To identify brain iron networks associated with healthy aging and biological sex.

The study of functional and structural covariance in the brain has revealed networks relevant to physiological differences such as age and sex. However, to date no studies have analyzed covariance in brain iron. Knowledge of how different brain areas covary in iron across healthy populations could reveal valuable information on the physiology of brain iron and provide novel biomarkers to study neurological diseases.

3T QSM scans from 170 healthy controls (85 females) age 9-81 years were co-registered to a common brain template and subjected to independent component analysis (ICA), a method which decomposes signals into statistically independently co-varying areas. Artifactual components (ICs) were manually excluded and the loading factors of remaining ICs were analyzed using generalized linear statistical models that included effects of age, sex, and age*sex interaction. Multiple comparisons were controlled using Bonferroni corrections.

Six of sixty-six non-artifactual ICs had significant age effects. Notably, age-dependent deep gray matter structures were captured in two components; one containing bilateral caudate, putamen, pulvinar, and red nuclei, and the other containing bilateral globus pallidus, substantia nigra, and red nuclei. One component had significant sex and sex*age effects and was predominantly right cerebellum.