RESUMO
We investigated the possibility to classify data from patients with different psychiatric disorders on the base of physiological measures of the brain activity. Event-related potentials during the execution of GO/NOGO task were recorded from age matched groups of patients with schizophrenia, major depression and obsessive-compulsive disorder. Separation of independent components for event-related potentials was performed by applying Independent Component Analysis method. Picked out components were used in the method of discriminant analysis to classify the data according to clinical diagnosis. Discriminant analysis of components for event-related potentials allows classifying the data of patients with a suitable amount of matches with an established diagnosis.