Cerebral Perfusion and Cortical Thickness Predict Cognitive Impairment in a Community‐dwelling Cohort

Abstract

AbstractBackgroundNeurodegeneration and cerebral perfusion are key neuroimaging biomarkers in the Alzheimer’s Disease continuum. ASL (arterial spin labeling) explores cerebral blood flow (CBF) and has aided in elucidating cerebrovascular health amongst people at risk of cognitive decline. Reduced cortical thickness (CT) on MRI is a neurodegeneration biomarker that also implicated in cognitive impairment. The influence of cortical thickness and perfusion in AD‐related regions on cognitive status is poorly understood. We explored markers to investigate the roles they had in cognitive impairment.MethodParticipants enrolled in the Wake Forest Alzheimer’s Disease Research Center (ADRC) Clinical Core cohort (N= 535, Table 1) underwent baseline T1/ASL MRI. Participants were adjudicated as cognitively impaired (CI: mild cognitive impairment or AD) or cognitively unimpaired (CU) using NIA‐AA criteria by a consensus panel of experts. CT in an AD‐related temporal lobe meta‐ROI (region of interest) was calculated on T1 using FreeSurfer v5.3 and standardized. A homologous set of AD‐sensitive temporal lobe ROIs was used to extract mean CBF, which was log transformed for normality. A one‐way MANOVA (covariates: age, education, sex) assessed effects of CT and CBF on cognitive impairment. Bivariate models evaluated the associations of CT and CBF in relation to age, education, sex, and cognitive impairment. Linear regression was used to analyze whether cognitive impairment was predicted by CT and CBF (adjusted).ResultWe found both CBF and CT were associated with age (r2= ‐0.142, p<0.01; r2= ‐0.339; p<0.01) and cognitive status (r2= ‐0.188, p<0.01; r2= ‐0.264; p<0.01); respectively. CBF was also associated with sex (r2= 0.083; p=0.055) such that women had higher CBF. We also found that higher CBF was marginally associated with higher CT (r2= 0.083; p=0.055); this relationship was not significant when age, sex, and education were included in the model (r2= 0.886; p=0.376). Finally, we found both CT and CBF predicted cognitive impairment (p<0.01; both unadjusted and adjusted, in models where each MRI marker was used to predict cognition), such that CBF and CT were higher in CU; further, both CBF and CT independently predicted cognitive status when together in a model (p<0.008).ConclusionCortical thickness and CBF are predictive of cognitive impairment.