EUVIMEDEUVIMED entdecken
Crossref

Graph convolution neural network channel selection with attention for motor imagery EEG decoding

Haoyu Li, Weidong Dang iD, Lei Liu, Peizhen Du, Xiaonan Cui, Yushi Hao, Jinchen Hu, Kun Han, Xi Wang iD, Zhongke Gao iD

Chaos: An Interdisciplinary Journal of Nonlinear Science · 2026

DOI: 10.1063/5.0319599 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Chaos: An Interdisciplinary Journal of Nonlinear Science, 2026. Verfasst von Haoyu Li und 9 weiteren Autor:innen. Beteiligte Einrichtung: Electronic and Information Engineering, School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China. Hauptthemen laut MEDLINE: Graph Neural Networks, Electroencephalography, Attention, Imagination. Die Arbeit nennt 44 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate decoding of motor imagery electroencephalography (MI-EEG) signals is critical for practical brain–computer interface (BCI) systems. However, conventional approaches typically rely on dense multi-channel recordings, which not only introduce data redundancy but may also incorporate noise, thereby hindering real-world deployment. To address this challenge, we propose a graph neural network–based co-optimization framework that simultaneously performs channel selection and MI classification. The framework comprises two core components: one is the Key Channel Locator (KCL), which models EEG electrodes as graph nodes and identifies a subject-specific, fixed-size subset of informative channels through a dual-perspective evaluation that integrates graph convolutional topology with self-attention-derived feature importance, and the other is the UniEEG-Net, which efficiently decodes MI tasks from the selected channels using multi-scale temporal convolutions, depthwise separable spatial projection, and a self-attention mechanism. We extensively validate the proposed method on three datasets, including BCI Competition IV 2a, High Gamma, and a newly collected dataset. Experimental results demonstrate that our approach achieves performance comparable to that obtained with all channels while using significantly fewer electrodes. Moreover, UniEEG-Net exhibits classification accuracy surpassing current state-of-the-art models. The entire system is thus well-suited for real-world BCI applications, particularly in neurorehabilitation.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Graph Neural NetworksElectroencephalographyAttentionImaginationHumansBrain-Computer InterfacesAlgorithmsSignal Processing, Computer-AssistedConvolutional Neural Networks

Autor:innen und Institutionen

  • Electronic and Information Engineering, School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China.
  • Yantai Port Railway Company, Yantai 326400, China.
  • International Joint Institute of Tianjin University, Tianjin University, Fuzhou 350205, China.

ORCID-Profile: Weidong Dang, Xi Wang, Zhongke Gao

Zitieren

APA

Li, H., Dang, W., Liu, L., Du, P., Cui, X., Hao, Y., Hu, J., Han, K., Wang, X., & Gao, Z. (2026). Graph convolution neural network channel selection with attention for motor imagery EEG decoding. Chaos: An Interdisciplinary Journal of Nonlinear Science. https://doi.org/10.1063/5.0319599

Vancouver

Li H, Dang W, Liu L, Du P, Cui X, Hao Y, et al. Graph convolution neural network channel selection with attention for motor imagery EEG decoding. Chaos: An Interdisciplinary Journal of Nonlinear Science. 2026. doi:10.1063/5.0319599

BibTeX
@article{Li2026,
  author = {Li, Haoyu and Dang, Weidong and Liu, Lei and Du, Peizhen and Cui, Xiaonan and Hao, Yushi and Hu, Jinchen and Han, Kun and Wang, Xi and Gao, Zhongke},
  title = {Graph convolution neural network channel selection with attention for motor imagery EEG decoding},
  journal = {Chaos: An Interdisciplinary Journal of Nonlinear Science},
  year = {2026},
  doi = {10.1063/5.0319599}
}
Kostenlos starten →