Excitatory synaptic transmission in CNS

谷氨酸受体激活介导中枢神经系统神经元之间的传递，从而产生兴奋性突触电流的机制还不是很清楚。然而，量子大小在长时程增强过程中会发生变化，这一过程通常被认为是最好的记忆模型之一。现在已经描述了包括AMPA通道在内的各种通道的选通模型。根据在稳态条件下估计的平均打开和关闭时间，可以确定存在多少状态，相邻事件之间的关联揭示了这些状态是如何相互关联的。然而，要评估单位电流(MEPSC)的产生机制，需要的不仅仅是AMPA通道的门控模型。这是一个复杂的扩散问题，在有限的空间内谷氨酸分子与突触后AMPA受体的随机相互作用。因此，它不仅需要实验方法，也需要理论方法。我建议从实验和理论两方面考察是什么因素塑造了么正量子事件。实验部分将沿用传统的全细胞膜片钳方法对海马神经元进行全细胞膜片钳。理论研究将从蒙特卡罗模拟开始，它将提供对mEPSC产生过程的平均行为的评估，以及它的变异性。随后的分子动力学将通过考虑谷氨酸分子形状的复杂性、它的电荷以及它与水和膜的相互作用(静电和非静电)来进一步推动我们对这些过程的理解。兴奋性突触电流的波动通常是非平稳的。谷氨酸浓度变化很快，其门控动力学方案的速率依赖于浓度。此外，AMPA受体的脱敏速度很快，有助于形成单一兴奋性突触后电流的时间进程。然而，可以使用现代信号处理技术，如小波和卡尔曼-AR方法来跟踪非平稳电流波动，甚至可以从单个mEPSC跟踪。鉴于兴奋性突触电流的非平稳波动在动态上非常丰富，门控动力学、谷氨酸释放或扩散常数的任何差异都会导致mEPSC的总体特性(幅度、时程和电流波动)有所不同，从而使其评估不仅可行，而且准确。我建议开发直接适应系统(特定的动力学模型)及其输入(突触间隙中的谷氨酸浓度)与突触电流的方法：a)全连接的前馈神经网络，b)模糊逻辑，c)支持向量机。这些方法将首先在蒙特卡罗模拟的mEPSCs上进行测试，然后在实验记录的大鼠海马区mEPSCs上使用。

The mechanisms involved in generation of excitatory synaptic currents, induced by glutamate receptor activation mediating transmission between neurons in the central nervous system, are not well understood. However, quantal size changes during long-term potentiation, which is a process usually considered as one of the best models for memory. The models of gating of a variety of channels including AMPA channels have now been described. How many states exist can be determined from the mean open and closed times estimated under steady-state conditions, and the correlations between adjacent events reveal how the states are inter-connected. The evaluation of mechanisms of generation of unitary currents (mEPSCs) however, requires more than just the model of gating of AMPA channels. It is a complex problem of diffusion in a restricted space with stochastic interactions of the glutamate molecules with postsynaptic AMPA receptors. As such it requires not only experimental but also theoretical approaches. I propose to examine both experimentally and theoretically what factors shape the unitary quantal events. Experimental part will follow the 'traditional' methods of whole-cell patch clamping of hippocampal neurons in hippocampus. Theoretical studies will start with Monte Carlo simulations, which will provide the assessment of the average behavior of the process of generation of mEPSCs, but also its variability. Molecular dynamics, which will follow will further advance our understanding of these processes by considering the complexities of shape of glutamate molecule, its charge, and its interactions (electrostatic and non-electrostatic) with water and membranes.The fluctuations of excitatory synaptic currents are generally non-stationary. Glutamate concentration changes rapidly, and the rates of their kinetic scheme of gating are concentration dependent. Moreover, the desensitization of AMPA receptors is fast and contributes to shaping the time course of unitary excitatory post-synaptic currents. Non-stationary current fluctuations can however be tracked using modern techniques of signal processing such as wavelet and Kalman-AR methods, even from individual mEPSCs. Given that the non-stationary fluctuations of excitatory synaptic currents are dynamically very rich, any difference of gating kinetics, glutamate release or diffusion constant will lead to mEPSCs whose overall properties (amplitude, time course and the current fluctuations) will differ in some respect from each other, rendering their evaluation not only feasible, but also accurate. I propose to develop methods that fit directly the system (particular kinetic model) and its input (glutamate concentration in the synaptic cleft), to synaptic currents: a) an all-connected, feed-forward neural network, b) fuzzy logic, and c) support vector machines. The methods will be tested first on Monte-Carlo simulated mEPSCs and use subsequently on experimentally recorded mEPSCs from rat hippocampus.