ELECTRONIC DEVICES FOR RECONSTRUCTION OF

HEARING

Albrecht Eiber

Institute of Engineering and Computational Mechanics, University of Stuttgart, Pfaffenwaldring 9, Stuttgart, Germany

Keywords: Hearing devices, mechanical models, sound transfer, feedback, distortion.

Abstract: The effect of specific hearing impairments can be alleviated or compensated using electrically driven

hearing aids. There is a broad variety of devices for stimulating the hearing either acoustically, mechanically

or electrically. Their applications depend on the type, the severity and the location of a particular

impairment. In order to get an optimal reconstruction, additionally to the medical aspects the dynamical

behavior of the implant has to be regarded together with the individual situation of the patient

simultaneously. Because of the broad variation of individuals and situations, the belonging parameters are

time dependent and vary in a wide range, too. Thus, the design of a hearing device must be robust or

insensitive against parameter variation and a sensitivity analysis with respect to parameters of the device,

variation of anatomy and variation of insertion is needed. By means of mechanical simulation models

enhanced by the behavior of the actuator and its control, the dynamical behavior of implants can be

calculated and optimized. Such simulations helps to shorten series of experiments in the lab and in clinical

practice and guidelines for the designer and for the surgeon can be derived.

1 INTRODUCTION

Hearing is a highly dynamical process, a sound

event expressed as fluctuation of pressure in the air

acts at the ear drum and excites the three tiny

ossicles malleus, incus and stapes to vibrate. Due to

the coupling of the stapes footplate, these motions

are transmitted to the fluid of the inner ear bringing

the basilar membrane in motion. The stereocilia of

the inner hair cells are bended and electrical spikes

are transferred to the brain by the nerves. For low

excitation, the outer hair cells of the organ of Corti

are activated serving as an amplifier.

Impairments may have various reasons:

mechanical defects in the middle ear like missing

ossicles, stiffened ligaments or damaged hair cells in

the inner ear. In order to get an appropriate

reconstruction, the surgeon must know the

mechanical behavior of the available implants, their

function and use as well as their dynamical behavior

when they are inserted in the body.

2 RECONSTRUCTION

Depending on the specific impairment, several

possibilities of reconstruction exist. Conventional

devices act at the outer ear canal driving the natural

ear acoustically. A mechanical reconstruction may

be a simple replacement of a natural structure or a

directly driven element of the hearing like an ossicle

of the middle ear or the skull. Moreover, a direct

electrical stimulation in the inner ear is possible.

Implants replacing missing ossicles are passive

mechanical elements. Actively driven devices

consist of a microphone, a sound processor, an

amplifier, a power source and an actuator.

Conventional devices produce an amplified sound

pressure applied to the outer ear canal, they can be

used in case of intact hearing with moderate hearing

loss.

Other implants are imposing mechanical

vibrations directly to the middle ear ossicles (middle

ear implants, MEI) or to the skull for transferring the

sound by structural vibrations to the inner ear (bone

anchored hearing aids, BAHA). These implants can

be used in case of a sensorineural defect in the inner

ear.

304

Eiber A. (2008).

ELECTRONIC DEVICES FOR RECONSTRUCTION OF HEARING.

In Proceedings of the First International Conference on Biomedical Electronics and Devices, pages 304-309

DOI: 10.5220/0001055503040309

 SciTePress

Another possibility is to stimulate the nerves of

the inner ear electrically via an electrode inserted

into the cochlea (cochlear implant, CI) or acting at

the brainstem. Such implants work even when the

hair cells are inactive or destroyed.

3 ACTIVE MIDDLE EAR

IMPLANTS

Depending on the degree of integration into the

body, totally and partially implantable hearing aids

are distinguished. Further categories are due to the

driving principle, the transfer of actuation to the

hearing organ, the point of actuation, the manner of

coupling to the driven ossicle, the type of suspension

to the base and of the location the elements between

the force is acting.

3.1 Driving Concept

Two different principles of driving are common:

magnetic coils like in classical loudspeakers and

piezoelectric elements. Such elements may work in

diverse modes of operation like bending effect of

beams or discs as well as elongation of staples.

3.2 Force/Displacement Transmission

Due to the transfer of force or displacement from the

actuator to the ossicles, the devices can be

categorized in groups working with an acoustical

coupling via air cushion, hydraulical coupling via

fluid in a tube and mechanical coupling.

Principally, an actuator can be attached at an

ossicle transferring its forces due to inertial effects

of a moved internal mass. This type is called floating

mass transducer (FMT) (Hong et al, 2007). It needs

a proper coupling to an ossicle.

If such type of actuator is attached to the skull,

the sound information is transferred to the inner ear

via structural vibrations of the skull and the device is

called bone anchored hearing aid (BAHA). It needs

a fixed coupling which is maintained by an osseo-

integrated screw (Tjellstroem, 1990).

Other actuators are placed between ossicle and

temporal bone, they may act even as force

transducer or as displacement transducer and need

an appropriate coupling at both ends.

3.3 Mechanical Coupling

One of the most important issue in reconstruction is

the mechanical coupling of the actuator to the

ossicle or/and skull because it governs significantly

the dynamical behavior of the reconstructed ear. It is

important to which ossicle and at which location the

actuator is coupled, in which spatial direction the

actuation takes place and how the coupling is

maintained.

Coupling of implants to the skull is commonly

achieved by means of screws in a mechanical ideal

way. More problematic is the connection to an

ossicle, it is achieved by gluing, crimping or

clamping. In the latter cases, mucosa, fascia or

cartilage is present as an intermediate layer between

the implant and the ossicle.

The mechanical description of such coupling

elements leads to nonlinear force/displacement laws

similar to the description of the ligaments or joints.

Generally, the relative displacements in the coupling

area are smaller than in the joints, but depending on

the specific design of coupling, pronounced

asymmetry in push-pull direction up to unilateral

coupling is possible. Moreover, kinks and

discontinuities may occur, leading to distorted sound

transfer.

4 MODELS AND MECHANICAL

BEHAVIOR

Classical models of the hearing organ have been

developed by acoustical or electrical engineers based

n electrical analogies. These circuit models are

mostly one directional described by scalar

differential equations (e.g. Zwislocki, 1963) and it is

difficult to describe the three dimensional nature of

ossicles motion (Hudde and Weistenhoefer, 1997).

Moreover, the parameters of such models are not

directly related to the geometrical dimensions or

physical quantities like mass and stiffness.

Due to the anatomy of the hearing organ and

the function based on the relation between forces

and displacements, three-dimensional mechanical

models are adequate for description and simulation,

well developed modelling techniques like Multibody

Systems approach (MBS) or Finite-Element-Method

(FEM) are available (Wada et al, 1992; Beer et al,

1999; Gan et al, 2004; Prendergast el al, 1999; Eiber

et al, 1999).

The mechanical considerations presented here

take into account the spatial arrangement of the

ELECTRONIC DEVICES FOR RECONSTRUCTION OF HEARING

305

ossicles and their three dimensional motions in a

vectorial description. They allow to regard the

spatially oriented direction of excitation, the

mechanical properties of the coupling of implants

and the nonlinear behavior of the coupling region, of

the joints and of the ligaments.

In Figure 1, a typical Multibody System model

is shown with its rigid bodies malleus, incus and

stapes. The ear drum, the air in the outer ear canal

and the fluid of the inner ear are modelled as lumped

mass points or bodies. The ligaments and the two

muscles tensor tympani and musculus stapedius are

described as visco-elastic elements with an active

part. The model has 83 degrees of freedom and the

vector z contains the generalized coordinates. The

crucial ones are indicated in the Figure.

Figure 1: Multibody System model of the middle ear with

its adjacent structures and an active middle ear implant

acting at the incus.

The dynamical behaviour of the system is

described by the equation of motion

),,(),()( tzzqzzkzzM

&&&&

+⋅

(1)

where M denotes the mass matrix, k the vector of

generalized forces, q the vector of applied

generalized forces. Generally, such models are

highly nonlinear due to nonlinear coupling elements,

i.e. ligaments and joints of ossicles or due to large

amplitudes of ossicles´ motion (Eiber and

Breuninger, 2004 a). The latter is not the case in the

physiological range of hearing (Eiber and

Breuninger, 2004 b). Figure 2 shows some typical

force/displacement relations of ligaments and joints

of the middle ear. Typically, coupling of implants

are similar to that of the incudo-stapedial joint with

asymmetry and kinks causing a distorted sound

transmission.

ear drum incudo -

malleolar joint

incudo -

stapedial joint

annular ring

Figure 2: Nonlinear characteristics of middle ear elements.

Static preloads acting at the ossicular chain deflect it

to a specific working position y

, where the small

physiological sound pressure variations are

superposed.

For the investigation of these small variations,

the equation of motion (1) can be linearized with

respect to y

(

)

(

)()

hyyKyyDyyM =⋅

⋅

WPWPWP

&&&

(2)

where M denotes the mass matrix, D the damping

matrix, K the stiffness matrix and h the vector of

applied forces. Thus, the dynamical behavior of the

chain depends on the working position and shows

changed natural frequencies. Such static prestress

may occur due to static pressure differences in the

outer ear canal, in the middle ear cavity or in the

inner ear, due to scar tissues of healing structures or

preloads from inserted implants. Particularly active

implants with a driving rod acting at the incus need

such a preload.

After linearization, the linear equation of

motion (2) is valid and frequency domain procedures

can be applied to investigate the dynamical behavior

represented by the linearized vector y. For specific

nonlinearities like unilateral constraints or kinks in

the force/displacement relation as illustrated in

Figure 2, linearization is not possible and there is a

distorted sound transfer even for small sound

pressure variations.

By means of time-integration the time history

of specific stimulations of hearing, in particular of

transient sound events, can be investigated based on

the equations of motion of the entire system

consisting of ossicular chain and the actuator with its

control elements.

xΔ

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

306

5 CASE STUDIES

Two different types of hearing aids driven by

magnetic coils and piezo elements are considered for

the case of totally implantable device. They are very

compact but the surgical effort for insertion is very

high. In case of partially implantable device the core

respectively the magnet are arranged in different

positions leading to a simpler surgery process.

5.1 Totally Implantable Hearing Aid

A magnetically driven device on the market is the

Otologics Middle Ear Transducer MET (OTOlogics,

LLC). All components even the microphone are

placed in a cavity of skull behind the ear. The

battery is inductively charged once per day. In

particular, the actuator is mounted in the mastoid

behind the ear canal, it pushes against the ossicular

chain by means of a driving rod (Waldmann et al,

2004).

Because of the unphysiological orientation of

the actuator, the motion imposed to the stapes

contains much higher components of rocking of

stapes around its short and the long axis of footplate

when compared to the natural hearing. Recent

studies revealed the influence of such motions on

hearing (Sequeira, 2007).

Coupling of the driving rod with the incus is

accomplished by pressing the rod against the ossicle.

In order to have a fixed attachment point, a notch

has to be burred using a drill or burned using a laser.

Mechanically, such a contact is nearly an unilateral

constraint, which needs a reasonable preload leading

to a pre-stressed ossicular chain. Due to the shifted

working point, the middle ear shows higher natural

frequencies but in particular, higher amplitudes of

the actuator are necessary to achieve the vibrations

of stapes which are demanded for compensation of a

hearing loss.

As a negative consequence, the malleus and the

ear drum are driven with larger amplitudes, too,

radiating sound into the outer ear canal. Picked up

by the microphone, this sound may lead to ringing

effects because of a loop. Such effects have been

observed particularly on implants driven by

piezoelectric actuators.

Thus, depending on nonlinear behavior of the

chain and the coupling, even powerful actuators are

not able to compensate a severe hearing loss.

For optimizing a reconstruction, the entire

system of ossicular chain, the particular impairment,

the characteristics of implant and its coupling to the

chain has to be regarded. A very sensitive entity in

the function is the adjustment of the actuator with

respect to the skull during surgery, which defines the

applied preload (Rodriguez Jorge et al, 2006), (Eiber

et al, 2007).

In Figure 3, the mechanical model of the

considered actuator is sketched showing the

adjustment of the device relative to the skull and the

inner stiffness between the mass of housing frame

and the driven mass. For a magnetically driven

actuator, this stiffness is relatively low in

comparison to the piezo actuator.

Figure 3: Model of actuator attached at the ossicle and the

skull.

Concerning the two driving principles there are

significant differences. The piezo elements can

produce high forces imposing prescribed

displacements to the chain. But due to its high inner

stiffness, such devices show a high sensitivity

against the adjustment of preload given by the

surgeon during insertion and also against static

pressure variations during daily use.

Therefore, a relatively high preload is

commonly applied by the surgeon, which leads to

extremely shifted working points with the effects of

feedback and distortion as described above. On the

other hand, ligaments exhibit a time dependent

mechanical behavior. In a long time range they

elongate under static load in a creeping effect

releasing the imposed preload.

Due to the lower internal stiffness, the implants

driven by magnetic coils are less sensitive against

variations in static pressure or preload. But even in

this case, the sensitivity against adjustment travel as

shown in Figure 3 remains and is crucial for the

surgeon. A high preload stiffens the chain and

requires higher force amplitudes with a significant

sound radiation by the ear drum as a consequence.

Modelling this feed back loop allows a stability

analysis and the calculation of a gain margin. It

defined as the maximal allowed amplification so that

the system is still stable. This amplification can be

used to compensate a hearing impairment of the

inner ear. In Figure 4 this gain margin is shown

qualitatively for various adjustment travel a

and it

suspension

housing

driving

rod

inner

stiffness

active

element

coupling

stiffness

matrix

adjustment

ossicle

ELECTRONIC DEVICES FOR RECONSTRUCTION OF HEARING

307

becomes clear, that there is a trade-off between a

good coupling and the risk of feedback particularly

in the lower frequency range. For that calculation,

the microphone was placed in the mastoid at the

posterior wall of the outer ear canal.

Figure 4: Gain margin for different travels of adjustment

screw. Increasing adjustment cause higher static preload.

A negative gain margin signifies risk of ringing.

Totally implantable hearing aids are a very

elegant reconstruction because there is no obvious

stigmatization of the patient. They are technically

sophisticated devices, their insertion is challenging

for the surgeon and needs some experience.

5.2 Partially Implantable Hearing Aid

In order to reduce the surgical effort, parts of the

active hearing aid are placed outside the skull like in

conventional devices. Another concept is to separate

the coil of actuator from the core.

In a current project, the placement of a

permanent magnet at the manubrium driven by a

magnetic coil placed behind the ear is investigated

as illustrated in Figure 5.

That concept ensures the kinematical

unconstrained motion of the ossicular chain free of

static preload. The chain is able to find its natural

working position and the actuator can work at this

particular position.

Figure 5: Partially implantable hearing aid with separated

coil and magnet. Force f and torque l imposed to the

magnet.

Depending on the orientation of the magnetic

field and the direction of magnetization of the

permanent magnet, there is a moment superposed to

the force acting at the magnet to drive the ossicular

chain in a quite physiological way. The position and

the mass of the magnet influence both the dynamical

behavior of the chain and the applied force effects.

Based on mechanical models of the chain together

with Finite Element models of the electrical part, an

optimization of the components could be

accomplished. First tests in the lab show an easy and

safe procedure to couple the magnet at the malleus

handle using an elastic clip. This is a mechanically

stable coupling, which guaranties a perfect transfer

of forces from the magnet to the ossicle.

6 CONCLUSIONS

Mechanical models serve as a base for simulating

the hearing process and for the design of active

hearing implants. Three-dimensional models are

able to describe the complex spatial motion of the

ossicular chain and their relation to the forces and

moments applied by an implanted hearing aid. A

crucial point is the coupling of implants to the

natural structures of the ear, which is governed by

nonlinear force/displacement relations. As a

consequence, the transmitted forces are limited

leading to an incomplete compensation of hearing

loss, the transfer of sound may be distorted leading

to a bad sound discrimination due to higher

overtones or the gain of amplification must be

restricted to avoid ringing due to feedback.

Due to their close relation to the anatomy,

mechanical models give a better insight into the

dynamic behavior of the middle ear than electrical

circuit models. Multibody system models with a low

number of degrees of freedom are well suited to

describe the global behavior of systems regarding

their nonlinear properties and to design the

mechanical part of hearing implants. Finite Elemente

models with a high number of degrees of freedom

are capable to describe distributed properties like

magnetic fields and mechanical stress. Simulations

with mechanical models facilitate an optimization

during design of hearing aids and may shorten the

clinical trial and error process.

coil

permanent

magnet

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

308

ACKNOWLEDGEMENTS

On this work participated Dipl.-Ing. C. Breuninger

and several students financed by the German

Research Council (DFG).

REFERENCES

Beer, H.J., Bornitz, M., Hardtke, H.J., Schmidt, R.,

Hofmann, G., Vogel, U., Zahnert, T., Huettenbrink,

K.B., 1999. Modelling of Components of the Human

Middle Ear and Simulation of Their Dynamic

Behaviour. Audiology and Neuro-Otology. 4:

156 -162.

Eiber, A., Freitag, H.-G., Burkhardt, C., Hemmert, W.,

Maassen, M., Rodriguez Jorge, J., Zenner, H.-P.,

1999. Dynamics of Middle Ear Prostheses -

Simulations and Measurements. In Audiology and

Neuro-Otology. 4: 178 - 184.

Eiber, A., Breuninger, C., 2004 a. Mechanical Problems in

Human Hearing. Proceedings in Applied Mathematics

and Mechanics, Vol. 4, No. 1, p. 51 - 54.

http://www.wiley-vch.de

Eiber, A., Breuninger, C., 2004 b. Nonlinear Properties of

the Middle Ear: Some Influences on Hearing and

Diagnosis. In: Gyo, K., Wada, H. (eds), Middle ear

Mechanics in Research and Otology, p. 19 - 28 .

World Scientific. Singapore.

Eiber, A., Breuninger, C., Rodriguez Jorge, J., Zenner, H.-

P., Maassen, M., 2007. On the Optimal Coupling of an

Implantable Hearing Aid – Measurements and

Simulations. In: Huber, A., Eiber, A. (eds), Middle ear

Mechanics in Research and Otology, p. 246 - 252.

World Scientific. Singapore.

Gan, R.Z., Feng, B., Sun, Q., 2004. Three-Dimensional

Finite Element Modelling of Human Ear for Sound

Transmission. Annals of Biomedical Engineering. 32,

(6): 847 – 859.

Hudde, H., Weistenhoefer, C., 1997. A Three-

Dimensional Circuit Model of the Middle Ear.

Acustica united with Acta Acustica. 83, (3): 535 – 549.

Hong, E.-P., Kim, M.-K., Park, I.-Y., Lee, S.-H., Choi, J.-

H., 2007. Vibration Modeling and Design of

Piezoelectric Floating Mass Transducer for

Implantable Middle Ear Hearing Devices. IEICE

Transactions on Fundamentals of Electronics,

Communications and Computer Sciences. 90,

(8):1620 - 1627.

OTOlogics LLC, Boulder.

http://www.otologics.com/

Prendergast, P.J., Ferris, P., Rice, R.J., Blayney, A.W.,

1999. Vibro-acoustic modelling of the outer and

middle ear using a finite-elemente method. Audiol

Neurootol, 4, 185 – 191.

Rodriguez Jorge, J., Pfister, M., Zenner, H.-P., Zalaman,

I., Maassen, M., 2006. In Vitro Model for

Intraoperative Adjustments in an Implantable Hearing

Aid (MET). In Laryngoscope, 116, 246 - 252.

Sequeira, D., Breuninger, C., Eiber, A., Huber, A., 2007.

The Effect of Complex Stapes Motion on the Cochlea

in Guinea Pigs. In: Huber, A., Eiber, A. (eds), Middle

ear Mechanics in Research and Otology, p. 130 - 135.

World Scientific. Singapore.

Suzuki, J. I. (ed), 1988. Middle Ear Implant: Implantable

Hearing Aids

, Karger. Basel.

Tjellstroem, A., 1990. Osseointegrated implants for

replacement of absent or defective ears. Clinics in

Plastic Surgery, 17: 355 – 365.

Wada, H., Metoki, T., Kobayashi, T., 1992. Analysis of

dynamic behavior of human middle ear using finite-

elemente-analysis. J Acoust Soc Am. 96: 3157 - 3168.

Waldmann, B., Maier, H.,Leuwer, R., 2004. Nonlinear

Properties of the Middle Ear: Some Influences on

Hearing and Diagnosis. In: Gyo, K., Wada, H. (eds),

Middle ear Mechanics in Research and Otology, p.

377-383. World Scientific. Singapore.

Zwislocki, J., 1963. Analysis of the Middle-Ear Function.

Part II: Guinea Pig Ear. J Acoust Soc Am. 35, (7),

1034 - 1040.

ELECTRONIC DEVICES FOR RECONSTRUCTION OF HEARING

309