Flanker task parameters are related to the strength of association between the ERN and anxiety: a meta-analysis

The error-related negativity (ERN)--an index of error monitoring--is associated with anxiety symptomatology. Although recent work suggests associations between the ERN and anxiety are relatively modest, little attention has been paid to how variation in task parameters may influence the strength of ERN-anxiety associations. To close this gap, the current meta-analysis assesses the possible influence of task parameter variation in the Flanker task--the most commonly used task to elicit the ERN--on observed ERN-anxiety associations. Here, we leveraged an existing open database of published/unpublished ERN-anxiety effect sizes, supplementing this database by further coding for variation in stimulus type (letter vs. arrow), response type (one-handed vs. two-handed), and block-level feedback (with vs. without). We then performed meta-regression analyses to assess whether variation in these Flanker task parameters moderated the effect size of ERN-anxiety associations. No evidence for an effect of stimulus type was identified. However, both response type and block-level feedback significantly moderated the magnitude of ERN-anxiety associations. Specifically, studies employing either a two-handed (vs. one-handed) task, or those with (vs. without) block-level feedback exhibited more than a two-fold increase in the estimated ERN-anxiety effect size. Thus, accounting for common variation in task parameters may at least partially explain apparent inconsistencies in the literature regarding the magnitude of ERN-anxiety associations. At a practical level, these data can inform the design of studies seeking to maximize ERN-anxiety associations. At a theoretical level, the results also inform testable hypotheses regarding the exact nature of the association between the ERN and anxiety.