Application of an Entropy-Assisted Optimization Model in Prediction of Agonist and Antagonist Muscle Forces

Abstract

Many existing optimization based biomechanical models fail to predict antagonist muscle activity. Some optimization models predict such a cocontraction, but either lack a compelling physiological basis or are computationally formidable. The current study takes advantage of the flexible definition of entropy as a scientific measure, and utilizes it in the objective function of an optimization formulation to construct a new optimization model for predicting agonist and antagonist muscle forces. In this model, the objective function of a nonlinear program consists of a weighted sum of two components: a linear or nonlinear term favoring agonist muscle exertions (reciprocal inhibition), and the entropy term enforcing cocontraction. The concept of the current optimization model is based on recent findings in neurophysiology that there exist two separate central nervous systems for generation of two motor patterns: agonist contraction and agonist-antagonist cocontraction.