Research project: Signal processing of otoacoustic emissions (OAEs)

OAEs are low level signals generated by the active amplifier of a healthy cochlea and are detectable in the ear canal using a sensitive microphone.  They are most easily measured after an evoking stimulus such as a click is presented to the ear.  The aim of this project theme is to develop signal processing techniques that can be applied to OAEs from humans in order to aid in the interpretation of the signals.  Questions that are being addressed are: what can the features of the OAE signal reveal about the patient's auditory system? What are the optimal evoking stimuli and signal processing strategies for recording OAEs, in order to minimise the problems of stimulus artefacts and noise, and thus, for example, to maximise sensitivity and specificity for screening?

Analysing OAE recordings:

What do OAEs sound like?  The sounds recorded in the ear canal are usually very brief (1/50 second).  To hear these clearly, we can breaking the recordings down into time-frequency components and then slowing down the occurrence of these components. Each ear produces its own distinctive (but repeatable) sound.

We are conducting research into how these patterns of components are related to cochlear phenomena.  One approach is to attempt to model the generation of OAEs from first principles.

One area of research is to compare simulated and measured OAEs in order to understand the link between processes occurring within the cochlea and the sounds that can be recorded in the ear canal.

Novel Measurement Techniques

Conventional recording techniques for transient evoked OAEs (TEOAEs) are not able to record frequency components above about 4.5 kHz, as these are obscured by a stimulus artefact.  One area of research is in developing improved measurement techniques to allow higher frequency components to be recorded, which is likely to provide clinically useful information.