Education is known to impact neuropsychological test performance, and self-reported years of education is often used in stratifying normative data. However, this variable does not always reflect education quality, particularly among underrepresented populations, and may overestimate cognitive impairment in individuals with low education quality. This cross-sectional study evaluated relative contributions of years of education and reading level to several verbally mediated assessments to improve interpretation of neuropsychological performance.

Data was obtained from the Vanderbilt Memory and Aging Project. Cognitively-unimpaired participants (n=175, 72±7 years, 59% male, 87% Non-Hispanic White, 16±2 years of education) completed a comprehensive neuropsychological protocol. Stepwise linear regressions were calculated using education and Wide Range Achievement Test (WRAT)-3 Reading subtest scores as predictors and letter fluency (FAS, CFL), category fluency (Vegetable and Animal Naming), the Boston Naming Test (BNT), and California Verbal Learning Test (CVLT)-II as outcomes to assess increase in variance explained by educational quality. Models covaried for age and sex. The False Discovery Rate (FDR) based on the Benjamini-Hochberg procedure (Benjamini & Hochberg, 1995) was used to correct for multiple comparisons.

The mean WRAT-3 score was 51±4 (range:37-57), indicating post-high school reading level. Education and WRAT-3 scores were moderately correlated (r=0.36, p<0.01). Both WRAT-3 and years of education independently predicted letter fluency (WRAT-3 p<0.001; education p<0.02), category fluency (WRAT-3 p<0.001; education p<0.05), and CVLT-II performance (WRAT-3 p-values<0.005; education p-values<0.02) in single predictor models. On BNT, WRAT-3 (p<0.001), but not education (p=0.06), predicted performance in single predictor models. In combined models with both WRAT-3 and education, WRAT-3 scores remained a significant predictor of FAS (WRAT-3 b=1.21, p<0.001; education b=0.006, p=0.99) and CFL performance (WRAT-3 b=1.02, p<0.001; education b=0.51, p=0.14). Both WRAT-3 (b=0.21, p=0.01) and years of education (b=0.35, p=0.03) predicted Animal Naming, while WRAT-3 (b=0.16,p=0.008), but not years of education (p=0.37), predicted Vegetable Naming. WRAT-3 was a significant predictor of BNT performance (b=0.21, p<0.001) but not years of education (p=0.65). WRAT-3 predicted CVLT-II learning (b=0.32, p=0.04), immediate recall (b=0.16, p=0.005), and delayed recall performances (b=0.15, p=0.005), while education did not (p-values>0.14). All significant results persisted after FDR correction. WRAT-3 scores explained an additional level of variance beyond the covariates and education alone for FAS (AR=18%), CFL (AR=13%), Animal Naming and Vegetable Naming (AR= 3%), BNT (AR=18%), and CVLT-II learning (AR=2%), immediate recall (AR=4%), and delayed recall (AR=3%).

Reading level more strongly associated with performance on several verbally mediated neuropsychological measures than years of education. For all measures, the addition of reading level significantly increased the amount of variance explained by the model compared to covariates and education alone, which aligns with existing research. However, most of this past work looks at individuals with lower levels of educational quality. Because our cohort was highly educated and at the upper end of the reading spectrum, our results suggest that reading level is important to consider even for more highly educated individuals. Therefore, reading level is a critical variable to consider when interpreting verbally mediated neuropsychological measures for individuals across the educational spectrum.

Area Deprivation Index (ADI) is a measurement of neighborhood disadvantage. Evidence suggests that living in a disadvantaged neighborhood has a negative impact on health outcomes independent of socioeconomic status, including increased risk for Alzheimer's disease (AD). However, less is known about the biological mechanisms that drive these associations. We examined how ADI influences structural imaging variables and cognitive performance in community-dwelling older adults. We hypothesized that greater neighborhood disadvantage would predict atrophy and worse cognitive trajectory over time.

Participants included the legacy cohort from the Vanderbilt Memory and Aging Project (n=295, 73±7 years of age, 16±3 years of education, 42% female, 85% non-Hispanic White) who lived in the state of Tennessee. T1-weighted and T2-weighted fluid-attenuated inversion recovery brain MRIs and a comprehensive neuropsychological assessment were acquired at baseline, 18-month, 3-year, 5-year and 7-year follow-up time (mean follow-up time=5.2 years). Annual change scores were calculated for all neuropsychological and structural MRI outcome variables. Baseline state ADI was calculated using the University of Wisconsin School of Medicine and Public Health Neighborhood Atlas (Kind & Buckingham, 2018) and was based on deciles where 1 represents the least deprived area and 10 represents the most. Mixed effects regression models related baseline ADI to longitudinal brain structure (volume, thickness, white matter hyperintensities) and neuropsychological trajectory (one test per model). Analyses adjusted for age, sex, race/ethnicity, education, Framingham Stroke Risk Profile score, (apolipoprotein) APOE-e4 status, cognitive status, and intracranial volume (for MRI outcomes). Models were repeated testing interactions with APOE-e4 status, sex, and cognitive status. A false discovery rate (FDR) correction for multiple comparisons was performed.

On average, the sample was from relatively less disadvantaged neighborhoods in Tennessee (ADI state decile=2.4±1.8). Greater neighborhood disadvantage at study entry predicted more thinning of an AD-signature composite over time (ß=-0.002, p=0.005, pFDR=0.06); however, all other models testing MRI and neuropsychological outcomes were null (p-values>0.05, pFDR-values>0.51). Baseline ADI interacted with sex on longitudinal cortical thinning captured on the AD-signature composite (ß=0.004, p=0.006, pFDR=0.08) as well as several longitudinal cognitive outcomes including an executive function composite score (ß=0.033, p<0.001, pFDR=0.01), naming (ß=0.10, p=0.01, pFDR=0.12), visuospatial functioning (ß=0.083, p=0.02, pFDR=0.09), and an episodic memory composite score (ß=0.021, p=0.02, pFDR=0.07). In stratified models by sex, greater ADI predicted greater cortical thinning over time and worse longitudinal neuropsychological performance among men only. All stratified models in women were null except for executive function composite score, which did not survive correction for multiple comparisons (ß=-0.013, p=0.03, pFDR=0.61). Interactions by APOE-e4 and cognitive status were null (p-values>0.06, pFDR-values>0.61).

Among community-dwelling older adults, greater neighborhood disadvantage predicted greater cortical thinning over the mean 5-year follow-up in anatomical regions susceptible to AD-related neurodegeneration. Neighborhood disadvantage also interacted with sex on cortical thickness and several cognitive domains, with stronger effects found among men versus women. By contrast, there were no interactions between neighborhood disadvantage and genetic risk for AD or cognitive status. This study provides valuable evidence for sociobiological mechanisms that may underlie health disparities in aging adults whereby neighborhood deprivation is linked with neurodegeneration over time.

Novel blood-based biomarkers for Alzheimer's disease (AD) could transform AD diagnosis in the community; however, their interpretation in individuals with medical comorbidities is not well understood. Specifically, kidney function has been shown to influence plasma levels of various brain proteins. This study sought to evaluate the effect of one common marker of kidney function (estimated glomerular filtration rate (eGFR)) on the association between various blood-based biomarkers of AD/neurodegeneration (glial fibrillary acidic protein (GFAP), neurofilament light (NfL), amyloid-b42 (Ab42), total tau) and established CSF biomarkers of AD (Ab42/40 ratio, tau, phosphorylated-tau (p-tau)), neuroimaging markers of AD (AD-signature region cortical thickness), and episodic memory performance.

Vanderbilt Memory and Aging Project participants (n=329, 73±7 years, 40% mild cognitive impairment, 41% female) completed fasting venous blood draw, fasting lumbar puncture, 3T brain MRI, and neuropsychological assessment at study entry and at 18-month, 3-year, and 5-year follow-up visits. Plasma GFAP, Ab42, total tau, and NfL were quantified on the Quanterix single molecule array platform. CSF biomarkers for Ab were quantified using Meso Scale Discovery immunoassays and tau and p-tau were quantified using INNOTEST immunoassays. AD-signature region atrophy was calculated by summing bilateral cortical thickness measurements captured on T1-weighted brain MRI from regions shown to distinguish individuals with AD from normal cognition. Episodic memory functioning was measured using a previously developed composite score. Linear mixed-effects regression models related predictors to each outcome adjusting for age, sex, education, race/ethnicity, apolipoprotein E-e4 status, and cognitive status. Models were repeated with a blood-based biomarker x eGFR x time interaction term with follow-up models stratified by chronic kidney disease (CKD) staging (stage 1/no CKD: eGFR>90 mL/min/1.73m2, stage 2: eGFR=60-89 mL/min/1.73m2; stage 3: eGFR=44-59mL/min/1.73m2 (no participants with higher than stage 3)).

Cross-sectionally, GFAP was associated with all outcomes (p-values<0.005) and NfL was associated with memory and AD-signature region cortical thickness (p-values<0.05). In predictor x eGFR interaction models, GFAP and NfL interacted with eGFR on AD-signature cortical thickness, (p-values<0.004) and Ab42 interacted with eGFR on tau, p-tau, and memory (p-values<0.03). Tau did not interact with eGFR. Stratified models across predictors showed that associations were stronger in individuals with better renal functioning and no significant associations were found in individuals with stage 3 CKD. Longitudinally, higher GFAP and NfL were associated with memory decline (p-values<0.001). In predictor x eGFR x time interaction models, GFAP and NfL interacted with eGFR on p-tau (p-values<0.04). Other models were nonsignificant. Stratified models showed that associations were significant only in individuals with no CKD/stage 1 CKD and were not significant in participants with stage 2 or 3 CKD.

In this community-based sample of older adults free of dementia, plasma biomarkers of AD/neurodegeneration were associated with AD-related clinical outcomes both cross-sectionally and longitudinally; however, these associations were modified by renal functioning with no associations in individuals with stage 3 CKD. These results highlight the value of blood-based biomarkers in individuals with healthy renal functioning and suggest caution in interpreting these biomarkers in individuals with mild to moderate CKD.

Alzheimer’s disease (AD) is highly heritable, and AD polygenic risk scores (AD-PRSs) have been derived from genome-wide association studies. However, the nature of genetic influences very early in the disease process is still not well known. Here we tested the hypothesis that an AD-PRSs would be associated with changes in episodic memory and executive function across late midlife in men who were cognitively unimpaired at their baseline midlife assessment..

We examined 1168 men in the Vietnam Era Twin Study of Aging (VETSA) who were cognitively normal (CN) at their first of up to three assessments across 12 years (mean ages 56, 62, and 68). Latent growth models of episodic memory and executive function were based on 6–7 tests/subtests. AD-PRSs were based on Kunkle et al. (Nature Genetics, 51, 414–430, 2019), p < 5×10−8 threshold.

AD-PRSs were correlated with linear slopes of change for both cognitive abilities. Men with higher AD-PRSs had steeper declines in both memory (r = −.19, 95% CI [−.35, −.03]) and executive functioning (r = −.27, 95% CI [−.49, −.05]). Associations appeared driven by a combination of APOE and non-APOE genetic influences.

Memory is most characteristically impaired in AD, but executive functions are one of the first cognitive abilities to decline in midlife in normal aging. This study is among the first to demonstrate that this early decline also relates to AD genetic influences, even in men CN at baseline.