This study used free-living accelerometry data to forecast cognitive decline in older adults without dementia. Two cohorts were used: one with hip accelerometers (7 days of continuous wear) and one with wrist accelerometers (72 hours of continuous wear). Classifier models predicted 1-year cognitive decline with over 85% accuracy using hip data and 5-year decline with nearly 70% accuracy using wrist data, exceeding the accuracy of models using demographics and comorbidities alone. The models show promise for clinical translation.
Publisher
npj Aging
Published On
Jun 06, 2022
Authors
Chengjian Shi, Niser Babiker, Jacek K. Urbanek, Robert L. Grossman, Megan Huisingh-Scheetz, Andrey Rzhetsky
Tags
cognitive decline
accelerometry
older adults
machine learning
predictive modeling
Related Publications
Explore these studies to deepen your understanding of the subject.