EEG motor imagery signals classification using maximum overlap wavelet transform and support vector machine

Abstract

A BCI system (Brain-Computer Interface) aims to the interpretation of brain signals perceived through electroencephalography (EEG) sensors in order to allow the user interaction with the environment through specific actions. In this paper we present an experiment of EEG signal classification under the motor imagery paradigm using two feature extraction methods for comparison purposes: discrete wavelet transform (DWT) and maximum overlap discrete wavelet transform (MODWT). The feature vectors are fed into a support vector machine (SVM) classification system. The results obtained show an accuracy of 98.81% in average.