Skip to content
DraconiticResearch
Search Sources API
203,948 studies31 sourcesUpdated 42 min ago
Peer-reviewedCC BYCyborg and bionic systems (Washington, D.C.)

Beyond Static Representations: Learning Electroencephalogram Spatiotemporal Dynamics for Enhanced Emotion Recognition.

  • Ke Liu1,2
  • Wenrui Zhu1,3
  • Rui Xu1,3
  • Wenlong Wang4
  • Zhuliang Yu5
  • Wei Wu4
5 affiliations
  1. 1Chongqing Key Laboratory of Computational Intelligence, Chongqing University of Posts and Telecommunications, Chongqing 400065, China.
  2. 2Key Laboratory of Big Data Intelligent Computing, Chongqing University of Posts and Telecommunications, Chongqing 400065, China.
  3. 3School of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China.
  4. 4Department of Neurology, Songjiang Hospital and Songjiang Research Institute, Shanghai Key Laboratory of Emotions and Affective Disorders, Shanghai Jiao Tong University School of Medicine, Shanghai 201600, China.
  5. 5College of Automation Science and Engineering, South China University of Technology, Guangzhou 510641, China.
Read full textPDFDOICite

Abstract

Electroencephalogram (EEG)-based emotion recognition plays a pivotal role in brain-computer interfaces and mental health monitoring. However, conventional approaches often neglect the intrinsic spatial correlations among EEG channels and fail to effectively model both global temporal dependencies and the continuity of neural oscillations. To overcome these challenges, we propose the Dynamic Spatial Embedding and Temporal Attention Mamba (DSE-TAM) framework. Our Dynamic Spatial Embedding (DSE) module leverages graph convolutional networks to dynamically capture inter-channel relationships, preserving spatial coherence in EEG representations. Subsequently, the Temporal Attention Mamba module integrates self-attention with state-space models to jointly optimize long-range temporal dependencies and smooth transition dynamics. Extensive experiments on 2 public benchmarks (DEAP and DREAMER) and an in-house CEED dataset demonstrate that DSE-TAM surpasses state-of-the-art methods, achieving accuracies of 79.75% (arousal) and 78.27% (valence) on DEAP, 80.55% (arousal) and 83.91% (valence) on DREAMER, and 73.75% (3-class) on CEED. These results validate the framework's efficacy in advancing EEG-based emotion decoding for brain-computer interfaces. The source code is available at https://github.com/Raizeroko/DSE-TAM.

Related

  1. Source-resolved electroencephalogram microstates reveal opposing prefrontal-posterior dynamics and a candidate temporal-to-precuneus pathway in disorders of consciousness
    Kun Li, Shiyu Zhang, Xu Hu +7
    Frontiers in NeurosciencePeer-reviewedCC BY 4.0Full textEpidemiology & public healthFrontiersCrossref
  2. Spatiotemporal Dynamics of Active Root Zone Storage Revealed From Hybrid Machine Learning
    Georgios Blougouras, Shijie Jiang, Alexander Brenning +7
    Water Resources ResearchPeer-reviewedCC BYFull text5 citedEnvironmental science & pollutionHydrology & soilsOpenAlex
  3. Learning Topological Representations of Protein Structure and Dynamics
    Dominik Geng, Florian Graf, Martin Uray +1
    arXivPreprintCC BY-SA 4.0Full textBiophysicsStatisticsarXiv
  4. Learning Stiffness Dependent Fluid Structure Dynamics from Coarse Flow Representations
    Chun-Jun Pu, Li-Wei Chen, Hai-Bo Huang
    arXivPreprintCC BY 4.0Full textFluid dynamicsarXiv
  5. Understanding Spatiotemporal Dynamics of Urban Water Bodies via Multi-Classifier Machine Learning Approaches in Semi-Arid Ahmedabad City
    Ankit Deshmukh, Shreekant Varshney
    EarthArXivPreprintCC BY-NC 4.0Full textHydrology & soilsEarthArXiv

Cite

BibTeXRIS
BibTeX
@article{liu2026beyond,
  title = {{Beyond Static Representations: Learning Electroencephalogram Spatiotemporal Dynamics for Enhanced Emotion Recognition.}},
  author = {Ke Liu and Wenrui Zhu and Rui Xu and Wenlong Wang and Zhuliang Yu and Wei Wu},
  journal = {Cyborg and bionic systems (Washington, D.C.)},
  year = {2026},
  month = oct,
  doi = {10.34133/cbsystems.0701},
  url = {https://europepmc.org/article/MED/42953001},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Beyond Static Representations: Learning Electroencephalogram Spatiotemporal Dynamics for Enhanced Emotion Recognition.
AU  - Ke Liu
AU  - Wenrui Zhu
AU  - Rui Xu
AU  - Wenlong Wang
AU  - Zhuliang Yu
AU  - Wei Wu
PY  - 2026
DA  - 2026/10/28
JO  - Cyborg and bionic systems (Washington, D.C.)
DO  - 10.34133/cbsystems.0701
UR  - https://europepmc.org/article/MED/42953001
AB  - Electroencephalogram (EEG)-based emotion recognition plays a pivotal role in brain-computer interfaces and mental health monitoring. However, conventional approaches often neglect the intrinsic spatial correlations among EEG channels and fail to effectively model both global temporal dependencies and the continuity of neural oscillations. To overcome these challenges, we propose the Dynamic Spatial Embedding and Temporal Attention Mamba (DSE-TAM) framework. Our Dynamic Spatial Embedding (DSE) module leverages graph convolutional networks to dynamically capture inter-channel relationships, preserving spatial coherence in EEG representations. Subsequently, the Temporal Attention Mamba module integrates self-attention with state-space models to jointly optimize long-range temporal dependencies and smooth transition dynamics. Extensive experiments on 2 public benchmarks (DEAP and DREAMER) and an in-house CEED dataset demonstrate that DSE-TAM surpasses state-of-the-art methods, achieving accuracies of 79.75% (arousal) and 78.27% (valence) on DEAP, 80.55% (arousal) and 83.91% (valence) on DREAMER, and 73.75% (3-class) on CEED. These results validate the framework's efficacy in advancing EEG-based emotion decoding for brain-computer interfaces. The source code is available at https://github.com/Raizeroko/DSE-TAM.
SN  - 2097-1087
SN  - 2692-7632
ER  -