Brain connectivity and cognitive impairment in Parkinson’s disease

Parkinson’s disease is a neurodegenerative process with several motor and non-motor manifestations. Among non-motor symptoms, cognitive decline is a major cause of disability, and its neural bases are poorly understood. In recent years, multimodal neuroimaging techniques have proven to be useful tools in the investigation of the bases of cognitive impairment related to neurological diseases. The objective of this thesis was to evaluate the neuroimaging substrates of Parkinson’s disease-related cognitive and neuropsychiatric manifestations through a network approach, using state-of-the art magnetic resonance imaging techniques to assess associated connectivity and structural changes. In this work, two samples of Parkinson’s disease patients and healthy controls underwent neuropsychological evaluation as well as structural and functional magnetic resonance imaging. One of these samples included 121 Parkinson’s disease patients and 49 healthy controls. The data obtained were used in 2 studies addressing the resting-state functional connectivity changes associated with mild cognitive impairment in Parkinson’s disease; one using a graph-theory approach, and the second assessing large-scale intrinsic connectivity networks through an independent-component analysis/dual regression approach. A third was performed to assess connectivity disruptions in frontostriatal circuits associated with the presence of apathy in Parkinson’s disease. And a fourth study was made assessing cortical thickness changes associated with the presence of mild cognitive impairment in Parkinson’s disease. The second sample included the longitudinal evaluation of 17 Parkinson’s disease patients and 15 healthy controls, and the data obtained resulted in two studies regarding progressive cortical thickness and subcortical volumes in early-stage Parkinson’s disease patients. A seventh study, using subjects from both samples, was performed to evaluate microstructural white matter changes (using diffusion tensor imaging) and gray matter changes (through voxel-based morphometry) related to the presence of facial emotion recognition deficits in Parkinson’s disease. We have found that a high percentage of Parkinson’s disease patients had mild cognitive impairment. These patients showed altered patterns of resting-state functional connectivity characterized by the loss of long range connections and strengthening of local connectivity. From a graph-theory perspective, these changes translated as increased small-world coefficients, modularity and clustering coefficients, which correlated with visuospatial/visuoperceptual and memory performance. The study of large-scale networks showed that patients with mild cognitive impairment had reduced connectivity between the dorsal attention network and the right frontoinsular region, and that this reduction was associated with attention/executive deficits. Additionally, parieto-occipital regions that belong to the dorsal attention network showed reduced connectivity with anterior task­positive regions and loss of the normal anticorrelated pattern with the default mode network; furthermore, these posterior regions showed structural degeneration. Finally, these posterior structural and functional changes were associated with the presence of visuospatial/visuoperceptual deficits. The analysis of Parkinson’s disease patients with mild cognitive impairment revealed a pattern of cortical thinning predominating in parieto­occipito-temporal regions. The longitudinal analysis of progressive structural changes in early Parkinson’s disease revealed that these patients had more marked cortical thinning in frontotemporal regions, even before the onset of clinically evident cognitive manifestations. The functional analysis of a recognition memory network likewise showed signs of progressive connectivity changes without overt clinical deterioration. We conclude that different types of cognitive decline in Parkinson’s disease are associated with different patterns of resting-state functional connectivity, structural connectivity and GM structural changes involving distinct neural systems. Different techniques and different conceptual frameworks can provide useful information in the characterization of the neural bases of cognitive deficits associated with Parkinson’s disease.