Citation

Ali, Md.Wajed and Refat, Kawsar Ahmed and Sadeque, Md Golam and Roy, Dipon and Sarkar, Md Tanjil and Ramasamy, Gobbi (2025) A LightGBM-Augmented Convolutional Transformer with Adaptive Feature Recalibration for EEG Motor Imagery Classification. In: 2025 Multimedia University Engineering Conference, MECON 2025, 21 July 2025 - 23 July 2025, Cyberjaya.

Text IEEE Xplore Full-Text PDF_5.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Interpreting motor imagery (MI) tasks in braincomputer interface (BCI) systems remains a significant challenge in both neuroscience research and clinical applications. One of the challenges in MI classification tasks is finding an easily handled Electroencephalogram (EEG) representation method that can preserve both temporal and spatial features. In this study, we suggest a compact Convolutional Transformer to combine global and local characteristics into a single EEG classification system. This architecture starts with the feature extraction module-based Convolution block and adaptive feature recalibration (AFR). In particular, throughout the one-dimensional temporal and spatial convolution layers, the convolution module picks up lowlevel local characteristics, and the AFR can improve the quality of the extracted features by modeling the inter-dependencies between the features. The global correlation within the local temporal features is extracted by simply connecting the selfattention module. Subsequently, the LightGBM classifier module is used to predict the categories for EEG data trained on the extracted features. Finally, we have carried out comprehensive tests to evaluate our approach on the publicly available BCI Competition IV 2a dataset in EEG-based motor imagery and emotion identification procedures. The experimental results indicate the potential of this approach for advancing BCI systems, achieving an accuracy of 74.73%, which outperforms the state-of-the-art methods.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Attention mechanism, motor imagery.
Subjects:	T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL1-484 Motor vehicles. Cycles
Divisions:	Faculty of Artificial Intelligence & Engineering (FAIE)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	17 Mar 2026 02:13
Last Modified:	17 Mar 2026 02:13
URII:	http://shdl.mmu.edu.my/id/eprint/15460

Downloads

Edit (login required)