Unsupervised multi-source domain adaptation via contrastive learning for EEG classification

• Published in: Expert systems with applications Vol. 261; p. 125452
• Main Authors: Xu, Chengjian, Song, Yonghao, Zheng, Qingqing, Wang, Qiong, Heng, Pheng-Ann
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.02.2025
• Subjects: Brain–computer interfaces (BCIs); Contrastive learning; Electroencephalography (EEG); Motor imagery (MI); Unsupervised domain adaptation; Brain–computer interfaces (BCIs); Unsupervised domain adaptation; Motor imagery (MI); Electroencephalography (EEG); Contrastive learning
• DOI: 10.1016/j.eswa.2024.125452
• ISSN: 0957-4174

Individual differences in electroencephalography (EEG) present significant challenges for subject-independent EEG classification in brain–computer interfaces (BCIs). Existing domain adaptation methods often address individual differences by merging all source domains indistinguishably into a single source and aligning features between this aggregate source and the target domain. Neglecting the relationships between different source domains would hinder the model’s adaptability and generalization. Therefore, we propose a method called Contrastive Learning-based Unsupervised multi-source Domain Adaptation (CLUDA) for learning subject-independent representations in motor imagery. Our method not only aligns the conditional distributions of each source domain with the target domain but also reduces discrepancies among the source domains using contrastive learning, thus learning more generalized domain-invariant representations. Specifically, CLUDA effectively eliminates semantic variances by maximizing the similarity between positive pairs (same class) and minimizing the similarity between negative pairs (different classes) across subjects. Finally, we validated CLUDA on four motor imagery datasets and consistently achieved state-of-the-art performance. •We propose a novel unsupervised multi-source domain adaptation framework to effectively learn subject-invariant representations for EEG-based motor imagery.•We utilize contrastive learning to address each source-target and inter-source variability in the multi-source domain adaptation process, facilitating learning subject-independent representations.•We have validated the proposed method on four motor imagery datasets. The experimental results demonstrate the superior performance of our method.