吉林大学学报(理学版)

• 计算机科学 • 上一篇    下一篇

基于关键参数反馈的神经元放电模式控制

王海洋1,2, 王江2   

  1. 1. 白城师范学院 机械工程学院, 吉林 白城 137000; 2. 天津大学 电气与自动化工程学院, 天津 300072
  • 收稿日期:2015-03-31 出版日期:2016-03-26 发布日期:2016-03-23
  • 通讯作者: 王江 E-mail:jiangwang@tju.edu.cn

Neurons Firing Mode Control Based on Key Parameters Feedback

WANG Haiyang1,2, WANG Jiang2   

  1. 1. School of Mechanical Engineering, Baicheng Normal University, Baicheng 137000, Jilin Province, China;2. School of Electrical Engineering and Automation, Tianjin University, Tianjin 300072, China
  • Received:2015-03-31 Online:2016-03-26 Published:2016-03-23
  • Contact: WANG Jiang E-mail:jiangwang@tju.edu.cn

摘要:

针对神经元数学模型中相关参数的异常变化问题, 先通过无迹Kalman滤波器(unscented Kalman filter)对这些关键参数进行实时估计, 再使用这些参数作为控制器的反馈信号完成神经元放电的有效闭环控制. 使用该方法对PinskyRinzel(PR)神经元由于关键参数(胞体和树突间的电导gc及树突刺激电流Ic)的病变导致的异常放电进行MATLAB仿真控制, 分别实现了单参数病变及两个参数同时病变引起的异常放电控制. 实验结果证明了该方法的有效性.

关键词: 闭环控制, 参数估计, 慢变量, 关键参数

Abstract:

Aiming at the problem of the abnormal change of the related parameters in the mathematical model of neuron, we first made a real time estimation of these key parameters by unscented Kalman filter (UKF), and then took these parameters as feedback signal of the controller to complete closed loop control of the neuron discharge. With this method, the MATLAB simulation was carried out for the abnormal discharge caused by the key parameters (the conductance
between cell bodies and dendrites gc, and the stimulated currents of dendrites Ic) of PinskyRinzel (PR)neuron, and the control of abnormal discharge caused by single parameter and two parameters were achieved respectively. The experimental results show the proposed method is effective.

Key words:  closed loop control, estimate of parameter, slow variable, key parameters

中图分类号: 

  • TP391