多尺度分割对脑连接组分析的影响  

Influence of Multi-Scale Parcellation on Connectome Analysis

作　　者：靳聪[1] 林岚[1] 付振荣[1] 宾光宇[1] 高宏建[1] 吴水才[1] 

机构地区：[1]北京工业大学生命科学与生物工程学院,北京100124

出　　处：《中国医疗设备》2015年第6期18-22,64,共6页China Medical Devices

基　　金：北京市自然科学基金(7143171)资助

摘　　要：脑连接组分析可以帮助我们认识、分析人类的大脑。为探究多尺度分割对脑连接组分析的影响,本文从3个尺度(32个节点,128个节点和512个节点)对脑网络中的特征进行分析,发现了稳定存在的小世界特征,以及一些网络特征的变化趋势。如最短路径长随节点个数的增加而增加,聚类系数随节点个数的增加而减少,节点度呈现先上升后下降的趋势。此外,我们发现多尺度分割对于网络的模块化结构影响不显著,但对于网络中的hub节点有一定的影响。The connectome analysis can help us to have a better understanding of human brain networks. In order to detect the inlfuence of the multi-scale parcellation on connectome analysis, this paper studied the brain network from three scales（32 nodes, 128 nodes and 512 nodes）. The results revealed that small-world property existed over those brain networks, and the network features showed some trends related to the parcellation scale. For example, the shortest path length increased with the decrease of the parcellation scale, clustering coefifcient decrease with the decrease of the parcellation scale, while the degree increased firstly and then decreased. Additionally, this paper also found that multi-scale parcellation had no signiifcant effect on the network modularity structure, but had somewhat impact on hub regions.

关 键 词：磁共振成像 脑连接组 多尺度分割模板 弥散张量成像 脑网络特征 

分 类 号：R197.39[医药卫生—卫生事业管理]