MODELLING OF HUMAN AUDITORY MECHANISM

Abstract

This Thesis summarizes the research and development work done on the problem of modelling of human auditory system. Such model representation have wide important app licabilities. Electrical modelling has distinct advantages over ether types. As a consenuence electrical models are presented for outer, middle and inner ear. Stressing the importance of study and general procedure adopted for cons truction of models in introduction chapter, necessary ana tomical and physiological background has been given in the second chapter. Functionally signal conditioning of input vibratory signal is done in the outer- and middle ear whereas inner ear converts this vibratory signal into electrical pulses which are in turn transmitted to the brain. Modelling has been discussed in three sections for the three above mention ed parts of the auditory mechanism. The representation of outer ear employs tee sections in series. Outer ear has been approximated in two possible ways viz uniform canal representation and tapered canal re presentation. Procedure for calculating numerical values of electrical elements in both these cases is on physical equivalence basis. It has been attempted to explore how well the sound pressure transfer characteristics of ear canal could be simulated by above mentioned electrical analogs of various degrees of complexity.

The ear canal is terminated by tympanic membrane. The properties of tympanic membrane as band pass filter have been analysed. VarimiS results obtained on such filter model hove been tabulated and plotted. The agreement of model with real ear data is reasonable. This possibly is the first attempt for having Such type of model for canal termination. It has been shown that analog for middle structure can be based on two considerations. Analog can be constructed by taking into account the working of middle ear parts or it might be based on anatomical, evidences. Various stages of model development under both cases have been analysed. The acoustic impedance of the ear has been of interest almost ex clusively in diagnosis of middle ear diseases. In both the models, therefore, impedance matching with real ear data has been given utmost importance. Tht discussed models compare favourably with available real ear data. Secondly based on the compiled physiological evidences, a new electronic model has been proposed. This model has capa bility to simulate the frequency response and limiting charac teristics of the actual middle ear. Output of model is shown to have reasonable agreement with experimental findings. Such model representation may find applications as prosthetic device. Vibratory sound signals are transmitted to the brain in the form of electrical signals. This conversion of vibra tory sound to electrical signal takes place in inner ear. A model is proposed to give the simple but important relation ship between externally recorded electrical activity and

actual generated signal in inner ear.It has been proved that generated hair cell potential of inner err closely corresponds the mechanical activity of the basilar membrane. The system at which this transduction takes place has been identified, and the transfer function by using difference equation tech nique has been obtained for the same. Validity of the trans fer' function has been checked by numerical calculations which indicated its close approximation to the system considered. The identified transfer function can be used to know the system behaviour- under variety of simulated conditions. Fur ther a lumped equivalent electrical network model is pre sented for human inner ear cochlea which is an improvement over existing formulations. The model is compared with re cent analogs. Qualitative evaluation of the model reveals its agreement with the experimental findings. Finally on the basis of existing literature on the cochlea, pattern process ing phenomenon has been explained. A synaptic matrix for audition has been proposed which is formed by inner and outer hair cells. Pattern discrimination, recognition and adapta tion phenomena are explained by considering such matrix. Assumptions involved at every stage in the above work are assessed with the present state of acquired knowledge on human auditory system. The main conclusions have been summari zed in the last chapter.