This project investigates how digital behavior markers derived from wearable devices can be harmonized with clinical cognitive assessments to improve the early detection and prediction of Alzheimer’s disease and related dementias (ADRDs). Our primary research questions are: (1) Can passive and semi-structured smartwatch-based digital markers predict cognitive status and clinical outcomes? (2) How can cognitive and behavioral data from different sources and measurement protocols be harmonized to enable cross-study analysis and model development? (3) How do behavioral and environmental contexts modulate everyday cognitive performance?
Our objective is to define and validate a continuum of digital cognitive and behavioral markers-from structured tasks (e.g., n-back, go/no-go) to passive sensing (e.g., sleep, mobility, routine)-and to develop machine learning models that predict clinical cognitive measures (e.g., MoCA, Trail Making Test) and forecast changes in cognition. We will apply harmonization techniques to align digital and cognitive measures across multiple datasets, including our studies and large-scale population data such as the UK Biobank.
Scientifically, this project addresses limitations of traditional assessments by capturing cognition in real-world settings using accessible, scalable technologies. It enables more ecologically valid and inclusive monitoring of cognitive health, particularly for underserved populations with elevated ADRD risk. Harmonizing diverse datasets will enhance statistical power and generalizability, supporting robust discovery and validation of digital biomarkers.