The brain and body undergo coordinated changes throughout the life span, yet studies of aging have traditionally examined these systems as separate entities. Here we ask how brain health relates to aging and peripheral biomarkers of metabolic and vascular function, including body mass index, blood pressure, and blood biochemistry. We use multivariate pattern learning to identify generalizable patterns of covariance between multi-modal neuroimaging data (structural, functional, diffusion, and arterial spin labeling MRI), demographic, and physiological markers in two large-scale deeply phenotyped datasets: the Human Connectome Project–Aging and UK Biobank. This data-driven approach isolates two principal axes of brain–body associations in both biological sexes. The first axis is driven by the dominant contribution of age. Across multiple brain measures, aging is associated with loss of brain structural integrity and cerebral vascular dysfunction. The second axis is driven by metabolic features, characterized by low high-density lipoprotein cholesterol, elevated body mass index, blood pressure, glycosylated hemoglobin, insulin, glucose, and alanine aminotransferase that predominantly converge on reduced cerebral perfusion. Importantly, the aging and the metabolic axes are independent of each other, meaning that age and metabolic dysfunction have separable influences on the brain. Finally, we show that deviations from a healthy metabolic profile are linked to cognitive deficits, particularly in females. Our study contributes to development of comprehensive translatable biomarkers for brain health assessment, and highlights the importance of metabolic health as a determinant of brain health in aging population.
Farahani et al. (2026) studied this question.