BRAIN ACTIVITY RECONGNITION THROUGH EEG CLASIFICATION

Abstract

Recent advances in computer hardware and signal processing have made possible the use of electroencephalogram (EEG) signals or “brain waves” for communication between humans and computers. The presence of artifacts, such as eye blinks, in EEG recordings obscures the underlying processes and makes analysis difficult. Large amounts of data must often be discarded because of contamination by eye blinks, muscle activity, and line noise. To overcome this difficulty, regression-based and blind source separation (BSS) based techniques are used to separate the artifacts from the EEG data of interest. To achieve the benefits of both techniques, a newly developed hybrid algorithms for EOG artifact rejection are FE-BSS, REG-ICA and PDAIC. REG-ICA removes successfully EOG activity and it also minimizes the distortion of the underlying brain activity as compared to above mentioned techniques. Peak detection algorithm of ICs (PDAIC) technique is to automatically remove eye-blink artifacts (EBA) based on ICA and peak detection algorithm for real time application. Removal of EMG contamination is a hard job for aforementioned algorithms because of noise like characteristics of EMG signals. A signal separation method based on canonical correlation analysis (CCA) is recently developed for EMG artifact rejection. After removal of artifacts, different methods based on Time Frequency Representations have been considered for the feature extraction of EEG.A grid based partition of time-frequency axis is purposed for the calculation of Power Spectral Density (PSD) of EEG segment. Classification of EEG segment (existence of epileptic seizure or not) is done through feedforward neural network. The feature extraction is done through short time fourier transform (STFT) and Choi-William (CW) distribution. The results indicate that classification performance of STFT is better as compared to CW distribution.