Download Diagnosis and Characterization of Middle

AUDIOLOGY·청능재활
ISSN 1738-9399
REVIEW PAPER
2013;9:95-112
Copyright ⓒ 2013 Korean Academy of Audilogy
Diagnosis and Characterization of Middle-ear and
Cochlear Functions
Biomedical and Mechanical Engineering, University of Oklahoma, OK, USA
Rong Z. Gan
ABSTRACT
Hearing impairment is one of the most common physical disabilities in the world. Conductive hearing loss results
from dysfunction of the middle ear, and sensorineural hearing loss results from abnormalities within the cochlea or
central nervous system. To improve the diagnosis of hearing loss and develop new technologies for restoration of
hearing, changes of the middle ear and cochlear functions in relation to ear structural disorders in hearing impaired
ears need to be characterized. This paper summarizes our biomedical engineering approach to characterize the middle
ear and cochlear functions in normal and diseased ears. With the experimental measurements in human temporal
bones and animal models and the finite element (FE) modeling of the human ear, a better understanding of
biomechanics of the ear for sound transmission is achieved. The acoustic-mechanical vibration and energy
transmission through the middle ear in normal and diseased ears can be visualized and quantified in a 3-dimensional
FE model. Four novel model-derived “auditory test curves” named as the middle ear transfer function, energy
absorbance, admittance tympanogram, and tympanic membrane holography are generated with clinically relevant
applications.
Keywords: Cochlea, Ear biomechanics, Finite element model, Hearing loss, Laser vibrometry, Middle ear, Middle ear
disorder
loss and the abnormalities within the cochlea or
INTRODUCTION
central nervous system results in sensorineural hearing
The human middle ear, including tympanic
loss. To improve diagnosis of hearing loss and
membrane (TM) and three ossicules (i.e., malleus,
develop new technologies for restoration of hearing,
incus and stapes) suspended in an air-filled middle ear
the changes of the middle ear and cochlear functions
cavity by suspensory ligaments/muscles, constitutes a
in relation to ear structural disorders in hearing
mechanical system for sound transmission from the
impaired ears need to be characterized.
external ear canal to the inner ear or cochlea. The
The first area for function characterization is the
middle ear dysfunction results in conductive hearing
middle ear transfer function. Sound entry to the
middle ear through the TM initiates the transfer
Received: Nov 28, 2013
Accepted: Nov 30, 2013
Corresponding Author: Rong Z. Gan, School of Aerospace and
Mechanical Engineering, University of Oklahoma, 865 Asp Avenue,
Room 200, Norman, OK 73019; E-mail: [email protected];
Tel: 1-405-325-1099
function. Output of the middle ear system is an
acoustic-mechanical transmission by the states footplate,
which sets into the oval window and transmits
vibrations into cochlear fluid. Since the transfer
95
96
AUDIOLOGY·청능재활 2013;9:95-112
function of the middle ear cannot be measured readily
middle ear pressure was varied from zero to positive
in living humans, various experimental measures using
or negative in human temporal bones (Dai et al., 2008a;
the fresh human cadaveric temporal bones (Dirckx &
Gan et al., 2006a; Ravicz et al., 2004).
Decraemer, 1992; Gan et al., 1997, 2001, 2004a; Gyo
In addition to studies in temporal bones, the
et al., 1987; Merchant et al., 1997; Peake et al., 1992;
diseased animal models provide a unique approach to
Ravicz et al., 2004; Voss et al., 2000), animal models
investigate the effects of changes of middle ear
(Dai & Gan, 2008b, 2010; Guan & Gan, 2011, 2013;
structure such as the tissue inflammation or infection
Larsson et al., 2005; Lee & Rosowski, 2001; Qin et
on METF. Animal models of otitis media have been
al., 2010; Turcann et al., 2009; von Unge et al.,
reported in the literature, but the changes of TM
1995), and different theoretical modeling methods
mobility and METF are recently reported by our group
(Feng & Gan, 2004b; Gan et al., 2004, 2006b, 2010;
in otitis media models of guinea pig and chinchilla
Goode et al., 1994; Kringlebotn, 1988; Sun et al.,
(Dai & Gan, 2008b, 2010; Guan & Gan, 2011, 2013a;
2002a; Zhang & Gan, 2011a; Zwislocki, 1962) have
Guan et al., 2013a, 2013b). The acoustic-mechanical
been used to describe middle ear mechanics in normal,
transmissions in otitis media ears are revealed through
diseased, and reconstructed ears.
those studies with a close relation to structural variations
The magnitude of stapes vibration is about 0.1 nm
in the middle ear.
at the threshold of hearing. To measure such small
Moreover, the auditory testing curves of sound
vibrations in human temporal bones, laser Doppler
energy transduction from the ear canal to middle ear
vibrometry (LDV) provides a critical technique for
are frequently measured in clinics (Dirks & Morgan,
accurately measuring any moving middle ear structures
2000; Feeny et al., 2003, 2009; Keefe et al., 1993;
such as the TM, ossicles, and round window membrane
Keefe & Simmons, 2003). The wideband acoustic
without contact (Dai et al., 2007a, 2007b, 2008a; Gan
admittance (tympanogram), middle ear energy absorbance,
et al., 2001, 2004a, 2006a, 2009; Hato et al., 2003;
and auditory brainstem response (ABR) are also
Murakami et al., 1997; Nuttall & Dolan, 1996; Rosowski
obtained from animal models of otitis media (Dai &
& Lee, 2002). Based LDV measurements of the ear
Gan, 2008b; Guan & Gan, 2011). The experimental
components in response to sound stimuli in the ear
results provide a correlation between the mechanical
canal, the middle ear transfer function (METF), defined
measurements and hearing level in normal and diseased
as the relationship between vibration of the stapes
ears. This is the second area of middle ear function
footplate and sound pressure level (SPL) in the ear
characterization: sound energy transduction through
canal or the vibration ratio of the stapes footplate to
the ear.
the TM, can be characterized over the auditory
To better understand the acoustic-mechanical
frequency range (Dai et al., 2007b, 2008a; Gan et al.,
transmission from the ear canal to middle ear and
2004a, 2006a; Guan & Gan, 2011). To further investigate
cochlea, different theoretical models have been
changes of METF in diseased ears such as otitis
developed to simulate the function of the ear. Early
media with effusion diagnosed with fluid in the middle
modeling of middle ear function used a transformer
ear cavity and middle ear pressure different from
analogy with circuit models or lumped parameter models
ambient pressure, the movements of the TM and
reported by Hudde & Weistenhöfer(1997), Kringlebotn
stapes footplate are measured while saline or silicone
& Gundersen(1985), and Zwislocki(1960). We also
fluid was introduced into the middle ear and the
published a lumped parametric model of the human ear
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions
97
to study the METF (Feng & Gan, 2004). However,
METF, energy absorbance (EA), admittance tympanogram
the circuit model has the limitation in simulating the
(AT), and TM holography are generated with clinically-
realistic acoustic-mechanical response in the ear
relevant applications.
involving complex geometry and an array of material
compositions.
The finite element (FE) model as a computer-based
numerical procedure has distinct advantages in
modeling complex biological system such as the ear.
FE method is always capable of modeling the
MATERIALS AND METHODS
1. Measurement of acoustic-mechanical
vibrations with laser Doppler vibrometry
complex geometry, ultrastructural characteristics, and
As one of few groups in 1990s using LDV to
non-homogenous and anisotropic material properties of
measure vibrations in the ear, the LDV (Polytec,
biological systems. FE models can also determine the
Model OFV-501) was first used for measuring the
detailed vibration modes, stress distributions, and
movement of stapes footplate at the frequencies of 200
dynamic behaviors at any locations in a system, which
~ 8,000 Hz to determine the normal middle ear
is not possible with analytical solutions and experimental
function and the effect of implant mass on stapes
measures (Gan et al., 2004b, 2006b; Koike & Wade,
movement (Gan et al., 1997, 2001). The input sound
2002; Lee et al., 2006; Qi et al., 2008; Sun et al,
level in the ear canal at 2 mm from the TM (umbo)
2002a, 2002b). A systematic literature search on
was increased from 80 to 90 dB and 100 dB. The
middle ear FE model and analysis with a focus on FE
increase of footplate displacement for every 10 dB
modeling of the middle ear transfer functions and
SPL change was equal to 10 dB over the whole
related pathologies was conducted by Zhao et al.
frequency range. The results showed the linearity of
(2009). This review paper indicated that the FE model
normal METF as the frequency response curves of the
of human ear developed by our group was the most
stapes footplate displacement in response to sound
comprehensive model to date (2009), which was based
input level at the TM. When the implant mass was
on a complete set of histological section images of the
applied at the incus-stapes joint (ISJ), this linearity did
temporal bone.
not change.
The present paper summarizes our biomedical
Two LDVs (Polytec, Models CLV-700 and OFV-
engineering approach to characterize the middle ear
501) were then used to measure vibrations of the TM,
and cochlear functions in normal and diseased ears.
stapes, and ISJ. One laser beam (CLV-700) was
The experimental measurements of METF in human
focused on the TM at the umbo through the ear canal
temporal bones and animal models, the measurements
and the second laser (OFV-501) was used to measure
of mechanical properties of ear tissues, and the finite
movement of the stapes footplate or ISJ caused by
element modeling of human ear are presented, which
acoustic vibration of the TM. Distribution of the
provide basic understanding of biomechanics of the ear
vibration along ossicular chain was revealed through
for sound transmission. The acoustic-mechanical vibration
the experiments (Gan et al., 2004a). Stapes and TM
and energy transmission through the middle ear in
displacement transfer functions were determined using
normal and diseased ears are visualized and quantified
dual LDV, provided accurate amplitude and phase
in a 3-dimensional (3D) FE model of the ear. Four
relationships from stapes footplate, ISJ, and TM.
novel model-derived “auditory test curves” named as the
After the normal function of the middle ear was
98
AUDIOLOGY·청능재활 2013;9:95-112
characterized in temporal bones, the middle ear disease
middle ear pressure and fluid levels in the cavity
such as otitis media with effusion (OME) was
with a correction factor for cochlear load. The results
simulated in temporal bones by injecting fluid (saline)
demonstrated that the middle ear pressure (positive and
into the middle ear cavity and adjusting pressure in
negative) reduced the TM motion at frequencies below
the middle ear. Two LDVs were used to detect the
1.5 kHz; fluid in the middle ear cavity reduced the
change of METF in OME model of temporal bones.
TM motion at frequencies over 1 kHz; a combination
The experimental setup displayed in Figure 1 is from
of fluid and pressure reduced the movement of the
Gan et al.(2006a). A hearing laser vibrometer
TM at all frequencies (Dai et al., 2008a; Gan et al.,
(Polytec, CLV-700) was used to measure the vibration
2006a). In addition to a single point LDV, we have
of the TM (umbo) and the second LDV (OFV-501)
used a scanning laser Doppler vibrometer (Polytec,
was used to measure the movement of stapes from the
PSV-400) to measure and image the full-field surface
medial side of the footplate. The experiment began
motion of the TM in temporal bones (Zhang et al,
with one laser only to measure the TM displacement
2013). The external ear canal was surgically removed
in bones with intact cochlea. The METF, which is
until about 90% of the TM surface was visible under
defined as the displacement transmission ratio (DTR)
the microscope.
of the TM to footplate, was derived under different
Figure 1. Schematic diagram of the experimental setup in human temporal bones with two laser vibrometers
for measuring vibrations at the TM and stapes footplate. One laser was aimed on the TM at the umbo and
another laser was aimed on the medial site of the stapes footplate with open cochlea. From Gan et al.
(2006a)
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions
99
TM displacements were measured under normal (no
following LPS injections. The change of the DTR
fluid in the middle ear cavity) and diseased conditions
from the TM to the round window occurred mainly in
(OME, fluid in the cavity) with different fluid levels over
chronic (e.g., 14 days post LPS injection) OME ears.
frequencies ranging from 0.2 to 8 kHz. The results from
To further investigate cochlear function in OME
experiments showed that in normal TM a simple
ears, displacement of the basilar membrane (BM) in
deflection shape with one or two major displacement
guinea pig ears with OME (3 and 14 days after LPS
peak regions was observed at low frequencies (1 kHz and
injection) were measured at the basal turn and the
below) while complicated ring-like pattern of the
apex (Dai & Gan, 2010). The measurement of BM
deflection shapes appeared at higher frequencies (4 kHz
movement was conducted in guinea pig temporal
and above). The middle ear fluid mainly affected TM
bones and all vibration measurements were completed
deflection shapes at the frequencies higher than 1 kHz.
within 2 h of the guinea pigs’ deaths. The results
showed that the displacement sensitivity of the BM at
the apex and the basal turn to sound pressure in the
2. Animal models of otitis media for
function characterization
Otitis media with effusion and acute otitis media
(AOM) are two main types of otitis media. OME
ear canal was reduced up to 25 dB at their characteristic
frequencies, respectively. Cochlear gain with respect to
umbo movement was also changed in ears with OME
in both groups.
describes the symptoms of middle ear effusion without
Combined measurements of METF and ABR
infection, and AOM is an acute infection of the
threshold in live guinea pigs with middle ear effusion
middle ear and caused by bacteria in about 70% of
(MEE) were reported by Guan & Gan(2011). The MEE
cases (Gould & Matz, 2010). With the long term goal
was created by injecting saline into the middle ear
of evaluating middle ear function with otitis media,
cavity from zero (control) to 0.1 and 0.2 ml (full fill of
animal models of OME and AOM were created in
the cavity). Vibrations of the TM (umbo), the tip of the
which middle ear transfer functions could be measured.
incus, and the round window membrane were measured
OME was created by injecting lipopolysaccharide
with a laser vibrometer (Polytec, HLV 1000) at
(LPS) into the middle ear in guinea pigs (Dai & Gan,
frequencies of 0.2 ~ 40 kHz when the middle ear fluid
2008b, 2010). The experimental animals were divided
increased from 0 to 0.2 ml. ABR waveform was
into 3 groups based on survival time after LPS
recorded as the middle ear fluid increased using the
injection (3, 7, or 14 days). These times reflect early,
TDT system III (TDT, FL). Repeated click stimuli and
medium, and chronic stages, respectively, of OME
tone burst stimuli with alternating polarity were
after LPS injection. Evidence of OME was assessed
delivered by the speaker (CF1, TDT) with the sound
by otoscopy, tympanograms, histological sections of
delivery tube placed 4 mm away from the umbo. In this
the TM, and by measuring the volume of fluid
study, the METF was represented by DTR of the TM
(effusion) in the middle ear. Vibrations of the umbo
to the incus tip, which varied with frequency and fluid
and round window membrane were measured with a
level. The MEE reduced the displacement of TM, incus
LDV (Polytec, Model HLV 1000) at the frequency
tip, and round window membrane mainly at high
range of 0.2 ~ 40 kHz in control and three groups of
frequencies and the reduction was related to the
OME ears. Displacement of both the umbo and round
volume of fluid in the cavity.
window membrane was reduced at all time-points
100 AUDIOLOGY·청능재활 2013;9:95-112
Figure 2. Observations of the TM and ossicles in control and diseased guinea pig ears after 3 days
of inoculation. Upper panel: otoscopic photographs of TM in control (A), OME (B), and AOM ear (C).
Lower panel: microscopic photographs of middle ear ossicles in control (D), OME (E), and AOM ear
(F). From Guan et al. (2013ae) (Color in online)
Acute otitis media was created in both guinea pig
behind the TM. In AOM ear (Figure 2C), purulent
and chinchilla by using two most common bacterial
effusion was observed in the middle ear and the TM
pathogens: Streptococcus pneumonia and Haemophilus
showed distinct redness resulting from dilation of blood
influenza reported by Guan & Gan(2013) and Guan et
vessels. Figures 2D ~ 2F display the typical ossicular
al.(2013a, 2013b). The AOM model of guinea pig was
chain and the round window niche in normal, OME, and
created by intrabullar injection of Streptococcus
AOM ears. In control ear (Figure 2D), there were no
pneumonia type 3 (Guan & Gan, 2013; Guan et al.,
signs of middle ear inflammation, and the middle ear
2013a) and the AOM model of chinchilla was created
ossicles were clearly identified. In OME ear (Figure 2E),
by injection of Haemophilus influenza (Guan et al.,
the ossicular chain appeared normal. In AOM ear (Figure
2013b). The microscopic observations and histological
2F), the purulent adhesions were formed between the
section images of the TM, middle ear cavity, and
ossicles and the middle ear bony wall and commonly
ossicles showed the morphological or middle ear
observed between the manubrium and promontory, and
structural changes in AOM ears. Figure 2 displays the
around the oval window and round window niche.
microscopic observations of the TM and middle ear in
The displacement of the TM at the umbo in
control, OME, and AOM ears of guinea pigs (Guan et
response to 80 dB SPL sound in the ear canal was
al., 2013a). In normal ear (Figure 2A), the TM was
measured in AOM ears using LDV and compared
translucent and no fluid was visible behind the TM.
with those obtained from control and OME ears. It
In OME ear (Figure 2B), serous effusion was present
was found that the TM mobility in AOM ears was
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions
101
lower than that in experimental OME ears at low
along our three models: from the TM and middle ear
frequencies, and the difference was mainly caused by
ossicular chain only (Model #1) to the complete
the middle ear ossicular structure changes during the
middle ear with the cavity and external ear canal
bacterial infection in AOM (Guan et al., 2013a).
(Model #2) and to the entire ear with the spiral
cochlea (Model #3). All the models have the accurate
3. 3D FE model of the human ear
Using the combined technologies of FE analysis
geometric configurations of the ear components and
the capability to analyze the transmission of sound
pressure-induced vibrations through the ear.
and 3D reconstruction, we have published three FE
Model #3 (male, age 52, left ear) was first reported
models of the human ear based on the complete sets
by Gan & Wang(2007) without cochlea attached to the
of histological section images of three adult temporal
oval window or stapes. This preliminary model provided
bones (Gan & Wang, 2007; Gan et al., 2002, 2004b,
insight into simulation of middle ear effusion and the
2006b, 2007; Gan et al., 2009, 2010; Sun et al., 2002a,
acoustic-structure-fluid coupled FE analysis. The model
2002b; Wang et al., 2007; Zhang & Gan, 2011a, 2013a).
included the external ear canal, middle ear ossicles
The 3D solid model was first constructed based on
with attached ligaments and muscle tendons, and
histological section images using the 3D geometry
middle ear cavity. By introducing saline or silicon
reconstruction software such as SolidWorks. Then, the
fluid into the middle ear cavity, Gan & Wang
model was meshed into a number of elements using
quantitatively evaluated how fluid level and viscosity
HyperMesh software. Finally, the meshed model was
affected the middle ear transfer function. Recently, this
input into ANSYS for FE analysis. Our long-term goal
model has been improved by adding the spiral cochlea
for modeling of the ear is to develop a comprehensive
with three chambers of the scala vestibule (ST), scala
3D model of the human ear with clinically-relevant
media (SM) and scala tympani (ST) separated by
applications.
cochlear basilar membrane and Reissner’s membrane as
The complexity of FE model was gradually increased
shown in Figure 3 (Zhang & Gan, 2011a, 2013a).
Figure 3. (A) FE model of the human ear in anterior-medial view. The model consists of external ear canal,
TM, 3 ossicles, 2 joints, 6 ligaments and muscle tendons, stapedial annual ligament, middle ear cavity, and
cochlea. (B) FE model of cochlea with the BM (SV, SM and RM are shown transparently). (C) Cross section of
cochlea model showing three chambers and two membranes. From Zhang & Gan (2011a) (Color in online)
102
AUDIOLOGY·청능재활 2013;9:95-112
The basilar membrane has two and half turns from
2012a, 2012b). Compared with the adult model, the ear
the base to the apex. Helicotrema is the place where
canal volume and middle ear cavity volume of the
the SV and ST chambers are connected with each
pediatric model are smaller than those in the adult
other. This is the most comprehensive ear model up
model. However, dimensions of cochlea in pediatric
to date and is capable to analyze middle ear and
model are similar to that of the adult model. The
cochlear functions and the interactions between the
validation process for the pediatric model is currently
middle ear and cochlea.
under investigation with a focus on identifying
In addition to the importance of anatomic structure
mechanical properties of soft tissues for young subjects.
for an ear model, the accuracy of FE model for
function analysis also depends on material properties
of model components such as the mechanical
RESULTS
properties of middle ear soft tissues. In the previously
published models (Gan & Wang, 2007; Gan et al.,
2002, 2004b, 2006b, 2007, 2009; Sun et al., 2002a),
1. Validation of the FE model in simulating
middle ear and cochlea functions
the TM, suspensory ligaments and muscle tendons
To validate the FE model of human ear, 90 dB
were all assumed as elastic materials. In fact, the
SPL pure tones were applied in the ear canal with 2
human ear works at the auditory frequency range from
mm away from the TM. Harmonic structure-acoustic
20 Hz to 20 kHz and the ear tissues show viscoelastic
-fluid coupled analysis was conducted over the auditory
or frequency-dependent dynamic properties which were
frequency range from 0.2 to 10 kHz. The METF
measured and confirmed in our lab (Gan et al., 2013;
defined as the displacement of the TM and stapes
Luo et al., 2009a, 2009b; Zhang & Gan, 2010, 2013b,
footplate (FP) were derived from the FE model and
2013c). Thus, the dynamic properties of middle ear
compared with the published experimental measurements
tissues should be included in FE model to achieve the
(Figure 4A). The sound pressure gain defined as the
accurate analysis for acoustic-mechanical transmission.
ratio between fluid pressure at the base of cochlear SV
In our comprehensive FE model of the human ear, the
and sound pressure at the ear canal was derived from
viscoelastic material properties are applied to the
the model and compared with the published experimental
middle ear soft tissues and the model has resulted in
data (Zhang & Gan, 2011a) (Figure 4B).
a better simulation of METF over the auditory
The model was further validated for middle ear
frequency range from 0.2 to 10 kHz (Zhang & Gan,
input impedance and energy absorbance in comparison
2011a, 2013a).
with the experimental results in Figure 5. Figures 5A
Considering otitis media is the most frequently
and 5B show the magnitude and phase of the model-
diagnosed illness in young children, a pediatric model
derived middle ear impedance at the TM compared
is needed for investigating the mechanisms of otitis
with the experimental results measured on human
media in pediatric population. Based on the histological
temporal bones by Rabinowitz(1981), Stepp & Voss
section images of a temporal bone (male, 4-year old,
(2005), and Voss et al.(2000).
left ear), a pediatric model is constructed in our lab.
A general agreement between the model and
This model consists of the ear canal, TM, middle ear
experimental results was observed over the frequency
ossicles with suspensory ligaments and muscle tendons,
range of 0.2 to 4 kHz.
middle ear cavity, and spiral cochlea (Zhang & Gan,
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions 103
Figure 4. (A) Displacement of the TM and stapes
footplate (FP) derived from the FE model compared
with the experimental measurements in human temporal
bones. (B) Sound pressure gain derived from FE model
compared with published experimental data. From
Zhang & Gan(2011a)
Figure 5C shows the model-derived EA curve
compared with the clinical measurements in patients’
ears of normal hearing at frequencies of 0.2 to 8 kHz.
The model-derived EA was generally consistent with
those reported by Feeney et al.(2003), Keefe et al.
(2003), Margolis et al.(1999), and Voss & Allen(1994).
The model-derived EA values were slightly higher
than the mean values of the published results from 3
to 5 kHz. However, it was still in the 5% to 95%
Figure 5. (A) and (B): The FE model-derived
middle ear (ME) input impedance at TM (thick solid
line) compared with the published experimental
data in magnitude (A) and phase (B). (C): The FE
model- derived energy absorbance (EA) (thick
solid line) compared with the published measurement
data on human ears. The thin dot lines represent the
5-95% range of EA measured on 75 normal-hearing
adults by Feeney et al. (2003). From Zhang & Gan
(2013a) (Color in online)
range of EA measured on 75 normal-hearing adults by
Feeney et al.(2003).
the model was used to investigate the efficiency of
After validation with the experimental measurements
the forward and reverse mechanical driving with middle
obtained in human temporal bones and the clinical data,
ear implant and the passive vibration of basilar
104
AUDIOLOGY·청능재활 2013;9:95-112
membrane with cochlear implant placed in the cochlear
(AT), and TM holography. Among these 4 curves,
scala tympani. The middle ear transfer function and
METF represents equivalent conductive hearing loss as
the cochlear function of the basilar membrane vibration
the displacement of stapes footplate per sound pressure
were derived from the model. This comprehensive ear
in the ear canal; AT represents the compliance of the
model provides a novel computational tool to visualize
middle ear as the ratio of volume velocity of the TM
and compute the implantable hearing devices and
to pressure; EA shows the ratio of energy absorbed by
surgical procedures (Gan et al., 2010; Zhang & Gan,
the middle ear; TM holography (time-average)
2011a).
illustrates a full-field view of TM motion with
displacement of each point on the TM. Each curve is
2. FE model-derived auditory test curves
- auditory test modeling system (ATMS)
correlated to patient age (adult or young children) and
disease with defined parameters of the middle ear
structure and tissue mechanical properties.
ATMS software is built based on two 3D FE
Users can input the patient information and
models of the human ear: the adult and pediatric
clinical observations into ATMS and select the
models and some preliminary results are reported
function they want to check, such as the METF, EA,
recently by Gan et al.(2014) and Ji et al.(2013). ATMS
AT, or TM holography. For example, for observations
allows audiologists and physicians to input the patient
of “TM stiffness increase”, the stiffness change of
information and their observations during the hearing
the TM (pars tensa and flaccida), or incus-stapes
test. The software package includes two normal
joint, or stapedial annular ligament can be selected
models and two diseased models named as “general
separately or together. Figures 6 ~ 8 show a few
middle ear disorders” and “otitis media” for both adult
examples of the auditory curves under “general
and pediatric ears. The outputs from ATMS are 4
middle ear disorders” directory in ATMS database.
novel model-derived “auditory test curves”: the middle
The curves reflect the effects of stiffness changes
ear transfer function (METF), wideband energy
of the middle ear on METF, AT, and EA.
absorbance (EA), wideband admittance tympanogram
Figure 6. (A) The METF curves in response to stiffness changes of IS joint in ATMS database (B) The AT
curves over ear canal pressure level from -300 to 300 daPa in response to TM stiffness changes in ATMS
database (Color in online)
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions 105
Figure 6A shows the METF (stapes footplate
not linearly related to the change of TM stiffness. At
displacement) vary with the IS joint stiffness. As can
frequencies above 4 kHz, there was no significant
be seen in this figure, the METF is not sensitive to
change of EA as TM stiffness varied. Figure 7B
stiffness increasing between 1.5 and 4 times the
shows the EA in response to the ossicular chain
normal value. However, when the stiffness increased to
stiffness changing into 0.2, 5, 25, and 100 times the
8 times the normal value, the stapes footplate
normal value. When the ossicular chain became stiffer
displacement was increased by 10 ~ 20 dB over the
(×5, ×25 or ×100), the EA decreased at frequencies
entire frequency range. The decrease of stiffness (×0.5)
below 4 kHz, while it increased above 4 kHz. When
resulted in decreasing of footplate displacement at
the ossicular chain stiffness increased more than 25
frequencies below 2 kHz. Figure 6B shows the AT–ear
times the normal value, there was only a minor
canal pressure curves at any frequency (e.g., 226 Hz)
change of EA. In that situation, the EA was dominated
as the TM stiffness varies. The admittance increases as
by the properties of the TM. As shown in Figure
the stiffness decreased and decreases as the stiffness
7B, the softer ossicular chain (0.2 times normal
increased.
stiffness) resulted in a second peak at 1 kHz, similar
Figure 7A shows the wideband EA presented as 2D
to the EA with disarticulated ossicular chain (Zhang
EA-frequency curves in response to the TM stiffness
& Gan, 2013a). The 3D EA-frequency-ear canal
changing into 0.5, 1.5, 2, 4, and 8 times the normal
pressure curves of the ear with the normal and stiffer
value. As can be seen in Figure 7A, the TM stiffness
TM (×4 times the normal value) are displayed in
has a significant effect on EA at frequencies below 4
Figure 8. The 3D curves visualize the change of EA
kHz. The increasing of TM stiffness resulted in the
with respect to both the ear canal pressure and
decreasing of EA. However, the reduction of EA was
frequency.
Figure 7. The EA curves over the frequency range from 200 to 8,000 Hz in response to TM stiffness changes
(A) and middle ear ossicular stiffness changes (B) in ATMS database (Color in online)
106 AUDIOLOGY·청능재활 2013;9:95-112
Figure 8. The 3D EA curves varying with frequency and ear canal pressure: (A) normal ear, (B) the ear
with TM stiffness increased by 4 times the normal value (Color in online)
Examples of auditory curves under “otitis media”
0.3 ml. When fluid was 0.1 ml, the pressure dominantly
directory in ATMS database are shown in Figure 9.
affected EA, especially at the frequencies lower than
Figure 9A shows the EA-frequency curves in
4 kHz. When fluid increased to 0.3 ml, EA was
response to negative middle ear pressure (MEP)
critically affected by the fluid. The low MEE with
levels: -50, -100, -150, and -200 daPa. As the MEP
high MEP mainly affected EA at low frequencies,
was reduced from 0 (normal) to -200 daPa, the
while the high MEE with low MEP significantly
slope of the EA curve at low frequencies (below 2
changed EA at high frequencies.
kHz) decreased. The EA curve became flatter under
larger MEP. EA reached its maximum value around
4 kHz, which did not change with the MEP level.
DISCUSSIONS AND
Figure 9B shows the change of EA-frequency curves
CONCLUSIONS
when middle ear effusion (MEE) levels reached 0.1,
0.3, 0.4, and 0.6 ml at zero MEP. When MEE
Auditory test tools commonly used in clinics
increased to 0.3 ml (up to the umbo), the EA was
include the measurements of audiogram, tympanogram,
significantly reduced at frequencies above 600 Hz.
energy absorbance or reflectance, and ABR. To exam
When MEE further increased to 0.6 ml (middle ear
the ear or structural changes of the ear, physicians
cavity filled), the EA-frequency curve was flat over
routinely check patient’s otoscopy and occasionally
the entire frequency range of 200 Hz to 8 kHz.
check the CT or MRI images to identify the status of
There was almost no EA predicted when the middle
ear disorders. However, the exact cause of hearing
ear cavity was filled with fluid, which means no
loss in the tested ear may not be visible on physical
acoustic energy was transferred into the middle ear.
or radiological examination. The anatomic constrains
Figure 9C shows EA curves at four combinations of
provide challenges for audiologists or physicians to
MEP and MEE levels: -50 daPa & 0.1 ml, -50 daPa
explain the diagnosis results and the treatment options
& 0.3 ml, -100 daPa & 0.1 ml, and -100 daPa &
for hearing impairment.
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions
Figure 9. The EA curves of otitis media ear over the frequency range from 200 to 8,000 Hz in
response to negative middle ear pressure (A), middle ear fluid volume (B), and combined
negative pressure and fluid (C)
107
108 AUDIOLOGY·청능재활 2013;9:95-112
As reported in this paper, the comprehensive FE
Acknowledgements
model of the human ear developed by our group
through a series of stages has demonstrated the
The work was supported by NIH R01DC006632
advantages of this model in simulating the anatomy
and R01DC011585. The author thanks her all former
and function of the ear. This model can assist
and current students, postdocs, and research collaborators
audiologists and physicians to interpret the diagnosis of
for their contributions to the work reported in this
ear function changes in relation to its structural
paper.
variations in diseased or disordered ears. It is
anticipated that when ATMS database is fully
REFERENCES
developed, a powerful diagnosis tool with 4 novel
“auditory test curves” will be available for audiologists,
ENT students and residents, and physicians to analyze
Aibara, R., Welsh, J. T., Puria, S., & Goode, R. L.
and visualize the function and structure relationship of
(2001). Human middle-ear sound transfer function
the human ear.
and cochlear impedance. Hearing Research, 152,
Our research efforts to characterize the middle ear
100-109.
and cochlear functions have been carried on for more
Cheng, T., Dai, C., & Gan, R. Z. (2007). Viscoelastic
than a decade. It has been demonstrated that the
properties of human tympanic membrane. Annals of
combination of experimental measurements in human
Biomedical Engineering, 35(2), 305-314.
temporal bones and animals with the theoretical
Cheng, T. & Gan, R. Z. (2007). Mechanical properties
modeling provides the best approach for investigating
of stapedial tendon in human middle ear. Journal
the mechanisms of hearing impairment over various
of Biomechanical Engineering, 129, 913-918.
dimensions. Note that in this paper, one research area
Cheng, T. & Gan, R. Z. (2008a). Experimental
of measuring mechanical properties of ear soft tissues
measurement and modeling analysis on mechanical
is not covered in details. However, the quasi-static
properties of tensor tympani tendon. Medical
nonlinear mechanical properties of the TM, IS joint
Engineering and Physics, 30, 358-366.
and stapedial annual ligament completed in our lab are
Cheng, T. & Gan, R. Z. (2008b). Mechanical properties
available in the literature (Cheng et al., 2007; Cheng
of anterior malleolar ligament from experimental
& Gan, 2007, 2008a, 2008b; Daphalapurkar et al.,
measurement and material modeling analysis.
2009; Gan et al., 2011; Huang et al., 2008; Zhang &
Biomechanics and Modeling in Mechanobiology,
Gan, 2011b). The dynamic properties of ear tissues
7, 387-394.
such as the TM and round window membrane over
Dai, C., Wood, M. W., & Gan, R. Z. (2007a).
the frequency range reported with several different
Tympanometry and laser doppler interferometry
methods are also available in the literature (Gan et al.,
measurements on otitis media with effusion model
2013; Luo et al., 2009a, 2009b; Zhang & Gan, 2010,
in human temporal bones. Otology and Neurotology,
2013b, 2013c). We will continue the biomedical
28(4), 551-558.
engineering approach to reach our long-term goal: to
Dai, C., Cheng, T., Wood, M. W., & Gan, R. Z.
improve the diagnosis of hearing loss and develop
(2007b). Detachment and fixation of superior and
new technologies for restoration of hearing.
anterior malleus ligament: Experiment and modeling.
Hearing Research, 230(1-2), 24-33.
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions 109
Dai, C., Wood, M. W., & Gan, R. Z. (2008a).
Combined effect of fluid and pressure on middle
ear function. Hearing Research, 236(1-2), 22-32.
interferometer. ENT-Ear, Nose and Throat Journal,
76(5), 297-309.
Gan, R. Z., Dyer, R. K., Wood, M. W., & Dormer,
Dai, C. & Gan, R. Z. (2008b). Change of middle ear
K. J. (2001). Mass loading on ossicles and middle
transfer function in otitis media with effusion
ear function. Annals of Otology, Rhinology, and
model of guinea pigs. Hearing Research, 243,
Laryngology, 110, 478-585.
78-86.
Gan, R. Z., Sun, Q., Dyer, R. K., Chang, K-H., &
Dai, C. & Gan, R. Z. (2010). Change in cochlear
Dormer, K. J. (2002). Three dimensional modeling
response in an animal model of otitis media with
of middle ear biomechanics and its application.
effusion. Audiology and Neurotology, 15, 155-
Otology and Neurotology, 23(3), 271-280.
167.
Gan, R. Z., Wood, M. W., & Dormer, K. J. (2004a).
Daphalapurkar, N. P., Dai, C., Gan, R. Z., & Lu, H.
Human middle ear transfer function measured by
(2009). Characterization of the linearly viscoelastic
double laser interferometry system. Otology and
behavior of human tympanic membrane using
Neurotology, 25(4), 423-435.
nanoindentation technique. Journal of the Mechanical
Behavior of Biomedical Materials, 2, 82-92.
Dirckx, J. J. J. & Decraemer, W. F. (1992). Area
change and volume displacement of the human
tympanic membrane under static pressure. Hearing
Research, 62, 99-104.
Gan, R. Z., Feng, B., & Sun, Q. (2004b). Threedimensional finite element modeling of human ear
for sound transmission. Annals of Biomedical
Engineering, 32, 847–859.
Gan, R. Z., Dai, C., & Wood, M. W. (2006a). Laser
interferometry measurements of middle ear fluid
Dirks, D. D. & Morgan, D. E. (2000). Tympanometry
and pressure effects on sound transmission. Journal
and acoustic reflex testing. In: Canalis, R. F.,
of the Acoustical Society of America, 120(6),
Lambert, P.R. (Eds.), The ear: Comprehensive
3799-3810.
Otology (pp. 223-229). Philadelphia, PA: Lipponcott
Williams & Wilkins.
Feeney, M. P., Grant, I. L., & Marryott, L. P. (2003).
Gan, R. Z., Sun, Q., Feng, B., & Wood, M. W.
(2006b). Medical Engineering and Physics, 28,
395-404.
Wideband energy reflectance measurements in
Gan, R. Z. & Wang, X. (2007). Multi-field finite
adults with middle-ear disorders. Journal of Speech
element analysis for sound transmission in otitis
Language and Hearing Research, 46, 901-911.
with effusion. Journal of the Acoustical Society of
Feeney, M. P., Grant, I. L., & Mills, D. M. (2009).
America, 122(6), 3527-3538.
Wideband energy reflectance measurements of ossicular
Gan, R. Z., Reeves, B. P., & Wang, X. (2007).
chain discontinuity and repair in human temporal
Modeling of sound transmission from ear canal to
bone. Ear and Hearing, 30, 391-400.
cochlea. Annals of Biomedical Engineering, 35(12),
Feng, B. & Gan, R. Z. (2004). Lumped parametric model
of human ear for sound transmission. Biomechanics
and Modeling in Mechanobiology, 3, 33-47.
2180-2195.
Gan, R. Z., Cheng, T., Dai, C., Yang, F., & Wood,
M. W. (2009). Finite element modeling of sound
Gan, R. Z., Wood, M. W., Ball, G. R., Dietz, T. G.,
transmission with perforations of tympanic membrane.
& Dormer, K. J. (1997). Implantable hearing
Journal of the Acoustical Society of America,
device performance measured by laser Doppler
126(1), 243-253.
110
AUDIOLOGY·청능재활 2013;9:95-112
Gan, R. Z., Dai, C., Wang, X., Nakmali, D., &
Pediatrics in Review, 31, 102-110.
Wood, M. W. (2010). A totally implantable hearing
Gyo, K., Aritomo, H., & Goode, R. L. (1987).
system–design and function characterization in 3D
Measurement of the ossicular vibration ratio in
computational model and temporal bones. Hearing
human temporal bones by use of a video measuring
Research, 263, 138-144.
system. Acta Oto-Laryngologica, 103, 87-95.
Gan, R. Z., Yang, F., Zhang, X., & Nakmali, D.
Hato, N., Stenfelt, S., & Goode, R. L. (2003). Three-
(2011). Mechanical properties of human stapedial
dimensional stapes footplate motion in human
annular ligament. Medical Engineering and Physics,
temporal bones. Audiology and Neurotology, 8,
33, 330-339.
140-152.
Gan, R. Z., Nakmali, D., & Zhang, X. (2013).
Hudde, H. & Weistenhöfe, C. (1997). A three-
Dynamic properties of round window membrane in
dimensional circuit model of the middle ear. Acta
guinea pig otitis media model. Hearing Research,
Acustica United with Acustica, 83(3), 535-549.
301, 125-136.
Huang, G., Daphalapurkar, N., Gan, R. Z., & Lu, H.
Gan, R. Z., Ji, X., & Zhang, X. (2013). Novel
(2008). Measurement of linear viscoelastic properties
auditory test curves derived from 3D finite element
of human tympanic membrane using nanoindentation.
models of human ear. ARO - Midwinter Meeting,
Journal of Biomechanical Engineering, 130, 014501-
37, San Diego, CA, February 22-26.
1 to -7.
Goode, R. L., Killion, M., Nakamura, K., & Nishihara,
Ji, X. D., Zhang, X., & Gan, R. Z. (2013). Development
S. (1994). New knowledge about the function of the
of auditory test modeling system (ATMS) software
human middle ear: Development of an improved
based on 3D finite element model of human ear.
analog model. American Journal of Otology, 15,
Proceedings from BMES 2013 Annual Meeting,
145-154.
Seattle, WA, September 25-28.
Guan, X. & Gan, R. Z. (2011). Effect of middle ear
Keefe, D. H., Bulen, J. C., Arehart, K. H., & Burns,
fluid on sound transmission and auditory brainstem
E. M. (1993). Ear-canal impedance and reflection
response in guinea pigs. Hearing Research, 277,
coefficient in human infants and adults. Journal of
96-106.
the Acoustical Society of America, 94, 2617-2638.
Guan, X. & Gan, R. Z. (2013). Mechanisms of
Keefe, D. H. & Simmons, J. L. (2003). Energy
tympanic membrane mobility loss in acute otitis media
transmittance predicts conductive hearing loss in
model of guinea pig. Journal of the Association for
older children and adults. Journal of the Acoustical
Research in Otolaryngology, 14, 295-307.
Society of America, 114, 3217-3238.
Guan, X., Li, W., & Gan, R. Z. (2013a). Comparison
Koike, T. & Wada, H. (2002). Modeling of the
of eardrum mobility in acute otitis media and otitis
human middle ear using the finite-element method.
media with effusion models. Otology and Neurotology,
Journal of the Acoustical Society of America,
34(7), 1316-1320.
111(3), 1306-1317.
Guan, X., Chen, Y., & Gan, R. Z. (2013b). Factors
Kringlebotn, M. & Gundersen, T. (1985). Frequency
affecting loss of tympanic membrane mobility in
characteristics of the middle ear. Journal of the
acute otitis media model of chinchillas. Hearing
Acoustical Society of America, 77, 159-164.
Research (Under Review).
Gould, J. M. & Matz, P. S. (2010). Otitis media.
Kringlebotn, M. (1988). Network model for the human
middle ear. Scandinavian Audiology, 17, 75-85.
R Gan : Diagnosis and Characterization of Middle-ear and Cochlear Functions
Larsson, C., Dirckx, J. J. J., Bagger-Sjöbäck, D., &
von Unge, M. (2005). Pars flaccida displacement
pattern in otitis media with effusion in the gerbil.
Otology and Neurotology, 26(3), 337-343.
Lee, C. F., Chen, P. R., Lee, W. J., Chen, J. H., &
Liu, T. C. (2006). Three-dimensional reconstruction
111
ossicular coupling in cat and human. Hearing
Research, 57, 245-268.
Qi, L., Funnell, W. R., & Daniel, S. J. (2008). A
nonlinear finite-element model of the newborn
middle ear. Journal of the Acoustical Society of
America, 124(1), 337-347.
and modeling of middle ear biomechanics by
Qin, Z., Wood, M., & Rosowski, J. J. (2010).
high-resolution computed tomography and finite
Measurement of conductive hearing loss in mice.
element analysis. Laryngoscope, 116(5), 711-716.
Hearing Research, 263, 93-103.
Lee, C-Y. & Rosowski, J. J. (2001). Effects of middle-ear
Ravicz, M. E., Rosowski, J. J., & Merchant, S. N. (2004).
static pressure on pars tensa and pars flaccida of
Mechanisms of hearing loss resulting from middle-ear
gerbil ears. Hearing Research, 153, 146-163.
fluid. Hearing Research, 195(1-2), 103-130.
Luo, H., Dai, C., Gan, R. Z., & Lu, H. (2009a).
Rabinowitz, W. M. (1981). Measurement of the
Measurement of Young’s modulus of human
acoustic input immittance of the human ear.
tympanic membrane at high strain rates. Journal of
Journal of the Acoustical Society of America, 70,
Biomechanical Engineering, 131, 064501-1 to -8.
1025-1035.
Luo, H., Dai, C., Gan, R. Z., & Lu, H. (2009b). A
Rosowski, J. J. & Lee, C-Y. (2002). The effect of
comparison of Young’s modulus for normal and
immobilizing the gerbil’s pars flaccida on the
diseased human eardrums at high strain rates.
middle-ear’s response to static pressure. Hearing
International Journal of Experimental and Computational
Research, 174, 183-195.
Biomechanics, 1(1), 1-22.
Margolis, R. H., Saly, G. L., & Keefe, D. H. (1999).
Wideband reflectance tympanometry in normal adults.
Journal of the Acoustical Society of America, 106,
265-280.
Merchant, S. N., Ravicz, M. E., Puria S, Voss, S. E.,
Whittemore, K. R. Jr., Peake, W. T., & Rosowski,
Stepp, C. E. & Voss, S. E. (2005). Acoustics of the
human middle-ear air space. Journal of the
Acoustical Society of America, 118, 861-871.
Sun, Q., Gan, R. Z., Chang, H-K., & Dormer, K. J.
(2002a). Computer-integrated finite element modeling
of human middle ear. Biomechanical Modeling in
Mechanobiology, 1, 109-122.
J. J. (1997). Analysis of middle ear mechanics and
Sun, Q., Change, K-H., Dormer, K. J., Dyer, R. K.,
application to diseased and reconstructed ears.
& Gan, R. Z. (2002b). An advanced computer
American Journal of Otology, 18(2), 139-154.
-aided geometric modeling and fabrication method
Murakami, S., Guo, K., & Goode, R. L. (1997). Effect
of middle ear pressure change on middle ear
mechanics. Acta Oto-laryngologica, 117, 390-395.
for human middle ear. Medical Engineering and
Physics, 24, 596-606.
Turcanu, D., Dalhoff, E., Muller, M., Zenner, H. P. &
Nuttall, A. L. & Dolan, D. F. (1996). Steady-state
Gummer, A. W. (2009). Accuracy of velocity distortion
sinusoidal velocity responses of the basilar
product otoacoustic emissions for estimating mechanically
membrane in guinea pig. Journal of the Acoustical
based hearing loss. Hearing Research, 251, 17-28.
Society of America, 99, 1556-1565.
Von Unge, M., Decraemer, W. F., Dirckx, J. J. J., &
Peake, W. T., Rosowski, J. J., & Lynch, T. J. III.
Bagger-Sjoback, D. (1995). Shape and displacement
(1992). Middle-ear transmission: Acoustic versus
patterns of the gerbil tympanic membrane in experimental
112
AUDIOLOGY·청능재활 2013;9:95-112
otitis media with effusion. Hearing Research, 82(2),
184-196.
35: 31, San Diego, CA, February 25-29.
Zhang, X. & Gan, R. Z. (2012b). Computational
Voss, S. E. & Allen, J. B. (1994). Measurement of
modeling of otitis media in pediatric ear.
acoustic impedance and reflectance in the human
Proceedings BMES Annual Meeting, Atlanta, GA,
ear canal. Journal of the Acoustical Society of
October 24-27.
America, 95, 372-384.
Zhang, X. & Gan, R. Z. (2013a). Finite element
Voss, S. E., Rosowski, J. J., Merchant, S. N., &
modeling of energy absorbance in normal and
Peake, W. T. (2000). Acoustic responses of the
disordered human ear. Hearing Research, 301, 146-
human ear. Hearing Research, 150, 43-69.
155.
Wang, X., Cheng, T., & Gan, R. Z. (2007). Finite
Zhang, X. & Gan, R. Z. (2013b). Dynamic properties
element analysis of middle ear pressure effects on
of human tympanic membrane based on frequency-
static and dynamic behavior of human ear. Journal
temperature superposition. Annals of Biomedical
of the Acoustical Society of America, 122(2),
Engineering, 41(1), 205-214.
906-917.
Zhang, X. & Gan, R. Z. (2013c). Dynamic properties
Zhang, X. & Gan, R. Z. (2010). Dynamic properties
of human round window membrane in auditory
of human tympanic membrane - Experimental
frequencies. Medical Engineering and Physics,
measurement and modeling analysis. International
35(3), 310-318.
Journal of Experimental and Computational Biomechanics,
1(3), 252-270.
Zhang, X., Guan, X., Nakmali, D., Palan, V., Pineda,
M., & Gan, R. Z. (2013). Imaging the human
Zhang, X. & Gan, R. Z. (2011a). A comprehensive
tympanic membrane motion using scanning laser
model of human ear for analysis of implantable
Doppler vibrometer and finite element model.
hearing devices. IEEE Transations on Biomedical
Journal of the Association for Research in
Engineering, 58(10), 3024-3027.
Otolaryngology (Under Review).
Zhang, X. & Gan, R. Z. (2011b). Experimental
Zhao, F., Koike, T., Wang, J., Sienz, H., & Meredith,
measurement and modeling analysis on mechanical
R. (2009). Finite element analysis of the middle
properties of incudostapedial joint. Biomechanics
ear transfer functions and related pathologies.
Modeling in Mechanobiology, 10, 713-726.
Medical Engineering and Physics, 31(8), 907-916.
Zhang, X. & Gan, R. Z. (2012a). A finite element
Zwislocki, J. (1962). Analysis of the middle ear
model of pediatric ear for sound transmission from
function. Part I. Input impedance. Journal of the
ear canal to middle ear. ARO - Midwinter Meeting,
Acoustical Society of America, 34, 1514-1523.