Experimental boundary conditions of reinstatement-induced return of fear in humans: Is reinstatement in humans what we think it is?

Experimental paradigms used to study reinstatement of fear in humans are characterized by procedural heterogeneity. Reinstatement protocols involve unexpected (re)-presentations of the unconditioned stimulus (USs) after fear extinction training. Here, we address the number of reinstatement USs administered as a potential boundary condition that may explain divergent findings in the field. A sample of 171 participants is exposed to a fear acquisition training, immediate extinction training, and reinstatement test experiment. Three groups differing in the number of reinstatement US are employed: one (n = 57) or four (n = 55) in experimental groups and zero (n = 59) in the control group. We adopt Bayesian statistical approaches beyond classical null hypothesis significance testing (NHST) to qualify evidence for or against this potential methodological boundary condition in reinstatement-induced return of fear. Startle potentiation to the reinstatement administration context was increased for the RI-US_one compared to the RI-US_zero group, supporting the role of context conditioning in reinstatement. This effect was weaker in the RI-US_four group. This, however, did not transfer to responding to conditioned stimuli during the return of fear-test: no evidence for an effect of the number of reinstatement USs (zero, one, four) was observed in behavioral or physiological measures. In sum, our results speak against the number of reinstatement USs as a potential boundary condition in experimentally induced return of fear in humans. This may challenge what we think we know about the reinstatement phenomenon in humans and call for critical reconsideration of paradigms as well as mechanisms that may underlie some reinstatement effects in the literature.