Scholarlink

<p>The diagnosis of major depressive disorder (MDD) urgently requires objective biomarkers for clinical translation. In this study, we established a multicenter cohort to date (<i>N</i> = 1,816; comprising 910 MDD patients and 906 healthy control subjects), using 23,608 standardized speech samples. Based on 6,373 acoustic-prosodic features, we develop a deep learning framework that employs a self-supervised architecture to leverage speech biomarkers. We perform a systematic comparative analysis among pretrained foundation models, including WavLM and HuBERT, and traditional acoustic features extracted from openSMILE. Model performance is evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Our framework achieves an AUC of 0.932 in internal validation (<i>n</i> = 333), significantly outperforming conventional methods, while maintaining efficacy in external validation (<i>n</i> = 160, AUC = 0.879). Self-supervised representations demonstrate robust diagnostic accuracy compared to other models. Leveraging the large speech biomarker dataset, our findings provide a rapid, cost-effective, and non-invasive approach for assisted depression assessment.</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

Speech as a biomarker for supported diagnosis of major depressive disorder using self-supervised representations

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月15日

查看全文
  • Yunhan Lin,
  • Biman Najika Liyanage,
  • Chuan Shi,
  • Zhengwen Zhu,
  • Yutao Sun,
  • Yaoyao Sun,
  • Ziwei Sunny Zhang,
  • Jun Yang,
  • Chenyang Xu,
  • Yundan Liao,
  • Yanbao Tao,
  • Zongfeng Li,
  • Guorui Zhao,
  • Weihua Yue

摘要

The diagnosis of major depressive disorder (MDD) urgently requires objective biomarkers for clinical translation. In this study, we established a multicenter cohort to date (N = 1,816; comprising 910 MDD patients and 906 healthy control subjects), using 23,608 standardized speech samples. Based on 6,373 acoustic-prosodic features, we develop a deep learning framework that employs a self-supervised architecture to leverage speech biomarkers. We perform a systematic comparative analysis among pretrained foundation models, including WavLM and HuBERT, and traditional acoustic features extracted from openSMILE. Model performance is evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Our framework achieves an AUC of 0.932 in internal validation (n = 333), significantly outperforming conventional methods, while maintaining efficacy in external validation (n = 160, AUC = 0.879). Self-supervised representations demonstrate robust diagnostic accuracy compared to other models. Leveraging the large speech biomarker dataset, our findings provide a rapid, cost-effective, and non-invasive approach for assisted depression assessment.