基于磁共振成像的静息态脑功能网络研究及其在重性抑郁症中的应用

发布时间：2018-03-28 18:49

  本文选题：静息态功能磁共振成像　切入点：重性抑郁症　出处：《第四军医大学》2017年博士论文

【摘要】：重性抑郁障碍(major depressive disorder,MDD)是一种常见的精神疾病。随着对于抑郁症的认识逐渐深入,目前的观点认为MDD是大脑中异常连接所导致的多脑区功能紊乱。基于功能磁共振成像(functional magnetic resonance imaging,fMRI)的研究发现大脑可以被分解为不同的功能网络,各个功能网络覆盖的脑区在结构上不相连,然而在功能上却存在紧密联系。功能网络在大脑处于任务状态和静息状态下都具有相似的分布模式。以功能网络为基础研究大脑功能的重组和交互,能够从网络连接的角度解释多脑区功能紊乱。因此静息态功能网络连接分析对精神疾病研究具有重要意义。MDD患者功能网络内的功能连接异常已经被许多研究报道过。默认网络(default mode network,DMN)、凸显网络(salience network,SN)和中央执行网络(central executive network,CEN)在MDD患者中均存不同程度的异常。脑区的异常能够通过神经连接传递到另外的脑区,因此研究异常脑区之间神经活动的作用方向,对理解疾病发病机制具有重要意义。功能连接反映脑区活动的相关性,却不能提供神经活动的方向信息。有效连接分析方法能够估计神经活动的连接方向及强度,但是需要基于先验知识构建模型。因此探究不同功能网络内部节点间的有效连接是功能网络研究的热点问题。大脑处于不同的功能状态时,功能网络的活跃程度也存在差异。比较功能网络的活跃程度在时间上的变化关系,能够得到网络之间的连接关系。功能网络之间的连接关系反映不同大脑功能之间的协同性。这种大脑功能网络的协同性在许多精神疾病中都发生了异常。功能网络之间的协同作用也是大脑功能网络研究的一个热点问题。皮层下深部脑区对大脑皮层具有广泛的神经投射。深部脑区与皮层脑区在功能上的连接模式被称为功能投射网络。深部脑区对功能投射网络的活动具有调节作用。根据深部脑区与功能投射网络的连接关系,能够对深部脑区进行功能划分。研究疾病相关脑区与功能投射网络的连接关系,以此进行脑区的功能划分,对深部脑刺激技术(deep brain stimulation,DBS)等疾病干预技术有重要的指导作用。针对上面提出的三个问题,本文的研究主要从功能网络内、功能网络间和深部脑区的功能投射网络三个角度研究了MDD患者的功能网络异常。论文的主要工作包括:第一章为静息态功能网络的有效连接分析。前期研究表明MDD患者在DMN、CEN和SN均出现连接异常。为了发现这些网络异常脑区间神经活动的作用方向和因果关系。本文首次采用频谱动态因果模型(spectral dynamic causal modeling,spDCM)分析方法对功能网络内的有效连接进行分析,并对比了MDD患者治疗前、治疗后和健康对照组之间的差异。通过独立成分分析(independent component analysis,ICA)方法分解出静息态功能网络,并选择出与MDD密切相关的静息态功能网络。再根据静息态功能网络的空间分布模式,分别选择出网络内部的关键节点。最后根据网络内部关键节点构建DCM模型,使用spDCM估计功能网络内关键节点之间的有效连接。结果发现MDD患者SN和DMN内部均存在有效连接异常。有效连接的结果表明,MDD患者DMN左侧顶叶皮层(left parietal cortex,LPC)神经活动的调控作用降低。SN腹侧子网络内的右侧前脑岛(right anterior insula,RAI)区域的神经活动调控能力降低。SN前部子网络的左侧额叶皮层(left frontal cortex,LFC)存在异常的调控作用。第二章引入DCM方法对SN前部子网络、SN腹侧子网络、DMN前部子网络、DMN后部子网络、CEN左侧子网络和CEN右侧子网络之间的有效连接进行建模,分别使用随机DCM(stochastic dynamic causal modeling,sDCM)和spDCM对模型参数进行估计。并与Pearson相关法和最大延迟相关在MDD患者治疗前、治疗后和健康对照组的结果进行了对比。DCM模型选择的结果显示,治疗前MDD组、治疗后MDD组和健康对照组的功能网络之间均存在因果联系。DCM的结果验证了SN对DMN与CEN相反的调节作用在MDD患者中出现异常,发现了SN与CEN之间的调节回路,以及SN与DMN后部子网络对DMN前部子网络相反的调控作用。同时也发现治疗前MDD患者的DMN后部子网络对CEN左侧子网络和CEN右侧子网络的异常激励作用,这种异常调节作用在治疗后也得到改善。第三部分使用功能投射网络分析方法研究了丘脑内体素与功能投射网络的连接关系。首先计算丘脑内每个体素的功能连接图。再使用ICA方法估计出丘脑的全脑功能投射模式,并从这些模式中选择出SN前部子网络、SN腹侧子网络、DMN前部子网络、DMN后部子网络、CEN左侧子网络、CEN右侧子网络和运动网络。对原有方法进行改进,使用有约束的多变量回归模型估计每个丘脑体素与各个功能投射网络的连接关系。最后根据丘脑体素与各个网络连接的分布情况,对比MDD和健康对照组的丘脑的功能分区差异。结果发现MDD组右侧丘脑内侧前部区域与CEN右侧子网络的连接出现异常降低。丘脑腹部后侧部分区域与运动网络之间的连接在MDD患者中也低于健康对照组,但是该结果在经过多重比较校正后并不显著。本文针对大脑功能网络分析问题,使用功能网络内有效连接分析、功能网络间的连接分析和深部脑区的功能投射网络分析方法,研究MDD患者的大脑功能网络异常。功能网络内有效连接分析定位了导致网络异常的关键节点。功能网络间的连接分析研究了网络间的异常调控作用关系。功能投射网络分析研究了丘脑内功能投射发生异常的区域。这三个种分析方法构成了一个较为全面的功能网络分析框架。
[Abstract]:Major depressive disorder (major depressive, disorder, MDD) is a common mental disease. As for depression gradually deepening understanding, the current view is that MDD is dysfunction in multiple brain areas caused by abnormal connections in the brain. Based on functional magnetic resonance imaging (functional magnetic resonance imaging, fMRI) the study found that the brain can be the network function is decomposed into different brain regions, each functional coverage of the network is not connected in the structure, but in function are closely linked. Functional networks have similar distribution patterns of task state and resting state in brain function. In the network as the foundation to the study of brain function reorganization and interaction, from the network the connection point of interpretation function of multiple regions of the brain disorder. Resting state functional connectivity so it is important to analyze the function of patients with.MDD in the network research work on mental illness Can an abnormal connection has been proved by many studies reported. The default network (default mode network DMN (salience), network SN, highlighting the network) and the central executive network (central executive network, CEN) in the patients with MDD were abnormal in different degrees. The abnormal brain regions by neural connections can transfer to other brain regions therefore, the direction of neural activities between abnormal brain regions, has important significance in understanding the pathogenesis of diseases. The correlation of functional connectivity reflects brain activity, but can not provide the direction information of neural activity. Effective connectivity analysis method to estimate the connection direction and strength of neural activity, but the need to model construction based on the prior knowledge. Therefore explore the effective connection of internal nodes of different function between networks is a hot topic in the research function of the network. The brain in different functional states, active function of the network also There are differences between the degree of active function. The changes of network in time, can get the network connections between the connection between the network function. Reflect the synergy between different brain functions. The brain functional network coordination in many psychiatric disorders are abnormal. A hot issue between the function of the network the synergistic effect is the brain functional network research. Subcortical deep brain regions with extensive neural projection to the cerebral cortex. The connection mode of deep brain area and cortex function is called functional projection activities on the network. The network function mapping in deep brain regions play a role in the regulation of the network connections. According to the deep Department of brain and the function of projection, can be divided the function of deep brain regions. The connection network of brain areas related to disease and function of projection, the brain function of stroke Points of deep brain stimulation (deep brain stimulation, DBS) and other disease intervention technology has an important guiding role. Aiming at the three problems mentioned above, this paper mainly from the function of network, network function and deep brain function network projection three angles to research the function of network with MDD abnormal. The main work includes: the first chapter is the effective connectivity of resting state functional network analysis. Previous studies showed that MDD patients in DMN, CEN and SN showed abnormal connections. In order to find these network abnormal brain nerve activity interval direction and causal relationship. For the first time by the spectrum of dynamic causal model (spectral dynamic causal modeling, spDCM) were effective connection to function in the network analysis, and compares the differences between the MDD patients before treatment, after treatment and healthy control group. Through independent component analysis (in Dependent component analysis, ICA) method of decomposition of resting state functional network, and the choice of the resting state functional network closely related with MDD. According to the distribution pattern of resting state functional network space, respectively select key nodes within the network. Finally, according to the internal network of key nodes to build DCM model, using spDCM to estimate the effective connection between the key function of nodes within the network. The internal SN and DMN in patients with MDD were abnormal. The effective connection of effective connection. The results showed that MDD patients with DMN left parietal cortex (left parietal cortex, LPC) Regulation of neural activity decreased.SN ventral sub network in the right anterior insula (right anterior, insula, RAI) nerve regulating the activities of regional capacity to reduce.SN sub network in front of the left frontal cortex (left frontal cortex, LFC) in regulation of anomalies. The second chapter introduces the method of DCM SN Sub network, SN ventral sub network, DMN front DMN rear sub network, sub network between CEN and CEN on the right side of the left sub network connection sub network effectively modeled respectively using stochastic DCM (stochastic dynamic causal modeling, sDCM and spDCM). The model parameters were estimated and correlated with Pearson method and the maximum delay in MDD patients before treatment, after treatment and healthy control group compared.DCM model selection showed that MDD group before treatment, there was a causal relationship between the function of.DCM network of MDD group and control group after treatment, the results show that the SN of DMN and CEN opposite role in abnormal MDD patients, were found between SN and CEN and SN and the DMN loop, the rear sub network regulation of DMN front sub network opposite. Also found before treatment in patients with MDD DMN CEN on the left rear sub network subnet The abnormal incentive effect of collaterals and the right of CEN sub network, the abnormal regulation is also improved after treatment. The third part studies the connection between the use and function of the thalamus voxel projection network analysis function mapping network. First calculate the function within the thalamus of each voxel connected graph. Then use the ICA method to estimate the total brain function projection patterns of the thalamus, and the choice of the SN front sub network from these patterns, SN ventral sub network, DMN front sub network, DMN CEN left rear sub network, sub network, CEN network and sub network. The right movement for improvement on the original method, using the constrained multivariable regression model estimation connection between each voxel and thalamus each function projection network. Finally according to the distribution of thalamic voxel connected to each network, function zoning difference comparison between MDD and thalamus in healthy control group. The results showed that MDD group On the right side of the front area of the medial thalamus and CEN right sub network is abnormal. Between the posterior portion of the thalamus decreased abdominal area and the movement of network connection in MDD patients was lower than that of the control group, but the results after correction for multiple comparisons was not significant after. According to the brain functional network analysis, network connection function analysis, link analysis and deep brain function mapping network function between the network analysis method, the brain functional network of MDD patients with abnormal function in the network. The network connection effective positioning analysis leads to different key nodes. The abnormal regulation of the relationship between the network connection between the network analysis function. The function of projection network analysis of the abnormal region of the thalamus. These three functional projection methods constitute a comprehensive analysis framework. The function of network

【学位授予单位】：第四军医大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：R445.2;R749.4
，

本文编号：1677589