Growing advances in imaging and fluid-based assays of Alzheimer's disease (AD) biomarkers including amyloid, tau and neurodegeneration, confirm that AD processes begin decades before clinical impairment in cognitive function. Subtle changes to cognition are also likely to co-occur years before a clinical diagnosis of dementia due to AD. There is an urgent need to develop sensitive measures of subtle cognitive decline associated with AD biomarkers, particularly for monitoring response to early intervention treatments in clinical trials. The proposed investigation is highly innovative and designed to leverage existing data from three longitudinal cohort studies?Wisconsin Registry for Alzheimer's Prevention, Wisconsin Alzheimer's Disease Research Center, and BIOCARD?using a classic and widely used measure of cognition: the story recall task. We developed a novel scoring system that we hypothesize targets semantic and associative memory processes: measures that capture lexical categories and serial position. Our preliminary data shows that proper name recall and serial position scores from story recall are significantly associated with beta-amyloid status from positron emission tomography (PET), while the traditional total score was not related to amyloid status. In this proposal, our central hypothesis is that item-level analysis of existing story recall data from several longitudinal cohorts will yield one or more new measures of cognition that are uniquely associated with underlying preclinical AD pathology.
The specific aims are:
Aim1 : Use data from multiple cohort studies to a) replicate preliminary findings that lexical-level and serial position markers from delayed story recall are associated with increased risk of amyloid positivity and b) extend analyses to investigate whether these variables are associated with PET tau, CSF A? and tau, or MRI neurodegeneration measures.
Aim 2 : Compare concurrent and predictive validity of measures to determine whether the novel measures are more strongly associated with biomarkers, cognitive decline, or progression to clinical levels of impairment than traditional total score measures.
Aim 3 : Enhance the lexical-level and serial position analysis with computational linguistic analysis of digitally recorded speech from story recall to determine whether semantic content, speech fluency, error-monitoring, and serial position recall explain unique variance in levels of amyloid and/or tau pathology. Impact: The proposed project leverages existing data and is expected to lead to the development of new outcome measures from a classic, commonly used test that has played a central role in detection of disease. We expect that our higher-level language and process-based measures will be sensitive to AD biomarkers in preclinical phases of cognitive decline. By utilizing existing resources from differing cohorts, we can validate our findings without adding participant burden, share these methods with other cohort studies, further develop a digital marker of speech and cognition, and contribute an improved understanding of the underlying mechanisms of memory and communication breakdowns in preclinical AD.
Early detection of subtle cognitive decline associated with developing Alzheimer's disease (AD) pathology is crucial for identifying individuals who are most likely to benefit from early interventions. In this innovative proposal, we leverage existing data from three cohort studies to examine novel measures from the widely-used story recall task, including lexical retrieval, serial position, and computational linguistic measures, and their associations with amyloid, tau, neurodegeneration, and progression to clinical impairment. By using existing data from multiple AD cohort studies, we can develop novel methodology for early detection without adding participant burden, and can contribute an improved understanding of the underlying mechanisms of memory and communication breakdowns in preclinical AD.
