Reproducible brain-wide association studies do not necessarily require thousands of individuals

In brain-wide association studies (BWAS), researchers correlate behavior with the inter-individual variability in functional or structural properties of distinct brain regions. Marek, Tervo-Clemmens, and colleagues (hereinafter, M&TC) empirically assess statistical power, replication rate, type I error, sign error, and effect size inflation in BWAS using data from three large-scale neuroimaging initiatives (i.e., ABCD, Human Connectome Project - HCP -, and UK Biobank). Their results indicate that reproducible brain-behavior associations require thousands of observations. Here, leveraging synthetic and HCP data, we demonstrate that their calculations overestimate the sample size needed to detect reproducible effects by one order of magnitude.