Spatio-temporal Spike and Slab Priors for Multiple Measurement Vector Problems

TL;DR

This paper introduces a probabilistic model with spatio-temporal spike and slab priors for the MMV problem, specifically targeting EEG source localization, and demonstrates its effectiveness through numerical experiments.

Contribution

It generalizes structured spike and slab priors for MMV problems and develops an Expectation Propagation inference scheme tailored for spatio-temporal data.

Findings

Model effectively captures spatio-temporal structure in EEG data

Approximate inference scheme viable for practical use

Numerical experiments confirm model's viability

Abstract

We are interested in solving the multiple measurement vector (MMV) problem for instances, where the underlying sparsity pattern exhibit spatio-temporal structure motivated by the electroencephalogram (EEG) source localization problem. We propose a probabilistic model that takes this structure into account by generalizing the structured spike and slab prior and the associated Expectation Propagation inference scheme. Based on numerical experiments, we demonstrate the viability of the model and the approximate inference scheme.