帕金森病患者步态障碍定量评估及量化分级评估方法研究

发布时间：2018-05-17 06:13

本文选题：帕金森病 + 步态障碍　；参考：《中国科学技术大学》2017年博士论文

【摘要】：帕金森病(Parkinson's disease, PD)是一种多发于中老年人群、以运动障碍为主要临床症状的神经退行性疾病,其临床上以静止性震颤、运动迟缓、肌僵直以及姿势步态障碍为四大主要症状。目前,PD患者步态障碍的分析评估主要依靠临床医师的经验观察、检测药物反应、国际通用量表以及调查问卷的方式。这种评估方法虽然简单易用,但是其易受医师临床经验,或患者瞬间症状变化的影响,客观定量性较差。近年来,基于信息传感技术的客观定量评价方法逐渐成为研究热点。由于PD患者步态障碍症状的复杂性和多变性,尽管已经存在多种分析方法与评估系统,但其往往关注的是步态障碍的某个单一信号的分析,量化分析的全面性和准确性得不到有效保障。此外,多数系统还停留在实验室使用,并没有在PD患者步态障碍的定量分析中得到普遍性应用。因此,针对PD患者所表现出的步态障碍,本文结合了步态运动生物力学特征分析和国内外现有步态获取装置研究的基础上,设计并实现了一种基于柔性力敏传感技术和三维测力平台的新型U型电子步道系统,并在此基础上探讨了 PD患者步态障碍客观定量评估及量化分级评估方法。本文主要围绕新型U型电子步道系统设计与实现、PD患者步态信号特征综合提取、PD患者步态障碍定量评估与量化分级评估以及PD患者步态识别与评估这几个方面开展研究,具体工作内容如下:(1)构建了一种基于柔性力敏传感技术和三维测力平台的新型U型电子步道系统,用以综合获取PD患者在静态站立和动态行走过程中的足底分区压力、步态运动学参数、步态动力学参数以及转弯情形下的步态特征。(2)针对足底压力图像中的足印分割问题,提出了一种分阶段的分割算法;利用足内弓和足外弓的外轮廓曲线波动性差异实现了步行足印的实时动态识别。(3)从静态和动态足底分区压力的对比分析、步态运动学和动力学参数统计分析、步态对称性与双边协调性评估以及步态的非线性特征分析这四个方面比较PD患者步态与正常对照组之间的差异性,并在此基础上探讨了 PD患者步态障碍客观定量评估及量化分级评估方法。此外,针对PD患者与健康人群组成的非平衡数据特点,提出了一种代价敏感支持向量机方法(CS-SVM)构建了步态信号分类模型,验证了本文设计的U型电子步道用于识别帕金森病患者步态的有效性和可行性。本文构建的新型U型电子步道系统虽然受空间上的限制,但是其采集获取的步态信息较为丰富,能够综合获取PD患者在静态站立、直行以及转弯情形下的步态特征。实验结果表明:本文提出的步态障碍客观定量评估方法对临床PD患者异常步态的诊断和治疗具有积极地指导意义。在此基础上构建的多分类SVM模型取得了初步成功,错误分类都集中在相邻的两个评分级别之间。此外,与k-近邻算法(KNN)、支持向量机(SVM)以及朴素贝叶斯分类器(NBC)构建的模型相比,本文提出的一种基于代价敏感支持向量机方法(CS-SVM)构建的步态信号分类模型,其综合性能最佳,不仅有效提高了本文PD患者的识别准确率,同时有效降低了临床医师对PD患者误诊而带来的误判代价。目前本文的U型电子步态系统及PD患者客观定量评估与量化分级评估方法已经成功应用于安徽中医药大学神研所附属医院的临床实践,效果显著。
[Abstract]:Parkinson's disease (Parkinson's disease, PD) is a kind of neurodegenerative disease which mostly occurs in the middle aged and elderly people, with the main clinical symptoms of motor disorders. The four major symptoms are static tremor, motion retardation, muscular stiffness and postural gait disorder. At present, the analysis and evaluation of gait disorders in PD patients mainly rely on clinicians. This evaluation method is easy to use, but it is easily affected by clinical experience of doctors, or changes in transient symptoms of patients. In recent years, the objective quantitative evaluation method based on information sensing technique has gradually become a hot topic. Because of the complexity and variability of the symptoms of gait disorders in PD patients, although a variety of analytical methods and evaluation systems have existed, it is often concerned about the analysis of a single signal of a gait disorder. The overall and accuracy of quantitative analysis is not effectively guaranteed. In addition, most systems are still used in the laboratory and are not in the PD. The quantitative analysis of gait disorders is universally applied. Therefore, based on the analysis of the biomechanical characteristics of the gait movement and the existing gait acquisition devices at home and abroad, a new type of U type based on flexible force sensing technology and three dimensional force platform is designed and realized in view of the gait disturbance shown by PD patients. On the basis of the electronic footpath system, the objective quantitative assessment and quantitative evaluation method of gait disorder in PD patients are discussed. This paper mainly focuses on the design and implementation of the new U type electronic footpath system, the comprehensive extraction of the gait signal characteristics of the PD patients, the quantitative assessment and quantitative assessment of the gait disorder of the patients with PD, and the gait recognition and evaluation of the patients with PD. The main contents are as follows: (1) a new U type electronic footpath system based on flexible force sensing technology and three dimensional force platform is constructed, which can be used to synthetically obtain the foot subarea pressure, gait kinematics parameters, gait dynamics parameters and turn of PD patients during static and dynamic walking. In the case of the gait characteristics. (2) a phased segmentation algorithm is proposed for the foot print segmentation in the foot pressure image. The real-time dynamic recognition of foot prints is realized by using the fluctuation of the outer profile curve of the foot bow and the foot arch. (3) the comparison and analysis of the static and dynamic foot subarea pressure, the gait kinematics and dynamics. The statistical analysis of parameters, gait symmetry and bilateral coordination evaluation and the analysis of the nonlinear characteristics of gait are the four aspects to compare the differences between the PD patients' gait and the normal control group, and on this basis, the objective quantitative assessment and quantitative evaluation method of the gait disorder in PD patients are discussed. In addition, the PD patients and the healthy population are compared. A cost sensitive support vector machine (CS-SVM) method (CS-SVM) is proposed to construct a gait signal classification model, which validates the validity and feasibility of the U electronic gait design in this paper to identify the gait of patients with Parkinson's disease. The gait information obtained by collection is abundant, and it can synthesize the gait characteristics of PD patients under static standing, straight line and turning. The experimental results show that the objective quantitative assessment method proposed in this paper has positive guiding significance for the diagnosis and treatment of abnormal gait in clinical PD patients. The multi classification SVM model has achieved preliminary success, and the error classification is concentrated between two adjacent grades. In addition, compared with the k- nearest neighbor algorithm (KNN), support vector machine (SVM) and the simple Bias classifier (NBC) model, a gait signal based on the generation price sensitive support vector machine (CS-SVM) method is proposed. The class model has the best comprehensive performance. It not only effectively improves the recognition accuracy of the PD patients in this paper, but also effectively reduces the misdiagnosis cost of the clinician for the misdiagnosis of PD patients. At present, the objective quantitative assessment and quantitative evaluation method of the U type electronic gait system and the PD patients have been successfully applied to the Anhui University of traditional Chinese medicine. The clinical practice of the Affiliated Hospital of the Institute is very effective.
【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：R742.5;TP274

【参考文献】