Two routine blood measurements were linked to an older-looking brain on MRI in a study of 852 adults. The findings offer research clues about body-brain connections, but neither marker is a validated clinical test of brain aging.
Could ordinary blood tests help explain why people of the same age have different patterns of brain structure? A new study in Scientific Reports examined that question using brain scans and 25 laboratory markers.
Higher alkaline phosphatase, or ALP, and a higher neutrophil-to-lymphocyte ratio, or NLR, were associated with a larger MRI-derived brain age gap after adjustment for several measured factors. That does not show that either marker causes the brain to age faster, or that lowering it would protect memory.
What an MRI brain age estimate means
Researchers used records from adults undergoing health examinations at Beijing Tiantan Hospital in 2019 and 2020. The final analysis included 852 people with suitable MRI scans and laboratory data, averaging about 49 years old. About 41% were women.
A deep learning model estimated brain age from structural magnetic resonance imaging, or MRI. Such models learn patterns from scans and estimate the age those patterns resemble. This model had previously been trained on scans from 9,794 healthy people.
The brain age gap is the difference between estimated brain age and chronological age. A positive gap means that the scan looks older than expected according to the model. The researchers also corrected for the tendency of prediction models to overestimate younger people’s ages and underestimate older people’s ages.
A brain age estimate is not a direct measurement of aging speed. A scan at one point in time cannot show how quickly that person’s brain is changing. Nor is an older-looking scan, by itself, a diagnosis of dementia.
ALP: a signal with several possible sources
Alkaline phosphatase is an enzyme, a protein that helps chemical reactions occur. Blood ALP can reflect processes involving the liver, bile ducts and bone, among other tissues. It is not specific to the brain.
In the fully adjusted analysis, a one-standard-deviation increase in ALP was associated with approximately 0.41 years higher corrected brain age gap. A standard deviation describes the spread of measurements within this study population. The 95% confidence interval ran from about 0.15 to 0.66 years.
This is an average statistical association, not a rule for converting an individual’s laboratory result into brain age. ALP was also associated with greater odds of exceeding the study’s brain age gap thresholds.
The authors discuss possible connections involving blood vessel mineralization, small-vessel injury and biological pathways involving tau, a protein relevant to some brain disorders. These are possible explanations informed by earlier work. This study did not demonstrate those mechanisms in its participants.
NLR: an immune-system clue
Neutrophils and lymphocytes are two types of white blood cell involved in immune defense. NLR is calculated by dividing the neutrophil count by the lymphocyte count. A higher ratio can accompany changes in immune activity or inflammation, but it is not a precise measure of inflammation inside the brain.
A one-standard-deviation increase in NLR was associated with approximately 0.26 years higher corrected brain age gap after adjustment. Its 95% confidence interval was about 0.02 to 0.50 years, and the statistical evidence was weaker than for ALP.
NLR was also associated with the stricter threshold of a brain age gap greater than three years. However, analyses using NLR cutoffs of 2 or 3 did not clearly separate risk after adjustment. The paper therefore does not establish a clinically useful NLR cutoff for brain aging.
Why inflammation and metabolism enter the picture
The brain depends on blood vessels and on the body’s handling of energy and nutrients. Immune activity, metabolic health and vascular health can overlap, giving researchers reasons to study signals in the blood alongside brain structure.
The authors propose pathways involving inflammation, blood vessel function and the blood-brain barrier, the selective boundary between circulating blood and brain tissue. These explanations remain hypotheses here. Measuring NLR in blood does not prove that a particular inflammatory pathway is damaging the brain.
Other metabolic measures showed initial associations or differences between groups, but several weakened with fuller adjustment. For example, the triglyceride-glucose index, a measure related to metabolic health, was not significantly associated with corrected brain age gap in the fully adjusted model.
Important limits of this study
This was a retrospective, cross-sectional study. Blood measurements and brain scans were compared without tracking subsequent brain changes. It cannot establish which changes happened first or whether one caused the other.
The models accounted for age, sex, body mass index, hypertension history, blood pressure and selected laboratory markers. But detailed smoking, alcohol, exercise and diet information was not collected. These unmeasured factors could help explain the associations. Medication history and other recorded medical information also had limitations.
The study involved a single Asian health-examination cohort. Selection related to who had available tests remains possible, and the results need independent replication. Brain age estimates also carry model uncertainty. The model’s reported average absolute error in its development and validation datasets was 2.63 years, which is different from the small average associations found here.
The publisher identifies this as a peer-reviewed article in press. The accepted paper is citable, but its final edited version may change.
What readers can take from it
ALP and NLR may help researchers investigate how systemic health relates to brain structure. They are not ready to tell an individual how old their brain is, and this study did not validate a screening tool or test an intervention.
It provides no evidence that changing either marker prevents brain aging, reverses biological age or reduces dementia risk. Longitudinal studies, independent validation and mechanistic research are needed before these associations can support clinical decisions.
Original research
Zhong, Y., Chai, L., Jiang, J. and colleagues. Peripheral serum biomarkers and MRI-derived accelerated brain aging: a cross-sectional study using deep learning-based brain age estimation. Scientific Reports. Published October 12, 2026. Read the original study, including its freely accessible accepted manuscript PDF.
Featured image: AI-generated conceptual editorial illustration, not an MRI image or a figure from the study.