Cognitive Resilience to Frailty and Mobility Adversities in Aging: Data-Driven Approach to Classification and Prediction by Risk Factors for Alzheimer’s Disease

• Author / Creator

  Thibeau, Sherilyn

• Objective: The overall aim of the current dissertation was to examine how and why some older adults maintain their cognitive performance for long periods of time, despite the presence of physical adversity factors known to be associated with the deterioration of cognitive function. We examined this general question in three longitudinal studies, organized into chapters. In Chapter Two, we examined the effect of frailty on three domains of cognitive performance and change (i.e., memory, speed, and executive function (EF)), as stratified by sex and Apolipoprotein E (APOE). In Chapter Three, we tested cognitive resilience to frailty across the three cognitive domains. Additionally, we investigated predictive factors that distinguish individuals with resilience to frailty from those without resilience. In Chapter Four, we investigated cognitive resilience to low mobility across the three cognitive domains. We also tested a set of predictive factors for distinguishing individuals with resilience to low mobility from those without resilience.
  Overall Method: For the three chapters, we assembled a sample of non-demented, community- dwelling, older adults (n = 632, M age = 71, age range = 53 – 95) from the Victoria Longitudinal Study (VLS). From this source sample we drew slightly different study samples for each chapter. In Chapter Two, we used latent growth modeling to establish the effect of frailty on cognitive performance and change as well as moderation analyses to establish the effects of APOE and sex on frailty-cognition relationships. For Chapter Three, we used two data-driven technologies, latent class growth analyses (LCGA) and random forest analyses (RFA), to (a) establish classes of relatively high and low frailty, (b) establish subclasses of resilience and non-resilience to frailty, and (c) identify salient predictors discriminating between the resilient and non-resilient classes. For Chapter Four, we used the same analytic technologies to test a similar set of research goals with respect to another physical health adversity, viz, low mobility.
  Results: We summarize the results separately for each chapter. Chapter Two. First, frailty levels predicted speed and EF performance and differential memory change trajectories. Second, change in frailty predicted rate of speed and EF decline. Third, sex moderation analyses showed that females were selectively sensitive to the effects of (a) frailty on memory change, and (b) changing frailty on speed slopes. Additionally, the effect of frailty on EF trajectories was stronger for males than females. Fourth, APOE genetic risk carriers were selectively sensitive to the effects of frailty on the rate of memory decline. Chapter Three. First, we differentiated between individuals with frailty from those who were non-frail, based on the LCGA with an algorithm of level and slope. Second, using the frail class, we used the same analytics to establish subclasses of cognitively resilient and non-resilient individuals, separately for memory, speed, and EF domains. Third, we used RFA to determine the best predictors discriminating the resilient from non-resilient subclasses in each of the three cognitive domains. The following predictors discriminated memory resilience to frailty: high education, female sex, being married, high cognitive activity, and alcohol use. Three factors distinguished EF resilience to frailty: younger age, high education, and high cognitive activity. Additionally, one factor discriminated between resilient and non-resilient subclasses, high cognitive activity. Chapter Four. First, we differentiated between individuals with low mobility from those with high mobility, based on the LCGA with an algorithm of level and slope. Second, using the class with low mobility, we established subclasses of cognitively resilient and non-resilient individuals in each of the cognitive domains. Third, the following factors differentiated between cognitively resilient and non-resilient older adults. For memory resilience to low mobility, high education, alcohol use, high cognitive activity, high physical activity, no depressive symptoms, high peak flow, and APOE non-risk status discriminated resilient from non-resilient subclasses. For speed resilience to low mobility, younger age, high education, high social activity, high peak flow, and high subjective health discriminated between resilient and non-resilient subclasses. For EF resilience to low mobility, younger age, high cognitive activity, high social activity, high volunteer activity, and low pulse pressure (PP) discriminated between resilient and non-resilient subclasses.
  Discussion: Frailty and mobility present typical adversities to aging adults. In Chapter Two, we showed that higher frailty was associated with worse cognitive performance and change, and this relationship differed according to sex and APOE genetic risk status. In Chapter Three, we empirically characterized cognitive resilience to frailty, and established factors that predicted resilience. In Chapter Four, we empirically characterized cognitive resilience to low mobility and established factors predictive of resilience to low mobility. For older adults, developing cognitive resilience despite the presence of physical health adversity offers great potential for AD risk reduction targets. Regarding frailty resilience, such potential modifiable targets include high cognitive activity and high education. Regarding mobility resilience, such potential modifiable targets include high education, high cognitive activity, high social activity, high peak flow, and younger age. Pinpointing and increasing conditions that are protective to cognitive functioning (and contribute to sustained cognitive resilience) has enormous potential to delay the onset of cognitive decline and dementia, despite the common aging adversities of frailty and mobility decrements.

• Subjects / Keywords

  • Cognitive Resilience
  • Frailty
  • Mobility
  • Data-Driven Prediction

• Graduation date

  Spring 2021

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-9c9c-np82

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.