Hands-free EEG-based control of a computer interface based on online detection of clenching of jaw

Abstract

This paper presents a method for characterizing electromyogram (EMG) contamination of electroencephalogram (EEG) signals caused by clenching of jaws. The goal of this study is to use volitional contraction of jaw muscles to generate command signals for triggering an external device. Frequency analysis of the EEG signal acquired from one electrode above the ear shows that clenching is demonstrated by high frequency components not observable during other jaw movements. Based on this analysis, a clenching index is defined and a threshold is chosen such that clenching index values above this threshold show the contraction of jaw muscles. The proposed algorithm is then paired with a virtual reality platform in which the user can navigate by sending commands through clenching her/his teeth. Our experimental results demonstrate that the sensitivity of the proposed method for online detection of clenching is 80% and its specificity is 88%.