THE EFFECT OF SEGMENT PARAMETER ERROR ON GAIT ANALYSIS

D J Pearsall (1), P A Costigan (2)
(1) Department of Physical Education, McGill University, Montréal, Québec, Canada, H2W 1S4
(2) School of Physical Health & Education, Queen's University, Kingston, Ontario, Canada, K7L 3N6

Presented at the 20th Annual Meeting of the American Society of Biomechanics
Atlanta, Georgia. October 17-19, 1996

INTRODUCTION

The extent to which errors in predicting body segment parameters(SP) affect the outcome of a biomechanical analysis of motion is not evident. The intent of this study was to address this question by quantitatively evaluating (1) the differences in SP estimates for a young, adult, male population based on predictive functions from various literature sources; and (2) the extent to which SP errors effect inverse dynamic analysis of walking motion.

REVIEW AND THEORY

Reducing systematic errors in motion analysis has focussed upon marker location, skin movement, and equipment accuracy; however, the significance of using generalized body SP has not been reviewed extensively. SP include measures of mass (MS), center of mass (CM) and moment of inertia (I) of body segments. Of the few quantitative reports, it has been noted that SP errors may be only as damaging as errors in acceleration data when estimating joint forces and moments1 while others have suggested that SP errors may be more serious2.

PROCEDURES

From a sample of 15 young, healthy males varying in body mass indices (18.8 to 27.3 cm/ kg2) SP estimates of the leg and thigh were calculated according to predictive functions found in the literature and include equations that use ratios, regressions and geometric descriptions3-5.

SP were varied to examine the effect of changing segment parameters on the kinetic output. The SP predict by Dempster's equations3 were taken as the baseline level. Each SP was varied in steps over nine levels by a defined percentage (-40% to +40%) of the original SP baseline value.

Walking kinematic and kinetic data were collected for each subject. The three-dimensional motion of the right lower limb was recorded at 50 Hz using the WATSMART motion tracking system (Northern Digital, Waterloo). The body model used (Figure 1) consisted of leg and thigh segments.

Figure 1. Lower limb model used for kinetic estimates of hip.

Inverse dynamic analysis was performed iteratively to compute hip forces and moments while simultaneously varying SP values over nine intervals within ± 40% of a baseline value. The output data were the forces and moments about the hip joint expressed in the thigh fixed-body coordinate system.

RESULTS

Substantial variation in SP were produced when various predictive formulae from the literature were used. In some cases, the differences were greater than 40% for MS and I values. A Duncan's multiple range test (p<0.05) was used to contrast the means of SP predicted. Significant differences were found for the MS, CM, and I segment parameters for both the leg and thigh.

Using repeated measures ANOVA, SP variations were found to significantly affect (p<0.05) most of the kinetic estimates, particularly during the swing phase. However, the absolute value of these differences were small, generally less than 1% of body weight. The one exception was the force in the distal-proximal direction where SP variations altered the force measures by up to 57% during stance and 62% during swing. (Figure 2)

DISCUSSION

Though nearly half of the kinetic measures were statistically affected by SP variation most effects were small. The SP influence may have been diminished by several factors including (1) unchanged kinematic and ground reaction force patterns, (2) the lack of interaction of SP used in the body model, and (3) the net mass of the thigh or the leg segments as compared to the whole body.

The small changes in the kinetic measures does not suggest that SP are unimportant in all human motion studies. In closed-chain situations, such as during gait stance, the limb accelerations are low when compared to the maximum achievable; thus the effect of SP are minimal. However, the effect of the SP would increase with skills that involve larger limb accelerations, such as running, and with open-chain skills that do not have large external loads, such as throwing. In addition, forward simulation of activities such as those just described would be more dependent on SP of the body model. The extent of this dependence is an issue for further investigation.

Figure 2. Hip forces during gait cycle as affected by changes in thigh mass.

REFERENCES

1. Davis B L. Uncertainty in calculating joint moments during gait. Proc 8th Euro Soc Biomech Conf, Rome, Italy, 276, 1992.

2. Wu G, Ladin Z. Proc 2nd Int Sympos on 3D Analysis of Human Movement. , 106-108, 1993.

3. Dempster W T. Space requirements for the seated operator. WADC (TR-55-159), 1955.

4. Clauser C E, McConville J T, Young J W. Weight, volume and center of mass of segments of the human body. AMRL (TR-69-70) 1969.

5. Zatsiorsky V, Seluyanov V. The mass and inerital characteristics of main segments of the human body. Biomechanics VIII-B. Matsui H, Kobayashi K. editors, 1152-1159, 1983.

ACKNOWLEDGMENTS

The authors would like to thank Dr. Urs Wyss for the use of the data from the Clinical Mechanics Group gait analysis laboratory (Queen's University) and the use of the QGAIT motion analysis software.