用肌电图+图神经网络提升帕金森病严重程度评估精度
Surface EMG Profiling in Parkinson's Disease: Advancing Severity Assessment with GCN-SVM
- 结合图卷积网络与支持向量机分析上臂肌电信号
- 模型准确率从83%提升至92%
- 适合临床神经评估与可穿戴设备研究者参考
帕金森病（PD）因进展性及症状复杂，诊断与监测存在挑战。本研究提出一种新方法，利用表面肌电图（sEMG）客观评估帕金森病严重程度，聚焦肱二头肌。对5名帕金森患者与5名健康对照的sEMG数据初步分析显示显著神经肌肉差异。传统支持向量机（SVM）模型准确率达83%，经图卷积网络-支持向量机（GCN-SVM）优化后提升至92%。尽管结果为初步验证，研究已建立详细实验流程，未来可通过更大样本队列验证并推动该方法在临床中的应用。该方法有望提升帕金森病严重程度评估水平，改善患者管理。
原文摘要 · Abstract (English)
Parkinson's disease (PD) poses challenges in diagnosis and monitoring due to its progressive nature and complex symptoms. This study introduces a novel approach utilizing surface electromyography (sEMG) to objectively assess PD severity, focusing on the biceps brachii muscle. Initial analysis of sEMG data from five PD patients and five healthy controls revealed significant neuromuscular differences. A traditional Support Vector Machine (SVM) model achieved up to 83% accuracy, while enhancements with a Graph Convolutional Network-Support Vector Machine (GCN-SVM) model increased accuracy to 92%. Despite the preliminary nature of these results, the study outlines a detailed experimental methodology for future research with larger cohorts to validate these findings and integrate the approach into clinical practice. The proposed approach holds promise for advancing PD severity assessment and improving patient care in Parkinson's disease management.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。