Design of adaptive EEG preprocessing algorithm for neurofeedback system

Epilepsy is a general neural disorder identified by manifestation of sudden transient electrical disturbance of the brain. The electroencephalogram (EEG) is an important parameter for analyzing brain actions. The understanding of complex behaviour of the neural system is possible with the help of EEG. The EEG is composed of electrical potentials arising from several neurons. A unique topography onto the scalp is result of projection of each source, including separate neural clusters and artifacts from blink or pulse. These scalp maps are mixed according to the principle of linear superposition. The paper explains the development of adaptive filtering technique for eliminating artifacts from EEG signals. Two adaptive techniques have been developed for applying to EEG signals. An adaptive algorithm is used because artifacts are non-periodic and non-stationary by nature. Using Mean Square Error (MSE) and the computational time a comparative study is made to assess the performances of the techniques. The results are compared with other EEG pre-processing methods such as Independent Component Analysis (ICA) and Principal Component Analysis (PCA). The adaptive filtering appears to be very fast but with compromise on accuracy. Therefore, for practical applications in which speed is important than accuracy, adaptive filtering can be employed.