Evaluation of motor unit tracking across hand movements by combining high-density surface electromyography with...

Motor unit (MU) activities provide the primary output from the central nervous system to muscles, and have found wide applications in neurophysiological investigation and human-machine interface. However, the current decoding methods usually struggle with MU tracking across movements. A common compromise is comparing motor unit action potentials extracted from electromyography (EMG) signals, but this only considers electrical responses and requires further validation. This study aims to simultaneously collect EMG and ultrasound (US) signals, extracting both electrical and mechanical responses of MU discharges, to validate their feasibility for MU tracking across movements.

Ten able-bodied subjects participated in experiments involving isometric contraction of six hand movements. High-density surface EMG signals and ultrafast US images were concurrently recorded from the forearm muscles. The EMG signals were decomposed into MU spike trains and the separation vectors were kept for MU decoding across movements. The three-dimensional activation zone of each MU, including the action potentials on the skin surface and the twitch area in the muscle cross-sections, was extracted to validate the tracking performance.

On average, 37 ± 17 MUs were identified from each decomposition, of which 20±16,2±2, and 2 ± 1 MUs were successfully tracked based on action potential, twitch area, and both, respectively.

These results demonstrate strong evidence for the feasibility of tracking MUs using pre-trained separation vectors, supported by both electrical and mechanical responses. In addition, the study provides a benchmark to track MU activities across hand movements, extending the potential applications of neural decoding in neurophysiological investigation and human-machine interfacing.