Currently, there are limited efficacious dementia treatments for the 40-80% of individuals with Parkinson’s disease (PD) at risk for developing dementia. Difficulties monitoring early cognitive impairments both outside of the clinic, and concomitant delays in treatment initiation, are major barriers to the accurate diagnosis of and the development of interventions for cognitive impairments in PD. A biomarker is any characteristic that can be measured objectively and indicates pathogenic processes and/or treatment responses. Biomarkers can take a number of forms including observable behaviors. Spoken discourse (i.e., language beyond single words and sentences) is sensitive to early cognitive changes in dementia making it a suitable biomarker target. Yet, to date no study has leveraged the sensitivity of spoken discourse for developing a biomarker of cognitive impairment in PD. The long-term objective of this research is to improve the early and accurate diagnosis of individuals with PD at risk for developing dementia, using a spoken discourse biomarker. Once the biomarker is systematically validated, computational approaches can be used to automate the biomarker analysis. There are two major barriers toward achieving this objective: 1) the lack of a rigorous, sufficiently-powered, hand-coded dataset of PD spoken discourse (with and without cognitive impairment) and 2) the absence of a spoken discourse biomarker that has been rigorously developed and tested on a meticulously-characterized cohort of healthy adults and individuals with PD. Using the Fit-for-Purpose biomarker framework, the goal of the proposed research is to eliminate these barriers by completing a Phase 2 spoken discourse biomarker study. Aim 1 will rigorously characterize the spoken discourse, cognitive, and motor speech profiles of 129 healthy adults and individuals with PD (with and without cognitive impairment) using a theoretically-grounded model of discourse production. Using methods refined in our pilot studies, Aim 2 will develop and evaluate the classification accuracy of an optimally weighted discourse classification function. Aim 2 will yield a single discourse measure that is comprised of multiple, optimally weighted individual measures. Individual discourse measure values can be ‘plugged’ into the classification function to yield a score, which when compared to a cut-off value, will determine whether the discourse sample was produced by a person with cognitive impairment. The proposed research is innovative in its use of an evidence-informed biomarker framework to leverage the sensitivity of spoken discourse to develop a clinically-grounded, robust biomarker of cognitive impairment in PD. The resultant biomarker will be further developed in a future Phase 3 study where the predictive accuracy of the biomarker will be tested in a longitudinal dataset. Immediately, the proposed research significantly advances research in PD, and neurodegenerative disorders more broadly, providing a rigorous discourse dataset that can serve as the foundation for developing automated analyses of biomarkers that monitor cognition in real world environments.
|Effective start/end date||5/1/19 → 4/30/22|
- National Institute on Deafness and Other Communication Disorders (5R21DC017255-02)