Nonlinear system identification for otoacoustic emissions
Nonlinear system identification for otoacoustic emissions
This thesis is concerned with the identification of a nonlinearity within the cochlea, which is at the heart of its fine frequency selectivity and it’s massive dynamic range. The source of the nonlinearity is believed to be an active process, which has a compressive response and also generates otoacoustic emissions (OAEs), that are acoustic signals produced within the cochlea and recorded in the ear canal. They exhibit similar nonlinear compression characteristics to that of the response of the cochlea, and as such can be used as a non-invasive research tool to characterise the nonlinearity.
A physiologically reasonable model is developed for the generation and emission of OAEs, based on a cascade structure with two linear elements surrounding a nonlinear block. A technique for decoupling and identifying the components of this type of nonlinear model, under certain conditions, has been developed. Two types of potential cochlear nonlinearity have been investigated, one with a static, instantaneous, power-law, and the second with a level-dependent nonlinearity known as an automatic gain control (AGC). These are both shown to be capable of producing physiologically realistic compression curves. The best way of distinguishing between these models appears to be in terms of their instantaneous input-output characteristics. A comparison of simulation results with previous experimental measurements suggests that the AGC model may be more appropriate for modelling cochlear nonlinearity.
A series of experiments, carried out on human subjects, is also presented which investigates the level-dependence of transient- and derived noise-evoked otoacoustic emissions, and provide further support for the theory that the inherent nonlinearity in the cochlea is level-dependent.
University of Southampton
2004
Harte, James Michael
(2004)
Nonlinear system identification for otoacoustic emissions.
University of Southampton, Doctoral Thesis.
Record type:
Thesis
(Doctoral)
Abstract
This thesis is concerned with the identification of a nonlinearity within the cochlea, which is at the heart of its fine frequency selectivity and it’s massive dynamic range. The source of the nonlinearity is believed to be an active process, which has a compressive response and also generates otoacoustic emissions (OAEs), that are acoustic signals produced within the cochlea and recorded in the ear canal. They exhibit similar nonlinear compression characteristics to that of the response of the cochlea, and as such can be used as a non-invasive research tool to characterise the nonlinearity.
A physiologically reasonable model is developed for the generation and emission of OAEs, based on a cascade structure with two linear elements surrounding a nonlinear block. A technique for decoupling and identifying the components of this type of nonlinear model, under certain conditions, has been developed. Two types of potential cochlear nonlinearity have been investigated, one with a static, instantaneous, power-law, and the second with a level-dependent nonlinearity known as an automatic gain control (AGC). These are both shown to be capable of producing physiologically realistic compression curves. The best way of distinguishing between these models appears to be in terms of their instantaneous input-output characteristics. A comparison of simulation results with previous experimental measurements suggests that the AGC model may be more appropriate for modelling cochlear nonlinearity.
A series of experiments, carried out on human subjects, is also presented which investigates the level-dependence of transient- and derived noise-evoked otoacoustic emissions, and provide further support for the theory that the inherent nonlinearity in the cochlea is level-dependent.
This record has no associated files available for download.
More information
Published date: 2004
Identifiers
Local EPrints ID: 465237
URI: http://eprints.soton.ac.uk/id/eprint/465237
PURE UUID: 5b444dd0-f92a-4558-a4e2-f1c85cdc2f96
Catalogue record
Date deposited: 05 Jul 2022 00:30
Last modified: 05 Jul 2022 00:30
Export record
Contributors
Author:
James Michael Harte
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics