DOI: 10.1002/alz.071546 ISSN: 1552-5260

Baseline cerebrovascular reactivity predicts longitudinal enlarged perivascular space burden in the basal ganglia among cognitively normal older adults

Timothy Justin Libecap, Christopher E. Bauer, Valentinos Zachariou, Colleen Pappas, Flavius D Raslau, Peiying Liu, Hanzhang Lu, Brian T. Gold
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



Increasing evidence suggests that enlarged perivascular spaces (ePVS) on magnetic resonance images (MRIs) are associated with cognitive dysfunction in aging. However, the etiology of ePVS remains unknown. Here we tested the possibility that cerebrovascular dysfunction, as assessed by an MRI measure of cerebrovascular reactivity (CVR), may contribute to ePVS development.


79 cognitively normal, older adults (46 women, age range 60‐84) were recruited to undergo baseline MRI scanning and follow‐up scanning approximately 2.5 years later. Of the 79 participants scanned at baseline, 50 returned for follow‐up. Participants were scanned on a 3T Siemens Prisma scanner with a 64‐channel head coil. All scanning methods were identical across time points and included collection of T1‐weighted, T2‐weighted FLAIR, and quantitative susceptibility mapping (QSM) images. Cerebrovascular dysfunction was assessed using CVR, a measure of the ability of cerebrovasculature to dilate in response to vasoactive stimuli. Participants underwent blocked administration of breathing either room air or hypercapnic gas, while blood‐oxygen‐level‐dependent (BOLD) functional magnetic resonance imaging (fMRI) was performed. CVR values were computed for baseline scans in specific regions by dividing change in BOLD signal by change in end‐tidal CO2. Baseline and follow‐up ePVS counts were performed by a single rater who was blinded to participant demographics. ePVS were individually and manually counted using STRIVE criteria in a representative, axial slice of the basal ganglia (BGePVS) and centrum semiovale (CSePVS).


BGePVS and CSePVS significantly increased during the follow‐up period. Mixed effects models indicated that baseline basal ganglia CVR values significantly predicted increased BGePVS counts after controlling for age, sex, education, and time point (p = <0.001). Furthermore, this effect was present bilaterally: Baseline CVR in the left hemisphere predicted increased BGePVS in the left hemisphere (p = <0.001) and baseline CVR in the right hemisphere predicted increased BGePVS in the right hemisphere (p = <0.001). Conversely, there was no longitudinal relationship between baseline CVR and increased CSePVS (p = 0.449).


These results demonstrate that low baseline CVR is a risk factor for later development of ePVS. Localization of these findings to the basal ganglia suggests that it is particularly vulnerable to cerebrovascular dysfunction.

More from our Archive