Brain–computer interface channel selection optimization using meta-heuristics and evolutionary algorithms

Many brain–computer interface (BCI) studies overlook the channel optimization due to its inherent complexity. However, a careful selection increases performance and users’ comfort while reducing cost of system. Evolutionary meta-heuristics, which have demonstrated their usefulness in solving complex problems, not been fully exploited yet this context. The purpose study is two-fold: (1) propose novel algorithm find an optimal set for each user compare it with other existing meta-heuristics; (2) establish guidelines adapting these strategies framework. A total 3 single-objective (GA, BDE, BPSO) 4 multi-objective (NSGA-II, BMOPSO, SPEA2, PEAIL) algorithms adapted tested public databases: ‘BCI competition III-dataset II’, ‘Center Speller’ ‘RSVP Speller’. Dual-Front Sorting Algorithm (DFGA), discrete method especially designed BCI framework, proposed as well. Results showed that all meta-heuristics outperformed full common 8-channel P300-based BCIs. DFGA significant improvement accuracy 3.9% over latter using also 8 channels; obtained similar accuracies mean 4.66 channels. topographic analysis reinforced need customize user. Thus, computes solutions different number channels, allowing select most appropriate distribution next sessions.