CLASSIFYING VARIOUS BRAIN ACTIVITIES BY EXPLOITING DEEP LEARNING TECHNIQUES AND GENETIC ALGORITHM FUSION METHOD

Abstract

The scan of functional Magnetic Resonance Imaging (fMRI) can provide three views for brain activities. These views are basically the X_axis (sagittal Plane), Y_axis (coronal plane) and Z_axis (axial plane). To the best of the obtained knowledge, studying brain activities for all of these views has not been considered before together with Deep Learning (DL) techniques. In this paper, various DL models named the X_axis Classification Model (XCM), Y_axis Classification Model (YCM) and Z_axis Classification Model (ZCM) are proposed. Each of these models is able to classify between the vision, movement and forward brain activities. Extensive experiments are performed for examining their parameters. The designed models have the capability to automatically detect the important features without any human supervision. In addition, they can provide intelligent decisions or classifications. Furthermore, effective combination method is suggested based on the Genetic Algorithm (GA) and Genetic Weighted Summation (GWS) rule, where high performances of outcomes can be achieved. After extensive experiments, the accuracies of 91.67%, 89.88% and 91.67% have been obtained for the XCM, YCM and ZCM, respectively.  In addition, the accuracy has been raised to 97.22% by applying the suggested fusion method.