The use of artificial neural networks in the motor program

Abstract

Though it is commonly assumed that the brain creates 'motor programs' which store the information essential to perform a motor skill, little direct evidence exists for such motor programs. Electromyography (EMG) provides a look into the motoneurons - level of a movement by measuring the electrical activity in relation to the muscle's involvement in the movement In this paper, artificial neural networks (ANNs) were applied to define the temporal patterns of EMG activity used by normal subjects in performing step-tracking tasks, and how such patterns change with practice. Our results demonstrate that ANNs could be trained to detect the input-output relationship between muscles' onset times and reaction times, and provided evidence to support the existence of a motor program.