A progression analysis of motor features in Parkinson's disease based on the mapper algorithm

Area of Science:

• Neuroscience
• Data Science
• Medical Informatics

Background:

• Parkinson's disease (PD) presents heterogeneous motor and non-motor symptoms.
• Challenges in PD management include predicting disease progression due to symptom variability and lack of reliable markers.

Purpose of the Study:

• To develop a dynamic model for predicting motor progression in early-stage Parkinson's disease.
• To enable quantitative comparison of disease progression based on medication usage.
• To provide an algorithm for predicting individual patient UPDRS III scores.

Main Methods:

• Application of the mapper algorithm, a topological data analysis tool, to Parkinson's Progression Markers Initiative (PPMI) data.
• Construction of a Markov chain on the mapper output graphs to model disease progression.
• Utilizing routinely gathered clinical assessments for model development.

Main Results:

• A novel progression model was developed, enabling quantitative comparison of patient outcomes under different medication regimens.
• An algorithm was created to accurately predict patients' Unified Parkinson's Disease Rating Scale Part III (UPDRS III) scores.
• The model demonstrated the ability to predict motor progression at an individual level in early PD.

Conclusions:

• The developed dynamic model, integrating topological data analysis and clinical assessments, accurately predicts motor progression in early Parkinson's disease.
• This predictive capability assists clinicians in tailoring intervention strategies and identifying suitable candidates for disease-modifying therapy clinical trials.
• The model offers a valuable tool for personalized medicine in Parkinson's disease management.