ORIGINAL RESEARCH article
Front. Neurosci.
Sec. Neuroprosthetics
A Multi-Domain Graph Convolutional Network-based Prediction Model for Personalized Motor Imagery Action
Provisionally accepted
- 1Hebei University of Technology, Beichen District, China
- 2Tianjin Hospital, Tianjin, China
- 3China Electronics Technology Group Corporation, Beijing, China
- 4Nankai University, Tianjin, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Motor imagery (MI)-based brain-computer interfaces (BCIs) provide a novel method to decode action imagination. Our previous study has demonstrated that actions serve as a key factor in causing individual differences, and cognitive EEG signals show a positive correlation with MI in reflecting these differences, providing a basis for predicting suitable MI actions for each individual. This study proposes a multi-domain graph convolutional network (M-GCN) to predict personalized MI action using cognitive data. The M-GCN extracts time, frequency, spatial domain features from cognitive tasks to construct multi-domain brain networks by different EEG quantization methods according to the characteristics of the three domains. Subsequently, the M-GCN utilizes spectral GCN to learn topology relationship between EEG channels by analyzing functional connection strength. Finally, for each action, the M-GCN has the capability to accurately map the cognitive data to the corresponding MI action and to output the personalized action for each subject. A subject-independent decoding paradigm with leave-one-subject-out cross-validation is adopted to validate the model on ten subjects. Compared to baseline and single-domain models, the M-GCN achieves the highest prediction accuracy 73.60% (p=7.1×10−3), improving by 15.87% (p=2.0×10−4) and by 7.2% (p=4.0×10−4), respectively. This study proves that the M-GCN can precisely predict personalized MI action, reflecting the efficiency of the multi-domain feature fusion based on cognitive task and GCN, and providing a novel method for personalized BCI.
Keywords: Brain-computer interface, Graph convolutional network, Feature fusion, brainnetwork, correlation between cognitive tasks and MI, MI prediction
Received: 28 May 2025; Accepted: 29 Sep 2025.
Copyright: © 2025 Ge, Wang, Zheng, Li, Wang and Xu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Mengfan Li, mfli@hebut.edu.cn
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.