Fusion Convolutional Neural Network for Cross-Subject EEG Motor Imagery Classification

• Published in: Computers (Basel) Vol. 9; no. 3; p. 72
• Main Authors: Roots, Karel, Muhammad, Yar, Muhammad, Naveed
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.09.2020
• Subjects: Accuracy; Analysis; Artificial neural networks; Brain research; brain–computer interface (BCI); Classification; Computational efficiency; Computing costs; convolutional neural network (CNN); Datasets; deep learning; Discriminant analysis; Electroencephalography; electroencephalography (EEG); fusion network; Human motion; Image classification; Indexing; motor imagery (MI); Motor learning; Motor skills; Neural networks; Signal processing; Estonia
• ISSN: 2073-431X; 2073-431X
• DOI: 10.3390/computers9030072

Brain–computer interfaces (BCIs) can help people with limited motor abilities to interact with their environment without external assistance. A major challenge in electroencephalogram (EEG)-based BCI development and research is the cross-subject classification of motor imagery data. Due to the highly individualized nature of EEG signals, it has been difficult to develop a cross-subject classification method that achieves sufficiently high accuracy when predicting the subject’s intention. In this study, we propose a multi-branch 2D convolutional neural network (CNN) that utilizes different hyperparameter values for each branch and is more flexible to data from different subjects. Our model, EEGNet Fusion, achieves 84.1% and 83.8% accuracy when tested on the 103-subject eegmmidb dataset for executed and imagined motor actions, respectively. The model achieved statistically significantly higher results compared with three state-of-the-art CNN classifiers: EEGNet, ShallowConvNet, and DeepConvNet. However, the computational cost of the proposed model is up to four times higher than the model with the lowest computational cost used for comparison.