BCIwiki (bciwiki.com) — The study was published in Computer Methods in Biomechanics and Biomedical Engineering on August 10, 2026. Motor imagery (MI)-based brain-computer interfaces decode EEG signals into control commands. However, fine-grained MI decoding within the same limb remains challenging due to highly similar neural patterns.
Researchers Zhu L., Yue Q., Huang A., Zhang J., and Yuan P. propose a Contrastive Learning Network based on a Multi-Scale Transformer (CLMT-Net) for fine-grained MI decoding. CLMT-Net integrates multi-scale temporal convolution, FFT-based frequency fusion, spatial convolution, and dual-path Transformer to learn complementary EEG representations. Supervised contrastive learning further improves feature discrimination.
On the MI-2 dataset, CLMT-Net achieves an accuracy of 76.13 +/- 6.77% with a 95% confidence interval of [73.33, 78.92], demonstrating competitive performance for same-limb MI decoding.