A neural network approach to EEG classification for the propose of differential diagnosis between epilepsy and normal EEG states

A neural network approach is presented to classify EEG signal for the propose of differential diagnosis between epilepsy and normal EEG states containing artifacts. The final set of features to be used as the input of the neural network is selected on the basis of the network intermediate performances. Computer simulations on both supervised and non-supervised networks show the fact that the features obtained in such a manner can be equally applied to diagnosis of complex disorders with no well-defined class description, such as the trauma resulted due to explosure wave.