Muscle torque-total torque relationships at the shoulder and elbow: Rules for initiating multijoint arm movements

Abstract

One concept central to theories of multijoint control concerns the selection of muscles for the appropriate joint motion. For multijoint movements, a given muscle torque at an individual joint can lead to flexion, extension, or very little motion, since mechanical effects coming from other segments interact with muscle torque. This study quantified the contribution of muscle torque to initial joint motion for horizontal arm movements throughout the workspace. Previous studies of arm mechanics have been limited to a few movements or have focused on one joint. In contrast, this study reports data for both the shoulder and elbow joints. Moreover, a large number of movements were used for which direction, excursion, and distance were manipulated. Using high speed video recordings and techniques of inverse dynamics, a ratio of muscle torque to total torque was computed for each movement as a measure of contribution of muscle torque to joint acceleration. One consistent finding was that the muscle torque contribution consistently differed between the shoulder and elbow for most of the workspace. At one joint, muscle torque directly contributed to acceleration with negligible interaction torque ('direct' muscle torque contribution), thus the joint appeared to act as the launcher of the arm. At the other joint, both muscle and interaction torques contributed to joint acceleration ('complex' contribution), thus the joint appeared to be responding to mechanical effects from motion of the launcher. This contrast between joints may provide a simplifying feature for multijoint arm control. Specifically, only one of the two joints has complex mechanics, while the other joint, surprisingly, has simplified mechanics similar to a single joint in isolation. Movements in this study also demonstrated a three fold covariance (muscle torque contribution, movement direction, and the relative excursions of the shoulder and elbow) regardless of distance. A covariance of movement features, historically viewed as a confound, may provide a further simplification for arm control by reducing the unknowns; namely, the muscle torque contribution is associated with a resultant direction and joint excursions, or a direction or set of excursions is achieved by the associated muscle torque contribution.