计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (14): 45-51.DOI: 10.3778/j.issn.1002-8331.1704-0203

• 理论与研发 • 上一篇    下一篇

基于时间自动机的脑网络时空演化建模方法

裴常福1,王  彬1,薛  洁2,刘  辉1,熊  新1   

  1. 1.昆明理工大学 信息工程与自动化学院,昆明 650500
    2.云南警官学院 信息网络安全学院,昆明 650500
  • 出版日期:2018-07-15 发布日期:2018-08-06

Spatio-temporal evolution modeling method of brain network based on time automata

PEI Changfu1, WANG Bin1, XUE Jie2, LIU Hui1, XIONG Xin1   

  1. 1.Faculty of Information Engineering & Automation, Kunming University of Science & Technology, Kunming 650500, China
    2.Faculty of Information Network Security, Yunnan Police College, Kunming 650500, China
  • Online:2018-07-15 Published:2018-08-06

摘要: 人类大脑本质上是不断变化的整体,但基于静息态功能磁共振成像重构技术的人脑网络动态特性研究尚在起步阶段,并且大多采用定性的方法描述。采用时间自动机理论对脑网络在时间和空间上的系统动态特征和演化过程展开了建模方法研究,首先通过对采样时间区间上血氧依赖水平信号的处理得到全脑脑区在单个采样点上的状态描述,然后通过无监督聚类获取其状态集,研究其状态随时间转换的可观测性,最后在此基础上结合时间自动机理论对脑网络状态的演化过程进行建模,从而达到对脑网络动态特性定量描述的目的。实验结果显示该方法可定量描述人脑网络的状态转换规律和演变进程,对不同被试数据具有普适性,并且可辨识出被试的异常演化过程,为脑网络动态特性的深入研究提供了理论基础。

关键词: 脑功能网络, 动态特性, 状态演化, 时间自动机, 血氧依赖水平

Abstract: As a whole system the human brain constantly maintains changing in essence, but the researches about dynamic characteristics of human brain network which is based on functional Magnetic Resonance Imaging(fMRI) reconstruction technology are still in the early stage, and those methods used to describe the brain system discipline are always be qualitative but not quantitative. A modeling method using time automaton theory is discussed for the description of human brain network dynamical characteristics and evolution process in this paper. Firstly the description of all the brain areas at one single time point is presented by analyzing the Blood Oxygen Level Dependent(BOLD) signal, then the unsupervised clustering technology is used to get all the status and the status transformation regular pattern with time changing is observed. Finally a time automation evolution model is given to describe the evolution process of brain network state quantitatively on the basis of all these works. The experimental results show that this method can give a precise description of the state transition and evolution rules of the human brain network. This method is universal to different subjects and able to identify the abnormal evolution process of subjects, and it also provides a theoretical basis for the further study of dynamic characteristics of brain network.

Key words: brain function network, dynamic characteristics, state evolution, timed automata, Blood Oxygen Level Dependent(BOLD)