# Spatial 3D Mat\'ern priors for fast whole-brain fMRI analysis

## TL;DR

This paper introduces a flexible and computationally efficient Bayesian framework for whole-brain fMRI analysis using spatial Matérn priors, improving interpretability and speed over previous methods.

## Contribution

It develops a novel inference framework based on Matérn covariance functions and SPDEs, enabling fast, interpretable, and more realistic spatial priors for fMRI analysis.

## Key findings

- Matérn priors outperform previous priors in activity map quality
- The proposed method achieves faster inference in high-dimensional data
- Empirical results validate the interpretability and effectiveness of the approach

## Abstract

Bayesian whole-brain functional magnetic resonance imaging (fMRI) analysis with three-dimensional spatial smoothing priors has been shown to produce state-of-the-art activity maps without pre-smoothing the data. The proposed inference algorithms are computationally demanding however, and the proposed spatial priors have several less appealing properties, such as being improper and having infinite spatial range. We propose a statistical inference framework for whole-brain fMRI analysis based on the class of Mat\'ern covariance functions. The framework uses the Gaussian Markov random field (GMRF) representation of possibly anisotropic spatial Mat\'ern fields via the stochastic partial differential equation (SPDE) approach of Lindgren et al. (2011). This allows for more flexible and interpretable spatial priors, while maintaining the sparsity required for fast inference in the high-dimensional whole-brain setting. We develop an accelerated stochastic gradient descent (SGD) optimization algorithm for empirical Bayes (EB) inference of the spatial hyperparameters. Conditionally on the inferred hyperparameters, we make a fully Bayesian treatment of the brain activity. The Mat\'ern prior is applied to both simulated and experimental task-fMRI data and clearly demonstrates that it is a more reasonable choice than the previously used priors, using comparisons of activity maps, prior simulation and cross-validation.