Global shared-AI network to improve Alzheimer's detection and prediction
Federated Deep Learning to Accelerate Alzheimer's Disease Research
This project builds a secure, shared AI system that learns from brain scans and medical data from people worldwide to help spot and predict Alzheimer's for patients and those at risk.
Quick facts
| Grant type | NIH-funded research |
|---|---|
| Study type | NIH-funded research |
| Funding institution | University of Southern California NIH-funded |
| Lab location | 1 site (Los Angeles, UNITED STATES) |
| Project ID | NIH-11194974 on NIH RePORTER |
What this research studies
Researchers are linking MRI, PET, genetic, and clinical records across international biobanks using a secure federated AI platform so raw data stays at each site. The AI will be trained on over 100,000 scans from sites in India, Japan, the U.S., and Europe to diagnose and subtype dementia and to estimate brain amyloid and tau from less invasive tests. The team will also build models to predict who will decline from normal cognition or mild cognitive impairment to Alzheimer's disease. The distributed approach lets many centers contribute to larger, more diverse models without sharing personal data outside their hospitals.
Who could benefit from this research
Good fit: Ideal candidates are people who have had or can provide brain imaging (MRI or PET), genetic data, and clinical memory evaluations, including those with Alzheimer's, mild cognitive impairment, or concerning memory symptoms.
Not a fit: People without relevant imaging or medical/genetic records, or whose memory problems are due to non-Alzheimer's causes, may not receive direct benefit from this project.
Why it matters
Potential benefit: If successful, this could enable earlier and more accurate diagnosis and better prediction of cognitive decline, helping patients get appropriate care and access to trials sooner.
How similar studies have performed: AI models trained at single centers have shown promise for detecting Alzheimer's from scans, but applying secure federated learning across many international biobanks at this scale is a novel advance.
Where this research is happening
Los Angeles, UNITED STATES
- University of Southern California — Los Angeles, United States (Active)
Researchers
- Principal investigator: Thompson, Paul M — University of Southern California
- Study coordinator: Thompson, Paul M
About this research
- This is an active NIH-funded research project — typically early-stage science, not a clinical trial accepting patient enrollment.
- Some NIH-funded labs run parallel clinical studies or seek volunteers for related work. To check, contact the principal investigator or institution listed above.
- For full project details, budget, and progress reports, visit the official NIH RePORTER page below.