Author ORCID Identifier

https://orcid.org/0000-0001-6738-9944

Date of Award

Spring 4-30-2024

Document Type

Thesis (Ph.D.)

Department or Program

Computer Science

First Advisor

Andrew Campbell

Abstract

This thesis presents a comprehensive exploration of enhancing mobile sensing capabilities to address various aspects of human behavior, mental health, personality, social functioning and beyond. We redesign the StudentLife app to improve its sensing efficiency and dependability, enabling support for multi-year-long studies. By adopting new app design, this study addresses the technical challenges of continuous sensing and enhances system robustness. The work is organized into several key studies that collectively aim to expand the scope of mobile sensing in diverse and complex environments.

The first study broadens the scope of mobile sensing to assess personality traits, exploring the potential of within-person variability in behavior to predict personality traits. By analyzing data from 646 college students, this study demonstrates significant correlations between sensed behaviors and self-reported personality traits, offering a novel approach for passive personality assessment.

The second study utilizes mobile sensing data to provide insights into the social functioning of individuals with mental health disorders, specifically schizophrenia. This study identifies behavioral patterns correlated with various aspects of social functioning, highlighting the potential for mobile sensing to inform new assessment and intervention strategies.

The third study investigates the integration voice diaries to enhance the prediction of auditory verbal hallucination (AVH) severity. This approach leverages deep learning models to analyze speech and mobility data, showcasing the feasibility of using mobile sensing for in-the-wild psychiatric symptom assessment.

The fourth study predicts the mental well-being of college students with a special emphasis on first-generation students, using longitudinal mobile sensing data to identify risk factors and behavioral patterns associated with mental health.

Finally, the thesis investigates the challenge of domain drift and model degradation over time, exploring adaptation technologies to maintain the effectiveness of mobile sensing frameworks for depression detection. By analyzing passive sensing data and self-reported surveys from undergraduate students over several years, this work demonstrates the efficacy of domain adaptation strategies in ensuring robust depression detection.

Together, these studies contribute to the development of power-efficient, scalable, and adaptable mobile sensing systems, pushing the boundaries of mobile sensing technologies, and offering new perspectives on assessing mental health and beyond.

Download

Share

COinS