hide
Free keywords:
Cognitive networks; Dorsal attention network; Eigenvector centrality; Meta-analytic co-activation maps; Mild cognitive impairment; Parkinson's disease; Resting-state functional MRI
Abstract:
Mild cognitive impairment (MCI) affects approximately one-third of non-demented Parkinson's Disease (PD) patients. We aimed at investigating the neural correlates of MCI in PD combining multimodal magnetic resonance imaging (MRI) with large-scale data from the literature. We analyzed 31 PD patients and 30 matched controls. The standard neuropsychological assessment of PD-MCI covered memory, attention, executive functions, language and visuospatial abilities. Following validated criteria, 16 patients were classified as showing MCI. Whole-brain functional connectivity and structural volume changes were assessed, respectively, by means of eigenvector centrality (EC) and voxel-based morphometry. To address the involvement of specific functional brain networks, we validated our results by building a meta-analytic co-activation map (MACM) based on the previous literature and then testing its overlap with the parcellation of functional networks derived from 1000 healthy controls. The EC comparison between PD with normal cognition and controls showed a selective decline in interconnectedness in the bilateral lentiform nuclei. Differently, comparing PD with MCI and controls revealed additional changes in non-motor areas. Directly comparing PD with and without MCI, we found a reduced interconnectedness in the bilateral superior parietal lobules and precuneus. No differences in brain volume were detected comparing these patient groups. The MACM and overlap analyses showed that the observed connectivity changes were localized in the hubs of the dorsal attention network. Notably, this aligned with the predominant attention deficit observed in our sample. Overall, functional impairment in the dorsal attention network seems to be the hallmark of MCI due to PD, thus extending previous findings of brain connectivity disruption in non-motor networks.
