Atlas of human brain imaging-derived phenotypes and disease risk.

Abstract

Background: The brain plays a central role in coordinating physiological processes across organ systems, yet population-scale evidence linking brain structure and function alterations, as captured by neuroimaging, to multisystem disease risk remains limited.

Methods: Leveraging multimodal magnetic resonance imaging (MRI) and linked health records from 64,836 participants, we performed a phenome-wide association study (PheWAS) to assess associations between 505 brain imaging-derived phenotypes (IDPs)-including T1-weighted, diffusion, and resting-state functional MRI measures-and 756 incident diseases spanning 15 organ systems, establishing the largest atlas of brain-disease risk to date.

Findings: Across approximately 380,000 tests, 1,500 significant IDP-disease pairs were identified, revealing that brain alterations relate not only to neurological but also to peripheral conditions, such as circulatory, digestive, and metabolic disorders. Clustering and network analyses highlighted white matter-related IDPs as a central hub linking brain and multisystem health. Prediction models combining IDPs with clinical covariates improved disease discrimination, and Mendelian randomization suggested potential causal roles of white matter-related IDPs in cerebrovascular disorders.

Conclusions: This PheWAS-based atlas advances population-level understanding of brain-body associations and provides a framework for exploring the potential of neuroimaging in cross-system disease risk assessment and therapeutic development. All associations are available through the open-access brain imaging-disease risk atlas (www.brainphewas.com).

Funding: This research was supported by the National Natural Science Foundation of China (82271220).