Retrieving motor unit depth using inverse approach on HD-sEMG signals

Abstract

The position identification of underlying activated motor units from the surface potential map that those motor units had produced constitutes a challenging inverse problem in the electromyography community. This field has gained wide interest due to the various medical applications that it enhances. Some of the most important medical applications are focused on the areas of prosthetic control enhancement and improving the efficiency of rehabilitation following an impairment. The proposed study includes the testing of an inverse problem methodology on electromyography data simulated on a vertical alignment of motor units (depth varying) using minimum norm estimation. This study also proposes a methodology for decreasing the number of sources needed in a simulation by formulating a fitting equation, which is formed by a limited number of sources of a definite configuration. The obtained results are promising and demonstrate the usefulness of such approaches.