Machine learning reveals distinct neuroanatomical signatures of cardiovascular and metabolic diseases in cognitively unimpaired individuals.

Pubmed ID: 40108173

Pubmed Central ID: PMC11923046

Journal: Nature communications

Publication Date: March 19, 2025

Affiliation: ['Laboratory of Behavioral Neuroscience, National Institute on Aging, National Institutes of Health, Baltimore, MD, USA.']

MeSH Terms: Humans, Male, Female, Aged, Aged, 80 and over, Cardiovascular Diseases, Risk Factors, Cohort Studies, Middle Aged, Magnetic Resonance Imaging, Diabetes Mellitus, Type 2, Cognition, Brain, Metabolic Diseases, White Matter, Machine Learning, Atrophy

Grants: RF1 AG054409, R01 AG067103, R01 AG083865, R01 AG085571, U24 NS130411, U19 AG033655, R01 AG080821, R01 AG054409, P30 AG066507, R01 AG062819

Authors: Launer LJ, Davatzikos C, Erus G, Doshi J, Bryan RN, Albert MS, An Y, Abdulkadir A, Nasrallah IM, Mamourian E, Wolk DA, Resnick SM, Shou H, Govindarajan ST, Melhem R, Tosun D, Bilgel M, Marcus DS, LaMontagne P, Benzinger TLS, Sotiras A, Espeland MA, Masters CL, Maruff P, Fripp J, Johnson SC, Morris JC, Habes M, Pomponio R

Cite As: Govindarajan ST, Mamourian E, Erus G, Abdulkadir A, Melhem R, Doshi J, Pomponio R, Tosun D, Bilgel M, An Y, Sotiras A, Marcus DS, LaMontagne P, Benzinger TLS, Espeland MA, Masters CL, Maruff P, Launer LJ, Fripp J, Johnson SC, Morris JC, Albert MS, Bryan RN, Resnick SM, Habes M, Shou H, Wolk DA, Nasrallah IM, Davatzikos C. Machine learning reveals distinct neuroanatomical signatures of cardiovascular and metabolic diseases in cognitively unimpaired individuals. Nat Commun 2025 Mar 19;16(1):2724.

Studies:

Abstract

Comorbid cardiovascular and metabolic risk factors (CVM) differentially impact brain structure and increase dementia risk, but their specific magnetic resonance imaging signatures (MRI) remain poorly characterized. To address this, we developed and validated machine learning models to quantify the distinct spatial patterns of atrophy and white matter hyperintensities related to hypertension, hyperlipidemia, smoking, obesity, and type-2 diabetes mellitus at the patient level. Using harmonized MRI data from 37,096 participants (45-85 years) in a large multinational dataset of 10 cohort studies, we generated five in silico severity markers that: i) outperformed conventional structural MRI markers with a ten-fold increase in effect sizes, ii) captured subtle patterns at sub-clinical CVM stages, iii) were most sensitive in mid-life (45-64 years), iv) were associated with brain beta-amyloid status, and v) showed stronger associations with cognitive performance than diagnostic CVM status. Integrating personalized measurements of CVM-specific brain signatures into phenotypic frameworks could guide early risk detection and stratification in clinical studies.