Ambient Pollution Components and Sources Associated with Hippocampal Architecture and Memory in Pre-Adolescents
Background: Ambient air pollution poses significant risks to brain health. Hippocampal structure and function are particularly vulnerable, yet the extent to which they are associated with air pollution in children remains unclear. We therefore conducted multi-pollutant mixture analyses to examine how air pollution influences hippocampal architecture and memory performance in late childhood. Methods: We used partial least squares correlation to explore cross-sectional associations between fifteen PM2.5 components, six PM2.5 source factors, NO2, and ozone exposures, and measures of hippocampal microstructure and volume in children aged 9-11 years (n = 7,940) We adjusted for demographic, socioeconomic, and neuroimaging confounds. We also tested whether air pollutants were associated with hippocampal-dependent list-learning memory performance to examine functional implications of air pollution exposure. Shared variance refers to the proportion of total covariance between variable sets captured by each latent dimension in the multivariate relationship. Findings: In the first latent dimension, greater exposure to organic carbon and ozone was associated with differential hippocampal diffusion (72% of shared variance), whereas the second latent dimension linked elemental carbon and iron to hippocampal diffusion (24% of shared variance). Source-based analyses identified biomass burning and traffic pollution as key contributors (61% and 32% variance, respectively). Volumetric analyses revealed higher copper and zinc exposure correlated with smaller hippocampal subregion volumes (left head, right body, tail; 77% variance), whereas lower nickel levels correlated with smaller right head volume (12% variance). Higher industrial and traffic pollutants were also associated with smaller hippocampal volumes (75% variance). We found two latent dimensions (67% and 23% variance, respectively) showing poorer learning, immediate recall, and mnemonic interference performance linked to higher calcium, elemental carbon, and zinc, and organic carbon, alongside lower copper exposure. Finally, hippocampal diffusion (higher free water/lower hindered extracellular diffusion; 83% variance) and smaller tail volumes (96% variance) were linked to poorer RAVLT recall. Interpretation: These results underscore the complex relationship between air pollution exposure and hippocampal architecture and cautions that such structural changes may either presage or reflect subtle differences in neurocomputational mechanisms associated with learning and memory performance in children.