Study Links AI-Estimated Biological Vascular Age to Subclinical Brain Injury and Cognitive Decline

A study of Framingham Heart Study participants found that an artificial intelligence-based biological vascular age metric correlates with subclinical brain injury and worse neuropsychological performance, suggesting a potential noninvasive biomarker for cognitive risk.

A cross-sectional study published in Arteriosclerosis, Thrombosis, and Vascular Biology evaluated the association of artificial intelligence-based vascular age (AIVA) with brain structure and neuropsychological performance.

AIVA was estimated using a validated convolutional neural network trained to predict carotid-femoral pulse wave velocity from a normalized peripheral pressure waveform.

The analysis included magnetic resonance imaging (MRI) data from 2,313 participants and neuropsychological test data from 3,001 participants, with a mean age of 62 years.

Higher AIVA was statistically associated with worse markers of cerebral small vessel disease, including white matter free water (β of 0.09, P < 0.001) and white matter hyperintensity volume (β of 0.08, P < 0.001).

Higher AIVA was also associated with increased odds of brain infarcts (odds ratio of 1.44, P = 0.04) and poorer neuropsychological performance, including lower global cognition scores (β of -0.07, P = 0.002) and higher odds of depressive symptoms (odds ratio of 1.24, P = 0.002).

The findings provide clinical evidence that peripheral vascular health, as captured by a noninvasive pressure waveform AI model, serves as a direct mirror of subclinical brain health and cognitive reserve.

Because vascular brain injury markers were shown to partially mediate the association between AIVA and global cognition, targeting systemic vascular aging may represent an upstream therapeutic pathway to prevent cognitive decline.

This approach could offer a low-cost, scalable alternative to expensive neuroimaging techniques like MRI and diffusion tensor imaging (DTI) for routine cognitive risk screening in aging populations.

The development of accessible, AI-enabled vascular biomarkers could refine patient stratification for clinical trials of therapeutics targeting neurodegenerative diseases and cognitive impairment.

The diagnostic and digital health sectors may see increased demand for integration of pulse wave analysis algorithms into consumer-grade and clinical wearable sensors to monitor biological aging.

Longitudinal clinical trials to establish whether changes in AIVA over time predict the clinical onset of dementia, stroke, or major depressive disorders.

Regulatory assessments of peripheral pressure waveform-derived AI models as validated secondary endpoints in clinical trials for cardiovascular and neurological longevity interventions.