Electrophysiological brain activity during the control of a motor imagery-based brain–computer interface

Abstract

This article considers the features of five electroencephalogram patterns that are most frequently extracted by the independent component analysis when subjects imagine the movement of their hands during the control of a brain–computer interface (BCI). The solution of the EEG inverse problem using the individual geometrical head model shows that the sources of the revealed patterns are located at the bottom of the left and right central sulci, as well as in the left premotor cortex, supplementary motor area, and precuneus. The functional value of the patterns is discussed by comparing the location results with the results of the metaanalysis of the published data that were obtained using a functional magnetic resonance imaging. The source locations are the same for seven healthy subjects and four poststroke patients with subcortical damage location. However, despite the same locations, the two groups of subjects significantly differed in the frequency characteristics of the revealed patterns; in particular, the patients had no clearly pronounced activity in the upper α-band and were characterized by a much lower level of inhibition of rates in the primary somatosensory areas during motor imagery.