Alzheimer’s: in people with a family history of the disease, the brain age gap measured by AI widens more rapidly as the disease progresses

A longitudinal study links an Alzheimer’s disease progression score to changes in ‘brain age’ estimated using covariance neural networks.

The difference between the apparent age of the brain, estimated from MRI images, and the actual age of a person has emerged as a promising marker of brain health. Jason Scheffel and Saurabh Sihag use covariance neural networks (VNNs), derived from graph signal processing, to calculate this, which have the advantage of being more interpretable than classical models.

Their study follows people with an increased risk of neurodegeneration over time due to a family history of Alzheimer's disease. The authors cross-reference this brain age gap with the Alzheimer's Progression Score, a multimodal indicator of the progression of the disease.

The result: the higher the risk of progression at the first visit, the faster the brain age gap increased during follow-up visits. For the researchers, this suggests that this marker could help distinguish people whose brains age faster according to their clinical and biological profile within a population at risk.