Scholarlink

<p>Accurate assessment of chronic pain is crucial for guiding clinical decisions, ensuring appropriate treatment, and optimizing patient well-being. However, current practice relies largely on self-reported measures, which are subjective and may be infeasible in patients with impaired communication. This study presents a novel dual-channel wearable EEG device integrated with machine learning for objective evaluation of chronic pain intensity. Prefrontal EEG signals were collected from 20 participants at a Pain Clinic and preprocessed to remove artifacts. Time-domain, frequency-domain, and nonlinear features were extracted and selected to train multiple classifiers for three-level pain grading. The proposed ensemble voting classifier achieved an average accuracy of 78% in distinguishing mild, moderate, and severe pain. Statistical analysis further identified alpha-band differential entropy as a promising EEG-based biomarker of pain severity, showing significant group differences in the Wilcoxon rank sum test. These findings demonstrate the feasibility of combining a compact, patient-friendly wearable EEG device with machine learning for objective pain assessment, underscoring this approach’s potential to enable point-of-care monitoring and support clinical translation.</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

Assessment and biomarker exploration of chronic pain using wearable EEG device

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月22日

查看全文
  • Yuxuan Chen,
  • Yingnan Tian,
  • Hao Zhou,
  • Jingyi Xu,
  • Yanjun Sun,
  • Chao Zhao,
  • Hong Liu

摘要

Accurate assessment of chronic pain is crucial for guiding clinical decisions, ensuring appropriate treatment, and optimizing patient well-being. However, current practice relies largely on self-reported measures, which are subjective and may be infeasible in patients with impaired communication. This study presents a novel dual-channel wearable EEG device integrated with machine learning for objective evaluation of chronic pain intensity. Prefrontal EEG signals were collected from 20 participants at a Pain Clinic and preprocessed to remove artifacts. Time-domain, frequency-domain, and nonlinear features were extracted and selected to train multiple classifiers for three-level pain grading. The proposed ensemble voting classifier achieved an average accuracy of 78% in distinguishing mild, moderate, and severe pain. Statistical analysis further identified alpha-band differential entropy as a promising EEG-based biomarker of pain severity, showing significant group differences in the Wilcoxon rank sum test. These findings demonstrate the feasibility of combining a compact, patient-friendly wearable EEG device with machine learning for objective pain assessment, underscoring this approach’s potential to enable point-of-care monitoring and support clinical translation.