A study of the Naive Bayes classifier for analyzing imaginary movement EEG signals using the Periodogram as spectral estimator

Abstract

This study aims at applying algorithms for the spectral estimation and the classification of EEG signals during imaginary movements. Accordingly, it was used a database created by Graz University of Technology for the BCI Competition II. The database was created through an experiment in which an individual was asked to imagine the movement of her right or left hand while the EEG signal reading (Electroencephalogram) was being performed directly on the scalp by using three bipolar electrodes. In order to analyze the EEG signals for spectral estimation, it was used the Periodogram method computed in three different frequencies in μ-rythm (8 a 13Hz) and β-rythm (14 a 25Hz) frequencies to generate the Naive Bayes classifier entries. This database enabled to obtain a hit rate higher than 80%, which is consistent with results from the literature. Systems based on Brain-Computer Interface (BCI) can appropriate these algorithms to trigger and control devices through mental intentions only.