特定群体手写运动定量分析研究

发布时间：2018-03-16 19:41

本文选题：手写运动　切入点：书写困难　出处：《中国科学技术大学》2016年博士论文　论文类型：学位论文

【摘要】：手写运动作为人体高级神经系统控制的特有精细运动,与人体脑功能、认知水平具有密切相关性。现有的手写运动分析方法,存在研究技术手段单一、对健康群体关注度较少、手写能力评价体系不全面等问题。本文重点针对身体各项机能处于快速发育的儿童,以及脑功能受到持续损伤的神经退行性疾病患者这两类群体,探讨手写运动在人体神经系统发育及病变阶段的表现特征,以理解手写运动机能的动态变化特点。本文针对手写运动发展、手写运动损伤开展特征分析以及量化分类工作,主要研究内容如下：(1)儿童手写运动能力发展研究：手写作为儿童进行思想表达、开展学业的一项基本方式,在儿童认知、心理发展过程中具有重要地位。对普通儿童的手写运动发展规律进行研究,有助于建立儿童手写运动常模数据,并为不同年龄段病患儿童的手写运动分析提供有效的对比数据源。本文基于手写三维力信息分析了儿童手写运动过程中的施力情况,并提出运动一致性特征参数,结合运动学、动力学特征对学龄期儿童的手写能力发展进行定量描述。结果表明高年级儿童具有更短的运动时间、更高的运动速度、速度曲线变化方向变化次数更少,更倾向于在书写平面中心附近抄绘图形、以施加更小的力及更小的能量完成手写任务。所设计的非直接视觉反馈实验方案,验证了高年级儿童具有更强的空间位置感。提出的特征指标在衡量手写运动熟练度发展方面具备有效性。(2)手写运动损伤特征分析：神经退行性疾病会造成脑功能的持续损伤,药物治疗对病情有一定控制作用但仍然不能阻止病程的发展,探索新型有效的生物标记,实现该类疾病的早期诊断以及病程跟踪,具有重要意义。本文以神经退行性疾病中具有代表性的帕金森病为对象,根据帕金森病的典型临床症状并结合传统量表的评价指标,设计了帕金森病患者手写运动的定量检测指标。帕金森病患者在手写运动患者在任务执行时间、速度、加速度、分段宽度及面积、极径曲线的方向变化次数NCR、极角频谱峰值频率PFA等指标上表现出显著差异,并验证了手写任务的设计对度量指标的影响。表明手写运动量化检测方法,有助于发现潜在的人体精细运动功能异常变化情况,实现基于手写运动的辅助诊断及病情发展跟踪。(3)基于手写运动特征和机器学习算法的辅助分类诊断：传统的评分量表难以便捷量化地监测病情进展,基于统计分析方法的单一性特征指标分析也无法准确地提供组间的有效区分,更具智能化的筛查检测工具有待探索。本文基于手写运动特征统计分析的结论,通过特征筛选和组合分类,构建手写运动特征集,引入机器学习方法实现群体分类,并分析了机器学习方法、手写实验任务、特征类型对分类效果的影响。利用手写运动检测结果进行手写障碍的自动化分类筛查,提高分类识别的准确度和有效性,有助于为相关疾病的诊断评估寻求新的技术手段提供思路。综上所述,本文对不同群体提出了若干手写运动评价检测指标,实现对儿童手写能力进化发育以及帕金森病患者手写运动能力退化的评估,并构建了手写运动采集与分析系统。本工作将持续积累健康群体手写运动样本,促进基于计算机的量化评分模型与临床应用相结合,为相关病患的辅助诊断及药效评估、手写运动障碍干预训练提供支撑。
[Abstract]:Handwriting movement as a unique fine motor control of human advanced nervous system, and the function of human brain, cognitive level has close relationship. The analysis of existing methods of handwriting movement, study the existence of technical means of a single, less attention to health groups, writing ability evaluation system is not comprehensive. This paper focuses on the function of the body in the rapid development of children well, the brain function by neurodegenerative diseases in patients with sustained damage in these two groups, to explore the features of handwriting movement in human nervous system development and the stage of disease, in order to understand the dynamic change characteristics of hand movement function. According to the development of handwriting handwriting movement, sports injury characteristic analysis and carry out quantitative classification, the main research contents are as follows: (1) study on the development of sports ability of children: children are thought as handwriting handwritten expression, to carry out a study The basic method in children's cognitive, plays an important role in the process of the development of psychological research on handwriting movement. Development of ordinary children, contribute to the establishment of children's handwriting movement norm data, and for the different age of sick children handwriting movement analysis provides comparative data source effectively. This paper analyzes children's handwriting three-dimensional force information handwritten application of force in the process of movement and put forward based on coherence characteristics of motion parameters, combining kinematics, dynamics of the quantitative description of handwritten ability development of school-age children. The results show that the movement time of high grade children has shorter, higher speed, less speed change curve changes direction, more inclined to the writing center near the plane drawing copy, complete the task force and applied to handwritten smaller less energy. The non direct visual feedback experiment scheme, verification The spatial position of high grade children has a stronger sense. Characteristic index proposed effective measure in handwriting movement proficiency development. (2) analysis of sports injuries of handwriting: damage sustained neurodegenerative diseases can cause brain function, drug therapy has certain effects on disease control but still can prevent the development of disease. To explore new effective biomarkers, the early diagnosis of disease and disease tracking, which is of great significance. In this paper, neurodegenerative diseases typical of Parkinson's disease as the object, according to the typical clinical symptoms of Parkinson's disease and combined with the traditional evaluation index scale, designed for patients with Parkinson's disease handwritten quantitative index movement. The patients with Parkinson's disease in handwriting movement patients in the task execution time, speed, acceleration, segment width and area, change the direction of polar radius curve number NCR The polar angle, the peak of the frequency spectrum of PFA and other indicators showed significant differences, and verified the design of the writing tasks on metrics. That handwriting movement quantitative detection method, is helpful to find the abnormal changes of the fine motor function of the human body potential, realize the auxiliary diagnosis and condition based on handwritten motion track (3.) auxiliary classification diagnosis handwriting movement features and machine learning algorithm based on scale to convenient quantitative monitoring the disease progression in the traditional single feature index analysis method of statistical analysis can accurately distinguish between groups provided for based on the screening tool more intelligent to be explored. Based on the analysis of handwriting movement statistics based on the conclusion, through feature selection and combination classification, construction of handwriting movement feature set, machine learning method is introduced to realize the population classification and analysis of machine learning Method of handwritten experimental task, influence of feature types on the classification results. By using the motion detection results to classify handwritten handwriting disorder screening automation, improve the accuracy and validity of the classification is helpful to diagnose diseases evaluation for new technology and ideas. In conclusion, this paper puts forward some handwritten motion evaluation the detection index for different groups of children's development and evolution ability of handwritten handwriting assessment in patients with Parkinson's disease exercise capacity degradation, and construct a handwritten motion data acquisition and analysis system. This work will continue to accumulate health groups handwriting movement samples, promote quantitative scoring model combined with clinical application of computer aided diagnosis and efficacy is based on Assessment of patient, handwriting movement disorder intervention training to provide support.

【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】：2016
【分类号】：R87;TP181

【参考文献】