RESUMEN
Optimal behavior depends on the ability to assess the predictive value of events and to adjust behavior accordingly. Outcome processing can be studied by using its electrophysiological signatures--that is, the feedback-related negativity (FRN) and the P300. A prominent reinforcement-learning model predicts an FRN on negative prediction errors, as well as implying a role for the FRN in learning and the adaptation of behavior. However, these predictions have recently been challenged. Notably, studies so far have used tasks in which the outcomes have been contingent on the response. In these paradigms, the need to adapt behavioral responses is present only for negative, not for positive feedback. The goal of the present study was to investigate the effects of positive as well as negative violations of expectancy on FRN amplitudes, without the usual confound of behavioral adjustments. A reversal-learning task was employed in which outcome value and outcome expectancy were orthogonalized; that is, both positive and negative outcomes were equally unexpected. The results revealed a double dissociation, with effects of valence but not expectancy on the FRN and, conversely, effects of expectancy but not valence on the P300. While FRN amplitudes were largest for negative-outcome trials, irrespective of outcome expectancy, P300 amplitudes were largest for unexpected-outcome trials, irrespective of outcome valence. These FRN effects were interpreted to reflect an evaluation along a good-bad dimension, rather than reflecting a negative prediction error or a role in behavioral adaptation. By contrast, the P300 reflects the updating of information relevant for behavior in a changing context.
Asunto(s)
Potenciales Evocados/fisiología , Retroalimentación Psicológica/fisiología , Aprendizaje Inverso/fisiología , Adulto , Análisis de Varianza , Electroencefalografía , Emociones/fisiología , Femenino , Humanos , Masculino , Reconocimiento Visual de Modelos , Estimulación Luminosa , Tiempo de Reacción/fisiología , Factores de Tiempo , Adulto JovenRESUMEN
BACKGROUND: Psychopathy (PP) is associated with a performance deficit in a variety of stimulus-response and stimulus-reinforcement learning paradigms. We tested the hypothesis that failures in error monitoring underlie these learning deficits. METHOD: We measured electrophysiological correlates of error monitoring [error-related negativity (ERN)] during a probabilistic learning task in individuals with PP (n=13) and healthy matched control subjects (n=18). The task consisted of three graded learning conditions in which the amount of learning was manipulated by varying the degree to which the response was predictive of the value of the feedback (50, 80 and 100%). RESULTS: Behaviourally, we found impaired learning and diminished accuracy in the group of individuals with PP. Amplitudes of the response ERN (rERN) were reduced. No differences in the feedback ERN (fERN) were found. CONCLUSIONS: The results are interpreted in terms of a deficit in initial rule learning and subsequent generalization of these rules to new stimuli. Negative feedback is adequately processed at a neural level but this information is not used to improve behaviour on subsequent trials. As learning is degraded, the process of error detection at the moment of the actual response is diminished. Therefore, the current study demonstrates that disturbed error-monitoring processes play a central role in the often reported learning deficits in individuals with PP.