Dimension constraints improve hypothesis testing for large-scale, graph-associated, brain-image data

TL;DR

This paper introduces GraphMM, an empirical Bayes method that leverages graph structure in large-scale brain imaging data to improve hypothesis testing power and false discovery rate control, especially for connected subgraph signals.

Contribution

The paper proposes GraphMM, a novel empirical Bayes approach that incorporates graph constraints to enhance detection power in large-scale brain imaging studies.

Findings

GraphMM controls false discovery rate effectively across various settings.

GraphMM yields higher detection power than traditional methods in brain MRI data.

Connected subgraph signals are more accurately identified using GraphMM.

Abstract

For large-scale testing with graph-associated data, we present an empirical Bayes mixture technique to score local false discovery rates. Compared to empirical Bayes procedures that ignore the graph, the proposed method gains power in settings where non-null cases form connected subgraphs, and it does so by regularizing parameter contrasts between testing units. Simulations show that GraphMM controls the false discovery rate in a variety of settings. On magnetic resonance imaging data from a study of brain changes associated with the onset of Alzheimer's disease, GraphMM produces substantially greater yield than conventional large-scale testing procedures.