Introduction and Basics

Kompis M, Caversaccio M-D (eds): Implantable Bone Conduction Hearing Aids.

Adv Otorhinolaryngol. Basel, Karger, 2011, vol 71, pp 10–21

Acoustic and Physiologic Aspects of Bone

Conduction Hearing
Stefan Stenfelt
Department of Clinical and Experimental Medicine, Linköping University, Linköping, Sweden

Abstract in the eardrum that are transmitted via the mid-

Bone conduction (BC) is the way sound energy is trans- dle ear ossicles and become a sound pressure in
mitted by the skull bones to the cochlea causing a sound the cochlea (scala vestibuli). This sound pressure
perception. Even if the BC sound transmission involves acts on the basilar membrane producing a travel-
several pathways including sound pressure induced in the ing wave that excites the sensory cells in the organ
ear canal, inertial forces acting on the middle ear ossicles of Corti causing an auditory sensation. Bone con-
and cochlear fluids, alteration of the cochlear space, and duction hearing is when the sound is transmitted
pressure transmission through the 3rd window of the through the skull bone, cartilage, skin and soft tis-
cochlea, the BC sound ultimately produces a wave motion sue, and fluids in the body, ultimately resulting
on the basilar membrane similar to that of air-conducted in a sound pressure in the cochlear scalae. This
sound. The efficiency of the BC stimulation is largely type of sound transmission is sometimes divid-
dependent on the skull bone where the skull acts as a rigid ed between body conduction and bone conduc-
body at low frequencies and incorporates different types tion, where the latter is only sounds transmitted
of wave transmission at higher frequencies. The interau- in the skull bone. Here, for simplicity, both body
ral stimulation difference is determined by the difference and bone conduction will be referred to as bone
between contralateral and ipsilateral BC sound transmis- conduction and abbreviated BC.
sion: the transcranial BC sound transmission. To benefit
from binaural processing, the transcranial transmission
should be low, while the same should be high when using Clinical Measurements using Bone
BC hearing aids for unilateral deaf subjects. By appropri- Conduction
ately positioning the stimulation, high or low transcranial
transmission can be achieved. Understanding the processes of BC sound was
Copyright © 2011 S. Karger AG, Basel early driven by its use for differential diagnosis
between conductive and sensorineural hearing
The conventional way of auditory stimulation is by loss. In the 19th century, the usage of the tuning
an airborne sound that is transmitted through the fork provided tests as Weber test and Rinne test
ear canal, where it induces mechanical vibrations [1]. After the introduction of the audiometer and
203.64.11.45 - 2/18/2015 11:05:12 PM
Kainan University
Downloaded by:
the BC transducer (e.g. Radioear B71) during the interface [9] have also been suggested to improve
20th century, most diagnosis of conductive hear- BC audiometry.
ing losses is by comparison of air conduction (AC)
and BC hearing thresholds (termed the air-bone
gap). Although BC hearing thresholds are affect- Basilar Membrane Stimulation
ed by an outer or middle ear lesion, the alteration
is small compared to the change in AC thresholds, One important issue related to perception of BC
and the air-bone gap is often used to quantify the sound is the sensory input to the neural system.
conductive loss. Since BC vibration causes motion of the whole
It is sometimes desirable to use BC stimula- skull and, most probably, also other parts of the
tion with other types of auditory measures such human body, the vibration perception is not nec-
as speech recognition, otoacoustic emissions, or essarily only auditory. At lower frequencies and at
electrophysiology-based tests such as brainstem higher stimulation levels, BC sound causes vibro-
response audiometry [2–4]. Even if it is possible, tactile excitation and the perception is multi-
the results are most often not directly comparable modal [10]. Moreover, BC vibration stimulation
with those obtained using AC stimulation. This can excite the vestibular system, which is espe-
difference is mostly related to the difference in cially noticeable in large vestibular aqueduct syn-
frequency response between earphones and the drome and semicircular canal dehiscence [11]. In
BC transducer, differences in hearing thresholds a study on loudness growth comparing AC and
(sound pressure for AC stimulation and force for BC sound, it was shown that the loudness of BC
BC stimulation) and the general poorer dynamic sound increased more compared with loudness of
range of a BC transducer (greater amount of non- AC sound, especially at the low frequencies [12].
linear distortion). Moreover, testing setups for AC This may be explained by multimodal excitation
may be incompatible with BC excitation. An ex- when stimulation is by BC resulting in a great-
ample of the latter is otoacoustic emissions [3], er growth of loudness compared to AC sound for
where normally the emissions are measured by a equal increase in sound intensity.
high-quality low-noise microphone tightly fitted It is well established that BC sound perception
in the ear canal giving a considerable occlusion is primarily caused by basilar membrane vibra-
effect, thereby affecting the result at frequencies tion. Firstly, BC pure tones can be cancelled by
below 2 kHz (see later section) [5]. AC tones [13–15], indicating that the two ways of
The most commonly used BC transducer for stimulation excite the same sensory cells for audi-
audiometry is the Radioear B71 which has well- tory perception. Secondly, direct measurement of
known characteristics and is used in the interna- basilar membrane motion shows similar excita-
tional standard for hearing thresholds [6]. The tion pattern independent of whether stimulation
drawback of this device is poor dynamic and is by AC or BC [16]. Thirdly, electrophysiology
frequency range: the great amount of nonlin- measures such as brainstem audiometry give sim-
ear distortion and housing resonances limits its ilar results for AC and BC stimulation when dif-
practical use to the frequency range 0.25–4 kHz ferences in stimulation spectrum are corrected for
and levels below 60–70 dB HL. These limitations [17] and, fourthly, distortion product otoacoustic
have often been addressed, and recently anoth- emissions can be generated by one AC and on BC
er BC transducer was proposed enabling high- tone [3]. Even if a BC sound influences the whole
frequency BC testing [7]. Other designs of BC cochlea and not only a local point as AC sound,
transducers providing greater dynamic range the inherent response of the basilar membrane
[8] and measurement of the transducer-skull is always a wave motion beginning at the base
203.64.11.45 - 2/18/2015 11:05:12 PM

Acoustic and Physiologic Aspects of Bone Conduction Hearing 11

Kainan University
Downloaded by:
 Air
Own voice
conduction

Bone
conduction
Skin Skull bone CSF

Soft Inertial Compression

tissue Bone forces Inertial Sound
expansion
vibration vibration forces pressure

Outer Middle Inner

ear ear ear Neural
Air response
conduction

Fig. 1. A model of the pathways for hearing BC sounds. A BC vibration on the com-
pressed skin of the skull bone causes vibrations of the skull and also produces a
sound pressure in the skull interior. The vibration of the skin and bone produces
a sound pressure in the ear canal, while inertial forces cause relative vibration be-
tween the ossicles and the surrounding bone. The sound is transmitted to the inner
ear from the outer and middle ear, but also directly through inertial forces acting
in the cochlear fluids, through compression and expansion of the cochlear space,
and, to some extent, through sound pressure transmission from the skull interior.
The own sound production is transmitted to the inner ear by both airborne sound
and BC.

traveling towards the apex [18]. Therefore, the component, and the inner ear component, respec-
assumed important parameter for BC-induced tively [19]. However, this categorization does not
wave motion on the basilar membrane is the pres- divide between different physical processes that
sure difference between scala vestibuli and scala are involved in the transformation from a skull
tympani. bone vibration to a pressure difference between
scala vestibuli and scala tympani and subsequent
basilar membrane wave motion. Tonndorf [20],
Pathways for Bone-Conducted Sound who investigated different mechanisms contrib-
uting to BC sound perception in the cat, identi-
Since it became generally accepted that humans fied seven components believed to be important
can perceive BC sound, theories of how a BC vi- for BC sound. In a previous paper, five compo-
bration in the skull bone becomes a basilar mem- nents were presented as the most important for
brane vibration have been presented. Early theo- BC sound perception in the normal and impaired
ries identified one or two pathways that resulted ear [21]. Below is a short summary of these five
in BC sound. Today, it is generally accepted that components (fig. 1).
BC sound transmission involves multiple path-
ways, and there is no obvious way to distinguish Sound Pressure in the Ear Canal and the
between them. One often used categorization is Occlusion Effect
the anatomical division where BC sound affect- When the skull is excited with BC sound, the ear
ing the outer, middle, and inner ear is referred canal deforms due to the vibrations, and airborne
to as the outer ear component, the middle ear sound is produced in the ear canal. This sound is
203.64.11.45 - 2/18/2015 11:05:12 PM

12 Stenfelt
Kainan University
Downloaded by:
transmitted to the eardrum and further transmit- determining the occlusion effect where greater
ted through the middle ear ossicles and produces volume results in less occlusion effect. It should
a sound pressure in the scala vestibuli causing a be noted that the occlusion effect measured as
basilar membrane traveling wave. Consequently, change in ear canal sound pressure is usually 5–10
this BC sound pathway is similar to an AC sound dB greater than the occlusion effect as measured
transmission and should be considered as a skull by alteration of BC hearing thresholds at frequen-
vibration-produced AC sound. This means that cies below 1–2 kHz [5].
this pathway is affected by the status of the outer
and middle ear and the transmission is altered by Inertia of the Middle Ear Ossicles
it. For the normal open ear, the contribution of the From a mechanical point of view, the middle ear
ear canal pathway of BC sound is some 10 dB be- ossicles are suspended in the middle ear space by
low other contributors at frequencies below 2 kHz the eardrum, ligaments, and tendons. Most im-
and less important at higher frequencies [22]. The portant for BC sound is the eardrum and the an-
outermost part of the ear canal comprises carti- nular ligament positioning the stapes footplate
lage and soft tissue, and is referred to as the carti- in the oval window: these two act as mechani-
lage part of the ear canal, while the innermost part cal springs attached to the free masses (ossicles).
is surrounded by skull bone known as the bony When the skull bone surrounding the middle ear
ear canal. If the BC transducer is positioned close cavity vibrates due to BC excitation, the spring
to the ear canal, for example on the mastoid, the effect causes the ossicles to vibrate in phase with
soft tissue is efficient in transmitting the sound to the skull at low frequencies. At higher frequen-
the ear canal at low frequencies, while the bony cies, the ossicles become vibrationally decoupled
part is the primary source for the BC-induced ear from the surrounding bone resulting in a relative
canal sound pressure at higher frequencies [22]. motion between the stapes footplate and the otic
The outer ear pathway of BC sound can be capsule. This behavior was experimentally veri-
dominating at lower frequencies if the open part fied in Stenfelt et al. [24], where the ossicles vi-
of the ear canal is closed, often termed the occlu- brated with the surrounding bone at low frequen-
sion effect. The effect of occluding the ear canal cies and showed large relative motion above the
depends on the type and position of the occluding middle-ear ossicle resonance frequency of 1.5
device, but can give up to 40 dB increased sensitiv- kHz. This finding was later verified in Homma et
ity to BC sound at low frequencies [5]. The origin al. [25] showing that the difference in middle ear
of the occlusion effect is the change in radiation ossicle resonance frequency between BC and AC
impedance at the ear canal opening. With the ear [e.g. 26] sound was caused by differences in vibra-
canal open and at low frequencies, the radiation tion modes of the ossicles.
impedance is low, and a large part of the sound It has been speculated that the middle ear os-
energy induced in the ear-canal by the BC skull sicle inertia is important for BC sound percep-
vibration leaks out of the ear canal. However, if tion at frequencies around the resonance fre-
the canal is occluded, the sound energy is trapped quency and below, i.e. approximately 2 kHz and
and transmitted to the eardrum. This phenom- below. Supporting this theory is, for example, the
enon was described by Tonndorf [20] and later worsening in BC thresholds around 2 kHz follow-
modeled in Stenfelt and Reinfeldt [5]. As predict- ing immobilization of the ossicles due to otoscle-
ed by von Békésy [23], the model showed that if rosis of the stapes (also known as the Carhart’s
the occlusion is deep enough, the occlusion ef- notch) [27] and increased BC sensitivity below
fect is insignificant [5]. Also, if circumaural de- 2 kHz with artificial mass loading of the ossicles
vices are used, the enclosed volume is primarily (lowering the resonance frequency) [28]. When
203.64.11.45 - 2/18/2015 11:05:12 PM

Acoustic and Physiologic Aspects of Bone Conduction Hearing 13

Kainan University
Downloaded by:
comparing ossicle vibration at hearing threshold static pressure can be transmitted from the cranial
with AC and BC stimulation, it was found that the space to the cochlea [34]; the same is also possible
ossicle inertia can be important for BC perception for dynamic pressure [33]. This type of transmis-
in the normal ear at and slightly above the reso- sion has primarily been attributed to a patent co-
nance frequency (2–3 kHz) [29]. chlear aqueduct, but the vestibular aqueduct and
other channels may well enable such mechanism.
Inertia of Cochlear Fluids and Fluid Pressure Although the sound pressure transmission from
Transmission the skull interior to the cochlea has been suggest-
The fluid in the cochlea is also subject to iner- ed as an important BC contributor, clinical re-
tial forces when the bone surrounding the cochlea sults as transcranial transmission and pathologi-
vibrates. The result of these forces is a pressure cal findings in semicircular canal dehiscence and
gradient across the basilar membrane that would large vestibular aqueduct syndrome suggest that
form a traveling wave. Since the fluid can be con- this contributor to BC sound perception is not the
sidered incompressible, a fluid flow in the co- most important one.
chlea would require compliant inlet and outlet of It has been estimated that only a small fraction
the fluid on both sides of the basilar membrane. of the fluid of the cochlea (less than one millionth)
The two obvious examples are the oval window is necessary to be displaced to give a hearing sen-
on the scala vestibuli side and the round window sation of 80–100 dB HL [21]. This indicates that
on the scala tympani side. However, there are the fluid inertia would be an efficient way to excite
several other structures that can act as compli- the cochlea. Moreover, in cases of immobilization
ant pathways of the cochlea. These structures are of the stapes, e.g. otosclerosis of the stapes, the AC
collectively known as the third window [30]. This hearing thresholds deteriorate significantly while
means that, as long as there is a pressure gradient the BC thresholds are only slightly affected in
over the basilar membrane, there will also be fluid the 2 kHz range. Since transmission through the
flow acting on the basilar membrane initiating a middle ear is severely reduced in such case, the
traveling wave. BC sound is transmitted directly to the cochlea.
The pathways other than the oval and round Moreover, the skull is believed to constitute near-
windows that enable pressure and fluid transmis- rigid body motion (wavelength longer than the
sion to and from the cochlea include the vestib- size of the skull) at lower frequencies, compres-
ular and cochlear aqueducts, nerve fibers, veins, sion and expansion of the cochlear walls cannot
and micro-channels entering the cochlea [31]. It explain the BC sound perception. Consequently,
has for example been noticed that an improved for lower frequencies, fluid inertia is likely the
fluid connection between the vestibular space and most important pathway for BC sound in the
the cranial space, known as a semicircular canal normal ear but probably less important at higher
dehiscence, improves the low-frequency BC sen- frequencies.
sitivity [11]. Similar results are found when the
dehiscence is made artificially into the middle Alteration of the Cochlear Space
ear space [32]. The concept of the third window When a transversal wave propagates in the skull
offers two different interpretations of BC stim- bone, the skull bone compresses and expands
ulation pathways. One is the inertial effects of with the wave. This means that the structure de-
the cochlear fluid resulting in a fluid displace- forms. When such wave affects the cochlea, the
ment mentioned above. The other interpretation deformations alter the cochlear space, causing
is sound pressure transmission from the cranial a fluid motion that gives a sound pressure. This
space to the cochlea [33]. It has been shown that pathway was termed inner ear compression by
203.64.11.45 - 2/18/2015 11:05:12 PM

14 Stenfelt
Kainan University
Downloaded by:
von Békésy [19] and distortional component by surface [19, 35, 36], while more recent studies
Tonndorf [20], and has often been central when have measured the response at the cochlea, ei-
explaining BC sound perception. The theory for ther in all three space dimensions [37, 38] or in
the compressional response is that the cochlea is one dimension [39, 40]. Even if these studies pro-
unsymmetrical regarding scala tympani and scala vide insight into the vibration characteristics of
vestibuli. Scala vestibuli space is about 50% great- the skull bone and differences in transmission ef-
er than scala tympani, and the impedance of the ficiency between stimulation positions and the
oval window is greater (stiffer) than the imped- cochlea, it is not clear how to relate a vibration
ance of the round window (more compliant) [20]. of the cochlea to a hearing sensation. However,
Consequently, when the cochlea is compressed, the hypothesis is that, in relative terms, a greater
excess fluid is forced from the scala vestibuli side vibration response of the cochlea at a specific fre-
to the scala tympani side and the round window. quency leads to a greater BC sound perception. It
The opposite happens when the cochlea expands, was shown in the Stenfelt and Goode study [38]
i.e. fluid flows from scala tympani towards scala that with BC stimulation at the skull surface, the
vestibuli, thereby exciting the basilar membrane. cochlea vibrates in all space dimensions without
The cochlea is coiled, and its dimension can any dominant direction.
be thought of like a sphere with approximately In normal BC hearing testing, the BC trans-
10-mm diameter. If the limit for effective com- ducer is placed on the compressed skin that trans-
pression response is set to a wavelength that is less mits the BC vibration to the skull bone. The effect
than ten times the size of cochlea, the lowest fre- of the skin, in terms of dynamic force transmis-
quency where the compressional response would sion, is small at low frequencies but will attenuate
be an effective excitation of the cochlea would be the BC sound at frequencies above approximately
4 kHz. This is in line with other estimations of its 2 kHz [41]. Moreover, the transmission through
importance in the normal ear of the human [21]. the skin improves with the area of the skin trans-
ducer interface and static pressure between trans-
ducer and skull [14]. If the vibration transducer
Bone Conduction Sound Transmission in the is coupled directly to the skull bone, as with the
Skull bone-anchored hearing aid [42] coupling system,
the attenuation caused by the skin is avoided.
The human cranial bone comprises dense corti- For hearing testing, the attenuation of BC sound
cal bone with fluid-filled spondaic bone in be- through the skin can often be overcome by an in-
tween, loaded on one side with skin and subcu- crease in the output from the audiometer. For BC
taneous tissue and fluid (cerebrospinal fluid) and hearing aids, where the output is limited, the skin
brain tissue on the other. Also, the bone in the attenuation is of greater importance. Since the at-
skull base, where the inner ear is situated, differs tenuation depends on both the BC hearing aid de-
from the cranial bone with thicker and denser sign and the skin and skull impedance, a specific
bone structure. To complicate things even more, number for this skin attenuation cannot be given.
the skull is not a single bone structure but con- But, a general 5- to 15-dB improvement in sensi-
sists of several parts fused with sutures. Due to tivity at frequencies above 1 kHz can be expected
its complexity, modeling efforts have so far only when the BC transducer is attached directly to the
provided limited insight to the BC sound trans- skull (as with the bone-anchored hearing aid) in-
mission in the skull. stead of the compressed skin.
Most earlier experimental studies of BC sound A general finding is that the closer the stimu-
transmission focused on the motion of the skull lation position is to the cochlea, the greater the
203.64.11.45 - 2/18/2015 11:05:12 PM

Acoustic and Physiologic Aspects of Bone Conduction Hearing 15

Kainan University
Downloaded by:
Fig. 2. Illustration of the vibration
modes of the human skull for fre-
quencies between 0.1 and 10 kHz.
For simplicity, the vibrations are
shown as one-dimensional mo-
tions, while the real human skull
shows vibration responses in all
space dimensions. The thick arrows
indicate the stimulation position,
a b
and the thin arrows indicate the re-
sponse directions. The rigid body
response at the lowest frequencies
is illustrated in a, while the mass-
spring system response at frequen-
cies between approximately 0.3 and
1.0 kHz is shown in b, where three
sections of the skull move sequen-
tially in opposite directions. c The
vibration responses for frequencies
above 2 kHz are illustrated differ-
ently for the skull base and the cra-
nial vault. At the skull base, longitu-
dinal wave propagation dominates
the response, while a mixture of
vibration modes including bending c
waves is present at the cranial vault.

BC sound stimulation of the cochlea. This is of the skull mechanical point impedance (150–
frequency dependent, and the greatest improve- 400 Hz), the skull behavior can be approximat-
ment is seen at the highest frequencies [38, 40]. ed with rigid body motion (fig. 2a) [38]. Above
In a study investigating the squamosal suture’s this region and up to about 1 kHz, where the first
influence on BC sound transmission, an aver- global skull resonance appears [35], the skull vi-
age attenuation of about 2 dB was found for fre- bration can be described as a mass-spring system,
quencies above 2 kHz [40]. However, it could where large parts of the skull move in phase (e.g.
not be concluded if the attenuation was a pure the petrous part of the temporal bone hosting the
effect of the suture, of moving the stimulation cochlea; fig. 2b). Above 1 kHz, wave transmission
position closer to the cochlea, or a combination appears in the cranial bone, and between 1 and
of both. 2 kHz the skull transitions from a mass-spring-
The vibration modes of the skull have been re- like behavior to be dominated by wave transmis-
ported in several studies [e.g. 19, 38], and for the sion. The wave transmission types differ between
frequency range 0.1–10 kHz four different types the cranial vault and the skull base (fig. 2c), where
of vibration modes appears (fig. 2). At the very the wave speed in the skull base seems constant
low frequencies, below the resonance frequency at approximately 400 m/s, while the wave speed
203.64.11.45 - 2/18/2015 11:05:12 PM

16 Stenfelt
Kainan University
Downloaded by:
in the cranial vault is frequency dependent with
around 250 m/s at 2 kHz increasing to 300 m/s at 20

Transcranial transmission (dB)

10 kHz [38].
10

0
Transcranial Transmission
–10
The BC stimulation at one mastoid is transmit-
–20
ted to both cochleae, and for testing BC thresh-
olds the non-test ear requires masking to ensure –30
the response from the test ear. For simplicity, in
0.2 0.4 0.7 1 2 4 7 10
audiometry the BC sound transmission to the Frequency (kHz)
contralateral cochlea is often assumed equal to
that of the ipsilateral cochlea (transcranial trans-
Fig. 3. Transcranial transmission of BC sound measured
mission equals 0 dB). However, there can, at cer-
as the ipsilateral sensitivity related to the contralateral
tain frequencies, be substantial interaural differ- sensitivity when stimulation is on the mastoid (normal
ences of up to 40 dB, where the sound is either BC audiometry position). A positive number indicates
greatest at the ipsilateral or the contralateral co- better transmission from the contralateral side. The dots
chlea [43]. When BC sound transmission is es- indicate individual results (a larger dot indicates several
individuals with equal result). The thick solid line (in the
timated from vibration measurements at the co- middle) is the median of all subjects, and the thinner solid
chlea, the transcranial transmission is close to 0 lines are the median ± 1 standard deviation.
dB for frequencies up to 1 kHz, where it starts to
decrease and becomes –15 to –20 dB at 10 kHz
[38, 39].
We measured the transcranial transmission in
30 unilateral deaf subjects using both the normal transmission (or the opposite, transcranial at-
mastoid stimulation position for BC audiometry tenuation) is measured as the difference in sen-
and the typical BAHA implant position some 55 sitivity between a contralateral and ipsilateral
mm behind the ear canal opening in the parietal stimulation at symmetrical positions, it is the ip-
bone. Bone-conducted hearing thresholds were silateral transmission that dominates the interau-
obtained at 31 1/6th octave frequencies between ral difference.
250 Hz and 8 kHz. From these data, it could be From the perspective of BC hearing aids, a
concluded that there is a general trend of greater low transcranial transmission (i.e. the stimu-
interaural differences at the higher frequencies lation of the ipsilateral cochlea is substantially
(fig. 3). However, the individual spread was great, greater than the contralateral cochlea) is ben-
and no general trends in the specific configura- eficial when bilateral application is used since
tion of the transcranial transmission were found. more binaural information can be extracted
It was also seen that a normal audiometric posi- from the two stimulation positions (less cross-
tioning of the BC transducers gave greater inter- over transmission, see below). However, from
aural differences than a positioning at the usual the perspective of using BC hearing aids for uni-
BAHA implant site. According to Eeg-Olofsson lateral deaf subjects, a high transcranial trans-
et al. [39], the transcranial transmission is al- mission is beneficial as more BC sound energy
most independent of the position of the contral- is transmitted from the deaf side to the healthy
ateral stimulation transducer. When transcranial cochlea.
203.64.11.45 - 2/18/2015 11:05:12 PM

Acoustic and Physiologic Aspects of Bone Conduction Hearing 17

Kainan University
Downloaded by:
Binaural Bone Conduction Hearing masking release, the stationary sinusoids from the
two sources add constructively or destructively
Binaural auditory processing are functions that depending on the signal phases, and the results
are superior by the use of two ears rather than one are influenced by the summation of two signals
including the ability to localize sound sources in as well as of binaural effects: these cannot be sep-
three dimensions and identify speech in a noisy arated [39]. Consequently, the binaural masking
environment. The general understanding is that level differences test using stationary sinusoids is
the human uses interaural temporal differences not suited for testing binaural processing with BC
at frequencies below 1 kHz and interaural level stimulation.
differences at the higher frequencies for efficient
binaural processing. Both the temporal and level
differences can be negatively affected by the BC Own Voice
transcranial transmission reducing the informa-
tion separation between the cochleae. The low Most people are familiar with the influence of BC
temporal differences between the cochleae at low on their own voice: when listening to a record-
frequencies with BC sound shown in Stenfelt and ing of one’s own voice, people are often struck by
Goode [38] impede the low-frequency binaural the difference between the recording and the way
processing. At the high frequencies, the tran- they normally perceive their own voice. The dif-
scranial transmission may limit the level sepa- ference is that while the recording only picks up
ration normally given by the head shadow; this the airborne sound, we hear ourselves through
could also impede binaural processing. So how both AC and BC sound (see fig. 1). The impor-
efficient is bilateral application of BC sound in tance of the two pathways of our own voice, AC
terms of binaural hearing, for example with BC and BC, have been estimated by von Békésy [48]
hearing aids? to be approximately equal; it was later estimated
It should be noted that the following discus- that the AC component is most important at low
sion is limited to two rather similarly functioning and high frequencies, while the BC component
cochleae; binaural auditory processing is not pos- dominated the own voice between 0.7 and 1.2
sible for unilateral profound deaf subjects. The re- kHz [49]. A recent thorough investigation showed
search into BC binaural hearing is sparse and has large differences in the AC and BC components
mainly been directed to binaural benefit in pa- for 10 speech sounds [50]. It was concluded that
tients with bilaterally fitted BAHA [44, 45]. Those the differences originated in the sound produc-
studies have shown that there are benefits in terms tion itself; sounds that are produced similarly and
of better spatial perception (sound localization) belong to the same speech sound groups had sim-
and release of masking for bilateral BAHA users ilar relation between AC and BC contribution of
compared with monaural application. Also, stud- one’s own voice. Even if later studies confirmed
ies in normal hearing subjects have indicated sim- von Békésy’s result of approximately equal pro-
ilar results: there are binaural benefits with bilat- portion of AC and BC sound for own voice per-
eral BC stimulation, but the binaural processing is ception, they also showed frequency dependency
not as good as when then the stimulation is pure- between the pathways as a function of the sound
ly by AC [46, 47]. One way to quantify the bin- produced [49, 50].
aural processing ability with BC stimulation has It has been argued, for example by Bárány
been to measure binaural masking level differ- [51], that own-made sounds such as vocaliza-
ences using low-frequency sinusoids in narrow- tion, chewing, etc., are less in animals with three
band masking noise [44, 45]. Besides the binaural middle-ear ossicles than in animals with only one
203.64.11.45 - 2/18/2015 11:05:12 PM

18 Stenfelt
Kainan University
Downloaded by:
bone. Due to the three-bone structure, the effec- Consequently, to achieve more than 60 dB atten-
tive mass, and consequently the inertial force, are uation, the BC sound transmission needs to be
less than in a single-bone structure. As a result, attenuated as well. This can be accomplished by
the influence of the middle ear ossicle inertia is helmets or body covers.
lower. Also, the stapedius muscle is elicited dur-
ing vocalization [52]. Such contraction affects the
low-frequency sound transmission through the Implications for Bone Conduction Hearing Aid
middle ear. It would also impede inertial force ef- Usage
fects on the middle ear ossicles. However, to what
extent the stapedius muscle contraction affects Recent understanding of BC sound and sound
the perception of BC sound and the own voice is transmission in the skull has indicated that the
currently not clear. gain in sensitivity by stimulating close to the
cochlea can overcome drawbacks of transcuta-
neous signal transmission for implantable BC
Bone Conduction in Relation to Hearing hearing aids [54]. Also, for bilateral applica-
Protection tion of BC sound a position close to the cochlea
results in greater interaural separation and
People who are exposed to high-level sounds would be beneficial for binaural processing abil-
should protect the ears with hearing protec- ities [38, 39]. However, when BC hearing aids
tion devices. In the extreme environments, the are used with patients suffering from unilateral
noise can exceed 150 dB SPL, and convention- deafness, a positioning close to the cochlea does
al hearing protection devices are no longer suf- not improve the transcranial transmission, but
ficient. Even if earplugs and earmuffs attenuate does not decrease it either. It should be noted
the AC pathway of sound to the cochlea, sound here that bilateral application of BC hearing aids
is reaching the cochlea by BC transmission. The does not only enable (reduced) binaural hear-
head and body in a sound field is affected by the ing, it also gives hearing from the non-hearing
sound field, and BC sound is induced and trans- side (in a monaural fitting) and thereby removes
mitted to the cochlea. This sound field-induced the head-shadow effect – the primary reason for
BC sound limits the possible attenuation by con- giving a BC hearing aid in unilateral deafness
ventional earmuffs and earplugs. According to [55]. Consequently, giving a second BC hearing
Reinfeldt et al. [53], the BC transmission is 50– aid to a bilateral conductive-impaired patient
60 dB below the AC sensitivity at frequencies be- provides greater benefit than giving a BC hearing
low 1 kHz, 40–50 dB at 2 kHz and 50–60 dB be- aid to a unilateral deaf patient since no binaural
low the AC sensitivity at the higher frequencies. processing is possible for the latter.

References
1 Ng M, Jackler R: Early history of tuning- 3 Purcell D, Kunov H, Madsen P, Cleghorn 4 Beattie R: Normative Wave V latency-
fork tests. Am J Otolaryngol 1993;14: W: Distortion product otoacoustic intensity functions using the EARTONE
100–105. emissions stimulated through bone 3A insert earphone and the Radioear
2 Valentine M, Stark E: Bone conducted conduction. Ear Hear 1998;19:362–370. B-71 bone vibrator. Scand Audiol 1998;
speech audiometry with normal-hearing 27:120–126.
and hearing-impaired children. J Aud
Res 1977;17:105–108.
203.64.11.45 - 2/18/2015 11:05:12 PM

Acoustic and Physiologic Aspects of Bone Conduction Hearing 19

Kainan University
Downloaded by:
 5 Stenfelt S, Reinfeldt S: A model of the 20 Tonndorf J: Bone conduction. Studies in 34 Yoshida M, Uemura T: Transmission of
occlusion effect with bone-conducted experimental animals. Acta Otolaryngol cerebrospinal fluid pressure changes to
stimulation. Int J Audiol 2007;46: 1966;(suppl 213):1–132. the inner ear and its effect on cochlear
595–608. 21 Stenfelt S, Goode RL: Bone conducted microphonics. Eur Arch Otorhinolaryn-
6 International Organization for Standard- sound: physiological and clinical aspects. gol 1991;248:139–143.
ization: Acoustics – reference zero for Otol Neurotol 2005;26:1245–1261. 35 Håkansson B, Brandt A, Carlsson P,
the calibration of audiometric 22 Stenfelt S, Wild T, Hato N, Goode RL: Tjellström A: Resonance frequency of
equipment. 3. Reference equivalent Factors contributing to bone conduc- the human skull in vivo. J Acoust Soc
threshold force levels for pure tones and tion: the outer ear. J Acoust Soc Am Am 1994;95:1474–1481.
bone vibrators. Geneva, International 2003;113:902–912. 36 Håkansson B, Carlsson P, Brandt A,
Organization for Standardization, 1994, 23 von Békésy G: Uber die Schallausbrei- Stenfelt S: Linearity of sound transmis-
ISO 389-3. tung bei Knochenleitung. Z Hals Nasen sion through the human skull in vivo.
7 Popelka G, Telukuntla G, Puria S: Ohrenheilk 1941;47:430–442. J Acoust Soc Am 1996;99:2239–2243.
Middle-ear function at high frequencies 24 Stenfelt S, Hato N, Goode RL: Factors 37 Stenfelt S, Håkansson B, Tjellström A:
quantified with advanced bone- contributing to bone conduction: the Vibration characteristics of bone
conduction measures. Hear Res 2010; middle ear. J Acoust Soc Am 2002;111: conducted sound in vitro. J Acoust Soc
263:85–92. 947–959. Am 2000;107:422–431.
8 Håkansson B: The balanced electromag- 25 Homma K, Du Y, Shimizu Y, Puria S: 38 Stenfelt S, Goode RL: Transmission
netic separation transducer: a new bone Ossicular resonance modes of the properties of bone conducted sound:
conduction transducer. J Acoust Soc Am human middle ear for bone and air con- measurements in cadaver heads.
2003;113:818–825. duction. J Acoust Soc Am 2009;125: J Acoust Soc Am 2005;118:2373–2391.
9 Weece R, Allen J: A method for calibra- 968–979. 39 Eeg-Olofsson M, Stenfelt S, Gran-
tion of bone driver transducers to 26 Hato N, Stenfelt S, Goode RL: Three- ström G: Implications for contralateral
measure the mastoid impedance. Hear dimensional stapes footplate motion in bone conducted transmission as mea-
Res 2010;263:216–223. human temporal bones. Audiol Neuroo- sured by cochlear vibrations. Otol Neu-
10 Richter U, Brinkmann K: Threshold of tol 2003;8:140–152. rotol 2010, Epub ahead of print.
hearing by bone conduction. Scand 27 Carhart R: Effects of stapes fixation on 40 Eeg-Olofsson M, Stenfelt S, Tjellström A,
Audiol 1981;10:235–237. bone-conduction response; in Ventry I, Granström G: Transmission of
11 Minor LB: Superior canal dehiscence Chaiklin J, Dixon R (eds): Hearing Mea- bone-conducted sound in the human
syndrome. Am J Otol 2000;21:9–19. surement: A book of Readings. New skull measured by cochlear vibrations.
12 Stenfelt S, Håkansson B: Air versus bone York, Appleton-Century-Crofts, 1971, pp Int J Audiol 2008;47:761–769.
conduction: an equal loudness investiga- 116–129. 41 Stenfelt S, Håkansson B: Sensitivity to
tion. Hear Res 2002;167:1–12. 28 Huizing EH: Bone conduction – the bone-conducted sound: excitation of the
13 von Békésy G: Zur theorie des hörens bei influence of the middle ear. Acta Otolar- mastoid vs the teeth. Scand Audiol 1999;
der schallaufnahme durch knochenlei- yngol 1960;suppl 155:1–99. 28:190–198.
tung. Ann Physik 1932;13:111–136. 29 Stenfelt S: Middle ear ossicles motion at 42 Snik AF, Mylanus EAM, Proops DW,
14 Khanna SM, Tonndorf J, Queller J: hearing thresholds with air conduction Wolfaardt JF, Hodgetts WE, Somers T,
Mechanical parameters of hearing by and bone conduction stimulation. Niparko JK, Wazen JJ, Sterkers O, Crem-
bone conduction. J Acoust Soc Am 1976; J Acoust Soc Am 2006;119:2848–2858. ers CWRJ, Tjellström A: Consensus
60:139–154. 30 Ranke O, Keidel W, Weschke H: Das statements on the Baha system: where
15 Stenfelt S: Simultaneous cancellation of Hören beim Verschluss des runden Fen- do we stand at present? Ann Otol Rhinol
air and bone conduction tones at two sters. Zeit Laryngol 1952;31:467–475. Laryngol 2005;114(suppl 195):1–12.
frequencies: extension of the famous 31 Kucuk B, Abe K, Ushiki T: Micro- 43 Nolan M, Lyon DJ: Transcranial attenua-
experiment by von Békésy. Hear Res structures of the bony modiolus in the tion in bone conduction audiometry.
2007;225:105–116. human cochlea: scanning electron J Laryngol Otol 1981;95:597–608.
16 Stenfelt S, Puria S, Hato N, Goode RL: microscopic study. J Electron Microsc 44 Privin C, Stenfelt S, Granström G,
Basilar membrane and osseous spiral 1991;40:193–197. Tjellström A, Håkansson B: Bilateral
lamina motion in human cadavers with 32 Songer J, Rosowski J: A superior semicir- bone-anchored hearing aids (Bahas):
air and bone conduction stimuli. Hear cular canal dehiscence-induced air-bone an audiometric evaluation. Laryngo-
Res 2003;181:131–143. gap in chinchilla. Hear Res 2010;269: scope 2004;114:77–84.
17 Mauldin L, Jerger J: Auditory brain stem 70–80. 45 Bosman A, Snik A, van der Pouw C,
evoked responses to bone-conducted 33 Sohmer H, Freeman S, Geal-Dor M, Mylanus E, Cramers C: Audiometric
signals. Arch Otolaryngol 1979;105: Adelman C, Savion I: Bone conduction evaluation of bilaterally fitted bone-
656–661. experiments in humans – a fluid path- anchored hearing aids. Audiology
18 von Békésy G: Paradoxical direction of way from bone to ear. Hear Res 2000; 2001;40:158–167.
wave travel along the cochlear partition. 146:81–88. 46 Rowan D, Gray M: Lateralization of
J Acoust Soc Am 1955;27:137–145. high-frequency pure tones with inter-
19 von Békésy G: Experiments in Hearing. aural phase difference and bone conduc-
New York, McGraw-Hill, 1960. tion. Int J Audiol 2008;47:404–411.
203.64.11.45 - 2/18/2015 11:05:12 PM

20 Stenfelt
Kainan University
Downloaded by:
47 MacDonald J, Henry P, Letowski T: 50 Reinfeldt S, Ostli P, Håkansson B, Stenfelt 54 Håkansson B, Reinfeldt S, Eeg-Olofsson
Spatial audio through a bone conduction S: Hearing one’s own voice during pho- M, Ostli P, Taghavi H, Adler J,
interface. Int J Audiol 2006;45:595–599. neme vocalization – transmission by air Gabrielsson J, Stenfelt S, Granström G:
48 von Békésy G: The structure of the and bone conduction. J Acoust Soc Am A novel bone conduction implant (BCI):
middle ear and the hearing of one’s own 2010;128:751–762. engineering aspects and pre-clinical
voice by bone conduction. J Acoust Soc 51 Bárány E: A contribution to the studies. Int J Audiol 2010;49:203–215.
Am 1949;21:217–232. physiology of bone conduction. Acta 55 Stenfelt S: Bilateral fitting of Bahas and
49 Pörschmann C: Influences of bone Otolaryngol 1938;(suppl 26):1–223. Baha fitted in unilateral deaf persons:
conduction and air conduction on the 52 Borg E, Zakrisson J: The activity of the acoustical aspects. Int J Audiol 2005;44:
sound of one’s own voice. Acustica 2000; stapedius muscle in man during 178–189.
86:1038–1045. vocalization. Acta Otolaryngol 1975;79:
325–333.
53 Reinfeldt S, Stenfelt S, Good T,
Håkansson B: Examination of bone-
conducted transmission from sound
field excitation measured by thresholds,
ear-canal sound pressure, and skull
vibrations. J Acoust Soc Am 2007;121:
1576–1587.

Stefan Stenfelt
Department of Clinical and Experimental Medicine
Technical Audiology, Linköping University
SE–58185 Linköping (Sweden)
Tel. +46 10 1032856, E-Mail [email protected]
203.64.11.45 - 2/18/2015 11:05:12 PM

Acoustic and Physiologic Aspects of Bone Conduction Hearing 21

Kainan University
Downloaded by: