Reduction of influence of task difficulty on perceptual decision making by STN deep brain stimulation

Abstract

Neurocomputational models of optimal decision making ascribe a crucial role-the computation of conflict between choice alternatives-to the subthalamic nucleus (STN). Specifically, these models predict that deep brain stimulation (DBS) of the STN will diminish the influence of decision conflict on decision making. In this work, patients with Parkinson's disease judged the direction of motion in random dot stimuli while ON and OFF DBS. To induce decision conflict, we varied the task difficulty (motion coherence), leading to increased reaction time (RT) in trials with greater task difficulty in healthy subjects. Results indicate that DBS significantly influences performance for perceptual decisions under high decision conflict. RT increased substantially OFF DBS as the task became more difficult, and a diffusion model best accounted for behavioral data. In contrast, ON DBS, the influence of task difficulty on RT was significantly reduced and a race model best accounted for the observed data. Individual data fits of evidence accumulation models demonstrate different information processing under distinct DBS states. Furthermore, ON DBS, speed-accuracy tradeoffs affected the magnitude of decision criterion adjustment significantly less compared to OFF DBS. Together, these findings suggest a crucial role for the STN in adjusting decision making during high-conflict trials in perceptual decision making.