Analysis of EEG Features for Brain Computer Interface Application

摘要

Electroencephalography (EEG) based assistive devices are the great support to the paralyzed patients to be in contact with their surroundings. These devices use Brain-Computer Interface (BCI) technology which is presently getting more attention by the related research community. In this paper, EEG features from multiple cognitive states have been explored for BCI applications. Here, Power Spectral Density (PSD), log Energy Entropy (logEE) and Spectral Centroid (SC) have been investigated as EEG feature. The EEG data have been captured from three different cognitive exercises; (ⅰ) solving math problem, (ⅱ) playing game and (ⅲ) do nothing (relax). The average PSD, average logEE and average SC of EEG Alpha and Beta band for three mental exercises are calculated in order to determine the best features that can be used for BCI application. The results of the research show that the EEG features when considering PSD, logEE and SC can be used to indicate the change in cognitive states after exposing the human to several cognitive exercises.