papersSEP 10 04:00 UTC
Deep Neural Networks Decode Finger Intent from sEMG for Post-Stroke Rehabilitation
An arXiv paper explores using deep learning to interpret finger-specific movement intentions from surface electromyography (sEMG) signals in stroke survivors. Because measurable muscle activity often persists even when movement is weak or incomplete, these signals can serve as control inputs for rehabilitation hardware. The study frames the problem as five-finger multilabel intent decoding.