Influence of age on dynamics of rising from a chair.

The ability to rise unassisted from a chair is a prerequisite to many activities of daily living. In teaching patients this activity, therapists have previously relied on qualitative conceptual models and clinical experience. We believe that without a quantitative description of what constitutes normal rising, there can be no demonstration of the efficacy of using clinical models.

Because of the complexity of testing and analyzing the task of rising, several studies have focused on specific aspects of the sit-to-stand movement. Several investigators (1-4) examined only the kinematics of rising, and others (5-14) studied only the lower-extremity dynamics. Consequently, there is a need for greater knowledge concerning integrated upper-body and lower-extremity kinematics and kinetics. Describing total-body kinematics of geriatric subjects in rising from a chair is necessary to further delineate differences between young and elderly subjects and provides data for future comparisons with kinematic data of elderly persons with pathology and dysfunction.

Literature Review

The use of different instruments and protocols limits the ability to compare results among studies. For example, chair height, speed of rising, initial body position, and use of the arms are all components of protocols that have been shown to affect joint torques, angles, and velocities. (6,12,13,15-20) These factors influence the dynamics of movement of the whole body, but few researchers have controlled all these elements in their study design. (1,15-20) Riley et al (17) and Schenkman et al (20) have provided a detailed description of total-body kinematics and kinetics of the sit-to-stand movement, with a controlled protocol, in a group of healthy young individuals.

The relationship of age to changes in functional tasks, such as rising from a chair, has not been clearly described in the literature. Significant data exist to confirm the reduction of function, independent of pathology, in many body systems as a result of the aging process. Several authors (21-23) have demonstrated decreases in range of motion (ROM) in elderly individuals. Variations in muscle histology, morphology, and performance have been demonstrated with aging, in addition to decreases in the conduction velocities of peripheral nerves. (24,25) Whether or how these changes in mobility and morphology translate to changes in functional activity, such as rising from a chair, has not been fully studied. Previous investigations of gait (26,27) have demonstrated characteristic alterations that occur with aging; thus, it is reasonable to question whether strategies of rising from a chair also differ.

The sit-to-stand movement requires total-body coordination, and there is potential for much variation and compensation within multiple body segments. For example, Fleckenstein et al (9) reported changes in hip flexion angles and hip extension peak moment with externally imposed restriction of knee flexion in rising from a chair. Berger et al (15) reported changes in upper-body flexion velocity in rising when initial foot position was altered. Changes in sit-to-stand kinematics and kinetics may also be anticipated in older persons with age-related restrictions in motion or position.

Previous studies have examined the effect of age on aspects of rising from a chair. Munton et al (28) conducted a survey and ascertained that 42% of 379 elderly individuals with and without arthritis reported difficulty getting out of their easy chairs. Wheeler et al (14) used a videotape system, electrogoniometry, and electromyography (EMG) to study the effect of age on standing from a chair. An elderly group (mean age=75 years, n=10) demonstrated significant differences in foot placement, knee flexion, and trunk forward lean and increased EMG activity of the vastus lateralis muscle. These findings were interpreted to indicate that the elderly group had more difficulty rising from the chair as compared with a younger group (mean age=24 years, n=10). (14) Yoshida et al (4) compared groups of hemiplegic subjects (mean age=60.7 years), young healthy men (mean age=27.9 years, n=10) and women (mean age=24.3 years, n=10), and elderly healthy men (mean age=67.5 years, n=10) and women (mean age=60.0 years, n=10). They found that the elderly subjects required more time to stand and to stabilize sway than did the younger subjects, but needed less time than the hemiplegic subjects. These studies used protocols that did not control for the chair height, the speed of rising, or the foot position, and the investigators did not analyze whole-body movements.

The purposes of this study were (1) to describe the whole-body (11-segment) sagittal-plane kinematics and lower-extremity kinetics of rising from a chair in a healthy geriatric sample and (2) to compare the geriatric group's data with data previously collected from healthy young subjects. We expected to find significant differences in head and trunk maximal angles and total ROM.

Method

Subjects

A sample of convenience, composed of 9 healthy elderly individuals (6 men, 3 women), participated with informed consent. The subjects were 61 to 74 years old ([MATHEMATICAL EXPRESSION OMITTED]=66.9, SD=4.5). They reported no diagnosed musculoskeletal disorders, and all subjects were independent in home and community activities. Data collected previously from a second sample of convenience, composed of 9 healthy young women aged 25 to 36 years ([MATHEMATICAL EXPRESSION OMITTED]=28.9, SD=3.4), were used for comparison. (20)

Instrumentation

Instrumentation consisted of four Selspot[TM] II optoelectronic cameras, (*) infrared light-emitting diodes (LEDs) embedded in arrays, two Kistler model 9281B piezoelectric force plates, (+1) TRACK[C] and NEWTON[C] software, (+2) a PDP 11/60 minicomputer, (*2) a MicroVax II work station, (*2) and an adjustable chair without arms or a backrest.

Arrays of LEDs were attached bilaterally to the feet, legs, thighs, pelvis, trunk, and arms and to the right side of the head (Fig. 1). The LED arrays were anchored securely by polypropylene molds and straps to reduce extraneous movement. Kinematic data were derived from tracking these LEDs using the Selspot[TM] system at a data-collection rate of 150 Hz.

The TRACK[C] (kinematic data acquisition and processing) software determined the position of the LED arrays with the accuracy of [+ or -] 1 mm. (29) Orientation accuracy for this system is within 1 degree. (29,30) This software system produced a three-dimensional analysis (6 [degrees] of freedom) of segmental translations and joint rotations. (29,30) This method of analysis assumed rigid body segments, and the joint angles were calculated using Cardan angles, as described by Tupling and Pierrynowski (31) andRiley and colleagues. (32)

The ground reaction forces were measured by the two Kistler force plates, which assessed anterior-posterior, lateral, and vertical forces in addition to the center of pressure. The NEWTON[C] software package (30) applied inverse Newtonian dynamics to the force-plate and kinematic data to determine net joint forces and torques.

Procedure

The protocol for this study was developed and used by Berger et al (16) and Schenkman et al (20) in previous studies of young subjects. A similar protocol was used by Jeng et al, (1) who demonstrated good test-retest reliability using a videotape system (intraclass correlation coefficients generally greater than .71). Schenkman and colleagues found similar values of performance repeatability using the data of this study, which were obtained with a protocol similar to that of Jeng et al (ML Schenkman, S-F Jeng, ER Ikeda, PO Riley, WA Hodge; unpublished research).

The subjects, barefoot and dressed in shorts, were seated on an armless, backless chair, which was adjusted to 80% of each subject's knee height (floor to lateral knee joint space). The elderly subjects' height ranged from 152 to 185 cm ([Mathematical expression omitted]=170.7), and the younger subjects' height ranged from 152 to 175 cm ([MATHEMATICAL EXPRESSION OMITTED]=160.9). Average chair height was 38.6 cm (SD=2.1). The participants were positioned near the front of the seat with their thighs and feet parallel. Each subject's ankles were placed in 18 degrees of dorsi-flexion. One foot was placed on each force plate, with the medial border of the feet 10.2 cm apart. Subjects were instructed to fold their arms across their chest and not to use their upper extremities to assist with the standing maneuver. This instruction was designed to minimize intersubject differences in use of the arms to contribute to upper-body momentum.

The movement was performed in synchrony with a metronome set at 52 beats per minute (1.15 seconds to complete the task). One of the investigators (MLS) gave the command "ready, set, start, stand" in time with the metronome. The task was initiated with the word "start" and completed (erect stance) at the word "stand."

The subjects were allowed to practice several times, until the investigators were assured that the movement was coordinated without excessive effort and accomplished within the prescribed time. Data were collected for twotrials.

Data Reduction

and Analysis

An approach to this descriptive analysis of the sit-to-stand movement was delineated in detail by Schenkman et al. (20) Four phases were established based on observations of kinematics, stability, and momentum. Maximum angles, torques, and velocities were temporally related to the time of buttocks lift-off (initial shift of weight bearing from buttocks to the feet, as detected by the force plate).

Phase 1, the flexion-momentum phase, began with initiation of hip or trunk flexion velocity and ended at lift-off of the buttocks. Phase 2, the momentum-transfer phase, began with lift-off of the buttocks and ended with maximal ankle dorsiflexion. Maximum dorsiflexion was the final kinematic event, which reached full flexion in phase 2 in the young group. (20) Phase 3, the extension phase, started with maximal dorsiflexion and ended with terminal hip extension (when hip extension velocity became zero). Phase 4, the stabilization phase, started with terminal hip extension and theoretically ended when all motion associated with the task was completed. This fourth phase was not addressed in this study because of the complexity of determining return to normal sway in individuals.

Although the Selspot[TM]/TRACK[C] system provided three-dimensional data, the sagittal-plane kinematics and kinetics were chosen for this investigation. The average values of two trials were used for all data analyses; averaged missing values were excluded from the data. At least eight values from each group were available for analysis, with the exception of head extension velocity (with four missing values). The maximum angles were derived from adjacent segments as well as the relationship of the trunk and head to the ground (horizontal). The duration of the three phases as well as the occurrence and order of key events (maximum angles, torques, and velocities) within each phase were evaluated for both groups. The percentages of difference of segmental velocities and torques between groups were calculated by dividing the mean difference by the mean of the older sample. Right- to left-side percentages of difference were determined by dividing the mean difference by the right-side value for velocities and torques. Independent t tests were conducted on maximum joint angles and total joint excursions.

Results

Phases

The sit-to-stand movement was divided into the four phases described previously by Schenkman et al (20) for a group of young healthy women (ie, flexion momentum, momentum transfer, extension, and stabilization). The first three phases were used for this analysis.

Phase 1, the flexion-momentum phase, was defined by the start of movement to initial weight shift from buttocks to feet (Fig. 2). This phase was characterized by the occurrence of maximum velocities of hip flexion, trunk flexion, and head-to-trunk extension in all of the young subjects (Fig. 3). These maximum velocities also occurred in this phase in the older group, with only a few exceptions. In two of the nine older adults, the maximum trunk flexion velocity occurred after lift-off of the buttocks (0.06 and 0.07 second after the end of this phase, respectively). Maximum hip flexion velocity was also attained after lift-off of the buttocks in two of the nine elderly subjects (0.02 and 0.03 second after lift-off of the buttocks, respectively).

The parameters of phase 2, the momentum-transfer phase, were defined as lift-off of the buttocks to achievement of maximum ankle dorsiflexion (Fig. 2). This phase was the most dynamic, with the body's center of mass moving off the chair and forward over the feet, the new base of support. The trunk position began to change from flexion toward extension. Phase 2 data were consistent for eight of the nine young participants, maintaining the maximal-angles sequences of (1) hip flexion, (2) trunk-to-ground flexion, (3) head-to-trunk extension, and (4) dorsiflexion. For the ninth young subject, maximal head extension was delayed by 0.12 second. The older group showed variation, with maximum trunk-to-ground flexion occurring after maximum dorsiflexion in two of nine individuals (0.05 and 0.2 second after maximum ankle dorsiflexion, respectively). Maximal head-to-trunk extension also developed late in five of nine elderly participants (0.04, 0.12, 0.13, 0.19, 0.22 second after maximal ankle dorsiflexion, respectively). In phase 2, maximum knee and hip torques were attained by both groups.

Timing of these components was closely related, within eight frames (0.05 second) in both groups, with the exception of one older individual (0.16 second). The timing of these maximum torques coincided with maximal hip flexion in full weight-bearing. All flexion velocities for both groups decelerated, and trunk extension velocity was initiated as the body's center of mass began to move vertically during this time period.

Phase 3, the extension phase, began with maximum ankle dorsiflexion and ended at the point at which velocity of hip extension reached zero (Fig. 2). This phase was characterized by vertical movement of the body to full stance. Full knee extension was reached in this phase for most participants, with the exceptions of one older subject and one younger subject (0.01 and 0.56 second late, respectively). Maximum velocities were attained for trunk extension, hip extension, and knee extension in both groups (Fig. 3). The sequence of these peak velocities varied considerably, with the most frequently observed order of hip, trunk, and knee for the young group (four of nine subjects) and knee, hip, and trunk for the older group (six of nine subjects). Flexion velocity of the head in relation to the trunk reached a peak during this phase for both groups, but the timing was very inconsistent (Fig. 3).

We agreed, based on our observations, that all subjects appeared to complete the activity in the allotted time of 1.15 seconds (metronome set at 52 beats per minute). The average time from the start of movement to the end of movement (designated as the point where hip extension velocity reached zero) for the older group was 1.95 seconds (SD=0.03), whereas the average time for the young sample was 1.86 seconds (SD=0.19). A comparison of the average timing for the individual phases is depicted in Figure 4.

Right-Left Differences

There were minimal right-left differences in the older group, and these values were comparable to the measured differences in the young group. Mean differences for maximum ankle and hip joint angles in the older group were 1% and 2%, respectively (Tab. 1). Right-left differences for total segment excursion ranged from 0.9% at the knee to 8% at the ankle (Tab. 1). Torque value differences from right to left for the knee and hip were 11% and 13%, respectively (Tab. 2).

Joint Angles

Independent t tests showed that maximum angles of trunk-to-pelvis flexion and head-to-trunk extension were significantly less in the older group than in the younger group (P<.05). The head-to-ground flexion angle was significantly greater in the older group than in the younger group (P<.05). The peak ankle, hip, and trunk-to-ground flexion did not differ between the two groups (Tab. 3).

The total excursion for head-to-trunk extension showed a significant difference between groups (P<.05), but the total head excursion relative to the ground did not vary significantly between the young and old groups. There was no statistically significant difference for other joint and segmental excursions between groups (Tab. 4).

Torques and Velocities

The torques were standardized by dividing by the subject's weight, multiplying by the subject's height, and then multiplying by 100. The percentages of difference (the mean differences divided by the younger group's values) between groups of maximum knee and hip torques, standardized to height and weight, were 7% and 6%, respectively (Tab. 5). The average peak velocities for all joints and body segments were within 7 [degrees]/s for the two groups. The standard deviations were greater in the elderly group than in the younger group. The range of percentages of difference for all velocities was between 0.9% for head extension velocity and 9% for trunk extension velocity (Tab. 6).

Discussion

Phases

The results of this study demonstrated that the performance of the older group of subjects was very comparable to that of the young group. The timing and extent of maximal joint and segmental angles, velocities, and torques were very similar between the two groups. Analysis of the older group's data also demonstrated the consistency of the phases delineated by Schenkman et al (20) in the young group of subjects. Dividing the task into flexion and extension phases, as was done previously, (3,12) does not give adequate information about the most physically challenging period of the sit-to-stand movement, that is, when the center of mass is initially far from the base of support (phase 2). The use of phases may help the clinician

[TABULAR DATA OMITTED]

decide whether a patient's inability to rise from a chair originates from inability to generate sufficient momentum (phase 1), inability to control balance when the center of mass is far from the base of support (phase 2), or ineffective extensor musculature (phase 3).

[TABULAR DATA OMITTED] [TABULAR DATA OMITTED]

Phase 2 was the most consistent for the younger group, but the older group had more variation in trunk and head position. This phase was also of shorter duration for the older group than for the younger group. Further investigation is under way to characterize the subtle differences in the dynamics of the sit-to-stand movement that account for the shortened duration of phase 2 observed in the older group.

The velocities showed little differences between the groups or between the right and left sides. This finding establishes that older persons are capable of performing this task similarly to younger persons when timed by a metronome. An uncontrolled protocol would be needed if differences in an individual's preferred speed of rising and timing of body movements was to be investigated.

Lower-extremity torques, when standardized to height and weight, were very similar for both groups. This finding is important, given that there was a 10-cm difference in height between groups, primarily attributed to gender differences. These results support the effectiveness of this protocol and analysis in providing consistency within and between groups, even when the group's gender composition is dissimilar. These results should be useful for torque comparisons of individuals with pathological conditions.

Overall, there were few kinematic differences between the younger group and older group. As predicted, the older group achieved less trunk-to-pelvis flexion than did the younger group, although no significant differences in trunk-to-ground flexion or total range of these segments were found. These findings suggest that the older individuals were seated in more trunk-to-pelvis flexion (in a posterior pelvic tilt) and thus the maximal angles reached were lower. A posterior pelvic tilt may have shiftfed the center of mass further posterior and thus increased the difficulty in this task. This conjecture was not, however, supported by evidence of other kinematic or kinetic changes.

Except for the study by Schenkman et al, (20) previous studies (1-3) have examined the head position with respect to horizontal or vertical. In our study, we examined the head position in relationship to the trunk as well as to the ground (horizontal). Other researchers (2,33) have hypothesized that the movement of the head with respect to the ground in the sit-to-stand movement is for generating momentum and "leading" the movement. We found variability in the sequencing of head angles and velocities in both groups in the first and third phases; thus, it was difficult to conclude that the head "led" any movement. A more likely possibility is that the head reacted to caudal movements and adjusted for balance maintenance, as

[TABULAR DATA OMITTED]

[TABULAR DATA OMITTED]

[TABULAR DATA OMITTED]

was seen in the sequence of the maximum head-to-trunk extension in phase 2.

The significant decreases of head-to-trunk maximum angle and total excursion and the significant increase in head-to-ground maximum angle in the older group may indicate a critical change in this sample. Because the head did not extend relative to the flexing trunk, the older group was facing down during phase 2, the momentum-transfer phase. In contrast, the young group maintained a fairly stable orientation of the head to the ground. It was unclear whether this change was related to loss of ROM (inability to extend), to a change in movement strategy (willful flexion), or to other factors. This position change may influence cervical proprioception and visual input, thus affecting the balance of the older subjects. We did not visually detect impairments in balance of the older group, nor did they report difficulty maintaining balance. Future analysis of phase 4 (stabilization phase, as described by Schenkman et al (20)), however, may be indicated to provide more detailed information about strategies and timing to regain normal standing balance in both populations.

Clinical Implications

Clinicians should be aware that few differences have been demonstrated in strategies of rising from a chair between young and elderly groups of subjects. We believe that problems with the sit-to-stand movement that are identified in older patients may be attributed to impairments in the musculoskeletal or neurologic system. A significant difference in head position between groups was found during this movement, which could potentially have a negative effect on balance. Additional loss of compensatory equilibrium mechanisms (eg, vision) in older patients may result in severe balance impairments.

Conclusion

This investigfation identified few differences between a group of healthy older individuals and a group of healthy younger individuals in the task of rising from a chair using a controlled protocol. The timing and order of key kinematic events were similar between groups. Significant differences in head position demonstrated functional movement changes in the older subjects and may have clinical implications for balance impairment in this population.

This study provides preliminary data about healthy older subjects that can be used in future investigations of older individuals with neuromuscular impairments. Future studies should include greater numbers of subjects, consider possible gender and anthropometric differences, and examine non-sagittal-plane angles during standing, because differences in these variables were observed by the investigators during this study. In addition, exploration of the relationships of musculoskeletal joint ROM to the kinematics of this task is recommended.

Acknowledgements

We thank Richard L Gajdosik, PhD, PT, and Robert W Mann, ScD, for their useful suggestions during the preparation of this article.

References

[1] Jeng S-F, Schenkman ML, Riley PO, Lin S-J. Reliability of a clinical kinematic assessment of the sit-to-stand movement. Phys Ther. 1990; 70:511-520.

[2] Jones FP, Hansen JA, Miller JF, Bossom. J. Quantitative analysis of abnormal movement: the sit-to-stand pattern. Am J Phys Med. 1963; 42:208-218.

[3] Nuzik S, Lamb RL, VanSant AF, Hirt S. Sit-to-stand movement patterns: a kinematic study. Phys Ther. 1986;66:1708-1713.

[4] Yoshida K, Iwakura H, Inoue F. Motion analysis in the movement of standing up from and sitting down on a chair. Scand J Rehabil Med. 1983;15:133-140.

[5] Bajd T, Kralj A. Standing up of a healthy subject and a paraplegic patient. J Biomech. 1982;15:1-10.

[6] Burdett RG, Habasevich R, Pisciotta J, Simon SR. Biomechanical comparison of rising from two types of chairs. Phys Ther. 1985;65:1177-1183.

[7] Ellis M, Seedhom BB, Wright V. Forces in the knee joint whilst rising from a seated position. J Biomed Eng. 1984;6:113-120.

[8] Ellis M, Seedhom BB, Amis AA, et al. Forces in the knee joint whilst rising from normal and motorized chairs. Eng Med. 1979;8:33-40.

[9] Fleckenstein SJ, Kirby RL, MacLeod DA. Effect of limited knee flexion range on peak hip moments of forces while transferring from sitting to standing. J Biomech. 1988;;21:915-918.

[10] Kelley DI, Dainis A, Wood GK. Mechanics and muscular dynamics of rising from a seated position. In: Komi PV, ed. Biomechanics V-B. Baltimore, Md: University Park Press; 1976; 127-134.

[11] Munton JS, Ellis MI, Wright V. Use of electromyography to study leg muscle activity in patients with arthritis and in normal subjects during rising from a chair. Ann Rheum Dis. 1984;43:63-65.

[12] Rodosky MW, Andriachhi TP, Andersson GBJ. The influence of chair height on lower limb mechanics during rising. J Orthop Res. 1989;7:266-271.

[13] Seedhom BB, Terayama K. Knee forces during the activity of getting out of a chair with and without the aid of arms. Biomed Eng. 1976;11:278-282.

[14] Wheeler J, Woodward C, Ucovich RL, et al. Rising from a chair: influence of age and chair design. Phys Ther. 1985;65:22-26.

[15] Berger RA, Riley PO, Mann RW, Hodge WA. Total body dynamics in rising from a chair. Presented at the Thirty-Fourth Annual Meeting of the Orthopaedic Research Society; 1988; Atlanta, Ga.

[16] Berger RA, Schenkman ML, Riley PO, Hodge WA. The chair model and its advantages for quantifying human functional performance. Orthopedic Transactions. 1989;13:275.

[17] Riley PO, Schenkman ML, Mann RW, et al. Mechanics of a constrained chair rise. J Biomech. In press.

[18] Schenkman ML, Berger RA, Riley PO, Hodge WA. Kinematics of rising to standing from sitting. Phys Ther. 1989;69:405. Abstract.

[19] Schenkman ML, Berger RA, Riley PO, Hodge WA. Normal total-body dynamics during rising from a chair. Presented at the annual meeting of the Biomechanical Society; August 1989; Burlington, Vt.

[20] Schenkman ML, Berger RA, Riley PO, et al. Whole-body movements during rising to standing from sitting. Phys Ther. 1990;70:638-651.

[21] Allander E, Bjornsson OJ, Olafsson O, et al. Normal range of joint movements in shoulder, hip, wrist, and thumb with a special reference to side. Int J Epidemiol. 1974;3:253-261.

[22] Boone DC, Azen SP. Normal range of motion of joints in male subjects. J Bone Joint Surg [Am]. 1979;61:756-759.

[23] Walker JM, Sue D, Miles-Elkousy N, et al. Active mobility of the extremities in older subjects. Phys Ther. 1984;64:919-923.

[24] Larsson L, Karlsson J. Isometric and dynamic endurance as a function of age and skeletal muscle characteristics. Acta Physiol Scand. 1978;104:129-136.

[25] Rowe JW, Besdine RW, eds. Geriatric Medicine. 2nd ed. Boston, Mass: Little, Brown & Co Inc; 1988.

[26] Hageman PA, Blanke DJ. Comparison of gait of young women and elderly women. Phys Ther. 1986;66:1382-1387.

[27] Imms FJ, Edholm OG. Studies of gait and mobility in the elderly. Age Ageing. 1981; 10:147-156.

[28] Munton JS, Ellis MI, Chamberlain MA, Wright V. An investigation into the problems of easy chairs used by the arthritic and the elderly. Rheumatol Rehabil. 1981;20:164-173.

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

[30] Riley PO, Mann RW, Hodge WA. Modeling the biomechanics of posture and balance. J Biomech. 1990;23:503-506.

[31] Tupling SJ, Pierrynowski MR. Use of Cardan angles to locate rigit bodies in three-dimensional space. Med Biol Eng Comput. 1987;25:527-532.

[32] 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.

[33] Jones FP, Gray FI, Hansen JA, O'Connell DN. An experimental study of the effect of head balance on patterns of posture and movement in man. J Psychol. 1959;47:247-258.

(*) Selective Electronics, Partille, Sweden.

(+) Kistler Instruments AG, Winterthur, Switzerland.

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

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

ER Ikeda, MS, PT, is Assistant Professor, Physical Therapy Department, University of Montana, 026 McGill Hall, Missoula, MT 59812-1076 (USA). At the time of this study, she was a graduate student at MGH Institute of Health Professions, Boston, MA. Address all correspondence to Ms Ikeda.

ML Schenkman, PhD, PT, is Associate Professor, Program in Physical Therapy, MGH Institute of Health Professions, 15 River St, Boston, MA 02108-3402, and a Fellow in Mechanical Engineering Newman Laboratory, Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02138.

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

WA Hodge, MD, is Assistant in Orthopaedics, Massachusetts General Hospital, 5 Lonfellow Pl, Ste 201, Boston, MA 02114.

This study was supported in part by Grant No. H133E0024-89 from the National Institute of Disability and Rehabilitation Research, US Department of Education.

This study was approved by the Subcommittee on Human Studies, Massachusetts General Hospital.

This article was submitted March 16, 1990, and was accepted January 28, 1991.