A novel mechanism of cochlear excitation during simultaneous stimulation and
pressure relief through the round window
Thomas D. Weddell1, Yury M. Yarin2, Markus Drexl3, Ian J. Russell1, Stephen J.
Elliott4 and Andrei N. Lukashkin1
1 School of Pharmacy and Biomolecular Sciences, University of Brighton, Brighton,
BN2 4GJ, UK
2 Clinic of Otorhinolaryngology, Department of Medicine, Universitätsklinikum
Dresden, Fetscherstr. 74, D-01307 Dresden, Germany
3 German Vertigo Center, Ludwig-Maximilians University Munich, Marchioninistr.
15, 81377 Munich, Germany
4 Institute of Sound and Vibration Research, University of Southampton,
Southampton, SO17 1BJ, UK
Short title: round window stimulation of the cochlea
Authors for correspondence:
Andrei N. Lukashkin
e-mail: [email protected]
Stephen J. Elliott
e-mail: [email protected]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Summary
The round window membrane (RW) provides pressure relief when the cochlea is
excited by sound. Here we report measurements of cochlear function from guinea pigs
when the cochlea was stimulated at acoustic frequencies by movements of a miniature
magnet which partially occluded the RW. Maximum cochlear sensitivity,
corresponding to subnanometer magnet displacements at neural thresholds, was
observed for frequencies around 20 kHz, which is similar to that for acoustic
stimulation. Neural response latencies to acoustic and RW stimulation were similar
and taken to indicate that both means of stimulation resulted in the generation of
conventional travelling waves along the cochlear partition. It was concluded that the
relatively high impedance of the ossicles, as seen from the cochlea, enabled the region
of the RW not occluded by the magnet, to act as a pressure shunt during RW
stimulation. We propose that travelling waves, similar to those due to acoustic far-
field pressure changes, are driven by a jet-like, near-field component of a complex
fluid-pressure field, which is generated by the magnetically vibrated RW. Outcomes
of research described here are theoretical and practical design principles for the
development of new types of hearing aids, which utilise near-field, RW excitation of
the cochlea.
Key words: cochlear round window; guinea pig cochlea; cochlear excitation; active
middle ear prosthesis; implantable hearing aid, near-field excitation.
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
1. Introduction
The round window membrane (RW) acts as a pressure relief valve for the almost
incompressible fluids of the cochlea, making possible movement of the stapes and,
hence, movement of the inner ear structures. It has long been known (e.g. see Culler et
al [1]) that cochlear function is impaired when the RW membrane is immobilised,
thickened, congenitally malformed or absent [2-4]. These observations signify the
physiological importance of the RW membrane for audition, and the necessity to
retain it, which has been revealed during surgery of the middle and inner ear [5].
Traumatic damage or rupture of the RW membrane can cause hearing loss and
deafness due to perilymph aspiration or loss of the normal pressure-releasing function
of a compliant RW [6-9].
It has been established, however, that while normal function of the RW is important
for effective stimulation of the cochlea through the conventional, oval window, route,
the cochlea can be stimulated successfully in non-conventional ways (e.g. through
bone conduction, through the RW, and through perforations in the cochlea’s apical
turn). All of these techniques produce similar patterns of cochlear sensitivity and
excitation [10-12]. In recent years significant recovery of hearing thresholds has been
achieved in human patients through the use of active middle ear implants (vibrating
electro-mechanical devices) to stimulate the cochlea through the RW at the
frequencies of the incoming sound [13-17]. The RW approach could be an effective
and preferable alternative to conventional hearing aids for hearing rehabilitation in
patients with, for example, chronic inflammatory middle ear diseases, recurrent
cholesteatoma, when the anatomy of the middle ear is highly distorted [14], and for
patients with congenital malformations of the outer and middle ear, i.e. in cases of
conductive hearing loss. In the latter situation dysplasia and immobilization of the
ossicles, often in combination with malformations of the oval window, can make
hearing reconstruction through the oval window impossible. The feasibility of the RW
approach for the above disorders, and for conditions when the oval window is
inaccessible, has been demonstrated in successful attempts to combine outer and
middle ear reconstruction with implantation of active middle ear prostheses on the
RW [13, 15, 16]. An understanding of the mechanisms of the cochlear excitation
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
through the RW is essential to ensure that the technique is to be predictable and
effective in the clinic.
One uncertainty relates to the mechanisms of basilar membrane (BM) excitation
using probes that cover only a part of the RW membrane, whereby the area of the RW
not covered by the probe provides an effective pressure shunt (Békésy [11], page
197). This might make excitation of the cochlea problematic through shunting
pressure changes that would otherwise excite the BM. Mobility of the stapes is
probably not essential for successful cochlear excitation through the RW if most of
the pressure relief is provided by the non-occluded area of the RW. The paper
presented here deals with the mechanisms of cochlear excitation, using a miniature
magnet that only partially occludes the RW. With the stapes mobility decreased, a
novel form of cochlear excitation is achieved with a sensitivity that matches that of
conventional acoustic stimulation.
2. Material and methods
Pigmented guinea pigs (280-390 g) were anaesthetised with the neurolept anaesthetic
technique (0.06 mg/kg body weight atropine sulphate s.c., 30 mg/kg pentobarbitone
i.p., 500 l/kg Hypnorm i.m.). Additional injections of Hypnorm were given every 40
minutes. Additional doses of pentobarbitone were administered as needed to maintain
a non-reflexive state. The heart rate was monitored with a pair of skin electrodes
placed on both sides of the thorax. The animals were tracheotomized and artificially
respired, and their core temperature was maintained at 38C with a heating blanket
and a heated head holder. The middle ear cavity of the ear used for the measurements
was opened to reveal the RW. Compound action potentials (CAPs) of the auditory
nerve were measured from the cochlear bony ridge in the proximity of the RW
membrane using Teflon-coated silver wire (S in figure 1). Thresholds of the N1 peak
of the CAP were estimated visually using 10 ms tone stimuli at a repetition rate of 10
Hz. Latencies of the N1 peak for acoustic and RW stimulation were estimated off-line
using recording of the CAP using 50 averages.
For acoustic stimulation sound was delivered to the tympanic membrane by a closed
acoustic system comprising two Bruel and Kjaer 4134 ½” microphones for delivering
tones and a single Bruel and Kjaer 4133 ½” microphone for monitoring sound
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
pressure at the tympanum. The microphones were coupled to the ear canal via 1 cm
long, 4 mm diameter tubes to a conical speculum, the 1 mm diameter opening of
which was placed about 1 mm from the tympanum. The closed sound system was
calibrated in situ for frequencies between 1 and 50 kHz. Known sound pressure levels
were expressed in dB SPL re 210-5 Pa.
A neodymium iron boron disk magnet (M in figure 1) (diameter 0.6 mm, thickness 0.2
mm), was placed on the RW to stimulate the cochlea through the RW. The magnet
covered about one fourth of the RW surface of total area of 1.18 mm2 [18]. A
miniature coil (C in figure 1) made of two turns of copper wire (0.15 mm in diameter)
was placed above the magnet. The magnet was driven with a magnetic field created
by AC current through the coil. Stimulating current through the coil was generated by
a Data Translation 3010 board, attenuated and fed to the coil through a current buffer.
Maximum voltage applied to the coil in this study was 10 V which corresponded to 0
dB attenuation. High-frequency cut-off of the system for electrical stimulation in situ
was above 100 kHz. No signals associated with the coil current were recorded in the
absence of the magnet which confirms an absence of electrical interference between
the coil and the CAP electrode. The efficacy of the floating magnet RW stimulation
varies between preparations (see below). The floating magnet does not, however,
impose a variable DC load on the RW, which undergoes large, slow, periodic,
movements caused by middle ear muscle contraction in anaesthetised animals. This
would not be the case for probes that employ a rigid lever in contact with the RW,
which would provide a variable DC load depending on the phase of the large RW
movement, with the possibility of causing damage to the RW.
Displacements of the magnet and the stapes were measured using a displacement-
sensitive laser diode interferometer without the need for reflective beads [19]. The
beam of the interferometer was focused centrally on the exposed surface of the
magnet or the head of the stapes. The output signal of the interferometer was
processed using a signal conditioning amplifier, digitised at 250 kHz with Data
Translation 3010 board and instantaneous amplitude and phase of the wave were
recorded and averaged 10 times using a digital phase-locking algorithm.
Voltage signals to the coil were sinusoidal stimuli of 40 ms in duration (with 110 ms
between stimuli) shaped with raised cosines of 0.5 ms duration at the beginning and at
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
the end of each sinusoidal pip. All acoustic stimuli in this work were shaped with
raised cosines of 0.5 ms duration at the beginning and at the end of stimulation. White
noise for acoustical calibration and tone sequences for auditory and mechanical
stimulation were synthesised by a Data Translation 3010 board at 250 kHz and
delivered to the microphones or to the coil through low-pass filters (100 kHz cut-off
frequency). Signals from the acoustic measuring amplifier were digitised at 250 kHz
using the same board and averaged in the time domain. Experimental control, data
acquisition and data analysis were performed using a PC with programmes written in
TestPoint (CEC, MA, USA). All procedures involving animals were performed in
accordance with UK Home Office regulations with approval from the local ethics
committee.
3. Results
3.1 Pressure distribution in a model of cochlear stimulation through partially
occluded RW
The RW is assumed to be driven by a piston (magnet) over a part of its area and the
remaining part of the RW is flexible and generates an equal and opposite volume
velocity, since the fluid is assumed to be incompressible and the cochlear wall is rigid.
The internal pressure in the cochlea is then made up of two components.
The first component is an average alternating pressure, PM, which is the same
throughout the cochlea. The magnitude of this will depend on the stiffness, S, of the
freely moving area of the RW. PM is required to be large enough to force this area to
have a volume velocity equal and opposite to that of the driving piston. The force on
the freely moving part is thus Sx, where x is its mean displacement. Its volume
velocity, q, is thus iωAx, where A is the freely moving area and ω is the stimulation
frequency. Because the force on the freely moving part is equal to PM A, we can
define
PM= Sqiω A2 . (1)
Since the average pressure is uniform throughout the volume, it does not cause any
excitation of an incompressible cochlear partition. It is worth noting that if the
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
stiffness, S, of the freely moving area of the RW is small then the mean pressure
generated within the fluid volume is also small.
Even if the volume velocity of the freely moving part of the RW is equal and opposite
to that of the piston, the second pressure component, a near-field pressure, is
generated close to the RW. This is due to the pressure required to accelerate the fluid
from below the piston into the freely moving part of the RW and is thus dependent on
the fluid inertia. The spatial distribution of this near-field pressure will depend on the
details of the geometry, but its magnitude will be proportional to the fluid density, ρ,
and to the acceleration of the piston, iωq. An indicative overall magnitude of PN can
then be defined as
PN∝ iωρq. (2)
For a given piston volume velocity, the average pressure (Eq. 1) decreases with
increasing the stimulation frequency, whereas the average near-field pressure (Eq. 2)
increases with the frequency.
It is worth noting that if the cochlear wall is not perfectly rigid, for example if the
stapes is not completely fixed, the internal pressure also generates a volume velocity
on the oval window, as illustrated in figure 2A. This would generate a pressure
component due to the flow of fluid across the cochlea that would depend on the fluid
inertia rather than the stiffness of the window. A near-field pressure would then be
generated in the vicinity of both the oval window and the RW.
The on-axis near-field pressure, PN, due to RW stimulation can be estimated using a
simple model of cancelling piston sources. The assumed geometry is shown in figure
2B in which an inner piston with a radius of a1 has a velocity of v1 and an annular
piston with an inner radius of a1 and an outer radius of a2 has a velocity of v2.
The volume velocity of the inner piston thus is
q1=π a12 v1 , (3)
and that of the outer piston is
q2=π (a22−a1
2 )v2 . (4)
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
If the cochlea is sealed, the volume velocities are equal and opposite so that q2=−q1,
and the ratio of the two linear velocities must be
v2
v1=
a12
a22−a1
2 ,(5)
which is plotted in figure 3.
Assuming free field conditions, the complex on-axis pressure a distance of r from a
single piston of radius a vibrating with velocity v is given (e.g. see Kinsler et al [20])
as
p (r )=−ρcv (e−ik √r2+a2
−e−ikr ) , (6)
where k=2 π /λ is the wavenumber, λ is the acoustic wavelength and c is the speed of
sound in the fluid. Distance between the RW and the BM, which is <1 mm in both
guinea pigs and humans [21], is much smaller than the wavelength, which is equal to
about 1.5m at 1 kHz assuming a speed of sound of 1500 ms-1 in the cochlear fluid.
Hence, we can assume λ≫ r , and
kr=2 πrλ
≪1. (7)
In this case we can take the first order series approximation to the exponentials in
equation (6) to give the pressure due to the inner piston as
p1(r )≈iρω v1 (√r2+a12−r ) . (8)
The pressure due to the annular piston is equal to that due to a piston of radius a2 and
velocity v2 minus that due to a piston of radius a1 and velocity v1
p2(r )≈iρω v2 (√r2+a22−√r2+a1
2 ) . (9)
The total near-field pressure PN(r ) is the superposition of those due to the inner and
outer pistons in equations (8) and (9) with v2/v1 given by equation (5)
PN(r )=iρω v1[(√r2+a12−r )− a1
2
a22−a1
2 (√r2+a22−√r2+a1
2 )] . (10)
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
Although this expression is not strictly valid in the present case unless the near-field
pressure has died away over the width of the cochlea, the modulus of equation (10) is
plotted in figure 4 as a function of distance from the centre of the source distribution,
r, for various values of a1 /a2 and with a constant value of v1. The pressure drops
quickly with the distance from the piston, especially for small a1/a2, but the shortest
distance between the RW and the BM is also only 0.2 mm in guinea pigs [21]. Hence,
excitation of the cochlea with a probe which covers just a part of the RW is possible
by the near-field pressure in vicinity of the RW. A natural consequence of the simple
model of cancelling piston sources is higher near-field pressure for larger a1/a2 ratios
for constant piston velocity. In the real cochlea, however, the RW piston velocity is
likely to drop for larger a1/a2 ratios as a consequence of an increase in cochlear
impedance as seen by the piston. A model is needed, which takes into account this
effect and measurements of the RW impedance for different proportions of the RW
covered by the magnet, in order to find an optimum a1/a2 ratio which provides the
most efficient stimulation of the cochlea through the RW. It also should be noted that
within the validity of equation (10), the total near-field pressure PN(r ) is proportional
to the acceleration of the inner piston. The BM is a pressure detector [22], it should,
therefore, be expected that BM stimulation is proportional to piston acceleration.
3.2 Placement of magnet on the RW does not alter the sensitivity of the cochlea to
acoustic stimulation
The cochlear neural threshold to acoustic stimulation, measured as a threshold for the
N1 peak of the CAP, did not change after placement of the magnet on the RW and
during the entire experiment (up to 3 hours after the placement) (figure 5). It is well
established that increases in the impedance of the RW leads to elevation of the
hearing threshold (e.g. see Culler et al [1]), thus the observed stability of the CAP
thresholds after placement of the magnet revealed that the RW impedance did not
increase and the area of the RW not covered by the magnet could provide effective
pressure relief at acoustic frequencies when the cochlea is stimulated conventionally.
The RW impedance is much lower than any other leakage impedance in the cochlea
(e.g. see Stieger et al [23]). Thus, in view of the stability of the N1 thresholds
following placement of the magnet, we can conclude that the RW impedance
remained the smallest impedance in the system, even after this manipulation.
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
3.3 Frequency dependence of neural thresholds in response to subnanometer RW
and acoustic stimulation are similar
The coil voltage - magnet displacement relationship was linear in all six preparations
where this characteristic was studied (an example is given in figure 6). In all
preparations, the frequency dependence of the magnet displacement resembled that of
a low-pass filter with the high-frequency slope being between 12 - 18 dB/octave
(figure 7A). A slope of 12 dB/octave is to be expected if the magnet is behaving as a
mass subject to a constant force caused by the magnetic field due to the voltage. In
this case, the magnet acceleration should not depend on frequency, which is indeed
observed for most of the frequency range used in this study (figure 7B). Because the
high-frequency cut-off of the system for electrical stimulation was above 100 kHz, the
relatively steep high-frequency slope of the coil voltage - magnet displacement
relationship for frequencies below 10 kHz (figure 7A) most likely reflects the multiple
degrees of freedom of the mechanical system formed by the inductively coupled coil
and the magnet inertia, together with the RW and the cochlear fluid.
There were no major differences between CAP threshold curves for acoustic
stimulation between all preparations used in this study (figure 7C). However, in the
same preparations, CAP threshold curves expressed as a function of the coil voltage
required to generate threshold CAP for magnetically driven RW stimulation varied by
~10 dB over the 8 – 20 kHz range (figure 7D). Likely bases for these CAP threshold
variations could include differences, between preparations, in the relative position of
the magnet and the coil and to slightly different locations of the magnet on the RW.
The smallest coil voltages required to generate the threshold CAP was observed for
frequencies between 10-11 kHz. Taking into account the linearity of the magnet
displacement (figure 6), the corresponding magnet displacements for the threshold
coil voltage could be readily derived from the iso-voltage response curves at 10 V
(figure 7A) and the CAP threshold curves (figure 7D) using the following equation:
displacement at threshold = displacement at 10 V / 10^(attenuation at threshold/20).
The derived magnet displacement threshold curves (figure 7E) reveal that the
sensitivity of the guinea pig cochlea to RW stimulation is in the sub nanometre range.
Maximum sensitivity, which corresponds to the smallest magnet displacements at
threshold, was observed for frequencies around 20 kHz, which is similar to the
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
frequency of maximum sensitivity for acoustic stimulation (figures 5, 7C, note
different frequency ranges in these figures). The absolute sensitivity of the magnet
displacement threshold curves varied between preparations (figure 7E), probably, as
pointed out above, because of variations between preparations in the relative locations
of the coil, the magnet and the RW membrane. Magnet displacement threshold curves
within the range of those shown in figure 7E were obtained for three further
preparations but, for clarity, are not shown.
Near-field pressure, which is likely to excite the BM responses in our experiments, is
proportional to the acceleration of the magnet (figure 7F) which, at threshold, was
calculated from the threshold displacement (figure 7E). The threshold acceleration
changes only by about 20 dB within the frequency range studied (figure 7F). This
change corresponds better to the changes in the threshold SPL for acoustic stimulation
within the same frequency range (figure 7C) while threshold displacement changes by
more than 40 dB for the same frequencies (figure 7E). This last observation provides
an additional confirmation that in our experiments the BM is excited by the near-field
pressure which is proportional to the magnet acceleration.
3.4 Latencies of neural responses to acoustic and RW stimulation are similar
To gain insight into the mechanisms by which the cochlea is excited through RW
stimulation, we compared the CAP latencies for suprathreshold stimulation within 10
dB above the thresholds for acoustic and RW stimulation of the cochlea (figure 8).
The latencies were essentially the same. The CAP latency for low-level acoustic
stimulation depends on the traveling wave delay and is inversely related to the
characteristic frequency of the fibre (e.g. see Goldstein et al [24]). Hence, we can
conclude that the mechanisms of energy propagation to the characteristic frequency
place were similar for both acoustic and RW stimulation.
3.5 Increase in the stapes impedance does not affect efficiency of cochlear
stimulation through the partially occluded RW
As suggested previously, the relatively high acoustic impedance of the middle ear
ossicles (as seen from the cochlea), combined with a magnet that only partially
occluded the RW, may allow the region of the RW not covered by the magnet to act
as a pressure shunt during RW stimulation. It may thus be suggested that stapes
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
mobility plays little or no role in the stimulation of the cochlea through the RW.
Indeed, when the cochlea was excited through the RW, movements of the stapes at the
frequency of stimulation were observed above the noise floor of the interferometer
only at low frequencies <7 kHz and high stimulation amplitudes (figures 9A-B).
Preparations with the largest and smallest stapes responses are shown in figure 9A-B.
Similar results were observed in two further preparations. When the stapes impedance
was increased by filling the ear canal with superglue, and stapes displacements could
not be detected above the noise floor of the interferometer regardless of the frequency
and magnitude of the control voltage to the stimulating coil (figures 9A-B), thresholds
of RW elicited CAPs before and after increase in the stapes impedance were largely
similar (figure 9C). According to this observation it is suggested that stapes mobility
is not required for effective excitation of the cochlea in our experimental
configuration, i.e. when relatively large part of the RW was not occluded.
4. Discussion
On the basis of CAP threshold measurements at stimulus frequencies that span almost
its entire auditory range, the guinea pig cochlea is sensitive to sub-nanometre
displacements and minute accelerations of the RW (figure 7). The CAP data, set in the
context of a model of cochlear stimulation through the partially occluded RW, support
the viability of RW stimulation as an effective route for exciting the mammalian
cochlea, further building on previous studies of the mechanical properties of the
cochlea and recent work concerning the implementation of RW implantable hearing
devices [13, 25, 26]. Quantification of the RW transducer displacement/acceleration
presented here provides parameters essential for the design of future devices, although
these parameters vary within individuals (figures 7 and 9). These factors may include
the shape of the transducer and the area of the RW it covers. Most significant is
whether the transducer completely covers the RW, as demonstrated by other groups
[27-29], or leaves a RW area partially free [26, 30-32]. The outcome of our
experiments reveals that partial coverage of the RW may not affect RW impedance
sufficiently to affect cochlear stimulation with the exposed portion of the RW
membrane acting as a pressure shunt (figure 5). Complete coverage would probably
remove this shunt, possibly resulting in a different mechanism of BM stimulation.
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
Our experiments demonstrate that middle ear prostheses, which partially cover the
RW, can be used effectively for simultaneous cochlear stimulation and pressure relief
through the RW. The existence of a “third window” (an additional pressure shunt,
Z3 , SV , Z3 , ST, figure 10) in the cochlea has been postulated to account for the efficiency
of RW stimulation because of the relatively high impedance, as seen from the cochlea,
presented by the ossicles [30]. The proposed identity of the third window ranges from
the vasculature of the cochlea to the cochlea and vestibular aqueducts [17, 30].
Evidence for the existence of the third window and its ability to shunt pressure is
claimed from experiments in which the RW membrane is completely blocked during
normal acoustic stimulation or the stapes is immobilized and the cochlea is stimulated
through the RW route. Despite blockage or fixation, the air conduction thresholds are
raised by between 20-50 dB depending on the study cited [3, 30, 33-35]. However,
even the comparatively low experimental value of a 20 dB rise in threshold level is
significant when compared with the low thresholds achieved in the current study
(figure 9). Additionally, it is unlikely that a hypothetical third window is responsible
and, indeed, required for pressure relief in our experiments. The RW impedance (ZRW,
figure 10A) is much lower than any other leakage impedance in the cochlea during
normal acoustic stimulation (e.g. see Stieger et al [23]), i.e. ZRW ≪ Z3 , ST and pressure
at point N is determined mainly by ZRW . Moreover, the neural thresholds are highly
sensitive to increases in RW impedance (e.g. see Culler et al [1]). In our experiments,
the neural thresholds during acoustic stimulation did not change after the magnet
placement (figure 5). We took this finding to indicate that the RW impedance
remained the smallest impedance in the system even after this manipulation (i.e.
ZRW ≪ Z3 , ST after the magnet placement) and provided an effective pressure shunt
minimising the pressure at point N (figure 10B) and the far-field pressure drop across
the impedance of the cochlear partition (ZBM, figure 10B). This outcome might also be
expected from the direct finding that intracochlear pressure differences between the
scala vestibuli and scala tympani remained low during RW stimulation [23]. On the
basis of our findings reported here, we propose that the pressure shunting occurred
through the area of the RW which remained exposed. This proposal is also supported
by the findings of Schraven et al [32] who tested different sized actuator tips and their
coupling to the RW. Schraven et al [32] found that stimulation by actuators with tip
sizes well below the dimensions of the RW resulted in reduced stapedial movements
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
compared with those tips that covered most of the RW. This result could be accounted
for if the free exposed area of the RW acted as a pressure shunt. This does not
discount other “windows” contributing to pressure relief but it is clear from results
presented here that their contribution appears to be negligible in our experiments.
We found that CAP thresholds did not change after stapes impedance (ZME, figure
10B) was increased (figure 9). CAP threshold elevation was, however, observed in
earlier studies using RW stimulation through a partially blocked RW [30].
Furthermore, where we found RW stimulation through the partially occluded RW
elicited stapes displacements only at high intensities below 7 kHz, under similar
stimulus conditions, Lupo et al [30] observed stapes movement in chinchillas across a
wider frequency range of 0.25 -16 kHz. We suggest that tighter hydromechanical
coupling between the RW transducer and the stapes, and consequent changes in the
hydromechanical properties of the cochlea after stapes immobilisation, may account
for elevation of the CAP thresholds after stapes fixation observed by Lupo et al [30].
Correspondingly, differences in transducer shape and the proportion of the RW
covered, may potentially account for differences in the transducer-stapes
hydromechanical coupling reported in this study and by Lupo et al [30].
Pre-tension of the RW, which can potentially increase stiffness of the exposed area of
the RW, and hence the average pressure within the cochlear (Eq. 1), can enhance
stapedial movement during RW stimulation [31, 36]. It is apparent that differences in
the outcomes of different studies, including ours presented here, depend on a number
of factors including the choice of both the size and shape of the transducer and pre-
tension of the RW. The finding by Stieger et al [23] that the stapes velocity is not a
good measure of the effectiveness of reverse stimulation of the cochlea, might reflect
sensitivity of the stapedial responses to variations in the parameters of RW
stimulation mentioned above. Control of these parameters is important to ensure that
mobility of the ossicular chain is reduced to levels that negate the need for additional
intrusive procedures that might cause sensorineural hearing loss.
In our experiments the RW provided an effective common pathway for both cochlear
stimulation and pressure relief. Therefore, we conclude that the mode of cochlear
excitation through the RW in this case is different from that observed during
conventional, acoustical cochlear stimulation or RW stimulation, which does not
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
allow pressure relief through the RW [28, 29]. Our results together with direct
measurements of the pressure in the cochlea during RW stimulation [23] using a
transducer that leaves the RW partially exposed, indicate that under this condition the
transducer displacement probably is not able to cause a far-field pressure difference
between the cochlear scalae, which is the normal stimulus for the cochlea [12]. Far-
field pressure differences between the scalae are unlikely to be caused by the type of
RW stimulation we employed because of the high impedance of the ossicular chain,
as seen from the cochlea [27, 28], and pressure backflow around the transducer [11].
We propose, instead, that during RW stimulation through the partially occluded RW,
the BM is stimulated via the near-field complex pressure generated in the vicinity of
the RW (figure 4), i.e. a fluid-jet flow, which, however, results in the generation of
conventional travelling waves along the cochlear partition. In this sense, the cochlear
excitation due to the RW stimulation in our experiments is similar to excitation due to
rocking movement of the stapes, which creates only local pressure gradients/fluid
motion in the cochlea [37, 38] resulting, nevertheless, in generation of the travelling
waves and neural excitation [39, 40]. In our experiments, the near-field pressure is
proportional to the acceleration of the transducer which allows effective cochlear
stimulation even at extremely small RW displacements at high frequencies. This
conclusion should be taken into account during the design of hearing-aid devices
which stimulate the cochlea through the partially occluded RW. The near-field
pressure variations will be limited to the vicinity of the RW membrane (figure 4), but
the close spatial relationship between the RW and the BM [21] make excitation of the
BM by these localised near-field pressure variations possible. Once the BM is
stimulated in the vicinity of the RW, a conventional travelling wave is generated
along the cochlear partition, which is supported by our CAP recordings. This
conclusion is based on our finding that the 12-25 kHz region of the cochlea, which is
most sensitive to RW magnet stimulation (figure 7E,F), is located in the middle of the
basal turn of the cochlea and this region is not adjacent to, or cannot be observed
through the RW. Indeed, the CAP latency for low-level acoustic stimulation, which
depends on the travelling wave delay (e.g. see Goldstein et al [24]), is similar for both
acoustic and RW stimulation (figure 8). Furthermore, CAP threshold tuning curves
derived through acoustic (figures 5 and 7C) and RW magnet (figure 7E,F)
stimulations have similar maxima in sensitivity between 12-25 kHz.
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
It should be noted that it is not essential for the excitation that the motion of the
magnet is strictly piston-like. It is likely that the magnet underwent some rocking
motion in our experiments. This possibility, however, does not affect our conclusion
about the mode of cochlear excitation and, in terms of the model (figure 2B), simply
corresponds to a different ratio a1 /a2 which should still provide effective stimulation.
Our investigation of the mechanisms underlying cochlea stimulation through the RW
using a transducer that partially covers the RW membrane has important significance
for future clinical use of such devices. RW stimulation does not even require mobility
of the ossicular chain, as demonstrated in this study, to provide input to the cochlea
that is similar to that obtained acoustically in a ‘normal’ ear. RW stimulation does,
however, involve a novel mechanism of generating travelling waves along the
cochlear partition, which we believe has not previously been considered. Travelling
waves are initiated as a consequence of near-field pressure in the immediate vicinity
of the RW, rather than far-field pressure differences between the scalae vestibule and
tympani.
ACKNOWLEDGMENTS
This work was supported by the Medical Research Council. We thank J. Hartley for
technical assistance.
REFERENCES
1. Culler E, Finch G, Girden E. 1935 Function of the round window in hearing. Am. J. Physiol. 111, 416-425.2. Martin C, Tringali S, Bertholon P, Pouget JF, Prades JM. 2002 Isolated congenital round window absence. Ann. Otol. Rhinol. Laryngol. 111, 799-801.3. Linder TE, Ma F, Huber A. 2003 Round Window Atresia and its Effect on Sound Transmission. Otol Neurotol. 24, 259-263.4. Borrmann A, Arnold W. 2007 Non-syndromal round window atresia: an autosomal dominant genetic disorder with variable penetrance? Eur. Arch. Otorhinolaryngol. 264, 1103-1108.5. Roland PS, Wright CG. 2006 Surgical aspects of cochlear implantation: mechanisms of insertional trauma. Adv. Otorhinolaryngol. 64, 11-30. 6. Goodhill V. 1971 Sudden deafness and round window rupture. Laryngoscope 81, 1462-1474.7. Mertens J. 1991 The reaction of the guinea pig cochlea to perforations of the round window membrane with and without perilymph aspiration. Eur Arch Otorhinolaryngol. 248, 395-399.8. Rozsasi A, Sigg O, Keck T. 2003 Persistent inner ear injury after diving. Otol. Neurotol. 24, 195-200.
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461462463464465466467468469470471472473474475476477478479480
9. Gedlicka C, Formanek M, Ehrenberger K. 2009 Analysis of 60 patients after tympanotomy and sealing of the round window membrane after acute unilateral sensorineural hearing loss. Am J Otolaryngol. 30,157-161.10. Wever EG, Lawrence M, Békésy Gv. 1954 A note on recent developments in auditory theory. Proc. Natl. Acad. Sci. USA 40, 508-512.11. Békésy Gv. 1960 Experiments in hearing. McGraw-Hill Book Company.12. Voss SE, Rosowski JJ, Peake WT. 1996 Is the pressure difference between the oval and round windows the effective acoustic stimulus for the cochlea? J. Acoust. Soc. Am. 100,1602-1616.13. Spindel JH, Lambert PR, Ruth RA. 1995 The round window electromagneticimplantable hearing aid approach. Otolaryngol. Clin. North. Am. 28, 189-205.14. Colletti V, Soli SD, Carner M, Colletti L. 2006 Treatment of mixed hearing losses via implantation of a vibratory transducer on the round window. Int. J. Audiol. 45, 600-608.15. Kiefer J, Arnold W, Staudenmaier R. 2006 Round window stimulation with an implantable hearing aid (Soundbridge®) combined with autogenous reconstruction of the auricle – a new approach. ORL 68, 378-385.16. Wollenberg B, Beltrame M, Schönweiler R, Gehrking E, Nitsch S, Steffen A, Frenzel H. 2007 Integration of the active middle ear implant Vibrant Soundbridge in total auricular reconstruction. HNO 55, 349-356.17. Koka K, Holland NJ, Lupo JE, Jenkins HA, Tollin DJ. 2010 Electrocochleographic and mechanical assessment of round window stimulation with an active middle ear prosthesis. Hear. Res. 263, 128-137.18. Ghiz AF, Salt AN, DeMott JE, Henson MM, Henson Jr OW, Gewalt SL. 2001 Quantitative anatomy of the round window and cochlear aqueduct in guinea pigs. Hear. Res. 162, 105-112.19. Lukashkin AN, Bashtanov ME, Russell IJ. 2005 A self-mixing laser-diode interferometer for measuring basilar membrane vibrations without opening the cochlea. J. Neurosci. Methods. 148, 122-129.20. Kinsler L, Frey A, Coppens A, Sanders JV. 1982 Fundamentals of acoustics. New York: John Wiley & Sons, Inc.21. Poznyakovskiy AA, Zahnert T, Kalaidzidis Y, Schmidt R, Fischer B, Baumgart J, Yarin YM. 2008 The creation of geometric three-dimensional models of the inner ear based on micro computer tomography data. Hear. Res. 243, 95-104.22. Peterson LC, Bogert BP. 1950 A dynamical theory of the cochlea. J. Acoust. Soc. Am. 22, 369-381.23. Stieger C, Rosowski JJ, Nakajima HH. 2013 Comparison of forward (ear-canal) and reverse (round-window) sound stimulation of the cochlea. Hear. Res. 301, 105-114.24. Goldstein JL, Baer T, Kiang NY. 1971 A theoretical treatment of latency, group delay, and tuning characteristics for auditory-nerve responses to clicks and tones. In Physiology of the Auditory System (ed. M. B. Sachs), pp. 133-141. Baltimore: National Educational Consultants.25. Zennaro O, Dumon T, Erre J-P, Guillaume A, Aran J-M. 1992 Piezo-electric middle ear implant hearing aid experimental model in guinea-pig. In Engineering in Medicine and Biology Society, 1992 14th Annual International Conference of the IEEE, pp.1316-1317.26. Lupo JE, Koka K, Jenkins HA, Tollin DJ. 2012 Third-window vibroplasty with an active middle ear implant: assessment of physiologic responses in a model of stapes fixation in Chinchilla lanigera. Otol. Neurotol. 33, 425-431.
481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530
27. Nakajima HH, Dong W, Olson ES, Rosowski JJ, Ravicz ME, Merchant SN. 2010 Evaluation of round window stimulation using the floating mass transducer by intracochlear sound pressure measurements in human temporal bones. Otol. Neurotol. 31, 506-511.28. Nakajima HH, Merchant SN, Rosowski JJ. 2010 Performance considerations of prosthetic actuators for round-window stimulation. Hear. Res. 263, 114-119.29. Tringali S, Koka K, Deveze A, Holland NJ, Jenkins HA, Tollin DJ. 2010 Round window membrane implantation with an active middle ear implant: a study of the effects on the performance of round window exposure and transducer tip diameter in human cadaveric temporal bones. Audiol. Neurotol. 15, 291-302.30. Lupo JE, Koka K, Holland NJ, Jenkins HA, Tollin DJ. 2009 Prospective electrophysiologic findings of round window stimulation in a model of experimentally induced stapes fixation. Otol. Neurotol. 30, 1215-1224.31. Schraven SP, Hirt B, Gummer AW, Zenner H-P, Dalhoff E. 2011 Controlled round-window stimulation in human temporal bones yielding reproducible and functionally relevant stapedial responses. Hear. Res. 282, 272-282.32. Schraven SP, Hirt B, Goll E, Heyd A, Gummer AW, Zenner HP, Dalhoff E. 2012 Conditions for highly efficient and reproducible round-window stimulation in humans. Audiol. Neurotol. 17, 133-138.33. Wever EG, Lawrence M. 1950 The acoustic pathways to the cochlea. J. Acoust. Soc. Am. 22, 460-467.34. Tonndorf J, Tabor J. 1962 Closure of the cochlear windows: its effect upon air-and bone-conduction. Ann. Otol. Rhinol. Laryngol. 71, 5-29.35. de Alarcon A, Jahrsdoerfer RA, Kesser BW. 2008 Congenital absence of the oval window: diagnosis, surgery, and audiometric outcomes. Otol. Neurotol. 29, 23-28.36. Maier H, Salcher R, Schwab B, Lenarz T. 2013 The effect of static force on round window stimulation with the direct acoustic cochlea stimulator. Hear. Res. 301, 115-124.37. Pozrikidis C. 2008 Boundary-integral modeling of cochlear hydrodynamics. J. Fluids Struct. 24, 336-365.38. Edom E, Obrist D, Henniger R, Kleiser L, Sim JH, Huber AM. 2013 The effect of rocking stapes motions on the cochlear fluid flow and on the basilar membrane motion. J. Acoust. Soc. Am. 134, 3749-3758.39. Huber AM, Sequeira D, Breuninger C, Eiber A. 2008 The effects of complex stapes motion on the response of the cochlea. Otol. Neurotol. 29, 1187-1192.40. Eiber A, Huber AM, Lauxmann M, Chatzimichalis M, Sequeira D, Sim JH. 2012 Contribution of complex stapes motion to cochlea activation. Hear. Res. 284, 82-92.
531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570
571
FIGURE LEGENDS
Figure 1. Schematic view of the cochlea (CH) through a lateral opening in the
temporal bone. M indicates a neodymium iron boron disk magnet (diameter 0.6mm,
thickness 0.2mm) placed on the surface of the round window (RW). C labels a
miniature coil situated above the magnet. S denotes a Teflon-coated silver wire used
for recording of activity of the auditory nerve. IS indicates incudostapedial joint.
Figure 2. Pressure distribution during cochlear stimulation through partially occluded
RW. A Pressure distribution in presence of a mobile stapes. The effect of the mean
internal pressure on an imperfectly fixed stapes is to generate a small volume velocity,
−αq, where α <1. In this case, the freely moving area of the RW generates a volume
velocity of −(1−α ) q and these two components cancel the volume velocity, q, of the
driving piston on the RW. Then the internal pressure is the sum of that due to a fixed
oval window and that due to two pistons with volume velocities +αq and −αq on
either side of the volume. B Sketch of cancelling piston sources.
Figure 3. Dependence of the ratio of the two linear velocities, v1/ v2, on the ratio of
inner piston to outer piston radii, a1/a2.
Figure 4. The modulus of the total near-field pressure, ¿ PN∨¿, as a function of the
distance from the piston for different ratio of inner piston to outer piston radii, a1/a2.
The inner piston velocity, v1, is 1 mm/s, the fluid density is 103 kg/m3, the frequency
of stimulation is 15 kHz. Cartoons along the a1/a2 axis show position of the magnet
(grey area) on the RW for different relative probe sizes.
Figure 5. Thresholds of acoustic stimulation (dB SPL) vs frequency (kHz) for N1
peak of the CAP of the auditory nerve measured before (open symbols) and 2.5 hours
after (solid symbols) placement of the magnet on the RW.
Figure 6. Displacement of the RW magnet as a function of voltage applied to the coil
for three different stimulus frequencies (5, 10 and 20 kHz) measured in situ on the
RW of a single preparation. The data at each frequency are pooled and normalized
from measurements at different gains of the signal conditioning amplifier.
Figure 7. Magnet displacement and neural thresholds as a function of the stimulation
frequency in three representative preparations. A Frequency dependence of the
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
magnet displacement at constant voltage (10 V) applied to the coil. B Frequency
dependence of the magnet acceleration at constant voltage (10 V) applied to the coil
calculated from data at panel A. C CAP thresholds during acoustic stimulation. Note a
shorter frequency range than in figure 2. D Dependence of the coil voltage to generate
threshold CAP during the RW stimulation. E Corresponding magnet displacement at
the CAP threshold. F Corresponding magnet acceleration at the CAP threshold
calculated from data at panel E.
Figure 8. CAP latency for suprathreshold stimulation within 10 dB of the threshold
for the acoustic and RW stimulation of the cochlea (mean SD, N=3 for each
condition of stimulation).
Figure 9. Effect of stapes fixation on the CAP thresholds during the RW stimulation.
A, B Stapes displacements for different attenuation of voltage (re 10V) applied to the
coil. Average of three measurements. Error bars are omitted for clarity. Noise floor of
the interferometer is indicated by a dashed line. C Corresponding CAP thresholds
measured before (solid symbols) and after (open symbols) stapes fixation.
Figure 10. Schematics of impedances of the cochlea during acoustic (A) and RW (B)
stimulation. ZBM and ZRW are impedances of the cochlear partition and the open area
of the RW respectively. ZME is the stapes impedance as seen from the cochlea. Z3 , SV
and Z3 , ST are impedances of “third windows” in the scala vestibuli and scala tympani,
respectively. q indicates a volume velocity generated by stimulation.
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
Figure 1
IS
M
S
C
RW
CH
21
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
Figure 2
A
B
22
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
Figure 3
0.0 0.2 0.4 0.6 0.8 1.010-3
10-1
101 -v
2/v1
a1/a2
23
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
Figure 4
24
669
670
671
672
673
674675
676
677
678
679
680
681
682
683
684
685
Figure 5
1 10-20
-10
0
10
20
30
CA
P th
resh
old
(dB
SP
L)
Tone frequency (kHz)
before magnet placement 2.5 hours after
rw2
25
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
Figure 6
50 40 30 20 10 00.01
0.1
1
10D
ispl
acem
ent a
mpl
itude
(AU
)
Attenuation (dB re 10V)
5kHz
10kHz 20kHz
1dB/dB
rw3
26
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
Figure 7
10 20 30
1
10
100
Mag
net d
ispl
acem
ent (
nm)
Frequency (kHz)
rw2 rw6 rw17
10V input
18dB/oct
12dB/oct
A
10 20 300
40
80
120
Mag
net a
ccel
erat
ion
(m/s
ec^2
)
Frequency (kHz)
rw2 rw6 rw17
B
5 10 15 20 25 30-20
-10
0
10
20
Sou
nd p
ress
ure
(dB
SP
L)
Frequency (kHz)
rw2 rw6 rw17
C
5 10 15 20 25 3070
60
50
40
30
Thre
shol
d vo
ltage
(-dB
re 1
0 V
)
Frequency (kHz)
rw2 rw6 rw17
D
5 10 15 20 25 301E-3
0.01
0.1
1
Mag
net d
ispl
acem
ent (
nm)
Frequency (kHz)
rw2 rw6 rw17
E
5 10 15 20 25 300.01
0.1
1
Mag
net a
ccel
erat
ion
(m/s
ec^2
)
Frequency (kHz)
rw2 rw6 rw17
F
27
717
718
719
720
721
722
723
724
725
726
727
Figure 8
1 10
1.0
1.5
2.0
2.5
CA
P la
tenc
y (m
s)
Frequency (kHz)
magnet stimulation air stimulation
28
728
729
730
731
732
733
734
735
736
737
738
Figure 9
1 10
0.1
1
10
Dis
plac
emen
t (nm
)
Frequency (kHz)
noiseattenuation
30att 20att 10att 0att
A
1 10
0.1
1
10
Dis
plac
emen
t (nm
)
Frequency (kHz)
noiseattenuation
30att 20att 10att 0att
B
5 10 15 20 25 30
70
60
50
40
30
Thre
shol
d vo
ltage
(-dB
re 1
0 V
)
Frequency (kHz)
stapes mobile stapes fixed stapes mobile stapes fixed
C
29
739
740
741
742
743
744
745
746
Figure 10
N
BZME
q
q
Z3,ST
ZRW
Z3,SV
ZBM
Z3,ST
ZRW
Z3,SV
ZBM
A
N
30
747
748
749
750
751
752
753