Zahra Montazeriani - Department of Medical Physics and Biomedical Engineering, School of Medicine, Tehran University of MedicalSciences, Tehran, Iran- Research center for Science and Technology in Medicine (RCSTIM), Tehran University of Medical Sciences, Tehran
Pezhman Pasyar - Department of Medical Physics and Biomedical Engineering, School of Medicine, Tehran University of MedicalSciences, Tehran, Iran- Research center for Science and Technology in Medicine (RCSTIM), Tehran University of Medical Sciences, Tehran
Vahid Sadeghi - Department of Medical Physics and Biomedical Engineering, School of Medicine, Tehran University of MedicalSciences, Tehran, Iran- Research center for Science and Technology in Medicine (RCSTIM), Tehran University of Medical Sciences, Tehran

Brain-computer interface is referred to a technique for communicating between human brain’s neural activityand an external device. The brain’s activity can be interpreted on the basis of electroencephalography signals whichhave been recorded by means of elastic cap and sensors. The main goal of this paper is to correct recognition andclassification of three different mental tasks by analyzing the specific pieces of electroencephalography signals. Aftersignal processing and finding features, we use a statistical feature selection method to reduce the data dimensions andconsequently enhance the accuracy of the classifier. The results show that the proposed method is a promising approachand has a good performance for brain activity levels determination or bispectral index usage, since the classificationaccuracy of 91.1% (Error of 8.9%) is obtained for the mentioned classes.

Artificial neural network, Brain computer interface, Classifier, Electroencephalography, Feature extraction, Feature selection