IMPORTANCE The eye is a sensory organ that is easily accessible for imaging techniques, allowing the measurement of the retinal nerve fiber layer (RNFL) thickness. The eye is part of the central nervous system, and its neurons may be susceptible to degeneration; therefore, changes in the RNFL thickness may reflect microstructural and volume alterations in the brain. OBJECTIVE To explore the association between the peripapillary RNFL thickness and brain alterations in the visual and limbic networks in elderly people without dementia. DESIGN, SETTING, AND PARTICIPANTS Cross-sectional analysis of the Three-City/Antioxydants, Lipides Essentiels, Nutrition et Maladies Oculaires (Alienor) Study cohort (April 2009 to December 2010). The dates of analysis were July 2017 to August 2018. The setting was a population-based study in France. The brain volume analysis included 104 participants, and the diffusion tensor imaging analysis included 79 participants. MAIN OUTCOMES AND MEASURES Global RNFL was assessed by spectral-domain optical coherence tomography. Brain volumes were assessed via T1-weighted magnetic resonance imaging by measurement of the global white and gray matter fractions and the hippocampal fraction. Brain microstructural alterations were assessed with diffusion tensor imaging at the level of the posterior thalamic radiations, the limbic system tracts (the fornix and cingulum bundles), and the posterior limb of the internal capsule (control region). Linear regression models adjusted for several confounders were performed. RESULTS Among a total of 104 participants, the mean (SD) age was 80.8 (3.9) years, and the cohort was 56.7% women (n = 59). The mean (SD) global RNFL thickness was 89.3 (12.9) mu m. A thicker RNFL was associated with a greater hippocampal fraction (quantity of increase beta = 0.013; 95% CI, 0.001-0.025 per 10-mu m increase in the RNFL thickness) and better diffusion tensor imaging variables in the global cingulum (mean diffusivity beta = -0.007; 95% CI, -0.015 to -0.000) and the hippocampal part of the cingulum (mean diffusivity beta = -0.009; 95% CI, -0.016 to -0.002 and radial diffusivity beta = -0.010; 95% CI, -0.018 to -0.002) and the posterior thalamic radiations (fractional anisotropy beta = 0.008; 95% CI, 0.000-0.017). No significant associations were found with other magnetic resonance imaging volumes or with other diffusion tensor imaging variables. In particular, there was no significant association with the control region of interest. CONCLUSIONS AND RELEVANCE Results of this study suggest that in elderly individuals without dementia, a thicker RNFL was associated with better magnetic resonance imaging variables both in a region that included the visual pathways and in regions particularly involved in the neurodegenerative processes of Alzheimer disease.< Réduire