Towards a Low Cost Mu-Rhythm Based BCI

The purpose of this paper is to evaluate whether mu rhythm based BCIs can be implemented using the low cost Emotiv Epoc EEG device. Synchronisation in the high alpha and low beta band caused by continuous imagery and real toes movement was recorded on 6 healthy subjects. We apply LDA and SVM classifiers in order to classify a trial as movement or non-movement by computing the power of the band of interest on C3, C4, FC3, FC4 standard positions. 10-cross validation results indicate that sensorimotor ERS takes some seconds to develop both in the case of imagery and real movement. The performance is better when classification occurs 5–7 seconds after the movement starts. A simple musical application where the user can move the pitch of a tone up and down in a musical scale is built based on real or imagery toes movement.