Longitudinal network-based factorization identifies the altered trajectory of motor symptoms in Parkinson's Disease

Parkinson's disease (PD) is the second most common neurodegenerative disease with progressive structural alterations throughout the brain, resulting in motor symptoms which seriously affects patients' daily life. The present study then aimed to explore the progressive changes in grey matter patterns in PD and identify the longitudinal neuroimaging biomarkers which could predict the progressive motor symptoms of PD. Nonnegative Matrix Factorization (NMF) was firstly used to decompose grey matter images into 7 latent factors from 199 healthy samples in OASIS-3, and then the latent factors were validated on an independent dataset (n=163) to verify the stability of the structural factors. Eventually, 7 potential grey matter factors were identified involving: 1) orbitofrontal; 2) supplementary motor and precentral; 3) middle temporal and inferior occipital; 4) pericalcarine and superior occipital; 5) basal ganglia; 6) amygdala/parahippocampal and inferior temporal; 7) cerebellum. Parkinson's patients (n=78, including baseline, 1-year follow-up, and 2-year follow-up data) and healthy controls (n=48) from PPMI were used to find the correlation between factor weights and motor-symptom related MDS-UPDRS scores. The decreasing trend of the factor weights with the increasing of the disease duration was found in first 6 factors. The random forest prediction model demonstrated that Factor 5 & 7 played pivotal roles in longitudinally predicting MDS-UPDRS-Ⅱ scores, whereas Factor 3 & 4 accounted for most change in MDS-UPDRS-Ⅲ. Our research indicated that the NMF factors could capture the progressive alterations of structural architectures in PD and the factor weights were capable of predicting the clinical motor symptoms. This provides new perspectives for exploring the neural mechanisms underlying the disease and future clinical diagnostic and therapeutic approaches.