EEG Data Classification Using CNN Method with Various Data Models

This research paper represents outcomes of the classification task aimed to distinguish three thoughts of turning off, turning on a light bulb and normal thought based on electroencephalogram (EEG). Participant, observing dynamic and static images on the screen showing turned on or turned off light bulb, generates EEG signals corresponding to the pictures. As a recording apparatus to obtain precise EEG signals the Emotiv Epoc X device with 14-channel was chosen. The classification of three types of thoughts was accomplished by applying Convolutional Neural Network (CNN) model the input of which is different types of features, namely EEG signal amplitudegenerated images, Power Spectrum Density-generated images and Wavelet transform-generated images. The results demonstrated that conversion of EEG signals from one-dimensional time series to image classification problem enhances the classification accuracy of the system. The proposed system achieved an average accuracy of classification 82.47%.