Dopaminergic medication impairs feedback-based stimulus-response learning but not response selection in Parkinson's disease

Cognitive dysfunction is a feature of Parkinson’s Disease (PD). Some cognitive functions are impaired by dopaminergic medications prescribed to address the movement symptoms that typify PD. Learning appears to be the cognitive function most frequently worsened by dopaminergic therapy. However, this result could reflect either impairments in learning (i.e., acquisition of associations among stimuli, responses, and outcomes) or deficits in performance based on learning (e.g., selecting responses). We sought to clarify the specific effects of dopaminergic medication on i) stimulus-response association learning from outcome feedback and ii) response selection based on learning in PD. We tested 28 PD patients on and/or off dopaminergic medication along with 32 healthy, age- and education-matched controls. In Session 1, participants learned to associate abstract images with specific key-press responses through trial and error via outcome feedback. In Session 2, participants provided specific responses to abstract images learned in Session 1, without feedback, precluding new feedback-based learning. By separating Sessions 1 and 2 by 24 hours, we could distinguish the effect of dopaminergic medication on a) feedback-based learning and response selection processes in Session 1 versus on b) response selection processes without feedback-based learning in Session 2, when performance accuracy across sessions was comparable. PD patients on medication learned stimulus-response associations more poorly than PD patients off medication and controls. Medication did not influence decision performance in Session 2 in the absence of feedback-based learning. We confirm that dopaminergic therapy impairs feedback-based learning in PD, discounting an alternative explanation that warranted consideration.