To investigate structural brain connectome abnormalities in a large population of Parkinson’s disease (PD) patients with (PD-MCI) and without mild cognitive impairment (PD-ncog).We enrolled 170 patients with idiopathic PD and 41 healthy controls. MRI scans were obtained on a 1.5 T Philips Achieva system. The following sequences were acquired: dual-echo turbo spin-echo, three-dimensional T1-Transient Field Echo , and pulsed gradient spin-echo single shot echo-planar (DT MRI). Briefly, T1-weighted images were processed using the Freesurfer suite (V 5.3 http://surfer.nmr.mgh.harvard.edu/), resulting in an 83 grey matter (GM) areas atlas that was used to define the brain nodes for the network analysis. The preprocessing of DT MRI data was performed using FSL (http://www.fmrib.ox.ac.uk/fsl/). White matter (WM) fiber tracts were reconstructed with Diffusion Toolkit/Trackvis (https://www.nitrc.org/projects/trackvis) using the Fiber Assignment by Continuous Tracking (FACT) algorithm. Tract-based spatial statistics (TBSS) version 1.2 (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/TBSS) was used to perform the multi-subject DT MRI analysis, as previously described.1 A brain network can be described as a graph, where the nodes are brain regions and edges represent their connections. Here the nodes were the 83 GM areas segmented by Freesurfer (Figure 1) and the edges were represented by the WM tracts created by the Diffusion Toolkit linking each pair of nodes. An individual brain network was obtained for each subject included in the study.2 Using the individual brain networks of patients and controls, we examined whether PD patient groups showed structural network alterations compared to healthy controls and whether structural network alterations were associated with cognitive deficits. Network based statistic (NBS)3 was used to identify brain network differences among groups. The analysis was performed on network fractional anisotropy (FA) and mean diffusivity (MD) to take into account the structural organization of the tracts in terms of both directionality and presence of barriers to water diffusion. The NBS statistics were performed to compare FA and MD network data between groups at two levels of significance (p<0.01 and 0.05). The largest (or principal) connected component and the smaller clusters of altered connections that were not included in the principal connected component were studied. A corrected p value was calculated for each component using a permutation analysis. The between patients group comparison was repeated including the Unified Parkinson's disease rating scale (UPDRS) III score as a nuisance variable (to adjust for disease severity). We also checked for differences in global number of streamlines (NOS), FA and MD across subjects groups using t test. Correlations of brain connectivity alterations with neuropsychological deficits and motor disability were tested using false discovery rate corrected Pearson’s correlations.Fifty-four PD patients were classified as PD-MCI, while the remaining PD patients were found to be cognitively unimpaired (PD-ncog). A subgroup of 54 PD-ncog patients, matched with PD-MCI subjects for demographical features, was selected to perform the between-patient group comparison, holding the same statistical power of the PD-MCI group. Despite both PD-MCI and PD-ncog patients had microstructural white matter damage at the voxel level, only PD-MCI showed significant network alterations compared to both controls and PD-ncog. Relative to controls, PD-MCI showed a large basal ganglia/frontoparietal network with decreased FA in the right hemisphere and a large subnetwork with increased MD involving similar regions bilaterally (Figure 2). Compared to PD-ncog, PD-MCI showed a principal network with decreased FA including basal ganglia and fronto-temporo-parietal regions bilaterally (adding the UPDRS III as a nuisance variable, altered connections in PD-MCI were still detectable, although at a lower significance) (Figure 3). FA of connections linking basal ganglia, fronto-temporo-parietal, and fronto-occipital regions correlated with visuospatial and executive scores, while correlations between MD networks and neuropsychological performances were more widespread.This study indicates that mild cognitive impairment in PD is associated with decreased FA and increased MD in structural networks connecting fronto-temporo-parietal areas and basal ganglia. On the contrary, PD without cognitive impairment is associated with frontoparietal WM microstructural alterations but not with structural network disruption, which correlated to cognitive deficits in PD. The distribution and pattern of the structural connectome alterations we found suggest that such an approach might allow to identify novel markers of PD-related cognitive impairment. Future longitudinal studies are warranted to confirm these findings and to describe the progression of structural network abnormalities over time in this condition and how they correlate with and/or predict functional network abnormalities and clinical outcome.