Alzheimer's disease (AD) is associated with the accumulation of neuropathological beta-amyloid (Ab) plaques, which is thought to be caused by an imbalance between Ab overproduction and dysfunctional Ab clearance. Both animal and human studies have shown that increased cerebrospinal fluid (CSF) levels of Ab peptides, especially Ab-38 and Ab-40 due to their high solubility, may be indicators of overall Ab dysregulation in preclinical AD, years before pathological Ab plaques begin to aggregate. This overabundance of Ab and later sequestration onto plaques eventually triggers a cascade of subsequent brain changes that may lead to cognitive decline. Indeed, alterations in gray matter integrity may play a role, as imaging studies have shown specific atrophy patterns in preclinical AD, particularly in language regions of the bilateral temporal lobes, which relate to cognitive performance. Here, we aimed to explore whether temporal lobe cortical volume is implicated in the relationship between increased CSF Ab levels and cognitive decline, as measured by confrontation naming performance -- an age-independent language task often impaired in preclinical AD -- in AD-vulnerable populations.

We selected 87 non-demented Veterans (Sex: 99% male; Age: M=68.2, SD=3.7; Education: M=15.5, SD=2.2) from the Alzheimer's Disease Neuroimaging Initiative-Department of Defense (ADNI-DOD) database based on available Boston Naming Test (BNT) scores, CSF measures of Ab-38 and Ab-40, and structural neuroimaging data. The 30-item BNT assessed confrontation naming performance. CSF Ab concentrations were measured using a 2D-UPLC-tandem mass spectrometry method outlined by ADNI-DOD. T1-weighted images were acquired on a 3T scanner and processed by ADNI to calculate cortical volumes (CVs) for regions of interest (ROIs); the present study focused on three bilateral ROIs in the temporal lobe (fusiform gyrus [FFG], inferior temporal gyrus [ITG], and middle temporal gyrus [MTG]). All CVs were adjusted (CV_adj) for intracranial volume (ICV) using the covariance formula (CV_adj = CV - b [ICV - mean(ICV)]). Linear regression models explored the relationship between CSF Ab peptides and BNT with temporal lobe ROIs as moderators using the PROCESS macro.

CV of the bilateral FFG significantly moderated the relationship between BNT performance and both CSF Ab-38 (p=.025, R2=.05, b=.0008) and Ab-40 (p=.016, R2=.06, b=.0002) levels. We then explored effects of the left and right FFG separately and found that the relationship between CSF Ab-38 and BNT was significantly moderated by the left FFG (p=.032, R2=.05, b=.0006) and nominally by the right FFG (p=.072, R2=.03, b=.0006). The relationship between CSF Ab-40 and BNT was significantly moderated by both the left (p=.032, R2=.05, b=.0001) and right (p=.038, R2=.04, b=.0001) FFG. CV of the bilateral ITG and MTG had no effect on any model (all p >.10).

Increased Ab may trigger alterations in neural gray matter integrity, specifically in the FFG of the temporal lobe, and these changes may in turn be implicated in AD-related cognitive decline, particularly in the language domain. These findings suggest that biomarker models incorporating CSF Ab and CV may aid early identification of disease and risk for cognitive decline in preclinical AD stages, which could help inform early interventions.