Abstract
Germani et al. (2025, Imaging Neuroscience) argue that pipeline variability inflates false positive rates in between-group mega-analyses that "compare populations whose data were processed differently at the subject level." While such variability is a confounder in this specific and rare type of analysis, it does not confound effects of interest in more common conventional mega-analyses, which test non-zero effects across studies. Instead, pipeline differences strengthen the generalizability of discoveries and protect against idiosyncratic pipeline-induced artifacts in conventional analyses. Even in between-group mega-analyses, it is context-dependent whether pipeline-induced differences are false positives. Recognizing these complexities will help understand when pipeline variability is advantageous and when it can cause problems.