Channel Reduction by Cultural-Based Multi-objective Particle Swarm Optimization Based on Filter Bank in Brain–Computer Interfaces

Applying many electrodes is undesirable for real-life brain–computer interface (BCI) application since the recording preparation can be troublesome and time-consuming. This chapter presented a novel channel selection method, named cultural-based multi-objective particle swarm optimization (CMOPSO) based on filter bank, which introduced a cultural framework to adapt the personalized flight parameters of the mutated particles. A filter bank was designed using a coefficient decimation (CD) technology. The broad frequency band (8–30 Hz) is divided into ten subbands with width 4 Hz and overlapping 2 Hz, and the channel selection algorithm was applied to each subband. The optimal channels were chosen from the best channels derived from each subband. The algorithm was tested on five four-class data sets and the experimental results showed that the approach outperforms the broad band approach in selecting a smaller subset of channels without the sacrifice of classification accuracy.