ON THE DEVELOPMENT OF A BRAIN-COMPUTER INTERFACE SYSTEM USING HIGH-DENSITY MAGNETOENCEPHALOGRAM SIGNALS FOR REAL-TIME CONTROL OF A ROBOT ARM

Abstract

This work describes a brain-computer interface (BCI) system using multi-channel magnetoencephalogram (MEG) signals for real-time control of a computer game and a robot arm in a motor imagery paradigm. Computationally efficient spatial filtering and time-frequency decomposition facilitate the extraction and classification of neurophysiologically meaningful and task-relevant signal components from all of the 275 channels comprising the high-density sensor array. To our knowledge, this is the first report of an MEG-based BCI system capable of real-time signal processing and control using the whole sensor array. The robust and reliable performance of this system was demonstrated several times in front of a large public audience at an open day celebrating the 5th anniversary of the F.C. Donders Centre for Cognitive Neuroimaging at Radboud University Nijmegen. 1