Real-time hybrid SSSEP-MI based brain computer interface system for upper limb rehabilitation

Abstract

Brain-computer interface (BCI) system is currently utilized to provide the neuro-rehabilitation especially for the stroke, brain injury, and cerebral palsy patients. In this work, we develop the prototype of the hybrid BCI system that simultaneously employs both motor imagery (MI) potential and steady-state somatosensory evoked potential (SSSEP) for motor and sensation of upper limbs rehabilitation. The MI potential can be observed from the event-related desynchonization (ERD) pattern when the user imagines the right or left hand movement. For the SSSEPbased BCI system, we attach the tactile stimulator at specific frequency of 23 Hz on the palm. The classification accuracy of our SSSEP-based and MI-based neuro-rehabilitation systems are approximately 82.5% and 75.27%, respectively. By systematically employ both systems simultaneously in one device; the rehabilitation process can be enhanced.