SDC-Net: A Domain Adaptation Framework with Semantic-Dynamic Consistency for Cross-Subject EEG Emotion Recognition

2025-07-23

AI Summary

To address the challenges of substantial inter-subject variability and unlabeled target-domain data in cross-subject EEG-based emotion recognition, this paper proposes a Semantic-Dynamic Consistency Domain Adaptation framework. Methodologically: (1) an intra-subject, intra-trial Mixup strategy is designed to preserve subject identity consistency; (2) a dynamic distribution alignment mechanism is formulated in a Reproducing Kernel Hilbert Space (RKHS) via multi-target kernel mean embeddings; and (3) a dual-domain similarity consistency learning scheme—requiring no temporal alignment—is introduced, integrating confidence-aware pseudo-labeling with latent-variable constraints. The framework significantly enhances unsupervised cross-domain semantic boundary modeling. Extensive experiments on SEED, SEED-IV, and Faced datasets demonstrate state-of-the-art performance in both cross-subject and cross-session transfer scenarios, achieving superior accuracy and generalization over existing methods.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Application Domains: Humanities & Computational Social Science
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Electroencephalography(EEG) based emotion recognition holds great promise for affective brain-computer interfaces (aBCIs), yet practical deployment remains challenging due to substantial inter-subject variability and the lack of labeled data in target domains. To overcome these limitations, we present a novel unsupervised Semantic-Dynamic Consistency domain adaptation network for fully label-free cross-subject EEG emotion recognition. First, we introduce a Same-Subject Same-Trial Mixup strategy that generates augmented samples via intra-trial interpolation, enhancing data diversity while explicitly preserving individual identity to mitigate label ambiguity. Second, we construct a dynamic distribution alignment module in reproducing kernel Hilbert space (RKHS), jointly aligning marginal and conditional distributions through multi-objective kernel mean embedding, and leveraging a confidence-aware pseudo-labeling strategy to ensure stable adaptation. Third, we propose a dual-domain similarity consistency learning mechanism that enforces cross-domain structural constraints based on latent pairwise similarities, enabling semantic boundary learning without relying on temporal synchronization or label priors. To validate the effectiveness and robustness of the proposed SDC-Net, extensive experiments are conducted on three widely used EEG benchmark datasets: SEED, SEED-IV, and Faced. Comparative results against existing unsupervised domain adaptation methods demonstrate that SDC-Net achieves state-of-the-art performance in emotion recognition under both cross-subject and cross-session conditions. This advancement significantly improves the accuracy and generalization capability of emotion decoding, and lays a solid foundation for real-world applications of personalized affective brain-computer interfaces (aBCIs). The source code will be released at https://github.com/XuanSuTrum/SDC-Net.

Problem

• Addresses cross-subject EEG emotion recognition without labeled target data
• Reduces inter-subject variability via semantic-dynamic consistency adaptation
• Enhances emotion decoding accuracy for personalized affective brain-computer interfaces

Innovation

• Same-Subject Same-Trial Mixup for label-free augmentation
• Dynamic distribution alignment in RKHS space
• Dual-domain similarity consistency learning mechanism