mlsanityData to 13 Oct 2026

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

arXiv 2511.08444 ↗First posted 11 Nov 2025cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Builds on 21 of its 53 references that are in the map. First posted on arXiv 11 Nov 2025.

Builds on

21 of 53 references.

References · 21

  1. 2017Attention Is All You Needcited by 60,106
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,304
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,223
  4. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,009
  5. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,318
  6. 2017Graph Attention Networkscited by 8,888
  7. 2018Representation Learning with Contrastive Predictive Codingcited by 7,712
  8. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,364
  9. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 5,021
  10. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,909
  11. 2021CSDI: Conditional Score-based Diffusion Models for Probabilistic Time Series Imputationcited by 579
  12. 2021TS2Vec: Towards Universal Representation of Time Seriescited by 469
9 more
  1. 2022GraphMAE: Self-Supervised Masked Graph Autoencoderscited by 446
  2. 2021Time-Series Representation Learning via Temporal and Contextual Contrastingcited by 343
  3. 2021Contrastive Learning of Subject-Invariant EEG Representations for Cross-Subject Emotion Recognitioncited by 59
  4. 2018EEG-GAN: Generative adversarial networks for electroencephalograhic (EEG) brain signalscited by 47
  5. 2019ICLabel: An automated electroencephalographic independent component classifier, dataset, and websitecited by 46
  6. 2022EEG based Emotion Recognition: A Tutorial and Reviewcited by 43
  7. 2020Deep Learning in EEG: Advance of the Last Ten-Year Critical Periodcited by 10
  8. 2024Pre-Training Graph Contrastive Masked Autoencoders are Strong Distillers for EEGcited by 4
  9. 2024A Supervised Information Enhanced Multi-Granularity Contrastive Learning Framework for EEG Based Emotion Recognitioncited by 2

Built on by

Not cited yet.