糖尿病患者轻度认知障碍脑电信号非线性动力学研究

发布时间：2018-04-26 22:04

本文选题：糖尿病 + 遗忘型轻度认知障碍　；参考：《燕山大学》2015年硕士论文

【摘要】：糖尿病及其引起认知功能损害的患病率正在逐年上升,严重影响到了患者的生活质量。糖尿病是导致认知功能下降和发展成痴呆的一个重要危险因素。研究糖尿病与认知功能相关的脑电信号,可加深人们对糖尿病认知功能损害机制的了解,对早期干预有重大意义。首先,研究经验模态分解的近似熵、样本熵、模糊熵、排序熵、功率谱熵和小波熵,利用Logistic映射模型进行仿真,仿真结果表明模糊熵的抗噪性较好。分析糖尿病患者遗忘型轻度认知障碍组和对照组脑电信号的非线性变化,采用单因素方差统计分析,基于曲线下最大面积的特征提取和支持向量机进行分类,对熵与神经心理学测试进行皮尔森线性相关分析。结果表明模糊熵的分类效果更好,前额、右颞和枕区为显著脑区,熵与神经心理学测试在枕区和颞区存在相关性。其次,将符号化模式与递归图相结合。符号化模式递归图对采样量及平稳性要求较低,受噪声影响小,可以直接对时间序列进行编码。利用Logistic映射和单通道双动态神经元群模型进行仿真,结果表明符号化模式递归图的确定性变量比排序递归图可以更好地刻画模型参数的变化。最后,对实际糖尿病和非糖尿病中的遗忘型轻度认知障碍组和对照组分析,结果表明符号化模式递归图比排序递归图的确定性变量能够更好地区别两组。在认知状态匹配下,发现糖尿病患者与认知功能存在某种联系。确定性变量与神经心理学测试在前额、颞区和枕区存在相关性。
[Abstract]:The prevalence of diabetes mellitus and its cognitive impairment is increasing year by year, which seriously affects the quality of life of patients. Diabetes is an important risk factor for cognitive impairment and the development of dementia. The study of the EEG associated with diabetes mellitus and cognitive function can deepen the understanding of the mechanism of diabetes cognitive impairment and have great significance for early intervention. Firstly, the approximate entropy, sample entropy, fuzzy entropy, sort entropy, power spectrum entropy and wavelet entropy of empirical mode decomposition are studied. The simulation results show that the fuzzy entropy has good noise resistance. The nonlinear changes of EEG in amnesia group and control group were analyzed. Single factor variance statistical analysis was used to extract the largest area of feature under the curve and classify it with support vector machine (SVM). Pearson linear correlation analysis of entropy and neuropsychological tests was carried out. The results showed that the classification effect of fuzzy entropy was better. The frontal, right temporal and occipital regions were significant brain regions, and the correlation between entropy and neuropsychological tests was found in occipital and temporal regions. Secondly, the symbolic pattern is combined with the recursive graph. The symbolic pattern recursion graph requires less sampling amount and stability, and is less affected by noise, so the time series can be coded directly. The simulation results show that the deterministic variables of symbolic pattern recursion graph can better describe the variation of model parameters than sort recursion graph by using Logistic map and single channel dual dynamic neuron group model. Finally, the analysis of amnesia mild cognitive impairment group and control group in actual diabetes mellitus and non-diabetes mellitus shows that the symbolic pattern recursion graph can better distinguish the deterministic variables between the two groups than the sequential recursive graph. Under the condition of cognitive state matching, it was found that diabetic patients had some relationship with cognitive function. Deterministic variables were correlated with neuropsychological tests in frontal, temporal and occipital regions.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R587.2;TP391.41

【参考文献】