A woman in her early forties came to see me about word-finding trouble. She runs a busy practice, has two school-age children, sleeps about six hours on a good night, and every routine laboratory value she brought with her was normal. Her MRI was normal as well, in the sense that a radiologist reading it for stroke or tumor would sign it out without comment, and she had left her previous physician's office with the word "stress" and no further plan. What she wanted to know was whether her brain was aging faster than she was, and for most of my career I have had no honest way to answer that with a number.
What a Brain Aging Index Actually Measures
On September 9, a group at UCLA led by Arpana Church, PhD, co-director of the Goodman-Luskin Microbiome Center, with first author Kanhao Zhao, published a study in eBioMedicine that attempts to supply that number. The team analysed resting-state functional MRI from 1,462 adults across three separate cohorts: a discovery cohort of 674, a replication cohort of 444, and an independent cohort of 344.
Rather than measuring the size or the thickness of brain structures, they measured communication. Whole-brain functional connectivity was computed across a 100-region Schaefer parcellation, and a Bayesian ridge regression model was trained to estimate each person's age from that pattern of communication alone. The Brain Aging Index, which the authors abbreviate as BAI, is the difference between the age the model estimated and the person's chronological age, corrected for the age bias that these models tend to introduce at the extremes.
Predicted brain age correlated with chronological age at r = 0.50 to 0.59 across the three cohorts. That is a moderate correlation, which means the connectivity pattern carries real information about age and also leaves a great deal of the variance unexplained.
What a Higher Brain Aging Index Was Associated With
In all three cohorts, adults whose brains were estimated to be older than their chronological age performed worse on testing of working memory and executive function, and they reported more depressive symptoms. The connectivity patterns that carried the association involved the posterior cingulate cortex, the precuneus, and medial frontal regions.
Those are default mode network hubs. They are the regions engaged during self-referential thought and episodic memory, and they have been among the earliest to show functional change in Alzheimer's disease, before atrophy is apparent on a structural scan. Seeing the same hubs carry a brain-age signal in young and mid-life adults, rather than in the older and already-diagnosed populations where most brain-age work has been done, is the part of this paper that should get a clinician's attention.
The age of the sample is what makes the finding useful, and as the authors note, the relevance of brain age deviation in younger adults has been insufficiently characterised because the bulk of this literature has been built in older cohorts. I would add one caution. These participants should not be assumed to be free of medical diagnoses, since the abstract does not describe the inclusion criteria and the laboratory that produced the work studies disorders of gut-brain interaction.
Where the Gut Microbiome Enters the Picture
In the independent cohort, the investigators also collected stool samples, ran metagenomic sequencing and metabolomics, and used multi-view sparse partial least squares to integrate those profiles with each person's Brain Aging Index. Among the metabolites that carried weight are ceramides, 24-hydroxycholesterol, dicarboxylic acids, and estetrol, the last of these associated inversely with BAI. Pathway enrichment analysis pointed toward neuroimmune signalling, vascular function, synaptic transmission, and mitochondrial energy production.
My interpretation is that three of those four metabolite families are ones a neurologist already has reason to think about. Ceramides are sphingolipids that tend to accumulate with insulin resistance and endothelial dysfunction. Dicarboxylic acids are products of fatty acid omega-oxidation, a pathway that tends to carry more traffic when mitochondrial beta-oxidation is under strain. 24-hydroxycholesterol is the principal form in which the brain exports cholesterol to the periphery.
Estetrol is the outlier, and it deserves caution rather than enthusiasm. It is a native human estrogen produced by the fetal liver during pregnancy, and since 2021 it has also been the estrogen component of an approved combined oral contraceptive, so in a cohort of this age range its appearance may reflect exogenous exposure. It may equally reflect the estrobolome, which is the gut's own machinery for deconjugating estrogens during enterohepatic recycling and the ordinary reason an estrogen turns up in stool at all. A rare steroid annotation in untargeted metabolomics can also simply be a misassignment. I would want this one confirmed against an authentic standard before anyone builds an argument on it.
What the paper does not show is direction. This is a cross-sectional study, the omics came from a single cohort, and nothing in the design can tell us whether gut biology influences brain aging, whether an aging brain reshapes the gut, or whether both are downstream of the same vascular and metabolic problem. The anatomy of that two-way conversation is covered in our earlier post on the vagus nerve and interoception.
What This Does and Does Not Support
The Brain Aging Index is a research instrument. It is not a test you can order, it has no established reference range, and no one has yet shown that changing it changes a clinical outcome. UCLA's announcement of the study carries the phrase "decades before symptoms." That describes the age range of the people who were studied, not a measured interval between a scan and a diagnosis, because no one in this study was followed forward in time.
What the paper does support is narrower and still worth having. The functional organisation of the brain varies between adults of the same chronological age, that variation is reproducible across three cohorts, it tracks with the two domains this study tested, and it carries a peripheral biological signature running through metabolism, inflammation, and blood vessels.
What I Do With This in Clinic Today
I cannot order a Brain Aging Index for the woman I described, and I cannot order any of those metabolites either. What I can order are the ordinary clinical proxies for the same pathways, and this is where a proper workup separates itself from a normal MRI and a reassurance. Fasting insulin and HbA1c tell me whether her brain is working in an insulin-resistant environment. ApoB captures the atherogenic particle burden that LDL cholesterol alone can understate, and lipoprotein(a) is not on a standard panel at all. High-sensitivity CRP speaks to the neuroimmune arm of the enrichment signal, and I add homocysteine on separate clinical grounds rather than because this study implicated it. Six hours of sleep a night in a woman with executive complaints is a finding rather than a lifestyle detail. That combination is the substance of the Intensive Brain Health Program, and it is available now, while the imaging science continues to mature.
Cognition is the asset that produces nearly everything else a person builds, which is the premise underneath the Neuroeconomy. This study suggests that the divergence between a person's brain and their chronological age is measurable well before anything has gone obviously wrong, and that part of what travels with that divergence can be measured in the blood and the gut, where we can already look. The interval between the first measurable change and the first symptom is the interval in which medicine is still inexpensive and still works.
Illustration created with AI image tools.
This article was drafted with the assistance of AI writing tools, then reviewed, edited, and approved by Dr. Sean C. Orr, M.D., who holds full editorial responsibility for its content.
References
- Zhao, K., Vignolle, G. A., Labus, J. S., Mayer, E. A., Vaughan, A., Dy, M., Vora, P., Hung, M. W., Vossel, K., Gill, C., Del Rio, D., Stanton, C., Ross, R. P., Cryan, J. F., Kaddurah-Daouk, R., Zhang, Y., & Church, A. (2026). Brain-gut crosstalk associated with brain ageing in young and mid-life adults: a multicohort cross-sectional study. eBioMedicine, 106468. https://doi.org/10.1016/j.ebiom.2026.106468
- UCLA Health. (2026, September 10). Brain aging may be detected decades before symptoms appear, with links to gut health. https://www.uclahealth.org/news/release/brain-aging-may-be-detected-decades-before-symptoms-appear