PanicToCalm: A Proactive Counseling Agent for Panic Attacks

2025-10-24

AI Summary

High-quality training data for panic attack intervention remains scarce. Method: We introduce PACE, the first first-person, high-stress dataset for panic episodes, alongside PanicEval—a dedicated evaluation framework. Grounded in psychological first aid principles, we propose PACER, a counseling agent that jointly optimizes crisis instruction following and empathetic emotional support via supervised fine-tuning and simulated preference alignment. Our approach establishes, for the first time, a closed-loop pipeline integrating domain-specific data curation, strategy-driven modeling, and multi-dimensional evaluation for panic intervention. Results: PACER significantly outperforms general-purpose LMs, CBT-oriented models, and GPT-4 across consultation quality and user affect improvement. Human evaluations confirm its state-of-the-art performance, establishing a reproducible, rigorously evaluable paradigm for AI-enabled psychological crisis intervention.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Cognitive Modeling & Cognitive Systems: Affective Computing
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Psychology-informed user models and recommender systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Panic attacks are acute episodes of fear and distress, in which timely, appropriate intervention can significantly help individuals regain stability. However, suitable datasets for training such models remain scarce due to ethical and logistical issues. To address this, we introduce PACE, which is a dataset that includes high-distress episodes constructed from first-person narratives, and structured around the principles of Psychological First Aid (PFA). Using this data, we train PACER, a counseling model designed to provide both empathetic and directive support, which is optimized through supervised learning and simulated preference alignment. To assess its effectiveness, we propose PanicEval, a multi-dimensional framework covering general counseling quality and crisis-specific strategies. Experimental results show that PACER outperforms strong baselines in both counselor-side metrics and client affect improvement. Human evaluations further confirm its practical value, with PACER consistently preferred over general, CBT-based, and GPT-4-powered models in panic scenarios (Code is available at https://github.com/JihyunLee1/PanicToCalm ).

Problem

• Addressing scarcity of panic attack intervention training datasets
• Developing counseling model for empathetic crisis support
• Evaluating effectiveness through multidimensional panic-specific framework

Innovation

• PACE dataset with high-distress first-person narratives
• PACER model trained with supervised learning methods
• PanicEval framework for multi-dimensional effectiveness assessment