Independent linear discriminant analysis : The first step in including covariates in EEG classification

International audience; The classifiers commonly used in BCIs do not take into account covariates such as changes in users’ mental states, eg., motivation or fatigue, which can influence EEG dynamic [1] and affect classification performances. We propose a linear discriminant analysis (LDA) variant [2] to account for these interfering covariates and improveperformance.