively early stages of PD is to characterize the relationship between COM and center of pressure (COP). Although sometimes confused, COM and COP represent 2 distinctly different quantities.9 The COM is the point on the body that moves in the same way that a single particle would move if subjected to the same external force, or the point at which the weight of the body can be considered to act.10 In contrast, the COP is the time-varying signal recorded from a force platform in the plane of the floor.11 During quiet standing, the fore-aft translation of the COP is largely influenced by net ankle moments associated with postural control, whereas the medial-lateral translation is largely influenced by hip control in the frontal plane. Changes in the COP reflect the central nervous system's response to movement in whole-body COM; in effect the COP describes the forces that must be produced to return the COM to a balanced position.
During quiet standing, the time-averaged vertical projections of the COM and the COP should coincide in the transverse plane.12,13 This situation represents a subject's most stable standing posture. With changes in body position, the distance between COM and COP projections increases, making the subject inherently less stable and necessitating active postural control to return the COM to a stable position within the base of support. The distance between vertical projections of the COM and the COP in the transverse plane (COM–COP distance) has been used to characterize chair rise and stair-climbing activities and, most recently, gait initiation.12,14,15 The greater the COM–COP distance, the more active postural control is needed. Therefore, individuals with intact postural control can more readily tolerate a larger COM–COP distance, whereas those with less effective postural control are likely to reduce this distance during transitional movements in an effort to reduce the need for active postural control.
Initiation of gait poses a demand for postural control because the person must make a transition from a relatively steady state (stance) to walking.13 During such a transition, controlled muscular effort is required to move the body COM away from the stable starting position, where COM and COP projections are aligned, producing an inherently unstable posture. Stability is further affected because the relatively wide and stable base provided by double-limb support is replaced by a more narrow and unstable base of single-limb support. Gait initiation represents a natural but deliberately destabilizing task during which conditions are established to promote forward progression while maintaining balance.16 During gait initiation, people with PD might be expected to limit the COM–COP distance, both because of the high demands on the postural control system and because of difficulty generating momentum. Therefore, we believe that gait initiation may be an ideal task for use in identifying and diagnosing abnormalities in the locomotor system, including those that may not be apparent with less sensitive measures taken during steady-state walking (eg, 10-m walk time, 6-minute walk distance).
A number of strategies have been used to characterize gait.8,12,13,17,18 Typical movements of the COM and the COP during gait initiation have been well characterized and are summarized in Figure 1.12,19,20 Gait initiation begins with the movement of the COP posterolaterally toward the extremity that will become the initial swing extremity, whereas the COM moves anterolaterally toward the initial stance extremity. The characteristic separation of the COM and the COP is an important feature of gait initiation, and the distance between them (COM–COP distance) can be used to quantify postural stability.12 Figure 1B illustrates 5 events that are commonly identified in the COP trajectory.13 Although the specific point marking the end of gait initiation and the beginning of steady-state walking varies among investigators,8,11–13,16–20 the point where the initial stance extremity breaks contact with the supporting surface is a reasonable place to end the gait initiation cycle. We refer to this point as "toe-off."
|
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
|
Instrumentation and Model Specifications
Kinematic data were collected by use of a Peak5 video motion measurement system* with 2 60-Hz Panasonic
DT5100 video cameras and synchronized video recorders conforming to the 400-line S-VHS standard. A total of 17 passive, retroreflective markers were attached bilaterally to a subject's fifth metatarsal heads, lateral malleoli, knee joint lines, greater trochanters, humeral heads, elbow joint lines, wrist joint lines, and temporomandibular joints and unilaterally to the forehead in the midsagittal plane. Each camera recorded the 2-dimensional positions of all markers and, following centroid calculation, a discrete linear transform was used to reconstruct 3-dimensional marker trajectories by use of software included with the Peak5 system. The 17 markers were then used to construct a simple 12-segment whole-body model, in which limb segments were defined by proximal and distal markers, the trunk was defined by trochanteric and humeral head markers, and the head was defined by the 2 lateral temporomandibular joint markers and the single midsagittal forehead marker. In our opinion, the model that we used is identical to the 19-marker whole-body spatial model described in the Peak5 manual because the heel markers used in the Peak5 description (and omitted in this experiment) do not contribute to the COM calculation. Estimates of segment mass centers and distributions were based on Dempster's anthropometric data22 and were expressed as distance from distal markers and percent segment distribution relative to whole-body mass. The individual segment COMs were then projected into the transverse plane and combined by use of superposition (weighted sum along each axis) to obtain an estimate of the whole-body COM.
Ground reaction force data were collected by use of 2 AMTI model OR6–5 force platforms with 6-channel strain gauge amplifiers.
The platforms were mounted side by side and oriented such that the laboratory coordinate system origin coincided with the posterior corner of the right-hand platform, with the y-axis aligned in the direction of forward progression. From each platform, force components in the 3 principal axes (Fx, medial-lateral; Fy, fore-aft; and Fz, vertical) and the moments about these axes (Mx, My, and Mz) were sampled at 600 Hz by use of the analog-to-digital interface included in the Peak5 system. Force and moment data from each platform were then combined by use of moment balance equations to determine the instantaneous COP across the platform array. These data were then synchronized and time divided to match the kinematic data, yielding time-locked COM and COP spatial trajectories at 16.67-millisecond intervals (60 Hz) suitable for further processing.
Experimental Protocol
Subjects began each trial standing barefoot in a relaxed position with feet on separate force platforms. Initial foot position was self-selected by the participants. They initiated forward locomotion at a self-selected pace after receiving a verbal cue from one of the investigators, who simultaneously triggered an electronic event marker identifying the start of a new trial. The completion of gait initiation brought the participant to the end of the force platform. For each subject, one practice trial was followed immediately by 2 data collection trials. In an effort to enhance between-trial consistency, the subject's feet were traced and the tracings were used prior to the start of each new trial for repositioning on the force platform. Force platform and kinematic data recording began prior to the event marker and continued until the subject cleared both platforms.
COM–COP Distance Calculation
The time-synchronized COM and COP trajectory data were exported onto a spreadsheet, and the numerical representation of each trajectory's starting point was offset in both x and y directions so that both had common spatial origins. This procedure resulted in a cluster of trajectory data around the origin during quiet standing followed by a smooth 2-dimensional trajectory after gait initiation. The temporal transition between quiet standing and gait initiation (t0) was defined as the first sample point in the COP trajectory in the direction of the initial swing limb at which no further point was closer to the spatial origin. The x and y coordinates of the COP trajectory prior to t0 were averaged, and the average x value and the average y value replaced the coordinate values at t0. All trajectory data for both COP and COM prior to t0 were discarded. The x and y coordinate values at t0 were used as an offset, effectively translating COP and COM origins to (0,0) at t0. With a common time base and a common spatial origin, the quantity COM–COP distance is easily determined by applying a conventional geometric distance formula between the coordinates of COM and the coordinates of COP at distinct points in time ti:
|
|
Events Identified in the Gait Initiation Cycle
The following 5 events in the COP trajectory were used to characterize the gait initiation cycle (Fig. 1B). Event A is the most lateral motion of the COP toward the extremity taking the first step (initial swing extremity); event B is the most posterior position of the COP as it moves in the direction of the initial swing extremity; event C is the first event after the COP crosses the midline as the COP moves laterally toward the initial stance extremity; event D is the shift in COP from lateral to anterior motion; and event E is the final event, which occurs with 100% of gait initiation, corresponding to the time when the initial stance limb breaks contact with the supporting surface (referred to as "toe-off"). The COM–COP distance was calculated for each trajectory when each event occurred by use of the equation; the result was used to test the primary hypothesis of this investigation. Figure 2 illustrates the COM–COP distance for event E.
|
Data Analysis
The primary hypothesis was that there would be differences in COM–COP distance among the 3 groups during the 5 events in gait initiation. A multivariate analysis of variance (MANOVA), which controls for type I errors,23 was used to test for overall group differences. An analysis of variance (ANOVA) then was used to determine which variables within the set showed group differences. Differences among subject groups were further examined by use of the Scheffe post hoc analysis for variables that were normally distributed and Kruskal-Wallis analysis for those that were not.
Secondary exploratory analyses (not driven by the primary hypothesis of this investigation) consisted of examination of differences during the 5 events for the following variables: (1) movement of COM and COP along the x-axis; (2) percentage of gait initiation cycle when the event occurred; and (3) elapsed time when the event occurred. We used an ANOVA to determine group differences and either Scheffe or Kruskal-Wallis post hoc analysis to further characterize the differences. Our secondary analyses were carried out at 20% increments during gait initiation.
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
|
To further understand the reasons for differences in COM–COP distance among subject groups, we carried out an exploratory analysis of motion of COM and COP along the x-axis as well as the elapsed time and the percentage of gait initiation at which the events occurred (Tab. 3). With regard to motion of the COM along the x-axis, the only difference among the groups was for event E, transition from lateral to anterior motion (subjects with PD versus younger subjects without PD). For all 5 events, the pattern was for the motion of the COM along the x-axis to be largest for the younger subjects without PD and smallest for the subjects with PD.
|
For the percentage of gait initiation at which any of the 5 events occurred, no differences were detected. For the elapsed time in gait initiation during which these events occurred, there was a difference between groups at event D, the shift from lateral to anterior (subjects with PD versus younger subjects without PD).
Finally, to determine how movement patterns differed among the 3 groups across equivalent events during gait initiation, trajectories were compared at 20% intervals. This analysis permitted a comparison among groups at equal intervals normalized over time, without regard to what event occurred at any particular time. For this analysis, gait initiation was normalized to the total elapsed time, and 4 variables (COM–COP distance, COM motion along the x-axis, COP motion along the x-axis, and total elapsed time) were determined for 20%, 40%, 60%, 80%, and 100% of gait initiation.
The COM–COP distance was inspected first for the 3 groups of subjects across the entire gait initiation process, as shown in Figure 3. This figure demonstrates that overall, the younger subjects without PD had the largest COM–COP distance, whereas the subjects with PD had the smallest COM–COP distance. The COM–COP distance was different among the 3 groups at 20%, 40%, and 100% of gait initiation (Tab. 4). Post hoc analysis revealed group differences between the subjects with PD and the older subjects without PD at 20% of gait initiation, between the subjects with PD and both of the other groups at 40% of the activity, and between the subjects with PD and the younger subjects without PD at 100% of gait initiation.
|
|
| Discussion and Conclusions |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
The results of our investigation demonstrate that the relationship between COM and COP provides a means to identify both spatial and temporal indicators of problems during initiation of gait by people in relatively early stages of PD. The subjects with PD were different from the older subjects without PD for several of variables examined and were different from the younger subjects without PD for almost all variables. For almost all variables tested, the pattern was consistent: the younger subjects without PD had the largest COM–COP distance, followed by the older subjects without PD and then the subjects with PD. These findings suggest to us that the COM–COP distance during gait initiation provides a useful tool for identifying subtle difficulties with movement performance and may differentiate subjects with and without PD or differentiate adults without PD but with and without subtle impairments of movement performance.
To further interpret the results, we considered conflicting requirements of forward progression and stance stability (Fig. 1B). The first 2 events (A and B) normally occur as force is exerted through the initial swing limb in order to propel the COM laterally toward the initial stance limb and then forward in preparation for forward progression. As the COM moves toward the initial stance limb, more and more of the subject's weight is transferred to the stance side, causing the COP path to move in that direction as well. By limiting the lateral displacement of the COM (subjects with PD had the smallest COM–COP distance), subjects with PD appeared to compensate for deficiencies in movements (Tab. 2).
As the COP path of progression crosses the midline (event C), both of the subject's feet are still in contact with the supporting surface. During this event, the subjects with PD again had the smallest COM–COP distance, followed by the older subjects without PD. The older subjects without PD, therefore, also appeared to experience subtle impairments of movement performance as they too limited COM–COP distance when the consequences of destabilization were greatest, that is, in the transition from double- to single-limb support.
As the COP path enters the initial stance side footprint, the initial swing side foot is being lifted from the supporting surface and swing limb toe-off occurs in close proximity to the time of event D. The trajectory of the COP is going from a predominantly lateral to a predominantly anterior translation. During this event, there were no differences among the 3 subject groups, and standard deviations were large. The observed variability may reflect the different strategies that individuals use to accomplish the 2 competing demands of this task—forward progression of the COM and stance limb stability.
Finally, as the initial swing limb is advanced through space, the COP travels in an anterior direction along the length of the initial stance limb footprint to final event (event E, corresponding to 100% of gait initiation). At this time, the initial swing limb makes contact with the supporting surface. This is the beginning of double-limb support. During event E, the COM–COP distance of the younger subjects without PD was different from that of the subjects with PD, being larger by an average of about 6 cm. Once the conditions used to easily control the COM are in place, that is, both feet are on or about to be on the supporting surface, the younger subjects without PD appeared to be willing to move through larger COM–COP displacements. In contrast, older subjects without PD and those with PD moved through smaller displacements of the COM relative to the COP. This shortening of COM–COP distance may reflect a need to preserve stability because of impairments of postural control mechanisms, may result from an inability to generate adequate momentum during initiation of gait, or may reflect some combination of these 2 factors. The shortening of COM–COP distance is consistent with the observation that the step length of people with PD is shorter than that of older adults without PD.7
In the exploratory secondary analyses, both spatial and temporal events that related to the control of the COM–COP distance were examined. These exploratory analyses allowed us to examine the x-axis displacements for the events previously identified by Jian et al.13 In addition, we performed similar analyses at 20% increments of the gait initiation cycle. In this way, we were able to remove the temporal factor from the initiation sequence and determine whether there were additional differences not previously identified.
For all spatial variables, differences were found between the younger subjects without PD and the subjects with PD, and in all instances, the COP or COM x-axis displacement of the younger subjects without PD was greater than that of the subjects with PD. Furthermore, for both of the COM variables that differed and for all except one of the COP variables that differed, the displacement of the older subjects without PD was midway in magnitude between those of the other 2 groups. These findings suggest to us that people with PD maintain stability by keeping the COM and the COP as close together as possible throughout the path of progression. The results of these exploratory analyses can be used to guide future research on gait initiation of people with a variety of disorders.
For all of the temporal variables, the younger subjects reached steady-state gait from quiet standing faster than the older subjects without PD or the subjects with PD. This finding is consistent with the observations of Halliday et al.8 The younger subjects without PD may have moved the COP farther as well as faster. Instantaneous 2-dimensional velocity vectors could be examined at several points in gait initiation and may prove to be another useful tool for quantifying and describing subtle differences in gait initiation among subject groups.
Several limitations of this study must be considered. The 12-segment model used to determine whole-body COM was based on individual segment lengths and kinematics calculated from surface markers placed at approximate joint centers, rather than a true 6-degree-of-freedom (6-df) model for each segment. In a 6-df model, each limb segment is defined as a separate rigid body and includes both segment translation and segment rotation about all 3 axes.24 Joint angular displacement is then calculated on the basis of the relative movements of individual segments. The 12-segment model used in this experiment is theoretically less accurate than a 12-segment 6-df model, because transverse rotations are ignored and joint angular displacement is described only by line segments connecting the skin markers placed near the surface projection of the true joint center. However, errors in the calculated COM are small when the simpler model is used because gait initiation involves primarily translational motions. In our opinion, the added complexity and more difficult marker placement associated with a 6-df model did not warrant the negligible improvement in COM determination for the purposes of this preliminary study.
To simplify presentation of a complex data set, we did not discuss the y-axis displacements during the secondary analyses. These displacements are more a function of stride length or walking velocity than postural control.8 These data are available and could be examined in the future.
Another limitation is that stance width among participants was not standardized because we chose to investigate self-selected gait initiation. Stance width may affect the absolute displacements of COM and COP and was controlled for within subjects by use of foot tracings to ensure the same placement in each trial. In our opinion, this variable should be examined in future investigations. Control of initial stance width, coupled with a measure of bradykinesia, may provide additional insight into the relative contributions of postural control and momentum generation of individuals in relatively early stages of PD.
Finally, the sample size was small, possibly contributing to our finding of some nonsignificant differences among groups. In addition, because the sample size was small, we were not able to differentiate among participants in Hoehn and Yahr stages 1, 2, and 3. Future studies should address these issues.
Despite the limitations, the methods we used were powerful enough to demonstrate differences between the subjects with PD and those without the disease. Our results demonstrate that individuals with relatively early PD systematically limit the inherently destabilizing COM–COP distance throughout gait initiation, allowing less separation of these 2 vertical projections than is seen in individuals without PD. These limitations in the COM–COP distance are apparent even in people with early to middle stages of PD. The smaller COM–COP distance may reflect a need to control stability, an inability to generate momentum, or some combination of the 2 factors. Furthermore, older participants without PD also had subtle alterations of gait initiation, as indicated by the COM–COP distance. Analysis of this variable in gait initiation, combined with an analysis of temporal variables, may lead to a better understanding of gait limitations in people with PD.
Our data indicate that the COM–COP distance measure may provide a potent indicator—for patients with PD, multiple sclerosis, and other progressive disorders—of impending difficulties with movement performance before overt problems are apparent clinically. With the method described in this report, postural control also may be used to compare active versus sedentary older individuals to further explore activity-related aspects of balance control, even among older individuals without specific disorders.
| Footnotes |
|---|
This work was supported by the National Institutes of Health, National Institute on Aging, Claude D Pepper Older Americans Independence Center, Grant No. 5 P60 11268.
* Peak Performance Technologies Inc, 7388 S Revere Pkwy, Ste 603, Englewood, CO 80112. 
Panasonic Industrial Products, 1 Panasonic Way, Secaucus, NJ 07094.
Advanced Mechanical Technology Inc, 176 Waltham St, Watertown, MA 02472.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussion and ConclusionsReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Related Article
This article has been cited by other articles:
|
|
 |
|
  A. Nieuwboer, R. Dom, W. De Weerdt, K. Desloovere, L. Janssens, and V. Stijn Electromyographic profiles of gait prior to onset of freezing episodes in patients with Parkinson's disease Brain, July 1, 2004; 127(7): 1650 - 1660. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
