The neurodegenerative process in PD moves along neural networks, possibly reflecting the spread of toxic alpha-synuclein molecules.

To decipher the mechanisms of network-based neurodegeneration in Parkinson’s disease (PD) investigating the relationship between functional network connectivity (FC) in healthy brain connectome and cortical thinning in patients at the early disease stage, and to develop a predictive model for atrophy spreading in PD.

86 early-stage PD patients performed 3D T1-weighted MRI at baseline and every year for 3 years. Resting state functional MRI (rs fMRI) was obtained from 60 age- and sex-matched controls at baseline.The brain was parcellated into 83 regions. Mean volumes for each region were calculated for PD patients and controls on 3D T1-weighted images. Regional percentage of atrophy in PD patients relative to controls was calculated at each time point: [1-(mean patient volumes/mean control volumes)]*100. Functional healthy connectome was estimated from rs fMRI in controls. In PD patients, the “disease exposure” (DE) to pathology of each brain region at each time point was defined as a function of FC of the region with the whole brain in healthy connectome and the severity of atrophy of the connected regions (i.e., the sum of the products of FC between i and all the other brain regions j in healthy subjects and atrophy percentage of the corresponding region j in patients). DE during the first and second year of follow up was used to predict atrophy at the subsequent time points.

In early-stage PD patients, regional DE at 1-year follow-up predicted atrophy accumulation at 2-year follow-up relative to baseline. Although atrophy at 3 years was increased relative to 2-year follow-up, its variation did not correlate with DE at any time point.

Our study suggests that disease propagation in PD follows functional network connectivity.