Modeling Resilience to Alzheimer's Disease Pathology in Cognitively Healthy Older Adults

点击翻译按钮获取中文摘要

Project Summary Cognitive decline in aging is common but not universal. In fact, some older adults have normal cognitive performance despite evidence of Alzheimer’s disease (AD) pathology in their brains. Resilience, or successfully coping with pathology, is a new concept that provides a framework for studying variability in the cognitive consequences of pathological changes in the aging brain. The critical idea is that individuals with high resilience will only experience cognitive decline as a result of pathology when the burden is high. In contrast, less pathology may cause changes in cognition in individuals with low resilience. This proposal addresses a major challenge to the study of resilience: the development of an accurate, quantitative measurement of the phenomenon. Using neuroimaging techniques for in vivo measurement of pathology combined with novel quantitative biological measurements and statistical approaches, the research described in this proposal represents a cutting-edge combination of concepts and tools. Specifically, spatial and temporal coupling measures will be calculated relating tau positron emission tomography (PET) imaging to neurodegeneration or atrophy at the individual level. The coupling of these steps in the AD pathological cascade may help to better define resilience, which is likely related to the dynamics of that cascade. Thus, another crucial novel aspect of this approach is the conceptualization of resilience as a biological phenomenon reflecting relationships between different variables on the causal pathway to cognitive decline. Spatial coupling will be measured using voxelwise spatial correlations while latent difference score models will be used to estimate the extent to which change in tau-PET predicts atrophy. Next, linear models will be used to predict baseline cognition in four domains: episodic memory, working memory, executive functioning and processing speed. Spatial and temporal coupling metrics will be included in these linear models. Residuals from linear models predicting cognition, which are a measure of cognitive resilience, will be extracted and used in the creation of the general resilience factor. Briefly, a partial least squares path model will be used to define three latent factors: cognitive resilience, cognitive reserve and brain reserve. A general resilience factor will be determined from these three measurement model latent factors. As a validation step, general resilience factor scores will be extracted and used to predict changes in cognition and measures of cognitive engagement. The expectation is that an accurate, quantitative resilience measure will be a useful predictor of future cognitive outcomes. A separate dataset from ADNI will be used as a replication sample to ensure reproducibility of this approach and its applicability to a larger multi-site cohort. Ultimately, a delay in the onset of AD, even just by several years, would greatly decrease the overall prevalence of the disease. Studying resilience, which promotes healthy cognitive functioning for longer, may reveal critical factors that could delay symptom onset.