Brain Potentials of Conflict and Error-Likelihood Following Errorful and Errorless Learning in Obsessive-Compulsive Disorder

Anke Hammer,Marcus Heldmann,Andreas Kordon,Thomas F Münte,Bartosz Zurowski,Antonio Verdejo García

doi:10.1371/journal.pone.0006553

Abstract

BackgroundThe anterior cingulate cortex (ACC) is thought to be overacting in patients with Obsessive Compulsive Disorder (OCD) reflecting an enhanced action monitoring system. However, influences of conflict and error-likelihood have not been explored. Here, the error-related negativity (ERN) originating in ACC served as a measure of conflict and error-likelihood during memory recognition following different learning modes. Errorless learning prevents the generation of false memory candidates and has been shown to be superior to trial-and-error-learning. The latter, errorful learning, introduces false memory candidates which interfere with correct information in later recognition leading to enhanced conflict processing.Methodology/Principal FindingsSixteen OCD patients according to DSM-IV criteria and 16 closely matched healthy controls participated voluntarily in the event-related potential study. Both, OCD- and control group showed enhanced memory performance following errorless compared to errorful learning. Nevertheless, response-locked data showed clear modulations of the ERN amplitude. OCD patients compared to controls showed an increased error-likelihood effect after errorless learning. However, with increased conflict after errorful learning, OCD patients showed a reduced error-likelihood effect in contrast to controls who showed an increase.Conclusion/SignificanceThe increase of the errorlikelihood effect for OCD patients within low conflict situations (recognition after errorless learning) might be conceptualized as a hyperactive monitoring system. However, within high conflict situations (recognition after EF-learning) the opposite effect was observed: whereas the control group showed an increased error-likelihood effect, the OCD group showed a reduction of the error-likelihood effect based on altered ACC learning rates in response to errors. These findings support theoretical frameworks explaining differences in ACC activity on the basis of conflict and perceived error-likelihood as influenced by individual error learning rate.

Highlights

In addition to other characteristic symptoms of obsessive compulsive disorder (OCD), such as chronic doubt, repetitive controlling, ruminations, and reduced behavioral flexibility [1], a number of neuropsychological studies have revealed altered memory functions [2,3,4,5,6,7] and executive dysfunction [8,9,10,11,12,13,14]
Performance measures Signal detection measures revealed that memory accuracy was significantly better for EL compared to errors in conflicting (EF)-learning in both groups (Table 4)
Both groups showed better performance measures for the EL learning compared to EF learning and no clear differences could be found between groups

Summary

Introduction

In addition to other characteristic symptoms of obsessive compulsive disorder (OCD), such as chronic doubt, repetitive controlling, ruminations, and reduced behavioral flexibility [1], a number of neuropsychological studies have revealed altered memory functions [2,3,4,5,6,7] and executive dysfunction [8,9,10,11,12,13,14]. Irrespective of the correctness of the response, Heldmann et al [31] observed an ERN for items classified as learned before (i.e. YES-responses: hits and false alarms) as compared to items classified as not learned before (i.e. NO-responses: correct rejections and misses) These results lead to the argument that variations of the ERN amplitude in EL/EF-learning might be partially explained by the subjects’ perceived likelihood of making an error. The present study was designed to answer the question which one of the following is true for OCD patients: Either the dysfunctional action monitoring system in OCD patients (1) is overactive resulting in increased ERN amplitudes for all stimuli [50,51,52,53,54,55,57] or (2) show different error-likelihood effects (i.e. increased following EL and decreased following EFlearning) as postulated by the error-likelihood model for individuals with altered error learning rates [48,49]

Materials and Methods

Results

Discussion