# Bayesian log-Gaussian Cox process regression: applications to meta-analysis of neuroimaging working memory studies

## TL;DR

This paper introduces a Bayesian log-Gaussian Cox process regression model for neuroimaging meta-analysis of working memory studies, effectively capturing spatial data and study heterogeneity to identify consistent brain activations.

## Contribution

It presents a novel Bayesian meta-regression approach using log-Gaussian Cox processes, with an efficient GPU-accelerated MCMC scheme for neuroimaging data analysis.

## Key findings

- Identified brain regions consistently activated during working memory tasks.
- Provided insights into interstudy variability in neuroimaging meta-analysis.
- Demonstrated the model's effectiveness on real neuroimaging data.

## Abstract

Working memory (WM) was one of the first cognitive processes studied with functional magnetic resonance imaging. With now over 20 years of studies on WM, each study with tiny sample sizes, there is a need for meta-analysis to identify the brain regions that are consistently activated by WM tasks, and to understand the interstudy variation in those activations. However, current methods in the field cannot fully account for the spatial nature of neuroimaging meta-analysis data or the heterogeneity observed among WM studies. In this work, we propose a fully Bayesian random-effects metaregression model based on log-Gaussian Cox processes, which can be used for meta-analysis of neuroimaging studies. An efficient Markov chain Monte Carlo scheme for posterior simulations is presented which makes use of some recent advances in parallel computing using graphics processing units. Application of the proposed model to a real data set provides valuable insights regarding the function of the WM.