Assessment of Metacognition and Reversal Learning in Parkinson's Disease: Preliminary Results.

Reversal learning (RL) has been widely used for assessment of behavioral adaptation, impulsivity, obsession, and compulsion in healthy controls as well as people suffering from psychiatric and neurological disorders such as Parkinson's disease (PD). Nevertheless, studies addressing high cognitive functions such as metacognition in PD are scarce. Here, we address for the first time the effect of levodopa and PD on metacognition within the framework of a RL paradigm. In agreement with previous reports, PD patients exhibited reversal shifting impairment with respect to healthy controls (CTRL) regardless of medication condition (MED-ON and MED-OFF), which was supported by a well-known model of learning conditioning (Rescorla-Wagner). In spite that we found a significant association between accuracy and decision confidence level for MED-OFF and CTRL, analysis of metacognitive sensitivity assessed by type 2 signal detection theory (SDT) revealed only a significant underperformance for patients without medication (MED-OFF). This finding points toward a non-compromising positive effect of dopaminergic medication on metacognition for PD.