复发缓解型多发性硬化患者默认网络小世界属性的fMRI研究

The ＂small-world＂ property of default mode network in relapsing-remitting multiple sclerosis：a graph theory resting-state functional network analysis

目的 采用复杂网络分析技术探讨复发缓解型多发性硬化（RRMS）患者默认网络（DMN）的“小世界”网络属性改变情况.方法 采集RRMS组和健康对照组（各24例）静息态功能磁共振（R-fMRI）数据,经过预处理后,选择DMN的20个感兴趣区（ROI）为节点,计算功能连接获得相关矩阵用于构建二值化网络,获取DMN的网络属性包括集群系数、特征路径长度和“小世界”指数,并分析存在差异的网络属性与患者临床参数如临床扩展残疾量表（EDSS）、同步听觉连续加法测验评分（PASAT）、脑实质分数（BPF）、T2可见病灶体积、病程的相关性.结果 RRMS组和健康对照组DMN均存在“小世界”属性（“小世界”指数大于1.0）.在网络密度范围（Dmin：0.01∶0.48,Dmin=0.42）内,RRMS患者DMN整体集群系数轻度下降、特征路径长度轻度增加及对应的“小世界”指数轻度下降,但差异均无统计学意义（P＞0.05）.左侧后扣带回的集群系数在最小网络密度（Dmin）条件下和网络密度范围曲线下面积测量（AUC）较对照组均呈显著升高改变（P值均为0.031）.RRMS组左侧后扣带回集群系数在Dmin时及其网络密度范围AUC均与PASAT呈负相关（P值分别为0.023和0.020）.结论 RRMS患者组DMN存在”小世界”属性,左侧后扣带回集群系数升高可能是“副效应”改变.

Objective To investigate the ＂small-world＂ property of default mode network in relapsingremitting multiple sclerosis （RRMS） compared with the （matched） healthy control subjects,using a graph theory for resting-state functional network analysis.Methods Twenty four patients with RRMS and 24 age-,and sexmatched healthy controls were exanined with resting-state functional MRI （R-fMRI） and 3D-T1WI on Trio 3.0 Tesla.The R-fMRI data were preprocessed,then 20 regions of interest （ROIs） was defined and extracted from the default mode network.The functional connectivity between a pair of regions was defined as the Pearson＇s correlation coefficient in the time courses.Graph theoretical analysis was implemented and compared with the property of binary functional network.Further,the relationships were examined between the topological property of functional networks and the clinical parameters.Results In this study,functional network in the default mode network was conducted,and the ＂small-world＂ characteristics were observed in both RRMS and healthy control group.In the range of densities （Dmin：0.01∶0.48 ;Dmin =0.42）,compared with healthy control group,RRMS groups had a slightly smaller normalized clustering coefficient （P＞0.05）,a slightly larger normalized path length （P＞0.05）,and a slightly decrease ＂small-world＂ index （P＞0.05）.The left posterior cingulated cortex showed a significantly larger clustering coefficient in RRMS group （P=0.031） on the network threshold at Dmin.The area under a curve （AUC） for clustering coefficient measure curves of the left posterior cingulated cortex （density range of 0.42：0.01：0.48） between two groups were compared.The AUC result was similar to the Dmin result （P=0.031）.Negative correlation was observed between the clustering coefficient of left posterior cingulated cortex and the Paced Auditory Serial Addition Test （PASAT） in both the Dmin（P=0.023） and the AUC （P=0.020）.Conclusion The default mode network has ＂small-world＂ property in RRMS group.The significantly increased clustering coefficience is observed in the left posterior cingulated cortex,which may be a ＂side effect＂.