journal article·2026-05-30Open access

PG-MCTFormer: A Prior-Guided Multi-Scale Convolutional Transformer for Interpretable Motor Imagery EEG Classification

Jiahui Yuan, Rui Zhang, Yazhou Zhao, Weidong Zhou, Lan Tian, Guoyang Liu

Biomimetics· Volume 11· Issue 6

Abstract

Motor imagery brain–computer interfaces (MI-BCIs) have important applications in neurorehabilitation, assistive communication, and non-muscular human–machine interaction. From a bionic neural-interfacing perspective, MI-BCI decoding provides a computational bridge between biological motor intention and external machine control. However, reliable motor imagery electroencephalography (MI-EEG) classification remains challenging due to the highly non-stationary features of MI-EEG and limited interpretability. In this work, we propose PG-MCTFormer, a prior-guided multi-scale convolutional Transformer for MI-EEG classification that integrates rhythm-aware temporal filtering, dual-scale spatial modeling, and contextual decoding within a unified architecture. We evaluated the model on the publicly available BCI Competition IV 2a dataset, achieving 85.08% average accuracy and a Cohen’s kappa of 0.80, with significant performance improvement over the traditional methods. Comprehensive multi-view interpretability analyses in the frequency, temporal, and spatial domains further show that the learned filters remain aligned with canonical MI-related bands, discriminative evidence concentrates in the middle-to-late imagery interval, and the spatial prior is refined into subject-adaptive sensorimotor topographic patterns. These results indicate that explicit neurophysiological priors can improve both the robustness and the interpretability of MI-EEG decoders for biomimetic neural-interface applications.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: EEG and Brain-Computer Interfaces · Epilepsy research and treatment · Functional Brain Connectivity Studies

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

School of Integrated Circuits, Shandong University, Jinan 250199, China

Department of Biomedical Engineering, New York University, New York, NY 10012, USA

Shenzhen Research Institute of Shandong University, Shenzhen 518000, China

Key Laboratory of Social Computing and Cognitive Intelligence, Dalian University of Technology, Ministry of Education, Dalian 116024, China

Funding

National Natural Science Foundation of China · Natural Science Foundation of Shandong Province · Guangdong Basic and Applied Basic Research Foundation · Shenzhen Fundamental Research Program · National Key Research and Development Program of China · Key Laboratory of Social Computing and Cognitive Intelligence (Dalian University of Technology), Ministry of Education

References (63)

Show source-provided references
1. Brain-Computer Interfaces for Communication and Control10.1145/1941487.1941506 ↗
2. Application of BCI Systems in Neurorehabilitation: A Scoping Review10.3109/17483107.2014.961569 ↗
3. Deep learning for electroencephalogram (EEG) classification tasks: A review10.1088/1741-2552/ab0ab5 ↗
4. Spatial Filtering and Neocortical Dynamics: Estimates of EEG Coherence10.1109/10.686789 ↗
5. Brunner, C., Leeb, R., Müller-Putz, G.R., Schlögl, A., and Pfurtscheller, G. (2008). BCI Competition 2008–Graz Data Set A, Institute for Knowledge Discovery, Graz University of Technology.
6. Tangermann, M., Müller, K.R., Aertsen, A., Birbaumer, N., Braun, C., Brunner, C., Leeb, R., Mehring, C., Miller, K.J., and Müller-Putz, G.R. (2012). Review of the BCI competition IV. Front. Neurosci., 6.10.3389/fnins.2012.00055 ↗
7. Lee, D.G., and Lee, S.B. (2025). Robust Motor Imagery–Brain–Computer Interface Classification in Signal Degradation: A Multi-Window Ensemble Approach. Biomimetics, 10.10.3390/biomimetics10120832 ↗
8. Premchand, B., Zhang, Z., Ang, K.K., Yu, J., Tan, I.O., Lam, J.P.W., Choo, A.X.Y., Sidarta, A., Kwong, P.W.H., and Chung, L.H.C. (2025). A Personalized Multimodal BCI–Soft Robotics System for Rehabilitating Upper Limb Function in Chronic Stroke Patients. Biomimetics, 10.10.3390/biomimetics10020094 ↗
9. A Review of Classification Algorithms for EEG-Based Brain–Computer Interfaces10.1088/1741-2560/4/2/R01 ↗
10. A Review of Classification Algorithms for EEG-Based Brain–Computer Interfaces: A 10 Year Update10.1088/1741-2552/aab2f2 ↗
11. Optimal Spatial Filtering of Single Trial EEG During Imagined Hand Movement10.1109/86.895946 ↗
12. Yong, X., Ward, R.K., and Birch, G.E. (2008). Robust Common Spatial Patterns for EEG Signal Preprocessing. Proceedings of the 2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, IEEE.
13. Improvement motor imagery EEG classification based on sparse common spatial pattern and regularized discriminant analysis10.1016/j.jneumeth.2020.108833 ↗
14. Regularized common spatial patterns with subject-to-subject transfer of EEG signals10.1007/s11571-016-9417-x ↗
15. Ang, K.K., Chin, Z.Y., Wang, C., Guan, C., and Zhang, H. (2012). Filter Bank Common Spatial Pattern Algorithm on BCI Competition IV Datasets 2a and 2b. Front. Neurosci., 6.10.3389/fnins.2012.00039 ↗
16. Local Temporal Common Spatial Patterns for Robust Single-Trial EEG Classification10.1109/TNSRE.2007.914468 ↗
17. Common Spatio-Time-Frequency Patterns for Motor Imagery-Based Brain Machine Interfaces10.1155/2013/537218 ↗
18. Multiband tangent space mapping and feature selection for classification of EEG during motor imagery10.1088/1741-2552/aac313 ↗
19. Improvement motor imagery EEG classification based on regularized linear discriminant analysis10.1007/s10916-019-1270-0 ↗
20. Motor imagery EEG classification based on ensemble support vector learning10.1016/j.cmpb.2020.105464 ↗
21. Deep learning in motor imagery EEG signal decoding: A Systematic Review10.1016/j.neucom.2024.128577 ↗
22. Enhancing motor imagery EEG signal decoding through machine learning: A systematic review of recent progress10.1016/j.compbiomed.2024.109534 ↗
23. Deep learning with convolutional neural networks for EEG decoding and visualization10.1002/hbm.23730 ↗
24. Deep learning for EEG motor imagery classification based on multi-layer CNNs feature fusion10.1016/j.future.2019.06.027 ↗
25. EEGNet: A compact convolutional neural network for EEG-based brain–computer interfaces10.1088/1741-2552/aace8c ↗
26. A Multi-Branch 3D Convolutional Neural Network for EEG-Based Motor Imagery Classification10.1109/TNSRE.2019.2938295 ↗
27. Ingolfsson, T.M., Hersche, M., Wang, X., Kobayashi, N., Cavigelli, L., and Benini, L. (2020). EEG-TCNet: An Accurate Temporal Convolutional Network for Embedded Motor-Imagery Brain-Machine Interfaces. Proceedings of the 2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC), IEEE.10.1109/SMC42975.2020.9283028 ↗
28. Physics-Informed Attention Temporal Convolutional Network for EEG-Based Motor Imagery Classification10.1109/TII.2022.3197419 ↗
29. EEG Conformer: Convolutional Transformer for EEG Decoding and Visualization10.1109/TNSRE.2022.3230250 ↗
30. CTNet: A convolutional transformer network for EEG-based motor imagery classification10.1038/s41598-024-71118-7 ↗
31. Temporal Convolutional Transformer for EEG Based Motor Imagery Decoding10.1038/s41598-025-16219-7 ↗
32. Compact convolutional transformer for subject-independent motor imagery EEG-based BCIs10.1038/s41598-024-73755-4 ↗
33. Multiscale time-frequency method for multiclass Motor Imagery Brain Computer Interface10.1016/j.compbiomed.2022.105299 ↗
34. Self-attention-based convolutional neural network and time-frequency common spatial pattern for enhanced motor imagery classification10.1016/j.jneumeth.2023.109953 ↗
35. A Parallel-Hierarchical Neural Network (PHNN) for Motor Imagery EEG Signal Classification10.1016/j.bspc.2023.105621 ↗
36. NexusNet: Lightweight Graph Modeling for Motor-Imagery-Based Brain–Computer Interfaces10.1109/JIOT.2025.3574180 ↗
37. A Brain Topography Graph Embedded Convolutional Neural Network for EEG-Based Motor Imagery Classification10.1016/j.bspc.2024.106401 ↗
38. A Topology-Aware Multiscale Feature Fusion Network for EEG-Based Motor Imagery Decoding10.1016/j.knosys.2025.114540 ↗
39. Fine-Grained Spatial-Frequency-Time Framework for Motor Imagery Brain-Computer Interface10.1109/JBHI.2025.3536212 ↗
40. Motor imagery activates primary sensorimotor area in humans10.1016/S0304-3940(97)00889-6 ↗
41. Event-Related EEG/MEG Synchronization and Desynchronization: Basic Principles10.1016/S1388-2457(99)00141-8 ↗
42. Motor Imagery Classification by Means of Source Analysis for Brain-Computer Interface Applications10.1088/1741-2560/1/3/002 ↗
43. The neural network of motor imagery: An ALE meta-analysis10.1016/j.neubiorev.2013.03.017 ↗
44. EEG Source Imaging Enhances the Decoding of Complex Right-Hand Motor Imagery Tasks10.1109/TBME.2015.2467312 ↗
45. Motor imagery and direct brain-computer communication10.1109/5.939829 ↗
46. Unraveling motor imagery brain patterns using explainable artificial intelligence based on Shapley values10.1016/j.cmpb.2024.108048 ↗
47. Task-related reconfiguration patterns of frontoparietal network during motor imagery10.1016/j.neuroscience.2025.05.035 ↗
48. Ainslie, J., Lee-Thorp, J., de Jong, M., Zemlyanskiy, Y., Lebron, F., and Sanghai, S. (2023). GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, Association for Computational Linguistics.10.18653/v1/2023.emnlp-main.298 ↗
49. RoFormer: Enhanced Transformer with Rotary Position Embedding10.1016/j.neucom.2023.127063 ↗
50. Bai, S., Kolter, J.Z., and Koltun, V. (2018). An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv.
51. Signal Processing Approaches to Minimize or Suppress Calibration Time in Oscillatory Activity-Based Brain–Computer Interfaces10.1109/JPROC.2015.2404941 ↗
52. Multi-Scale Convolutional Transformer Network for Motor Imagery Brain-Computer Interface10.1038/s41598-025-96611-5 ↗
53. EMG contamination of EEG: Spectral and topographical characteristics10.1016/S1388-2457(03)00093-2 ↗
54. EMG and EOG artifacts in brain computer interface systems: A survey10.1016/j.clinph.2006.10.019 ↗
55. Beta-band oscillations—Signalling the status quo?10.1016/j.conb.2010.02.015 ↗
56. Mu and beta rhythm topographies during motor imagery and actual movements10.1023/A:1023437823106 ↗
57. Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., and Batra, D. (2017). Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization. Proceedings of the IEEE International Conference on Computer Vision, IEEE.10.1109/ICCV.2017.74 ↗
58. EEG-based intention recognition with deep recurrent-convolution neural network: Performance and channel selection by Grad-CAM10.1016/j.neucom.2020.07.072 ↗
59. Feature-aware domain invariant representation learning for EEG motor imagery decoding10.1038/s41598-025-95178-5 ↗
60. Kim, Y.K., Park, E., Lee, A., Im, C.H., and Kim, Y.H. (2018). Changes in network connectivity during motor imagery and execution. PLoS ONE, 13.10.1371/journal.pone.0190715 ↗
61. Channel selection from source localization: A review of four EEG-based brain–computer interfaces paradigms
62. Leeuwis, N., Yoon, S., and Alimardani, M. (2021). Functional Connectivity Analysis in Motor-Imagery Brain Computer Interfaces. Front. Hum. Neurosci., 15.10.3389/fnhum.2021.732946 ↗
63. Visualizing Data Using t-SNE