IntroductionOur ability to learn associations between events with rewarding or punishing outcomes plays an important role in guiding our behaviour.1 Of equal importance is our capacity to successfully alter established associations to maximize performance when the environment requires us to change our behaviour. When an event that previously led to reward (event A) currently leads to punishment and when a previously punished event (event B) now yields a reward, the optimal behaviour would be to reverse the initial response tendencies by starting to respond to event B instead of event A. This principle is a driving force behind a form of behavioural adaptation known as response reversal. Response reversal can be defined as a change of behaviour following a reversal of the previously established relationships between events and their reinforcing value.2 Response reversal contributes to flexibility in behaviour, both under social and nonsocial circumstances .3 The amygdala and orbitofrontal cortex (OFC) have been shown to play a crucial role in response reversal.
3,4The amygdala has been linked to the establishment of stimulus-outcome associations, 5 whereas the OFC has been Background: Psychopathy is a severe personality disorder that has been linked to impaired behavioural adaptation during reinforcement learning. Recent electrophysiological studies have suggested that psychopathy is related to impairments in intentionally using information relevant for adapting behaviour, whereas these impairments remain absent for behaviour relying on automatic use of information. We sought to investigate whether previously found impairments in response reversal in individuals with psychopathy also follow this dichotomy. We expected response reversal to be intact when the automatic use of information was facilitated. In contrast, we expected impaired response reversal when intentional use of information was required. Methods: We included offenders with psychopathy and matched healthy controls in 2 experiments with a probabilistic cued go/no-go reaction time task. The task implicated the learning and reversal of 2 predictive contingencies. In experiment 1, participants were not informed about the inclusion of a learning component, thus making cue-dependent learning automatic/ incidental. In experiment 2, the instructions required participants to actively monitor and learn predictive relationships, giving learning a controlled/intentional nature. Results: While there were no significant group differences in acquisition learning in either experiment, the results revealed impaired response reversal in offenders with psychopathy when controlled learning was facilitated. Interestingly, this impairment was absent when automatic learning was predominant. Limitations: Possible limitations are the use of a nonforensic control group and of self-report measures for drug use. Conclusion: Response reversal deficits in individuals with psychopathy are modulated by the context provided by the instructions, according to the distinction between ...