RECAP: Resistance Capture in Text-based Mental Health Counseling with Large Language Models

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Expert, quick

Summary

The RECAP framework addresses challenges in detecting client resistance within text-based mental health counseling. It leverages PsyFIRE, a theoretically grounded framework that identifies 13 fine-grained resistance behaviors and collaborative interactions. Based on this, the ClientResistance corpus was built, comprising 23,930 annotated utterances from real-world Chinese text-based counseling, each with context-specific rationales. RECAP, a two-stage framework, detects resistance and its fine-grained types with explanations, achieving 91.25% F1 for distinguishing collaboration/resistance and 66.58% macro-F1 for fine-grained categories. It outperforms leading prompt-based LLM baselines by over 20 points. Expert evaluations confirm high faithfulness and reliability of explanations. A pilot study with 62 counselors demonstrated RECAP's potential to improve understanding and intervention strategies.

Key takeaway

For mental health counselors utilizing AI-assisted tools, RECAP offers a robust method to identify and understand client resistance in text-based interactions. You can leverage its fine-grained detection and explanatory capabilities to refine your intervention strategies and improve therapeutic relationships. Consider integrating such advanced NLP frameworks to enhance your analytical precision and client outcomes.

Key insights

Detecting fine-grained client resistance in text-based counseling improves therapeutic understanding and intervention.

Principles

• Recognizing client resistance is critical for effective counseling.
• NLP must move beyond oversimplified resistance categories.

Method

The PsyFIRE framework defines 13 fine-grained resistance behaviors. RECAP is a two-stage framework for resistance detection and classification with explanations.

In practice

• Build corpora with context-specific rationales.
• Use two-stage models for nuanced behavior detection.

Topics

• Mental Health Counseling
• Client Resistance
• Natural Language Processing
• RECAP Framework
• PsyFIRE
• Text-based Therapy

Best for: AI Scientist, NLP Engineer, Research Scientist

Related on AIssential

• An Agentic LLM-Based Framework for Population-Scale Mental Health Screening — An agentic LLM framework enhances mental health screening by ensuring robust, adaptable, and trustworthy AI pipelines.
• When PCOS Meets Eating Disorders: An Explainable AI Approach to Detecting the Hidden Triple Burden — Explainable AI can detect co-occurring health conditions like PCOS and eating disorders from social media data.
• RSPC: A Benchmark for Modeling Stress and Psychiatric Conditions in Digitally Mediated Relationships using Psychiatrist Annotations — Modeling mental health in NLP benefits from relational context, moving beyond individual-centric approaches.
• A domain-adapted large language model to support clinicians in psychiatric clinical practice — Domain-adapted LLMs can significantly enhance psychiatric clinical workflows, improving accuracy and efficiency.
• TherapyGym: Evaluating and Aligning Clinical Fidelity and Safety in Therapy Chatbots — TherapyGym evaluates and aligns therapy chatbots using clinical fidelity and safety metrics, improving performance through RL with expert-validated feedback.
• MiqraBERT: Regression-Based Sentence-BERT Finetuning for Biblical Hebrew Parallel Detection — Finetuning a BERT model with regression for semantic similarity effectively detects textual reuse in ancient languages beyond lexical overlap.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.