Replication Research Symposium

Concerns about the replicability of neuroimaging findings in mental health research have grown in recent years. Factors thought to undermine replicability—such as small effect sizes, limited sample sizes, and a high number of researcher degrees of freedom—are particularly pronounced in this field. Yet, direct replication studies remain extremely rare, and the true state of replicability is largely unknown. The Replicability of Findings in Neuroimaging in Major Depressive Disorder (ReFiNe-MDD) project addresses this gap by conducting 60 replication attempts of semi-randomly selected published findings on gray matter correlates of depression-related variables. To this end we utilize large-scale MRI cohort datasets comprising over 6,000 participants from both clinical and non-clinical samples, covering a broad range of psychological, biological, and lifestyle measures. In this talk, I will present the study design and pilot results from the ReFiNe-MDD project and discuss field-specific challenges we have encountered in replicating neuroimaging findings, related to preregistration, open data, and the definition of replication success in three-dimensional brain results.