Classification Method of EEG Signals Based on Wavelet Neural Network

A new wavelet neural network (WNN) is constructed combining wavelet transform and neural network theory to classify electroencephalogram (EEG) signals. The new WNN takes nonlinear mother wavelet as neuron instead of traditional nonlinear sigmoid function. It owns the merits of good generalization ability and high converging speed. In addition, multi-resolution and self-adaptation are also its advantages. Experimental results have shown that our method performs well for the classification of mental tasks from EEG data compared with the approaches based on traditional neural network. It can provide a new way for the EEG automation classification when the EEG is used as input signal to a brain computer interface (BCI).