Artificial intelligence-assisted visual elicitation in anorexia nervosa : Qualitative case studies.

Area of Science:

• Psychiatry
• Psychology
• Artificial Intelligence in Healthcare

Background:

• Anorexia nervosa (AN) is associated with difficulties in emotional processing and expression.
• Traditional therapeutic approaches may benefit from innovative methods to enhance emotional insight and engagement.

Purpose of the Study:

• To explore the feasibility and therapeutic potential of combining artificial intelligence (AI)-assisted visual elicitation with sensory-attuned guided reflection.
• To support emotional expression and engagement in individuals with anorexia nervosa (AN).

Main Methods:

• A two-session, therapist-led intervention was conducted with two adults diagnosed with restrictive AN.
• Participants engaged in guided reflection using metaphor and sensory language, with narratives translated into AI-generated images (DALL·E v3).
• Reflexive thematic analysis and cross-case synthesis were used to analyze the data.

Main Results:

• Visual metaphors facilitated the externalization and communication of emotions, eliciting embodied responses.
• The co-creative process involving AI-generated imagery enhanced therapeutic engagement and participants' sense of agency.
• Participants reported refined affective descriptions and deeper emotional exploration.

Conclusions:

• AI-assisted visual elicitation, integrated within a structured therapeutic framework, shows promise as an adjunct to traditional talking therapy for AN.
• This method may improve emotional insight and communication, particularly for individuals with challenges in emotion labeling and regulation.
• The approach offers a novel pathway for visual expression to support mental health treatment.