Computational Intelligence for Pattern Recognition in EEG Signals

Electroencephalography (EEG) captures brain signals from the scalp and has applications in medicine, psychology, rehabilitation, and many other areas. EEG signals are inherently noise-prone, however, and patterns are rarely visible in raw signals. This chapter presents an overview of how computational intelligence is used to discover patterns in brain signals throughout a typical BCI pipeline: preprocessing, feature extraction, feature selection, and classification. Because EEG is the outcome of a complex, nonlinear, non-stationary stochastic biological process with many internal and external noise sources, appropriate computational intelligence methods are required at each step and can significantly improve end results.