One Model for All: Universal Pre-training for EEG based Emotion Recognition across Heterogeneous Datasets and Paradigms

2025-11-11

AI Summary

EEG-based emotion recognition suffers from poor model generalizability and limited cross-paradigm transferability due to dataset heterogeneity—particularly variations in channel configurations and subject-specific characteristics. To address this, we propose the first unified pretraining framework for multi-source EEG data. Our approach first standardizes inputs via a Unified Channel Schema (UCS), then employs a two-stage decoupled learning strategy to model cross-dataset knowledge. We further integrate an Adaptive Resampling Transformer (ART) with a Graph Attention Network (GAT) to jointly capture spatiotemporal dynamics and enhance noise robustness. Additionally, self-supervised contrastive learning is incorporated to improve representation discriminability. Extensive experiments on SEED, DEAP, and DREAMER achieve state-of-the-art performance. Crucially, our cross-dataset transfer results significantly outperform in-domain pretraining: improvements of +7.65% on DEAP and +3.55% on DREAMER demonstrate superior generalization across heterogeneous EEG sources.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

EEG-based emotion recognition is hampered by profound dataset heterogeneity (channel/subject variability), hindering generalizable models. Existing approaches struggle to transfer knowledge effectively. We propose'One Model for All', a universal pre-training framework for EEG analysis across disparate datasets. Our paradigm decouples learning into two stages: (1) Univariate pre-training via self-supervised contrastive learning on individual channels, enabled by a Unified Channel Schema (UCS) that leverages the channel union (e.g., SEED-62ch, DEAP-32ch); (2) Multivariate fine-tuning with a novel'ART'(Adaptive Resampling Transformer) and'GAT'(Graph Attention Network) architecture to capture complex spatio-temporal dependencies. Experiments show universal pre-training is an essential stabilizer, preventing collapse on SEED (vs. scratch) and yielding substantial gains on DEAP (+7.65%) and DREAMER (+3.55%). Our framework achieves new SOTA performance on all within-subject benchmarks: SEED (99.27%), DEAP (93.69%), and DREAMER (93.93%). We also show SOTA cross-dataset transfer, achieving 94.08% (intersection) and 93.05% (UCS) on the unseen DREAMER dataset, with the former surpassing the within-domain pre-training benchmark. Ablation studies validate our architecture: the GAT module is critical, yielding a +22.19% gain over GCN on the high-noise DEAP dataset, and its removal causes a catastrophic -16.44% performance drop. This work paves the way for more universal, scalable, and effective pre-trained models for diverse EEG analysis tasks.

Problem

• Addresses dataset heterogeneity in EEG emotion recognition across different studies
• Overcomes knowledge transfer limitations between diverse EEG recording setups
• Develops universal pre-training for generalizable cross-dataset EEG analysis

Innovation

• Universal pre-training framework for EEG across datasets
• Univariate contrastive learning with Unified Channel Schema
• Multivariate fine-tuning using ART and GAT architecture