The erring brain: Error‐related negativity as an endophenotype for OCD—A review and meta‐analysis

Abstract

Obsessive‐compulsive disorder (OCD) is a complex and heterogeneous disorder that is associated with high personal and societal costs. Feelings of doubt, worry, and repetitive behavior, key symptoms of OCD, have been linked to hyperactive error signals in the brain. The error‐related negativity (ERN) represents a validated marker of error processing in the ERP. Increased ERN amplitudes in OCD have been reported very robustly over the last 20 years. This article integrates results from 38 studies analyzing the ERN in OCD, using a quantitative meta‐analysis. Meta‐regressions were used to examine potential moderators such as task type, symptom severity, age, and sample size. The meta‐analysis reveals a robust increase of ERN in OCD patients compared to healthy participants in response‐conflict tasks (SMD −0.55) that is not modulated by symptom severity and age. No increase in ERN in OCD was observed in tasks that do not induce response conflict (SMD −0.10). In addition to the meta‐analysis, the current article reviews evidence supporting that increased ERN amplitudes in OCD fulfill central criteria for an endophenotype. Further, the specificity of increased ERN amplitudes for OCD and its suitability as a potential transdiagnostic endophenotype is discussed. Finally, the clinical utility and clinical applications are examined. Overall, the evidence that increased ERN amplitudes represent a promising endophenotype indicating vulnerability for OCD is compelling. Furthermore, alterations in ERN are not limited to OCD and may constitute a transdiagnostic endophenotype. Altered neural error signals might serve as a diagnostic or predictive marker and represent a promising target for interventions.

Null results from a dimensional study of error-related negativity (ERN) and self-reported psychiatric symptoms

“…In terms of statistical power, our study has one of the highest sample numbers (N = 196) investigating the ERN in psychiatry. Despite the ERN's link to psychiatry for two decades (Gillan et al, 2017;Olvet and Hajcak, 2009;Riesel, 2019;Weinberg et al, 2015a), there are only five studies with total N > 150 to date (Hanna et al, 2018(Hanna et al, , 2016Meyer and Klein, 2018;Riesel et al, 2019;Weinberg et al, 2015b). Our data suggest that in order to have 80% power to detect a significant association between OCD symptoms and the ERN in a future study, a sample size of N = 729 is required.…”

“…This year, several authors have highlighted the potential for a transdiagnostic framework at reconciling the broad range of ERN patterns in the literature (Gillan et al, 2017;Pasion and Barbosa, 2019;Riesel, 2019). The present paper is timely, being the first study to apply an expansive and empirically robust transdiagnostic approach that directly addresses the issue of co-occurring symptoms in a large sample.…”

“…Contextualising these data with the broader literature, ERN abnormalities may be more sensitive to the categorical comparison of patients versus controls than dimensional variation in the general population. Several OCD patient studies did not find any correlation with symptom severity and ERN amplitude within patient groups (Carrasco et al, 2013a;Endrass et al, 2008;Riesel, 2019;Riesel et al, 2017Riesel et al, , 2014. The ERN remains elevated in OCD despite successful treatment (Hajcak, 2006;Ladouceur et al, 2018;Schrijvers et al, 2009) and elevations are also observed in unaffected first-degree relatives of patients (Carrasco et al, 2013a;Riesel et al, 2019).…”

“…Likewise, the trend was for alcohol addiction to be associated with a blunted ERN, consistent with the previous literature. Recent reviews add support to this conclusion: two meta-analyses noted that overall effect sizes for anxiety or OCD traits for enhanced ERN were relatively small (Cavanagh and Shackman, 2015;Pasion and Barbosa, 2019) and another that assessed the effect size of ERN amplitude shifts in OCD (Riesel, 2019) noted that larger effect sizes were associated with smaller sample publications, suggesting publication bias.…”

“…The ERN remains elevated in OCD despite successful treatment (Hajcak, 2006;Ladouceur et al, 2018;Schrijvers et al, 2009) and elevations are also observed in unaffected first-degree relatives of patients (Carrasco et al, 2013a;Riesel et al, 2019). As such, the ERN has been couched a psychiatric vulnerability endophenotype (Riesel, 2019). Nonetheless, our individual differences approach should have been able to pick up these trait effects along the continuum of scores, regardless of the subtleties of state-based fluctuations.…”