基于UWDRS的肝豆状核变性患者震颤评估

Evaluation of Tremor in Patients with Wilson’s Disease Based on UWDRS

震颤是肝豆状核变性(Wilson’s disease, WD)最常见的症状之一,可以应用于WD的诊断.基于统一肝豆状核变性分级量表(UWDRS)的WD震颤评估模型为医生辅助诊断和患者自我监测提供了帮助.采用智能手机加速度传感器采集52例WD患者的手部震颤数据,进行短时傅立叶变换后提取最优特征,通过四种分类器对UWDRS评分数的平均值进行了分类对比,在MATLAB上实现的实验结果显示朴素贝叶斯分类器的准确率最高,分类的精度达到将近100%.定量分析结果可以代替UWDRS的临床诊断,可进一步将算法应用到嵌入式系统中,便于WD患者进行家庭健康监护,医生实时监测治疗效果.

Tremor is one of the most common symptoms of Wilson’s Disease(WD),which can be used in the diagnosis of WD.The accurate evaluation model of WD tremor based on Unified Wilson’s Disease Rating Scale(UWDRS) is helpful for doctor assisted diagnosis and patient self-monitoring.The hand tremor data of 52 patients with WD were collected by smart phone acceleration sensor, and the optimal feature is extracted after short time Fourier transform.By comparing the average UWDRS score of four classifiers, the experimental results on MATLAB show that the accuracy of naive Bayes classifier is the highest, and the classification result reaches 100%.The results of quantitative analysis can replace the clinical diagnosis of UWDRS,and the algorithm can be further applied to the embedded system, which is convenient for WD patients to carry out family health monitoring and doctors to monitor the treatment effect in real time.