AI Reads Alzheimer’s Risk in a Routine Eye Photo

University of Florida researchers trained AI on 62,876 retinal photos from 44,501 people to read 12 Alzheimer’s risk factors, like blood pressure and smoking, from a single eye image. The signals showed up about 8.5 years before diagnosis, though the model flags risk rather than diagnosing the disease.
artificial-intelligence
Author

Kabui, Charles

Published

2026-06-30

Keywords

alzheimers-disease, retinal-imaging, medical-ai, early-detection, deep-learning

Download as PDF

University of Florida researchers trained AI to read ordinary photographs of the retina, the tissue at the back of the eye, for signs of Alzheimer’s risk. Using 62,876 eye photos from 44,501 people in the UK Biobank, a large UK health database, two deep-learning models learned to predict 12 factors tied to the disease from a single image, among them blood pressure, smoking, alcohol use, sleeplessness, and blood sugar. The AI nailed some factors and barely beat a coin flip on others. The telling part: when it marked which pixels mattered, it kept pointing at the optic nerve and retinal blood vessels, the structures a specialist inspects by hand.

Retinal photos are already everywhere. People with diabetes, glaucoma, or cataracts get them routinely, and a checkup for new glasses can capture one. That makes this far cheaper than an MRI, a PET scan, or a spinal tap, the usual ways to probe for Alzheimer’s. The retina is the one piece of the central nervous system a camera can see directly, because it grows from the same tissue as the brain. An everyday eye photo doubles as a low-cost sensor of brain health.

The model flags risk, it does not diagnose. The signals turned up in people an average of 8.5 years before an Alzheimer’s diagnosis, the window when lifestyle changes or treatment can still help. Like Mayo Clinic’s AI reading routine CT scans years early, the shift is from late diagnosis to early warning.

Sources:

Disclaimer: For information only. Accuracy or completeness not guaranteed. Illegal use prohibited. Not professional advice or solicitation. Read more: /terms-of-service

Reuse

GNU GENERAL PUBLIC LICENSE v3.0(View License)

Citation

BibTeX citation:
@misc{kabui2026,
  author = {{Kabui, Charles}},
  title = {AI {Reads} {Alzheimer’s} {Risk} in a {Routine} {Eye} {Photo}},
  date = {2026-06-30},
  url = {https://toknow.ai/posts/ai-retinal-photo-alzheimers-risk-factors-early-detection/},
  langid = {en-GB}
}
For attribution, please cite this work as:
Kabui, Charles. 2026. “AI Reads Alzheimer’s Risk in a Routine Eye Photo.” https://toknow.ai/posts/ai-retinal-photo-alzheimers-risk-factors-early-detection/.