“…Lesser number of electrodes also increases the ITR by reducing the computational complexity. To select the most effective electrode configuration for classification, different electrode/channel selection methods such as recursive A c c e p t e d M a n u s c r i p t 8 channel elimination (Schröder et al, 2005), jump-wise regression (Colwell, Ryan, Throckmorton, Sellers, & Collins, 2014), Gibbs sampling (Speier, Deshpande, & Pouratian, 2015), Multi-ganglion ANN (Gao, Guan, Gao, & Zhou, 2015), Particle Swarm Optimization (PSO) (Jin, et al, 2010) and Genetic Algorithm (GA) (Kee, Ponnambalam, & Loo, 2015;Kee, Chetty, Khoo, & Ponnambalam, 2012) have been developed in recent years. The research shows that the best electrode configuration is subject dependent and hence the best configuration need to be selected for each subject (Blankertz et al, 2008;Rivet, Cecotti, Maby, & Mattout, 2012;Xu et al, 2013).…”