Particle.news

Download on the App Store

Single MRI-Scan AI Tool Predicts Brain Aging Rate and Dementia Risk

Validated on over 50,000 worldwide scans, the open-source algorithm offers early risk forecasts for dementia years before symptoms appear

Image
Image
Image
Image

Overview

  • DunedinPACNI was trained on T1-weighted MRI scans of 860 Dunedin Study participants at age 45, linking brain structure to nearly two decades of biomarker-derived Pace of Aging scores
  • Published July 1 in Nature Aging, the tool identified that faster MRI-derived aging rates correspond to a 60% higher dementia risk, an 18% increase in chronic disease likelihood, and a 40% rise in mortality
  • External validation across ADNI, UK Biobank, and BrainLat cohorts confirmed consistent predictive accuracy across diverse socioeconomic, racial, and geographic groups
  • The open-source algorithm is now freely available on GitHub, enabling researchers to analyze existing brain MRI datasets for aging rate assessments
  • Researchers stress that further studies and clinical trials are required to confirm causal pathways and to establish practical clinical applications