nm000177
Multi-session longitudinal motor imagery EEG dataset (Kumar et al. 2024)
EEG
A longitudinal motor imagery EEG dataset from 18 brain-computer interface (BCI)-naive subjects acquired across 6 sessions (1 offline + 5 online) to investigate transfer learning and domain adaptation for calibration-free BCI training. Subjects performed left/right hand motor imagery tasks with visual feedback using 22 EEG channels sampled at 512 Hz. The dataset compares Generic Recentering (unsupervised) and Personally Assisted Recentering (supervised) domain adaptation frameworks, with features extracted as covariance matrices and classified using Riemannian geometry-based methods.
- Participants
- 18
- HED
- v8.4.0
- Size
- 1.66 GB
- Version
- v1.0.1
- Updated
- Jun 23, 2026