Study On The Application Of Stochastic Resonance In Detecting Neural Weak Signal

It is usually believed that noise is detrimental to signal detection. It weakens the response of system to a signal, making the signal on the background of strong noise difficult to be detected. However, the theory of stochastic resonance (SR) points out that by the addition of a particular level of noise, the response of nonlinear system will be enhanced, so that strengthening the system's weak signal detection. Focusing on neural weak signal detection, this thesis presents the principle and practicable method to utilize SR for neural weak signal detection.First, based on the comprehensive review of the theoretical and experimental research on SR, the thesis, according to the requirement of weak signal detection, systematically studies the relationship between system, signal and noise when SR phenomenon appears.Second, according to the characteristic of the instantaneous change of action potential accompanied by the nerve impulse, we use H-H equation to describe such change, and conduct simulation combined with SR theory. The main process is as follows:1. Since the H-H equation consists of four nonlinear differential equations, and random Gaussian white noise should be added to the equation to study SR, we adopts a new algorithm to solve the random differential equations, which is easy to handle. By the simulation results the algorithm is proved to be correct and feasible.2. Based on the H-H model, the thesis studies the reciprocity of random noise and signal in neuron, and the rule of the phenomenon of SR resulting from their coordinating effect. It is concluded that by the addition of certain level of noise, the response of the nonlinear system to a weak signal is enlarged. That is to say, noise strengthens the response of nonlinear system to signal.Finally, after the study of simulation on the SR phenomenon in the neuron, we conduct a psychophysical experiment about the effect of SR on body's electric sense. Combining the basic laws of psychophysics and signal detection theory, we use two-alternative forced-choice(2AFC) tostudy the threshold change of human electric sense under different noise levels. The result indicates that, as the noise level ascends, the threshold of human electric sense descends; noise strengthens the weak-signal detectability in human body. However, the trend does not appear that the threshold ascends as the noise level continues ascending. Further research is needed.