EEG data classification with several mental tasks

The main contribution of this paper is the development of neural network models for classification of temporal data from subjects. The study focuses on the analysis of electroencephalogram (EEG) signals. A headband with three embedded electrodes was used to record EEG signals from a subject's forehead. Neural networks were trained to classify 8 seconds segments of 10-subbands, EEG data into classes corresponding to several mental tasks performed by subjects. Two layer backpropagation neural networks were trained using cross-validation. The average percentage of test segments correctly classified ranged from 70% to 90% for each mental task.