Rochester Institute of Technology
Human-Machine Interactions for Behavioral Health Care
Abstract
dc:description.abstractAbout one in four American adults (26%) and 20.3% of children aged 8 to 16 years reported behavioral health concerns in 2023. However, a substantial gap remains between demand for behavioral healthcare services and access to such services. In response, research efforts have been made to advance effective and accessible approaches to provide assessments and treatments for populations with behavioral health needs. Yet, two primary challenges exist in current practice and research: 1) self-reports and manual observation-based assessments are subjective and labor-intensive, as evaluations are limited to the availability of clinical professionals; 2) they only target the general experiences but cannot capture specific details, such as the exact time when particular behavioral health concerns occurred and have minimal capability for continuous monitoring of an individual’s behavioral patterns. Therefore, this dissertation work developed human-machine interaction technologies to discover objective behavioral patterns through automated assessment of continuous signals (e.g., eye gaze). Three interaction modes – human-human, human-machine, and hybrid human-machine – were studied across three populations (intimate couples with moderate alcohol consumption, children with autism, and the primary caregivers of autistic children) to illustrate important design aspects and the preliminary efficacy. Novel multimodal measurements based on non-verbal behaviors, physiological signals, and interaction dynamics were proposed to capture unique behavioral cues that were important to the targeted populations. New data analysis pipelines were introduced, incorporating behavioral response quantification, signal processing, machine learning, and statistical modeling, for systematic evaluations. This dissertation work demonstrated the potential of using human-machine interaction as an effective tool for behavioral health assessments.
Degree
thesis:*- Name thesis:degree_name
- Engineering (Ph.D.)
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Biomedical Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Yu, Zhiwei
- Contributors dc:contributor
-
- Zhi Zheng
Subjects
dc:subject × 6Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.rit.edu/theses/11789