Bauermann Jonathan, Lindner Benjamin
Bernstein Center for Computational Neuroscience Berlin, Philippstr. 13, Haus 2, 10115 Berlin, Germany; Department of Physics, Humboldt Universität zu Berlin, Newtonstrasse 15, 12489 Berlin, Germany.
Bernstein Center for Computational Neuroscience Berlin, Philippstr. 13, Haus 2, 10115 Berlin, Germany.
Biosystems. 2019 Apr;178:25-31. doi: 10.1016/j.biosystems.2019.02.002. Epub 2019 Feb 5.
We study simple integrate-and-fire type models with multiplicative noise and consider the transmission of a weak and slow signal, i.e. a signal that evokes a small modulation of the instantaneous firing rate on time scales that are much larger than the membrane time scale and the mean interspike interval. The specific question of interest is whether and how the state-dependence of the noise can be optimized with respect to information transmission. First, in a simple model in which the noise intensity varies linearly with the state variable, we show analytically that multiplicative fluctuations may benefit the signal transfer and we elucidate the mechanism for this improvement. In a conductance-based integrate-and-fire model with synaptically filtered shot-noise input, we show by means of extended numerical simulations that also in a biophysically more relevant situation, multiplicative noise can enhance the signal-to-noise ratio. Our results shed light on a so far unexplored aspect of stochastic signal transmission in neural systems.
我们研究了具有乘性噪声的简单积分发放型模型，并考虑了微弱缓慢信号的传输，即一种在比膜时间尺度和平均峰峰间隔大得多的时间尺度上引起瞬时发放率小幅调制的信号。我们感兴趣的具体问题是，噪声的状态依赖性能否以及如何针对信息传输进行优化。首先，在一个噪声强度随状态变量线性变化的简单模型中，我们通过解析表明乘性涨落可能有益于信号传递，并阐明了这种改善的机制。在一个具有突触滤波散粒噪声输入的基于电导的积分发放模型中，我们通过扩展的数值模拟表明，在生物物理上更相关的情况下，乘性噪声也可以提高信噪比。我们的结果揭示了神经系统中随机信号传输迄今为止尚未探索的一个方面。
