Strategies in probabilistic feedback learning in Parkinson patients OFF medication. (21st April 2016)

• Record Type: Journal Article
• Authors: Bellebaum, C.; Kobza, S.; Ferrea, S.; Schnitzler, A.; Pollok, B.; Südmeyer, M.
• Is Part Of: Neuroscience. Volume 320(2016)
• Page Start: 8
• Page End: 18
• Publication Date: 2016-04-21
• Subjects: BDI Beck Depression Inventory -- DA dopamine -- PD Parkinson's Disease -- RTs reaction times -- Sn substantia nigra -- UPDRS Unified Parkinson's Disease Rating Scale; Parkinson's Disease -- dopamine -- feedback learning -- observation; Neurochemistry -- Periodicals; Neurophysiology -- Periodicals; Neurology -- Periodicals; Neurochimie -- Périodiques; Neurophysiologie -- Périodiques; Neurochemistry; Neurophysiology; Electronic journals; Periodicals; Electronic journals; 612.8
• DOI: 10.1016/j.neuroscience.2016.01.060
• Languages: English
• ISSNs: 0306-4522

Abstract

Highlights: Parkinson's Disease (PD) patients show more lose-shift behavior than controls. Acquisition in PD patients correlates negatively with lose-shift responses. Acquisition in PD patients correlates positively with win-stay responses. PD patients and controls do not differ on general acquisition. PD patients and controls do not differ on strategies in observational learning. Abstract: Studies on classification learning suggested that altered dopamine function in Parkinson's Disease (PD) specifically affects learning from feedback. In patients OFF medication, enhanced learning from negative feedback has been described. This learning bias was not seen in observational learning from feedback, indicating different neural mechanisms for this type of learning. The present study aimed to compare the acquisition of stimulus–response–outcome associations in PD patients OFF medication and healthy control subjects in active and observational learning. 16 PD patients OFF medication and 16 controls were examined with three parallel learning tasks each, two feedback-based (active and observational) and one non-feedback-based paired associates task. No acquisition deficit was seen in the patients for any of the tasks. More detailed analyses on the learning strategies did, however, reveal that the patients showed more lose-shift responses during active feedback learning than controls, and that lose-shift and win-stay responses more strongly determined performance accuracy in patients than controls. For observational feedback learning, the performance of both groups correlated similarly with the performance in non-feedback-based paired associates learning and with the accuracy of observed performance. Also, patients and controls showed comparable evidence of feedback processing in observational learning. In active feedback learning, PD patients use alternative learning strategies than healthy controls. Analyses on observational learning did not yield differences between patients and controls, adding to recent evidence of a differential role of the human striatum in active and observational learning from feedback.