Preictal seizure analysis on the Siena Scalp EEG dataset

Useful functions for pre-processing steps and extracting features from iEEG recordings

EEG analysis code for quantifying subject-identity effects in spectral features and showing how identity leakage can inflate machine-learning performance. Includes variance decomposition, biometric evaluation, cross-session and cross-paradigm tests, and MDD leakage simulations with subject-wise controls.

End‑to‑end simulated EEG signal decoding with feature engineering and transformer models.