Advancing Cross-Subject Domain Generalization in Brain-Computer Interfaces with Multi-Adversarial Strategies

作     者：Liu, Yici Qin, Lang Chen, Xin Jeannes, Regine Le Bouquin Coatrieux, Jean Louis Shu, Huazhong 

作者机构：Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications Southeast University Ministry of Education Nanjing210096 China Southeast University Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing Nanjing210096 China  RennesF-3502 France Southeast University INSERM Nanjing210096 China The First Affiliated Hospital With Nanjing Medical University Department of Radiology China Universit de Rennes 1 Laboratoire Traitement du Signal et de l’Image Rennes35000 France Centre de Recherche en Information Biomedicale Sino-Francais Rennes35042 France National Institute for Health and Medical Research Rennes35000 France 

出 版 物：《IEEE Transactions on Instrumentation and Measurement》 (IEEE Trans. Instrum. Meas.)

年 卷 期：2025年第74卷

学科分类：0810[工学-信息与通信工程] 0808[工学-电气工程] 08[工学] 

基　　金：National Key Research and Development Program of China National Natural Science Foundation of China Innovation Project of Jiangsu Province 

主　　题：Signal reconstruction 

摘      要：A cross-subject domain generalization (DG) approach with multi-adversarial strategies (DGMA) is introduced to reduce brain-computer interfaces (BCIs) systems’ dependency on high-quality, subject-specific EEG data, making it adaptable to unseen domains. DGMA leverages annotated training data from other subjects and consists of three modules: (1) Prefeature extraction, enhancing EEG signal separability through preprocessing, data augmentation, and tangent space mapping;(2) Distribution feature updater, aligning inter-subject feature distributions with marginal maximum mean discrepancy (MMD);(3) Multi-adversarial training, initially using gradient reversal layer (GRL) to amplify domain differences and classification loss, allowing the model to learn diverse domain-specific features before minimizing these differences to balance domain transferability and discriminability. DGMA is capable of better capturing domain-specific features while achieving stronger generalization compared to traditional methods focused solely on minimizing domain differences. Validated on four motor imagery datasets, DGMA achieved state-of-the-art accuracies of 76.1% on BCI Competition IV 2a and 72.4% on the 002-2014 dataset. Additional tests on a private fatigue dataset and the SEED dataset yielded accuracies of 99.5% and 86.6%, respectively. © 1963-2012 IEEE.