Researchers developed Quest2Dx, a language model that screens for Parkinson's disease using responses to standard self-reported health questionnaires, and identified which individual questions carry the most predictive weight. The approach offers a low-cost, non-invasive route to earlier detection without imaging, biomarkers, or specialist assessment. Earlier identification matters because the window for intervention in neurodegeneration narrows substantially once motor symptoms dominate.
Key Points
- AI model screens Parkinson's from routine self-reported questionnaire responses
- Identifies highest-weight predictive questions, improving interpretability over opaque models
- Low-cost, non-invasive alternative to imaging and specialist evaluation
Longevity Analysis
Parkinson's disease begins years before diagnosis, with sleep disturbance, constipation, reduced smell, and mood changes often preceding tremor by a decade or more. A model that extracts diagnostic signal from questions patients already answer turns ordinary self-report into a surveillance tool, catching subtle shifts in movement, digestion, and autonomic regulation while they remain reversible or modifiable. The constraint shifts from access to technology toward the discipline of repeated, honest self-assessment over time.
Original published by Nature Aging, by Diego Machado Reyes.