Classification of drowsy and controlled EEG signals

Abstract

Electroencephalogram (EEG) signal analysis provides ground for evaluation of various neurological disorders and implementation of Brain Computer Interface (BCI) for such neurological disabilities. These capabilities of BCI system enable patients suffering from severe motor disability to control variety of applications by simply generating commands using BCI channel like, brain controlled arm or wheel chair. Successful realization of an efficient Brain Computer Interface depends upon accuracy maintained during EEG signals recording, processing, feature extraction and classification. The patients with more alcoholic medicines are seems to be drowsy. In that case, it is very difficult to extract and classify the brain signals accurately. In this work, a comparative study of EEG signals, recorded during drowsiness condition and controlled condition for same mental task, is performed for successful implementation of a BCI system. For classifying between recorded EEG signals for both situations, Fast Fourier Transform (FFT) and Power Spectral Density (PSD) are calculated. Comparison between FFTs and PSDs of EEG signals for both mental conditions shows clear difference between two mental conditions.