Feature alignment and enhancement network with guided tuning for non-stationary EEG classification
Journal of Neural Engineering · 01.04.2026 · 23 (2) · 026022
DOI: 10.1088/1741-2552/ae512e ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Neural Engineering, online erschienen am 31.03.2026. Verfasst von Donglin Li und 3 weiteren Autor:innen. Beteiligte Einrichtung: The College of Electrical Engineering, Shenyang University of Technology, Shenyang 110000, People's Republic of China. Hauptthemen laut MEDLINE: Electroencephalography, Brain-Computer Interfaces, Imagination, Neural Networks, Computer. Die Arbeit nennt 35 Literaturangaben. Gefördert von: Science and Technology Department of Liaoning Province, National Natural Science Foundation of China, Key Scientific Research Project of Liaoning Provincial Department of Education.
Abstract (Originalsprache)
Objective. Electroencephalogram (EEG) signal variability caused by external factors and subject differences limits the adaptation of motor imagery (MI) classification models in brain–computer interfaces (BCIs). Existing domain alignment methods often inadequately utilize critical source and target domains information, leading to negative transfer problems. This paper proposes a Feature Alignment and Enhancement Framework for cross-domain MI-EEG classification to address these limitations. Approach. First, by aligning the covariance matrices of the source and target domains, the spatial distributions of the two domains are preliminarily aligned, establishing a consistent foundation for feature mapping. Second, a conditional domain adversarial network optimizes cross-domain representations, reducing distribution discrepancies while enhancing discriminability. Finally, this paper introduces an EEG feature-based guided tuning method. This method extracts high-confidence features from both the source and target domains and generates centroid features to construct cross-domain feature banks. The input feature representations are dynamically optimized by attending to the relationships between centroid features, thus enhancing the model’s adapt-ability to target domain tasks. Main results. Experimental data show that in the four-class MI task of the BCI Competition IV-2a dataset, the cross-session and cross-subject model classification accuracies were 76.89% and 57.91%, respectively. The model achieved accuracy rates of 84.61% and 82.78% on the BCI Competition IV-2b datasets and the High Gamma Datasets, respectively, as well as 84.09% and 70.81%. Significance. The proposed framework effectively mitigates cross-domain variations, providing a reliable solution for cross-session and cross-subject MI-EEG classification.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- The College of Electrical Engineering, Shenyang University of Technology, Shenyang 110000, People's Republic of China.
- The School of Engineering Training and Innovation, Shenyang Jianzhu University, Shenyang 110168, People's Republic of China.
- The Post-Doctoral Station, Northern Heavy Industries Group, Shenyang 110141, People's Republic of China.
ORCID-Profile: Donglin Li, Jiacan Xu
Förderung
- Science and Technology Department of Liaoning Province
- National Natural Science Foundation of China
- Key Scientific Research Project of Liaoning Provincial Department of Education
- Doctoral research project funded by the Liaoning Provincial Department of Education
- Liaoning Province Doctoral Research Start-up Foundation
Zitieren
APA
Li, D., Wang, J., Xu, J., & Zhang, Y. (2026). Feature alignment and enhancement network with guided tuning for non-stationary EEG classification. Journal of Neural Engineering, 23(2), 026022. https://doi.org/10.1088/1741-2552/ae512e
Vancouver
Li D, Wang J, Xu J, Zhang Y. Feature alignment and enhancement network with guided tuning for non-stationary EEG classification. Journal of Neural Engineering. 2026;23(2):026022. doi:10.1088/1741-2552/ae512e
BibTeX
@article{Li2026,
author = {Li, Donglin and Wang, Jingyu and Xu, Jiacan and Zhang, Yuxian},
title = {Feature alignment and enhancement network with guided tuning for non-stationary EEG classification},
journal = {Journal of Neural Engineering},
year = {2026},
volume = {23},
number = {2},
pages = {026022},
doi = {10.1088/1741-2552/ae512e}
}
EUVIMED