EEG electrode selection method based on BPSO with channel impact factor for acquisition of significant brain signal

Abstract

A brain–computer interface (BCI) based on motor imagery is a system that transforms a subject's intentions into control signals by classifying electroencephalograph (EEG) signals obtained from imagining the movement of a subject's limbs. On imagining a limb's movement, the primary motor cortex area is prominently activated. For our new paradigm, however, we do not know which positions are activated or not. In that case, a simple approach is to use as many electrodes as possible. The problem is that using many electrodes also causes other problems. When applying a common spatial pattern (CSP), which is an EEG extraction method, using many electrodes causes an overfitting problem. In addition, there is difficulty using this technique for medical analysis. To overcome these problems, we suggest an optimal electrode selection method using binary particle swarm optimization (BPSO) with a channel impact factor. We examined optimal selected electrodes among all electrodes using four optimization methods and compared the classification accuracy and the number of selected electrodes between BPSO, BPSO with a channel impact factor, a genetic algorithm (GA), and a harmony search (HS) using a support vector machine (SVM). The results showed that BPSO with a channel impact factor selected five fewer electrodes and even improved accuracy by 20.4% compared with HS, GA, and BPSO.