Reply to Chen et al.: Parametric methods for cluster inference perform worse for two-sided t-tests

TL;DR

This paper demonstrates that parametric methods are less effective for two-sided t-tests in neuroimaging, while non-parametric methods perform consistently across both test types, advocating for the default use of two-sided tests.

Contribution

The study extends previous work on false positive rates to compare parametric and non-parametric methods for one-sided and two-sided tests in neuroimaging.

Findings

Parametric methods perform worse for two-sided t-tests.

Non-parametric methods perform equally well for both test types.

Two-sided tests should be the default in neuroimaging analyses.

Abstract

One-sided t-tests are commonly used in the neuroimaging field, but two-sided tests should be the default unless a researcher has a strong reason for using a one-sided test. Here we extend our previous work on cluster false positive rates, which used one-sided tests, to two-sided tests. Briefly, we found that parametric methods perform worse for two-sided t-tests, and that non-parametric methods perform equally well for one-sided and two-sided tests.