MASSCHTISES
MASSACHUSETTS INSTIT TE
OF TECHNOLOLGY
B.S., University of Massachusetts Lowell (2010)
Submitted to the Harvard-MIT Health Sciences and Technology
Speech and Hearing Bioscience and Technology in partial fulfillment of the requirements for the degree of
Master of Science in Health Sciences and Technology at the
Feb 2015
Massachusetts Institute of Technology 2015. All rights reserved.
A
Author.........
V V
Harvard-MIT Health Sciences and Technology
Speech and Hearing Bioscience and Technology
14, 2015
Certified
L-/
John J. Rosowski
Professor of Otology and Laryngology, & Health Sciences and
Technology, HMS
Accepted by............
Thesis Supervisor
Emry N. Brown, MD, PhD
Director, Harvard-MIT Program in Health Sciences and Technology
Professor of Computational Neuroscience and Health Sciences and
Technology

2

David Chhan
Submitted to the Harvard-MIT Health Sciences and Technology
Speech and Hearing Bioscience and Technology on January 14, 2015, in partial fulfillment of the requirements for the degree of
Master of Science in Health Sciences and Technology
While much is known about the process of how airborne sound is conducted to the inner-ear via the outer ear and middle ear, so-called air conduction (AC), the mechanisms by which vibrations of the head and body, so-called bone conduction
(BC), produce an auditory response are not well understood. It is clear that the inner ear is the sensory site of auditory stimulation by bone conduction, and that the resultant activation of the inner ear has many features in common with airconduction stimulation; however, bone conduction is known to stimulate the inner ear through multiple pathways. The relative significance and frequency dependence of these different pathways have not been well defined. Our previous work on bone conduction in chinchillas suggested inner-ear mechanisms are the dominant sources in
BC. This thesis builds upon the early work by investigating inner ear mechanisms with stapes fixation and ear canal occlusion. Results of stapes fixation show a decrease in scala vestibuli sound pressure Psv and little change in scala tympani sound pressure
PST in bone conduction. Ear canal occlusion produces an increase in ear canal sound pressure PEC with a similar amount of increase in PsV, but almost no change in
PST.
We attributed the differences in the change between PsV and PST in bone conduction after these manipulations to the existence of compressible cochlear structures or third window pathways, e.g. the cochlear aqueduct. While ear canal compression and middle ear inertia sources may contribute to the total bone conduction response (a
10 dB decrease in Psv after middle ear interruption and stapes fixation, and a 10 dB increase after ear canal occlusion), inner ear mechanisms are still the most significant sources in bone conduction because the changes in Psv and PST in BC are much smaller than the changes in AC.
Thesis Supervisor: John J. Rosowski
Title: Professor of Otology and Laryngology, & Health Sciences and Technology, HMS
3

4

Foremost, I would like to express my utmost thanks and deepest gratitude to my advisor John Rosowski for his support and guidance throughout this research project.
His mentorship is beyond the work presented here and his dedication to teaching and caring for his students deserves much recognition. John, when most people would have already given up on me, you continued to find something in me to believe in.
For that, I am forever grateful.
I want to acknowledge the people in the Middle Ear group and Eaton-Peabody
Laboratory at the Massachusetts Eye and Ear Infirmary for their support. I cannot thank Melissa enough for her help in pressure transducer assembly, animal surgery and experimental preparation. I am also grateful to Heidi and Mike for their insightful comments and advice on my work and teaching me how to fabricate pressure transducers. My lab colleagues Tao, Jeremie and Gabby are always there when needed. I want to also thank NIH and NIDCD for the financial support throughout my tenure in the program.
Lastly, I want to thank my wife, my parents, brothers and sister for their unconditional love and support. They are great sources of inspiration and encouragement.
5

6

1 Introduction 13
1.1 Anatomy of the Peripheral Auditory System . . . . . . . . . . . . . . 14
1.2 The Conduction of Airborne Sound . . . . . . . . . . . . . . . . . . . 16
1.3 The Conduction of Bone Vibrations . . . . . . . . . . . . . . . . . . . 17
1.4 Vibration Modes, Pathways and Components in Bone Conduction .
.
18
1.4.1 Ear-Canal Cartilaginous Wall Compression . . . . . . . . . . . 19
1.4.2 Middle-Ear Ossicular Inertia . . . . . . . . . . . . . . . . . . . 19
1.4.3 Inner-Ear Mechanisms: Cochlear Fluid Inertia, Cochlear Bone
Compression and Third-Window Pathways . . . . . . . . . . . 19
1.5 Aims of This Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2 Materials and Methods 25
2.1 Fiber-Optic Pressure Sensors . . . . . . . . . . . . . . . . . . . . . . . 26
2.1.1 Sensor Tip with Gold-Coated Diaphragm Construction . .
. .
26
2.1.2 Fiber-Optic Pressure Sensor Probe Assembly . . . . . . . . . . 27
2.1.3 Calibration: A Test of Sensitivity and Stability . . . . . . . .
27
2.2 Animal Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.3 Acoustic and Vibration Stimuli . . . . . . . . . . . . . . . . . . . . . 31
2.4 Experimental Procedure . . . . . . . . . . . . . . . . . . . . . . . . . 31
3 Results 33
3.1 A Potential Artifactual Sound Pressure in Bone Conduction Stimulation 34
3.2 Handling of the Pressure Sensor Sensitivity and Stability . . . . . . . 38
7

3.3 Intracochlear Sound Pressure Measurements in AC and BC . . . . . . 39
3.4 Effects of Scala Vestibuli Hole . . . . . . . . . . . . . . . . . . . . . . 43
3.5 Effects of Scala Tympani Hole . . . . . . . . . . . . . . . . . . . . . . 43
3.6 Frequency Response of the AC and BC Stimulus . . . . . . . . . . . . 48
3.7 Effects of Ear Canal Occlusion . . . . . . . . . . . . . . . . . . . . . . 52
3.8 Effects of Stapes Fixation in AC and BC . . . . . . . . . . . . . . . . 55
4 Discussion
4.1 Were PST Measurements Influenced by Noise or Artifacts, Thus Lim-
61 iting Its Change After the Manipulations? . . . . . . . . . . . . . . . 64
4.2 Can a Third Window Like the Cochlear Aqueduct Explain the Differences In the Change between Psv and PST in Bone Conduction After the M anipulations? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
5 Conclusions 71
8

1-1 Anatomy of the peripheral auditory system (skidmore.edu) . . . . . . 14
1-2 Schematic of the cochlea (en.wikipedia.org). . . . . . . . . . . . . . . 16
1-3 Schematic of the compressional BC mechanism after Bekesy
. . .
20
2-1 Schematic of the fiber-optic pressure sensor assembly. . . . . . . . . . 27
2-2 Schematic of a calibration setup. . . . . . . . . . . . . . . . . . . . .
28
2-3 Frequency dependence and sensitivity of a representative sensor before and after an experiment. . . . . . . . . . . . . . . . . . . . . . . . . . 29
2-4 Schematic of the animal preparation and experimental setup. . . . . . 30
3-1 Scala vestibuli sound pressure Psv with hole unsealed (Hole), sealed with .Jeltrate (JT), sealed with dental cement (DC). Measurements are from Experiment CH20_130702. . . . . . . . . . . . . . . . . . . . . . 36
3-2 Sensor fiber velocity (Vfiber) with hole unsealed (Hole), sealed with Jeltrate (JT), sealed with dental cement (DC) and velocity of the cochlear bone (Vbone). Measurements are from Experiment CH20_130702. .
.
37
3-3 Intracochlear Sound Pressures Psv, PST and AP with AC stimulation
(Experiment DC9.141106). The measured pressures are normalized by ear canal sound pressure. . . . . . . . . . . . . . . . . . . . . . . . . . 41
3-4 Intracochlear Sound Pressures Psv, PsT and AP with BC stimulation
(Experiment DC9_141106). The magnitude of the measured pressures are normalized by the stimulus voltage to the BAHA. The phase angle of the measured pressures are normalized by the phase angle of the velocity of the petrous bone evoked by the BAHA . . . . . . . . . . . 42
9

3-5 Measurements of Psv with AC stimulation made before and after sealing the hole around the transducer with Jeltrate and Dental Cement
(Experiment DC9_141106). . . . . . . . . . . . . . . . . . . . . . . . . 44
3-6 Measurements of Psv with BC stimulation made before and after sealing the hole around the transducer with Jeltrate and Dental Cement
(Experiment DC9_141106). . . . . . . . . . . . . . . . . . . . . . . . . 45
3-7 Measurements of PST with AC stimulation made before and after sealing the hole around the transducer with Jeltrate and Dental Cement
(Experiment DC9_141106). . . . . . . . . . . . . . . . . . . . . . . . . 46
3-8 Measurements of PST with BC stimulation made before and after sealing the hole around the transducer with Jeltrate and Dental Cement
(Experiment DC9_141106). . . . . . . . . . . . . . . . . . . . . . . . . 47
3-9 Ear canal sound pressure PEC generated by air conduction stimulation
(Experiment DC8-141023). Shown in gray is the range of responses to the 20 dB stimulus range used for measurements. . . . . . . . . . . .
49
3-10 Cochlear bone velocity generated by bone conduction stimulation (Experiment DC8_141023). Shown in gray is the range of responses to the
20 dB stimulus range used for measurements. Note that the phase is relative to the stimulus voltage. . . . . . . . . . . . . . . . . . . . . . 50
3-11 Ear canal sound pressure PEC generated by bone conduction stimulation in an occluded ear canal (Experiment DC8_141023). Shown in gray is the range of response to the 20 dB stimulus range used for measurements. Note that the phase is relative to the stimulus voltage. 51
3-12 Magnitude of ear canal sound pressure I PEC I, intracochlear sound pressures JPsv and |PSTI and differential sound pressure |API before and after the occlusion in one ear (Experiment DC5_140821). . . . . . . .
53
3-13 The change in the magnitude of ear canal sound pressure IPEC I, intracochlear sound pressures IPsVy and IPSTI and differential sound pressure JAP after the occlusion in four ears and the average. . . . . . .
54
10

3-14 Magnitude of ear canal sound pressure PEC I, intracochlear sound pressures
and IPSTI and differential sound pressure JAPI before and after stapes fixation with AC stimulation in one ear (Experiment DC1_140710). 56
intracochlear sound pressures JPSVI and
IPST and differential sound pressure
before and after stapes fixation with BC stimulation in one ear (Experiment DC1-140710). 57
3-16 The change in magnitude of ear canal sound pressure
IPEC
, intracochlear sound pressures
and
and differential sound pressure IAPI after stapes fixation with AC stimulation in six ears and the average. ........ .................................. 58
3-17 The change in magnitude of ear canal sound pressure PEC , intracochlear sound pressures Psv
and
I and differential sound pressure IAPI after stapes fixation with BC stimulation in six ears and the average. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4-1 Ratio of the magnitude of scala tympani sound pressure PST I to scala vestibuli sound pressure
in AC and BC in normal ears. . . . . . 63
4-2 Top: Magnitude of scala tympani sound pressure
when the sensor hole unsealed (Hole), sealed with Jeltrate (JT) and with dental cement
(DC). Bottom: Magnitude of fiber velocity (Vfiber) when hole sealed with Jeltrate (JT) and dental cement (DC) and cochlear bone velocity
(Vbone). Measurements are from Experiment DC5_140821. . . . . .
65
4-3 Top: IPsvI and
normalized by
PEC before and after partial round window fixation and IS-joint interruption in air conduction. Bottom:
and JPST before and after partial round window fixation and
IS-joint interruption in bone conduction (Experiment DC9_141106). .
67
4-4 Effects of middle ear interruption on scala vestibuli sound pressure Psv and scala tympani sound pressure PST in AC and BC (figure from [4]). 68
11

12

This thesis investigates the mechanisms that contribute to responses of the chinchilla's middle ear and inner ear when their skull bones are subjected to vibratory stimulation: so-called "bone conduction". Of particular interest is the determination of the relative importance of the different mechanisms of bone conduction. There are a number of mechanisms that have been hypothesized based on clinical observations and experimental work in human cadevers as well as animals. However, there is no consensus of what the dominant mechanisms are in bone conduction among past and present researchers and investigators working on the subject. While others pointed out the importance of the middle ear in bone conduction processes, previous work on chinchilla in our lab suggested the inner-ear mechanisms play a dominant role [22]
[3]. This thesis aims at further testing our hypothesis of the dominance of inner-ear mechanisms in bone conduction. In this section, we'll discuss the anatomy and function of the different components of the peripheral auditory system that are involved in the process of air conduction and bone conduction. We'll also briefly describe the mechanisms of air conduction versus bone conduction before introducing the specific aims of the study.
13

The major components of the human peripheral auditory system are schematized in
Figure 1-1. It consists of the outer, middle and inner ear.
Pinna
/1
$En*stac er
Figure 1-1: Anatomy of the peripheral auditory system (skidmore.edu).
The outer ear is subdivided into a pinna and external auditory canal and is terminated by the tympanic membrane, or the eardrum. The pinna or the ear flap is a cartilaginous structure that serves as a sound collector directed to the ear canal. It also plays a role in spectral transformations that assist in sound localization in the vertical plane [11]. The ear canal is a 2-3 mm long tube-like structure whose outer
1/3 is supported by cartilage and the inner 2/3 is embedded in the temporal bone.
The external ear canal can be considered as a passive sound transformer
[11] in which sound pressure measured near the surface of the tympanic membrane is higher than at the entrance of the ear canal. The external sound pressure gain can be as large as
20 dB near 4 kHz [17].
The ear canal is terminated by a cone-shaped membrane called the tympanic membrane (TM). The TM functions as an acoustical-to-mechanical transformer that converts airborne sound energy in the ear canal to mechanical vibration of the middle-
14

ear ossicles. Anatomically, the TM is attached to the manubrium (the handle) of the malleus at the umbo (at the center of the TM) and near the TM rim. The attachment of the TM to the manubrium appears to draw the center of the TM into the tympanic cavity. In some mammals, including the chinchilla, the middle-ear or tympanic cavity is extended by a thin-walled bony capsule called the "bulla".
In this air-filled cavity are the three tiny middle-ear bones: malleus, incus and stapes, whose Latin names mean hammer, anvil, and stirrup respectively. The ossicles are coupled together and loosely supported by ligaments and tendons. There are three ligamentous joints, the incudo-mallear joint (between the head of the malleus and the body of the incus), the incudo-stapedial joint (between the lenticular process of the incus and the head of the stapes) and the stapedio-vestibular joint (the annular ligament that couples the stapes to the oval window of the inner ear). Sound pressure in the ear canal sets the TM into motion that is transmitted to the middleear ossicles. In a simplistic view, the malleus and incus rotate as rigid bodies about an axis defined by supporting ligaments; this motion produces in-and-out piston-like motion of the stapes within the oval window to the inner ear. The three middle-ear ossicles together act as a mechanical lever that helps match the low impedance of the air at the external ear canal to the high impedance of the inner ear, which is filled with water-like fluid bounded by bone and soft tissues. This impedance matching is greatly assisted by the 20 times difference in the area of the TM and the stapes footplate, where this area difference produces a hydraulic lever that trades a drop in sound-induced volume displacement between the TM and the footplate for an increase in sound pressure.
Movement of the middle-ear ossicles can be reduced by two muscles that act to stiffen the middle-ear mechanism: the tensor tympani (attached to the manubrium of the malleus) and the stapedius (attached to the head of the stapes). These muscles contract in response to loud sounds that can excessively vibrate the middle ear, thus reducing the response transmitted to the inner ear. This protective function is called the acoustic reflex.
The last link of the ossicular chain, the stapes, is connected to the cochlea (the
15

auditory portion of the inner ear) at the oval window by an annular ligament. The oval window is an entrance to the spiral-shaped cochlea in which three lymph-filled chambers surround the sensory structures that transform sound pressures within the lymphs to sensory potentials and action potentials on the auditory nerves. The sensory components are housed in the organ of corti located in the scala media that is sandwiched between scala vestibuli and scala tympani. The oval window is placed in the basal end of scala vestibuli. Scala vestibuli and scala tympani are linked at the helicotrema in the apex of the cochlea. The basal end of scala tympani is terminated
by the round window which provides a low-impedance pressure release that permits the back and forth motion of the scalae flulids when sound causes motion of the stapes
(Fig. 1-2). The horizontal dashed line of Figure 1-2 indicates a fluid path between the two chambers.
Oval
Window
Round
Window
Helicotrema
Nerve
Organ of Corti
Cochlear
Partition
Fluid path
Figure 1-2: Schematic of the cochlea (en.wikipedia.org).
"Air conduction" refers to the conduction of air-borne sound in the ear canal through the middle ear to the inner ear. This is the most efficient way and common mode for sound energy to reach the inner ear. The process starts with sound pressure in the
16

ear canal. The associated compression and rarefaction generate motion of the TM that sets the three middle-ear ossicles into vibration. Vibration of the stapes causes motion of the fluids in the cochlea. Fluid stimulation results in a pressure wave that travels from the oval window, where the stapes is attached to, to the helicotrema, an opening at the apex of the cochlea and out through the highly compliant round window. These pressure waves produce a pressure distribution in both scalae and throughout the cochlea. The pressure distribution in the cochlea can be described in terms of longitudinal and transverse components. Traveling at the speed of sound in fluid, the compressional longitudinal pressure component propagates nearly instantaneously from the oval window to the round window. The transverse component of the intracochlear pressure arises from the differential pressure between the scala vestibule and scala tympani across the cochlear partition (cochlear partition and basilar membrane are often used interchangeably when their mechanical behavior is concerned because the mechanical property of the former is dominated by that of the later).
The transverse differential pressure initiates a vibration of the basilar membrane and sets up a traveling wave from base to apex. The unique property of the basilar membrane whose stiffness decreases from base to apex leads to stimulus-frequency dependent variations in the location of maximal vibration on the basilar membrane.
Maximum vibration occurs at the base of the partition with high frequency stimulation and at the apex with low frequency stimulation. As a result, the cochlea is often described as a frequency analyzer that codes high-frequency sound nearer to the base and low-frequency sound closer to the apex. The mechanical vibration of the basilar membrane stimulates the hair cells in the organ of corti, which produce sensory potentials that evoke action potentials in the dendrites of the auditory nerves.
The electrical action impulses are then transmitted to the brain.
In the audiology clinic, bone conduction (BC) describes a collection of mechanisms that produce a hearing sensation in response to vibration of the bones of the skull.
17

Early investigations of the sound-conduction processes in bone conduction were largely driven by clinical application. Use of BC hearing in clinical practice dates back to the 19th century when Weber and Rhinne developed hearing tests that employed a tuning fork to separate conductive (middle ear based) and sensorineural (inner ear based) hearing losses [14]. Such a separation depended on the hypothesis that the conduction of skull vibrations to the inner ear was independent of the middle ear.
Patients who showed a decrease in auditory sensitivity to a tuning fork held in air near the ear canal (an AC stimulus), but had normal sensitivity to the tuning fork coupled to the bone near the ear (a BC stimulus) had a conductive hearing loss; while patients with sensory neural hearing loss showed an equal decrease in sensitivity to the tuning fork held in air and coupled to the skull.
Despite the early use of bone conduction as a tool in hearing test, its mechanisms were not investigated in any depth until the work of Birany [1], Wever and Lawrence
[30], Bek6sy [21, Huizing [10], and Tonndorf [29]. Their clinical observations and experimental results have established the basis of current theories of how bone vibration is perceived as sound.
According to Wever and Lawrence [30], there are three major modes that contribute to bone conduction hearing. The modes correspond to the components of the auditory periphery and pathways in which bone vibration transfers to fluid motion in the cochlea. In brief, the first mode relates to sound energy that radiates into the ear canal through compression of the cartilaginous ear-canal wall. The second is a translatory mode that causes ossicular motion due to differences between the inertia of the skull and the loosely-coupled middle ear ossicles. The third relates to the inner-ear based mechanisms in which motion of the cochlear fluids results from the compression and expansion as well as back-and-forth shaking of the bony cochlear capsule.
18

1.4.1 Ear-Canal Cartilaginous Wall Compression
Vibration of the skull and attached jaw produces vibrations of the cartilaginous ear canal that rarify and compress the air within the canal. The resultant sound energy induced in the ear canal produces motion of the tympanic membrane and the coupled ossicles that stimulate the inner ear in a manner similar to AC sound. The ear-canal sound pressures and sensations produced by this mechanism increase significantly when the normally open lateral end of the cartilaginous canal is occluded. This observation leads to naming this pathway 'the occlusion effect'
The translatory motion of the head induces relative motion between the head and the coupled ossicles. These relative motions result because of differences in the inertia of these structures, together with the elastic coupling of the ligaments that support the ossicles in the skull. From a mechanical standpoint, the middle-ear ossicles act as free masses, which are loosely attached to the head by ligaments, tendons and the TM which act as mechanical springs. At low frequency, below the middle-ear resonance frequency of 1-2 kHz, the middle-ear ossicles vibrate in phase with the bones surrounding the middle ear. At higher frequencies; however, vibration of the bone and the middle-ear ossicles are no longer in phase thus causing a relative motion that stimulates the fluids in the cochlea. As pointed out by Biriny [1] this stimulation depends on the direction of vibration, where the anatomy of the ear is most sensitive to side-to-side shaking of the head.
* Cochlear fluid inertia: Similar to the middle-ear ossicular inertia mechanism, the nearly-incompressible inner-ear fluids can move in response to vibrations produced by BC stimulation. The relative motion of the fluid bounded by the two compliant windows within the moving surrounding bone can stimulate the
19

inner ear in a manner similar to the motion of the stapes in AC.
* Cochlear bone compression: The bony cochlear capsule may undergo compression and expansion when the skull bones vibrate. Figure 1-3 illustrates this mechanism at work. In the figure, the cochlea is uncoiled and viewed as a rectangular chamber constrained by the bony wall and the flexible oval window (OW) and round window (RW) to the left. The chamber is split in two
by the cochlear partition which separates scala vestibuli (SV) on the top and scala tympani (ST) on the bottom (Fig. 1-3 A). The compressions and rarefactions of the bone around the cochlea induce volume changes in the inner ear spaces. These volume changes in the presence of the flexible cochlear windows produce motion of the nearly incompressible cochlear fluids. Differences in the impedance of the two flexible cochlear windows produce a differential fluid flow in the cochlea that can set up traveling wave patterns just as when the cochlea is stimulated by air conducted sound (Fig. 1-3 B). Differences in motion of the two scalae can be enhanced by asymmetries in inner-ear structure and by compressible structures within the inner ear (Fig. 1-3 C).
A B C
Figure 1-3: Schematic of the compressional BC mechanism after B6kesy [2].
* Third-window pathways: Third-window pathways involve the induction of motion within cochlear fluid via motion of cranial soft tissues and fluids coupled to the inner ear via neural, lymph or vascular channels in the bone and cerebro-spinal fluid [7]. This mechanism has features similar to cochlear compression in that it allows a differential flow of sound-induced volume velocity in
20

and out of the inner ear which essentially changes the volume of the cochlear fluid. The difference is that the dynamic increases and decreases in fluid volume are produced by vibration sources outside of the bony inner ear rather than compression of the surrounding bone.
The three vibratory modes with their involved components and pathways coexist and independently contribute to bone-conducted sound. Due to the complex interaction of these multiple pathways, the relative contributions of each of the pathways to bone-conducted sound are still poorly understood. B6arny
and Huizing [10] stressed the importance of middle-ear inertia as the dominant mode in bone conduction while Tonndorf [28] [29] argued that the compressional mode is more significant.
More recently, Stenfelt and coworkers [25] have argued for the primacy of inner-ear fluid inertia. One point of agreement among these authors is their admission of the complexity of the effects produced by bone conduction stimulation.
Recent work on bone conduction in both clinical and experimental settings suggests the contribution of the different stimulus pathways are highly frequency dependent. Stenfelt and coworkers suggest middle-ear and inner-ear fluid inertia contribute to the perception of bone-conducted sounds between the frequencies of 1.5 and 3.5 kHz [26]. They compared ossicle displacements at hearing thresholds for both AC and
BC and showed that between 1.5 and 3.5 kHz the motions in both cases are within
5 dB indicating middle-ear ossicular motions contribute to BC perception in that frequency range. They also found a decrease in stapes velocity after incudo-stapedial joint interruption in BC stimulation [23].
Evidence consistent with a significant cochlear compression mode was presented
by Songer and Rosowski who studied superior canal dehiscence syndrome (SCD) [22].
Patients with SCD [13] present not only with vestibular symptoms, but also auditory symptoms that include improved sensitivity to bone conducted stimuli along with decreased sensitivity to air conducted sound. Rosowski and coworkers suggest a model of SCD effects that includes inner ear compression at frequencies below 1 kHz. Such a model is consistent with their animal studies on the effects of SCD syndrome in both
AC and BC sound [22]. Their model suggests SCD acts like a third window shunting
21

acoustic energy from the ear to the braincase thus reducing hearing sensitivity to air conducted sound. Such shunting allows relative motions of the fluids in scala vestibuli and tympani, and is thought to increase relative fluid flow in bone conduction and improve hearing sensitivity to BC sound. Our previous work on bone conduction in chinchillas using intracochlear sound pressure measurements also pointed to the inner-ear mechanisms being the dominant sources in bone conduction [3]. We found the intracochlear sound pressures in both scalae in response to bone conduction were less affected by the middle-ear osscilar interruption compared to a 40-50 dB reduction of the intracochlear sound pressures when the ear was stimulated though air conduction. The unequal change of the scala vestibuli sound pressure and scala tympani sound pressure after ossicular interruption in bone conduction provided additional evidence that BC ear canal compression and middle ear inertia are not the primary sources that drive the fluid in the cochlea. These results pointed to the significant contribution of the inner ear sources in bone conduction. We will further test this hypothesis by introducing different manipulations of the middle ear and inner ear: fixing the stapes and round window.
Unlike air conduction, where one distinct sound pathway has been identified, the multiple conduction pathways in bone conduction and their interactions are poorly understood. In particular, the frequency range and the relative importance of each pathway to the total BC sensation are not well described. The overall goal of this project is to improve our understanding of the contribution of the different pathways.
We will use anesthetized chinchillas to further test our hypothesis of the inner-ear mechanisms as being the dominant sources. The animal model allows us to measure acoustical and mechanical responses of the cochlea to both air- and bone-conduction stimulation in live preparations. In addition, the measurements are made in a whole animal preparation instead of the extracted cadaver temporal bone. Specifically, we will quantify the frequency range and significance of the inner-ear mechanisms to
22

bone-conducted sound by comparing measurements made before and after fixation of the oval and round windows. Our previous work showed that interrupting the ossicular chain had little effect on the inner-ear responses (a 10 dB drop in scala vestibular sound pressure Psv and little change in scala tympani sound pressure
PST) to bone-conducted sound compared to the 40 to 50 dB decrease in these pressures when the ear is stimulated with AC. The large reduction in the response to AC sound suggests any contributions of the outer ear and middle ear sources in BC should also be reduced by ossicular interruption. The small changes in the sensitivity to
BC stimuli that we did observe suggest the inner-ear based mechanisms in BC are significant. We further hypothesized the 10 dB decrease in Psv in BC after middleear interruption was associated with an interrupt induced decrease in reverse middleear input impedance (the load at the oval window that would oppose the action of inner-ear BC sources), while the almost-no-change in PST was consistent with the impedance of the round window being unaffected by the manipulation. If our hypotheses regarding the dominance of inner-ear mechanisms and the contribution of alterations on the load at the cochlear windows are correct, fixation of the stapes or oval window should produce an increase in Psv in BC stimulation.
23

24

In this thesis work, as in our previous work [3] [4], we used chinchillas as an animal model to study the mechanisms of bone conduction. Chinchillas have relatively large middle-ear airspaces that allow easy access to the middle ear and cochlear windows.
They have a hearing range and sensitivity similar to humans and the size of their middle-ear structures are also human-like. In addition, we are well versed in the surgical techniques associated with measuring mechanical and acoustical responses produced by AC and BC stimulation in this animal.
Our approach is different from past experimental studies of bone conduction: we employed acoustical measurements in a live-anesthetized animal while manipulating the structures of the middle ear and inner ear. Stenfelt and coworkers [23] [24] used mechanical measurements (velocity measurements) in cadaveric humans, while Tonndorf [29] used physiological measurements that allowed estimation of the size of different stimulus components but did not provide adequate measurements of mechanism. Another difference in approach is our use of a Bone Anchored Hearing Aid
(Cochlear BAHA) implanted on the animal skull as a vibrator. This device permits better stimulus control and reproducibility between animals. Stenfelt used a shaker that vibrates the whole cadaveric temporal bone in one particular direction favoring one mode of vibration over others. We believe a vibration source applied to the whole skull will stimulate all of the potential stimulus modes that exist in clinical bone conduction.
25

In this chapter, we describe measurement tools and techniques necessary to our experiments.
With their small size (150 pm diameter), high sensitivity (2.8 mnV/Pa) and broad frequency bandwidth (100 Hz to 50 kHz), the in-house built fiber-optic pressure sensors provided adequate measurements of the pressure in the cochlea (scala vestibuli and scala tympani). The need for such a small size hydrophone is driven by the limited space and delicacy of the chinchilla cochlea [21]. The fiber-optic pressure sensors were designed, fabricated and calibrated as described by Olson [15]. There are three stages to the process:
1. Construction of the sensor tip and reflective gold-coated diaphragms
2. Assembly of the fiber-optic pressure sensor probes
3. Calibration of the probes to test its sensitivity and stability
The sensor tips are composed of 1-2 cm long hollow core glass capillary tubes that are closed at one end by gold-coated diaphragms. A diaphragm is made from a UV-cured thin film polymer approximately 0.5 pm in thickness. A tiny drop of Norland Optical
Adhesive 68 (NOA68) is placed on a still surface of deionized water, and cured by exposure to UV light. The cured adhesive film is affixed to one end of the capillary tube. The outer surface of the diaphragm is coated with a thin (450-500 pm) film of gold. The gold deposition is carried out at the Microsystem Technology Laboratories at MIT with its electron-beam evaporator: a high vacuum chamber capable of thinfilm deposition through vaporization of the depositing materials. The addition of the gold makes the diaphragm highly reflective.
26

2.1.2 Fiber-Optic Pressure Sensor Probe Assembly
A single optical fiber (outer diameter of 120 pm) is inserted into the open end of the sensor tip till there is about a 50 pm gap between the fiber end and the inner surface of the gold diaphragm. The fiber and the glass capillary tube are then glued together with UV-cured Norland adhesive. The optical fiber is spliced to a 'Y' coupling.
One branch of the coupler is attached to a light-emitting diode (LED) that produces incoherent light, and the other branch is attached to a photodiode that measures the light reflected from the diaphragm (Fig. 2-1). The diaphragm flexes in response to sound and modulates the light reflected back down the fiber.
old deposition air gap adhesive diaphragm hollow capillary tube
single optical fiber
Figure 2-1: Schematic of the fiber-optic pressure sensor assembly.
Calibration: A Test of Sensitivity and Stability
The completed fiber-optic miniature microphone is calibrated in a vibrating water column using the method described by Schloss and Strasberg [19]. The pressure
27

sensor is immersed in a vibrating column of liquid sitting on a shaker head and accelerometer (Briiel and Kjwr 429) (Fig. 2-2). The pressure at the diaphragm is related to acceleration of the shaker a and immersion depth h given by Equation 2.1
where p is the density of the water.
P = pha
(2.1)
-hkeinput
Figure 2-2: Schematic of a calibration setup.
Unfortunately, the sensors are fragile and prone to sensitivity change, thus careful handling is necessary. Preliminary tests of the stability are performed via repeated calibrations. During the experiment, calibration of the sensor sensitivity is performed repeatedly before and after each series of measurements to ensure that the sensitivity does not change significantly. Only measurements from stable sensors are used in our data analysis and reported here. The frequency dependence and sensitivity of a representative sensor are plotted in Figure 2-3. The solid and dashed curves represent the sensitivity of the sensor before and after the experiment. The frequency response of the sensor is generally flat up to 10 kHz. The determination of the sensor sensitivity is usually done during the first series of in-the-ear measurements where the pre and post calibration could be obtained. In later stages where sealing of the holes and minimizing the relative motion between the sensor and the bone by dental cement application are necessary, post-measurement calibration is not always possi-
28

ble. Therefore, we relied on the information about the sensor in the first series of measurements and over-time stability measurements during the later stage.
6-
'a
U
4-
3
2-
1 -
0-
10
-130 -
-
.140
FS
-160
1u
I
I
100
-
Bfr
...-After
101
100
Frequency (kHz)
101
Figure 2-3: Frequency dependence and sensitivity of a representative sensor before and after an experiment.
All animal procedures were approved by the Animal Care and Use Committee of the
Massachusetts Eye and Ear Infirmary and MIT. The animals were anesthetized with
Pentobarbital (50 mg/kg) and Ketamine (40 mg/kg), and their tracheas cannulated.
The superior and posterior middle-ear cavities were opened. The tendon of the tensor tympani muscle was cut and the tympanic-portion of the VIIth nerve was sectioned to immobilize the stapedius muscle to prevent random contraction that can affect results [18]. The bony wall lateral and posterior to the round window (RW) was removed to visualize the surface of the round window and stapes footplate. Reflective beads were applied on the stapes crus, footplate and surrounding bones for velocity measurements. A laser Doppler vibrometer (LDV) was used to measure motion of
29

the stapes crus or footplate Vs, cochlear bone velocity Vbone and pressure sensor transducer fiber velocity Vfiber. Sensitivity of the LDV was set at 2mm/s/V. The external ear was left intact, but a probe-tube microphone was sealed in the lateral end of the intact cartilaginous and bony ear canal near the tympanic membrane to measure ear canal sound pressure
PEC within 1 to 2 mm of the umbo of the malleus
(Fig. 2-4). To measure scala vestibuli pressure Psv, a small hole (less than 250 pm in diameter) was made in the vestibule adjacent to the saccule after removal and retraction of a portion of the cerebellum posterior and medial to the vestibule. A hole made in the cochlear base near the round window was used to place the scala tympani pressure sensor Psr. Both Psv and
PsT measurements were made with our in-house built micro-optical pressure sensors.
INTACT SEALED
PINNA & SUPERIOR
EAR CANAL BULLA HOLE
FACIAL
ACCANAL
SOUND SOURCE
OVAL
PROBE
MICROPHONE
FOR
PEC
10 MM
TMMPANE
IN
SCALA
VESTIBULI
HOLE
WPRESSURE
P
FIBER-OPTIC
SENSORS
PST
FIBER-OPTIC
PRESSURE SENSORS
OPEN
POSTERIOR
BULLA
HOLE
Figure 2-4: Schematic of the animal preparation and experimental setup.
30

Air conducted sound was delivered to the ear canal via an ER-3A insert earphone (Etymotic) while bone conduction stimulation was delivered by a BAHA, Bone-Anchored
Hearing Aid (Cochlea BAHA BP100) sound processor attached to a titanium screw implanted on the animal's skull near the vertex. The use of the BAHA stimulator allowed us to move the animal's head during experiments without affecting the reproducibility of our stimuli. The frequency response of the BAHA is bandpass in shape from 200 Hz to 8 kHz where the maximum peak is around 800 Hz. Since the chinchilla's skull is thin, the standard 3 mm titanium screw was inserted only 1 mm deep into the skull and dental cement was used as a foundation and space filler to hold the screw in place. The BAHA was in its 'direct audio input' mode where electrical stimuli are presented directly to the vibration processor and the microphone input to the processor is muted. The BAHA was programmed for 'linear' operation over the range of stimulus levels we used. The stimulus voltage input to either the ER-3A or the
BAHA are generated via a National Instruments I/O board controlled by a custombuilt LabView program. A TDT-programmable attenuator was used to control the stimulus levels. The measurement protocol called for simultaneous measurements of
PEC, Vs or Vbone or Vfiber, PSV, and PST during repeated presentations of stepped pure tones of 100 Hz to 10 kHz at 24 points/octave.
Measurements of PEC, PSV, and PST were made with tone sequences in both air conduction and bone conduction before and after ear canal occlusion and manipulation of the middle ear and inner ear. Measurements made at a variety of stimulus levels
(over a 20 dB range) were consistent with linear responses. Once we established a good series of stable pressure sensor measurements in the normal condition, we proceeded to apply acrylic cement on either the posterior stapes crus, posterior part of the footplate or the round window. After stapes fixation and round window fixation,
31

measurements of fEC, Sv, and sT were repeated in both AC and BC. The aiimal remained anesthetized throughout the whole procedure and was euthanized after completion of the experiment.
32

In this chapter, we show results of intracochlear sound pressure measurements in air conduction as well as bone conduction before and after ear canal and middle ear manipulations have taken place. Manipulations consisted of opening and occluding the ear canal using the foam plug of the ER-3A insert earphone, fixing the stapes crus and footplate with dental cement. We hypothesize these manipulations will significantly affect certain sound conduction pathway or alter the impedances of certain sources thus allowing us to compare the relative contribution of the different bone conduction sources.
For the most part, data reported in this thesis work only includes those measurements in which the sensitivity of the sensor varied by less than 3 dB between preand post-measurement calibrations. In some cases, adjustments to the sensitivity of the sensors were made based on the information of an early set of stable intracochlear sound pressure measurements. Pre-measurement calibration was also used for measurements in which the sensors are glued to the bones by dental cement and unrecovered at the end of the experiment. In addition, only data gathered with a signal-to-noise ratio of 10 dB or higher are included in the analysis (signal to noise was estimated by observation of the magnitude of off-stimulus frequency components determined when we computed Fourier transforms of the responses to sinusoidal stimulation). For comparison with other measurements taken in different ears, our AC data were normalized by ear canal sound pressure PEC. BC magnitude data were nor-
33

malized by the voltage used to drive the BAHA, while their phases were normalized
by the phase of the cochlear bone velocity.
Intracochlear sound pressure measurements are not only central to our work as a measurement tool but are also critical to our interpretation of the results. Therefore it is important to take a closer look at the way the intracochlear sound pressures were measured in particular with bone conduction stimulation.
The intracochlear sound pressures measured during calibration with our in housebuilt fiber optic sensors (fabrication and assembly process described in Chapter 2) are calculated from the formula given in Eq. 2.1. It is a function of the acceleration, immersion depth and fluid density. The acceleration is the relative acceleration of the moving fluid (motion that we want to sense) against a fixed stationary sensor fiber. When the ear is stimulated with air conduction, the cochlear bone and the sensors which are held fixed are assumed to be stationary while the cochlear fluid is stimulated by motion of the ossicular chain. The recorded pressures are related to the interaction of the fluid set in motion by the ossicles and the impedance of the fluid path as the pressure sensors are held immobile within the immobile cochlear bone. However in bone conduction, the cochlear bone is no longer stationary when it is stimulated with a bone vibrator. While the bone is moving, holding the sensors fixed, relative accelerations between the bone and the sensors may produce unwanted artifactual pressures that complicate our estimates of the pressures due to the normal bone conduction mechanisms. To test the idea of how relative motions between the cochlear bone and the sensor affact our measurements, we have carried out a series of tests. Figure 3-1 shows scala vestibuli sound pressure Psv measured with the hole unsealed, sealed with Jeltrate (a rubbery impression material) and sealed with rigid dental cement. Jeltrate application covered the sensor hole, thus preventing
34

fluid from accumulating. It also made dental cement application possible because dental cement does not cure well in a wet environment. Figure 3-2 shows sensor fiber velocity in the corresponding hole conditions and velocity of the cochlear bone.
Using dental cement, we were able to bring the velocity of the sensor fiber closer to the velocity of the cochlear bone thus minimizing their relative motion and any artifactual pressure this relative motion may contribute to the real scala vestibuli sound pressure measurements. In fact the measured scala vestibuli sound pressure stayed unchange regardless of the relative velocity between the cochear bone and the sensor fiber.
From this exercise, we can safely assume that the effect or the contribution of the relative motion between the bone and the sensor to the measured intracochlear sound pressure in bone conduction is negligible and insignificant.
Adding dental cement on top of the Jeltrate minimizes relative motion but it comes at the expense of potentially losing the ability for post-measurement calibration of the sensor (the cement may hold the sensor so well that it can't be removed without damage). To check whether the sensor is stable, we generally rely on its pre and post measurement calibration. In cases where post-measurement calibration is not possible we employed the in-the-ear stability check and use that to indicate the sensor performance over time. In-the-ear stability of a pressure sensor is monitored in two ways: by the sensor's direct current (DC) readings, and the results of repeated measurements at each stage of our procedures (before and after each manipulation).
A scheme of how we determine and make sure of the sensitivity and stability of a sensor during measurements is detailed in the next section.
35

U
0 ca.
C
0
40
0
30
20
10
0
0
60
50 -
10 0
Psv.Hole
-- PsvJT
-- Psv.DC
-3
10
Frequency (kHz)
Figure 3-1: Scala vestibuli sound pressure PsV with hole unsealed (Hole), sealed with
Jeltrate (JT), sealed with dental cement (DC). Measurements are from Experiment
CH20-130702.
36

80
70
60
-
E
W 50
I'
I ~
C
'~40
I
30
44
II I I
ILI
%I s
Vfiber.Hole
- - -VfiberJT
Vfiber.DC
Vbone i
.
t i 'i
-1 1 -
-.
Ii
20 f-
10
1 t)
0
10
-1
0
-2
-3
do
-'
-4
100
Figure 3-2: Sensor fiber velocity (Vfiber) with hole unsealed (Hole), sealed with
Jeltrate (JT), sealed with dental cement (DC) and velocity of the cochlear bone
(Vbone). Measurements are from Experiment CH20.130702.
37

To check sensitivity and stability of each sensor probe, we performed pre- and postmeasurement calibration. The scheme is as follow.
1. A pre-measurement calibration was made prior to the insertion of the sensor into the vestibule to measure Psv or into the scala tympani behind the round window membrane through the round window cochlear wall to measure
PST.
2. Psv and Ps were measured with AC and BC stimulation with both Psv and
PST holes opened.
3. A post-measurement calibration was made and compared to the pre-measurement calibration. If the sensitivity was within our stability criterion of 3 dB, we proceed with the same sensors, otherwise new sensors were introduced and the process repeats. At the end of this stage, we have a set of repeatable Psv and
PST measurements in the normal condition with both AC and BC stimulation.
4. Psv and PST sensors were reinserted into their respective holes. The holes were then sealed with Jeltrate and detal cement. Jeltrate was used to seal the hole and prevent fluid from collecting. Dental cement was used to provide further sealing in addition to affixing the sensor to the cochlear bone. The use of dental cement ensures minization of the relative motion between the sensor and the bone when the ear is stimulated with bone conduction thereby reducing any artifactual pressures introduced by such motion as explained in the early section. Once the dental cement covering the holes firms up, the sensor fibers were released from the manipulator so they could move freely with the bone.
5. After sealing the sensor holes with Jeltrate and dental cement, another set of
Psv and PsT measurements were made in the normal condition before manipulations (i.e. stapes fixation, ear canal occlusion) occured. These measurements were compared to the earlier set with the hole unsealed to make sure that we
38

have a consistent PSV and PST measurements before we proceed with the manipulations.
6. Manipulation of the ear canal and middle ear were performed. Psv and PST were measured repeatedly after.
7. If possible, we calibrated the sensors after the experiment. Post-measurement calibrations at the end of the experiment were not always possible because of the dental cement affixing the sensor fiber and the cochlear bone, thus we sometimes relied on the pre-measurement calibration and in-the-ear measurements to track the sensitivity change of the sensors if there is any.
Intracochlear differential sound pressure AP (the difference between scala vestibuli and scala tympani sound pressures, AP = PsV PsV), the acoustic input drive to the cochlea, is a useful tool in experimental work to study sound transmission with different conditions of the auditory conductive apparatus (the tympanic membrane and the middle ear) under different stimuli (acoustic and non-acoustic). It has been demonstrated that this input drive proportionally relates to the sensory output of the cochlea, and is a good indicator of hearing sensitivity [6] and [5]. Figure 3-
3 and 3-4 show scala vestibuli sound pressure Psv, scala tympani sound pressure
PST and differential sound pressures AP with air conduction and bone conduction stimulation respectively from one representative experiment. For air conduction, the intracochlear sound pressures plotted were normalized by the ear canal sound pressure while in bone conduction, they are normalized by the voltage input drive to the BAHA sound processor. Because BAHA stimulation produces a large phase accumulation over many cycles as frequency increases, phase of the intracochlear sound pressure measurements are normalized by bone velocity to simplify phase reading. In air conduction, the two intracochlear sound pressures differ largely at low frequency but
39

the differences narrow as frequency increases. The scala vestibuli and scala tympani sound pressures in air conduction are related to the mostly-resistive cochlear input impedance and the highly compliant round window impedance. The notch at around
3 kHz resulted from a resonance that is caused by the opened middle ear cavity [9].
The intracochlear differential sound pressure AP resembles the scala vestibuli sound pressure Psv because PSV magnitude is larger than scala tympani sound pressure
PST. Only at higher frequencies where the two are coming closer, do we see the differential sound pressure become more similar to PST. In bone conduction, the two sound pressures differed by about 10 dB across a wide range of frequency. The difference in shapes and frequency responses of the intracochlear sound pressure in air and bone conduction is related to the normalization of the air-conduction response by the sound pressure at the TM and the normalization of the magnitude of the boneconduction response by the stimulus voltage to the BAHA stimulator. The later (as will be shown later) responds to constant voltage with a bandpass response that is similar to bandpass of the two measured scala pressures.
40

C:
20
3 15
10
40
35
30-
25
5
1
00
0.5
0-
100
-
-
Pst
AP
-0.5-
0~0
-1
100
Figure 3-3: Intracochlear Sound Pressures Psv, Psr and AP with AC stimulation
(Experiment DC9..141106). The measured pressures are normalized by ear canal sound pressure.
41

60
50
-- Psv
~20
W 0
10-
-1.5-
-0
.-
0
> 0.5
2
0 a.
00
10
Figure 3-4: Intracochlear Sound Pressures Psv, PsT and AP with BC stimulation
(Experiment DC9-141106). The magnitude of the measured pressures are normalized
by the stimulus voltage to the BAHA. The phase angle of the measured pressures are normalized by the phase angle of the velocity of the petrous bone evoked by the
BAHA.
42

Sealing the holes around the pressure sensors not only serves to minimize the relative motion between the bone and sensor during bone conduction stimulation, but is necessary to prevent a leak or shunting of the acoustic energy at low frequency and to return the cochlea to its natural state. Neither Slama [20] nor Ravicz [16] found any consistent relationship between the hole size and low frequency scala vestibuli sound pressure. Ravicz [16] did not attempt to seal the scala vestibuli hole due to the limitation of his technique. With our new technique in accessing the vestibule from the paraflocculus, we were able to drill a smaller hole (less than 250 pm in diameter) for the pressure sensor. In addition, fluid generally accumulated in the space around the hole serving as a load to the opened hole. The new technique also made it possible to seal the Psv hole. Figure 3-5 and 3-6 show the measured scala vestibuli sound pressure Psv before and after the hole sealed in AC and BC stimulation respectively.
It doesn't appear from this experiment that the hole affected our Psv measurements, though we sometimes saw improvement in Psv at frequencies less than 1000 Hz.
Like the scala vestibuli hole, we were able to keep the size of the scala tympani hole small, just enough for the sensor tip to fill in. Furthermore, a small hole placed near the highly compliant, low-impedance round window membrane has little effect on the mechanics that contribute to the scala tympani sound pressure PST. In general, we do not see a significant effect of the scala tympani hole on the measured intracochlear sound pressures. Figure 3-7 and 3-8 show the measured scala tympani sound pressure PST before and after the hole has been sealed with AC and BC stimulation respectively. There is no noticeable change between the two conditions.
43

40
25 w z
4-,
20
15
10
35
30
5
0
1 1
10 0
-- Unsealed
- - Sealed
0.5 1-
(U
C-
'I)
~I)
U 0
-0.51
-1
100
Figure 3-5: Measurements of Psv with AC stimulation made before and after sealing the hole around the transducer with Jeltrate and Dental Cement (Experiment
DC9_141106).
44

40 o.
C)
30 rs
20
10
60
50
-
Unsealed
Sealed
0
2
U
1.5 F
0
..
0.5
1
0
0
-0.5 F
-1
10
100
Figure 3-6: Measurements of Psv with BC stimulation made before and after sealing the hole around the transducer with Jeltrate and dental cement (Experiment
DC9-141106).
45

25
4->
M
20
15
10
5
1
0
40
35
30
0.5 F
'V
I-)
'V
Cr, a-
-0.5
0
Unsealed
- - Sealed
10 0
-1
100
Frequency (kHz)
Figure 3-7: Measurements of PST with AC stimulation made before and after sealing the hole around the transducer with Jeltrate and dental cement (Experiment
DC9_141106).
46

40
Q)
30
201
60
50
-- Unsealed
- Sealed
10
~~1
0.5
4-,
U
0
-0.5
0 c
0
-1
(U
-c
0~
-1.5
-2
1
0
10 0
10
Figure 3-8: Measurements of PST with BC stimulation made before and after sealing the hole around the transducer with Jeltrate and dental cement (Experiment
DC9-141106).
47

We used an Etymotic ER-3A insert earphones for the acoustic stimulus. The insert earphone has a foam eartip that can be inserted into the animal's intact ear canal.
For bone vibration, we used a Cochlear Bone Anchored Hearing Aid BAHA BP 100 implanted on the animal's vertex. The benefit of having the BAHA is it allows movement of the animal's head during experiment without having to reposition the bone conduction stimulator on the skull. AC and BC stimulus frequency responses are shown in figure 3-9, 3-10 and 3-11. Figure 3-9 shows the range of sound pressure levels we used as AC stimulus in one representative experiment. In bone conduction, frequency response of the bone anchored hearing aid was measured by velocity of the cochlear bone (see figure 3-10). The bandwidth in which the BAHA produced a considerable amount of output is between 200 Hz to 8 kHz, the range over which all of our data are plotted. The gray range indicates a 20 dB stimulus range used for measurements. Like other vibration sources, the BAHA also emits airborne acoustic sound. However, the BAHA airborne sound pressure is unlikely to be significant given the occluded ear canal with the insert earphone used for AC stimulus. The sound pressure recorded in the occluded ear canal during BAHA stimulation is shown in figure 3-11. This sound pressure is a product of the volume velocity resulted from the compression of the ear canal walls due to bone vibration and the impedance that terminates the ear canal (In this case, the termination that is most relevant is the middle-ear input impedance). This sound pressure is comparable to the ear canal sound pressure produced by the acoustic source between 500 Hz and 5 kHz, and thus could potentially contribute to the total bone conduction responses through the normal air conduction pathway.
48

120
1101-
100
F
CO
=3
-o
:E)
90F
80
70
1
60
10 0
-2
%Aj
-4
U fa
-C
-6
-8
0
-10
100
Frequency (kHz)
Figure 3-9: Ear canal sound pressure PEC generated by air conduction stimulation
(Experiment DC8_141023). Shown in gray is the range of responses to the 20 dB stimulus range used for measurements.
49

BC Stimulus: Cochlear Bone Velocity
-5
U
CL
-10
15
20
10
0
0
50 tA
E
=n
M)
40
0
30
70
60
F
100
-20
100
Frequency (kHz)
Figure 3-10: Cochlear bone velocity generated by bone conduction stimulation (Experiment DC8_141023). Shown in gray is the range of responses to the 20 dB stimulus range used for measurements. Note that the phase is relative to the stimulus voltage.
50

2
0~
90
80
-o
70
(0
60
50
40
30
120
110
100
0
100
-5
(U
-10
%n
-c
-15
-20
10
Frequency (kHz)
Figure 3-11: Ear canal sound pressure PEC generated by bone conduction stimulation in an occluded ear canal (Experiment DC8_141023). Shown in gray is the range of response to the 20 dB stimulus range used for measurements. Note that the phase is relative to the stimulus voltage.
51

The ear canal occlusion effect is a well known phenomenon that plays a significant role in low frequency bone conduction hearing in human [10]. It has been demonstrated that with occlusion, ear canal sound pressure increases by about 10 dB between 0.3 and 2 kHz with a similar increase in perceived stimulus level by about the same amount between 300 Hz and 2 kHz [27]. The mechanism attributed to the occlusion effect is the transmission of sound through air conduction pathway with the increase in ear canal sound pressure resulting from the increase in ear canal terminating impedance with the occlusion. Pertinent to our intracochlear sound pressure measurements, the increase in ear canal sound pressure caused by occlusion should produce the same amount of increase in scala vestibuli and scala tympani sound pressures if the air conduction pathway is the dominant BC transmission route. Our results suggest a more complicated picture. Figure 3-12 shows the magnitude of ear canal sound pressure
FPEC
scala vestibuli sound pressure IPsV , scala tympani sound pressure JPST and differential sound pressure
before and after the occlusion from one individual experiment. Figure 3-13 shows the change in magnitude
(A denotes the change of the measurements after the occlusion) of four experiments with one having no PST measurement and the average data after the occlusion. While there are some variations among animals, ear canal sound pressure PEC I increases
by 10 dB for frequencies below 2 kHz after occlusion. Scala vestibuli sound pressure Psv I increases in a similar amount over a similar frequency range. Interestingly scala tympani appears little affected by the occlusion. If the intracochlear sound pressures PsvI and
IPST are primarily driven by the outer and middle ear sources, we would expect the change in PsVy and JPST would equal the change in ear canal source, the IPEC!. The fact that we see a different change between JPsv and
suggests more than one source contributes to these intracochlear sound pressures in bone conduction, and that PST is especially influenced by the other sources.
52

30
20
10
-M a' 0
4-
2
-10
-20
I
.1
I
I
I
-30
I
-40
U
11
60
50
40
0M
-L
30
Ca
M) 20
10
-10
0
|API
%4
- --
-- Unoccluded
Occluded
100
100
60
50
40
Co
30
C
20
10
-10
0
I
"-
I
60
40
30
C
20
0,
'U
10
-10
0
100
Frequency (kHz)
100
Frequency (kHz)
Figure 3-12: Magnitude of ear canal sound pressure
IPECG, intracochlear sound pressures
and IPSTI and differential sound pressure
before and after the occlusion in one ear (Experiment DC5.140821).
53

40
30
140717
140724
140807
140821
40
30-
CO
20
10
Cn
0
-S
10
0
-10
-20 i
-10
-20
100
AIPsvI
100
40 40
30 30
20
-S a,
10*
CD
0
It
-10
-
--
-
20
~0 a,
-o
C
10 1
0
-V
101
-20 -20
100
Frequency (kHz)
100
Frequency (kHz)
Figure 3-13: The change in the magnitude of ear canal sound pressure
, intracochlear sound pressures
and IPsTI and differential sound pressure
after the occlusion in four ears and the average.
54

Stapes fixation was carried out by applying dental cement between the posterior crus of the stapes, posterior footplate and the cochlear bone where access is possible through the already opened posterior bulla. Because of the wet environment around the stapes area, applying dental cement to fix the stapes is sometimes a challenge.
Careful removal of the fluid is necessary to enable curing of the cement and preventing the flow of cement to unwanted areas (e.g. round window membrane). Air conduction measurements are used to quantify the degree of fixation. Figure 3-14 and 3-15 show the magnitude of the ear canal sound pressure PECl, intracochlear sound pressures
IPSV and
JPST and differential sound pressure
before and after stapes fixation from one animal with AC and BC stimulation respectively. In AC stimulation, the increase in PEC at frequencies below 1 kHz is associated with the increase in middle-ear input impedance produced by the stapes fixation while the 30 dB decrease in JPSvI and
IPST is the result of a reduction in the transmission of sound from the ear canal to the inner ear due to the stiffened stapes. In BC, while the change in I PEC
I is similar to that in AC, the change in scala vestibuli sound pressure IPsvy and scala tympani sound pressure
IPST are different and much smaller. Figure 3-16 and 3-17 show the change in magnitude of
1P
ECl, PSvI, SPSTJ and |API in six experiments with one experiment having no
JPSTJ measurement. Stapes fixation produced a significant effect on AC measurements, but less effect on BC. This is not surprising because several
BC sound transmission pathways do not depend on the normal working of the middle ear. However, the change in Psvy after stapes fixation in bone conduction is not consistent with the hypothesis we developed in our previous work [3] and [4] in which was suggested inner ear mechanisms are the main BC sources and Psy is a function of the fluid volume velocity (driven by inner ear mechanisms) and the impedance of the oval window looking out from the cochlea. Such a hypothesis predicts an increase in Psv after stapes fixation not the decrease we observed. An alternative hypothesis consistent with the new data is that fixation reduces the transmission of airborne sound introduced into the ear canal to scala vestibuli, but not the sound pressure in
55

scala tympani. Such a hypothesis is consistent with the differential effect on these sound pressures we saw when opening and occluding the ear canal with BC stimulation (Fig. 3-13). jPecl
40
30
~0 cii
~0
20
10 r
U)
(V
0
40
30
20
[ a
CU
Arv
10
0 b -i
-10
-20
-20 -
_______
-
I
;
10
-- Normal
- - - Stapes Fixation
PsvI
I'
V
I
40
30
I -I ~0 a)
~0
4-,
C
U)
(V
20
10
0
-10
-20
40
30 a
0
4-J
E)
20
10
0
10 [
-20
100
Frequency (kHz)
-It tI
10
10
Frequency (kHz)
Figure 3-14: Magnitude of ear canal sound pressure IPEC1, intracochlear sound pressures JPsvj and JPST and differential sound pressure IAPI before and after stapes fixation with AC stimulation in one ear (Experiment DC1-140710).
56

C
-10
-20
-30
30
20
10
I,
40
CO
4)
30
Cn
EU
20
10
0
60
50
I'%
I-
_
S-
--Normal
- - Stapes
Fixation
100
60
501
.-%
40
30
C
Im
20
10
-
-
-
0
100
IPst|
60
30
C
MU
20
50
40
10
0
%~ i
I
,'if
I
Ii
10I
Freqency(k~z i
100
Frequency (kHz)
Figure 3-15: Magnitude of ear canal sound pressure
intracochlear sound pressures
and IPSTI and differential sound pressure
before and after stapes fixation with BC stimulation in one ear (Experiment DC1.140710).
-,
57

- - - - - - -
40
30
AlPecl
140710
140717
---
--
140724
140807
140821
- 141023
Average
IM
Su toa
M
-10
-20
-30
10
0
-
-1
-
/ I \
\l
0
Mv
20
10
0
-10
1
-20
-
-40
-50
100
100
APst|
10
0
/1 ~
-10
-v
GD
-u
.2
C to
-20
-30
-50
-40
.4
\
''I
I
100
Frequency (kHz)
.-
-10
-20 t:
-30
-40
10
0
-50
\~
/
<'
'A u-j~~
\~ I
100
Frequency (kHz)
Figure 3-16: The change in magnitude of ear canal sound pressure sound pressures IPsvI and IPsT| and differential sound pressure
after stapes fixation with AC stimulation in six ears and the average.
58

40
30
0
20
M0
10
0
-
140710
-- 140717
--- 140724
140807
140821
- - - 141023
-----Average
-- --
30
20
10
C
0
-10
' /
-201
-10
-30
-20
100
AIPsvI
100
30
30
20 [
20 [
10
0
Cn
-10
*O
M3
C0
4.M
10
0
10
-20
-201 V
10
0
Frequency (kHz)
-30
100
Frequency (kHz)
Figure 3-17: The change in magnitude of ear canal sound pressure
intracochlear sound pressures
and |PST| and differential sound pressure
after stapes fixation with BC stimulation in six ears and the average.
59

60

In our early work on bone conduction in chinchillas in which intracochlear sound pressures were measured before and after middle-ear ossicular chain interruption [3] and [4], we provided initial evidence of multiple mechanisms that contribute to the total bone conduction response and that the inner-ear mechanisms appear to be the dominant sources. In this thesis, we continue and extend our previous work by measuring intracochlear sound pressures before and after ear canal occlusion and stapes fixation while enhancing our measurement technique of the intracochlear sound pressures that minimizes the contribution of potential artifacts such as relative motion between the cochlear bone and the sensor fiber used to measure the intracochlear sound pressures. Stapes fixation was used to test our previous hypotheses that inner ear mechanisms are the main drive to PSv and PST in bone conduction and that reduction in PSv after middle ear interruption is the result of a decrease in impedance looking out from the oval window. Ear canal opening and occluding were introduced to investigate how a simple change in the BC ear canal source effects the two scalae sound pressures produced by bone conduction stimualtion.
Results of intracochlear sound pressures measurements in the normal intact middle ear before the manipulations support the early work that there are multiple mechanisms that contribute to bone conduction responses. Figure 4-1 shows the magnitude ratio of scala tympani sound pressure JPSTJ to scala vestibuli sound pressure
61

that is
in AC and BC. The magnitude ratio between the two intracochlear sound pressures in air conduction is similar to that reported by Ra vicz et al. [16]. At frequencies less than 1 kHz, the ratio is very small, consistent with
JPSV >> JPSTJ
Psv and PSj in air conduction can be estimated from volume velocity of the oval window and round window and their respective impedances namely cochlear input impedance and round window impedance. Assuming the same volume velocity at the oval window and at the round window for air conduction, the ratio of the two sound pressures is equivalent to the ratio of their respective impedances. Cochlear input impedance is primarily resistive at frequencies above 300 Hz while the round window impedance is highly compliant. At lower frequencies the cochlear impedance shows contributions of the helicotrema and the round window [12]. Results of IPST/PSVI in air conduction are consistent with an influence of round window stiffness on the cochlear impedance at frequencies below 300 Hz while at higher frequencies the resistence of the cochlear input impedance becomes dominant. In bone conduction, complexities in the frequency dependence, shape and size of IPST/PSV are not consistent with a single volume flow from oval to round window and suggest the presence of multiple BC sound sources.
It is clear that BC ear canal source and middle ear inertial source alone cannot completely explain the data we obtained with BC stimulation. If either the ear canal source or the middle ear inertial source dominated the bone conduction stimulus to the inner ear, we would expect the pressures ratio curve in BC to be similar to that in
AC. Could the BC results be explained by the inner ear mechanisms alone such that
PSV and PST depends on volume velocity at the two windows and their impedances looking out from the cochlea? Data from stapes fixation 3-17 are not consistent with this hypothesis. Stapes fixation which increases the impedance of the oval window looking out from the cochlea failed to produce an increase in the scala vestibuli sound pressure PSV. If inner ear drives determined PSV, we would expect to see the opposite. However, as we have seen with all of our middle ear manipulations, the change in bone conduction produced
IPST is smaller than the change in JPsvy after manipulations, and such a result is consistent with a dominant inner ear source for
PST.
62

Therefore, our data suggest that while Psv in bone conduction is significantly influenced by sound transmitted by the ear canal and middle ear, PST is not and appears to be dominated by inner-ear sources.
Air Conduction Pst/Psv
-10
0
30
Bone Conduction Pst/Psv
20
10
-20
1~
-10 -30
-40
10*
Frequency (kHz)
-___140710
140717
--
140724
140821
-
141023
141106
Average
-20
10'0
Frequency (kHz)
Figure 4-1: Ratio of the magnitude of scala tympani sound pressure
to scala vestibuli sound pressure
in AC and BC in normal ears.
63

PST
Our Psr and Psv measurements in both AC and BC before and after the manipulations were well above noise at most of the measured frequencies. As pointed out earlier in the Results section, all data shown here are at least 10 dB above the noise floor (where noise was defined by the off-stimulus frequency components in the Fourier transforms of the responses to a sinusoidal stimulation). In bone conduction before any manipulations, the two scalae sound pressures differ by only 10 to 15 dB below
2 kHz unlike in air conduction where the two differ by as large as 30 dB, and since we can clearly distinguish the effects of our manipulation on Psv and PST in air conduction, we do not expect that noise will limit PsT in bone conduction.
Could artifacts produced by relative motion between the cochlear bone and sensor fiber used to measure PST affact its measurements? We have shown in an earlier section that this relative motion did not effect our estimates of Psv. Because the insertion direction of PST fiber sensor was not paralell to that of Psv, one could argue the effects of relative motion between the cochlear bone and sensor fiber is different in the two sensors. Thus we could not apply what we found on Psv and assume that the relative motion has no effect on PST measurements. Figure 4-2 shows the same measurement procedure to investigate the effect of relative motion between the cochlear bone and sensor fiber in PST as we did with Psv. The top panel is the magnitude of scala tympani sound pressure IPsTI with the sensor hole unsealed, sealed with Jeltrate and with dental cement.
in the three conditions were the same even though the velocity of the fiber in the hole sealed with dental cement gets closer to the cochlear bone velocity (see bottom panel of figure 4-2). As we have seen in the test in Psv measurements, these results are strong evidence that relative motion between the sensor and the bone does not affact our measurements of either
Psv or PST.
64

s0
40 -
-Pst.Hole
- - -
PstJT
Pst.DC
30-
'W 20
10
-10
0 --
80
100
Frequency (kHz)
70
E
-60-
:, so -
On40
30 -
--
VfiberJT
-- Vfiber.DC
Vbone
20
100
Frequency (kHz)
Figure 4-2: Top: Magnitude of scala tympani sound pressure IPSTI when the sensor hole unsealed (Hole), sealed with Jeltrate (JT) and with dental cement (DC). Bottom: Magnitude of fiber velocity (Vfiber) when hole sealed with Jeltrate (JT) and dental cement (DC) and cochlear bone velocity (Vbone). Measurements are from
Experiment DC5.140821.
65

What else could limit PST measurements? Another idea worth investigating is that PST in the normal middle ear condition before manipulatiion is already at its minimum. Thus if we increase it to a level closer to Psv, we might be able to see a bigger change in Psi after manipulation. Figure 4-3 shows
and IPST| before and after partial round window fixation and incudostapedial joint (IS joint) interruption in air conduction (top panel) and bone conduction (bottom panel). Partial round window fixation was achieved by applying a tiny drop of dental cement on the posterior part of the round window membrane. This partial round window fixation increased IPsTI by more than 10 dB for air conduction and less than 10 dB for bone conduction at frequencies below 1 kHz. The increase in round window impedance has little effect on IPsvI. With this manipluation we were able to increase
in bone conduction to a level similar to
at frequencies below 800 Hz. We then interrupted the incudostapedial joint to see if the change in PsT is similar to the change in
IPsvl after IPSTI was increased. What we found was the change in
was smaller than the change in tPsvI after IS-joint interruption, consistent with our earlier work where middle ear interruption was carried out without partial round window fixation [3] and [4] where the two scalae sound pressure decrease to the same level in air conduction. From these exercises, we confidently concluded that Ps, measurements in bone conduction was not affected by either the noise, the artifacts or limited by our techniques.
66

cc
20
V
-a
C tM
10
0
-10
-20
50
40
30
-30
60
50
40a-
-U a,
-U
30
C
VI
20
10
-10
0
AC
100
BC
Cn
40
30 o4-
C
20
10
0
-10
-20
-30
---
-- Normal
-- Partdal.RW.Fixation
Ijoint.interruption
60
50
0-1
40
30
C
20
10
-10
0
AC
100
BC |Pst|
100
Frequency (kHz)
100
Frequency (kHz)
Figure 4-3: Top: |PsvI and
normalized by PEC before and after partial round window fixation and IS-joint interruption in air conduction. Bottom: IPsvI and
before and after partial round window fixation and IS-joint interruption in bone conduction (Experiment DC9.141106).
67

30
101
We have data of scala vestibuli sound pressure Psv as well as scala tympani sound pressure PsT before and after three manipulations: ear canal occlusion (see figure 3-
13), middle ear interruption [3] [4] (for reference, figure 4-4 is included here) and stapes fixation (see figure 3-17). In bone conduction, middle ear interruption, opening the ear canal and stapes fixation produced a small decrease of 10 dB in IPsvI with little effect on IPsrI. Furthermore, since the changes in both Psv and Psr after these manipulations were much smaller in bone conduction than in air conduction we conclude inner ear mechanisms are the dominant sources in bone conduction.
AC
APsv APst
CB
-3s
-
-
Frequency (kHz)
Figure 4-4: Effects of middle ear interruption on scala vestibuli sound pressure Psv and scala tympani sound pressure PsT in AC and BC (figure from [4]).
68

While it is clear that inner ear mechanisms contribute the most to Psv and
PsT, we cannot rule out the middle ear source contribution because of the decrease in PSV after middle ear interruption, stapes fixation and opening the ear canal. A significant question is why was PST not affacted the same way as PSV by bone-conducted sound coming from the ear canal and middle ear. In air conduction, much data is consistent with the two-window model that describes the generation of the window pressure difference PSV PST from sound transfer between the oval and round windows in an otherwise rigid-walled cochlea. In the two-window model air conduction stimulation produces an equal magnitude of volume velocity but opposite in phase between the oval window and the round window. Such a model is consistent with measurements of oval and round window volume displacement made in a human temporal bone model
[24]. In bone conduction, however, the same study found there was a difference of 5 to 15 dB of volume displacement between the two windows where the round window volume displacement is larger at frequencies below 2 kHz and the oval window volume displacement becomes larger at higher frequencies. They attributed the differences in the two windows volume displacement to the inner ear mechanisms of bone conduction: cochlear bone compression, compressibility of the fluid and third-window pathways. Our data are consistent with their results suggesting the inner ear mechanisms are responsible for the differences we see in PSV and PST measured near the oval window and round window respectively.
Is it possible that PST is strongly related to the inner ear mechanisms while PSV is a combination of multiple mechanisms: ear canal, middle ear, and inner ear? Our data are consistent with this hypothesis, but what mechanism causes either the ear canal or middle ear source to contribute to PSV but not PST? Such phenomenon can not be explained by the two-window cochlear model because in that model the sound flow that produces Psv equals the flow that evokes PST. If there is a compressible component within the inner ear or another outlet or window where sound energy can flow in and out, then maybe the flow of sound through that outlet counteracts the contribution of the ear canal or the middle ear source, thus protecting PST from this influence. A sound source positioned in the scala tympani and near the round
69

window would be consistent with the second option. The cochlear aqueduct, a narrow channel between the inner ear on the scala tympani and the brain, appears to be a good candidate for such a source. This relative high impedance outlet does not affect the flow of sound from oval window to round window in air conduction because the impedance of the round window is much lower than the impedance of the cochlear aqueduct. However in bone conduction where one of the sources originates in the inner ear, this third-window pathway can act as an additional window allowing the flow of sound into and out of the inner ear. This hypothesis is consistent with a hypothesis
Stenfelt et al. [24] used to explain the differences in the volume displacement of the oval and round windows in bone conduction.
70

This thesis builds upon our previous work with a primary goal to define and quantify dominant sources of bone conduction in chinchillas. In our early work, results from intracochlear sound pressure in air conduction and bone conduction after middle ear interruption suggested inner ear mechanisms are the dominant sources in bone con- duction stimulation. In this work we continued and extended the investigation of inner ear mechanisms by measuring intracochlear sound pressures with air conduction and bone conduction stimulations before and after stapes fixation to test our hypothesis that inner ear mechanisms are the dominant sources in bone conduction and that Psv was a function of the inner ear fluid volume velocity and the oval window impedance looking out from the cochlea. We also used ear canal opening and occluding to investigate the occlusion effect in bone conduction. Series of tests were also carried out to look for potential sources of artifacts that could limit our measurements of intracochlear sound pressures in bone conduction. The following is a summary of results and interpretations from this work.
" The potential artifact due to relative motion between the cochlear bone and sensor fiber used to measure intracochlear sound pressures does not contribute to our measurements of sound pressure inside the cochlea in bone conduction.
" The noise defined as off-frequency stimulus responses did not seriously limit our
71

ability to measure scala vestibuli Psv and scala tympani PST sound pressures in AC and BC.
" In AC, JPsvj is 20-30 dB larger than JPsl at low frequencies and their difference decreases as frequencies increase. The magnitude ratio of PST to PsV is consistent with a report from Ravicz et al., [16] suggesting the presence of round window membrane stiffness at frequencies below 300 Hz and a significant resistive component at higher frequencies in the cochlear input impedance.
* In BC, IPsvI is only 10-15 dB larger than JPSTJ and their ratio is different in shape and size compared to AC suggesting multiple mechanisms involved.
* Stapes fixation on BC produced Psv was not consistent with our earlier hypothesis that inner ear mechanisms are the main driving sources to PsT, in that increases in the stiffness of the stapes produced decreases in Psv rather than the increases predicted for inner-ear sources.
* Ear canal occlusion increased PSv by about 10 dB at frequencies below 2 kHz, consistent with a similar amount of increase in ear canal sound pressure PEC.
However, the occlusion induced change in the ear canal sound pressure failed to affect PST.
* All three manipulations (ear canal occlusion, middle-ear interruption, stapes fixation) produce small but measurable changes in Psv but little change in PST in bone conduction as compared to large changes in both Psv and PST in air conduction. These results suggest the two-window cochlear model which fits air conduction data does not explain our bone conduction results.
" The existence of compressible cochlear structures or third window pathways, e.g. the cochlear aqueduct may explain why we see differences in the change between Psv and PST in bone conduction stimulation. This suggests a role of the cochlear aqueduct in BC responses, consistent with Sohmer's idea of non-osseous mechanisms in bone conduction [7].
72

* While ear canal compression and middle ear inertia sources may contribute to the total bone conduction response (10 dB decrease in Psv after middle ear interruption and stapes fixation, and 10 dB increase after ear canal occlusion), inner ear mechanisms appear to be highly significant sources for bone-conducted sound, especially when considering the effect on
PST.
73

74

[1] Biriny E. A contribution to the physiology of bone conduction. Acta Otolaryn-
gol, Suppl. 26, 1-223, 1938.
[2] B6kesy G von. Experiments in Hearing. McGraw-Hill, New York, Chap 6, 127-
203, 1960.
[3] Chhan D, Rb5sli C, Mckinnon ML, Rosowski JJ. Evidence of inner ear contribution in bone conduction in chinchilla. Hear Res, 301:66-71, 2013.
[4] Chhan D. Role of the middle ear inertial component of bone conduction in chinchilla. M.S. thesis. MIT, Cambridge, MA, 2013.
[5] Dallos P. Cochlear potentials and cochlear mechanics in basic mechanisms of hearing. Academic, New York, 335-376, 1973.
[6] Dancer A and Franke R. Intracochlear sound pressure measurements in guinea pigs. Hear Res 2, 191-205, 1980.
[7] Freeman S, Sichel JY, Sohmer H. Bone conduction experiments in animals evidence for a non-osseous mechanism. Hear Res, 146:72-80, 2000.
[8] Goldstein D and Hayes C. The occlusion effect in bone-conduction hearing.
Hearing Measurement: A Book of Readings. Appleton-Century- Crofts, New
York, pp. 150-157, 1971.
[9] Guinan JJ, Peake WT. Middle-ear characteristics of anesthetized cats. J Acoust
Soc Am, 41:12371261, 1967.
[10] Huizing EH. Bone conduction- the influence of the middle ear. Acta Otolaryngol,
Suppl 155, 1960.
[11] Jahn AF and Santos-Sacchi J. Physiology of the ear, 2nd Edition, Singular,
2001.
[12] Lynch TJ III, Nedzelnitsky V, and Peake WT. Input impedance of the cochlear in cat. J Acoust Soc Am, 72:108-130, 1982.
[13] Minor LB. Superior canal dehiscence syndrome. Am J Otolaryngol, 21(1):9-19,
2000.
75

[14] Ng M, Jackler R. Early history of tuning-fork tests. Am J Otolaryngol, 14:100-
105, 1993.
[15] Olson ES. Observing middle and inner ear mechanics with novel intracochlear pressure sensors. J Acoast Soc Am, 103:3445-3463, 1998.
Ravicz ME, Slama MC and RIosowski JJ. Middle-ear pressure gain and cochlear partition differential pressure in chinchilla. Hear Res, 263(1-2):16-25, 2010.
[17] Rosowski JJ. Outer and middle ear. In: Fay R.R, Popper AN, editors. Comparative Hearing in Mammals. Springer-Verlag, New York, pp. 172-247, 1994.
[18] Rosowski JJ, Ravicz ME, and Songer JE. Structures that contribute to middleear admittance in chinchilla. J Comp Physiol, 192(12):1287-1311, 2006.
[19] Schloss F and Strasberg M. Hydrophone calibration in a vibrating column of liquid. J A coast Soc Am, 34(7):958-960, 1962.
[20] Slama MCC. Middle ear pressure gain and cochlear input impedance in the chinchilla. M.S. thesis. AlIT, Cambridge, MA, 2008.
[21] Slama MCC, Ravicz ME and Rosowski JJ. Middle ear function and cochelar input impedance in chinchilla. J Acoust Soc Am, 127(3):1397-1410, 2010.
[22] Songer JE, Rosowski JJ. A mechano-acoustic model of the effect of superior canal dehiscence of hearing in chinchilla. J Acoust Soc Am, 122(2):943-51, 2008.
[23] Stenfelt S, Hato N, Goode RL. Factors contributing to bone conduction: The middle ear. J Acoust Soc Am, 111:947-959, 2002.
[24] Stenfelt S, Hato N and Goode R. Fluid volume displacement at the oval and round windows with air and bone conduction stimulation. J Acoust Soc Am,
115(2):797-812, 2004.
[25] Stenfelt S, Goode RL: Bone conducted sound: Physiological and clinical aspects.
Otol Neurotol, 26:1245-1261, 2005.
[26] Stenfelt S. Middle ear ossicles motion at hearing thresholds with air conduction and bone conduction stimulation. J Acoust Soc Am, 119:2848-2858, 2006.
[27] Stenfelt S, Reinfeldt S. A model of the occlusion effect with bone-conducted stimulation. International Journal of Audiology, 46:595-608, 2007.
[28] Tonndorf J. Compressional bone conduction in cochlear models. J Acoust Soc
Am, 34, 1127-1131, 1962.
[29] Tonndorf J. Bone conduction; Studies in experimental animals; A collection of seven papers. Acta Oto-Laryngol, Suppl. 213, 1-132, 1966.
[30] Wever EG and Lawrence M. Physiological acoustics. Princeton, NJ: Princeton
University Press, 1954.
76