基于格兰杰因果的精神分裂听觉门控脑网络的研究

发布时间：2018-06-15 16:40

本文选题：精神分裂症 + 脑网络　；参考：《天津医科大学》2017年硕士论文

【摘要】：研究目的:精神分裂是一种严重的精神障碍类疾病,相关研究表明,脑区之间的不正常连接、紊乱的振荡活动和同步性的改变是精神分裂症的显著特征。为了研究精神分裂患者在听觉门控过程中的脑网络连接状态,本文利用基于频域的格兰杰因果方法分析了听觉条件-测试范式下门控过程中脑电信号因果连接特性,并利用图论参数分频段量化分析了精神分裂患者门控过程中的全局和局部网络连接特征,探索是否存在脑网络的特异性标志对精神分裂症患者在临床上的诊断和治疗进行评估,为定量分析精神分裂的脑网络和脑功能状态提供理论基础和技术支持。研究方法:本论文基于听觉条件-测试范式下32通道EEGs,应用基于Granger因果分析的DTF方法计算两组受试者的因果连接矩阵,并根据矩阵构建脑功能连接网络并定量分析了两组受试者的连接特性差异。研究方法包括:(1)实验数据:利用听觉条件-测试听觉刺激范式,分别采集20名精神分裂患者和20名正常受试者32通道的脑电;(2)对32通道的EEGs数据进行预处理,去除工频干扰、基线漂移等噪声和伪迹;(3)对预处理后的数据利用基于频域的格兰杰因果的直接传递函数方法计算从通道j到通道i的因果连接矩阵,构建出了因果连接网络;(4)计算全脑的平均因果连接强度,并计算额区、中央区及顶区的脑区内部和脑区之间的因果连接强度;(5)计算两组受试者的集群系数和全局效率等因果网络评价参数;(6)计算整个脑区的信息流增益情况,画出信息流增益的空间分布;(7)计算两组受试者的rich-club系数,判断是否存在rich-club特性。研究结果:1.各频段的平均因果连接强度在High-beta和Gamma频段,SZ组(精神分裂受试组)的DTF_(mean)显著高于HC组(健康对照组),并且存在显著性差异。2.各脑区平均因果连接强度在中央区内部及中央区到顶区的连接中,在High-beta和Gamma频段SZ组的DTF连接值显著大于HC组;在中央区到额区的连接中,在各频段SZ组的DTF连接值均显著大于HC组;在顶区内部的连接中,在Theta和Alpha频段SZ组的DTF连接值均显著大于HC组。3.因果网络的定量描述在High-beta和Gamma频段,SZ组的集群系数值显著高于HC组,且SZ组的全局效率值显著高于HC组。4.各频段信息流增益两组受试者的主要信息活跃区均在中央-顶叶区域,SZ组的PO_3通道附近的信息流增益低于HC组,而SZ组Cz通道的增益值高于HC组,并且随着频段的增加,SZ组Fz通道的信息流增益变大。5.Rich-club系数SZ组和HC组均存在rich-club特性,且除High-beta频段外,SZ组的rich-club系数均显著低于HC组。研究结论:在听觉门控抑制过程中,High-beta和Gamma频段下大脑因果网络存在着过连接现象,这可能是精神分裂患者无法有效的进行听觉门控抑制的潜在原因。精神分裂患者的脑区内部与脑区之间的连接存在异常,提示精神分裂受试者的因果网络连接的异常主要在中央区内部及中央区到额区和顶区的连接。精神分裂受试者的聚集系数和全局效率在High-beta和Gamma频段均显著高于正常受试者,表明在高频段,精神分裂患者脑网络的局部连接特性和全局连接特性参数提高,提示精神分裂患者在听觉门控抑制过程中,无法有效的区分外界的有效信息和干扰信息,部分脑区被干扰信息激活,复杂网络特性出现异常。信息流增益研究进一步发现,精神分裂受试者的Cz通道和高频段额区的信息流活动增加,顶枕区的信息流活动减少。计算网络的rich-club系数发现,精神分裂患者和正常受试者都具有rich-club特性,说明在网络中存在节点度大的节点且相互之间更趋向于紧密连接。
[Abstract]:Research objectives: schizophrenia is a serious mental disorder. Related studies have shown that abnormal connections between the brain regions, chaotic oscillations and synchrony changes are significant characteristics of schizophrenia. In order to study the brain network connection status of schizophrenic patients in the auditory gating process, this paper uses the frequency domain. The Grainger causality method analyses the causality connection characteristics of the EEG in the threshold control process under the auditory condition test paradigm, and quantifies the global and local network connection characteristics in the gated process of schizophrenic patients by using the graph parameter frequency division to explore whether there is a specific sign of the brain network for the schizophrenic patients in clinical practice. The diagnosis and treatment are evaluated to provide theoretical basis and technical support for quantitative analysis of schizophrenia brain networks and brain functions. Research methods: Based on the 32 channel EEGs under the auditory conditioned test paradigm, this paper uses the DTF method based on Granger causality analysis to calculate the causal connection matrix of the two groups of subjects, and according to the matrix structure. The brain functional connection network and quantitative analysis of the differences in connection characteristics between the two groups were made. The research methods included: (1) experimental data: using auditory condition - Testing auditory stimulus paradigm, 20 schizophrenic patients and 20 normal subjects were collected separately in 32 channels of brain electricity; (2) the EEGs data of the 32 channels were preprocessed to remove the frequency interference, base Line drift and other noise and artifacts; (3) the preprocessed data is calculated using the direct transfer function method of Grainger causality based on the frequency domain to calculate the causal connection matrix from channel J to channel I, and construct a causal connection network; (4) calculate the average causal connection intensity of the whole brain, and calculate the frontal area, the inner and brain regions of the central and top areas. The strength of causal connection between the two groups; (5) calculating the parameters of the causal network evaluation of the cluster coefficient and the global efficiency of the subjects; (6) calculating the information flow gain in the whole brain, drawing the spatial distribution of the information flow gain; (7) calculating the rich-club coefficient of the two groups of subjects and judging whether there is a rich-club characteristic. The results of the study: 1. frequency bands The average causal connection intensity was in the High-beta and Gamma bands, and the DTF_ (mean) of the SZ group (the schizophrenic group) was significantly higher than that of the HC group (the healthy control group), and there was a significant difference in the mean causal connection between the.2. regions and the central region and the central area to the top area. The DTF connection value of the SZ group in High-beta and Gamma frequency bands was significantly larger. In the HC group, in the central area to the frontal area, the DTF connection value in each band SZ group is significantly greater than that of the HC group; the DTF connection value of the SZ group in the Theta and Alpha band SZ group is significantly larger than the HC group.3. causality network in the top region. The local efficiency value is significantly higher than that of the HC group.4., the main active areas of the two groups of information flow gain are in the central parietal lobe region, the information flow gain near the PO_3 channel in the SZ group is lower than that of the HC group, while the gain value of the Cz channel in the SZ group is higher than that of the HC group, and with the increase of the frequency band, the gain of the information flow of the SZ group Fz channel becomes larger.5.Rich-club coefficient S. There are rich-club characteristics in both group Z and HC group, and the rich-club coefficient of SZ group is significantly lower than that of group HC except High-beta band. Conclusion: in the process of auditory gating suppression, there is a phenomenon of over connection in the brain causality network under High-beta and Gamma frequency bands, which may be the potential of the schizophrenic patients to be unable to effectively carry out auditory gated suppression. There is an abnormal connection between the brain area and the brain area of the schizophrenic patients, suggesting that the abnormal connections between the schizophrenic subjects are mainly in the central and central areas and the central and the top areas. The aggregation coefficient and the global efficiency of the schizophrenic subjects are significantly higher in the High-beta and Gamma frequency bands than in the normal subjects. It is shown that in the high frequency segment, the local connection characteristics and the global connection characteristic parameters of the brain network in the schizophrenic patients are improved, suggesting that the schizophrenic can not effectively distinguish the external effective information and interference information in the auditory gated control process. Some brain regions are disturbed by interference and the characteristics of the complex network are abnormal. The information flow increases. It is further found that the information flow activity in the Cz channel and the high frequency section of the schizophrenic subjects is increased and the information flow activity in the top pillow area is reduced. The rich-club coefficient of the computing network shows that both the schizophrenic and the normal subjects have the rich-club characteristics, indicating that there are nodes with large nodes in the network and tend to be closer to each other in the network. Close to close connections.
【学位授予单位】：天津医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R749.3;TN911.7

【参考文献】