DreamDirector: Designing a Generative AI System to Aid Therapists in Treating Clients' Nightmares
Authors
Nightmares can adversely affect individuals' mental health and well-being, necessitating timely psychological intervention. Current nightmare therapy has set high requirements for therapists, appeared abstract to clients, and showed poor interaction between them, due to its extensive information input, lack of sensory stimulation, and exclusive reliance on one-on-one conversation. We proposed DreamDirector, a visual-interactive and narrative generative system powered by generative AI. Based on Imagery Rehearsal Therapy (IRT) and Nightmare Deconstruction and Reprocessing (NDR), the system can (1) recollect the nightmare scene, (2) interpret the dream with LLM, (3) reprocess the nightmare by generating therapeutic dream visuals using AI painting alongside meditative texts, and feedback with a picture book. Finally, we verified the usability of this system in terms of efficiency enhancement and interaction promotion through a user study with 2 therapists and 16 clients. The results revealed emotional relief among clients, with a positive and impressive attitude toward visual interaction.
Research Questions / Practical Problems
Question signals indexed for this paper.
- 67%
HCI and Affective Health: Taking stock of a decade of studies and charting future research directions
CHI '19· Mental Health Apps & Online Support Communities
- 67%
Emergent Self-Regulation Practices in Technology and Social Media Use of Individuals Living with Depression
CHI '20· Mental Health Apps & Online Support Communities
- 67%
Semantic Gap in Predicting Mental Wellbeing through Passive Sensing
CHI '22· Mental Health Apps & Online Support Communities
- 67%
Knowing How Long a Storm Might Last Makes it Easier to Weather: Exploring Needs and Attitudes Toward a Data-driven and Preemptive Intervention System for Bipolar Disorder
CHI '23· Mental Health Apps & Online Support Communities
- 67%
LumaDreams: Designing Positive Dream Meaning-Making for Daily Empowerment
CHI '25· Generative AI (Text, Image, Music, Video) +1
- 67%
Studying Self-Care with Generative AI Tools: Lessons for Design
DIS '24· Generative AI (Text, Image, Music, Video) +1
- 67%
LuciEntry: Towards Understanding the Design of Lucid Dream Induction
DIS '25· Mental Health Apps & Online Support Communities
- 67%
Symptom Detection with Text Message Log Distributions for Holistic Depression and Anxiety Screening
UbiComp '24· Mental Health Apps & Online Support Communities
Based on Jaccard similarity of research subtopics & professions (≥60%)