基于逆问题的膜片钳新技术研究

发布时间：2018-02-04 04:18

本文关键词： 膜片钳技术反卷积系统辨识白噪声互相关技术子空间法自动补偿　出处：《华中科技大学》2012年博士论文　论文类型：学位论文

【摘要】：膜片钳技术是细胞离子通道记录方法的金标准。20年来，，神经科学的发展体现出与信息技术（计算机）越发紧密结合的趋势，而膜片钳技术中计算机的参与度相对较低。膜片钳技术中的信号处理主要使用模拟电路中的概念，如滤波、检波、补偿等。为了适应神经科学发展的总体趋势，本文致力于以膜片钳系统的数学模型为基础，为膜片钳的信号处理引入正、逆问题的概念，以便经典的膜片钳技术从现代信息技术的丰硕成果中获益。本文的工作主要有两点：1）讨论膜片钳放大器、电极、细胞构成的统一电系统的数学模型，并总结膜片钳电路设计及信号处理中的正问题和逆问题；2）对于膜片钳系统中的逆问题用反卷积进行描述，并提出采用白噪声驱动的时间序列分析及基于子空间的系统辨识法对反卷积问题进行求解。用正、逆问题的概念对膜片钳技术中信号处理问题进行分类，可使信号之间的关系更为清晰，使问题的描述更系统，使问题的求解有坚实的理论基础和可借鉴的方法。根据正、逆问题的基本概念，细胞通道电流记录本质上是个逆问题（信号还原）。若细胞以单室模型建模，通道电流与膜电容被动响应电流相并联。于是，通道电流的求解可看作加法问题的逆问题，即减法问题。待减去的膜电容电流，其大小与膜电容参数有关，从测量数据获得膜电容参数信息是个逆问题（系统辨识）。线性时不变系统中的逆问题就是反卷积。反卷积的求解以数学模型为基础。本文讨论了膜片钳放大器和细胞构成的电系统的若干种数学模型，如微分方程模型、传递函数模型、卷积模型和状态空间方程模型，以及它们之间的关系。文章从信号与系统的角度，讨论了膜片钳实验中的几个典型的反卷积问题，包括信号复原和系统辨识。受基于ARMA模型的时间序列分析的启发，可用白噪声激励膜片钳系统，通过互相关技术求解系统的时域特性，即冲激响应（本文中称为卷积核），并开发了基于卷积核的非迭代快电容自动补偿算法——K-method；为了确定膜片钳探头反馈电阻的杂散电容，可采用子空间法对其进行系统辨识，并以此为基础开发了高值串联电阻估计方法和软件高频补偿方法——SHB。本文详细介绍了这些具体应用的原理和实现步骤。相比原迭代方法，K-method具有简单、准确、抗饱和的优点；SHB有利于减小膜片钳体积、提高膜片钳的集成度。本文工作不但为膜片钳技术结合现代信号处理技术打下了理论基础，还提供了两个具体范例展示了膜片钳技术的发展方向。
[Abstract]:Patch clamp technology is the gold standard of cell ion channel recording. In the past 20 years, the development of neuroscience has shown a trend of closer integration with information technology (computer). The signal processing in patch clamp technology mainly uses the concept of analog circuit, such as filtering, detection, compensation, etc. In order to adapt to the general trend of neuroscience development. Based on the mathematical model of patch clamp system, this paper introduces the concepts of forward and inverse problems for the signal processing of patch clamp, so that the classical patch clamp technology can benefit from the fruitful results of modern information technology. In this paper, there are two points: 1) the mathematical model of the unified electrical system composed of patch clamp amplifier, electrode and cell is discussed, and the forward and inverse problems in the design of patch clamp circuit and signal processing are summarized. 2) the inverse problem in patch clamp system is described by deconvolution, and the white noise-driven time series analysis and system identification method based on subspace are proposed to solve the deconvolution problem. Using the concepts of positive and inverse problems to classify the signal processing problems in patch clamp technology can make the relationship between signals clearer and the description of problems more systematic. So that the solution of the problem has a solid theoretical basis and can be used for reference. According to the basic concepts of forward and inverse problems, the recording of cell channel current is essentially an inverse problem (signal reduction). If the cell is modeled by a single cell model, the channel current is parallel with the passive response current of membrane capacitance. The solution of channel current can be regarded as the inverse problem of the addition problem, that is, the subtraction problem. The magnitude of the membrane capacitance current to be subtracted is related to the parameters of the membrane capacitance. It is an inverse problem to obtain membrane capacitance parameter information from measurement data (system identification). The inverse problem in linear time-invariant system is deconvolution. The solution of deconvolution is based on mathematical model. In this paper, some mathematical models of electric system composed of patch clamp amplifier and cell are discussed. For example, differential equation model, transfer function model, convolution model and state space equation model, and their relations. Several typical deconvolution problems in patch clamp experiments, including signal recovery and system identification, are discussed. Inspired by the time series analysis based on ARMA model, white noise excited patch clamp system can be used to solve the time-domain characteristics of the system by cross-correlation technique, that is, impulse response (in this paper called convolution kernel). A non-iterative fast capacitor compensation algorithm based on convolution kernels is developed. In order to determine the stray capacitance of the feedback resistance of the patch clamp probe, the subspace method can be used to identify the system. On this basis, the high value series resistance estimation method and the software high frequency compensation method, SHB, are developed. The principle and implementation steps of these specific applications are introduced in detail, compared with the original iterative method. K-method has the advantages of simplicity, accuracy and anti-saturation. SHB can reduce the size of patch clamp and improve the integration of patch clamp. This work not only lays a theoretical foundation for patch clamp technology combined with modern signal processing technology, but also provides two concrete examples to show the development direction of patch clamp technology.
【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：R329

【共引文献】