EEG Band Separation Using Multilayer Perceptron for Efficient Feature Extraction and Perfect BCI Paradigm

For treatment of mental and brain diseases and diagnosis of abnormalities electroencephalogram (EEG) is an important measurement of brain activity. Feature extraction is vital in brain-computer interface (BCI) in the zone of biomedical and bioinformatics research alongside developing and adopting advanced signal processing techniques. Nonstationary and the nonlinear behavior of the EEG signal is the main challenge in feature extraction process. For the betterment of healthcare services, effective and affordable interpretation methods are the emerging keys. In this paper, the main focus is to separate different frequency band from EEG signal to extract features more efficiently using Multilayer Perceptron (MLP). B-Alert X10 is used for EEG acquisition and for analyzing the signal data, a virtual platform MATLAB has been used. For the classification of EEG bands Multilayer Perceptron Neural Network has been implemented which has been proved to be a more effective method with 95.47% accuracy for the classification.