Discrimination between mental and motor tasks of EEG signals using different classification methods

This paper presents comparison of different classification algorithms which are Linear Discriminant Analysis, Support Vector Machines and Neural networks for EEG signals recorded during mental and motor tasks from a subject. The purpose was to determine an optimum classification scheme that could be efficiently used in a brain-computer interface application. Each EEG data set were first excluded from noise and after that a feature selection procedure has been applied and then the data given to the classifier. This research demonstrated that according to the selected features, among these classification methods Support Vector Machines has the highest accuracy rates. On the other hand the training time of Linear Discriminant Analysis is the shortest of all. While the classification accuracy of mental and motor task comparison is high for all three classification methods, it decreases in case of two motor task comparison.