This derivative dataset redistributes 20 per-patient feature tables computed from the Freiburg intracranial EEG database. Each row is a 2-second sliding window (1-second step) with 1080 univariate and bivariate features extracted from six electrodes across six frequency bands. Windows are labelled as pre-ictal, ictal, post-ictal or inter-ictal, so the tables can be used to train machine-learning models for epileptic seizure analysis. The CSV files are byte-identical to the source Zenodo record, and no raw recordings are included.
- Participants
- 20
- Size
- 27.7 GB
- Version
- vv1.0.2
- Updated
- Oct 16, 2026