Are errors more aversive in an uncertain world? Testing the influence of uncertainty on the error-related negativity in a randomized controlled trial.

Abstract

Overactive error monitoring-as measured by the error-related negativity (ERN)-is a candidate transdiagnostic risk marker for internalizing psychopathology. Previous research reported associations of the ERN and individual differences in intolerance of uncertainty (IU). These findings imply associations between the subconstructs of IU (prospective and inhibitory IU) and the ERN, which we sought to replicate and extend by testing for causal influences that might elucidate specific mechanisms underlying this association. To test associations of uncertainty and the ERN, a preregistered, randomized-controlled design was employed. After measuring the baseline ERN of N = 120 university students, a subsample was randomly assigned to two groups: While an intervention group (n = 30) performed an unsolvable probabilistic "learning" task intended to induce state uncertainty, a passive control group (n = 30) rested. Subsequently, the ERN was assessed again. Self-reported uncertainty was assessed before and after the intervention. To further increase the statistical power of the replication attempt, we performed a correlation analysis (non-preregistered) by including data from two additional samples collected at different study sites. This analysis comprised psychophysiological data from a total of N = 355 participants. Cross-sectionally, no effects of IU on the ERN emerged. Regarding the state uncertainty induction, the intervention group displayed increased self-reported uncertainty after the intervention, but no evidence emerged for ERN alterations attributable to the intervention. The link between individual differences in IU and the ERN might be smaller and less robust than previous findings suggest, reflecting the understudied character of this association. The absence of evidence for mechanistic changes in the ERN due to a successful induction of state uncertainty further questions a link between IU and the ERN. In line with previous studies that linked increased ERN to anxiety, independent of current clinical status, the ERN seems unaffected by short-term changes such as symptom provocations in non-clinical populations.