AUTHOR=Song Jiuxiang , Zhai Qiang , Wang Chuang , Liu Jizhong 

TITLE=EEGGAN-Net: enhancing EEG signal classification through data augmentation

JOURNAL=Frontiers in Human Neuroscience

VOLUME=Volume 18 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2024.1430086

DOI=10.3389/fnhum.2024.1430086

ISSN=1662-5161

ABSTRACT=Brain-computer interfaces (BCIs) represent a burgeoning technological frontier poised to enhance the quality of life for individuals grappling with disabilities. Despite their potential, widespread adoption has encountered challenges rooted in the limited accuracy of electroencephalogram (EEG) signal classification. In response to this predicament, we introduce a novel EEG signal classification model termed EEGGAN-Net, leveraging a data augmentation framework. By incorporating Conditional Generative Adversarial Network (CGAN) data augmentation, a cropped training strategy, and a Squeezeand-Excitation (SE) Attention mechanism, EEGGAN-Net adeptly assimilates crucial features from the data, consequently enhancing classification efficacy across diverse BCI tasks. Comparative analyses against other prominent models underscore EEGGAN-Net's superiority in terms of classification accuracy and volatility prediction. Ablation experiments provide further corroboration of the efficacy of CGAN data enhancement, the cropped training strategy, and the SE Attention mechanism. In conclusion, the amalgamation of data augmentation and attention mechanisms proves instrumental in acquiring generalized features from EEG signals, ultimately elevating the overall proficiency of EEG signal classification.