Cochlear modelling of distortion product otoacoustic emissions J How, S Elliott

1
Cochlear modelling of distortion product
otoacoustic emissionsJ How, S Elliott B
Lineton
SPCG away day July 2009
Introduction The cochlea, situated in the inner
ear (figure 1), produces sounds which can be
detected in the ear canal. Distortion product
otoacoustic emissions (DPOAEs) are emitted by the
cochlea when it is stimulated by two pure tones
(f1 and f2 where f2gtf1) simultaneously, and occur
at frequencies which are integer combinations of
the stimulus frequencies DPOAEs can be used
clinically as an objective screening test for
hearing loss, but further applications have been
impeded by limited understanding of the
mechanisms by which the DPOAEs are generated and
travel out of the cochlea. The largest DPOAE in
humans (fdp2f1-f2) is thought to arise from two
difference sources within the cochlea a
wave-fixed (distortion) component located near
the f2 characteristic place and a place-fixed
(reflection) component originating from the fdp
characteristic place. We have developed a
classical, lumped-parameter cochlear model of the
type illustrated in figure 3, to study the
cochlea and simulate the production of DPOAEs.
The model incorporates the nonlinear behaviour of
the outer hair cells and so the simulations
exhibit known cochlea response characteristics
such as compressive amplitude growth with
increasing stimulus level as well as harmonic and
intermodulation distortion. We have applied the
model to confirm that the fine structure, which
can be seen when DPOAE amplitude is plotted as a
function of fdp, can essentially be attributed to
the mixing of the wave-fixed and place-fixed
components of the DPOAE in the ear canal.

Methods
Figure 2 Representation of a one-dimensional
lumped parameter cochlear model
Results
Figure 3 The amplitude and phase of the 2f1-f2
DPOAE as a function of emission frequency for
fixed f2/f1 ratio (1.2). A) Measurement in a
human ear canal using L151 and L230 dB SPL.
Adapted from figure 1 of 5. B) Model simulation
using L1L230dB SPL. In each case the grey solid
line represents the total DPOAE, and the black
solid and dotted lines represent the wave-fixed
(distortion) and place-fixed (reflection) DPOAE
components respectively.
Conclusions The introduction of random
perturbations into the active micromechanics of a
nonlinear active cochlear model can simulate
DPOAE fine structure, consistent with
experimental observations. This confirms that
fine structure can arise due to mixing of
wave-fixed and place-fixed emission components in
the ear canal. Figure 3 illustrates that both the
reflection component, and the total DPOAE,
exhibit fine structure in physiological
measurements and in the model simulation. Further
work is needed to establish if this result is
also valid for other DPOAE frequencies in
addition to 2f1-f2.
References 1 Neely Kim (1986). A model for
active elements in cochlear biomechanics. JASA,
79, 1472-1480. 2 Ku (2008). Modelling the
Human Cochlea. University of Southampton, ISVR,
PhD Thesis. 3 Kanis de Boer (1993).
Self-suppression in a locally active nonlinear
model of the cochlea A quasilinear approach.
JASA, 94, 3199-3206. 4 Mauermann Kollmeier
(2004). Distortion product otoacoustic emissions
(DPOAE) input/output functions and the influence
of a second DPOAE source. JASA, 116,2199-2212
Figure 1 The ear. Adapted from
http//www.ich.ucl.ac.uk/factsheets/families/F0100
74/images/ear.gif