An artificial intelligence (AI) analysis of genetic, neurological, cardiovascular and other data from multiple studies has found three distinct patterns of brain aging and dementia risk outcomes. Researchers recently identified two subgroups with normal and accelerated brain aging. These findings, published in JAMA Psychiatry, suggest a path to better predict who will develop Alzheimer’s disease and vascular dementia, and could lead to more efficient future research into brain aging and dementia.
The study was conducted by an international collaborative team that included NIA-funded researchers at the University of Pennsylvania and scientists from the NIA Intramural Research Program. The researchers used AI and machine learning methods to analyze more than 20 years of neuroimaging, clinical, and cognitive data from more than 27,000 participants across a range of studies collected by the Image-Based Coordinate System for Aging and Neurodegenerative Diseases (iSTAGING) International Consortium. The researchers investigated how brain structural changes are associated with multiple factors, including genetics, cardiovascular risk, beta-amyloid, cognitive decline, smoking, and white matter hyperintensities (WMHs), a lesion associated with Alzheimer’s disease and cognitive impairment.
The research team identified three distinct patterns of brain aging: typical aging (A1) and two accelerated aging subgroups (A2 and A3) that become particularly evident after age 65. The A1 group tended to have milder brain atrophy, fewer cardiovascular genetic risk factors, and normal amounts of WMH. The average brain age (a measure of the physical changes in the brain that typically occur with age) in the A1 group was several years younger than the participants’ chronological age.
The A2 subgroup had the highest and fastest increasing levels of WMH, as well as higher rates of cardiovascular disease-related genetic risk factors and amyloid plaques in the brain. Participants in the A2 subgroup had brain ages 2-3 years older than their chronological age. The A3 subgroup had more extensive brain atrophy, faster rates of cognitive decline, more intermediate cardiovascular risk factors, and brain ages 3-5 years older than their chronological age.
The researchers saw these results as a first step toward a future system to more accurately predict and classify the trajectory of brain aging and cognitive decline associated with Alzheimer’s disease in research trials and precision medicine. They hope to expand this study to a more diverse sample of participants and over a longer period of time so that they can be further followed up on physical and cognitive outcomes.
This research was supported by NIA grants P30 AG066444, K23 AG063993, P01 AG003991, P01 AG026276, R01 AG080635, R35 AG071916, R01 AG063887, and P30 AG072947.
These activities are linked to NIH AD+ADRD Research Implementation Milestone 9M, “Multietiology Dementias: Establishing Presymptomatic Diagnosis and Biomarkers.”
Reference: Skampardoni Ioanna, et al. Genetic and clinical correlates of AI-based brain aging patterns in cognitively unimpaired individuals. JAMA Psychiatry. 2024; 81(5):456-467. doi:10.1001/jamapsychiatry.2023.5599.
