Objective: To characterize differences in choice patterns, and differentiate the effects of dopaminergic (DA) treatment on motor and cognitive function, during dynamic decision-making in Parkinson’s disease (PD), with the aim of developing behavioral biomarkers of impulsivity and advancing therapeutic outcomes.

Background: During strategic games, i.e., Rock-Paper-Scissors (RPS), each player’s actions and associated outcomes change dynamically based on their opponent’s actions. Action selection during RPS requires inhibition of one action in favour of another; and impulsivity results in a tendency to produce habitual actions prematurely, without adequately computing reward history. In reinforcement learning models, a reward prediction error signal drives learning the value of actions, dependent on midbrain DA neurons, and PD patients show decreases in striatal prediction error signalling. Further, in PD, DA treatment increases learning rates in reinforcement guided tasks, potentially contributing to impulse control disorders.

Methods: Patients with stage 1-3 PD and age-matched controls played a visuosaccadic version of RPS against a computer opponent that exploited biases in player’s choice patterns. Players maximized reward by minimizing bias (i.e., choosing stochastically). Results were contrasted with a control task wherein reward was linked to a specific visual stimulus, but was similar to RPS in terms of sensory input, motor output, and reward rate, allowing us to dissociate cognitive from motor impairment during RPS. Choices were made with either a saccade or button press. Both groups completed 2 sessions; patients both on- and off- DA medication.

Results: Patients were more variable than controls, had faster reaction times, and showed a directional bias during RPS, which was exacerbated in saccade trials, and did not differ on- and off- medication. Overall reward rate was similar in patients and controls during RPS, but tended to be higher in patients off- than on- medication.

Conclusions: We propose a novel paradigm to probe decision-making deficits in PD. Computational modeling of choice patterns in PD is critical to understanding how medication modulates learning rates during RPS, and may provide insight into the incidence of impulse control disorders.

Preliminary data & analysis for the current project was presented at the Society for Neuroscience’s 45th Annual Meeting in Chicago, IL., October, 2015.

« Back to 2016 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/impairment-of-decision-making-processes-in-patients-with-parkinsons-disease-during-a-strategic-game/