Synergy matrices to estimate fluid wrist movements by surface electromyography

Although many efforts have been undertaken to develop an interface using surface electromyography (sEMG) to connect the gap between a human and a wrist prosthesis, most of these efforts have offered only static positioning (ON/OFF) of the prosthesis. This study introduced synergy matrices to extract fluid wrist movement intents by sEMG to allow individuals with wrist amputations to use wrist prostheses. A non-negative muscle synergy matrix was used to map muscle activities in the forearm into four predefined wrist movement intents (flexion/extension and radial/ulnar deviation). The directions of the predefined intents were constrained to two perpendicular axes, so each movement spanned only a one-dimensional space. A joint synergy matrix was used to span the whole two-dimensional space by combining the four wrist movement intents. Ten healthy subjects volunteered for a validation experiment, which was built as a virtual environment in which people with wrist amputation could receive myoelectric control training. The results showed that proportional two-degree-of-freedom (DOF) movements could be estimated by sEMG. This work could be useful not only for wrist prostheses but also for alternative computer interfaces and studies to examine motor adaptation by sEMG.