Objectives: To test the hypothesis that brain arterial diameters are associated with cognitive performance, particularly in arteries supplying domain-specific territories. Methods: Stroke-free participants in the Northern Manhattan Study were invited to have a brain MRI from 2003–2008. The luminal diameters of 13 intracranial arterial segments were obtained using time-of-flight magnetic resonance angiogram (MRA), and then averaged and normalized into a global score and region-specific arterial diameters. Z-Scores for executive function, semantic memory, episodic memory and processing speed were obtained at MRI and during follow-up. Adjusted generalized additive models were used to assess for associations. Results: Among the 1034 participants with neurocognitive testing and brain MRI, there were non-linear relationships between left anterior (ACA) and middle cerebral artery (MCA) diameter and semantic memory Z-scores (χ2=10.00; DF=3; p=.019), and left posterior cerebral artery (PCA) and posterior communicating artery (Pcomm) mean diameter and episodic memory Z-scores (χ2=9.88; DF=3; p=.020). Among the 745 participants who returned for 2nd neuropsychological testing, on average 5.0±0.4 years after their MRI, semantic memory change was associated non-linearly with the left PCA/Pcomm mean diameter (χ2=13.09; DF=3; p=.004) and with the right MCA/ACA mean diameter (χ2=8.43; DF=3; p=.03). In both cross-sectional and longitudinal analyses, participants with the larger brain arterial diameters had more consistently lower Z-scores and greater decline than the rest of the participants. Conclusions: Brain arterial diameters may have downstream effects in brain function presenting as poorer cognition. Identifying the mechanisms and the directionality of such interactions may increase the understanding of the vascular contribution to cognitive impairment and dementia. (JINS, 2018, 24, 335–346)