Suprathreshold Stochastic Resonance In Neural System

Noise always is harmful in the transmission of signal, but noise also can improve the performance for the signal due to the nonlinear effect of stochastic resonance. The method for signal processing and the mechanism of neural signal processing that are based up on stochastic resonance are still in developing. The conventional studies of stochastic resonance usually require the input signal being below the threshold of the system, and noise is still considered to be harmful in the transmission of suprathreshold signal. However, our studies have shown that noise also can optimize the performance of nonlinear system in the transmission of suprathreshold signal. This study extends the conventional stochastic resonance to suprathreshold stimulus, we denote it in term of suprathreshold stochastic resonance. Suprathreshold stochastic resonance can be important and applicable in communication, neural signal processing and etc.In chapter one, we briefly introduce the mechanism, basic model,method of research, and it's important application in many fields of classical stochastic resonance.In chapter two, we detailedly introduce the suprathreshold stochastic resonance in a comparator. The conventional studies have shown that a single comparator can transmit subthreshold signals by the assistance of noise; but noise is harmful when it transmits suprathreshold signals. Our study has shown that noise also can improve the performance of comparator when it transmits strongly suprathreshold signals. This kind of stochastic resonance which happens in the case of the suprathreshod input signal extends the study of stochastic resonance.In chapter three, we consider a classical leaky integrate-and-fire (IF) neuron model, and reveal that suprathreshold stochastic resonance also is exist when it transmits suprathreshod signals. Up to now, stochastic resonance in neuron systems is available only when the input signal is subthreshold. But, when a neuron which described by IF neuron model transmits suprathreshold signals, our study has shown thatSNR of output can have a maximum value in the presence of an optimal noise, and we also get suprathreshold stochastic resonance. This result can be important to explain how a real neuron works.At last, we discuss the significance of suprathreshold stochastic resonance in neuron signal processing and applicability of it.