An Unsupervised Machine Learning Framework for Parkinson's Disease Progression Analysis and Subtyping

Tianjie Cheng, Beilei Xu, Wencheng Wu1, Lei Lin, Trevor Richardson, Edgar Bernal

  • 1University of Rochester / Goergen Institute for Data Science

Details

12:15 - 14:15 | Wed 20 November | Upper Foyer Balcony | A1P-E.13

Session: [A1P-E] Poster Session - Early Detection of Disease or Toxicity 1

Category: Poster Session
Theme: Early Detection of Disease or Toxicity

Full Text

Abstract

Parkinson's disease (PD) is a neurological disorder caused by a dopamine deficiency in the brain. Because of the considerable variations in the expression of the symptoms of PD, it is often difficult to pinpoint the underlying disease mechanisms. Subtyping of PD is essential to better understand the disease, predict outcomes of treatments, and design clinical trials. We aim at identifying subgroups of PD patients who might have different intrinsic disease progression rates. We explore different feature representations capturing different aspects of the disease and perform various types of cluster analysis to quantify disease progression relative to motor and cognitive symptoms.