Article Details

Title · Authors · Abstract · Keywords · DOI · References

A Lightweight EEG Signal Classification Model Based on Riemannian Manifold

WANG Long, YANG Zhanhai

(School of Mathematics and Computer Science, Yan'an University, Yan'an 716000, China)

Section：Information Application · Vol. 2026, Issue 10 · Pages：155-159 · 2096-4706（2026）10-0155-05 · CLC number：TN91 · Document code：A · DOI：10.19850/j.cnki.2096-4706.2026.10.029

Abstract

To improve the accuracy and robustness of Motor Imagery EEG (MI-EEG) decoding, this paper proposes a Dense Riemannian Attention Network. The model integrates efficient feature encoding, second-order statistical modeling, and a hierarchical Attention Mechanism. Firstly, temporal-spatial features are extracted using one-dimensional convolution along with spatial-spectral attention. Then, a stackable hierarchical Riemannian attention block is introduced to perform deep modeling on covariance features, while dense connections are employed to enhance feature reuse. Finally, a multi-scale structure is utilized to capture information across different temporal scales. Experimental results demonstrate that the proposed method achieves an average classification accuracy of 82.79% on the BCI Competition IV-2a dataset, significantly outperforming EEGNet (71.95%) and FBCNet (76.52%). These results verify the effectiveness and generalization capability of the proposed model for MI-EEG decoding tasks.

Keywords

EEG signal; motor imagery; Attention Mechanism; Deep Learning; brain-computer interface

References

  1. [1] VAFAEI E，HOSSEINI M. Transformers in EEG Analysis：A Review of Architectures and Applications in Motor Imagery，Seizure，and Emotion Classification [J/OL].Sensors，2025，25（5）：1293（2025-02-20）.https://doi.org/10.3390/s25051293.
  2. [2] YANG X，JIA Z. Spatial-Temporal Mamba Network for EEG-Based Motor Imagery Classification [C]//International Conference on Advanced Data Mining and Applications.Singapore：Springer，2024：418-432.
  3. [3] WANG J，YE W，HE J，et al. Integrating Biological and Machine Intelligence: Attention Mechanisms in Brain Computer Interfaces [J/OL].Information Fusion，2025：103417（2025-07-02）.https://doi.org/10.1016/j.inffus.2025.103417.
  4. [4] LIU H，YANG S，ZHANG Y，et al. LIBEER: A Comprehensive Benchmark and Algorithm Library for EEG-Based Emotion Recognition [J].IEEE Transactions on Affective Computing，2025，16（4）：3596-3613.
  5. [5] MA X，CHEN W，PEI Z，et al. A Temporal Dependency Learning CNN With Attention Mechanism for MI-EEG Decoding [J].IEEE Transactions on Neural Systems and Rehabilitation Engineering，2023，31：3188-3200.
  6. [6] JIN M，DU C，HE H，et al. PGCN：Pyramidal Graph Convolutional Network for EEG Emotion Recognition [J].IEEE Transactions on Multimedia，2024，26：9070-9082.
  7. [7] LI C，WANG F，ZHAO Z，et al. Attention-Based Temporal Graph Representation Learning for EEG-Based Emotion Recognition [J].IEEE Journal of Biomedical and Health Informatics，2024，28（10）：5755-5767.
  8. [8] XU B，ZHANG X，ZHANG X，et al. DDNet: A Hybrid Network Based on Deep Adaptive Multi-Head Attention and Dynamic Graph Convolution for EEG Emotion Recognition [J/OL].Signal，Image and Video Processing，2025，19（4）：293（2025-01-26）.https://doi.org/10.1007/s11760-025-03876-4.
  9. [9] MIAO Z，ZHAO M，ZHANG X，et al. LMDA-Net：A Lightweight Multi-Dimensional Attention Network for General EEG-Based Brain-Computer Interfaces and Interpretability [J/OL].NeuroImage，2023，276：120209（2023-06-06）.https://doi.org/10.1016/j.neuroimage.2023.120209.
  10. [10] ZHU X，LIU C，ZHAO L，et al. EEG Emotion Recognition Network Based on Attention and Spatiotemporal Convolution [J/OL].Sensors，2024，24（11）：3464（2024-05-27）.https://doi.org/10.3390/s24113464.
  11. [11] MA X，CHEN W，PEI Z，et al. Attention-Based Convolutional Neural Network With Multi-Modal Temporal Information Fusion for Motor Imagery EEG Decoding [J/OL].Computers in Biology and Medicine，2024，175：108504（2024-06-01）.https://doi.org/10.1016/j.compbiomed.2024.108504.
  12. [12] DING Y，LI Y，SUN H，et al. EEG-Deformer：A Dense Convolutional Transformer for Brain-Computer Interfaces [J].IEEE Journal of Biomedical and Health Informatics，2025，29（3）：1909-1918.
  13. [13] DEVARAJAN K，PONNAN S，PERUMAL S. Hybrid CNN-Transformer Architecture for Enhanced EEG-Based Emotion Recognition:Capturing Local and Global Dependencies With Self-Attention Mechanisms [J/OL].Discover Computing，2025，28（1）：87（2025-05-22）.https://link.springer.com/article/10.1007/s10791-025-09596-0.
  14. [14] LUO J，WANG Y，XIA S，et al. A Shallow Mirror Transformer for Subject-Independent Motor Imagery BCI [J/OL].Computers in Biology and Medicine，2023，164：107254（2023-07-25）.https://doi.org/10.1016/j.compbiomed.2023.107254.
  15. [15] HAN C，LIU C，WANG J，et al. A Spatial–Spectral and Temporal Dual Prototype Network for Motor Imagery Brain–Computer Interface [J/OL].Knowledge-Based Systems，2025，315：113315（2025-03-18）.https://doi.org/10.1016/j.knosys.2025.113315.
  16. [16] LIANG W，ALLISON B Z，XU R，et al. SecNet：A Second Order Neural Network for MI-EEG [J/OL].Information Processing & Management，2025，62（3）：104012（2024-12-06）.https://doi.org/10.1016/j.ipm.2024.104012.

Funding

Shaanxi Provincial Science and Technology Plan Project (2023-JC-QN-0744); Research Plan Project of Shaanxi Provincial Department of Education (23JK0738); Horizontal Project (YDYZH20241231)

About the authors

Wang Long (1999-), male, Han ethnicity, from Yan'an, Shaanxi Province, is currently pursuing a master's degree. His research interests include deep learning and EEG signals; Corresponding author: Yang Zhanhai (1973-), male, Han ethnicity, born in Jingyang, Shaanxi, associate professor, master's student, research direction: software engineering, network control.