Impaired Structural Connectivity In Parkinson's Disease Patients With Mild Cognitive Impairment: A Study Based On Probabilistic Tractography

Abstract

Background: Probabilistic tractography, in combination with graph theory, has been used to reconstruct the structural whole-brain connectome. Threshold-free network-based statistics (TFNBS) is a useful technique to study structural connectivity in neurodegenerative disorders; however, there are no previous studies using TFNBS in Parkinson's disease (PD) with and without mild cognitive impairment (MCI). Methods: Sixty-two PD patients, 27 of whom classified as PD-MCI, and 51 healthy controls (HC) underwent diffusion-weighted 3T MRI. Probabilistic tractography, using FSL, was used to compute the number of streamlines (NOS) between regions. NOS matrices were used to find group differences with TFNBS, and to calculate global and local measures of network integrity using graph theory. A binominal logistic regression was then used to assess the discrimination between PD with and without MCI using non-overlapping significant tracts. Tract-based spatial statistics (TBSS) were also performed with FSL to study changes in fractional anisotropy (FA) and mean diffusivity (MD). Results: PD-MCI showed 37 white matter (WM) connections with reduced connectivity strength compared to HC, mainly involving temporo-occipital regions. These were able to differentiate PD-MCI from PD without MCI with an area under the curve of 83-85%. PD without MCI showed disrupted connectivity in 18 connections involving fronto-temporal regions. No significant differences were found in graph measures. Only PD-MCI showed reduced FA compared with HC. Discussion: TFNBS based on whole-brain probabilistic tractography can detect structural connectivity alterations in PD with and without MCI. Reduced structural connectivity in fronto-striatal and posterior corticocortical connections is associated with PD-MCI.