Pitch And Noise Nonlinear Process In Auditory Systems

Pitch perception and noise procession are important features in a auditory system,they gave the auditory system the ability to feel the acoustics characteristics and cope with noise.The study of their mechanics have great value in hearing rehabilitation and acoustics information processing based on auditory.This research starts from the nonlinearity of auditory system and makes corresponding progress.In the pitch perception field,by theoretical research,I found that the three frequency resonance only occurs while the drive signal is a series of periodical functions.In this condition,the system has three frequency resonance and the resonance frequency that is coincident to the residue pitch which is obtained in psychological experiments.If the system is driven by combined sinusoidal signals,the three frequency resonance phenomenon does not happen.With a series of laser interference measurement on isolated and in vivo guinea pig cochlea's basilar membrane,I found that in a cochlea driven by combined sinusoidal signals,there does not exists a frequency response that three frequency resonance theory respects.On the contrary,the time pattern of the basilar membrane oscillation waveform expressed the pitch information that fits the psychological experiments' result.This research have proved experimental and theoretic that the pitch information expression in cochlear is not through the nonlinear resonance,but the time series variation caused by nonlinearity.This work may guide the improvement of the cochlea implants' coding strategy.For the noise process in hearing,previous researchers have discovered the stochastic resonance in nerve or auditory systems,which means the noise can enhance signal transmitting,but this phenomenon only exists in the weak signal condition,and the researches only considered one noise source.In this research,through the Integrate-and-Fire neuron model,for the first time,the signal transmitting in a neuron with both internal and external noise have been researched.In this research,a discrete method is applied,solved the first passage problem in an I-F neuron with two independent noise,which cannot be solved by existing classical theory.Based on this method,the effect of noise on firing timing precision and the effect of noise on the periodic signal's expression have been studied.It has been found that the intrinsic noise of a neuron can improve the firing time precision,this enhance effect on firing time precision is based on the internal noise that have short correlation time than the external noise.In the neuron with noisy periodic input,this research found that the internal noise can enhance the periodic signal while suppressing the external noise.This enhancement on the signal and suppressing on the noise is caused by mode-lock,which is an effect of the internal noise.The condition for this effect is that the variation of membrane potential caused by internal noise have greater roughness,which can be demonstrated as a smaller H?lder exponent,than the variation caused by the external noise.The discrete method used in this research for solving the first passage problem could be applied in a series of nonlinear problems.The neuron internal noise's enhance effect and suppress effect not only provided a possible mechanism to explain how neuron process noise,but also give the information engineering a new way to use internal noise suppress the external noise in signal processing.