Memory impairment in Parkinson's disease: The retrieval versus associative deficit hypothesis revisited and reconciled

Objective: Our study explored the retrieval deficit and the associative deficit hypotheses of memory impairments in Parkinson's disease (PD). The former supports a memory deficit mediated by attention/executive dysfunctions, whereas the latter hypothesizes a hippocampal memory impairment in PD. Method: We studied 31 controls and 34 PD patients classified as PD with normal cognition (PD-NC; n = 18) and PD with mild cognitive impairment (PD-MCI; n= 16). To test the retrieval deficit hypothesis, we measured the performance in encoding, retention, and recognition in verbal and visual domains; to test the associative deficit hypothesis, we used a specific associative binding measure. Using resting-state functional-MRI, we compared the functional connectivity of different hippocampal subfields between PD patients and controls, and we related it to memory performance. Results: Consistently with the retrieval deficit hypothesis, PD-MCI, and PD-NC, were impaired in free recall encoding and retention in comparison to controls, especially in the visual domain. However, as predicted by the associative deficit hypothesis, PD-MCI and, to a lesser extent, PD-NC, showed also significant associative and binding deficits in cued recall. Notably, PD patients compared to controls did not show structural differences, although they had lower connectivity between the anterior hippocampi and the precuneus/superior parietal cortex. Worse performance in memory was associated with a more severe disruption of the hippocampal connectivity. Conclusions: The pervasive pattern of memory impairment in PD supports both hypotheses. The interplay between the hippocampus, related to associative memory deficits, and the precuneus, related to attentional control, provides a neural signature that reconciles them.