Brain age prediction using deep learning uncovers associated sequence variants
Problem
Predicting brain age from structural MRI and identifying genetic sequence variants associated with the gap between predicted and chronological age.
Why it matters
Brain-age gaps signal aging and brain disease, but their genetic basis was largely unknown.
What the AI did
A deep learning model predicted age from T1-weighted MRI; the predicted-age difference was used as a phenotype in a genome-wide association study.
Results
Two variants found: rs1452628-T (β=-0.08, P=1.15×10⁻⁹) and rs2435204-G (β=0.102, P=9.73×10⁻¹²); discovery N=12,378, replication N=4,456.
Impact
Peer-reviewed with an independent replication dataset; 394 citations indicate research impact.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2019 |
| Work type | article |
| Authors | Benedikt A. Jónsson, Gyða Björnsdóttir, Thorgeir Elís Thorgeirsson, Lotta M. Ellingsen, G. Bragi Walters, Daniel Fannar Gudbjartsson, Hreinn Stefánsson, Kāri Stefánsson et al. |
| Citations | 393 (OpenAlex, at import time) |
| References | 93 |
| Model or method | Deep learning (architecture not stated in source) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-019-13163-9 |
| Open access | gold |
| Keywords | brain age, deep learning, MRI, GWAS, genetic variants, UK Biobank |
Importance rating
significant — thousands to millions affected
Predicted age difference is a phenotype relevant to aging and brain disease; identifying its genetic variants advances understanding, but abstract gives no scale of affected populations.