Perivascular space mediated the interaction between sleep, and brain functional connectivity in the healthy aging population

Abstract

Perivascular space (PVS) surrounds the perforating arteries or draining veins of the cerebral cortex as part of the brain clearance system. Previous studies showed that sleep affects both brain clearance function and brain functional connectivity (FC). However, the impact of PVS characteristics on brain FC remains unclear. This study investigated these associations and their link to cognition. We utilized cross-sectional structural MRI and resting state-fMRI data from 512 health aging population in the HCP-Aging dataset, together with Pittsburgh Sleep Ǫuality Index questionnaire and NIH cognitive tests. Our results showed that basal ganglia (BG)-PVS volume fraction (VF) was positively correlated with FC in the right anterior medial temporal gyrus (aMTG) and right temporal regions, while centrum-semiovale (CSO)-PVS VF was positively correlated with FC in the left hippocampus and right frontal regions. Increased CSO-PVS VF in early middle-aged adults showed higher hippocampal FC and better cognitive performance. Interestingly, individuals with longer time spent in bed had larger BG-PVS VF linked to higher FC in the right aMTG. Additionally, older adults with better sleep quality had larger BG-PVS VF linked to higher FC in the right aMTG. These findings suggest that PVS morphology may reflect changes in neural connections involved in memory-related regions.

Keywords: Cognition; Functional connectivity; Magnetic resonance imaging; Perivascular space; Sleep.

Figure 1.. Brain functional connectivity (FC) related to PVS in HCP-Aging cohorts.

(A) Brain activation map showing areas with BOLD response related to BG-PVS in an HCP-Aging cohorts. After adjustment for age, sex, and years of education, the functional connectivity between the right aMTG and right temporal cluster positively correlated with BG-PVS volume fraction(VF). (B) Brain activation map showing areas with BOLD response related to CSO-PVS in an HCP-Aging cohorts. After adjustment for age, sex, and years of education, the functional connectivity between the left hippocampus and right frontal cluster positively correlated with CSO-PVS VF. The color-bar represents the t-value for the statistical test. Significant ***p<0.0001.

Figure 2.. Demonstration of the mediation analysis of the CSO-PVS, FC and cognitive performance in the early middle-aged and middle-aged participants.

(A) In early middle-aged population, there is mediating effect of hippocampus functional connectivity on the association between CSO-PVS volume fraction (VF) and the MoCA score. (B) In the older population, there is no mediating effect of aMTG functional connectivity on the association between CSO-PVS VF and MoCA score. Mediation (%) = (c-c’)/c. ACME: Average Causal Mediated Effect; ADE: Average Direct Effect. Significant *p<0.05.

Figure 3.. Demonstration of the mediation analysis of the PVS, time in bed/sleep quality and functional connectivity in the early middle-aged and older participants.

(A) In the early middle-aged cohort, there is mediating effect of BG-PVS volume fraction (VF) on the association between time in bed and the aMTG functional connectivity; however, (B) there is no mediating effect of CSO-PVS VF on the association between time in bed and the hippocampus functional connectivity. (C) In the older cohort, there is mediating effect of BG-PVS VF on the association between time in bed and the aMTG functional connectivity. Mediation (%) = (c-c’)/c. ACME: Average Causal Mediated Effect; ADE: Average Direct Effect. Significant *p<0.05.