Cognitive reserve (CR) refers to factors that moderate the effects of age-related brain and pathologic changes on cognitive outcomes associated with Alzheimer’s disease (AD). One widely used indicator of CR is level of education. White matter hyperintensities (WMH) are a radiological feature of small vessel cerebrovascular disease that are associated with cognitive in aging and AD. We hypothesized, based on the concept of CR, that greater regional WMH would be more strongly associated with worse memory in those with lower education. Furthermore, as an extension of CR, we hypothesized that the associations between WMH and other pathology measures such as ptau-217 and hippocampal volume would be stronger in those with lower education.

A sample of 112 older adults (mean age (SD) = 70.77 (6.28) years, age range: 59-86 years, 75 (67%) women, mean education (SD) = 16.5 (2.3), education range = 12-20) who underwent MRI scanning were included. Regional WMH volumes were derived from T2-FLAIR images using ANTs, a vascular territory atlas, and manual editing. The vascular territory atlas segmented WMH into anterior cerebral artery (ACA), middle cerebral artery (MCA) and posterior cerebral artery (PCA) regional volumes. P-tau217 was assessed in plasma (n=63). From FreeSurfer, we derived average hippocampal volumes, adjusted by estimated total intracranial volume. The mnemonic discrimination task (MDT) was administered to assess episodic memory and pattern separation ability within both the object (MDT-O) and spatial (MDT-S) domains. Adjusting for age and sex, we ran moderations using PROCESS SPSS v.6 (model 1) to test whether years of education moderated the associations among vascular-defined WMH volumes and measures of cognition and pathology.

Years of education moderated the association between PCA WMH and object and spatial domains of memory. In those with lower education, higher PCA WMH were associated with worse object domain memory performance (MDT-O: interaction b=0.002, p=0.01) and with worse spatial domain memory performance (MDT-S: interaction b=0.002, p=0.02). Years of education also moderated the association between ACA WMH and ptau-217, such that those with higher education showed a positive association between ACA WMH and ptau-217 (interaction b=0.001, p=0.03). Years of education did not moderate the association between WMH and hippocampal volume.

In those with low education, PCA WMH volume was related to worse performance on tasks of episodic memory and pattern separation. In those with high education, greater ACA WMH burden is associated with higher ptau-217. The hypothesis of CR supports our finding that education moderates the associations between WMH and worse cognitive outcomes in older adults. However, while the finding of ACA WMH and ptau-217 being stronger among those with higher education was unexpected, this may be explained by depletion of CR in those who are also older or more advanced in the disease. Our finding demonstrates that education can also moderate the associations of regionally specific WMH and memory measures, as well as how reserve (education) may affect the association between regional WMH and downstream markers of AD pathology.