International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | A Hybrid Multi-Scale Deep Learning Framework for Robust Motor Imagery EEG Classification in Brain–Computer Interfaces |
|---|---|
| ABSTRACT | EEG-based Brain–Computer Interfaces sit at a fascinating intersection of neuroscience and engineering, offering a pathway through which the brain can directly manipulate external devices without relying on the conventional motor system. This capability is especially meaningful for individuals living with severe physical disabilities, where restoring even partial communicative agency can be life-changing. Within this landscape, the classification of motor imagery (MI) signals — patterns of brain activity generated when a person mentally rehearses a movement rather than physically performing it — occupies a central role in translating intent into action. Yet the problem is far from solved. EEG recordings are inherently messy: they are corrupted by biological and environmental artefacts, they shift in character from one recording session to the next, and they vary dramatically between individuals. These characteristics place hard limits on how well conventional machine learning pipelines can perform, since such pipelines typically depend on hand-engineered features that do not travel well across experimental conditions. This paper investigates the landscape of existing classification approaches and, drawing on identified shortcomings, introduces a hybrid multi-scale deep learning architecture. The model couples multi-resolution convolutional blocks with an attention mechanism, jointly exploiting the spatial topology of EEG electrodes and the temporal dynamics of neural oscillations. Evaluation against widely used benchmark datasets reveals that the proposed framework improves classification accuracy and cross-subject robustness while keeping computational demands manageable — a balance that prior work has rarely achieved simultaneously. |
| AUTHOR | PROF. USHA K, DARSHAN G S, DEEPAK P, OMKAR SHANKAR HIREMATH, DEVARADDI RABBANAHALLI Assistant Professor, Dept. of CSE, Jain Institute of Technology, Davangere, Karnataka, India UG Students, Dept. of CSE, Jain Institute of Technology, Davangere, Karnataka, India |
| VOLUME | 184 |
| DOI | DOI: 10.15680/IJIRCCE.2026.1405093 |
| pdf/93_A Hybrid Multi-Scale Deep Learning Framework for Robust Motor Imagery EEG Classification in Brain–Computer Interfaces.pdf | |
| KEYWORDS | |
| References | [1] V. J. Lawhern, A. J. Solon, N. R. Waytowich, S. M. Gordon, C. P. Hung, and B. J. Lance, "EEGNet: A compact convolutional neural network for EEG-based brain–computer interfaces," Journal of Neural Engineering, vol. 15, no. 5, p. 056013, 2018. [2] P. Andrikopoulos and S. Mehrkanoon, "EEG-MFTNet: Enhanced EEGNet with multi-scale temporal convolutions and transformer fusion," arXiv preprint, 2026. [3] A. Das, S. Sharma, and R. Gupta, "Enhanced EEG signal classification in brain–computer interfaces using hybrid deep learning models," Scientific Reports, vol. 15, 2025. [4] C. S. Chen, Y.-H. Lin, and J.-W. Liaw, "QEEGNet: Quantum machine learning for EEG signal encoding in brain–computer interfaces," IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2025. [5] O. W. Gómez-Morales, A. Quiroga, and H. Torres, "EEG signal prediction for motor imagery classification in reduced-channel BCIs," Sensors, vol. 25, 2025. [6] R. Lyu, "Deep learning approaches for EEG-based healthcare applications: A systematic review," Frontiers in Human Neuroscience, vol. 20, 2026. [7] W. Luo et al., "EEG-based brain–computer interfaces: Fundamentals, methods, applications, and emerging challenges," IEEE Internet of Things Journal, 2025. [8] X. Liu et al., "Recent applications of EEG-based brain–computer interfaces in the medical field," Military Medical Research, vol. 12, 2025. |
