Personalized AI-Driven Models in Cognitive Behavioral Therapy for Anxiety

NCT07430800 | Recruiting

• 2 other identifiers: R01MH139134-At-HomeR01MH139134
• Study Type: interventional
• Target: 140
• Locations: 1 country
• Sites: 1

Brief Summary

Untreated anxiety undermines long-term physical and emotional wellbeing, especially among college students, with rates worsening since the onset of the COVID-19 pandemic. Cognitive Behavioral Therapy (CBT) is the leading evidence-based intervention for anxiety, but many students fail to complete exercises between CBT sessions, reducing its effectiveness. Socially assistive robots (SARs) help promote adherence to home-based practice in the context of elder care, social skill learning, and physical therapy, but it is unknown how SARs can enhance CBT. The specific objective of this research is to develop personalized CBT SARs that can support CBT compliance for college students with anxiety. To meet the goals of the proposed work, these studies will determine how SAR personalization based on implicit and explicit feedback can help promote greater CBT compliance and anxiety reduction outcomes for students.

Conditions

Generalized Anxiety Disorder (GAD)

Outcome Measures

Primary Outcomes (7)

• Number of user-spoken words during interactions with SAR

  Using transcriptions of the interactions, the investigators will count the number of words uttered by the participant.

  During each session, through study completion, approximately 6 weeks.

• Overall length of interaction sessions

  Using the interaction data from the robot interface, the investigators will collect the time that the participant started the interaction with the robot and the time that the interaction ended.

  Through the duration of the study, approximately 6 weeks.

• Number of days participants complete CBT exercises (adherence)

  Using the interaction data from the robot interface, the investigators will collect the number of days that the participant interacts with the system at least once during the day

  Through study completion, approximately 6 weeks.

• Outcomes Questionnaire-45.2

  The total reported symptom distress value from the symptom distress subscale. Question scale (0-4). A higher total score indicates higher distress overall, and higher scores for question subsets indicate worse outcomes for symptom distress, interpersonal relations, and social role.

  Once in weeks 1, 3, 6 of the study.

• Modified Working Alliance Inventory-Short Revised

  Repeated daily measure for assessing the quality of the relationship with the SAR companion. Question scale (1-5). Higher scores indicate stronger alliance.

  Once in weeks 2, 4, 6 of study.

• Sustem Usability Scale

  The total reported usability score for evaluating participant perception of SAR companion's usability. Question responses scale from strongly agree to strongly disagree. Responses should not indicate trends in participant outcomes, but should instead help the investigators understand participants' perception of the system.

  Once on Week 2 Day 4 and Week 6 Day 4.

• Semi-structured interview

  Focus group interviews with the investigators and participants will be used to process for collecting qualitative feedback and insights. The investigators will use this data to perform thematic analysis of participant responses while using the system.

  One interview for 30 minutes - 120 minutes at the end of the 6 week study.

Other Outcomes (1)

• GAD-7 Questionnaire

  Once in weeks 2, 4, 6 of the study.

Study Arms (3)

Explicit CBT SAR Personalization for 6 weeks

EXPERIMENTAL

College student participants with clinically elevated anxiety will engage in 6 weeks of in-home CBT daily exercises with an explicitly personalized CBT SAR in the participants' homes. Each participant will receive personalized re-engagement feedback delivered by the SAR will be based on explicit user feedback regarding the participant's subjective preferences related to the robot attributes and engagement features.

Behavioral: Explicit CBT SAR Personalization for 6 weeks

Implicit CBT SAR Personalization for 6 weeks

EXPERIMENTAL

College student participants with clinically elevated anxiety will engage in 6 weeks of in-home CBI daily exercises with an implicitly personalized CBT SAR in the participants' homes. The personalized re-engagement feedback provided by the SAR will be based on machine learning methods applied to implicit visual and auditory cues.

Behavioral: Implicit CBT SAR Personalization for 6 weeks

Control CBT SAR for 6 weeks

PLACEBO COMPARATOR

College student participants with clinically elevated anxiety will engage in 6 weeks of in-home CBT daily exercises with a non-personalized CBT SAR in the participants' homes. Participants will not have the capability to personalize the robot's attributes, and this condition will be a control baseline comparison group for the personalized intervention conditions described above.

Behavioral: Control CBT SAR for 6 weeks

Interventions

Explicit CBT SAR Personalization for 6 weeks BEHAVIORAL

College student participants with clinically elevated anxiety will engage in 6 weeks of in-home CBT daily exercises with an explicitly personalized CBT SAR in the participants' homes. Participants will receive personalized re-engagement feedback delivered by the SAR will be based on explicit user feedback regarding their subjective preferences related to the robot attributes and engagement features.

Explicit CBT SAR Personalization for 6 weeks

Implicit CBT SAR Personalization for 6 weeks BEHAVIORAL

College student participants with clinically elevated anxiety will engage in 6 weeks of in-home CBI daily exercises with an implicitly personalized CBT SAR in the participants' homes. The personalized re-engagement feedback provided by the SAR will be based on machine learning methods applied to implicit visual and auditory cues.

Implicit CBT SAR Personalization for 6 weeks

Control CBT SAR for 6 weeks BEHAVIORAL

College student participants with clinically elevated anxiety will engage in 6 weeks of in-home CBT daily exercises with a non-personalized CBT SAR. Participants will not have the capability to personalize the robot's attributes, and this condition will be a control baseline comparison group for the personalized intervention conditions described above.

Control CBT SAR for 6 weeks

Eligibility Criteria

• Age: 18 Years+
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• Are 18 years of age or older
• Are university students
• Are able to communicate in English
• Have corrected-to-normal vision and hearing
• Consent to have audio/video/interaction data recorded as part of the study
• Have a GAD-7 (Generalized Anxiety Disorder-7 item) score of 5 or greater, indicating mild to elevated levels of self-reported symptoms of anxiety
• Have a lower than minimal level of suicidality risk as measured by the Columbia Suicide Severity Rating Scale (C-SSRS) during screening ("NO" responses to items 3, 4, 5, and 6 on C-SSRS)
• Having access to home WiFi

You may not qualify if:

• Are less than 18 years of age
• Are not university students
• Are not able to communicate in English
• Do not have corrected-to-normal vision and hearing
• Do not consent to have audio/video/interaction data recorded as part of the study
• Have a GAD-7 (Generalized Anxiety Disorder-7 item) score of 4 or lower.
• Higher than a minimal level of suicidality risk as measured by the Columbia Suicide Severity Rating Scale (C-SSRS) during screening ("YES" responses to items 3, 4, 5, or 6 on C-SSRS)
• Don't have access to home WiFi

Sponsors & Collaborators

• University of Southern California: lead
• National Institute of Mental Health (NIMH): collaborator

Study Sites (1)

University of Southern California

Los Angeles, California, 90089, United States

RECRUITING

MeSH Terms

Conditions

Generalized Anxiety Disorder

Condition Hierarchy (Ancestors)

Anxiety Disorders; Mental Disorders

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: NONE
• Purpose: BASIC SCIENCE
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Chan Soon-Shiong Chair and Distinguished Professor of Computer Science, Neuroscience, and Pediatrics

Study Record Dates

• First Submitted: February 18, 2026
• First Posted: February 24, 2026
• Study Start: March 30, 2026
• Primary Completion (Estimated): August 1, 2028
• Study Completion (Estimated): August 1, 2028
• Last Updated: May 4, 2026

Record last verified: 2026-02

Locations

US (1)