The static relationship between elbow joint angle and torque at constant IEMGs was investigated with the aid of an artificial neural network technique. A normal volunteer was allowed to move his upper extremity freely on a horizontal plane, and was instructed to maintain his forearm posture at a target position against an external load applied to his wrist. Firstly, seven channel IEMGs, joint angles and torques around elbow and shoulder were measured. Secondly, the neural network model was constructed; inputs were IEMGs and joint angles, outputs were torques. Finally, elbow torque was estimated at various elbow angles with constant IEMGs by using the model. The torque of extensor muscles monotonically increased as flexion of the elbow joint. This characteristic was similar to that of monkey's triceps brachii. However the torque of flexor muscles increased and then decreased as extension of the elbow joint. This characteristic does not agree with the fundamental property of the equilibrium point hypothesis.
展开▼