Analysis and Classification of Motor Imagery Using Deep Neural Network

Isah Salim Ahmad (State Key Laboratory of Reliability and Intelligence of Electrical Equipment, Hebei University of Technology Tianjin 300130, China)
Shuai Zhang (State Key Laboratory of Reliability and Intelligence of Electrical Equipment, Hebei University of Technology Tianjin 300130, China)
Sani Saminu (Hebei University of Technology, Biomedical Engineering Department, University of Ilorin-Nigeria.)
Isselmou Abd El Kader (State Key Laboratory of Reliability and Intelligence of Electrical Equipment, Hebei University of Technology Tianjin 300130, China)
Jamil maaruf musa (Department of Computer Science and Technology, School of Artificial Intelligence. Hebei University of Technology, Tianjin 300401, China)
Imran Javid (State Key Laboratory of Reliability and Intelligence of Electrical Equipment, Hebei University of Technology Tianjin 300130, China)
Souha Kamhi (State Key Laboratory of Reliability and Intelligence of Electrical Equipment, Hebei University of Technology Tianjin 300130, China)
Ummay Kulsum (State Key Laboratory of Reliability and Intelligence of Electrical Equipment, Hebei University of Technology Tianjin 300130, China)

Journal of Applied Materials and Technology
Vol. 2 No. 2 (2021): March 2021

Motor imagery based on brain-computer interface (BCI) has attracted important research attention despite its difficulty. It plays a vital role in human cognition and helps in making the decision. Many researchers use electroencephalogram (EEG) signals to study brain activity with left and right-hand movement. Deep learning (DL) has been employed for motor imagery (MI). In this article, a deep neural network (DNN) is proposed for classification of left and right movement of EEG signal using Common Spatial Pattern (CSP) as feature extraction with standard gradient descent (GD) with momentum and adaptive learning rate LR. (GDMLR), the performance is compared using a confusion matrix, the average classification accuracy is 87%, which is improved as compared with state-of-the-art methods that used different datasets.