Estimators based on non-convex sparsity-promoting penalties were shown to yield state-of-the-art solutions to the magneto-/electroencephalography (M/EEG) brain source localization problem. In this paper we tackle the model selection problem of these estimators: we propose to use a proxy of the Stein's Unbiased Risk Estimator (SURE) to automatically select their regularization parameters. The effectiveness of the method is demonstrated on realistic simulations and 30 subjects from the Cam-CAN dataset. To our knowledge, this is the first time that sparsity promoting estimators are automatically calibrated at such a scale. Results show that the proposed SURE approach outperforms cross-validation strategies and state-of-the-art Bayesian statistics methods both computationally and statistically.
@article{arxiv.2112.12178,
title = {Electromagnetic neural source imaging under sparsity constraints with SURE-based hyperparameter tuning},
author = {Pierre-Antoine Bannier and Quentin Bertrand and Joseph Salmon and Alexandre Gramfort},
journal= {arXiv preprint arXiv:2112.12178},
year = {2021}
}