Selective inference for fMRI cluster-wise analysis, issues, and recommendations for critical vector selection: A comment on Blain et al

TL;DR

This paper compares two permutation-based methods, Notip and pARI, for cluster-wise fMRI analysis, highlighting their differences, advantages, and drawbacks to guide better critical vector selection.

Contribution

It provides an extensive comparison of Notip and pARI methods, revealing their relative performance and informing best practices for critical vector choice in fMRI analysis.

Findings

pARI outperforms Notip under recommended settings

Each method has unique advantages and drawbacks

Comparison clarifies method selection for fMRI studies

Abstract

Two permutation-based methods for simultaneous inference on the proportion of active voxels in cluster-wise brain imaging analysis have recently been published: Notip (Blain et al. 2022) and pARI (Andreella et al. 2023). Both rely on the definition of a critical vector of ordered p-values, chosen from a family of candidate vectors, but differ in how the family is defined: computed from randomization of external data for Notip and determined a priori for pARI. These procedures were compared to other proposals in the literature, but an extensive comparison between the two methods is missing due to their parallel publication. We provide such a comparison and find that pARI outperforms Notip if both methods are applied under their recommended settings. However, each method carries different advantages and drawbacks.