Elamur Research⌕
☰
home / papers / 2026 / Empathy in Diversity: Personalized Depression and Anxiety Therapy via Dialogue State Tracking and Patient-Aware Planning

Empathy in Diversity: Personalized Depression and Anxiety Therapy via Dialogue State Tracking and Patient-Aware Planning

Paper · 2026 · Xinwei Yang, Junyi Fan, Yuqing Liu et al.

Summary

LLM-based therapy chatbots tend to apply the same generic interventions to every user, even though patients with depression and anxiety vary widely in traits, emotions and symptoms. The authors trace this to two gaps: weak tracking of how a patient's state changes during a session, and a lack of explicit patient modelling when choosing strategies. They make three contributions. MindEval is a role-play benchmark in which GPT-4o simulates patients built from real counseling reports. MindData is a de-identified corpus of real CBT sessions, annotated by experts at both turn and session level. MindApt is a framework that jointly tracks CBT stage, personal traits and emotional dynamics, then selects a fitting intervention from a CBT strategy library. On MindEval, MindApt beats strong baselines in therapeutic success and dialogue quality while needing fewer turns. In a clinical study with real patients, MindApt-guided care performed about as well as care directed by therapists, and therapists using MindApt's suggestions did best. This supports using LLMs as decision-support tools rather than as replacements for clinicians.

Key points

  • MindEval: About 6k patient profiles were built from counseling reports on Yidianling and Psy525. Each profile has twelve facets and belongs to one of 16 trait-defined categories. Evaluation uses 800 of these profiles (50 per category), with dialogues capped at 50 turns.
  • Metrics: Success rate (SR) and average turns (AT) are computed automatically. Dialogue quality is rated on CTRS plus Empathy, Coherence, Helpfulness and Safety. Ratings come from three LLM judges and from ten licensed CBT therapists, who scored a stratified 10% sample.
  • MindData: The corpus contains 2,573 sessions and 63,348 turns. Turn-level labels cover CBT stage, traits, emotions and therapist strategy. Session-level labels include CTRS and quality scores.
  • Results: MindApt improves success rate by 13.6% over strong baselines and remains robust across patient subgroups.
  • Clinical study: A 16-week trial approved by the Sichuan University IRB measured outcomes with BDI and BAI. The hybrid setting, combining therapist judgment with MindApt's recommendations, produced the strongest results.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Hong Kong Polytechnic UniversityHong Kong · China106 works · 5,190 citations🎓Sichuan UniversityChengdu · China20 works · 1,293 citations🎓Hong Kong Baptist UniversityHong Kong · China10 works · 246 citations🏛️Ministry of Education of the People's Republic of ChinaBeijing · China4 works · 117 citations

 Fields, topics and concepts

📚 Shelves 7The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksPsychology · 1,649 worksSocial Psychology · 330 worksApplied Psychology · 25 works

🧭 Topics 1Research subjects and languages the work is about.Dialogue systems · 654 works

Cite

Cite as Xinwei Yang, Junyi Fan, Yuqing Liu, Jiaxuan Wang, Jiashuai Zhang, Hongru Liang, Wenqiang Lei and Yao Song. 2026. Empathy in Diversity: Personalized Depression and Anxiety Therapy via Dialogue State Tracking and Patient-Aware Planning. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.424

@inproceedings{yang2026empathy,
  title = {{Empathy in Diversity: Personalized Depression and Anxiety Therapy via Dialogue State Tracking and Patient-Aware Planning}},
  author = {Xinwei Yang, Junyi Fan, Yuqing Liu, Jiaxuan Wang, Jiashuai Zhang, Hongru Liang, Wenqiang Lei and Yao Song},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.424},
  url = {https://doi.org/10.18653/v1/2026.acl-long.424}
}

References

42 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. American Psychiatric Association. 2022. Diagnostic and Statistical Manual of Mental Disorders. find
  2. Judith S Beck. 2020. Cognitive behavior therapy: Basics and beyond. Guilford Publications. find
  3. Yirong Chen, Xiaofen Xing, Jingkai Lin, Huimin Zheng, Zhenyu Wang, Qi Liu, and Xiangmin Xu. 2023. Soulchat: Improving llms’ empathy, listening, and comfort abilities through fine-tuning with multi-turn empathy conversations. arXiv preprint arXiv:2311.00273. find
  4. Yujia Chen, Changsong Li, Yiming Wang, Qingqing Xiao, Nan Zhang, Zifan Kong, Peng Wang, and Binyu Yan. 2025. Mind: Towards immersive psychological healing with multi-agent inner dialogue. arXiv preprint arXiv:2502.19860. find
  5. Yu Ying Chiu, Ashish Sharma, Inna Wanyin Lin, and Tim Althoff. 2024. A computational framework for behavioral assessment of llm therapists. arXiv preprint arXiv:2401.00820. — [arXiv:2401.00820](https://arxiv.org/abs/2401.00820) find
  6. Yujin Cho, Mingeon Kim, Seojin Kim, Oyun Kwon, Ryan Donghan Kwon, Yoonha Lee, and Dohyun Lim. 2023. Evaluating the efficacy of interactive language therapy based on llm for high-functioning autistic adolescent psychological counseling. arXiv preprint arXiv:2311.09243. — [arXiv:2311.09243](https://arxiv.org/abs/2311.09243) find
  7. Gheorghe Comanici, Eric Bieber, Mike Schaekermann, Ice Pasupat, Noveen Sachdeva, Inderjit Dhillon, Marcel Blistein, Ori Ram, Dan Zhang, Evan Rosen, and others. 2025. Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities. arXiv…
  8. Pim Cuijpers, David D Ebert, Ceren Acarturk, Gerhard Andersson, and Ioana A Cristea. 2016. Personalized psychotherapy for adult depression: A meta-analytic review. Behav. Ther., 47(6):966–980. find
  9. Joshua E. Curtiss, Daniella S. Levine, Ilana Ander, and Amanda W. Baker. 2021. Cognitive-Behavioral Treatments for Anxiety and Stress-Related Disorders. Focus, 19(2):184–189. find
  10. Yang Deng, Wenxuan Zhang, Wai Lam, See-Kiong Ng, and Tat-Seng Chua. 2023. Plug-and-play policy planner for large language model powered dialogue agents. In The Twelfth International Conference on Learning Representations. find
  11. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.
  12. Zhuoyun Du, Lujie Zheng, Renjun Hu, Yuyang Xu, Xiawei Li, Ying Sun, Wei Chen, Jian Wu, Haolei Cai, and Haohao Ying. 2024. Llms can simulate standardized patients via agent coevolution. arXiv preprint arXiv:2412.11716.
  13. Susan Folkman and Judith Tedlie Moskowitz. 2004. Coping: Pitfalls and Promise. Annual Review of Psychology, 55(1):745–774. find
  14. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, and others. 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948.
  15. He Hu, Yucheng Zhou, Juzheng Si, Qianning Wang, Hengheng Zhang, Fuji Ren, Fei Ma, and Laizhong Cui. 2025. Beyond empathy: Integrating diagnostic and therapeutic reasoning with large language models for mental health counseling. arXiv preprint arXiv:2505.15715. — [arXiv:2505.15715](https://arxiv.org/abs/2505.15715) find
  16. Sheri L Johnson. 2005. Mania and dysregulation in goal pursuit: a review. Clin. Psychol. Rev., 25(2):241– 262. find
  17. OpenAI. 2025. Gpt-5 system card. https://cdn.openai.com/gpt-5-system-card.pdf. Accessed: 2025-09-23. ↗
  18. Luoma Ke, Song Tong, Peng Cheng, and Kaiping Peng. 2025. Exploring the frontiers of llms in psychological applications: A comprehensive review. Artificial Intelligence Review, 58(10):305. find
  19. James W. Pennebaker. 1997. Writing about emotional experiences as a therapeutic process. Psychological Science, 8(3):162–166. find
  20. Klara Knauer, Anne Bach, Norbert Schäffeler, Andreas Stengel, and Johanna Graf. 2022. Personality traits and coping strategies relevant to posttraumatic growth in patients with cancer and survivors: A systematic literature review. Current Oncology, 29(12):9593–9612. find
  21. Suyeon Lee, Jieun Kang, Harim Kim, Kyoung-Mee Chung, Dongha Lee, and Jinyoung Yeo. 2024a. Cocoa: Cbt-based conversational counseling agent using memory specialized in cognitive distortions and dynamic prompt. arXiv preprint arXiv:2402.17546. — [arXiv:2402.17546](https://arxiv.org/abs/2402.17546) find
  22. Suyeon Lee, Sunghwan Kim, Minju Kim, Dongjin Kang, Dongil Yang, Harim Kim, Minseok Kang, Dayi Jung, Min Hee Kim, Seungbeen Lee, and others. 2024b. Cactus: towards psychological counseling conversations using cognitive behavioral theory. arXiv preprint arXiv:2407.03103. find
  23. June M Liu, Donghao Li, He Cao, Tianhe Ren, Zeyi Liao, and Jiamin Wu. 2023. Chatcounselor: A large language models for mental health support. arXiv preprint arXiv:2309.15461. — [arXiv:2309.15461](https://arxiv.org/abs/2309.15461) find
  24. Mounica Maddela, Megan Ung, Jing Xu, Andrea Madotto, Heather Foran, and Y-Lan Boureau. 2023. Training models to generate, recognize, and reframe unhelpful thoughts. arXiv preprint arXiv:2307.02768. find
  25. Mason McClay, Matthew E Sachs, and David Clewett. 2023. Dynamic emotional states shape the episodic structure of memory. Nature Communications, 14(1):6533. find
  26. Elizabeth M Muran and Robert W Motta. 1993. Cognitive distortions and irrational beliefs in post-traumatic stress, anxiety, and depressive disorders. Journal of Clinical Psychology, 49(2):166–176. find
  27. Huachuan Qiu, Hongliang He, Shuai Zhang, Anqi Li, and Zhenzhong Lan. 2023. Smile: Single-turn to multi-turn inclusive language expansion via chatgpt for mental health support. arXiv preprint arXiv:2305.00450.
  28. Huachuan Qiu and Zhenzhong Lan. 2024. Interactive agents: Simulating counselor-client psychological counseling via role-playing llm-to-llm interactions. arXiv preprint arXiv:2408.15787.
  29. C. R. Rogers. 1957. The necessary and sufficient conditions of therapeutic personality change. Journal of Consulting Psychology, 21(2):95–103. find
  30. Edmund T. Rolls. 2013. What are Emotional States, and Why Do We Have Them? Emotion Review, 5(3):241–247. find
  31. Karsten Roth, Biagio Brattoli, and Bjorn Ommer. 2019. Mic: Mining interclass characteristics for improved metric learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV). find
  32. A. John Rush, Jan Weissenburger, and Greg Eaves. 1986. Do thinking patterns predict depressive symptoms? Cognitive Therapy and Research, 10(2):225–235. find
  33. Samuel Schmidgall, Rojin Ziaei, Carl Harris, Eduardo Reis, Jeffrey Jopling, and Michael Moor. 2024. Agentclinic: a multimodal agent benchmark to evaluate ai in simulated clinical environments. arXiv preprint arXiv:2405.07960. — [arXiv:2405.07960](https://arxiv.org/abs/2405.07960) find
  34. Helmut Schöller, Kathrin Viol, Wolfgang Aichhorn, Marc-Thorsten Hütt, and Günter Schiepek. 2018. Personality development in psychotherapy: a synergetic model of state-trait dynamics. Cognitive Neu-rodynamics, 12:441–459. find
  35. Hongbin Na. 2024. Cbt-llm: A chinese large language model for cognitive behavioral therapy-based mental health question answering. arXiv preprint arXiv:2403.16008. find
  36. Ashish Sharma, Kevin Rushton, Inna Lin, David Wadden, Khendra Lucas, Adam Miner, Theresa Nguyen, and Tim Althoff. 2023. Cognitive reframing of negative thoughts through human-language model interaction. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volum… find
  37. Hongbin Na, Yining Hua, Zimu Wang, Tao Shen, Beibei Yu, Lilin Wang, Wei Wang, John Torous, and Ling Chen. 2025. A survey of large language models in psychotherapy: Current landscape and future directions. arXiv preprint arXiv:2502.11095. find
  38. Martin D van den Broek, Linda Monaci, and Jared G Smith. 2019. Personal problems questionnaire (ppq): Normative data and utility in assessing acquired neurological impairment. Archives of Clinical Neuropsychology, 34(5):625–636. find
  39. OpenAI. 2024. Gpt-4o system card. https://cdn.openai.com/gpt-4o-system-card.pdf. Accessed: 2025-09-23. ↗
  40. Laurens van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-sne. Journal of Machine Learning Research, 9(86):2579–2605. find
  41. Jiashuo Wang, Yang Xiao, Yanran Li, Changhe Song, Chunpu Xu, Chenhao Tan, and Wenjie Li. 2024a. Towards a client-centered assessment of llm therapists by client simulation. arXiv preprint arXiv:2406.12266. conducted a secondary manual review to verify that no sensitive private data remained in the c… — [arXiv:2406.12266](https://arxiv.org/abs/2406.12266) find
  42. Ruiyi Wang, Stephanie Milani, Jamie C Chiu, Jiayin Zhi, Shaun M Eack, Travis Labrum, Samuel M Murphy, Nev Jones, Kate Hardy, Hong Shen, and others. 2024b. Patient-{\\Psi}: Using large language models to simulate patients for training mental health professionals. arXiv preprint arXiv:2405.19660. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yang-2026-empathy-diversity-personalized-depression-anxiety/