Citation

Ho NCW, Dunlop K. Biol. Psychiatry Cogn. Neurosci. Neuroimaging 2023; 8(4): 347-348.

Copyright

(Copyright © 2023, Society of Biological Psychiatry, Publisher Elsevier Publishing)

DOI

10.1016/j.bpsc.2023.01.005

PMID

37028903

Abstract

Brain aging as a field of study harnesses interindividual variability in structural and/or functional correlates of aging. In most studies, machine learning models are trained on large normative structural magnetic resonance imaging (MRI) or functional MRI datasets to predict chronological age. The prediction error between an individual's chronological age and their predicted age (the brain-predicted age difference [PAD]) is thought to reflect brain health, with a lower predicted age than the chronological age reflecting slowed brain aging, and vice versa for accelerated aging. Accelerated brain aging is associated with negative health outcomes, including an increased mortality rate and an increased risk of developing dementia, and is repeatedly observed in numerous neurological conditions relative to unaffected control subjects.

Language: en