Note2Narr: a neuro-symbolic multi-agent system for synthesizing high-fidelity training narratives from sparse counselor notes
- 1HDU-ITMO Joint Institute, Hangzhou Dianzi University
- 2College of Computer Science and Technology, Zhejiang University of Technology
Abstract
Introduction: Early identification of mental health risks among university students requires high-quality, privacy-conscious dialogue data for developing and evaluating computational screening models. Existing synthetic datasets often lack the contextual nuance and diagnostic consistency needed to preserve clinically relevant psychological signals.
Methods: We introduce Note2Narr, a neuro-symbolic framework for synthesizing high-fidelity training narratives from sparse counselor notes. A Dynamic Counseling Tree guides inquiry from broad topic exploration to in-depth probing of personal experiences. A Supervisor Agent equipped with a Chain of Psychological Clues audits the safety and consistency of generated narratives, while a Fictional Detail Generator incorporates diverse, disorder-neutral personal experiences. A Vendi Score-guided 1-to-X strategy adaptively allocates the number of narrative variants. Evaluation combines blinded human assessment, LLM-based assessment, student-disjoint downstream prediction, and a label-leakage audit with masking ablation.
Results: The framework produced 981 multi-turn counseling dialogues comprising 11,145 question-answer turns from 224 source students. Blinded human comparisons with MDD-5K identified statistically significant improvements in Professionalism, Communication, Participation, Repetition, and Authenticity after Holm correction. Downstream text classifiers achieved macro-AUROC and macro-AUPRC values of up to 0.739 and 0.405, respectively. Masking explicit diagnostic cues changed macro-AUPRC by at most 0.005, suggesting that predictive utility was not driven solely by direct diagnostic wording.
Discussion: These findings support Note2Narr as a framework for constructing synthetic counseling corpora for privacy-conscious dialogue modeling and screening-oriented research. The results demonstrate learnable associations in synthetic narratives; prospective and external validation remains necessary before applied use.
Related
- Relationships between Counselor Statements and Student Connection to Mental Health Care: A Cross-sectional Analysis of Digital Screening Platform Conversations
- Ayurvedacharya K. Kunjuraman Vaidyar: A Kerala-based Physician-Academician and Gandhian Apostle, synthesizing Ayurveda with modern clinical insights.
- Retrospective Narratives of Childhood and Adolescent Trauma Among Help-Seeking Adult men with High ACE Exposure
- Patient Narratives as Testimonies to Structural Inequity.
- Narratives about Breaking the Cycle of Intergenerational Violence in Families: Different Routes, Same Goal
BibTeX
@article{chen2026note2narr,
title = {{Note2Narr: a neuro-symbolic multi-agent system for synthesizing high-fidelity training narratives from sparse counselor notes}},
author = {Hanrui Chen and Cheng Yang and Yilin Hu and Jiong Lin and Beining Wu and Hanyin Zhang and Fu Li},
journal = {Frontiers in Psychiatry},
year = {2026},
month = nov,
doi = {10.3389/fpsyt.2026.1948130},
url = {https://www.frontiersin.org/articles/10.3389/fpsyt.2026.1948130/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Note2Narr: a neuro-symbolic multi-agent system for synthesizing high-fidelity training narratives from sparse counselor notes
AU - Hanrui Chen
AU - Cheng Yang
AU - Yilin Hu
AU - Jiong Lin
AU - Beining Wu
AU - Hanyin Zhang
AU - Fu Li
PY - 2026
DA - 2026/11/04
JO - Frontiers in Psychiatry
DO - 10.3389/fpsyt.2026.1948130
UR - https://www.frontiersin.org/articles/10.3389/fpsyt.2026.1948130/full
AB - Introduction: Early identification of mental health risks among university students requires high-quality, privacy-conscious dialogue data for developing and evaluating computational screening models. Existing synthetic datasets often lack the contextual nuance and diagnostic consistency needed to preserve clinically relevant psychological signals. Methods: We introduce Note2Narr, a neuro-symbolic framework for synthesizing high-fidelity training narratives from sparse counselor notes. A Dynamic Counseling Tree guides inquiry from broad topic exploration to in-depth probing of personal experiences. A Supervisor Agent equipped with a Chain of Psychological Clues audits the safety and consistency of generated narratives, while a Fictional Detail Generator incorporates diverse, disorder-neutral personal experiences. A Vendi Score-guided 1-to-X strategy adaptively allocates the number of narrative variants. Evaluation combines blinded human assessment, LLM-based assessment, student-disjoint downstream prediction, and a label-leakage audit with masking ablation. Results: The framework produced 981 multi-turn counseling dialogues comprising 11,145 question-answer turns from 224 source students. Blinded human comparisons with MDD-5K identified statistically significant improvements in Professionalism, Communication, Participation, Repetition, and Authenticity after Holm correction. Downstream text classifiers achieved macro-AUROC and macro-AUPRC values of up to 0.739 and 0.405, respectively. Masking explicit diagnostic cues changed macro-AUPRC by at most 0.005, suggesting that predictive utility was not driven solely by direct diagnostic wording. Discussion: These findings support Note2Narr as a framework for constructing synthetic counseling corpora for privacy-conscious dialogue modeling and screening-oriented research. The results demonstrate learnable associations in synthetic narratives; prospective and external validation remains necessary before applied use.
SN - 1664-0640
ER -