Scholarlink

Background <p>Early detection of cognitive impairment remains a critical public health challenge. While biomarkers such as neuroimaging and cerebrospinal fluid analyses offer high sensitivity, their limited accessibility hampers widespread screening, especially in underserved settings. Speech-based markers have emerged as promising, noninvasive indicators of cognitive decline.</p> Objective <p>To develop and validate SpeechDETECT, an end-to-end speech-processing pipeline that captures fine-grained acoustic and temporal markers of cognitive impairment and provides interpretable outputs suitable for large-scale screening.</p> Methods <p>SpeechDETECT comprises six modules: (1) noise reduction / amplitude normalization; (2) an eight-domain voice-analysis framework (e.g., frequency parameters, speech fluency); (3) 50&#xa0;ms segment-level feature extraction; (4) feature visualization; (5) dimensionality reduction / selection (Joint Mutual Information Maximization, LassoNet, PCA); and (6) classifier training with SHapley Additive exPlanations (SHAP). Performance was benchmarked against six acoustic toolkits (e.g., GeMAPS) on two English datasets: the DementiaBank Pitt corpus (train = 166, test = 71) with single cookie-theft picture description task and NIA PREPARE Phase 2 corpus (train = 1 064, test = 267) with multiple speech tasks.</p> Results <p>A Multi-Layer Perceptron trained on PCA-derived SpeechDETECT features achieved an&#xa0;F1-score = 0.81% and AUC-ROC = 0.80&#xa0;on the Pitt test set, outperforming the best competing toolkit (AUC = 0.76). On the PREPARE test set—comprising ≤ 30&#xa0;s recordings from four speech tasks—the same model attained&#xa0;F1 ≈ 0.67% and AUC-ROC = 0.70<b>,</b> demonstrating good generalizability. Cumulative-gains analysis showed that screening the top 40% of ranked participants captured ~ 70% of cognitively-impaired (CI) cases in Pitt and ~ 63% in PREPARE. SHAP revealed speech-fluency metrics (hesitation rate, pause ratio) and high-frequency formant dynamics as the most discriminative features.</p> Conclusion <p>SpeechDETECT delivers accurate (AUC up to 0.80) and interpretable detection of early cognitive impairment across both structured and multi-task speech settings. Its fully automated, domain-informed approach enables scalable, speech-based screening and provides a foundation for multimodal systems that combine acoustic markers with clinical or biomarker data to further improve diagnostic precision<i>.</i> The SpeechDETECT toolkit is openly available on GitHub at <a href="https://github.com/SpeechCARE/SpeechDETECT-Toolkit">https://github.com/SpeechCARE/SpeechDETECT-Toolkit</a> for researchers and clinicians.&#xa0;A demo tutorial video showing pipeline usage is available at <a href="https://github.com/SpeechCARE/SpeechDETECT-Toolkit/blob/main/SpeechDETECT.mp4">https://github.com/SpeechCARE/SpeechDETECT-Toolkit/blob/main/SpeechDETECT.mp4</a>.</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

SpeechDETECT: an explainable automated speech processing pipeline for early detection of neurological and health changes

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月24日

查看全文
  • Maryam Zolnoori,
  • Elyas Esmaeili,
  • Mehdi Naserian,
  • Ali Zolnour,
  • Sina Rashidi,
  • Tahoura Morovati,
  • Hossein Azadmaleki,
  • Zhihong Zhang,
  • James M. Noble,
  • Margaret V. McDonald

摘要

Background

Early detection of cognitive impairment remains a critical public health challenge. While biomarkers such as neuroimaging and cerebrospinal fluid analyses offer high sensitivity, their limited accessibility hampers widespread screening, especially in underserved settings. Speech-based markers have emerged as promising, noninvasive indicators of cognitive decline.

Objective

To develop and validate SpeechDETECT, an end-to-end speech-processing pipeline that captures fine-grained acoustic and temporal markers of cognitive impairment and provides interpretable outputs suitable for large-scale screening.

Methods

SpeechDETECT comprises six modules: (1) noise reduction / amplitude normalization; (2) an eight-domain voice-analysis framework (e.g., frequency parameters, speech fluency); (3) 50 ms segment-level feature extraction; (4) feature visualization; (5) dimensionality reduction / selection (Joint Mutual Information Maximization, LassoNet, PCA); and (6) classifier training with SHapley Additive exPlanations (SHAP). Performance was benchmarked against six acoustic toolkits (e.g., GeMAPS) on two English datasets: the DementiaBank Pitt corpus (train = 166, test = 71) with single cookie-theft picture description task and NIA PREPARE Phase 2 corpus (train = 1 064, test = 267) with multiple speech tasks.

Results

A Multi-Layer Perceptron trained on PCA-derived SpeechDETECT features achieved an F1-score = 0.81% and AUC-ROC = 0.80 on the Pitt test set, outperforming the best competing toolkit (AUC = 0.76). On the PREPARE test set—comprising ≤ 30 s recordings from four speech tasks—the same model attained F1 ≈ 0.67% and AUC-ROC = 0.70, demonstrating good generalizability. Cumulative-gains analysis showed that screening the top 40% of ranked participants captured ~ 70% of cognitively-impaired (CI) cases in Pitt and ~ 63% in PREPARE. SHAP revealed speech-fluency metrics (hesitation rate, pause ratio) and high-frequency formant dynamics as the most discriminative features.

Conclusion

SpeechDETECT delivers accurate (AUC up to 0.80) and interpretable detection of early cognitive impairment across both structured and multi-task speech settings. Its fully automated, domain-informed approach enables scalable, speech-based screening and provides a foundation for multimodal systems that combine acoustic markers with clinical or biomarker data to further improve diagnostic precision. The SpeechDETECT toolkit is openly available on GitHub at https://github.com/SpeechCARE/SpeechDETECT-Toolkit for researchers and clinicians. A demo tutorial video showing pipeline usage is available at https://github.com/SpeechCARE/SpeechDETECT-Toolkit/blob/main/SpeechDETECT.mp4.