Abstract Body: Introduction: Perivascular Spaces (PVS) visible on brain MRI are fluid-filled compartments surrounding cerebral blood vessels. They are key indicators of cerebral small vessel disease (CSVD) and are associated with mild cognitive impairment (MCI) and dementia. This study aims to measure the effect of PVS location and severity on cognitive performance in adults from a community-based population.

Methods: Framingham Heart Study participants from all cohorts with available MRI scans and neuropsychological (NP) test data within one year of each other were included in a cross-sectional analysis. PVS burden in the basal ganglia (BG) and centrum semiovale (CSO) was categorized into grades I-IV based on PVS counts. PVS burden was further dichotomized into high burden (Grades III & IV) vs low (Grades I & II). NP tests included were the Boston Naming Test (BNT30), Hooper Visual Organization Test (HVOT), Visual Reproduction-Delayed (VRD), Logical Memory-Delayed (LMD), and Trials B (TRB). Multiple linear regression models tested associations between PVS burden and each NP test, adjusting for age, sex, smoking, diabetes, hypertension, and education.

Results: Among 3777 participants (mean age 60.6 ± 14.5, 46.4% male, 49% college educated, 17.5% CSO high burden, 9.8% BG high burden), we observed a general trend toward an inverse relationship between PVS burden and cognitive performance, although this trend was not significant in every model. Subjects with either high CSO or BG burden had slower completion time on TRB (β estimate 0.135, 95% Confidence Interval (CI): [0.036, 0.234], p-value: 0.008, and β: 0.346, 95% CI: [0.218, 0.475], p: <.0001, respectively). High CSO burden was also associated with lower test scores in VRD (β: -0.289, 95% CI: [-0.543, -0.034], p: 0.026), LMD (β: -0.342, 95% CI: [-0.681, -0.002], p: 0.049), and HVOT (β: -0.337, 95% CI: [-0.597, -0.077], p: 0.011).

Conclusion: PVS high burden, notably CSO high burden, is associated with poorer cognitive performance in several cognitive domains. These findings suggest a role for MRI-visible PVS as an early biomarker for the identification of those at risk of MCI or dementia. Furthermore, PVS topography may be an important factor in understanding and predicting cognitive impairment. Further research tracking cognition across several exam periods is warranted, as well as possible effect modifications by genetic and environmental factors.