Jaelim Cho^a, Heeseon Jang^a, Hyunji Park^b, Young Noh^c, Jungwoo Sohn^d, Sang-Baek Koh^e, Seung-Koo Lee^f, Sun-Young Kim^g, Changsoo Kim^a,h,i

^aDepartment of Preventive Medicine, Yonsei University College of Medicine, Seoul, Republic of Korea
^bDepartment of Public Health, Yonsei University College of Medicine, Seoul, Republic of Korea
^cDepartment of Neurology, Gil Medical Center, Gachon University College of Medicine, Incheon, Republic of Korea
^dDepartment of Preventive Medicine, Jeonbuk National University Medical School, Jeonju, Republic of Korea
^eDepartment of Preventive Medicine, Wonju College of Medicine, Yonsei University, Wonju, Republic of Korea
^fDepartment of Radiology, Severance Hospital, Yonsei University College of Medicine, Seoul, Republic of Korea
^gDepartment of Cancer Control and Population Health, Graduate School of Cancer Science and Policy, National Cancer Center, Goyang, Republic of Korea
^hInstitute for Environmental Research, Yonsei University College of Medicine, Seoul, Republic of Korea
ⁱInstitute of Human Complexity and Systems Science, Yonsei University, Incheon, Republic of Korea

Corresponding author: Changsoo Kim

Little is known about the effect of air pollution on Alzheimer's disease (AD)-specific brain structural pathologies. There is also a lack of evidence on whether this effect leads to poorer cognitive function. We investigated whether, and the extent to which, AD-like cortical atrophy mediated the association between air pollution exposures and cognitive function in dementia-free adults. We used cross-sectional data from 640 participants who underwent brain magnetic resonance imaging and the Montreal Cognitive Assessment (MoCA). Mean cortical thickness (as the measure of global cortical atrophy) and machine learning-based AD-like cortical atrophy score were estimated from brain images. Concentrations of particulate matter with diameters ≤ 10 μm (PM₁₀) and ≤ 2.5 μm (PM_2.5) and nitrogen dioxide (NO₂) were estimated based on each participant's residential address. Following the product method, a mediation effect was tested by conducting a series of three regression analyses (exposure to outcome; exposure to mediator; and exposure and mediator to outcome). A 10 μg/m³ increase in PM₁₀ (β = -1.13; 95 % CI, -1.73 to -0.53) and a 10 ppb increase in NO₂ (β = -1.09; 95 % CI, -1.40 to -0.78) were significantly associated with a lower MoCA score. PM₁₀ (β = 0.27; 95 % CI, 0.06 to 0.48) and NO₂ (β = 0.35; 95 % CI, 0.25 to 0.45) were significantly associated with an increased AD-like cortical atrophy score. Effects of PM₁₀ and NO₂ on MoCA scores were significantly mediated by mean cortical thickness (proportions mediated: 25 %-28 %) and AD-like cortical atrophy scores (13 %-16 %). The findings suggest that air pollution exposures may induce AD-like cortical atrophy, and that this effect may lead to poorer cognitive function in dementia-free adults.