This code reproduces the numerical results about fitting real EEG data with Wishart and t-Wishart distributions
To get the figures and the p-values of the statistical tests provided in the paper, please run "main.py"
The repository contains:
| Name | Description |
|---|---|
| main | Plot figures of fitting provided in the paper |
| preprocess_ssvep | Load the ExoSkeleton dataset, filter the SSVEP recordings, and cut them into trials |
| tWishart | Draw random samples from the t-Wishart distribution and derive the MLE for the center parameter given a degree of freedom |
| manifold | Framework for Riemannian optimization needed to compute the MLE of t-Wishart samples: manifold of the center parameter |
| fitting | Compute the empirical cumulative density function (cdf) of EEG samples and the cdfs of the fitted Wishart and t-Wishart samples and yield the Kolmogorov-Smirnov statistical tests for the Wishart and t-Wishart distributions |
numpy - scipy - matplotlib - moabb - pymanopt - mne - tqdm - joblib