摘要
目的 采用可穿戴式惯性传感器采集的运动数据进行帕金森病(Parkinson’s Disease,PD)患者运动症状自动评分系统的研究,以提高疾病诊断的准确性和便捷性。方法 选取2020年5月至2021年12月在江苏省老年病医院神经内科就诊的PD患者作为研究对象。根据临床需求设计传感器采集数据的范式动作,并与统一帕金森病评分量表(Unified Parkinson’s Disease Rating Scale,UPDRS)运动评分项目建立对应关系。基于SQLite数据库,开发自动导出、自动管理的患者信息、评分与运动信号管理工具,获取平稳动作和重复性动作的特征性运动参数,探索基于支持向量机(Support Vector Machine,SVM)的自动评分模型,最终建立基于粒子群优化算法超参数优化的SVM自动评分模型。结果 使用分段参数序列组合作为输入,震颤最优准确度为0.90,平衡性最优准确度为0.87,上肢灵活性、下肢灵活性准确度分别可以达到0.83和0.81。直接使用主成分分析方法降维方法,步态最优准确度为0.78。结论 基于可穿戴式惯性传感器的PD运动功能自动量化系统与临床运动症状评估量表有较好的相关性,为临床工作带来便捷的同时,在一定程度上提高诊断的精准性,临床应用价值较高。
Objective To improve the accuracy and convenience of diagnosis of Parkinson’s disease,to use the motion data collected by wearable inertial sensors to study the automatic scoring system of motor symptoms in patients with Parkinson’s disease(PD).Methods PD patients treated in the Department of Neurology of Jiangsu Province Geriatrics Hospital from May 2020 to December 2021 were selected as the study subjects.According to the clinical requirements,the normal motion of sensor motion acquisition experiment was designed,and the corresponding relationship was established with the unified Parkinson’s disease rating scale(UPDRS)motion score project.Based on SQLite database,an automatic export,automatic management of patient information,score and movement signal management tool were developed.The characteristic motion parameters of stationary and repetitive movements were obtained,and the automatic scoring model based on support vector machine(SVM)was explored.Finally,the SVM automatic scoring model based on particle swarm optimization hyperparameter optimization was established.Results Using the piecewise parameter sequence combination as input,the optimal accuracy was 0.90 in the tremor test and 0.87 in the balance test.The accuracy of upper limb flexibility test and lower limb flexibility test could reach 0.83 and 0.81 respectively.The optimal accuracy of gait test was 0.78 by principal component analysis method.Conclusion The automatic quantification system of PD motor function based on wearable inertial sensor has a good correlation with the clinical motor symptom assessment scale,which brings convenience to clinical work and improves the accuracy of diagnosis to a certain extent,and the clinical application value is high.
作者
孙奕
荣哲
汪丰
徐畅
郑慧芬
SUN Yi;RONG Zhe;WANG Feng;XU Chang;ZHENG Huifen(Department of General Practice,Jiangsu Health Vocational College,Nanjing Jiangsu 210029,China;Department of Neurology,Jiangsu Province Geriatric Hospital,Nanjing Jiangsu 210024,China;School of Biological Sciences and Medical Engineering,Southeast University,Nanjing Jiangsu 210009,China)
出处
《中国医疗设备》
2023年第10期27-32,共6页
China Medical Devices
基金
江苏省卫生健康委科研项目(H2019052)
院级人才建设基金(IR2019101)。
关键词
可穿戴式
惯性传感器
帕金森病
统一帕金森病评分量表
支持向量机
wearable
motion sensor
Parkinson’s disease
unified Parkinson’s disease rating scale
support vector machine
