The abstract is read at the publisher; the summary is JClub's.
SE-RTBNet enables intra-subject cross-task EEG transfer from motor execution to motor imagery, achieving average classification accuracies of 69.39%, 74.61%, and 98.94% across three public datasets.
- Why it matters: MI-BCIs are constrained by limited reliability of MI calibration because participant engagement during imagery is hard to verify, while ME offers more supervised motor-related signals.
- What they did: The Shared-Encoder Residual Task-Bridge Network uses a shared encoder for stable multi-band EEG motor representations and a performance-aware residual task-bridge routing module to align ME and MI features.
- The result: Across three datasets, the method attains the reported accuracies and outperforms existing transfer-learning algorithms overall, suggesting it can reduce dependence on extensive MI calibration data.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.