Enhanced electroencephalogram signal classification: A hybrid convolutional neural network with attention-based feature selection

• Published in: Brain research Vol. 1851; p. 149484
• Main Authors: Liu, Bao, Wang, Yuxin, Gao, Lei, Cai, Zhenxin
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 15.03.2025
• Subjects: Attention - physiology; Brain - physiology; Brain-Computer Interfaces; Convolutional Neural Networks; Convolutional neural networks (CNN); Deep Learning; EEG signals; Electroencephalography - methods; Humans; Imagination - physiology; Male; Motor imagery (MI); Neural Networks, Computer; Neurology; Signal Processing, Computer-Assisted; Temporal convolutional networks (TCN); Deep learning; Motor imagery (MI); EEG signals; Convolutional neural networks (CNN); Temporal convolutional networks (TCN)
• ISSN: 0006-8993; 1872-6240; 1872-6240
• DOI: 10.1016/j.brainres.2025.149484

[Display omitted] •Our model improves MI-EEG classification with deep learning-based feature extraction.•It uses CNN and attention mechanisms to enhance MI-EEG features for better accuracy.•The model reduces MI-EEG signal noise via wavelet transform and referencing.•The model extracts MI-EEG features effectively, outperforming all compared models. Accurate recognition and classification of motor imagery electroencephalogram (MI-EEG) signals are crucial for the successful implementation of brain-computer interfaces (BCI). However, inherent characteristics in original MI-EEG signals, such as nonlinearity, low signal-to-noise ratios, and large individual variations, present significant challenges for MI-EEG classification using traditional machine learning methods. To address these challenges, we propose an automatic feature extraction method rooted in deep learning for MI-EEG classification. First, original MI-EEG signals undergo noise reduction through discrete wavelet transform and common average reference. To reflect the regularity and specificity of brain neural activities, a convolutional neural network (CNN) is used to extract the time-domain features of MI-EEG. We also extracted spatial features to reflect the activity relationships and connection states of the brain in different regions. This process yields time series containing spatial information, focusing on enhancing crucial feature sequences through talking-heads attention. Finally, more abstract spatial–temporal features are extracted using a temporal convolutional network (TCN), and classification is done through a fully connected layer. Validation experiments based on the BCI Competition IV-2a dataset show that the enhanced EEG model achieves an impressive average classification accuracy of 85.53% for each subject. Compared with CNN, EEGNet, CNN-LSTM and EEG-TCNet, the classification accuracy of this model is improved by 11.24%, 6.90%, 11.18% and 6.13%, respectively. Our work underscores the potential of the proposed model to enhance intention recognition in MI-EEG significantly.