Classification of imaginary tasks from three channels of EEG by using an artificial neural network

Abstract

We used an artificial neural network to recognize imaginary left or right hand movements from scalp recorded EEC signals. Subjects were asked to imagine moving their left or right hand when indicated by a visual cue. Three channels were used in the present study to test the feasibility of a practical brain computer interface system. C3, C4, and Fz were selected based on the fact that they showed distinct difference between power spectrum density (PSD) of imaginary left and right hand movements. The PSD features of the three channels were fed onto the artificial neural network and the output was left or right imaginary movement. Testing results in three subjects with 90 trials show an average success rate of 72.2%.