Pooling 65 MRI studies, researchers show that the “brain age gap” — how much older your brain looks than your birthday says — is consistently elevated across Alzheimer’s, MCI, multiple sclerosis, Parkinson’s, schizophrenia, stroke and bipolar disorder, ranging from roughly 3.5 to nearly 8 extra years. But the effect varies so wildly between studies that the authors caution it is, for now, a promising group-level marker rather than a personal diagnostic.
There is a number hidden inside a routine brain MRI that may matter more than your chronological age: your brain age. By training machine-learning models on tens of thousands of healthy scans, researchers can predict how old a brain “looks.” Subtract a person’s real age, and you get the brain age gap (BAG). A positive gap means the brain appears prematurely old; a negative gap suggests resilience.
A new meta-analysis from the University of Queensland, published in GeroScience, asks a deceptively simple question: is this gap a genuinely transdiagnostic signature of brain aging, or just noise that looks different in every disease? The team pooled 65 MRI studies — drawing on giant resources like the UK Biobank, ADNI and OASIS — spanning seven major neurological and psychiatric conditions.
The headline is that the brain age gap is elevated in every single disorder examined. Multiple sclerosis topped the list, with brains appearing on average about 7.8 years older than healthy controls (a large statistical effect). Alzheimer’s followed at roughly 5.6 years, then bipolar disorder (~4.8 years) and stroke (~4.6 years), schizophrenia (~4.4 years), mild cognitive impairment (~3.8 years), and Parkinson’s (~3.5 years). Crucially, in the neurodegenerative diseases the gap did not grow with chronological age — implying the signal reflects disease-driven damage, not normal aging. In schizophrenia, by contrast, the gap widened with age, hinting at a slow, cumulative neurobiological toll.
That is the seductive “big idea”: one MRI-derived number, computed the same way regardless of diagnosis, that tracks brain vulnerability across the entire neuro-psychiatric map. It raises the prospect of using BAG as a universal yardstick — for early detection, for staging disease, and as an endpoint to test whether drugs or lifestyle changes actually slow brain aging.
But the authors are unusually disciplined about the catch. Between-study variability was enormous — statistically, 92–98% of the variation in most disorders came from genuine differences between studies, not chance. When they calculated “prediction intervals” (the range a future study would likely fall in), those intervals crossed zero for Alzheimer’s, MCI, MS and schizophrenia. In plain terms: although the average points firmly upward, the next individual cohort could in principle show almost no gap at all. Only Parkinson’s held a gap reliably above zero. The verdict: a real, reproducible group-level phenomenon, but not yet a number you can hand to an individual patient.
Actionable Insights
Be clear-eyed: this paper tests no treatment. It quantifies a biomarker. So the “actionable” content is indirect, and the only hard effect sizes it offers are the disease-vs-healthy gaps themselves:
- MS: +7.81 years (Hedges’ g = 0.89, large)
- AD: +5.57 years (g = 0.66, medium-to-large)
- Stroke: +4.63 years (descriptive, 2 cohorts)
- BD: +4.84 years (descriptive, 1 cohort)
- SCZ: +4.40 years (g = 0.42, small-to-medium)
- MCI: +3.77 years (g = 0.43, small-to-medium)
- PD: +3.52 years (g = 0.37, small)
The real-world takeaway: a brain disorder can add the structural equivalent of half a decade or more of aging — and in MS the gap reaches a striking ~18 years in the worst cohort.
For prevention, the paper names (but does not test) modifiable levers it considers mechanistically plausible: aerobic exercise (linked to BDNF and neuroplasticity), Mediterranean/MIND diets, sleep, and avoiding fine-particulate air pollution, plus cognitive reserve from education. No effect size is provided for any of these — they remain hypotheses for future causal studies. Treat them as reasonable bets, not proven BAG-reducers. [Confidence: Low for lifestyle claims; High for the disease-gap magnitudes.]
Source:
- Open Access Paper: Aging beyond diagnosis: the MRI brain age gap across disorders
- Institution: The University of Queensland (lead), Brisbane, Australia. Co-affiliations: Isra University (Amman, Jordan) and Helmholtz-Zentrum Dresden-Rossendorf (Leipzig, Germany).
- Country: Australia (lead).
-
Journal: GeroScience (Springer; official journal of the American Aging Association).
Impact Evaluation: Recent metrics for GeroScience: Journal Impact Factor ≈ 5.3–5.4 and CiteScore ≈ 8.3 (Q1 in Geriatrics & Gerontology). The impact score of this journal is ~5.3 (JIF) / 8.3 (CiteScore), evaluated against a typical high-end range of 0–60+ for top general science, therefore this is a Medium-impact journal