BNCI 2014-001 Motor Imagery dataset

The BNCI 2014-001 Motor Imagery dataset is a widely-used benchmark for brain-computer interface research, comprising EEG recordings from 9 healthy subjects performing four-class motor imagery tasks (left hand, right hand, feet, and tongue). Each subject completed two sessions with 6 runs per session, yielding 200 training and 240 test trials. The dataset features 22 EEG channels with two versions: original at 1000 Hz and downsampled to 100 Hz (using Chebyshev Type II filtering). Minimal preprocessing includes bandpass filtering (0.05-200 Hz) and 50 Hz notch filtering, making it a standard resource for evaluating multi-class motor imagery classification algorithms and cross-session transfer learning approaches.