Single-trial prediction of unilateral isometric force levels from EEG signals

Abstract

The extent of physiological circuits involved in the increasing isometric force levels during unilateral contraction has been shown to be different. Here, we explore the possibility of force level prediction based on electroencephalogram (EEG) activity recorded during a single trial ofunilateral 5% or 40% of maximal isometric voluntary contraction (MVC) into right hand grip movement and grip imagination. There are eleven healthy subjects involved in this study. The subjects were required to randomly perform 20 trials of each force level during right hand grip movement or grip imagination. We proposed to use common spatial patterns and coherence between EEG signals as the features to support vector machine for force level prediction. The results shown that the force levels can be predicted through single-trial EEG during unilateral isometric hand grip (mean accuracy=69.88± 7.24%) as well as during grip imagination (mean accuracy=64.8 9.05%). This may provide additional control commands for robot being controlled by users through brain-computer interface.