Abstract:Background: The substantial heterogeneity of clinical symptoms and lack of reliable progression markers in Parkinson's disease (PD) present a major challenge in predicting accurate progression and prognoses. Increasing evidence indicates that each component of the neurovascular unit (NVU) and blood-brain barrier (BBB) disruption may take part in many neurodegenerative diseases. Since some portions of CSF are eliminated along the neurovascular unit and across the BBB, disturbing the pathways may result in changes of these substances. Methods: Four hundred seventy-four participants from the Parkinson's Progression Markers Initiative (PPMI) study (NCT01141023) were included in the study. Thirty-six initial features, including general information, brief clinical characteristics and the current year's classical scale scores, were used to build five regression models to predict PD motor progression represented by the coming year's Unified Parkinson's Disease Rating Scale (MDS-UPDRS) Part III score after redundancy removal and recursive feature elimination (RFE)-based feature selection. Then, a threshold range was added to the predicted value for more convenient model application. Finally, we evaluated the CSF and blood biomarkers' influence on the disease progression model. Results: Eight hundred forty-nine cases were included in the study. The adjusted R 2 values of three different categories of regression model, linear, Bayesian and ensemble, all reached 0.75. Models of the same category shared similar feature combinations. The common features selected among the categories were the MDS-UPDRS Part III score, Montreal Cognitive Assessment (MOCA) and Rapid Eye Movement Sleep Behavior Disorder Questionnaire (RBDSQ) score. It can be seen more intuitively that the model can achieve certain prediction effect through threshold range. Biomarkers had no significant impact on the progression model within the data in the study. Conclusions: By using machine learning and routinely gathered assessments from the current year, we developed multiple dynamic models to predict the following year's motor progression in the early stage of PD. These methods will allow clinicians to tailor medical management to the individual and identify at-risk patients for future clinical trials examining disease-modifying therapies.

Bridging Imaging and Clinical Scores in Parkinson's Progression via Multimodal Self-Supervised Deep Learning

An Automated Hybrid Approach Via Deep Learning and Radiomics Focused on the Midbrain and Substantia Nigra to Detect Early-Stage Parkinson’s Disease

MC-RVAE: Multi-channel recurrent variational autoencoder for multimodal Alzheimer's disease progression modelling

Revealing Patterns of Symptomatology in Parkinson's Disease: A Latent Space Analysis with 3D Convolutional Autoencoders

Parkinson's Disease Classification and Clinical Score Regression via United Embedding and Sparse Learning From Longitudinal Data

Enhancing early Parkinson's disease detection through multimodal deep learning and explainable AI: insights from the PPMI database

A deep generative multimodal imaging genomics framework for Alzheimer's disease prediction

Personalized Progression Modelling and Prediction in Parkinson’s Disease with a Novel Multi-Modal Graph Approach

Deep Neural Networks with Broad Views for Parkinson's Disease Screening

Multimodal and multicontrast image fusion via deep generative models

Deep Learning and Artificial Intelligence Applied to Model Speech and Language in Parkinson's Disease

Optimizing Parkinson's Disease Diagnosis with Multimodal Data Fusion Techniques

Multi-modal Variational Autoencoders for normative modelling across multiple imaging modalities

An interpretable generative multimodal neuroimaging-genomics framework for decoding Alzheimer's disease

Predicting the multi-domain progression of Parkinson’s disease: a Bayesian multivariate generalized linear mixed-effect model

Motor Progression in Early-Stage Parkinson's Disease: A Clinical Prediction Model and the Role of Cerebrospinal Fluid Biomarkers

Early diagnosis and clinical score prediction of Parkinson's disease based on longitudinal neuroimaging data

A generative-discriminative framework that integrates imaging, genetic, and diagnosis into coupled low dimensional space

Predicting Individualized Clinical Measures by a Generalized Prediction Framework and Multimodal Fusion of MRI Data.

Probabilistic prediction of neurological disorders with a statistical assessment of neuroimaging data modalities

Multimodal Predictive Modeling: Scalable Imaging Informed Approaches to Predict Future Brain Health