多模态MRI影像在矿难创伤后应激障碍研究中的应用

发布时间：2018-04-05 15:06

本文选题：脑皮层厚度　切入点：血流灌注　出处：《第四军医大学》2012年博士论文

【摘要】：创伤后应激障碍(Post-traumatic Stress Disorder, PTSD)指由突发性、威胁性或灾难性事件引起人体延迟出现并长期持续存在的精神障碍。常常表现为再度体验创伤，易激惹和回避行为，并常伴有睡眠障碍、注意力难以集中、警觉性过高、过度惊吓反应等症状。大部分人在创伤后数天至六个月内发病，经过一段时间调整后都可以恢复正常。据报道，在美国，约有3.6%的人曾患有PTSD，而其中7.8%的患者终身患病。引发创伤的事件包括战争、暴力犯罪、性侵害、交通事故、失业、离婚、自然灾害等，PTSD患者多为直接或间接接触创伤事件的幸存者、目击者和救援者。有研究表明，不同类型的创伤可能影响PTSD严重程度和症状表现。目前，PTSD已成为流行病学、心理学、社会医学等多个学科的研究热点。尤其是，近年来，矿难、地震、海啸、恐怖袭击频发，这些突发的灾难性事件所引起的PTSD越来越受到人们的关注。目前，各国科学家已通过遗传、神经、生化、内分泌、心理及社会等多方面的研究，对PTSD病理、病因学机制进行了探讨，但具体机制仍不甚明晰。同时，由于对介导PTSD症状的大脑环路缺乏清晰的认识，目前仍未有针对PTSD的有效治疗方法。近年来，影像医学的发展为脑部疾病的研究提供了新的手段。特别是磁共振成像(Magnetic Resonance Imaging, MRI)已逐渐从结构成像扩展到功能成像，从单模态成像发展到多模态成像。从多方面来分析患者的结构变化、血流灌注、功能代谢和功能连接等信息，将MRI的多种模态相结合，并应用到脑部疾病的诊断和评价，是目前的研究热点和难点。现已有研究通过影像手段，初步发现了PTSD患者在脑部结构和功能方面不同程度的病理性损害。对于精神类疾病，在疾病引发实质性病变之前，往往会先出现功能代谢、血流灌注和功能区形状(如脑皮层厚度)的改变。因此，对脑皮层厚度和脑血流的分析往往可以检测出疾病发生早期对人体结构和功能的影响。各向异性分数(fractional anisotropy，FA)和表面弥散系数(apparent diffusion coefficient,ADC)可以有效反映微结构及分子水平的病理、生理学改变，能够检测早期的病变引起的脑损伤。但目前，尚未见基于MRI影像，对矿难创伤后PTSD所引起脑皮层厚度、血流及DTI指标变化的研究。本研究针对上述问题，利用矿难幸存者的MRI结构、灌注和弥散张量成像(Diffusion Tensor Imaging, DTI)数据，进行脑皮层厚度、脑血流及DTI指标分析，检测矿难创伤后早期PTSD(MRI扫描时间为事故发生后6个月)对脑部结构和功能的影响；并进一步研究这些结构和功能的变化与PTSD症状的严重程度的相关关系及相关程度。本研究被试者为2007年河南省三门峡市东风井，因河床水通过采空区涌入井下，而被困1400米深地下72小时的幸存者。参加本研究的20名幸存者中10人满足DSM-IV中描述的PTSD指标，为PTSD组；另外10人不满足，为非PTSD组。PTSD病情的严重程度通过创伤后应激障碍量表(Clinician-administered PTSD Scale, CAPS)进行评价。本研究主要包含以下研究内容： 1、脑皮层厚度的准确测量及分析在已有研究的基础上，本文构建了基于拉普拉斯方程的脑皮层三维厚度检测流程和方法，提高了计算速度和测量的准确性。首先考虑到DARTEL标准化算法对脑部细微结构进行配准具有更高的准确性，检测细微结构的异常变化具有更高的敏感性，本研究中利用VBM-DARTEL工具包对MRI的T1序列图像进行预处理，得到分割后的脑灰质、白质和脑脊液数据。并在预处理结果的基础上，采用改进的基于拉普拉斯偏微分方程的三维厚度测量方法测量脑皮层厚度，该方法通过在脑皮层内部构建势能场，并利用垂直等势面场线的长度来定义皮层厚度，使内、外表面具有唯一可逆对应点，避免了最短欧氏距离引起的测量误差，是目前脑皮层厚度测量最准确、有效的测量方法之一。但由于大脑脑沟区域的复杂形态变化和数据采集过程中各种因素的影响，容易造成分割误差。因此，在测量脑皮层三维厚度之后，采用脑沟厚度校正方法对可能存在误分割的脑沟区域进行校正，并根据校正后的灰质图像，重新测量脑皮层厚度，从而得到更准确的三维厚度值，并构建三维脑皮层厚度图像。在构建脑皮层结构和三维厚度图像的基础上，利用SPM对两组间的体积及皮层厚度进行统计学分析。由于在人口学特征分析中，两组除年龄外，其他特征都不存在统计学差异，因此，本研究以年龄作为协变量，通过两样本t检验进行分析，未发现两组的体积存在统计学差异的区域。但对脑皮层厚度的统计分析表明，较非PTSD组，PTSD组在左脑顶叶、右脑额下回和右脑海马旁回区域脑皮层厚度明显变薄。为了探索PTSD严重程度与脑皮层厚度间的相关关系，本研究采用穿过厚度变薄区域的场线上的所有体素，作为感兴趣区域(Regional of Interest,ROI)。通过ROI分析发现，PTSD组的10名被试者的右脑额下回的平均厚度与PTSD症状的严重程度(CAPS值)明显负相关。 2、脑血流(Cerebral Blood Flow, CBF)的准确测量及分析通过灌注成像扫描序列的筛选，本研究使用无创、标记效率高的脉冲式动脉自旋标记(Pulsed Artery Spin Labeling, PASL)扫描序列来分析矿难后早期PTSD患者的CBF变化。为此，，本研究在已有研究的基础上，构建了PASL序列CBF准确测量方法。首先，由于PASL序列图像成像速度快，图像分辨率较低，本研究利用SPM分别对PASL序列图像和T1序列结构图像进行预处理，并将结构像信息应用到功能像当中，提高标准化的准确性和功能图像的分辨率。经过预处理后，图像分辨率有所提高，但仍受到部分容积效应(Partial Volume Effect,PVE)的严重影响。为此，本研究首次利用PASL序列的时间信息，构建了PASL时间序列，并使用MAP-EM算法分析时间序列中每个体素点中的混合组织参数，并结合CBF的测量方法，在对图像进行部分容积校正的基础上，获得更准确的CBF分布图。获得CBF分布图后，通过分析，未发现两组的平均CBF统计学差异。但考虑到年龄对CBF的影响，本研究以年龄作为协变量，采用SPM进行两样本t检验，发现PTSD组，较非PTSD组，在右脑额叶区域CBF明显增高。但通过ROI分析，未发现PTSD组在该区域的平均CBF值与PTSD症状严重程度(CAPS值)存在相关关系。 3、DTI数据分析本研究使用DTI Studio和SPM联合分析了矿难幸存者的DTI序列数据，发现PTSD组，较非PTSD组，在右脑侧脑室、右脑梭状回、右脑颞上回和右脑颞中回区域FA值明显降低，在右脑颞上回和右脑胼胝体区域ADC值明显增高。但通过ROI分析，未发现PTSD组在这些区域的平均FA值或ADC值与PTSD症状严重程度(CAPS值)存在相关关系。这可能是由于PTSD早期引起的结构及功能变化不明显，FA和ADC指标尚无显著变化所导致的。 4、小结本研究首先构建了基于拉普拉斯方程的脑皮层厚度检测流程和方法，提高了计算速度和测量的准确性；考虑到灌注成像的分辨率低及PVE对CBF测量的影响，将基于MAP-EM算法的部分容积校正理论与CBF测量方法相结合，得到更准确的CBF分布图，上述方法为研究神经系统或脑部疾病引起的脑皮层厚度和血流变化提供了有效的手段。基于提出的方法，本研究对经历矿难的20名幸存者构成的PTSD组和对照组的脑皮层厚度和CBF分布进行了统计分析，发现了多个厚度或CBF存在差异的的区域。同时，使用DTI Studio和SPM联合分析了DTI序列数据，发现了多个FA和ADC值异常的脑区。以上结果一方面可为进一步理解PTSD病理、病因学机制奠定基础，另一方面，提出的区域或指标，有望进一步成为早期PTSD临床诊断和评价的依据。
[Abstract]:Posttraumatic stress disorder (Post-traumatic Stress, Disorder, PTSD) refers to a sudden, threatening or catastrophic events caused by human mental disorders in delayed and long-term existence. Is often re experience the trauma, irritability and avoidance behavior, and often accompanied by sleep disorders, difficulty concentrating, high alertness, excessive shock the reaction and other symptoms. Most of the people in a few days to six months after trauma in the onset, after a period of time after the adjustment can be restored to normal. It is reported that in the United States, about 3.6% of people who suffer from PTSD, of which 7.8% of the patients. The lifetime risk for incident triggered trauma including war, violence, sexual assault, traffic accident, unemployment, divorce, natural disasters, PTSD patients are direct or indirect contact with trauma survivors, witnesses and rescuers. Studies have shown that different types of trauma may influence the severity of PTSD And symptomatic performance. At present, PTSD has become a research hotspot in many disciplines such as epidemiology, psychology, social medicine and so on. Especially in recent years, PTSD caused by disasters, earthquakes, tsunamis and terrorist attacks has been more and more concerned.
At present, many scientists have been using genetic, biochemical, neural, endocrine, research on psychological and social aspects of PTSD pathology, etiology, mechanism is discussed, but the mechanism is not very clear. At the same time, because of the symptoms of PTSD mediated brain circuits lack a clear understanding, at present there is no effective treatment for PTSD.
In recent years, the development of medical imaging provides a new tool for the study of brain diseases. Especially, magnetic resonance imaging (Magnetic Resonance, Imaging, MRI) has gradually changed from structure imaging is extended to functional imaging, from single modality imaging to multi modality imaging. To analyze the structure changes of patients with blood perfusion function from several aspects. The metabolic and functional connectivity information, combine multi modality of MRI, and applied to the diagnosis and evaluation of brain diseases, is a research hotspot and difficulty at present. We have studied by imaging techniques, preliminary findings in patients with PTSD in brain structure and function of different degree of pathological damage.
For mental illness, before causing substantive changes in the disease, often appear before the function of metabolism, shape perfusion and functional areas (such as cortical thickness) change. Therefore, the analysis of the thickness and blood flow to the brain cortex can often be detected early influence to the human body structure and function of the anisotropic disease. Score (fractional anisotropy, FA) and apparent diffusion coefficient (apparent diffusion, coefficient, ADC) can effectively reflect the pathological microstructure and molecular level, physiological changes, can cause brain damage, early detection of the lesions. But at present, has not yet been based on MRI image, the thickness of the mine caused by cerebral cortex PTSD after trauma. Study on the changes of blood flow and DTI index.
Based on the above problems, the use of MRI structure mining survivors, perfusion and diffusion tensor imaging (Diffusion Tensor, Imaging, DTI) data of cortical thickness analysis, cerebral blood flow and DTI index, early detection of PTSD mine after trauma (MRI scan time was 6 months after the accident) effects on brain structure and function the degree of correlation; and the severity of changes and further study of these structures and functions with PTSD symptoms.
The research subjects for Dongfeng well Sanmenxia city of Henan province in 2007, because the river water through the mined out area into the underground, trapped 1400 meters deep underground for 72 hours. The survivors of the 20 survivors of 10 people participated in this study meet the PTSD criteria described in DSM-IV and PTSD group; the other 10 were not satisfied the PTSD severity for non PTSD group.PTSD severity scale (Clinician-administered PTSD Scale, CAPS) were evaluated.
The main contents of this study are as follows:
1, accurate measurement and analysis of cerebral cortex thickness
On the basis of the existing research, this paper constructs the 3D cortical thickness detection method and process based on the Laplasse equation, the computation speed and improves the accuracy of measurement. Firstly, considering the DARTEL standard algorithm has a higher accuracy on the fine structure of the brain registration, detection of abnormal changes in the fine structure has higher sensitivity, T1 image sequence of MRI pretreated by VBM-DARTEL kit in this study, obtained after segmentation of brain gray matter, white matter and cerebrospinal fluid data. And based on the pretreatment result, the improved thickness measurement method based on 3D Laplasse partial differential equation of cortical thickness measurement, the method by constructing the potential field in the cerebral cortex inside, and the vertical equipotential surface field line length to define the cortical thickness, the inner, outer surface has only reversible corresponding points and avoid the shortest Euclidean distance The measurement error caused by the cortical thickness measurement is the most accurate, one of the effective measures. But due to the influence of various brain sulcal regions complex morphological changes and factors in the process of data collection, easy to cause the segmentation error. Therefore, after measuring the cortical thickness of three-dimensional, by sulcus thickness correction method for possible correction of sulcal regions segmentation error, and according to the gray image after correction, re measurement of cortical thickness, so as to obtain more accurate values of 3D thickness, cortical thickness and the construction of three-dimensional images.
In the foundation of the structure of the cortex thickness and three-dimensional image, the statistical analysis on the volume and cortical thickness between the two groups using SPM. The analysis on the demographic characteristics of the two groups except for age, other characteristics have no significant differences, therefore, in this study, age as a covariate, through the analysis of two samples t test, no significant differences between the two groups of the volume of the region found. But the statistics of cortical thickness analysis showed that compared with non PTSD group, PTSD group in the left parietal lobe, right inferior frontal gyrus and right parahippocampal gyrus regional cortical thickness was significantly thinner.
In order to explore the relationship between PTSD and the severity of cortical thickness, this study used across the thickness of the field line all voxels as region of interest (Regional of, Interest, ROI). The ROI analysis showed that the severity of the average thickness of the PTSD group of 10 subjects of right frontal gyrus the symptoms of PTSD (CAPS) was significantly negative correlation.
2, accurate measurement and analysis of cerebral blood flow (Cerebral Blood Flow, CBF)
By screening the perfusion imaging scanning sequence, this study using noninvasive, pulsed arterial spin labeling labeling efficiency high (Pulsed Artery Spin Labeling, PASL CBF) changes in patients with early PTSD scan sequence analysis after the accident. Therefore, the study on the basis of existing research, construct the accurate measurement method of PASL sequence CBF.
First, the imaging speed of PASL image sequence, image resolution is low, the research on PASL sequence structure image and T1 image sequences are processed by SPM, and will be applied to function like structures like information, improve the standardization and accuracy of image resolution. After preprocessing, image resolution improved, but is still subject to partial volume effect (Partial Volume, Effect, PVE) is seriously affected. Therefore, this study is the first to use the time information of PASL sequences, constructed the PASL time series, hybrid organization parameters using MAP-EM algorithm and analysis of time series in each voxel point, combined with the method of CBF. Based on the partial volume correction of the image, get the CBF distribution more accurately.
CBF map, through the analysis, the average CBF was not different between two groups. But considering the influence of age on CBF, based on the age as a covariate, using SPM two sample t test group PTSD, compared with non PTSD group, in the right frontal lobe area increased obviously. But by CBF the analysis of ROI, not PTSD group in the region average CBF value and the PTSD symptom severity was found (CAPS) correlated.
3, DTI data analysis
This study used DTI Studio and SPM combined analysis of DTI sequence data mining accident survivors, PTSD group, compared with non PTSD group, in the right side of the right ventricle, fusiform gyrus, lower back area FA value of right temporal gyrus and right middle temporal, right superior temporal gyrus and right brain in the corpus callosum area ADC value was significantly higher. But through the analysis of ROI, PTSD group in the average FA in these regions or ADC values and the severity of PTSD symptoms was found (CAPS) correlated. This may be due to early PTSD caused by the change of structure and function is not obvious, caused by the FA and ADC index had no significant change.
4, a summary
This study first constructs the cortical thickness of the Laplasse equation and the detection process based method, improves the accuracy and computation speed measurement; considering the influence of low resolution and PVE perfusion imaging on CBF measurement, the partial volume correction algorithm of MAP-EM theory and CBF measurement method based on the combination of figure CBF distribution more accurately the changes of cerebral cortex thickness and blood flow, the method for the study of nervous system and brain diseases caused by providing effective means.
The proposed method based on the study of 20 survivors experience mine consisting of PTSD group and control group in cerebral cortex thickness and the distribution of CBF was analyzed, found a number of regional differences in the thickness or CBF. At the same time, the use of DTI Studio and SPM combined analysis of DTI sequence data, found a number of FA and ADC value of brain abnormalities. The above results on the one hand to further understand the etiology of PTSD pathology, lay the foundation of mechanism, on the other hand, regional or index is put forward, is expected to become the early clinical diagnosis of PTSD and evaluation basis.

【学位授予单位】：第四军医大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：R749.5

【引证文献】