A New Approach to Classify the Schizophrenic and Normal Subjects by Finding the Best Channels and Frequency Bands

In this paper, EEG signals of ten schizophrenic patients and ten age-matched control participants are analyzed with the objective of determining which frequency bands have the more discriminative information. Our signals are caught from 22 channels according to 10-20 international recording system. First, the frequency range of 0-72 Hz is divided to 18 non-overlap intervals. Then, for each channel, power of EEG signal in different frequency bands is calculated in each time frame. In order to find the more informative frequency bands, genetic algorithm (GA) is employed. Support Vector Machine (SVM) classifier is applied to assess our features, and its classification error is considered as the fitness function of GA. The results show the best frequency bands for most of channels are 0-4, 4-8 and 48-52 Hz. It is shown that Cz, Pz, T₃, T₄, F₃, F₈, T₆, O₁, O₂, and A2 channels have the higher accuracy and more discriminative information than the other channels. Most of these channels are located on or around the temporal lobes containing the limbic system that confirms the neuro-phychological differences in these areas between the schizophrenic and normal participants.