ISSN 0371-0874, CN 31-1352/Q

过刊浏览

基于神经网络的振荡模式分析及其应用

李群, 程宁, 张涛

南开大学生命科学学院,天津 300071;南开大学数学学院,天津 300071;喀什师范学学院数学系,喀什 844008

摘要

振荡现象以动态形式普遍存在于神经系统中,并且与大脑的信息处理、传递和整合、巩固记忆等高级认知活动密切相关。神经振荡的特定活动模式往往关联认知功能及其变化,因此如何量化分析神经振荡活动模式成为了计算神经生物学的研究热点之一。结合作者实验室近年来的研究工作,本文对在神经生物学和认知科学研究中常用的多种分析算法进行了详细而全面的综述,并试图按照度量指标及耦合或同步方式的差异进行归类。通过算法比较,给出计算特点及算法适用情形。最后对将来有潜在应用价值的几种多维算法进行了深入的探讨。

关键词: 神经网络; 振荡分析算法; 信息流; 振荡同步

分类号:Q42; R338

[Analysis of oscillatory pattern based on neural network and its applications.] [Article in Chinese]

Li Qun, CHENG Ning, ZHANG Tao

College of Life Sciences, Nankai University, Tianjin 300071, China; College of Mathematics, Nankai University, Tianjin 300071, China; Department of Mathematics, Kashgar Normal College, Kashgar 844008, China

Abstract

Neural oscillatory phenomenon generally exists in the nervous system through a dynamic form. It plays a very important role in the brain, especially in the higher cognitive activities, such as information processing, transfer and integration, consolidating memory and so on. Furthermore, the specific activity pattern of neural oscillations is often associated with cognitive functions and their alterations. Accordingly, how to quantitatively analyze the pattern of neural oscillations becomes one of the fundamental issues in the computational neuroscience. In this review, we addressed a variety of analytic algorithms, which are commonly employed in our recent studies to investigate the issues of neurobiology and cognitive science. In addition, we tried to classify these analytic algorithms by distinguishing their different metrics, synchronization and coupling modes. Finally, multidimensional analytic algorithms for potential application have also been discussed.

Key words: neural network; analytic algorithms; neural information flow; oscillatory synchronization

收稿日期:2014-11-25  录用日期:2014-12-29

通讯作者:张涛  E-mail: zhangtao@nankai.edu.cn

引用本文:

李群, 程宁, 张涛. 基于神经网络的振荡模式分析及其应用[J]. 生理学报 2015; 67 (2): 143-154.

Li Qun, CHENG Ning, ZHANG Tao. [Analysis of oscillatory pattern based on neural network and its applications.] [Article in Chinese]. Acta Physiol Sin 2015; 67 (2): 143-154 (in Chinese with English abstract).