Preclinical Alzheimer disease pathology partially mediates the relationship between cortical thickness and cognitive control

Abstract

AbstractBackgroundPrevious research has shown that with age and disease, the integrity of grey matter structures decline. One such measure of grey matter integrity is cortical thickness, where lower values represent thinning of the cortical surface. Given that cognitive performance tends to decline with age, we were interested in investigating whether performance on attention‐based tasks was related to cortical thickness measured using magnetic resonance imaging (MRI) in an unimpaired older adult cohort. Further, although often unmeasured in studies of cognitive aging, many older adults will present with Alzheimer Disease (AD) pathologies such as beta‐amyloid (Aβ) deposition, despite being asymptomatic. Therefore, we were also interested in quantifying how much of the behavior‐structure relationship seen in older adult cohorts is driven by preclinical AD.Method511 cognitively unimpaired (CDR™=0) older adults (Age: M=68.2, SE=0.38) enrolled in studies at the Knight Alzheimer Disease Research Center, were included in these analyses. All participants completed an MRI and several cognitive assessments (Stroop, Simon, and CVOE); a subset of individuals (n=396) also underwent Aβ‐positron emission tomography (PET). A spatial cortical map highlighting regions where cortical thickness and cognition were related was created (Figure 1). This map was entered into subsequent analyses as our MRI variable. We used an established attentional control composite as our behavioral measure of interest. Linear models were conducted to elucidate the relationship between cortical thickness and cognition, and a mediation analysis was performed to examine whether Aβ deposition impacted this relationship.ResultUsing vertex‐wise general linear models, we found that reduced cortical thickness in temporal, medial‐ and lateral‐parietal, and dorsolateral prefrontal cortex, predicted worse performance on an attention task composite (Figure 2). Interestingly, mediation analysis showed that levels of Aβ partially account for the observed relationship between cortical thickness and attentional control (Figure 3).ConclusionThese results demonstrate that behavioral variability seen in older adult populations reflects corresponding changes in brain anatomy, specifically cortical thickness. Furthermore, our results support the hypothesis that preclinical AD, as measured by Aβ deposition, is partially driving what would otherwise be considered general aging in a cognitively normal older adult population.