Measurements and model of the cat middle ear: evidence of tympanic membrane acoustic delay.

Abstract

In order to better understand the mechanics of tympanic membrane (TM) transduction at frequencies above a few kHz, the middle-ear (ME) impedance measured near the tympanic membrane is studied for three anesthetized cat ears after widely opening the ME cavities (MEC). Three conditions were measured: intact ossicles, drained cochlea, and disarticulated stapes. When the cochlear load is removed from the ME by disarticulating the stapes, the impedance magnitude varies by about +/- 25 dB in the 5- to 30-kHz range, with peaks and valleys at intervals of approximately 5 kHz. These measurements suggest middle-ear standing waves. It is argued that these standing waves reside in the TM. In contrast, the magnitude of the impedance for the intact case varies by less than +/- 10 dB, indicating that for this case the standing waves are damped by the cochlear load. Since the measurements were made within 2 mm of the TM, standing waves in the ear canal can be ruled out at these frequencies. Although the ME cavities were widely opened, reflections from the ME cavity walls or surrounding structures could conceivably result in standing waves. However, this possibility is ruled out by model predictions showing that such large standing waves in the ME cavity space would also be present in the intact case, in disagreement with the observation that standing waves are damped by cochlear loading. As a first-order approximation, the standing waves are modeled by representing the TM as a lossless transmission line with a frequency-independent delay of 36 microseconds. The delay was estimated by converting the impedance data to reflectance and analyzing the reflectance group delay. In the model the ossicles are represented as lumped-parameter elements. In contrast to previous models, the distributed and lumped parameter model of the ME is consistent with the measured impedance for all three conditions in the 200-Hz to 30-kHz region. Also in contrast with previous models, the ear-canal impedance is not mass dominated for frequencies above a few kHz. Finally, the present model is shown to be consistent, at high frequencies, with widely accepted transfer functions between (i) the stapes displacement and ear-canal pressure, (ii) the vestibule pressure and ear-canal pressure, and (iii) the umbo velocity and ear-canal volume velocity. An improved understanding of TM mechanics is important to improve hearing aid transducer design, ear-plug design, as well as otoacoustic emissions research.