Unbiased ROI selection in neuroimaging studies of individual differences

Individual differences studies are those that investigate the interaction between a subject level covariate, such as sex, age, or performance, and the effect of an experimental task. Commonly, a brain region is selected on the basis of the task effect, and the signal in this region correlated with individual covariates. It is shown here that, provided that data are identically and independently distributed between subjects, the selection of the region on the basis of the task effect is unbiased in two cases: when selection is based on a one-sample t test of the task effect, or when the subject level covariate is centered. This result is meant to contribute to clarifying when using the same data for ROI selection and testing leads to valid tests in studies of individual differences.