Foot force direction in an isometric pushing task: prediction by kinematic and musculoskeletal models

The abilities of a kinematic model and a muscle model of the human lower limb to predict the stereotyped direction of the muscular component of foot force produced by seated subjects in a static task were tested and compared. Human subjects ( n=11) performed a quasi-static, lower-limb pushing task against an instrumented bicycle pedal, free to rotate about its own axis, but with the crank fixed. Each pushing trial consisted of applying a force from the resting level to a force magnitude target with the right foot. Ten force target magnitudes were used (200, 250, ..., 650 N) along with 12 pedal positions. For each pushing effort, the muscular contribution to the measured foot force was determined from push onset to peak attained force. This segment was well characterized by a straight line across subjects, pedal positions, and force target magnitudes. The linear nature of the muscular component allowed a characteristic direction to be determined for each trial. A three-joint (hip, knee, and ankle) and a two-joint (hip and knee) net joint torque optimization was applied to a sagittal-plane kinematic model to predict the characteristic force direction. A musculoskeletal model was also used to create a feasible force space (FFS) for the lower limb. This FFS represents the range of possible forces the lower limb could theoretically produce. From this FFS, the direction of the maximum feasible foot force was determined and compared with the characteristic direction of subject performance. The muscle model proved to be the most effective in predicting subject force direction, followed by the three-joint and two-joint net joint torques optimizations. Similarities between the predictions of the kinematic and muscle model were also found.