The Suprathreshold Stochastic Resonance In A Integrate And Fire Neuron

In this paper, we use integrate and fire model as the researching object to studying why and when suprathreshold stochastic resonance appears. We found that when the signal frequency are near to the inherence frequency which is decided by the model's parameter such as current I D and threshold Vt h. When the two frequency is near, the appearance of noise can make them form a 1:1 phase-locking sometimes. When the 2 frequency is in a 1:1 phase-locking, the neuron would fire at the frequency of the signal frequency. This phenomenon also can be shown in the firing phase distribution of the spike train of the I-F neuron. This theory of phase-locking can tell why stochastic resonance only appears when the 2 frequency are near.In this paper, we also introduced the stochastic resonance of weak signal in a I-F neuron with low resolving rate. In such a neuron, when the input signal's amplitude is too small, the neuron can not code it with the interspike interval. The output of such a neuron is completely a spike train with same interval. Obviously, the frequency information of the weak signal can not be demonstrate by the interval. But when nemoral noise is added to the signal, in one hand the completely same interval is broken, in the other hand, the frequency information of weak signal appears. This is to say that the noise also improved the neuron model's resolving power of frequency. This may explain the reason why noise can improve humans hearing to the sound signal upper than hearing threshold.