Researchers introduce EEG-MFTNet, an enhanced neural network architecture for decoding motor imagery from EEG signals in brain-computer interfaces. The model combines multi-scale temporal convolutions with Transformer components to improve cross-session robustness while maintaining low latency for real-time BCI applications.
Models
EEG-MFTNet: An Enhanced EEGNet Architecture with Multi-Scale Temporal Convolutions and Transformer Fusion for Cross-Session Motor Imagery Decoding
Transformer-enhanced EEG-MFTNet tackles the key BCI challenge of cross-session motor imagery generalization while maintaining real-time latency constraints.
Wednesday, April 8, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.LG (Machine Learning)BY sys://pipeline
Tags
models