不同节点尺度下基于共同邻居的功能脑网络建模方法研究

Research on Method of Brain Functional Network Modeling Based on Common Neighbor with different node scales

采用脑网络的结构特性与功能特性相结合的建模方法,探索了3种不同节点尺度下的建模效果好坏。结构特性采用解剖距离;功能特性采用共同邻居这一相似性指标,从脑网络的全局属性与局部属性的角度分析了建模效果,并提出了一种评估构建的模型网络与真实网络的拟合相似程度的指标E值。结果表明,采用共同邻居相似性这一指标,在3种尺度下构建的脑网络与真实网络的属性拟合程度不同,90个节点下的拟合程度最好,其他2种次之。

This paper adopts a modeling method combining brain network structural and functional characteristics and explores the modeling effect under three node scales.The structural characteristic uses anatomical distance and functional characteristic uses common neighbor similarity index.We analyze the modeling effect from the global properties and local properties of five kinds of brain network and put forward an index Eas the evaluation of the fitting similarily between model and real networks.Results show the fitting degree under the three node scales is different.The fitting degree for 90 nodes scale is best and the other two is slightly inferior.