From the Lab to Everyday Life: Wearable Sensors for Robust, Longitudinal Assessment of Health

Speaker : Dr. Paula Lago, an emerging researcher at the CRIUGM and an Assistant Professor at Concordia University

Abstract

Wearable sensors offer continuous insight into how older adults move, sleep, and organize their days, but turning these data into reliable measures of health and healthy aging requires addressing two main challenges.

The first is robustness. Models for human activity recognition and gait analysis are sensitive to inter-subject variability and to changes in device, position, and orientation, and these challenges are amplified in older adults, whose movement patterns often differ from those the models were trained on. We will examine the sources of this variability, their effects on model performance, and strategies for developing models that generalize to real-world conditions.

The second is longitudinal analysis. Meaningful behavioral patterns emerge over weeks and months, rather than in isolated moments. We will present computational approaches for characterizing long-term behavior, including routines, rhythms, and stability, and discuss findings linking these measures to health outcomes. We will also consider strategies for modeling diverse temporal dynamics and detecting meaningful changes in behavior over time.

The Zoom link below will be active on the day of the event.