Tracking Event-Related Potentials During BMI Driven Rehabilitation

Abstract

Current brain-machine-interface (BMI) rehabilitation approaches typically focus on a specific aspect of neural activity. Auxiliary signals, derived from independent measures of neural activity and recorded in parallel might be useful in quantifying and tracking a subjects mental state and performance. In this work, we demonstrate that event-related potentials can be reliably observed in stroke survivors with chronic paralysis during a BMI intervention. The averaged evoked response remains stable over sessions and varies between subjects. A prominent negativity, positivity complex emerges whose features can be tracked across subjects and sessions.