Power dissipation in the cochlea Prodanovic et al.

Title:
Power dissipation in the cochlea is beneficial for frequency selectivity

Authors:

S. Prodanovic1, S. M. Gracewski1, 2, and J.-H. Nam1, 2

Key Words:

Organ of Corti, outer hair cell, cochlear impedance, frequency tuning, viscous damping
Power dissipation in the cochlea Prodanovic et al.

ABSTRACT (< 300 words)

In the cochlea, acoustic energy is transmitted toward the apex through the vibrations of a viscoelastic
partition known as the organ of Corti complex. The dimensions of the vibrating structures range from a
few hundred micrometers (e.g., the basilar membrane) to a few micrometers (e.g., the stereocilia bundle).
Vibrations of micro-structures in viscous fluid are subjected to energy dissipation. Because the viscous
dissipation is considered to be detrimental to the function of hearingsound amplification and frequency
tuning, the cochlea is believed to use cellular actuators to overcome the dissipation. Compared to the
extensive investigations on the cellular actuators, the dissipating mechanisms have not been given
appropriate attention. For example, many theoretical studies use an inviscid fluid approximation, and
lump the viscous effect to viscous damping components. Others neglect viscous dissipation in the organ
of Corti, but consider fluid viscosity. We have developed a computational model of the cochlea that
incorporates viscous fluid dynamics, organ of Corti micro-structural mechanics, and electro-physiology of
the outer hair cells. The model is validated by comparing with experimental results in the literature, such
as the viscoelastic response of the tectorial membrane, and the cochlear input impedance. Using the
model, we investigated how dissipation components in the cochlea affect its function. Our results suggest
that most energy dissipation occurs within the organ of Corti complex, not in the scalar fluids. We found
that, quite contrary to the long-held belief, the dissipation in the cochlea is beneficial for high quality
tuning. Our results suggest that there exists an optimal dissipation level for tuning quality. Appropriate
dissipation enhances the tuning quality by confining the spread of energy from the amplification site.
 Power dissipation in the cochlea Prodanovic et al.

1 INTRODUCTION
2 In the cochlea, acoustic energy is transmitted through the vibrations of the viscoelastic cochlear partition
3 (1). Vibrating structures in the cochlea include the 100-500 m-wide basilar membrane, and the 1-10 m-
4 tall stereocilia of the sensory receptor cells. Because of the micron-scale, fluid dynamics in the cochlea is
5 highly viscous (Reynolds number < 1). A generally accepted but seldom voiced notion is that viscoelastic
6 dissipation in the cochlear partition is detrimental to tuning and amplification of acoustic signals. To
7 achieve physiological levels of tuning and amplification, the cochlear system is thought to need active
8 mechanical feedback (2). The outer hair cells, one of two different types of sensory receptor cells in the
9 mammalian cochlea, generate mechanical force to compensate for the dissipation (3).

10 Transmission of acoustic energy along the length of cochlear partition has been actively investigated. An
11 isolated outer hair cell can generate electromotile force on the order of 10-10 N (4, 5). According to the
12 measurement of cochlear input impedance (6-8), the acoustic power delivered through the stapes is on the
13 order of 10-10 W at 40 dB sound pressure level. In theory, the outer hair cells can generate more power
14 than the acoustic input power at 60 dB sound pressure level (9, 10). On the other hand, how power is
15 dissipated in the cochlea has been investigated at the cellular level (11-14), tissue level (15, 16), and
16 systems level (10). Existing theories on how/where the power is dissipated in the cochlea are divergent.
17 For example, most theoretical studies use an inviscid fluid and lump the viscous effect into damping
18 components, with damping properties adjusted to match the desired tuning quality and amplification
19 factor. In contrast, a series of studies by Steele and his colleagues (17, 18) explicitly incorporated the fluid
20 viscosity while neglecting the dissipation within the organ of Corti. The different assumptions for
21 dissipating mechanisms are partly responsible for the divergence in cochlear amplification theories.

22 We developed a computational model of cochlear mechano-transduction emphasizing the micro-

23 mechanics of the organ of Corti (19, 20). In an effort to better describe the dissipating mechanisms in the
24 cochlea, fluid viscosity has been incorporated in this study. We investigated where and how the acoustic
25 power is dissipated, and how dissipating properties affect the frequency tuning quality of the cochlea.

26
27

28

Page 1 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 METHODS

2 Model overview
3 Three physical systems are solved simultaneously (Fig. 1): the fluid dynamics in the cochlear ducts, the
4 structural mechanics of the organ of Corti complex (OCC: the complex of the organ of Corti, tectorial
5 membrane and basilar membrane), and the electro-physiology of the outer hair cells. A similar approach
6 has been established by Grosh and his colleagues (21, 22). Model properties represent the gerbil cochlea.
7 The longitudinal, radial and transverse directions correspond to the x-, y- and z- axes, respectively.

8 The fluid domain is separated by the OCC into two chambers (Fig. 1A). The bottom chamber representing
9 the scala tympani and the top chamber representing the scala media and vestibuli. Note that most existing
10 cochlear studies assume symmetry to reduce the two fluid chamber system into a one-chamber system,
11 but several recent studies relaxed the symmetry assumption (23-25). The fluid in the top and bottom
12 chambers interacts independently with the top and the bottom surfaces of the OCC, respectively. The
13 entire domain is 0.6 mm tall and 12.3 mm long including the 0.3 mm helicotrema.

14 The OCC mechanics is represented by 3-D finite element (FE) model (Fig. 1B). Structural details of the
15 OCC, such as the basilar and tectorial membranes, Dieters and pillar cells, the reticular lamina and the
16 outer hair cells, are incorporated (26). Those cells and tissues are represented with beam elements that can
17 bend or stretch. The tectorial and basilar membranes are represented by a meshwork of beams consisting
18 radial and longitudinal components. Considering the strong anisotropy of the two membranes due to
19 collagen fibers aligned in the radial direction (along the y-axis), the Youngs moduli of the radial beams

Figure 1. The cochlear model. (A) Cochlear viscous fluid dynamics: Top and bottom fluid chambers
are separated by elastic OCC. Scale bars (x, z) (1, 0.2 mm). Pressure values are in mPa. (B) OCC
micro-mechanics and OHC electrophysiology: The 3-D finite element model of the OCC incorporated
realistic geometrical and mechanical characteristics of the gerbil cochlea. The OCC micro-structures
repeat with a longitudinal grid size of 10 m. Outer hair cell mechano-transduction and electro-
physiology: Two active forces were incorporated with the outer hair cellsthe force originating from
mechano-transduction in the hair bundle (fMET) and the electromotive force of the cell membrane (fOHC).
(C) Interaction between the three dynamic systems.

Page 2 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 are greater than the longitudinal beams (see Table S1 in Supporting Information). The basilar membrane
2 is hinged along the spiral limbus and clamped along the spiral ligament. The edge of the tectorial
3 membrane along the spiral limbus is clamped. The apical and the basal extremities of the OCC (x = 0 and
4 12 mm) are clamped. For a more detailed description of the FE analysis, see (26).

5 The mechano-transduction and electro-motility of the outer hair cells incorporated in this work is the same
6 as in several previous studies (20, 27, 28). Compared to our previous studies, the mechno-transduction
7 sensitivity was increased slightly (the gating swing of the transduction channel was increased to 0.7 nm).

8 Fluid-structure interaction
9 The effect of fluid viscosity is incorporated into this study, distinguishing it from the previous work (20).
10 As a result, the momentum conservation and continuity equations are solved independently (Eq. 1-3). The
11 viscous fluid space is discretized into 2-D grid (grid size of 20 m). After assuming small amplitude
12 harmonic motion at radial frequency , the 2-D Navier-Stokes and continuity equations for an
13 incompressible viscous fluid are

14 ju 2u p / x 0 , (1)

15 jv 2v p / z 0 , (2)

16 u / x v / z 0 , (3)

17 where 2 is the Laplace operator, j is the imaginary unit, p is the fluid pressure, u and v are the
18 longitudinal and transverse velocity components and and are dynamic viscosity and density of the
19 fluid. No-slip boundary conditions are applied along all fluid-structure boundaries. This no-slip boundary
20 condition, which cannot be enforced for an inviscid fluid, creates a velocity gradient normal to the wall.
21 The velocity gradient is responsible for the energy dissipation in the viscous fluid. The longitudinal fluid
22 velocity at the oval window defines the input stimulation. At the round window, the fluid pressure is zero
23 (pressure release boundary) and the longitudinal velocity gradient is zero. Along the top and bottom walls
24 of the fluid chambers and along the wall next to the helicotrema, the fluid velocity is zero. Because the 2-
25 D fluid-domain interacts with 3-D structural domain, two longitudinal lines of the OCC represent the
26 interacting surfaces: the edge of the tectorial membrane, and the center-line of the basilar membrane.
27 Along the fluid-OCC interacting surfaces, the fluid velocity is determined by the velocity of the
28 interacting OCC structures. Or,

29 u 0 , and (4)
30 v a32 j xIS , (5)

31 where xIS is the z-displacement at the interacting surface of the OCC. The parameter a32 accounts for 3-D
32 to 2-D conversion assuming the half-sine radial velocity profile of the interacting surface. The parameter

Page 3 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 is the ratio between the basilar membrane width and the cochlear chambers width. The conversion
2 satisfies conservation of energy on the fluid-structure interface: the rate of work done within the fluid
3 along the interacting lines equals the rate of work done by the pressure forces on the OCC.

4 The pressure force acting on the OCC along the interacting lines is
5 f fluid a23 AIS p , (6)

6 where AIS is the surface area of the OCC at the location of contact with the fluid. The parameter a23
7 converts the pressure of the 2-D fluid domain into nodal forces acting on the 3-D structural domain.

8 Dissipating components
9 Dissipation-related parameters will be described later in Results and Discussion, but a brief summary is
10 given here: The acoustic energy is dissipated both in the fluid and solid domains. The dissipation in the
11 viscous fluid domain will be discussed later in the Results section. For the solid domain (the FE model),
12 the damping matrix is defined by multiplying element stiffness matrices by coefficients (stiffness-
13 proportional Rayleigh damping coefficients). In choosing the damping coefficients, three structures are
14 independently considered. First, the viscous dissipation in the sub-tectorial space is analytically derived
15 and added to the dissipating component of the stereocilia bundle of the outer hair cells. Second,
16 reasonable viscoelastic properties of the outer hair cells are determined based on the literature (13). The
17 details of these two components will be presented in the Discussion section. Third, the viscoelastic
18 properties of the tectorial membrane are assigned after comparing with existing experimental data. The
19 tectorial membrane simulation results are presented in the Results section. For other organ of Corti
20 structures, the damping coefficients varying along the length were applied. The standard coefficients were
21 2.4 and 180 s-1 at x = 2 and 10 mm, respectively. Various parametric study results are presented in the
22 Results section.

23 Electro-physiology of the outer hair cell

24 The mechano-electrical transduction theory from previous studies (29, 30) was used to obtain the
25 relationship between the normalized transduction current (po) and hair bundle displacement (xHB).
26 According to the gating compliance theory (31), the action of transduction channels results in a force
27 (fMET) proportional to the transduction current, or
28 f MET g MET po xHB , (7)

29 where, gMET is the maximum gating force that occurs in the stereociliary bundle. The constant gMET ranges
30 from 100 pN at the cochlear base to 10 pN at the apex (29). The fMET is applied as a pair of equal and
31 opposite forces at the tip and base of the hair bundle so that it deflects the bundle. The outer hair cells
32 somatic force, fOHC, is proportional to the change of the transmembrane voltage, Vm

Page 4 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 fOHC gOHCVm po , (8)

2 where gOHC is a constant representing the outer hair cells electro-mechanical gain. A constant value of
3 gOHC = 0.1 nN/mV (5) was used independent of the location. The outer hair cells somatic force fOHC was
4 applied along the axis of the outer hair cell as a pair of equal and opposite forces at the extremities of the
5 cell body so that depolarization (positive Vm) contracts the cell. In this paper, the term active/passive
6 refers to the state with/without outer hair cells active feedback, i.e. gMET = gOHC = 0 when passive.

7 Further details on the relationship between xHB, po and Vm, and the related model properties are presented
8 in our previous paper (20).

9 Numerical implementation
10 For small harmonic excitation at radial frequency , the discretized equation of motion of the OCC can
11 be written in terms of stiffness, mass and damping matrices, K, M and C, respectively, as

12 (K 2M j C)x f fluid fOHC f MET . (9)

13 where x is the column vector of nodal displacements, and ffluid, fOHC and fMET are force vectors due to fluid
14 pressure, outer hair cells somatic force and stereociliary bundle active force, respectively.

15 Viscous fluid mechanics, structural mechanics and outer hair cell electro-physiology were solved
16 simultaneously in the frequency domain. Unknown variable vectors are: fluid pressure p, fluid
17 longitudinal and transverse velocity u and v, OCC nodal displacement x, electro-physiological variables e
18 (po and Vm of the outer hair cells).

19 The governing equations (Eq. 1, 2, 3 and 9), fluid-structure interaction relationship (Eq. 4 and 5), and all
20 the forces (Eq. 6, 7, and 8) can be written in combined matrix form as

A f 11 0 A f 13 0 0 u b1
0 A f 22 A f 23 A fx1 0 v 0

21 A f 31 A f 32 0 A fx 2 0 p b 2 . (10)

0 0 A xf A xx A xe x 0
0 0 0 A ex A ee e 0

22 Submatrices on the left hand side of the Eq. 10 are discretized forms of the governing equations. Af11 and
23 Af22 represent the first two terms in the Navier-Stokes equations (Eq. 1 and 2, respectively). Af13 and Af23
24 represent the third term (the pressure gradient term) in the equation Eq. 1 and 2, respectively. Afx1
25 represents the Navier-Stokes equation boundary condition along the OCC-fluid interacting surfaces, Eq.
26 5. Af31 and Af31 represent the first and second terms of the continuity equation, Eq. 3, respectively. Afx2
27 represents the fluid-structure interaction of Eq. 5. Axf represents the force boundary condition of the OCC
28 due to fluid pressure, Eq. 6. Axx represents the OCC mechanics equation, Eq. 9. Axe defines the outer hair

Page 5 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 cell forces acting on the OCC according to Eq. 7 and 8. Aex and Aee represent the relationship between the
2 outer hair cells stereociliary bundle displacements and mechano-transduction currents. The right hand
3 side of Eq. 10 is the vector representing the oval and round window boundary conditions: b1 for Navier-
4 Stokes equation, b2 for the continuity equation.

5 The program is written in Matlab (MathWorks, Natick MA). No specific toolbox was used. The
6 assembled system has approximately 220,000 degrees of freedom. In a personal computer (Intel i7-6700
7 processor, 16 GB memory), it takes about 10 seconds to solve for one frequency. The majority of time is
8 spent on updating frequency dependent components of the submatrices in Eq. 10. The code is available
9 upon request.
10

Page 6 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 RESULTS
2 Typical model responses are presented in Fig. 2. The contributions of different dissipating components
3 are analyzed in Fig. 3. To investigate how model parameters affect cochlear power dissipation, and to
4 support our choice of key model parameters relevant to power dissipation, simulation results are
5 compared with experimental measurement such as tectorial membrane viscoelastic vibrations (Fig. 4) and
6 cochlear input impedance (Fig. 5). A series of parametric studies show the effects of model parameters on
7 the cochlear input impedance (Fig. 6), the shape of traveling waves (Fig. 7), and the tuning and
8 amplification (Fig. 8).

9 Fluid viscosity minimally affects model responses

10 A general way to determine dissipating parameters for a cochlear model is to compare tuning and
11 amplification characteristics with physiological observations. Representative model responses to pure
12 tone stimulations are shown in Fig. 2. Several features commonly observed in physiological
13 measurements are reproduced by our model. For example, the active responses are more sharply tuned
14 and amplified compared to passive responses (Fig. 2A). Unlike the gain plots, the phase-frequency
15 relationships are similar regardless of the active feedback of outer hair cells (solid and broken curves, Fig.
16 2B). The amplification and tuning quality are comparable to known values of the gerbil cochlea (Fig. 2C
17 and D) (32-34).

Figure 2. Pure tone responses. (A) Magnitude and (B) phase of the basilar membrane displacement
gain. The gain is with respect to the stapes displacement. Solid and broken curves are active and
passive responses, respectively. Filled and empty circles indicate the phase at the peaks of the active
and passive responses, respectively. In (A), the red arrow defines the amplification. (C) Amplification
level along the distance from the base. (D) Tuning quality along the distance from the base. Results
from viscous and inviscid fluids simulations are shown together. (E) Pressure, streamlines and particle
traces resulting from the passive model subjected to 0.6 kHz stimulation, and (F) the active model
subjected to 18.5 kHz stimulation (bottom). Plots in panels (E) and (F) are obtained for the maximum
inward stapes velocity of 1 m/s. The color bar indicates pressure level in Pa.

Page 7 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 Fluid viscosity minimally affected the simulated responses. The results in Fig. 2A were obtained from the
2 analysis with viscous fluid. Similar results were presented previously for the inviscid model (20). The
3 amplification and tuning quality of the viscous and the inviscid models were compared in Fig. 2B and C.
4 The difference between the two sets of simulations is ascribed to modeling artifacts (e.g. differences in
5 fluid domain variables, meshing, and boundary conditions), rather than to true physical differences due to
6 fluid viscosity. For example, decreasing fluid viscosity by one order of magnitude had a negligible effect
7 on the simulation results. All the results hereafter are from the analyses with the viscous fluid. The
8 trajectory of fluid particles depends on the stimulating frequency. For high frequencies (Fig. 2 F), fluid
9 particle motions were circular, while for low frequencies (Fig. 2E), they were elliptical, elongated along
10 the longitudinal direction. When passive, the maximum pressure amplitude occurs at the oval window.
11 When active, the maximum pressure occurs near the cochlear partition. This implies that the outer hair
12 cell active feedback can overwhelm the input from the stapes (10).

13 Primary power dissipation occurs in the OCC, not in the scala fluid
14 To investigate where power dissipation occurs in the cochlea, the total power dissipation was divided into
15 different components (Fig. 3). First, the total power dissipation is divided into the dissipation in the fluid
16 domain and structural domain (or PTotal =Pfluid + POCC). Then, the structural dissipation is divided again
17 into the dissipations in the tectorial and basilar membranes and in the organ of Corti (or POCC = PTM-BM
18 +POoC). Dissipated power is obtained from the rate of work done by the dissipating forces averaged over
19 one cycle.

20 Dissipated power in the OCC is calculated from

21 POCC 0.5 v CT
OCC C v OCC , (11)

Figure 3. Power dissipation in the cochlea. The

cochlear model was subjected to a series of pure tone
simulations through the stapes (velocity amplitude of
1 m/s regardless of frequency). (A) Total power
dissipation in the cochlea for the active and passive
responses. The total dissipated power was divided
into three components depending on where the
dissipation occurs: in the scala fluid (Fluid), in the
organ of Corti (OoC), and in the basilar and tectorial
membranes (BM-TM). Those three components
were normalized by the total dissipated power, and
plotted versus frequency for the (B) Active and (C)
Passive cases.

Page 8 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 where C is damping matrix of the finite element model, and vOCC is the OCC velocity. The superscript CT
2 indicates conjugate-transpose. Power dissipation in the scala fluid per cycle is calculated from the volume
3 integral of the dissipation function

4 PFluid b x dxdz , (12)

5 where is the dissipation function, b(x) is the width of the basilar membrane, and is the ratio between
6 the widths of the fluid chamber and basilar membrane. The dissipation function is the power dissipation
7 per unit volume defined as

u u CT v v CT u v u v CT
8 .
2 z x z x (13)
x x z z

9 To show the relative contribution of different dissipating mechanisms, the three dissipation components
10 in the active and the passive cochlea were computed (Fig. 3B and C). The dissipation in the scala fluid
11 was negligible, except when the passive cochlea was simulated at low (< 2 kHz) frequencies.

12 There are several differences between the power dissipation patterns in the active and passive cochlear
13 models. First, amplified vibrations in the active case result in greater power dissipation. When subjected
14 to the same stapes velocity, the active cochlea dissipates 102 to 105 times more power than the passive
15 case (Fig. 3A). This greater dissipation in the active case is mostly ascribed to the increased vibration
16 amplitude (between 15 and 50 dB, Fig. 1B). Second, unlike the passive case where the total dissipation
17 remained at a similar level over different stimulating frequencies, in the active case, the power dissipation
18 increased as the stimulating frequency increases from 1 to 20 kHz. Third, the fraction of power dissipated
19 in the organ of Corti (labeled as OoC in Fig. 3B and C) was greater when active. The increased
20 dissipation of the active organ of Corti is ascribed to the increased dissipation of the outer hair cells due to
21 their motility.

22 It must be noted that our power dissipation analysis depends critically on model parameters, especially on
23 the OCC damping properties. Theoretical studies including ours chose the damping properties so that the
24 model responses (amplification and tuning quality) become comparable to experimentally measured
25 values. In the following, we show that our chosen damping properties are reasonable as compared to
26 existing experiments, and present how different damping values affect the model responses.

27 Comparing the viscoelastic response of the tectorial membrane with experiments

28 The viscoelastic responses of a section of the tectorial membrane were simulated and compared to
29 experimental measurements to assess our choice of dissipation parameters. The tectorial membrane
30 accounts for approximately a quarter of the cross-sectional area of the OCC (35). The viscoelastic
31 responses of the tectorial membrane have been investigated for the rodent cochlea (36, 37). Ghaffari and

Page 9 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 his colleagues applied sinusoidal stimulation to one end of the tectorial membrane section, excised from
2 the mouse cochlea. From the vibrating patterns of the tissue, the wavelength and decaying space constant
3 were measured.

4 The experimental results were reproduced with the tectorial membrane components of our model (Fig. 4).
5 A 1 mm long section of the tectorial membrane was simulated, centered at x = 2 mm where the best
6 responding frequency is comparable to the experiment (18 kHz). The tectorial membrane model consists
7 of beams running in the longitudinal and radial directions. Following the experiment, one short end of the
8 tectorial membrane was subjected to time harmonic radial displacement while the other end was fixed.
9 The tectorial membrane formed propagating waves with amplitude decaying with distance from the
10 excited edge. Three parameters determine the wavelength and decay space constant of the propagating
11 waves ( and in Fig. 4C). They are the elastic modulus of longitudinally running elements, the damping
12 coefficient and the mass. The mass property was assigned using available anatomical information (1.4 g
13 for 1 mm section, estimated from a microscopic image in (38)). Within the tested parameter range, there
14 exists a unique combination of Youngs modulus and damping parameter that fits best with the
15 experimental results (Fig. 4A and B). Our standard model properties were in reasonable agreement with
16 the best-fit properties. The Youngs modulus and the damping coefficient of our model are 1/2 and 1/3 of
17 the best-fit properties (filled circles, Fig. 4C and D).

Figure 4. Viscoelastic properties of the

tectorial membrane. FE model of the
tectorial membrane simulated the experiment
of Ghaffari et al. (2007). See text for
simulation details. (A) The response of a
tectorial membrane section subjected to radial
vibrations at its best frequency (18 kHz).
Simulated vibrating pattern (solid curve) is
presented together with the experimental data
(). The decaying space constant () and the
wavelength () represent the viscoelastic
properties of the tectorial membrane. (B)
Wavelength as the stimulating frequency
increases. (C) Wavelength as a function of
damping and stiffness properties. The Youngs modulus of longitudinally running elements represents
the stiffness. (D) Space constant as a function of damping and longitudinal coupling properties. The
damping and stiffness properties were normalized with the values used to fit the experimental data. The
standard parameters of our study are indicated with filled circles in (C) and (D).

Page 10 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 This comparison shows that the damping parameters used to obtain the OCC power dissipation are
2 reasonable and possibly underestimated. This assessment of tectorial membranes damping property
3 corroborates the conclusion of Fig. 3 that the OCC is the primary place of cochlear power dissipation.

4 Cochlear input impedance does not reflect how the energy is dissipated in the cochlea
5 When considering a vibrating mechanical system, the real and imaginary parts of its impedance represents
6 dissipating and conservative component in the system, respectively. To investigate the dissipating and
7 conservative components of cochlear mechanics, the input impedance of our cochlear model was
8 compared to existing experimental data for verification, then a series of parametric studies was
9 performed. To conclude in advance, our computational study confirms the analytic prediction of
10 Zwislocki (39)the cochlear input impedance is not affected by damping properties.

11 In Fig. 5, computed input impedance was compared to an experimental estimation from intra-cochlear
12 pressure measurements by de la Rochefoucauld and her colleagues (8). The cochlear input impedance (Z0)
13 is the ratio between input pressure (p0) and the resulting volume velocity (U0) at the pressure entrance (the
14 stapes) of the cochlea,
15 Z 0 p0 /U 0 . (14)

16 The volume velocity was obtained by multiplying the stapes velocity with the oval window area of 0.8
17 mm2 (Mongolian gerbil cochlea, 40). Our simulated result is in agreement with the measured data. For
18 example, the computed impedance amplitude of 5-71010 Pas/m3 is comparable to the measured values
19 of 3-61010 Pas/m3. The amplitude remains flat over the audible frequency range. The pressure was in
20 phase with the stapes velocity (within 15 degrees over the simulated frequency range between 0.5 and 25
21 kHz). At stimulating frequencies below 0.5 or above 25 kHz, the simulated results were unreliable as they

Figure 5. Cochlear input impedance. Simulated

cochlear input impedance is shown together with
the experimental results by de la Rochefoucauld et
al. (8). In the experiment, intra-cochlear fluid
pressure near the stapes, and the stapes velocity
were measured to obtain the input impedance. The
same condition was simulated with our model. (A)
Magnitude, and (B) phase of the cochlear input
impedance. Presented curves are obtained from the
passive cochlear simulation, but the active results
are similar to these curves.

Page 11 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 were strongly affected by the boundary conditions. This simulation is with the passive cochlea, which has
2 the best frequencies at the basal and apical end of about 25 kHz, and 0.5 kHz respectively. The impedance
3 amplitude and phase for the active cochlea are similar to those for the passive cochlea (data now shown).

4 Neither structural damping nor the fluid viscosity affects the cochlear input impedance (Fig. 6).
5 According to the literature (39, 41, 42), the cochlear input impedance is related to the energy used to
6 generate slow traveling waves propagating toward the apex of the cochlea. When the energy of these
7 waves is not reflected back, the wave energy must be dissipated in the cochlea. In that sense, the cochlear
8 input impedance carries information about the energy entering the cochlea. Although the cochlear input
9 impedance is resistive in that it is in phase with the stapes velocity, it is not determined by the
10 dissipating terms of the cochlea such as the structural damping and fluid viscosity. Instead, it is primarily
11 determined by inertial and elastic components such as the fluid mass density (), and the OCC
12 compliance near the stapes (C0), or Z 0 0 S0 / C0 (39). Here, S0 is the area of the oval window.

13

Figure 6. Cochlear input impedance does not represent

the dissipating components. A series of parametric studies
were performed to test how stiffness, damping and inertial
parameters of the model affect the cochlear input impedance.
The parametric study is done by varying one of the
parameters over 4 orders-of-magnitude range. (A) Magnitude
and (B) phase of the cochlear input impedance evaluated at 3
kHz stimulating frequency as a function of OCC damping,
cochlear scalae fluid viscosity and density and OCC stiffness.

Page 12 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 Existence of equivalent parameter set with reduced structural mass and damping
2 We investigated whether there exists an equivalent parameter set that results in the same tuning quality.
3 According to vibration theory for an under-damped mechanical system, the tuning quality is inversely
4 related to the damping factor , or Q 0.5/ 0.5 / , where the c, m and k represent the
5 damping, mass and stiffness of the system, respectively. It is possible that the organ of Corti tissue or the
6 tectorial membrane is so soft that they behave more like a fluid than a solid, or that the structural mass of
7 the organ of Corti is negligible. Indeed, there are studies that neglected the OCC structural mass (17, 43).
8 Had we overestimated the structural mass so that structural mass became more dominant than fluid
9 inertia, the structural damping property could be overestimated, too. We investigated whether the power
10 dissipation in the OCC was exaggerated by incorporating excessive structural mass.

11 We found that there was no equivalent parameter set of reduced mass and damping that results in the
12 same tuning quality. Regarding the tuning quality, decreasing the damping factor does not necessarily
13 increase the tuning quality. In one series of simulations (Fig. 7A), the damping was reduced by a factor of
14 4, and different structural masses were tested. In the other series of simulations (Fig. 7B), the structural
15 mass was reduced by a factor of 16, and different levels of damping were tested. With the reduced
16 structural mass, we were unable to achieve high tuning quality despite lowering the damping.

17 Increased spreading of wave envelope in the low damping-low mass case is responsible for the reduced
18 tuning quality. In Fig. 8, the traveling waves that peak at x = 3.5 mm in the passive and the active
19 cochlear model are shown. When damping and mass are reduced, as quite expected, the amplitude of
20 vibrations increases. Because the vibration amplitude increases both in the passive and the active cochlea
21 models, the amplification factor can either increase or decrease (top panels of Fig 7A and B) as compared
22 to the standard case. With the reduced damping and mass properties, the wavelength of cochlear traveling
23 wave decreases (thin red curves). Notably, the acoustic energy disperses over a wider span. The darker

Figure 7. Effect of mass and damping

properties on amplification and tuning. It was
examined if there exists an equivalent reduced
mass-damping set that results in similar
amplification and tuning levels to the standard
case. Amplification (in dB re. the passive case) and
tuning quality (in Q10dB) were obtained for the
best-responding frequencies along the distance (x).
(A) Effect of different masses. One fourth of
standard damping property was used, and three
different mass properties were simulated. (B)
Effect of different damping levels. One sixteenth
of standard mass value was used and three
different damping properties were simulated.

Page 13 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 shaded spans indicate the 3 dB bandwidth of the traveling wave envelope. The reduced damping-mass
2 resulted in broadened bandwidth of the traveling waves, when the cochlea is active.

3 Reduced structural mass has greater effect toward the basal end. For example, as the mass decreases by
4 factor of 4, the best responding frequency at x = 2 mm increased from 18.6 kHz to 23 kHz, while at x = 10
5 mm it remains at 0.7 kHz (result not shown). This is in agreement with the theory that fluid mass becomes
6 more dominant in the low frequency locations where the wavelength of traveling waves is longer (44),
7 and with the observation that structural mass contributes to cochlear mechanics in the basal turn of the
8 cochlea (45).

Figure 8. Reduction of mass and damping

broadens the traveling wave span. Traveling
patterns of the passive (top) and active responses. In
addition to the standard responses (thick black
curves), a cochlear model with reduced mass-
damping (thin red curves) was simulated. For the
test case, the mass and damping properties were
reduced to 1/16 and 1/4 of the standard values,
respectively. The shaded areas indicate the wave
envelopes.

9 Dissipation can enhance tuning quality

10 A parametric study on the damping property clearly reveals a difference between passive and active
11 cochlear mechanics (Fig. 9). With all other model parameters fixed, only the damping properties were
12 changed over two orders of magnitude around the standard value. When passive, as the damping increases
13 the envelope of traveling wave gets wider (Fig. 9A and C), which corresponds to decreasing quality factor
14 (Fig. 9D). Interestingly, when the cochlea is active, the damping versus tuning relationship is not
15 monotonic. As the damping increases, the bandwidth of traveling wave decreases and then increases (Fig.
16 9B and C). That is, our study predicts that there exists an optimal damping property for best tuning
17 quality (Fig. 9D). For the shown frequencies that peak at x = 2 mm, the optimal property is close to the
18 standard value. The spatial bandwidth of 0.34 mm of our study is 40 percent greater than the experimental
19 observation of 0.24 mm by Ren (46).

20 The difference between the passive and the active cases originates from different modes of acoustic
21 energy transmission. In the passive case, the acoustic energy enters at the oval window and dissipates

Page 14 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 monotonically toward the apex. At reduced damping, the energy travels further, or in the aspect of a fixed
2 location, the best responding frequency increases and wavelength gets shorter. In the active case, in
3 addition to the acoustic input at the stapes, the energy is generated within the organ of Corti. This
4 generated energy can be much greater than the input energy (10). The stream line patterns and pressure
5 field in Fig. 3C show this energy flow. In the passive case (top panel), the greatest pressure amplitude
6 occurs at the oval window (x = 0). But, in the active case (bottom panel), the greatest pressure amplitude
7 is at the best responding site. The dissipation of the OCC has two contradictory effects. In one, it opposes
8 the active power generation. In the other, it arrests the dispersion of energy from the generation site. Until
9 a certain level of dissipation, the latter role (arresting the dispersion) is more prominent so that the tuning
10 quality increases as the damping level increases. However, above the optimal damping level, the damping
11 begins to suppress the active power generation itself.
12

13

14

15

Figure 9. Optimal damping for tuning quality.

Different damping levels were simulated. (A)
Passive case: Normalized spatial envelopes of the
basilar membrane vibration when the model is
subjected to a pure tone. The labels indicate relative
damping compared to the standard value. As the
damping decreases, the envelope becomes
narrower. (B) Active case. (C) Spatial bandwidth at
3 dB below the peak along different damping levels.
(D) Tuning quality along different damping levels.

Page 15 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 DISCUSSION
2 The most prominent conclusion of this study is that OCC damping is necessary to achieve high tuning
3 quality (Figs. 8 and 9). We found that there exists an optimal level of damping for tuning quality (Fig. 9).
4 Our results support the hypothesis that acoustic power in the cochlea is dissipated within the OCC rather
5 than in the scala fluid (Fig. 3). The dissipation in the viscous scala fluid becomes comparable to the OCC
6 dissipation only when the passive cochlea is subjected to low frequency stimulations (< 0.5 kHz). As a
7 result, cochlear amplification and tuning are minimally affected by the viscosity of the extra-OCC fluid
8 (Fig. 2B). To justify the choice of OCC damping properties, we compared viscoelastic responses of the
9 tectorial membrane with experimental results (Fig. 4). Both the comparison with the measured data and
10 the parametric study suggest that the OCC damping properties of this study are reasonable. In the
11 following, we discuss the damping properties in the literature, and present several first-order
12 approximations of different dissipating mechanisms.

13 Damping properties in the literature

14 Damping properties used in different theoretical studies are summarized in Table 1. Nearly all other
15 studies used fewer degrees-of-freedom to
16 represent the cochlear mechanics than used in Table 1. Damping coefficients in the literature. Unit in
(kNs/m3)
17 our study, but all models have a degree-of- Reference Case: Base, Apex Species
18 freedom representing the transverse BM: 15, 0.3
Neely & Kim, 1986 TM: 0.1, 0.0004 Cat
19 displacement of the basilar membrane. OoC: 0.02, 0.003
20 Therefore, for the sake of comparison, our Diependaal et al. 1987 5, 0.03 Human
Mammano & Nobili
21 OCC damping can be converted to an 1.6, 0.38 Guinea pig
1993
22 effective damping coefficient with respect to Kolston & Ashmore BM: 2.0
Human
1996 TM: 17~170
23 the basilar membrane vibrations. The BM: 1.9, 13
Lu et al., 2006 Guinea pig
24 effective damping coefficient of the OCC per RL: 0.5, 3.5
Liu & Neely, 2010 15, 86 Human
25 10 m-long cross-section is computed from BM: 0.85, 0.43
Meaud & Grosh, 2012 Guinea pig
26 the dissipated power within the section, TM: 1.5, 0.75
Passive: 2.2 to 0.23
27 POCC, and the basilar membrane velocity, This study Gerbil
Active: 7.5 to 0.92
28 vBM.
29 cOCC 2POCC / v CT
BM v BM (15)

30 Note that, because this effective property is dependent on vibration patterns, the value is dependent on
31 vibrating frequency and active outer hair cell feedback. When passive (i.e. when all the elements in the
32 cross-section vibrate approximately in phase), the values of cOCC range from 3.5 Ns/m at the base, to
33 0.64 Ns/m at the apex. When considered per unit basilar membrane area, these values correspond to 2.2
34 and 0.23 kNs/m/m2. When active, increased relative (out-of-phase) motion within the organ of Corti

Page 16 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 increases overall dissipation. The effect of the relative motion increases as stimulating frequency
2 decreases as shown in Fig. 3B. As a result, when active, the value of cOCC increases to 12 Ns/m at the
3 base and to 2.6 Ns/m at the apex.

4 As summarized in Table 1, theoretical studies have used a wide range of dissipating properties. Some
5 studies used greater damping value toward the base, others considered the opposite trend. In a model not
6 included in the table (10), the intra-OCC damping was not incorporated. Instead, they incorporated fluid
7 viscosity as the dissipating mechanism. It is clear that good agreement does not exist on dissipating
8 properties of the cochlea. Considering that the dissipating property is crucial to reconcile different
9 hypothetical mechanisms of cochlear amplification, more quantitative experimental data regarding the
10 organ of Corti viscoelastic responses will be very valuable in settling current debates on the operating
11 mechanisms of cochlear mechano-transduction.

12 Origin of power dissipation in the OCCRough estimation of OCC damping

13 Although it is difficult to make an accurate estimation of OCC damping, at least potential dissipating
14 mechanisms within the OCC can be discussed. In the following, three mechanisms are listedthe
15 dissipation in the sub-tectorial space, the dissipation due to fluid flow within the organ of Corti, and the
16 dissipation in viscoelastic cells.

17 First, the sub-tectorial space (the micrometer-thick fluid layer between the tectorial membrane and the
18 reticular lamina) has been considered to be a major location for energy dissipation (47). If the sub-
19 tectorial space is represented with a fluid layer between a pair of parallel plates, the damping coefficient is
20 determined by the viscosity and geometry as cSTS = Lw/h, where L, w and h are the length, width and
21 height of the space. The value may range between 0.1 and 1 Ns/m for a length of L = 10 m (the
22 longitudinal spacing between outer hair cells). In our recent study (16), it was shown that this dissipation
23 in the sub-tectorial space can further increase at low frequencies due to the additional viscous friction of
24 the inner hair cell stereocilia bundle. In the present work, we incorporated only the simplistic parallel
25 plate estimations (cSTS). Therefore, the damping in the sub-tectorial space could be underestimated,
26 especially for low stimulating frequencies.

27 Second, deformation of the organ of Corti due to outer hair cell motility may induce viscous friction.
28 Recent measurements (48, 49) clearly demonstrated that there exists considerable out-of-phase motion
29 between the top and bottom surfaces of the organ of Corti. This deformation may be substantial enough to
30 induce longitudinal fluid flow along the tunnel of Corti (50). According to the mechanical impedance of
31 the pulsating flow of viscous fluid in a pipe (51), the effective damping coefficient along the tunnel of
32 Corti can be approximated with cTC rL 2 , where r is the effective radius of the tube.
33 Considering the geometry and best frequency of different locations, the estimated damping coefficient
34 will be a fraction of 1 Ns/m for a length of L = 10 m. This value will define a lower limit because

Page 17 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 there must be other fluid motion within the organ of Corti besides the longitudinal flow along the tunnel
2 of Corti.

3 Finally, deforming viscoelastic cells in the organ of Corti will dissipate power. In particular, the outer hair
4 cells themselves must deform because of their electromotility. Based on the estimation of viscoelasticity
5 of a red blood cell membrane (52), the axial damping coefficient of an outer hair cell was estimated to be
6 on the order of 0.1 Ns/m for a 60 m-long cell (13). Considering three rows of outer hair cells, this
7 corresponds to cOHC = 0.3 Ns/m per 10 m section. In the present work, cOHC ranges between 0.3
8 Ns/m in the base and 0.75 Ns/m in the apex. This study does not incorporate any supporting cells in
9 the organ of Corti other than the Deiters cell and the pillar cells where further dissipation can occur.

10 These three damping coefficients correspond to different motionssub-tectorial space shear velocity,
11 longitudinal fluid flow in the tunnel of Corti, cellular deformations in the organ of Corti. Therefore, for
12 better comparison with those values in Table 1, they need to be converted with pertinent kinematic factors
13 relating to the basilar membrane velocity. Based on available evidence, the kinematic factors seem not far
14 from unity. For example, the stereocilia bundle displacement that equals the sub-tectorial space shear
15 displacement, and the outer hair cell somatic deformation are comparable to the basilar membrane
16 displacement (20, 27, 49, 53).

17 These rough estimations are on the order of 1 Ns/m per 10 m section, supporting our conclusions that
18 the major power dissipating place in the cochlea is the OCC, and that the power dissipation depends on
19 vibration modes of the OCC modulated by the outer hair cell motility.

20

21 AUTHOR CONTRIBUTIONS
22 SP and JN conceived the research. SP and JN wrote and run the computer programs. SP, SG, and JN
23 analyzed the data, and wrote the paper.

24

25 ACKNOWLEDGEMENTS
26 This work was supported by NIH NIDCD R01 DC014685.

27

28
29

Page 18 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 REFERENCES
2 1. Olson, E. S., H. Duifhuis, and C. R. Steele. 2012. Von Bekesy and cochlear mechanics. Hear Res
3 293:31-43.

4 2. Gold, T. 1948. Hearing .2. The physical basis of the action of the cochlea. Proc R Soc Ser B-Bio
5 135:492-498.

6 3. Ashmore, J. 2008. Cochlear outer hair cell motility. Physiol Rev 88:173-210.

7 4. Hallworth, R. 1995. Passive compliance and active force generation in the guinea pig outer hair cell.
8 J Neurophysiol 74:2319-2328.

9 5. Iwasa, K. H., and M. Adachi. 1997. Force generation in the outer hair cell of the cochlea. Biophys J
10 73:546-555.

11 6. Overstreet, E. H., 3rd, A. N. Temchin, and M. A. Ruggero. 2002. Basilar membrane vibrations near
12 the round window of the gerbil cochlea. J Assoc Res Otolaryngol 3:351-361.

13 7. Puria, S. 2003. Measurements of human middle ear forward and reverse acoustics: implications for
14 otoacoustic emissions. J Acoust Soc Am 113:2773-2789.

15 8. de la Rochefoucauld, O., W. F. Decraemer, S. M. Khanna, and E. S. Olson. 2008. Simultaneous

16 measurements of ossicular velocity and intracochlear pressure leading to the cochlear input
17 impedance in gerbil. J Assoc Res Otolaryngol 9:161-177.

18 9. Ramamoorthy, S., and A. L. Nuttall. 2012. Outer hair cell somatic electromotility in vivo and power
19 transfer to the organ of Corti. Biophys J 102:388-398.

20 10. Wang, Y., C. R. Steele, and S. Puria. 2016. Cochlear Outer-Hair-Cell Power Generation and Viscous
21 Fluid Loss. Sci Rep-Uk 6:19475.

22 11. Tolomeo, J. A., and C. R. Steele. 1998. A dynamic model of outer hair cell motility including
23 intracellular and extracellular fluid viscosity. J Acoust Soc Am 103:524-534.

24 12. Rabbitt, R. D., S. Clifford, K. D. Breneman, B. Farrell, and W. E. Brownell. 2009. Power efficiency
25 of outer hair cell somatic electromotility. PLoS computational biology 5:e1000444.

26 13. Liao, Z., A. S. Popel, W. E. Brownell, and A. A. Spector. 2005. Modeling high-frequency
27 electromotility of cochlear outer hair cell in microchamber experiment. J Acoust Soc Am 117:2147-
28 2157.

29 14. Iwasa, K. H. 2016. Energy Output from a Single Outer Hair Cell. Biophys J 111:2500-2511.

30 15. Steele, C. R., and S. Puria. 2005. Force on inner hair cell cilia. Int J Solids Struct 42:5887-5904.

31 16. Prodanovic, S., S. Gracewski, and J. H. Nam. 2015. Power dissipation in the subtectorial space of the
32 mammalian cochlea is modulated by inner hair cell stereocilia. Biophys J 108:479-488.

33 17. Lim, K. M., and C. R. Steele. 2002. A three-dimensional nonlinear active cochlear model analyzed
34 by the WKB-numeric method. Hear Res 170:190-205.

Page 19 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 18. Yoon, Y. J., C. R. Steele, and S. Puria. 2011. Feed-forward and feed-backward amplification model
2 from cochlear cytoarchitecture: an interspecies comparison. Biophys J 100:1-10.

3 19. Nam, J. H., Y. Liu, and S. M. Gracewski. 2014. A computational study on traveling waves in the
4 gerbil cochlea generated by electrical impulse. In Mechanics of Hearing 12th International
5 Workshop. D. Corey, and K. D. Karavitaki, editors, Attica, Greece.

6 20. Liu, Y., S. M. Gracewski, and J. H. Nam. 2015. Consequences of Location-Dependent Organ of
7 Corti Micro-Mechanics. PLoS One 10:e0133284.

8 21. Ramamoorthy, S., N. V. Deo, and K. Grosh. 2007. A mechano-electro-acoustical model for the
9 cochlea: response to acoustic stimuli. J Acoust Soc Am 121:2758-2773.

10 22. Meaud, J., and K. Grosh. 2014. Effect of the attachment of the tectorial membrane on cochlear
11 micromechanics and two-tone suppression. Biophys J 106:1398-1405.

12 23. Lamb, J. S., and R. S. Chadwick. 2011. Dual traveling waves in an inner ear model with two degrees
13 of freedom. Phys Rev Lett 107:088101.

14 24. Cormack, J., Y. Liu, J. H. Nam, and S. M. Gracewski. 2015. Two-compartment passive frequency
15 domain cochlea model allowing independent fluid coupling to the tectorial and basilar membranes.
16 The Journal of the Acoustical Society of America.

17 25. Dong, W., and E. S. Olson. 2016. Two-Tone Suppression of Simultaneous Electrical and Mechanical
18 Responses in the Cochlea. Biophys J 111:1805-1815.

19 26. Nam, J. H., and R. Fettiplace. 2010. Force transmission in the organ of Corti micromachine. Biophys
20 J 98:2813-2821.

21 27. Nam, J. H., and R. Fettiplace. 2012. Optimal electrical properties of outer hair cells ensure cochlear
22 amplification. PLoS One 7:e50572.

23 28. Nam, J. H. 2014. Microstructures in the organ of Corti help outer hair cells form traveling waves
24 along the cochlear coil. Biophys J 106:2426-2433.

25 29. Beurg, M., J. H. Nam, A. Crawford, and R. Fettiplace. 2008. The actions of calcium on hair bundle
26 mechanics in mammalian cochlear hair cells. Biophys J 94:2639-2653.

27 30. Nam, J. H., and R. Fettiplace. 2008. Theoretical conditions for high-frequency hair bundle
28 oscillations in auditory hair cells. Biophys J 95:4948-4962.

29 31. Howard, J., and A. J. Hudspeth. 1988. Compliance of the hair bundle associated with gating of
30 mechanoelectrical transduction channels in the bullfrog's saccular hair cell. Neuron 1:189-199.

31 32. Ren, T., and A. L. Nuttall. 2001. Basilar membrane vibration in the basal turn of the sensitive gerbil
32 cochlea. Hear Res 151:48-60.

33 33. Temchin, A. N., and M. A. Ruggero. 2010. Phase-locked responses to tones of chinchilla auditory
34 nerve fibers: implications for apical cochlear mechanics. J Assoc Res Otolaryngol 11:297-318.

35 34. Cooper, N. P., and W. S. Rhode. 1995. Nonlinear mechanics at the apex of the guinea-pig cochlea.
36 Hear Res 82:225-243.

Page 20 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 35. Edge, R. M., B. N. Evans, M. Pearce, C. P. Richter, X. Hu, and P. Dallos. 1998. Morphology of the
2 unfixed cochlea. Hear Res 124:1-16.

3 36. Ghaffari, R., A. J. Aranyosi, and D. M. Freeman. 2007. Longitudinally propagating traveling waves
4 of the mammalian tectorial membrane. Proc Natl Acad Sci U S A 104:16510-16515.

5 37. Ghaffari, R., A. J. Aranyosi, G. P. Richardson, and D. M. Freeman. 2010. Tectorial membrane
6 travelling waves underlie abnormal hearing in Tectb mutant mice. Nat Commun 1:96.

7 38. Cheatham, M. A., R. J. Goodyear, K. Homma, P. K. Legan, J. Korchagina, S. Naskar, J. H. Siegel, P.

8 Dallos, J. Zheng, and G. P. Richardson. 2014. Loss of the tectorial membrane protein CEACAM16
9 enhances spontaneous, stimulus-frequency, and transiently evoked otoacoustic emissions. J Neurosci
10 34:10325-10338.

11 39. Zwislocki, J. 1963. Analysis of some auditory characteristics. Special Report LSC-S-1, Laboratory of
12 Sensory Communication, Syracuse University, Syracuse, New York.

13 40. Cohen, Y. E., C. K. Bacon, and J. C. Saunders. 1992. Middle ear development. III: Morphometric
14 changes in the conducting apparatus of the Mongolian gerbil. Hear Res 62:187-193.

15 41. Koshigoe, S., W. K. Kwok, and A. Tubis. 1983. Effects of perilymph viscosity on low-frequency
16 intracochlear pressures and the cochlear input impedance of the cat. J Acoust Soc Am 74:486-492.

17 42. Puria, S., and J. B. Allen. 1991. A parametric study of cochlear input impedance. J Acoust Soc Am
18 89:287-309.

19 43. van der Heijden, M. 2014. Frequency selectivity without resonance in a fluid waveguide. Proc Natl
20 Acad Sci U S A.

21 44. Lighthill, J. 1981. Energy-Flow in the Cochlea. J Fluid Mech 106:149-213.

22 45. de La Rochefoucauld, O., and E. S. Olson. 2007. The role of organ of Corti mass in passive cochlear
23 tuning. Biophysical journal 93:3434-3450.

24 46. Ren, T. 2002. Longitudinal pattern of basilar membrane vibration in the sensitive cochlea. Proc Natl
25 Acad Sci U S A 99:17101-17106.

26 47. Mammano, F., and R. Nobili. 1993. Biophysics of the cochlea: linear approximation. J Acoust Soc
27 Am 93:3320-3332.

28 48. Lee, H. Y., P. D. Raphael, J. Park, A. K. Ellerbee, B. E. Applegate, and J. S. Oghalai. 2015.
29 Noninvasive in vivo imaging reveals differences between tectorial membrane and basilar membrane
30 traveling waves in the mouse cochlea. Proc Natl Acad Sci U S A 112:3128-3133.

31 49. Ren, T., W. He, and D. Kemp. 2016. Reticular lamina and basilar membrane vibrations in living
32 mouse cochleae. Proc Natl Acad Sci U S A.

33 50. Karavitaki, K. D., and D. C. Mountain. 2007. Evidence for outer hair cell driven oscillatory fluid
34 flow in the tunnel of corti. Biophys J 92:3284-3293.

35 51. Kinsler, L. E. 2000. Fundamentals of acoustics. Wiley, New York.

Page 21 of 24 December 21, 2016

 Power dissipation in the cochlea Prodanovic et al.

1 52. Waugh, R., and E. A. Evans. 1976. Viscoelastic properties of erythrocyte membranes of different
2 vertebrate animals. Microvasc Res 12:291-304.

3 53. Lee, H. Y., P. D. Raphael, A. Xia, J. Kim, N. Grillet, B. E. Applegate, A. K. Ellerbee Bowden, and J.
4 S. Oghalai. 2016. Two-Dimensional Cochlear Micromechanics Measured In Vivo Demonstrate
5 Radial Tuning within the Mouse Organ of Corti. J Neurosci 36:8160-8173.
6

Page 22 of 24 December 21, 2016