Abstract
Voluntary eye movements are sensitive to reward contingencies (e.g. Madelain et al, 2011). Here we used smooth pursuit eye movements to investigate responses to visual targets associated with different probabilities of monetary gain or loss. Smooth pursuit allows a continuous read-out of processing cognitive information, such as reward, from the earliest phase of the response prior to target onset (anticipatory pursuit) to visually-guided steady-state pursuit of a selected target. In a novel task, inspired by the Iowa Gambling Task (IGT, Bechara et al., 1994), observers had to fixate in the screen center while two targets moved towards fixation from different directions. Once the targets reached fixation, observers had to select one and track it with their eyes. Importantly, in the main experiment, each target's direction was associated with a different stochastic reinforcement rule, either advantageous, yielding an overall gain across trials, or disadvantageous, yielding an overall loss. In a control experiment, the target was explicitly instructed (e.g. "Follow the black target") with no association between target selection and reward. Participants were patients diagnosed with Parkinson's disease (PD), a neurodegenerative movement disorder that is also frequently associated with impaired ability to assess risk, tested both ON and OFF medication. We also tested age-matched and young healthy controls. For all groups, choice latency (the delay after which the oculomotor target selection becomes evident) was clearly shortened in the IGT-pursuit task compared to the control-task. Moreover, eye movements deviated toward the selected target direction already in the anticipatory phase. However, early visually-guided smooth pursuit underwent a significantly stronger bias toward the selected direction in young controls than in PD patients the age-matched controls. The analysis of target selection strategy with respect to the reinforcement rule revealed a consistent impairment in decision-making for PD patients.
Meeting abstract presented at VSS 2018