SOURCE RECORD / S0193
A multi-day and high-quality EEG dataset for motor imagery brain-computer interface
A multi-day and high-quality EEG dataset for motor imagery brain-computer interface
- Authors or organization
- Yang B, Rong F, Xie Y, Li D, Zhang J, Li F, Shi G, Gao X
- Published date
- 2025-03-23
- Source type
- peer_reviewed_dataset
- Technology route
- non-invasive 58-channel EEG motor imagery
- Function or setting
- dataset and algorithm validation
- Rights status
- mirror_allowed / CC BY 4.0
Evidence summary
Evidence: Reusable multi-day MI-BCI dataset and within-dataset benchmark. Key figures: 58 channels; three sessions; 85.31% two-class and 76.90% three-class average accuracy. Limitations: Healthy contest cohort; reported 10-fold cross-validation does not itself prove held-out cross-subject or cross-session generalization.
Key figures
58 channels; three sessions; 85.31% two-class and 76.90% three-class average accuracy
Limitations
Healthy contest cohort; reported 10-fold cross-validation does not itself prove held-out cross-subject or cross-session generalization