The segment principal moments of inertia of a sample of 7 elderly males and 12 elderly females were estimated using a model based on stacked elliptical cylinders at 2-cm intervals. Apart from the thigh, all male parameters were larger than female parameters. The largest differences were for the lower trunk and hand and for the forearm. The inertia parameters of the thigh for the males were about 12% smaller than the females. Nonlinear estimations of segment principal moments were then determined. The effect of the differences was tested by cross validating cadaver results against the elliptical model results. The regressions were then cross validated using an independent sample of 6 subjects. The standard errors of fit given as a percentage of the mean, Sf, were smaller than the cross validation results for the cadaver regressions and the differences were attributed to differences between cadavers and living subjects.
Robert K. Jensen and Paula Fletcher
Robert K. Jensen, Tina Treitz and Han Sun
The purpose of the study was to use the elliptical cylinder model adapted for infants (Sun & Jensen, 1994) with a cross-sectional sample to select appropriate multiple linear regression equations for predicting masses and nonlinear regression equations for predicting principal moments of inertia (Yeadon & Morlock, 1989). The linear and nonlinear predictions were evaluated with an independent cross-validation sample of infants and a sample where inertias ranged below and above the cross-sectional sample. The cross-validation for masses was compared to a cross-validation of four linear regressions for masses developed by Schneider and Zernicke (1992). It is recommended that the linear regression equations developed in this study be used to predict infant segment masses. It is also recommended that the nonlinear regression equations developed in this study be used to predict the principal moments of inertia of all infant segments, other than head Ix and lower trunk Ix and Iy.
Robert K. Jensen, Tina Treitz and Sylvie Doucet
The purpose of this study was to develop prediction equations to estimate mass, radius to the center of mass (CM), and principal moments of the segments during pregnancy. Nonlinear regression equations were determined for the lower trunk, upper trunk, and thigh. The third sampling month of a longitudinal study was used (Sample 1, n = 15). The nonlinear regressions were then used to predict segment inertias above and below the third sampling month (Sample 2, the remaining 74 measurements). For the remaining segments, body mass and segment lengths were used as predictor variables for mass, radius to CM, and radius of gyration about the centroidal axes. The remaining seven segments did not change substantially during pregnancy, and the means of the repeated measures were used for the simple linear regressions. Eighteen of the 28 regressions and all of the CM regressions were significant. With pregnant subjects it is recommended that these regressions be used if application of the elliptical cylinder model is not possible.
Ross H. Sanders, Barry D. Wilson and Robert K. Jensen
This study investigated whether force data could be derived accurately using segment inertia data determined by the elliptical zone method (Jensen, 1976), automatic digitizing from high-speed video using a Motion Analysis VP110 system, and for an activity that does not require flexion of the thorax. The criterion fonctions were the force-time records of the jumps recorded at 500 Hz by a Kistler 9281B force platform. A second-order Butterworth digital filter was used to smooth the derived data, with frequency cutoffs being selected on the basis of root mean square error of the smoothed function with respect to the criterion force function. In a second procedure, the criterion function was the directly measured force-time record after filtering with a second-order Butterworth digital filter at 5 Hz to remove the high frequency part of the force signal. The closeness of fit of the derived data to the low frequency part of the criterion force was then assessed. It was concluded that, using the techniques described, the low frequency components of the ground reaction forces of drop jumps could be derived accurately.