Brain Age

Motivation

The deviation between chronological age and age predicted from neuroimaging data has been identified as a sensitive risk marker of cross-disorder brain changes, growing into a cornerstone of biological age research.

Research Project

Brain Age is a so-called “normative modeling” task, where a normative group is used to fit a model which is then used to generate predictions for samples not contained in the normative group to estimate deviation from the predefined normative group. Specifically for Brain Age there have been found associations between accelerated brain aging and markers of physiological aging (e.g., grip strength, lung function, walking speed), cognitive aging, life risk, and poor future health outcomes including progression from mild cognitive impairment to dementia, mortality, and a range of neurological diseases and psychiatric disorders. We use uncertainty modeling to consider data density and availability for predicting Brain Age. We provide the first model to predict uncertainty aware brain age and provide evidence that uncertainty estimation is required for Brain Age prediction caused by uncertainty confounded effects if not corrected for uncertainty.

Additionally we provide a second model which is not only able to capture the empirical prediction uncertainty, but also provides guarantees about the predicted uncertainty quantiles based on conformal prediction theory. This enables individual evaluation of Brain Age deviation for the first time, as individual uncertainty quantiles are provided.