Scholarlink

<p>Accurate psychiatric diagnosis and assessment are crucial for effective treatment. However, current diagnostic approaches heavily rely on subjective observations constrained by time and clinical resources. This study investigates the potential of using Large Language Models (LLMs) to identify the symptoms in psychiatrist-patient dialogues and use them as intermediate features to predict the diagnostic labels. We collected audio recordings of 1160 outpatients with depressive disorder and anxiety disorder. LLMs were trained and utilized to identify clinical symptoms, rate assessment scales, and an ensemble learning pipeline was designed to classify diagnostic results and symptoms with 10-fold cross-validation. The system achieved 86.9% accuracy for identifying the appearance of clinical annotations and 74.7% (77.2%) accuracy for identifying symptoms of anxiety (depression). In addition, analysis of LLM-generated features shows that depression cases exhibited prominent markers of anhedonia and decreased volition, whereas anxiety disorders were characterized by tension and an inability to relax.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Identifying psychiatric manifestations in outpatients with depression and anxiety: a large language model-based approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月2日

查看全文
  • Shihao Xu,
  • Yiming Yan,
  • Yanli Ding,
  • Feng Li,
  • Shu Zhang,
  • Haoyun Tang,
  • Chao Luo,
  • Yan Li,
  • Hao Liu,
  • Yu Mei,
  • Wenjie Gu,
  • Hong Qiu,
  • Yong Wang,
  • Jianyin Qiu,
  • Tao Yang,
  • Zike Wang,
  • Qing Zhang,
  • Haiyang Geng,
  • Yunyun Han,
  • Jun Shao,
  • Nils Opel,
  • Lidong Bing,
  • Min Zhao,
  • Yifeng Xu,
  • Xun Jiang,
  • Jianhua Chen

摘要

Accurate psychiatric diagnosis and assessment are crucial for effective treatment. However, current diagnostic approaches heavily rely on subjective observations constrained by time and clinical resources. This study investigates the potential of using Large Language Models (LLMs) to identify the symptoms in psychiatrist-patient dialogues and use them as intermediate features to predict the diagnostic labels. We collected audio recordings of 1160 outpatients with depressive disorder and anxiety disorder. LLMs were trained and utilized to identify clinical symptoms, rate assessment scales, and an ensemble learning pipeline was designed to classify diagnostic results and symptoms with 10-fold cross-validation. The system achieved 86.9% accuracy for identifying the appearance of clinical annotations and 74.7% (77.2%) accuracy for identifying symptoms of anxiety (depression). In addition, analysis of LLM-generated features shows that depression cases exhibited prominent markers of anhedonia and decreased volition, whereas anxiety disorders were characterized by tension and an inability to relax.