Brain–Computer Interface Channel Selection Optimization Using Meta-Heuristics and Evolutionary Algorithms

Abstract

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

Note:
Funding Information: This study was partially funded by projects PID2020-115468RBI00 and RTC2019-007350-1 of the ‘Ministerio de Ciencia e Innovación’ and European Regional Development Fund (ERDF); 0702_MIGRAINEE_2_E (‘Análisis y correlación entre la epigenética y la actividad cerebral para evaluar el riesgo de migraña crónica y episódica en mujeres’) of the European Commission, as well as by CIBER-BBN through ‘Instituto de Salud Carlos III’ co-funded with ERDF funds. Eduardo Santamaría-Vázquez was in receipt of a PIF grant by the ‘Consejería de Educación de la Junta de Castilla y León’.

Conflict of Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Keywords: Brain–computer interface (BCI), Channel selection, Multi-objective optimization, Evolutionary algorithms, P300 event-related potentials