Are torque-driven simulation models limited by an assumption of monoarticularity?

Subject-specific forward-dynamics simulation models of human movement incorporating torque generators have typically represented the torque at a joint based on the kinematics of only that joint. Representing the net torque exerted at a joint by considering the kinematics of two joints ought to better represent the torques exerted by the two-joint biarticular muscles. In this study two-joint torque generator representations including monoarticular and biarticular torque components and single-joint torque generator representation of the lower limb were determined for a single male participant using strength measurements collected on an isovelocity dynamometer. For plantar flexion, knee flexion, knee extension torques, a two-joint representation was found to offer better agreement with torques measured on an isovelocity dynamometer than a single joint representation when the joint angle of a secondary joint was changed by 37˚ or more. In addition a two-joint representation should account for the biarticular knee flexor and extensor contributions to hip joint torques. Two subject-specific simulation models of squat jumping, one with single-joint torque generators and another with two-joint torque generators were created to establish the necessity for two-joint torque representation in whole-body simulation models of human movement. Including the two-joint torque generators was shown to provide better agreement with a participant performance than a single-joint torque-driven simulation model. The simulation of a maximal effort human movement should include two-joint torque representations when both of the following characteristics are present in the task to be simulated: a) the task does not have a large initial whole-body momentum; b) the task involves multiple joint kinematics with a large biarticular muscle contribution. A single-joint torque representation was effective when the kinematics of the secondary joint were similar to those which were present during torque measurements. Therefore if the simulated task involves similar kinematics at both joints then a single-joint representation may be sufficient.