A COMPARISON OF METHODS TO MEASURE INTRA-INDIVIDUAL COGNITIVE VARIABILITY IN NON-HISPANIC WHITES AND AFRICAN AMERICANS IN THE WISCONSIN REGISTRY FOR ALZHEIMER'S PREVENTION (WRAP)

Abstract

Inexpensive, easily-disseminated, and reliable markers are needed to detect preclinical Alzheimer's disease (AD). Intra-individual cognitive variability (IICV) in cognitively healthy individuals, calculated as the standard deviation of cognitive scores, predicts later cognitive impairment; yet little is known about how IICV performs in groups typically underrepresented in research. Recent research suggests that incorporating the pattern or direction of discrepancies may yield stronger IICV indicators for AD risk. In this study, we compared novel approaches for estimating IICV, which considered the direction and pattern of differences (Pattern-IICV) to a previously-used simple dispersion IICV, estimating only the magnitude of variability (Dispersion-IICV). Secondly, we compared performance of the IICV indices in African Americans (AA) and non-Hispanic whites. Participants were from the Wisconsin Registry for Alzheimer's Prevention (WRAP) and included non-Hispanic whites (n=1154) and AA (n=48) with >2 study visits. IICV estimates included a previously-studied Dispersion-IICV and Pattern-informed IICVs, using either robust normed z-scores or z-scores derived from raw scores. Table 1 describes IICV estimates. IICV indices were compared across race. Associations between IICV indices and other midlife risk factors, and baseline (normal, early MCI) and later cognitive status (normal, early MCI, clinical MCI/Dementia) were examined. AAs showed significantly higher variability than whites on IICV indices. All IICV formulations predicted concurrent cognitive status; however, only Pattern-IICVs were associated with longitudinal cognitive status. See Table 2. Conversely, Dispersion-IICVs were associated with midlife AD risk factors. Comparing indices between whites and Blacks, a raw score Pattern-IICV predicted risks similarly; whereas Dispersion-IICVs appeared under-powered to detect risk in AAs. A non-invasive, easily-disseminated marker of preclinical AD would be a valuable tool to reach populations reticent to under-go procedures in academic medical centers. Overall, AAs demonstrated more cognitive variability at baseline than whites. Although, it is unclear whether IICV is detecting true elevated risk for later impairment or over-pathologizing normal dispersion in AAs’ cognitive performance. Validity checks suggest a mixed picture for the three IICVs. Pattern-IICVs hold promise for measuring risk for incident decline, whereas Dispersion-IICV was most closely associated with other midlife risk factors. Overall, IICV needs further validation, especially in underrepresented groups.