November 18, 2026
Health news, explained with evidence
LIVE

New Machine Learning Framework May Help Predict Parkinson's Disease Progression

🥗 Nutrition 2 min read 👁 3

Researchers at the University of Florida have developed a machine learning framework called DEGAS that uses brain imaging to identify distinct Parkinson's disease subtypes and predict three-year outcomes.

🥗Nutrition

Key points

  • DEGAS framework uses functional MRI and graph theory to classify Parkinson's disease subtypes.
  • The study analyzed baseline scans from 336 early-stage patients and tracked three-year outcomes.
  • External validation is required before this approach can be considered for routine clinical use.

A team of investigators has introduced a novel machine learning framework designed to address the variable progression of Parkinson's disease by identifying distinct patient subgroups using brain connectivity data. Published in Nature Health Sciences, the study developed DEGAS—disease endophenotyping through graph and adjacency metrics—to analyze baseline functional MRI scans from 336 early-stage patients categorized by suspected REM sleep behavior disorder, an indicator of more aggressive disease progression.

The models incorporating graph theory features achieved a balanced accuracy of 0.91 to 0.93, outperforming traditional approaches. Researchers noted that metrics linked to the basal ganglia-thalamic, temporal-limbic, and cerebello-cortical circuits served as the most powerful predictors. Patients identified as positive for suspected REM sleep behavior disorder experienced more severe motor complications, faster autonomic dysfunction, and increased rates of depression, hallucinations, and dyskinesias over a three-year period. However, the study is limited by its reliance on early-stage observational data, and external validation is strictly necessary before any clinical implementation.

What to know: The DEGAS framework successfully mapped baseline network signatures to clinical outcomes in early Parkinson's patients, though it remains in the investigational phase. Readers should consult a qualified physician for personal medical advice and individualized care decisions.

Source Nature Health Sciences — nature.com/articles/s41531-026-01602-6 ↗

Contact the newsroom

Got a tip, a proposal or a partnership question? Write to us and we will reply shortly.