Affiliations: [a] U1171, INSERM, Université de Lille, Lille, France
| [b] Service de Neurologie et Pathologie du Mouvement, Centre Hospitalier Regional Universitaire, Lille, France
| [c] Service de Neurophysiologie Clinique, Centre Hospitalier Regional Universitaire, Lille, France
Correspondence:
[*]
Correspondence to: Dr. Nicolas Carriere, Service de Neurologie et Pathologie du Mouvement, Hôpital Roger Salengro, Centre Hospitalier Régional Universitaire de Lille, Rue Emile Laine, F-59037 Lille cedex, France. Tel.: +33 320446730; Fax: +33 320446680; E-mail: nicolas.carriere@chru-lille.fr.
Abstract: Background: Impulse control disorders (ICDs) in Parkinson’s disease (PD) are related to treatment with dopamine agonists, which is thought to deregulate the dopaminergic mesolimbic pathway and impair reward evaluation. EEG studies in healthy controls (HCs) have suggested that the increase in theta power observed after negative outcome is a marker of reward processing. Objective: To compare outcome-locked, event-related spectral perturbation in a gambling task in PD patients with and without ICDs and in HCs. Methods: Twelve PD patients with ICDs, 12 PD patients without ICDs and 14 HCs underwent EEG while performing a gambling task. The groups were compared in terms of (i) the peak EEG power in the theta (4–7 Hz), alpha (8–14 Hz) and beta (15–30 Hz) frequency bands between 200 and 500 ms after the outcome, and (ii) time-frequency plots at Fz, FCz and Cz. Results: Positive outcomes were associated with greater theta power than negative outcomes in patients without ICDs and in HCs, but not in patients with ICDs. Patients with ICDs and HCs displayed greater theta power following unexpectedly high outcomes. HCs displayed greater beta power following high amplitude than low amplitude outcomes, whereas patients with ICD showed the opposite pattern. Conclusions: In PD, ICDs are associated with (i) weaker modulation of frontocentral theta power by reward valence, (ii) greater frontocentral theta power following unexpected, high outcomes, and (iii) a reversal of the effect of risk on beta oscillations. These observations are consistent with an impairment in prediction error computation in the medial prefrontal cortex.
