Title: Modeling and characterizing stochastic neurons based on in vitro voltage-dep endent spike probability functions Show Chinese title

Title: 基于体外电压依赖性尖峰概率函数的随机神经元建模与表征

Abstract: Neurons in the nervous system are submitted to distinct sources of noise, such as ionic-channel and synaptic noise, which introduces variability in their responses to repeated presentations of identical stimuli. This motivates the use of stochastic models to describe neuronal behavior. In this work, we characterize an intrinsically stochastic neuron model based on a voltage-dependent spike probability function. We determine the effect of the intrinsic noise in single neurons by measuring the spike time reliability and study the stochastic resonance phenomenon. The model was able to show increased reliability for non-zero intrinsic noise values, according to what is known from the literature, and the addition of intrinsic stochasticity in it enhanced the region in which stochastic resonance is present. We proceeded to the study at the network level where we investigated the behavior of a random network composed of stochastic neurons. In this case, the addition of an extra dimension, represented by the intrinsic noise, revealed dynamic states of the system that could not be found otherwise. Finally, we propose a method to estimate the spike probability curve from in vitro electrophysiological data. Show Chinese abstract

Abstract: 神经系统中的神经元会受到不同的噪声源影响，例如离子通道噪声和突触噪声，这会在重复呈现相同刺激时引入其反应的可变性。 这促使使用随机模型来描述神经元行为。 在本研究中，我们表征了一个基于电压依赖性尖峰概率函数的固有随机神经元模型。 我们通过测量尖峰时间的可靠性来确定内在噪声对单个神经元的影响，并研究了随机共振现象。 根据文献中的知识，该模型能够显示出非零内在噪声值时的可靠性增加，并且在其内部随机性增加后，增强了随机共振存在的区域。 我们继续在网络层面进行研究，调查由随机神经元组成的随机网络的行为。 在这种情况下，由内在噪声表示的额外维度揭示了系统动态状态，否则无法找到这些状态。 最后，我们提出了一种从体外电生理数据中估计尖峰概率曲线的方法。