Replication of fMRI group activations in the neuroimaging battery for the Mainz Resilience Project (MARP).

ABSTRACT

Motivated by the recent replicability crisis we tested replicability of functional magnetic resonance imaging (fMRI) group activations in two independent samples. An identical behavioral and fMRI test battery for the longitudinal investigation of stress resilience mechanisms was developed for the Mainz Resilience Project (MARP) and conducted in a discovery (N = 54) and a replication sample (N = 103). The test battery consisted of a stress reactivity task, a reward sensitivity task, a fear conditioning and extinction paradigm, two volitional reappraisal tasks and an emotional interference inhibition task. Replicability of group activations was tested with the Jaccard index and the Intra Class Correlation (ICC). Overall, we observed good to excellent replicability of activations at the whole brain level. Only a minority of contrasts showed unsatisfactory replicability. Replicability at the level of individual regions of interest (ROIs) was generally lower. Tasks with stronger activation in the discovery sample showed better replicability.