用信号检测理论和泊松过程仿真感觉系统中随机共振现象

Simulation of stochastic resonance of sensory system by signal detection theory and Poisson process.

运用心理物理学的实验方法来研究噪声对某一感觉通路的作用是神经信息学中的重要手段,一些研究者们已经借用此类实验方法证实了感觉通路中存在有噪声作用的随机共振现象,但是相应的理论研究工作仍不完善.基于信号检测理论中的理想观测器模型,考察了在保守阈值下的阈下随机共振现象和宽松阈值下的阈上随机共振现象.另外,文中从更接近实际神经电活动的角度出发,使用神经元放电的泊松模型来详细分析噪声对感觉神经系统的作用。

Some experimental methods of psychophysics were used to study the effect of noise on sensory systems detecting signal, but the corresponding theoretical explains were lacking. In this paper, based on the paradigm of signal detection theory in psychophysics, their capacity of signal detection were analyzed when the sensory systems were affected by noise. Firstly, the ideal observer was used to model the sensory neurons. It was found that the stochastic resonance phenomenon would be produced under condition of the conservative criterion or the liberal criterion. Suprathreshold stochastic resonance has been called on the liberal criterion. This paper also exploited the neuronal Poisson model to analyze the effect of noise and obtained the conclusion that stochastic resonance phenomenon was proved.