Thousand Voices of Trauma: A Large-Scale Synthetic Dataset for Modeling Prolonged Exposure Therapy Conversations

2025-04-16

AI Summary

AI development for trauma therapy is hindered by a scarcity of high-quality, clinically grounded dialogue data. Method: We introduce the first large-scale synthetic dialogue dataset (3,000 dialogues across 500 cases and six treatment phases) tailored to prolonged exposure therapy for PTSD. Our novel clinical-protocol-driven, multi-perspective emotional evolution synthesis framework integrates demographic attributes, 20 trauma types, and 10 trauma-related behavioral patterns. It jointly employs deterministic rules and probabilistic generation, embedding clinical guidelines, symptom comorbidity patterns, and empirically calibrated real-world prevalence sampling (e.g., witnessing violence: 10.6%; nightmares: 23.4%). Contribution/Results: We release an expert-validated emotional trajectory evaluation benchmark, ensuring privacy-preserving yet clinically faithful dialogues. Clinical experts confirm the dataset’s emotional depth and therapeutic fidelity, enabling quantifiable assessment of trauma-informed dialogue understanding and response generation.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Humans and AI: Emotional Intelligence
• Application Domains: Humanities & Computational Social Science

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

The advancement of AI systems for mental health support is hindered by limited access to therapeutic conversation data, particularly for trauma treatment. We present Thousand Voices of Trauma, a synthetic benchmark dataset of 3,000 therapy conversations based on Prolonged Exposure therapy protocols for Post-traumatic Stress Disorder (PTSD). The dataset comprises 500 unique cases, each explored through six conversational perspectives that mirror the progression of therapy from initial anxiety to peak distress to emotional processing. We incorporated diverse demographic profiles (ages 18-80, M=49.3, 49.4% male, 44.4% female, 6.2% non-binary), 20 trauma types, and 10 trauma-related behaviors using deterministic and probabilistic generation methods. Analysis reveals realistic distributions of trauma types (witnessing violence 10.6%, bullying 10.2%) and symptoms (nightmares 23.4%, substance abuse 20.8%). Clinical experts validated the dataset's therapeutic fidelity, highlighting its emotional depth while suggesting refinements for greater authenticity. We also developed an emotional trajectory benchmark with standardized metrics for evaluating model responses. This privacy-preserving dataset addresses critical gaps in trauma-focused mental health data, offering a valuable resource for advancing both patient-facing applications and clinician training tools.

Problem

• Limited access to therapeutic conversation data for trauma treatment
• Need for diverse synthetic dataset modeling PTSD therapy progression
• Lack of standardized metrics for evaluating AI therapy responses

Innovation

• Synthetic dataset with 3000 PTSD therapy conversations
• Diverse demographics and trauma types incorporated
• Privacy-preserving emotional trajectory benchmark developed