Biomedical Engineering Reference
In-Depth Information
directly stimulated, bypassing the natural functioning of the outer, middle and inner
ear. Current is applied to one electrode mimicking the action of the basilar membrane
and the inner hair cells. In a natural cochlea, a pure tone produces neural excitation at a
specific region of the cochlea corresponding to a characteristic frequency (tonotopy)
[
9
]. In CIs however, a broad excitation is produced, mainly because the fluids in the
cochlea are highly conductive causing the charge to spread along the cochlea. This
phenomenon is commonly referred to as spread of excitation. Spread of excitation
in the cochlea causes that different electrodes in the cochlea interact with each other,
this is known as channel interaction. It has also been demonstrated that spread of
excitation might explain some of the performance variability in speech intelligibility
observed in CI users [
19
]. Current CI devices use monopolar stimulation mode, this
means that an electrode in the scala tympani is activated and current flows to an
external electrode, typically situated close to the skull or the housing of the implant.
Additionally, it has been shown that the amount of spread of excitation is influenced
by the anatomy and the conductivity of the tissues in the cochlea [
3
,
6
,
8
,
17
]. This
has been confirmed using in vivo measurements of the cochlea with an implanted
electrode, using resistive models and solving analytical equations of the 3D volume
conduction problem.
Resistance networks introduced by von Békésy (1951) as transmission line mod-
els, have been used to calculate the voltage in the scala tympani as a function of
distance from the cochlear base. It has been reported that these type of models are of
value in the estimation of current interactions but they do not provide with the res-
olution necessary to simulate individual excitation process. In order to obtain more
accurate simulations, boundary element method (BEM) and finite element method
(FEM) have been proposed. Finley et al. [
6
] were the first to present an integrated
3D neuron field model of a segment of an unrolled cochlea, using the FEM. In [
7
,
8
]
they presented a rotationally symmetric cochlear geometry to calculate neural exci-
tation patterns for different electrode configurations and stimulation patterns. In [
3
]
the model was improved using a more refined helical representation of the cochlea.
In [
10
] an FEM model based on a spiral shaped cochlea was presented which was
used to model spread of excitation with simultaneous electrode stimulation. In [
16
]
a simplified spiraled model of the human cochlea was developed from a cross sec-
tional microphotography. More recently in [
17
] a three-dimensional FEM model of
the cochlea was developed to obtain the voltage distribution at positions closer to the
site of neural stimulation. This model was used to demonstrate the way the voltage
distribution varies with the geometry of the cochlea and the electrode array.
As mentioned above, current commercial CI systems use monopolar stimula-
tion. This produces a broad voltage distribution in the cochlea causing large channel
interaction. Different stimulation modes have been proposed to try to compensate
the negative effects of such channel interaction. For example, bipolar and tripo-
lar stimulation can be used. The idea of these methods is to narrow the electrical
field and produce more place-specific neural activation in the cochlea [
1
]. In bipolar
stimulation mode two adjacent electrodes, the active and the reference electrode are
used to focus the electrical field. In tripolar stimulation mode three intracochlear
electrodes are activated simultaneously. One electrode acts as the active electrode
