Aging beyond diagnosis: the MRI brain age gap across disorders.
Authors
Affiliations (9)
Affiliations (9)
- School of Human Movement and Nutrition Sciences, Faculty of Health, Medicine and Behavioural Sciences, The University of Queensland, St Lucia, Brisbane, QLD, Australia.
- Department of Medical Imaging, Faculty of Allied Medical Sciences, Isra University, Amman, Jordan.
- ARC Training Centre for Innovation in Biomedical Imaging Technology, University of Queensland, Brisbane, Australia.
- Australian Institute for Bioengineering and Nanotechnology, The University of Queensland, Brisbane, QLD, Australia.
- School of Psychology, The University of Queensland, Brisbane, QLD, Australia.
- Reactive Transport Department, Institute of Resource Ecology, Helmholtz-Zentrum Dresden-Rossendorf, Leipzig Research Site, Permoserstr. 15, 04318, Leipzig, Germany.
- School of Human Movement and Nutrition Sciences, Faculty of Health, Medicine and Behavioural Sciences, The University of Queensland, St Lucia, Brisbane, QLD, Australia. [email protected].
- Reactive Transport Department, Institute of Resource Ecology, Helmholtz-Zentrum Dresden-Rossendorf, Leipzig Research Site, Permoserstr. 15, 04318, Leipzig, Germany. [email protected].
- ARC Training Centre for Innovation in Biomedical Imaging Technology, University of Queensland, Brisbane, Australia. [email protected].
Abstract
The brain age gap (BAG), the difference between magnetic resonance imaging-predicted brain age and chronological age, is a proposed marker of neurobiological aging, yet its transdiagnostic significance remains uncertain. This meta-analysis evaluated BAG in Alzheimer's disease (AD), mild cognitive impairment (MCI), multiple sclerosis (MS), Parkinson's disease (PD), schizophrenia (SCZ), stroke, and bipolar disorder (BD) to determine shared and disorder-specific patterns of accelerated brain aging. Sixty-five MRI-based studies were included. For each cohort, BAG was extracted or calculated, and standardized mean differences (Hedges' g) were computed between patient and healthy control groups using random-effects models. Subgroup analyses and meta-regressions were conducted to assess the influence of age and disorder characteristics. Publication bias and heterogeneity were evaluated using Egger's test, I<sup>2</sup> statistics, and trim-and-fill procedures. BAG was significantly elevated across all disorders. The largest effects were observed in MS (mean adjusted BAG = 7.81 years; g = 0.89) and AD (5.57 years; g = 0.66). Meta-analyses of adequately powered disorders showed moderate elevations in SCZ (4.40 years; g = 0.42) and MCI (3.77 years; g = 0.43), while PD showed smaller but significant acceleration (3.52 years; g = 0.37). Descriptively, stroke (2 cohorts; mean adjusted BAG = 4.63 years) and BD (1 cohort; mean adjusted BAG = 4.84 years) also showed elevated BAG, though insufficient study numbers precluded formal meta-analysis for these disorders. BAG was largely independent of chronological age in neurodegenerative disorders but increased with age in SCZ. These findings support BAG as a transdiagnostic indicator of apparent accelerated brain aging in conditions where the meta-analytic evidence base is robust (AD, MCI, MS, PD, SCZ), while stroke and BD remain preliminary owing to limited cohort numbers. The magnitude of BAG varies substantially across conditions, and very high between-study heterogeneity (I<sup>2</sup> > 90% in most disorders) means pooled effect sizes should be interpreted alongside prediction intervals.