Classificação Automática de Sinais Anormais de EEG por meio de Microestados e Aprendizado de Máquina

Abstract

Automatic classification of electroencephalography (EEG) signals into normal or abnormal is the first step for the automatization in the detection of neuropathologies, and has the potential to considerably reduce the time between signal capture and medical report. A technique that has not yet been explored for this specific task, but which has shown a good ability to detect mental disorders, due to its capacity to capture spatial and temporal information, is the EEG microstate analysis. This work proposes a methodology for detecting abnormal EEG signals combining the use of microstates and a Learning Vector Quantization (LVQ) network for better discrimination of the microstates. Experimental results in a public database suggest that microstate analysis, which uses the topographic characteristics of the signal, are promising for the normal/abnormal EEG classification, regardless of a neuropathology specified a priori.