Trunk kinematics during locomotor activities.

Trunk kinematics are critically important to the maintenance of body equilibrium and should be examined as a component of locomotion analysis.(1,2)

Three-dimensional, normal trunk kinematic data obtained during stair and chair locomotor daily activities, as well as during gait, enable therapists to make comparisons with pathological locomotion characteristics such as "gluteus medius" limp following hip surgery, trunk abduction lurch associated with knee arthroplasty or above-knee amputation, and spinal fusion. To date, however, most locomotor studies have only described lower-extremity kinematics.(3-8) Comparatively little information exists on the kinematics of upper-body movement during gait, and no three-dimensional kinematic data have been reported for other locomotor activities of daily living such as stair and chair activities.(9)

In the 1960s, Murray and colleagues,(7,8) described the transverse (medial/lateral [internal/external]) rotation of the trunk for free-speed walking of 60 healthy men (20-65 years of age) to be 6.9 [+ or -] 1.9 degrees ([closure of the set][+ or -]SD). Recently, Opila-Correia(10) reported three-dimensional trunk kinematics observed during the free-speed gait of 14 women, 21 to 54 years of age ([closure of the set]=35.0, SD=10.4). In gait trials with subjects wearing low-heeled footwear, trunk flexion/extension, abduction/ adduction, and medial/lateral total angular excursions relative to the pelvis averaged 11.1, 12.6, and 17.5 degrees, respectively. These same displacements relative to room coordinates were 9.2, 5.2, and 11.2 degrees, respectively. Treadmill gait was studied by Thorstensson et al(11) and Stokes et al.(12) Although Thorstensson et al did not report transverse-plane rotation, they did report net frontal-plane trunk range of motion (ROM) in 7 healthy 18- to 34-year-old subjects to be 2 to 9 degrees and net flexion/extension ROM to be 2 to 12 degrees in walking. Stokes and colleagues analyzed trunk movement of 3 female and 5 male subjects during treadmill walking (with shoes on) and reported "small" flexion/extension amplitudes, with a mean trunk abduction/adduction of 4.9 [+ or -] 1.8 degrees and a mean transverse-plane trunk rotation of 4.7 [+ or -] 1.6 degrees. It is clear that free walking differs from treadmill gait(13) and that room-referenced trunk kinematics differ from pelvis-referenced kinematics. Because the locomotor control system, using sensory input from the vestibular system, is aware of global, gravity-referenced coordinates, we examined both selfreferenced (pelvis-referenced) and gravity-referenced (room-referenced) trunk motions during locomotor activities of daily living.

The purposes of our study were (1) to determine a sample of trunk ROM and maximum angular orientation in flexion/extension, abduction/ adduction, and medial/lateral rotation in healthy subjects during daily activities, including free-speed and paced gait, during stair climbing and descending, and while rising from a chair and (2) to examine these kinematic findings in two frames of reference-the trunk relative to the pelvis and the trunk relative to room (gravity) coordinates.

Method

Subjects

Eleven volunteers who were free of musculoskeletal and neurological disease participated in the study. Subjects were recruited from the community and the laboratory staff. All subjects, to be eligible for the study, must have been (1) able to walk at least 1.6 km (1 mile) without stopping and to climb and descend stairs and arise from a chair without personal or upper-extremity assistance and (2) free of neuro-musculo-skeletal pathology and deformity as determined by history and physical examination. All subjects provided written informed consent. Participants' ages, sex, heights, and weights are presented in Table 1.

Instrumentation

Gait trials were conducted on a 10-m walkway. Stair trials involved the use of a weighted modular staircase with four stairs. The first of the four steps was 2.5 cm tall and was provided merely to help initiate steady-state stepping activity prior to data collection on three 18-x 28-cm stairs. Chair trials involved the use of an armless and backless chair with an adjustable-height rigid seat.

Four Selspot II optoelectric cameras,(*) in addition to PDP 11/60(t) and MicroVAX [TM](tt) computers, were used to acquire bilateral kinematic data from the subjects. The Selspot system's infrared light-emitting diodes (LEDs) are tracked by an infrared detector within each camera; the system accuracy is [is not greater than] 3 mm.(14) Camera placement resulted in a viewing volume of 1.8 m per side (Fig. 1). Kinematic data were acquired for 3 seconds at 153 Hz.

(*) Selspot AB, Flojelbergsgatan 14, S-431 37 Molndal, Sweden, and Selspot System Ltd, Troy, MI 48093.

(t) Digital Equipment Corp, 146 Main St, Maynard, MA 01754.

(tt) Developed at the Massachusetts Institute of Technology, Cambridge, Mass.

The LEDs were mounted in rigid arrays secured to 11 body segments: head, trunk, pelvis, thighs, shanks, feet, and upper arms (Fig. 2). Each segment was modeled as a rigid body having 6 degrees of freedom (three translations and three rotations), with its kinematics determined in part using TRACK(C) software(tt)(14) and the technique described by Riley et al.(15) The orientation of each body segment in space and the associated joint angles were calculated using 3-1-2 Cardan angles.(16,17)

Procedure

General. Barefoot subjects performed all activities in a single session. To prevent the preferred cadence from being influenced by the paced cadence, free (preferred-speed) gait was followed by paced gait. Stair climbing, then stair descending, then arising from a chair were performed with at least 1.5 minutes' rest between trials.

To approximate easily reproduced natural cadences and to prevent velocity-dependent kinematic changes from confounding the within- and between-subject comparisons, the paced cadence chosen for each activity was determined from pilot studies and previous literature.(7,8) All activities, except rising from a chair, were performed with unrestricted arm-swing.

Free gait. Subjects walked along a 10-m walkway at a comfortable speed; that is, subjects were instructed to "walk the way you usually do, as if you were taking a brisk stroll in the park."At least three strides were completed before the subjects entered the data-collection portion of the walkway. Several acclimation trials were completed before two trials of data were collected.

Paced gait. Subjects walked using an identical procedure to that of free gait, but each foot-strike was synchronized with a metronome set at 120 beats per minute (bpm). Subjects practiced this paced gait cadence until synchrony was comfortably achieved for each step, as determined by the heel-strike occurring at each metronome beat, without awkward-appearing "marching."

Stair climbing. Subjects climbed stairs to the beat of the metronome set at 80 bpm, with the left foot first striking the small step and the right foot striking the next full riser. Subjects stopped when they reached the top platform, which was the fourth step. Several practice trials were performed before the two data-collection trials to ensure smoothness and cadence synchronization with the metronome.

Descending stairs. Subjects started on the fourth (top) step and descended in time with the beat of a metronome set at 80 bpm and with the left foot striking the third step. Several practice trials preceded two data-collection trials, again to ensure smoothness and synchronization with the metronome.

Rising from a chair. Subjects were seated such that the chair edge was approximately 4 cm distal to their greater trochanters. Foot distance from the chair was determined by setting the angle between the floor and the tibia to 72 degrees (18 [degrees] of dorsiflexion). The subjects' feet were equidistant from the chair and 10 cm apart Chair height was adjusted to the height from the floor to the lateral knee joint line when the tibia was perpendicular to the floor (Fig. 3).(9) Subjects rose in time with the beat of a metronome set at 52 bpm, with arms folded to minimize the upper extremities' involvement in the movement. Subjects began rising at 1 beat and came to a full stand 1.2 seconds later at the next beat. Several practice trials were performed until subjects could perform the trials smoothly within the 1.2-second time limit. Data from two trials were collected for each subject.

Data Analysis

The trunk's peak (maximum or minimum) angle and total range of angular displacement were determined for both the trunk segment relative to room orientation and the trunk segment relative to the pelvic segment. Using a computer program that displayed maximum, minimum, and total ROM in the sagittal, frontal, and transverse planes (flexion/extension, abduction/adduction, and medial-lateral rotation, respectively), the angular values of each trial were obtained from a "window" corresponding to a complete cycle (right heel-strike to right heel-strike for gait and stair activities) or from the complete chairrising activity. Velocity and cycle time were determined from center-of-mass (CM) displacements in this same "window." Means, standard deviations, and repeated-measures multivariate analysis of variance (MANOVA) results were calculated with the Statistical Analysis System (version 6.03)[sections of] for the IBM PC.[~~] Multivariate statistics for multiple dependent variables were used to preserve the experimentwise alpha level at [is not greater than] .05.

Results

Descriptive linear velocity and cycle-time data are shown in Table 2. Results from a typical subject are presented in Figure 4. Angular displacement results are illustrated in Figures 5 and 6. Velocity, pattern of motion, and trunk angular displacements did not differ between free and paced gait trials; hence, the remainder of the "Results" section and the "Discussion" section focus on paced gait trials, as compared with paced stair and paced chair activities of daily living. Between-trial reliability within subjects was very high ([closure of the set] [is not greater than] 3 [degrees] difference between trials for any activity/motion combination and trial-to-trial Pearson r [equal or greater than] .88), but moderate between-subject variability is apparent in the standard deviations in Figures 5 and 6, being greater in peak motions than in ROM (peak-to-peak amplitude). No age-related peak motion or ROM differences were found in any plane or coordinate system (F [is not greater than] 1.2, P [is greater than] .10) in any activity.

Trunk ROM during rising from a chair in all planes relative to either the pelvis or room coordinates was significantly different from stair and gait ROM (Hotelling-Lawley F [is greater than] 70.1, P [is not greater than] .01) with the exception of chair versus gait abduction/adduction (F=.18, P=.68). Gait trunk ROM was similar (F [is not greater than] 5.6, P [is greater than] .05) to stair-descending trunk ROM in all planes, but gait ROM differed significantly from stair-climbing ROM in all planes (F [is greater than] 10.7, P [is not greater than] .01). Trunk ROM did not differ significantly between stair climbing and descending, except in medial/lateral rotation (F= 11.5, P= .01).

(sections of) SAS Institute Inc, PO Box 8000, Cary, NC 27511.

(~~) International Business Machines Corp, PO Box 1328, Boca Raton, FL 33432.

The greatest trunk ROM relative to the pelvis occurred for flexion/extension during rising from a chair (22.9 [degrees] [+ or -] 9.5 [degrees]) and was quadruple that of gait sagittal ROM (Fig. 5A). Room-referenced flexion/extension during rising from a chair (36.2 [degrees] [+or -] 7.7 [degrees]) exceeded by [is greater than] 50% (F=19.8, P [is not greater than] .001) that relative to the pelvis (Fig. 5B). Nonsagittal trunk ROM relative to the pelvis was greatest in stair climbing (abduction/adduction, 14.7 [degrees] [+ or -] 5.7 [degrees]) and least in rising from a chair (medial/ lateral rotation, 3.2 [degrees][+ or -] 1.1[degrees]) (Fig. 5A). Indeed, trunk ROM relative to the pelvis was greater than that relative to the room coordinates across all activities and motions (F [is greater than] 5.3, P [is not greater than] .05), except in chair and stair medial/lateral rotation (Figs. 5A and 5B).

During gait, trunk maximum sagittal orientation was about 3 degrees more toward extension relative to the pelvis (Fig. 6A) than relative to the room coordinates (Fig. 6B); this difference was not significant. Maximum trunk flexion relative to the room coordinates (Fig. 6B) during stair climbing was double that recorded during stair descending and sextuple that recorded during gait (F [is greater than] 15.3, P [is not greater than] .01). Relatively small, and statistically insignificant, differences existed in peak trunk lateral and transverse maximal positions among the five activities.

Discussion

We examined the kinematics of the trunk referenced to room coordinates and to the pelvis during common 10comotor activities of daily living (ie, gait, stair, and chair activities). We chose a sample that was heterogeneous in height, weight, age, and gender to enhance the generalizability of the findings from this study. Because of the small sample size, however, firm conclusions await future studies. We found no differences in trunk kinematics with respect to age. We stipulated rather high functional locomotor levels, however, in the sample inclusion criteria for all subjects, including our elderly subjects. These high locomotor levels may have attenuated the possibly disease-induced differences that Murray and colleagues(7,8) observed in their elderly subjects compared with younger healthy subjects. As in the lower extremities, trunk kinematic patterns are very repeatable within subjects, but each subject used slightly idiosyncratic movement patterns and amplitudes, particularly in stair climbing and descending (note the standard deviation magnitudes in Figs. 5 and 6). In general, the standard deviations of peak motions exceeded those of ROM, because maximum angular displacements depend on each person's posture (segment anatomical orientations), which varies across subjects, and on movement amplitude. Range of motion depends only on movement peak-to-peak amplitude. For example, subjects could attain 36 degrees of trunk flexion ROM during rising from a chair by an excursion from 10 degrees of extension to 26 degrees of flexion, or 5 degrees of flexion to 41 degrees of flexion.

Kinematic Patterns

Trunk kinematic patterns are defined by Stokes et al (12) as the relative invariance of the number and cycle location of curve inflections and reversals. For example, in the chair-rising activity depicted in Figure 4, flexion relative to the room coordinates has a major reversal at about 55% of the cycle (abscissa) and an inflection (mathematically, where the second derivative of the curve changes sign) at about 35% of the cycle. The following discussion highlights only those movement patterns that were typical of the subjects investigated according to the criteria of Stokes et al. Future studies should consider larger representative samples, to delineate the extent of variations from the typical patterns determined in our study.

Control of the trunk is important for posture and balance, in part because of the upper body's relatively large mass (about two thirds of the body mass lies superior to the waist). It is difficult to depict three-dimensional

  Table 2.  Average Forward Velocities
and Cycle Times ([+ or -] Standard Deviation)
for Each Event
                      Velocity         Cycle
Activity              (m/s)            Time (s)
Free gait             1.12 [+ or -] 0.13       1.08 [+ or -] 0.07
Paced gait            1.18 [+ or -] 0.12       1.06 [+ or -] 0.06
Stair climbing        0.39 [+ or -] 0.05       1.60 [+ or -] 0.29
Stair descending      0.36 [+ or -] 0.07       1.58 [+ or -] 0.25
Rising from a chair ...                        1.32 [+ or -] 0.12

real-life patterns in two-dimensional written reports such as this; however, we include a comparative kinematic depiction to illustrate selected aspects of how trunk kinematic patterns may differ in subjects with known balance and locomotor control dysfunction (Fig. 7).

Gait. Sagittal-plane patterns typically included a flexion peak near each heel-strike, with maximum extension occurring during single-limb support (Fig. 4), but the amplitude of these motions was small (Figs. 5A and 5B). Apparently, the vertical CM excursion during gait previously reported(8) is directly linked with the trunk sagittal angular displacements. The CM was highest during single-limb support and lowest during double-limb support, the periods of peak trunk extension and flexion, respectively.

Frontal-plane trunk motions relative to the pelvis tended to occur toward the stance limb, reaching their maximum at the time of opposite side toe-off (Fig. 4). That is, at right foot contact, the trunk was midway in its movement from left to right leaning, and this motion continued until left toe-off, at which time a reversal occurred and the trunk began to lean toward the left side. Relative to room vertical, the abduction/adduction amplitude was lower and the curve had fewer inflection points than did pelvis-referenced motions; in general, we observed pelvis-referenced trunk abduction toward the soon-to-be stance limb, which reversed about 5% of a cycle following ipsilateral foot contact. The greater trunk-to-pelvis ROM was due to independent pelvis motions moving out of phase with the trunk.

Transverse trunk rotation relative to room coordinates was also 180 degrees out of phase with the pelvis and achieved maxima about 10% of a cycle after each heel-strike, rotating so that the ipsilateral shoulder was posterior to the heel-strike limb, nearly directly over the foot at mid-stance, and maximally anterior to the stance limb near toe-off (Fig. 4). A slight timing difference was apparent in the pattern relative to room coordinates in that curve reversals were achieved directly coincident with each heel-strike. That is, the trunk-pelvis transverse-plane maxima that occurred after each heel-strike were apparently caused by contravening pelvic rotation continuing after the trunk's intrinsic rotation period was tion/adduction ROM during stair climbing or descending (Figs. 5A and 5B) and peak flexion relative to room coordinates (Fig. 6B) substantially exceeded those values obtained during gait. Peak flexion values obtained during stair climbing reflect a more inclined posture assumed as subjects oriented the trunk to roughly parallel the stairs' 33-degree slope. After assuming this flexed posture, subjects demonstrated a trunk flexion/extension ROM relative to the pelvis that was similar to their gait ROM (Fig. 5A). Subjects descended the stairs with much less maximum trunk flexion relative to the room coordinates (Fig. 6B) than during stair climbing, apparently to maintain stability by shifting the trunk's mass away from the stairs' declension.

Trunk frontal-plane ROM relative to the pelvis during stair climbing was greater than that of gait, apparently to help clear the swing foot over the step and thus minimize lower-limb flexion requirements (Fig. 5A). These data also suggest that greater trunk abduction/adduction ROM relative to the room coordinates during stair locomotion, as compared with gait (Fig. 5B), stems from the relatively greater vertical displacement of the body's mass required to clear each stair.(8) The lesser transverse rotation relative to the pelvis during stair locomotion, as compared with gait, is apparently due to the shorter step and stride lengths required during stair climbing and descending. Motion patterns relative to the room coordinates during stair climbing were essentially opposite those of gait: the trunk was abducted away from the limb approaching stance, apparently to assist limb clearance over the step; the trunk then traversed laterally toward the stance limb until opposite footstrike occurred (Fig. 4). Trunk-topelvis frontal-plane patterns, however, were similar to gait patterns.

Rising from a chair. Very little lateral or transverse-plane motion was apparent in these subjects, and no appreciable pattern was discernable in these motions. By contrast, the trunk followed a very typical sagittal-plane pattern, both relative to the pelvis and to room coordinates. Between movement initiation and lift-off from the seat, the trunk and pelvis flexed in synchrony, apparently to aid in lift-off from the seat. Almost no flexion relative to the pelvis occurred, but substantial trunk flexion relative to the room coordinates was evident. After lift-off, trunk-pelvis synchrony decreased. The pelvis flexed more than the trunk, inducing relative spinal extension (Fig. 5A), apparently to aid attainment of the body's final vertical position.(9)

Dally Activity Demands on Spinal Mobility

These data suggest the extent of spinal mobility required by healthy subjects to perform activities of daily living. Trunk movements result from forces induced by trunk and limb muscles, or forces induced by ground reactions and inertia from the moving segments.(11) Patients with spinal fusion or muscle dysfunction limiting trunk mobility would presumably compensate with greater lower-limb or neck motions to maintain postural and head stability during daily activities. Combined spinal and pelvifemoral restriction could significantly disrupt all locomotor activities of daily living, particularly given the importance of pelvic kinematics suggested by these data. Trunk motion impairments would be expected to impede sagittal gait kinematics slightly, but stair and chair activities would probably be substantially impaired. This assertion deserves future study.

Normal whole-body kinematics apparently depend on smooth spinopelvic coordination. For example, the chief function of the erector spinae and anterolateral abdominal muscles during erect standing or locomotion is to resist gravity and to stabilize the trunk relative to gravity.(19) Unconstrained trunk flexion would be expected to occur during locomotion if muscle and hip movement restraints were absent, as in quadriplegia or if ligamentous integrity were surgically compromised. One function of the greater trunk flexion posture found during stair activities than during gait (Figs. 6A and 6B) may thus be that in stair locomotion, greater trunk inclination helps to project the wholebody CM forward along the moving base of support, providing a source of stable anteriorly directed momentum. Greater trunk inclination would also reduce lower-limb stress by damping the body's vertical CM oscillations. This hypothesis, although consistent with these data, should be investigated in future studies.

Trunk abduction/adduction ROM relative to the pelvis (Fig. 5A) was generally greater than that observed relative to room coordinates (Fig. 5B). These data support the assertions of Saunders et aP that lateral and transverse-plane pelvic rotations are particularly important for decreasing trunk-to-room oscillations that could excessively displace and thus destabilize the subject during locomotion. These data suggest that gravity's potentially destabilizing effects may be decreased by permitting relatively larger trunk-to-pelvis motions than trunk-to-room oscillations (Figs. 4 and 5B). For example, during the stance phase of normal gait, the pelvis and trunk shift toward the stance limb, and may approximate (abduct) on the swing-limb side, to assist foot clearance without necessitating excessive trunk and CM displacements. In gait, relatively little trunk abduction/adduction is required to effectively transfer weight and assist the swing leg. In stair locomotion, however, the greater lateral displacements may be needed to ensure that the trunk's substantial mass is securely over the supporting (stance) limb. Transverse-plane data support this contention. Medial/lateral rotation ROM relative to the room coordinates was greatest in stair negotiation (Fig. 5B), but transverse rotation ROM relative to the pelvis was relatively small because the pelvis moved in greater synchrony with the trunk in stair locomotion (Fig. 5A).

Because the chief task during rising from a chair is to transport the CM vertically, and only a small horizontal (forward) CM displacement is required,(4) it is not surprising that our subjects displayed little transverse-and frontal-plane trunk motions but large sagittal-plane displacements.

Comparison with Previous Results

Murray(8) found the range of thoracic transverse rotation for free-speed walking to be 6.9 degrees, compared with 9.0 degrees for our subjects. Stokes et al (12) reported the mean trunk range of abduction/adduction ROM in eight subjects during free-speed walking was 4.9 degrees, compared with 5.4 degrees for our subjects in the same movement. Stokes et al, however, provided no descriptive data other than gender for their subjects, and those data were collected during treadmill walking only. Thorstensson et al(12) reported the amplitude of apparently room-referenced trunk flexion/extension movements during treadmill gait of seven young subjects to be 2 to 12 degrees. Flexion/extension ROM in gait varied from 2 to 8 degrees relative to room coordinates and from 2.5 to 12.6 degrees relative to the pelvis in our study. Movement in all three planes of trunk ROM in Opila-Correia's study(10) exceeded that demonstrated by our subjects. Prior reports of trunk ROM during gait are presented in Table 3.

The data obtained in this study are within the range of the previously published results for gait, but no comparable data for stair or chair activities have been published, to our knowledge.

Further studies are needed (1) to investigate whether widely reported lower-limb phenomena such as increasing amplitude ROM with increasing Iocomotor velocity are also present in the trunk and (2) to investigate the effects of various diseases and impairments on trunk kinematics. Data on the effects of these angular rotations on translatory positions would complement the findings of previous studies and would be likely to further the understanding of whole-body stability (balance) control.

Clinical Considerations

We believe the primary utility of the data reported is in using these data to identify the relative demands of gait and activities of daily living on trunk mobility. Familiarity with trunk kinematics of healthy subjects should assist clinicians in designing therapeutic programs for patients with trunk and lower-limb impairments, such as those that occur after back surgery or following spinal cord injury with trunk weakness. Restoration of trunk mobility relative to room coordinates can be attained to some extent even in the absence of spinal mobility, through compensatory strategies. Restoration of normal trunk-to-pelvis kinematics, however, demands thoracolumbar spinal mobility. Hence, these data indicate that although lower-extremity gait kinematic studies will continue to be useful, full-body kinematic analyses during gait, stair climbing and descending, and arising from a chair are feasible and contribute unique information to locomotor assessments that solely lower-extremity studies cannot provide.

Summary

Because the trunk possesses the largest mass of any body segment and because we believe that computing resources continue to increase, whole-body assessments should be routinely included in "gait lab" analyses. Arising from a chair required the greatest trunk ROM in flexion/extension and the least trunk ROM in abduction/adduction and mediaL/lateral rotation. During gait and stair locomotion, trunk abduction/adduction ROM relative to the pelvis was significantly greater than that relative to room orientation, probably because of the downward tilting of the stanceside pelvis. Flexion/extension ROM was four times greater in rising from a chair than in stair activities, which in turn exceeded gait ROM. Because motions relative to the pelvis differ from room-referenced kinematics, observational locomotor assessments must carefully distinguish the two frames of reference. Each reference frame provides distinct information. The overall goals of locomotion probably include transporting the body with the least energy and the greatest stability, which are probably accomplished by optimizing the motions of the trunk relative to room coordinates and relative to the pelvis simultaneously.

These trunk ROM results during gait are consistent with previous findings, but they are the first three-dimensional kinematics to be reported on stair and chair activities. The functional trunk motions required by healthy subjects to perform activities of daily living should be considered in treating patients with 1ocomotor disability, and especially when surgical procedures limit ROM of any of the body's segments. These findings should be considered tentative, however, until verified in larger, representative population studies.

DE Krebs, PhD, PT, is Associate Professor, MGH Institute of Health Professions, 15 River St, Boston, MA 02108-3402 (USA). Address all correspondence to Dr Krebs.

D Wong, MD, PT, is Resident in Rehabilitation Medicine, Kingsbrook Jewish Medical Center, 585 Schenectady Ave, Brooklyn, NY 11203. He was an NIDRR Advanced Rehabilitation Fellow at the MGH/MIT Rehab Engineering Center when this work was conducted.

D Jevsevar, MD, is Resident in Orthopaedics, Tufts New England Medical Center, 750 Washington St, Boston, MA 02111. He was an NIDRR Advanced Rehabilitation Fellow at the MGH/MIT Rehab Engineering Center when this work was conducted.

PO Riley, PhD, is Technical Director, MGH Biomotion Laboratory, Massachusetts General Hospital, Boston, MA 02114.

WA Hodge, MD, is Assistant in Orthopaedics, Massachusetts General Hospital.

This study was approved by the Massachusetts General Hospital Institutional Review Board.

This work was supported in part by Grants HI33P90005 and HI33GOO025 from the National Institute of Disability and Rehabilitation Research.

This article was submitted June 12, 1991, and was accepted March 18, 1992.

References

1 Thorstensson A, Nilsson J, Carson H, Zomlefer MR. Trunk movements in human locomotion. Acta Physiol Scand. 1984;121:9-22.

2 Inman VT. Human locomotion. Can Med Assoc. J. 1966;94:1047-1054.

3 Saunders M, Inman VT, Eberhart HD. The major determinants in normal and pathological gait. J Bone Joint Surg [Am]. 1953;35: 543-558.

4 Pai Y-C, Rogers MW. Segmental contributions to total body momentum in sit-to-stand. Med Sci Sport Exerc. 1991;23:225-230.

5 Andriacchi TP, Andersson GBJ, Fermier RW, et al. A study of lower-limb mechanics during stair climbing. JBoneJoint Sung [Am]. 1980;62: 749-757.

6 Craik RL. Gait and aging. In: Woollacott M, Shumway-Cook A, eds. Posture and Gait Across the Life Span. Columbia, SC: University of South Carolina Press; 1989:176-201.

7 Murray MP, Drought AB, Kory RC. Walking patterns of normal men. J Bone Joint Sung IAm]. 1%4;46:335-360.

8 Murray MP. Gait as a total pattern of movement. AmJ Phys Med. 1%7;46:290-333.

9 Ikeda ER, Schenkman ML, Riley PO, Hodge WA. Influence of age on dynamics of rising from a chair. Phys Ther. 1991;71:473-481.

10 Opila-Correia KA. Kinematics of high-heeled gait. Arch Phys Med RehabiL 1990;71: 304-309.

11 Thorstensson A, Carlson H, Zomlefer MR, Nilsson J. Lumbar back muscle activity in relation to trunk movements during locomotion in man. Acta Pbysiol Scand. 1982;116:13-20.

12 Stokes VP, Andersson C, Forssberg H. Rotational and translational movement features of the pelvis and thorax during adult human 10comotion. J Biomech. 1989;22:43-50.

13 Strathy GM, Chao EY, Laughman RK. Changes in knee function associated with treadmill ambulation. J Biomech. 1983; 16: 517-522.

14 Antonsson EK, Mann RW. Automatic 6-D.O.F. kinematic trajectory acquisition and analysis. Journal of Dynamic Systems, Measurements, and Control 1989;111:31-39.

15 Riley PO, Fijan RS, Hodge WA, Mann RW. Determination of joint centers for posture studies. In: Stein JL, ed. The Biomechanics of Normal and Prosthetic Gait: BED.ASME4. New York, NY: American Society of Mechanical Engineers; 1987:131-136.

16 Tupling SJ, Peirrynowki MR. Use of Cardan angles to locate rigid bodies in three dimensional space. Med Biol Eng Comput. 1987;25: 527-532.

17 Riley PO, Mann RW, Hodge WA. Modelling of the biomechanics of posture and balance. J Biomech. 1990;23:503-506.

18 Thurston AJ. Spinal and pelvic kinematics in osteoarthrosis of the hip joint. Spine. 1985; 10:467-471.

19 Basmajian J.V, DeLuca CJ. Muscles Alive: Their Functions Revealed by Electromyograpby. Baltimore, Md: Williams & Wilkins Co; 1985:358, 384.