基于rs-fMRI数据的脑功能网络构建与分析

Construction and Analysis of Brain Functionality Network Based on rs-fMRI Data

从脑网络的角度研究大脑功能脑区之间的连接关系,对于理解大脑的工作方式乃至探究精神疾病的病理机制具有重要意义。本文基于静息态功能磁共振成像(rs-fMRI)数据,计算264个脑区间的相关性,提出了3个合理的假设来确定相关系数阈值,构建出相应的脑功能网络。通过计算网络的聚类系数和平均最短路径长度等属性,结果表明脑功能网络具有小世界特性。针对脑区节点数大于信号时间序列长度情况下的偏相关计算,提出了一种矩阵变换法,获得脑区间的偏相关系数,能够消除其他节点的间接影响。最后在标准脑图上实现了脑功能网络连接关系的可视化。实验证明本文的构建和分析算法是可行的,为脑功能网络分析提供了有益的探索。

It has important significance for understanding the brains work and exploring the pathological mechanism of mental disease that research the functional connectivity of the human brain regions from the viewport of brain network.By using the data of resting state functional Magnetic Resonance Imaging（rs-fMRI）,this paper calculates the correlation among 264 brain regions.And then,by determining the available threshold of correlation coefficient via three reasonable assumptions,this paper constructs the brain functionality network.The experiment results via computing clustering coefficient and average minimum path length show that the brain functionality network has the feature of small world.Considering the number of brain nodes greater than the length of signal sequence,this paper proposes a matrix transformation algorithm to obtain the partial correlation algorithm and eliminate the indirect effects of other nodes.Finally,the visualization of brain nodes connectivity is constructed based on the standard brain images.The experiments illustrate that the proposed algorithm is feasible and beneficial for the exploration in the field of brain function connectivity.