Abstract
Boukhechba, M., Baglione, A. N., & Barnes, L. E. (2020). Leveraging mobile sensing and machine learning for personalized mental health care. Ergonomics in Design: The Quarterly of Human Factors Applications, 28(4), 18–23. https://doi.org/10.1177/1064804620920494
This article was printed in the October 2020 issue of Ergonomics in Design: The Quarterly of Human Factors Applications. Important references to the work which inspired the article were omitted and are presented here. Figure 1 should have been labeled “Generalized hierarchical sensing framework. The figure is inspired by Mohr et al. (2017).” The corrected first paragraph of the “Generalized Architecture for Personalized Mental Health Interventions” section is published below. The online version of the article has been updated to reflect these changes.
Generalized Architecture for Personalized Mental Health Interventions
Inspired by Mohr et al.’s (2017) previous research, which surveyed the use of mobile sensing to inform mental health, we present a layered, hierarchical sensemaking framework that (1) illustrates how to extract knowledge of mental health states from sensors and (2) lays the groundwork for delivering effective interventions based on this knowledge. In this framework, shown in Figure 1, biomarkers are extracted from raw sensor data. These biomarkers remove noise and add more meaningfulness to the data (e.g., GPS coordinates are augmented by semantic labels such as home and work) and can be used to investigate health states such as anxiety and social avoidance. Our framework’s novelty is inherent in its bridge-building between health states and interventions, such as cognitive bias modification, and in its allowance for ongoing, dynamic health state monitoring: Interventions are delivered and their effectiveness is measured by sensing the users’ context again. We will now discuss each layer of this framework in detail and in tandem with key design considerations for mobile sensing researchers to address.