Association between the Harvard automated phone task and Alzheimer’s disease pathology in clinically normal older adults

Abstract

AbstractBackgroundWith the advent of prevention trials in preclinical Alzheimer’s disease (AD), evaluating novel and sensitive methods to measure instrumental activities of daily living (IADL) is becoming increasingly important in order to detect these clinically meaningful changes at the earliest stages of disease. The objective of the current exploratory study was to examine the cross‐sectional relationship between a phone performance‐based IADL test, the Harvard Automated Phone Task (APT), and regional cerebral tau and cortical amyloid burden in clinically normal (CN) older adults.MethodThirty‐one CN participants (mean age 70.0 (SD 4.8), 70% female) underwent flortaucipir (FTP) tau and Pittsburgh Compound B (PiB) amyloid positron emission tomography (PET). IADL were assessed using the 3 Harvard APT tasks: refilling a prescription (APT‐Script), calling the health insurance company to select a new primary care physician (APT‐PCP), and making a bank account transfer and payment (APT‐Bank). Linear regression models were used in the primary analyses to determine the associations between each APT task rate (correct responses/time) as the dependent variable and cortical amyloid aggregate, entorhinal cortex, inferior temporal, or precuneus tau as predictors of interest in separate analyses. In secondary analyses, models were repeated with an interaction between cortical amyloid and regional tau. Analyses included age, sex, and education as covariates.ResultA significant association was found between the APT‐Script rate and entorhinal cortex tau (β=‐0.08, 95% CI=‐0.14, ‐0.21, t value=‐2.78, p=0.01). No significant associations were found between other APT tasks and amyloid and tau regions. Additionally, no significant associations were found with the interaction between amyloid and tau regions.ConclusionOur preliminary findings suggest an association between a region of early tau accumulation in CN adults, the entorhinal cortex, and a real‐life IADL test. Moreover, APT‐Script task was more normally distributed, and had overall larger variability compared to the other two tasks, suggesting that it may be a more sensitive IADL measurement in CN individuals. Further analyses should explore this relationship in larger samples and longitudinally in order to determine whether the Harvard APT can serve as a reliable IADL outcome measure for preclinical AD prevention trials and ultimately in the clinic setting.