神经元与神经元回路的模型分析

发布时间：2018-01-01 05:39

本文关键词：神经元与神经元回路的模型分析　出处：《华南理工大学》2011年硕士论文　论文类型：学位论文

【摘要】：神经回路是大脑中普遍存在的结构,也是脑信息处理的主要功能单元,而神经元则是组成神经回路的基本单元,是神经信息处理的最小载体.在本文中,我们采用由小到大、层层递进的方式,系统地分析了神经元单房室模型、两房室模型、多房室模型、电缆模型、神经元模型中ISI倍周期分岔的数学机理及闸门控制环路中神经纤维的活动特性和信息编码机制,并在此基础上,通过建立与实际结构相吻合的嗅球神经回路、小脑皮层神经回路、海马神经回路、视网膜神经回路及基底神经节神经回路,来模拟了这些回路对神经信号的处理和加工,考察了神经信息经这些回路整合后的编码特性,最后对大脑内其他几种重要的神经回路也做了简要的介绍,具体结果有: 1、单房室神经元模型:主要是通过计算机模拟的手段,考察了CA1锥体神经元所具有的丰富的发放模式,分析了随着某些电生理参数的改变,这些发放模式之间的转移,发现了几种典型的非线性动力学分岔现象,如:倍周期分岔、加周期分岔等,进一步理解了CA1锥体神经元的信息编码特性. 2、两房室神经元模型:首先,对皮层锥体神经元两房室模型中所固有的几种离子通道进行了考察,分析了离子通道电导系数和离子平衡电位的扰动所引起的皮层锥体神经元多样性的发放模式及周期性的分岔现象.其次,鉴于两房室模型中胞体和树突在形态和功能上的区别,我们分析了两房室中膜电容和泄漏电导单独改变和同时改变对皮层锥体神经元活动特性的影响,进一步理解了神经元胞体和树突在信息编码上的差异.再次,考察了内部时滞和外部温度所带来的皮层锥体神经元电位发放特性上的变化.最后,研究了另一种可产生Ghostbursting的两房室神经元模型,并总结了一般性的两房室神经元模型中所具有的丰富的动力学分岔现象. 3、多房室神经元模型:主要是通过NEURON软件构造了三种神经元的多房室模型,其中有:嗅球僧帽细胞、小脑皮层浦肯野细胞和海马锥体细胞,利用发放频率和峰峰间距(ISI)图,比较了三种神经元电位发放特性的不同以及它们在各自神经回路中信息处理作用上的差异. 4、电缆模型:以神经元的整段树突为基础,根据Hodgkin-Huxley离子通道模型和Rall一维电缆模型,建立了三维空间上的神经元树突电缆模型,对应得到整段树突的空间信号传递理论.数值结果首次揭示了树突内神经信号的空间传递过程和规律. 5、ISI倍周期分岔的数学表达:针对神经元HH化模型的构造特点以及神经元动作电位的轨迹,通过积分和微分两种手段,对峰峰间距(ISI)的数学表达式进行了推导,并将其运用到实际的神经元模型中,发现了大量的倍周期分岔现象,通过与传统LOGISTIC映射引入的倍周期分岔对比,阐述了我们提出的ISI倍周期表达式的新颖之处. 6、闸门控制回路:在本节,我们通过考察具体的神经元回路--闸门控制回路,分析了回路中三种不同传递疼痛的神经纤维在传递速度、承载疼痛类型上的差异,并研究了不同疼痛刺激类型所造成的回路传出神经元的反应. 7、嗅球神经回路:首先,通过参考嗅球的解剖结构,建立了与实际结构相吻合的嗅球网络模型.其次,对嗅球中存在的两种实验现象进行了验证,包括:动作电位在单个僧帽细胞上的产生和传递,嗅球网络中动作电位产生和传递的四种模式,说明了所建网络模型的合理性.最后,对模型的一些数值结果进行了分析,包括:嗅球中不同神经元的动作电位发放序列、不同平面刺激类型所造成的嗅球输出结果的不同以及颗粒细胞对僧帽细胞的侧抑制所造成的僧帽细胞侧树突上动作电位传递的衰减和中断. 8、小脑皮层神经回路:首先,通过参考小脑皮层的解剖结构,建立了与实际结构相吻合的小脑皮层网络模型.其次,考察了来自两种不同传入系统的刺激所造成的浦肯野细胞发放频率上的差异,并通过给定这两种传入系统一定的刺激,得到了小脑皮层回路中5种神经元的电位发放图.再次,利用所建模型考察了小脑皮层回路中存在的两种重要的现象:时间聚焦现象和空间聚焦现象.最后,针对实际中存在的兔子眼闪烁的条件反射现象,我们也通过所建网络模型再现了这一过程. 9、海马神经回路:首先,通过参考海马的解剖结构,建立了与实际结构相吻合的海马结构网络模型.其次,通过给定一外界刺激,得到了海马结构中几种神经元的电位发放图.最后,重点分析了连接左右海马片区的胼胝体对海马回路中信息处理和加工的影响,得出CA1片区受同侧CA3输入的影响要大过对侧CA3对它的输入. 10、视网膜神经回路:首先,通过参考视网膜的解剖结构,构建了与实际结构相吻合的视网膜网络模型.其次,考察了光刺激强度的改变所引起的光感受器反应的不同.再次,借助于平面刺激,研究了光刺激范围的不同和光刺激类型的不同所造成的视网膜神经回路输出结果的差异.最后,讨论了视网膜神经回路中网间细胞的反馈抑制对视网膜神经节细胞发放模式的影响. 11、基底神经节神经回路:从基底神经节的功能展开,构造了基地神经节网络模型,通过采用5神经元制赋予了网络五个不同的信号通道,同时模拟5种不同情况的刺激下,正常人、帕金森(PD)病人和DBS治疗病人的活动特性(膜电位发放图).结合帕金森病病例,运用WLC模型模拟了PD的发病机制以及相应的治疗方法,通过建立相图指标对正常人、PD病人和DBS治疗病人三种情况进行对比,验证了治疗方法的合理性. 12、大脑内其他几种神经回路的介绍:在本节,我们简要的介绍了大脑内存在的其他几种被广泛研究和考察的神经回路模型:决策选择网络、丘脑皮层网络和运动皮层对随意运动的控制网络.并展望了大规模的脑神经网络的模型及其计算. 通过模型模拟大脑内各种神经回路中神经元动作电位的发放序列是理论研究大脑神经系统的第一步,如何利用所建网络来分析这些数值结果并与实际实验现象进行对比,进而提出网络中所隐藏的新的特性就显得十分重要了. 本文数值计算采用NEURON和MATLAB软件,并利用ORIGIN对部分数据进行了处理,结果真实可靠.
[Abstract]:Neuronal circuits are commonly exist in the brain, the main function is the brain information processing unit, and the neurons is composed of the basic unit of the neural circuit is the smallest vector of neural information processing. In this paper, we use small, step-by-step way, systematically analyzes the single compartment neuron model, two compartment model, multi compartment model, cable model, nerve fiber mathematical mechanism and gate control loop of ISI bifurcation in the neuron model activity characteristics and information encoding mechanism, and on this basis, through the establishment of olfactory bulb neural circuits are in good agreement with the actual structure of the cerebellar cortex neural circuits, hippocampal neural circuits, retinal neural circuits and the basal ganglia neural circuits, to simulate the processing and processing of neural signals of these loops were investigated by neural information encoding characteristics of the circuit integration, finally Several other important neural circuits in the brain were briefly introduced, and the results were as follows:
1, the single compartment neuron model: mainly by means of the computer simulation, the effects of CA1 pyramidal neurons has rich firing pattern, analyzed with some electrophysiological parameters, transfer between these payment modes, found several typical nonlinear dynamic bifurcation phenomena, such as period doubling bifurcation, period adding bifurcation so, further
The solution of the information encoding characteristics of CA1 pyramidal neurons. 2, two compartment neuron model: firstly, several ion channels inherent in cortical pyramidal neurons in the two compartment model were investigated. The analysis of distribution pattern and cycle diversity cortical pyramidal neurons of the bifurcation phenomena caused by the disturbance of ion channel conductance and ion balance potential. Secondly, in view of two compartment model in the soma and dendrites in the form and function of the difference, we analyzed two chamber membrane capacitance and conductance change alone and at the same time the change of cortical pyramidal neuron activity characteristics, further understanding of the differences between the soma and dendrites of neurons in the information encoding on again. On the issue, change characteristics of cortical pyramidal neurons caused by the effects of internal potential delay and external temperature. Finally, research on another can produce Ghostbursting two The model of the atrioventricular neuron and the rich dynamic bifurcation phenomena in the general model of the two chamber neurons are summarized.
3, the multi compartment neuron model is mainly a multi compartmental model, three kinds of neurons is constructed by NEURON software including: olfactory mitral cells, cerebellar cortical Purkinje cells and pyramidal cells, the distribution frequency and interspike interval (ISI), compared with three kinds of neuronal potentials of the different characteristics and their differences in their respective neural circuits in the information processing function.
4, cable model: a whole based neuron dendrites, according to the Hodgkin-Huxley channel model and Rall one-dimensional cable model, established the dendrite cable model in three-dimensional space, the corresponding space signal throughout the dendritic transfer theory. Numerical results first revealed neural signals within dendrites space transfer process and the law.
5, the expression of ISI bifurcation of Mathematics: according to the structural characteristics of HH neurons and the trajectory model of neuronal action potentials, through integral and differential two means of interspike interval (ISI) mathematical expressions were derived and applied to the actual neuron model, found the phenomenon of double period a large number of bifurcation bifurcation, by comparing with the traditional LOGISTIC mapping is introduced, elaborated the novelty of ISI times the cycle we present expressions.
6, gate control circuit: in this section, we analyze the specific neural circuits, gate control circuit, analyzes the circuit in three different transmitting pain nerve fibers in the transmission speed, bearing differences in pain types, and the effects of different types of pain caused by the loop efferent neuron response.
7, the olfactory bulb neural circuits: first, through the reference anatomical structure of the olfactory bulb, established olfactory network model are in good agreement with the actual structure. Secondly, two kinds of experimental phenomena exist in the olfactory bulb were verified, including: the generation of action potentials in single mitral cells and the transfer of four modes action potentials of olfactory bulb network and transfer, which proves the rationality of the network model. Finally, some numerical results of the model are analyzed, including: the olfactory bulb in different neuronal action potential firing sequence, the mitral cell lateral dendrites caused by olfactory bulb output results of different types of stimulation caused by plane the different and granular cells of the mitral cell lateral inhibition on action potential propagation and attenuation of the interrupt.
8, the cerebellar cortex neural circuits: first, through the reference anatomical structure of cerebellar cortex cerebellar cortex, established the network model are in good agreement with the actual structure. Secondly, the difference caused by the frequency distribution were investigated from two different afferent system stimulation of Purkinje cells, and by stimulating certain given the two afferent system the potential of the 5 kinds of neurons, the cerebellar cortex in the loop distribution diagram. Thirdly, the model analyzes two important phenomena existing in the cerebellar cortex in the loop: time and space focusing focusing phenomenon. Finally, according to the condition of reflection phenomenon in reality of the rabbit eyes, we also through the network the model reproduces this process.
9, the hippocampal neural circuits: first, through the reference anatomical structure of the hippocampus, established the hippocampal structure network model is consistent with the actual structure. Secondly, through a given stimulus, the potential of several neurons in hippocampal structure in the distribution graph. Finally, analyzed the influence of the corpus callosum is connected to the information about the hippocampus area treatment and processing of hippocampal circuits, the CA1 area is affected by the same side CA3 input is greater than the contralateral CA3 on its input.
10 retinal neural circuits: first, through the anatomical reference of the retina, constructs the retina network model are in good agreement with the actual structure. Secondly, effects of photoreceptor responses to stimulus light changes caused by the different. Again, with the help of plane stimulation, the research results of differences in retinal nerve loop output light stimulation range different types and light stimulation caused by the different. Finally, discussed the feedback inhibitory effects on retinal ganglion cell firing pattern of retinal nerve cells in the loop between.
11, the basal ganglia circuits: start from the function of basal ganglia, construct basal ganglia network model, by using the 5 neuron system gives five different signal channel, and simulated 5 different conditions under the stimulation of normal people, Parkinson (PD) and DBS disease activity characteristics of patients (film potential distribution map). Combined with Parkinson's disease, the WLC model was used to simulate the pathogenesis of PD and the corresponding treatments, through the establishment of a phase diagram index of normal people, compared with PD patients and DBS patients three cases, verified the rationality of the treatment method.
12, several other neural circuits in the brain: in this section, we briefly introduce several other existing neural circuits in the brain has been widely studied and investigated the model: decision network, control network and motor cortex of thalamocortical network free movement. And look a large-scale neural network model and its calculation the.
Through the issuance of series model action potentials of neurons in the brain of various neural circuits is the first step in theoretical research of brain neural system, how to use the network to analyze the numerical results were compared with the actual experimental phenomena and then puts forward the hidden network of new features is very important.
In this paper, NEURON and MATLAB software are used in the numerical calculation, and some data are processed with ORIGIN. The results are true and reliable.

【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：R311

【参考文献】