基于Kendall改进的同步算法癫痫脑网络分析

An improved synchronous algorithm based on Kendall for analyzing epileptic brain network

提出了一种基于Kendall等级相关改进的同步算法IRC(inverse rank correlation).Kendall等级相关是非线性动力学分析的一般化算法,可有效地度量变量间的非线性相关性.复杂网络的研究已逐渐深入到社会科学的各个领域,脑网络的研究已经成为当今脑功能研究的热点.利用改进的IRC算法,基于脑电EEG(electroencephalogram)数据来构建大脑功能性网络.对构建的脑功能网络的度指标进行了分析,以调查癫痫脑功能网络是否异于正常人.结果显示:使用该改进的算法能够对癫痫和正常脑功能网络显著区分,且只需要记录很短的脑电数据.实验结果数据表明,该方法适用于区分癫痫和正常脑组织网络度指标,它可有助于进一步地加深对大脑的神经动力学行为的研究,并为临床诊断提供有效工具.

In this study, we propose a kendall rank correlation based synchronous algorithm inverse rank correlation （IRC）. The kendall rank correlation is a generalized algorithm of nonlinear dynamics analysis which can effectively measure nonlinear correlations between variables. The study of complex networks has gradually penetrated into various fields of the social sciences. We use our algorithm to construct functional brain networks based on the data from electroencephalogram （EEG）. The average node degree of complex brain networks is analyzed to investigate whether epileptic functional brain networks are distinctly different from normal brain networks. Results show that our method can distinguish between epileptic and normal functional brain networks and needs to record a very small number of EEG data. Experimental data show that our method suited to distinguish between epilepsy and normal brain node degree, which may contribute to further deepening the study of the brain neural dynamic behaviors, and provide an effective tool for clinical diagnosis.