Objective: To predict longitudinal brain changes in Parkinson’s disease (PD) using an agent-based model.

Background: Considerable evidence suggests that α-synuclein(α-syn) behaves in a prion-like fashion, with misfolded α-syn spread through the brain’s connectome and act as a template inducing further misfolding of normal proteins[1]. Recently we showed that a Susceptible-Infectious-Removed (SIR) metapopulation structured agent-based model, where individual α-syn represents the agent, predicts atrophy in de-novo PD patients using white matter connectivity and gene expression[2].

Method: T1-weighted MRI scans of PD at baseline and after 1, 2, and 4 years of follow-up were obtained from PPMI database. Deformation-based morphometry (DBM) was performed using CAT12 toolbox, and Jacobian maps were computed as an atrophy measure. After quality control, samples of 318, 120, 108, and 85 PD were selected at baseline(61.0 ± 10.0years; 37%females) and after 1, 2, and 4 years of follow-up. Average DBM values were extracted from an atlas of 42 cortical and subcortical ROIs using MarsBaR toolbox. W scores of atrophy values were calculated for each subject at each ROI by standardizing regional deformation for age and gender based on a sample of 157 healthy controls (61.82 ± 10.17years; 34%females). The SIR model was used to simulate regional atrophy at each timestep. Mean differences in W scores were calculated between baseline and each time point, and correlated with the simulated atrophy to assess model fit.

Results: The SIR model replicates the atrophy found at baseline in early PD patients, especially at earlier time steps, with a peak at around T=400 (r=0.7, p=3.46e-05). The 1-year atrophy was also correlated with the data generated by the SIR model, in particular at later timesteps of the simulated spreading process (T= ~3000th , r=0.3, p=0.034). After 2 years, the atrophy observed in early PD patients was replicated by the SIR model, again later during spreading process (T= ~3000th , r=0.3, p= 0.046). In contrast, the correlation between the observed atrophy after 4 years and the simulated atrophy was not significant.

Conclusion: The SIR model replicates the atrophy observed in PD patients at baseline and after 1 and 2 years. The lack of correlation at 4 years indicates that other factors will be incorporated into the model. Further studies should model these, including cell and synaptic death that may slow down protein synthesis and propagation.

References: 1. Walsh DM, Selkoe DJ. A critical appraisal of the pathogenic protein spread hypothesis of neurodegeneration. Nature Reviews Neuroscience. 2016;17(4):251. Pmid:26988744 2. Zheng YQ, Zhang Y, Yau Y, Zeighami Y, Larcher K, et al. Local vulnerability and global connectivity jointly shape neurodegenerative disease propagation. PLOS Biology. 2019 17(11): e3000495.

« Back to MDS Virtual Congress 2020

MDS Abstracts - https://www.mdsabstracts.org/abstract/predicting-longitudinal-atrophy-in-parkinsons-disease-using-sir-model/