Deep Learning Based Motor Imagery Intention Classification Using Electroencephalogram Signal

Recently, various machine learning and deep learning techniques have been used in motor imagery intention classification to keep up with the rapid advances in human-machine interaction. Nevertheless, there is a limited understanding of feature selection to assist in identifying the most informative and discriminatory characteristics of neural activity in different spatial locations. This study proposes a deep neural network that has the ability to learn to classify motor imagery intention using selected electroencephalogram features across different channels.  We extracted three sets of features from each channel and used sequential forward feature selection with support vector machines to select the best features. The proposed model was evaluated with a publicly available dataset after training and testing the deep neural network with the selected features. Our system model achieved an average of 85.85 ± 1.35% accuracy for classifying two motor imagery scenarios. These results demonstrate that the proposed method effectively identifies the most informative and discriminative characteristics of neural activity at different spatial locations to differentiate motor imagery intentions and contribute to future prosthetics and brain-computer interface (BCI) technology.