Effects of Cricothyroid and Thyroarytenoid Interaction On Voice Control: Muscle Activity, Vocal Fold Biomechanics, Flow, and Acoustics

15/10/2023 22:31 Effects of cricothyroid and thyroarytenoid interaction on voice control: Muscle activity, vocal fold biomechanics, flow,
cs, flow, and aco… 15/10/2023 22:31 Effects of cricothyroid and thyroarytenoid interaction on voice control: Muscle activity, vocal fold biomechanics, flow, and aco…
The human voice originates from complex neuromuscular

JULY 02 2021
contractions of intrinsic laryngeal muscles (ILM), which set the
Effects of cricothyroid and pre-phonatory posture (shape, position, tension, and stiffness) of
thyroarytenoid interaction on voice the vocal folds. The pre-phonatory posture, vocal fold tissue
control: Muscle activity, vocal fold properties, and airflow from the lungs interact to generate self-
biomechanics, flow, and acoustics  sustained oscillations of the vocal folds, which mechanically
Mohammadreza Movahhedi; Biao Geng; Qian Xue; Xudong Zheng convert expiratory airflow into pulsatile airflow, forming the
primary sound source of the voice.1–3
 Author & Article Information Despite the important role of vocal fold posturing in voice
J. Acoust. Soc. Am. 150, 29–42 (2021)
production, its relation with muscle activity and voice outcome is
https://doi.org/10.1121/10.0005275 Article history 
not fully understood. Early studies mostly employed in vivo
An MRI-based three-dimensional computer model of a canine electromyography (EMG) measures to relate muscle activity and
larynx was used to investigate the effect of cricothyroid (CT) and voice outcome;4–9 yet, the vocal fold posture was not measured
thyroarytenoid (TA) muscle activity on vocal fold pre-phonatory so that the overall system relation cannot be understood.
posturing and glottic dynamics during voice production. Static Furthermore, the EMG technique allows only coarse
vocal fold posturing in the full activation space of CT and TA measurement of muscle activity; thus, it cannot provide
muscles was first simulated using a laryngeal muscle mechanics systematic understanding. Chhetri et al.10 developed a technique
model; dynamic flow-structure-acoustics interaction (FSAI) using graded stimulation of ILMs for precise control of vocal fold
simulations were then performed to predict glottal flow and voice posturing and voice outcomes in in vivo canine larynges.11–20 A
acoustics. The results revealed that TA activation decreased the major limitation of this technique was that the observation was
length and increased the bulging, height, and contact area of the limited to an endoscopic view due to the inaccessibility of the
vocal fold. CT activation increased the length and contact area larynx; thus, the three-dimensional (3D) properties of vocal fold
and decreased the height of the vocal fold. Both CT and TA posture and vibration cannot be measured. While the limitation is
activations increased the vocal fold stress, stiffness, and closure possible to be overcome by a hemi-larynx setup,20 properties
quotient; and only slightly affected the flow rate and voice such as tension and stiffness of vocal fold tissues are still difficult
intensity. Furthermore, CT and TA showed a complex control to obtain in in vivo models.
mechanism on the fundamental frequency pattern, which highly
correlated with a combination of the stress, stiffness, and stretch Computational models can complement experiments by
of the vocal fold. providing measurements that are difficult in experiments. While
many computational models have been developed for simulating
Topics voice production, very few of them incorporated the relationships
Speech communication, Acoustics, Vocal folds, Larynx, between vocal fold posturing and voice production. For example,
Speech analysis, Speech production, Computer simulation, Bhattacharya and Siegmund21 introduced a three-dimensional
Magnetic resonance imaging, Musculoskeletal system, continuum vocal fold model and simulate the flow-structure
Biomechanics interactions (FSI) of vocal fold oscillations by solving Navier-
Stokes and equilibrium equations to evaluate the effects of
I. INTRODUCTION mechanical stress on the vocal fold dynamics and hydration.
https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 1/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 2/44
15/10/2023 22:31 Effects of cricothyroid and thyroarytenoid interaction on voice control: Muscle activity, vocal fold biomechanics, flow, and aco… 15/10/2023 22:31 Effects of cricothyroid and thyroarytenoid interaction on voice control: Muscle activity, vocal fold biomechanics, flow, and aco…
Although both continuum mechanics and lumped-element types using graded muscle stimulation on a canine model showed
of computer models were developed for understanding FSI that,13 across all CT activation levels, TA activation always first
3,22–30
during vocal fold vibrations, most of them did not incorporate increased f 0 and then decreased it. Computer models were also
active muscle mechanics. A few studies31,32 related the stiffness used to relate muscle activity and f 0. Results from a three-mass
parameters in the lumped-element vocal fold models to muscle model32 suggested that at low CT activation levels, TA activation
activity, but the model cannot describe the changes of vocal fold first raised and then slightly lowered f 0; at high CT level, TA
posture due to an abstract representation of the vocal fold activation always decreased f 0. A continuum model of vocal fold
structure. Farley33 constructed a simplified mathematical larynx was used to study the effect of CT and TA activation on the
model to study the effects of laryngeal muscle activations on the eigenfrequencies of vocal fold vibration through pre-stressed
glottal width and fundamental frequency estimated by the eigenanalysis by Yin and Zhang.38 It was found that TA always
frequency equation of a stretched string. Similarly, a number of lowered the eigenfrequencies under high CT activation; under
computer models of laryngeal muscle mechanics were mild CT activation, TA first lowered and then increased the
developed for studying vocal fold posturing, but vocal fold eigenfrequencies; the eigenfrequencies increased monotonically
vibrations and sound production were not included.23,34–36 with TA activation only when CT was not activated.
The role of CT and TA in laryngeal control of posturing and voice Despite previous studies, the role of CT and TA in controlling f 0
production was studied in animal, human, and computational is not fully understood. While it is known that TA activation could
models. These studies found that TA activation causes vocal fold either raise or lower f0 depending on the state of the vocal folds, it
shortening, adduction, and vertical height enhancement, while is still not clear under what conditions TA activity would raise or
CT activation has the opposite effect by elongating, mildly lower f0 and the mechanisms behind it. It has been hypothesized
abducting, and decreasing the vertical height of the vocal that the complex control of TA on f0 might be associated with the
fold.20,37,38 In terms of the stiffness and stress, TA activation complex effect of TA on the vocal fold stiffness as TA activity
decreases them in the cover layer due to the vocal fold reduces the passive stress in the cover and body layers but
shortening, but increases them in the body layer due to active increases the active stress in the body layer. However, the
muscle stress; CT activation increases them in both the cover stiffness and stress were never measured in in vivo studies;
and body layers due to vocal fold elongation.38,39 therefore, the relationship cannot be understood. Moreover, while
CT activation was observed to always increase f0, its
CT and TA are known as the two primary controllers of the
effectiveness appeared different at different TA levels.13 The
fundamental frequency of a voice. CT activation was observed to
interaction strategy of CT and TA in effective control voice
always increase f 0 except during register change when it could
outcome, such as f0 and sound pressure level (SPL), remains
lower f 0.13 TA activation, however, has been known to have
unknown. To inform such strategy and, more importantly,
complex effects on f 0. Using EMG measures, Titze et al.7 found
understand the associated underlying mechanism, it is necessary
that TA activation raised f 0 at lower fundamental frequencies and to establish relationships between muscle activity, vocal fold
lower vocal intensities, and reduced f 0 at higher fundamental biomechanics, vocal fold vibration, and voice acoustics.
frequencies and lower vocal intensities. This suggested that TA
activation would lower f 0 under high CT activation, and increase In this study, we employed a 3D high-fidelity continuum model of
it under low CT activation. However, a more recent in vivo study canine larynx which incorporated both active muscle mechanics
and FSIs to investigate the major links between CT/TA activity,
vocal fold posture, vocal fold vibration, and voice outcomes for active stress, was implemented and validated in our previous
during voice production. In a recent work,36 we developed a 3D work.40 A brief description of the model is provided below.
muscle mechanics based laryngeal model which generated the
In the fiber-reinforced material model, the passive strain energy is
geometries based on magnetic resonance imaging (MRI) of a
defined as
canine larynx and incorporated all the major laryngeal features
including the conus elasticus (CE), the paraglottic space (PGS) U=C10(Ĩ1−3)+1D1(J−1)2+k12k2ek2(Ĩ4−1)2−1, (1)
of adipose tissue, and all the relevant cartilages and muscles.
where C10 is the isotropic neo-Hookean parameter, Ĩ1 and Ĩ4 are
The model was used to simulate vocal fold posturing under
the first and fourth invariants of the reduced Cauchy-Green
activation of various combinations of laryngeal muscles. The
tensor, respectively, J is the Jacobian determinant of the
simulation results showed extensive agreement with in vivo
deformation, k1 is a modulus-like parameter, k2 is a
canine laryngeal experiments. In this work, we fully coupled the
dimensionless parameter that accounts for the degree of
laryngeal model with a FSAI computer model to simulate muscle
nonlinearity, and D1 is the compressibility factor, which was set to
control of vocal fold vibration and voice production through vocal
a value close to zero [1/20C10] to approximate the
fold posturing. While the model was developed for general
incompressibility of vocal fold tissues.
simulations of any combinations of ILMs, this work focused on the
role of CT and TA. In the 1D Hill-based model,41 the active stress is defined as
For the rest of the paper, the numerical algorithms of the σA=a(t)σmaxσTL*σTV*, (2)
laryngeal muscle mechanics model and FSAI model are briefly where σA is the active stress, at is the activation level, σmax is
described in Sec. II A. The simulation setup and procedure of the the maximum active muscle stress, and σTL* and σTV* are the
coupled vocal fold posturing and FSAI was described in Sec. II B. normalized scaling factors as functions of muscle stretch and
The parametric simulations with systematic variations of CT and stretch rate, respectively. σTL* is calculated using a Gaussian
TA activations are also described in Sec. II B. The simulation function42 as
results including the geometric and mechanical properties of
σTL*=exp−λ−λoptsf2+mλ, (3)
vocal fold pre-phonatory postures, glottal flow waveform and
voice acoustics are presented in Sec. III. A discussion about the where λ is the stretch ratio of the muscle, defined as the current
relationship between these features is also presented in Sec. III. length over the initial length and thus deformation dependent.
Finally, the conclusion is provided in Sec. IV. λopt is the optimal stretch ratio, at which maximum stress is
achieved. Furthermore, the shape and asymmetry of the relation
II. METHODS are controlled by the shape factor sf and the slope parameter m.
λopt, sf, and m were set to 1.5, 0.35, and 0.01, respectively.40,41
A. Numerical models
The value of σTV* was set to 1 as a constant.36
1. Laryngeal muscle and non-muscle material models

The Mooney-Rivlin model,43 which is a hyperelastic nonlinear
A laryngeal muscle mechanics model, which combines a fiber- material model, was used for the non-muscle materials, including
reinforced model for passive stress and a 1D Hill-based model the cover, CE, and PGS tissues. In this model, the strain energy
is defined as
U=C10I¯1−3+C01I¯2−3+1D1(J−1)2, (4) ρ∂2 d →∂t2=σ¯¯. +ρf →, ∇ (8)
where C01 is the material constant and I¯2 is the second where ρ is the vocal fold density, d → is the displacement vector,
invariant of the reduced Cauchy-Green tensor. When C01=0, the σ¯¯ is the stress tensor, and f → is the body force. The equation
model is reduced to Neo-Hookean. For the dynamic simulation, a was solved using the finite element method. The details of the
Rayleigh damping model40 was used. solver are referred to Liu et al.45
The relationship between the stress tensor and the deformation 3. Glottal aerodynamics model
for this strain energy function is given by taking the derivative of
The glottal aerodynamics was modeled by the one-dimensional
the strain energy function with respect to the right Cauchy-Green
(1D) Bernoulli equation. The flow separation was assumed to
tensor as
occur at the minimum glottal area. Downstream the flow
SKLPassive=2C10∂I¯1∂CKL+2D11−1I3∂I3∂CKL+2k1I¯4−1ek (5) separation, the flow pressure was assumed to be zero gage
where K and L are the subscript for indicial notation, SKLPassive pressure. Upstream the flow separation, the flow pressure was
is the second Piola-Kirchhoff stress tensor, and I3 is the third calculated using the Bernoulli equation
invariant of the right Cauchy-Green deformation tensor CKL. The Py=Psub−12ρairQA(y)2, (9)
readers are referred to Dhondt44 for detailed derivation of the
where Py is the intraglottal pressure at the vertical location of y,
stress-strain relationship. For an incompressible specimen under
A(y) is the cross-sectional area, Psub is the subglottal pressure,
uniaxial stretch along the fiber direction, an analytical stress-
Q is the flow rate, and ρair is the air density. The flow rate was
stretch relation can be obtained following some straightforward
calculated using
derivation as
Q=2PsubρairAmin, (10)
σ=2C10λ2−1λ+2k1λ2−1λ2ek2λ2−1, (6)
where Amin is the minimum glottal area. To measure this
where σ is the true tensile stress and λ is the stretch. It shows the
parameter, the area between the vocal fold medial surface and
highly nonlinear nature of the fiber reinforced material.
contact plane was calculated at each discretized section in the
In the finite element calculation, the uniaxial active stress inferior-superior direction. The minimum value multiplied by two
computed using Eq. (6) was transformed into a stress tensor corresponds to the minimum glottal area, where the flow
based on the muscle orientation (defined by the origin and separation is assumed to occur. If the glottis is completely closed,
insertion of the muscle) and added to the passive stress as the Amin becomes zero.
total stress that satisfies the equilibrium. Specifically,
4. Vocal tract acoustics model
STotal=SPassive+SActive=[SPassive]+qTσA00000000q, (7)
The sound pressure propagation in the vocal tract was modeled
where [q] is the coordinate transformation matrix.
by a 1D wave-reflection equation developed by Story,46
2. Vocal fold dynamics model f1=a1b1+QρaircA1, (11)
The above material models were implemented into a Navier where f1 and b1 are, respectively, the pressures of the forward-
equation solver to simulate vocal fold dynamics. The Navier traveling and backward-traveling sound waves at the vocal tract
equation is inlet, a1 is the attenuation factor, Q is the glottal flow rate
obtained from the glottal aerodynamics model, c is the speed of the volume and insertion, were realistic by following literature
sound, and A1 is the cross section area of the vocal tract inlet. data.47,48 The material parameters of the muscle and non-muscle
Calculation of f1 is explicit using the value of b1 from the tissues are provided in Tables I and II. Note that the material
previous time step in the simulation. parameters of the muscles were the same as in the previous work
and they were determined by curve fitting experimental data of
5. Flow-structure-acoustics coupling canine tissues.40,49 Specifically, the curve fitting for the passive
The coupling between the aerodynamics model and vocal fold parameters of the CT and TA muscles was described in a
dynamics model was implemented on the vocal fold surfaces previous work40 using experimental measurements.50,51 Note that
where they were in contact with the glottal airflow. At each time the same parameters were used for CT and TA because they
step, the glottal aerodynamic pressures were first calculated with have almost identical passive response.49 Passive parameters
the medial surface shape of the vocal folds obtained from for the other muscles [lateral cricoarytenoid (LCA), interarytenoid
previous time step, and then fed back to the vocal fold dynamics (IA), and posterior cricoarytenoid (PCA)] were obtained in a
solver to calculate the new deformation of the vocal folds. The similar way based on experimental data.49 The maximum active
simulation then marched one step by using the updated vocal stresses of the canine laryngeal muscles were measured
fold shape. The coupling between the aerodynamics model and extensively by Alipour-Haghighi et al.;50–52 CT50,52 and TA51 both
acoustics model was one way that the flow rate obtained from the have a maximum active stress around 100 kPa. The other
aerodynamics model was used to calculate the vocal tract muscles (IA, LCA, and PCA) seem to have lower maximum
acoustic pressures, while the acoustic pressures were not fed active stress,53 but the authors mentioned that it could be due to
back to the aerodynamics model. the samples being not as fresh. Previous models23,35 have also
assumed for these muscles the same maximum active stress as
B. Simulation setup
that of TA. However, the maximum active stress of the IA, LCA,
Figure 1(a) shows the geometric model which was reconstructed and PCA muscles does not affect the current simulations since
from a high-resolution MRI of a dissected canine larynx. The they were not active. Two parameter Mooney-Rivlin models were
reconstruction method was described in detail in our previous used for the non-muscle tissues, and the material parameters
work.36 Briefly, the model contained major anatomic components were determined based on previous vocal fold models.54,55
of a larynx, including the lamina propria (cover); the two TA
muscle bundles (thyrovocalis and thyromuscularis); the conus
elasticus; the lateral cricoarytenoid; the paraglottic space of
adipose tissue; the thyroid, cricoid, and arytenoid (paired)
cartilages; the two bundles of the cricothyroid muscle (rectus and
oblique); the interarytenoid muscle; and the posterior
cricoarytenoid muscle. Note that while most tissues are
segmented directly from the MRI scans, simplified geometries
were used for the CT and interarytenoid (IA) muscles because
their segmentations were very difficult due to a lack of clear
boundaries in the images; but their anatomic properties, such as
FIG. 1. IA 2.035 10 2.5 200
TABLE II.
Material parameters for non-muscle tissues.
Tissue C10 [kPa] C01 [kPa]
Cover 0.117 0.295
PGS (adipose) 0.117 0.295
CE 2.000 1.500
Inter-muscular tissue 0.117 0.295
To accurately predict vocal fold posturing with CT/TA activation,

modeling the motion of the cricoarytenoid joint (CAJ) and
cricothyroid joint (CTJ) is critical. Following the analysis of the
joint mechanism56,57 and previous modeling practice,23,58 the CAJ
(Color online) Simulation setup and procedure. (a) The MRI-
was modeled as a rocking-sliding motion, where the arytenoid
reconstructed canine laryngeal model which contains major anatomic
cartilage can rotate around and slide along the long axis of the
components including muscles, cartilages, and vocal folds. (b) The
laryngeal muscle mechanics model was used to simulate static joint facet on the cricoid. The CTJ was treated similarly that the
deformations of the vocal folds upon muscle activations. (c) The vocal thyroid cartilage was allowed to rotate around and move
fold deformation obtained from posturing simulation, along with TA horizontally along the joint. Our previous work showed that this
activation, were applied on the resting vocal fold configuration to implementation was able to simulate wide ranges of vocal fold
generate the pre-phonatory shape, tension, and stiffness of the vocal
postures and the results were in good agreement with in vivo
folds. The vocal folds were then coupled with the glottal aerodynamics
experiments in terms of the trends and ranges of the vocal fold
model and vocal tract acoustics model to simulate vocal fold vibrations
and sound generation. strain and vocal process distance under various combinations of
muscle activation.11,13,38,59
TABLE I.
Figure 1 illustrates the simulation procedure of vocal fold
Material parameters for muscle tissues.
posturing and FSAI under a certain CT/TA activation level. The
Muscle C 10 [kPa] k 1 [kPa] k2 σmax [kPa]
simulation consisted of two steps. In the first step [Fig. 1(b)], the
TAv 3 0.15 9 100 laryngeal muscle mechanics model was used to simulate static
TAm 3 0.15 9 80
deformation of the vocal folds under the CT/TA activation. Note
that since the canine larynx has been artificially adducted using a
CTo & CTv 3 0.15 9 100
suture at the vocal processes when it was scanned, the vocal
LCA 2.7 2 5.5 100 folds in the current model are already at the adducted
PCA (all bellies) 2.7 2 5.5 100

configuration; therefore, activation of LCA and IA was not

considered. In the second step [Fig. 1(c)], the vocal fold The cases in a parametric set are referred to by the activation
deformation obtained in the first step, along with TA activation, levels (aCT, aTA), where both aCT and aTA are in the range of [0, 1].
were applied on the resting vocal fold configuration to generate For example, case (0.5,0.4) is the case in which the activation of
the pre-phonatory shape, tension, and stiffness of the vocal folds. CT is 50% and TA 40%. Cases (*, 0.4 ∼ 0.6) include all the cases
The vocal folds were then coupled with the glottal aerodynamics with the TA level of 0.4 to 0.6 and CT levels from 0 to 1.
model and vocal tract acoustics model to simulate vocal fold
vibrations and sound generation. The geometry of the vocal tract III. RESULTS
was generated by following Refs. 46 and 60 using the cross
A. Effects of CT/TA activation on vocal fold medial
section areas of a /ah/ vowel model superimposed on an MRI- surface shape
based vocal tract centerline profile. Note that, to save
Figure 2(a) shows the 3D shape of the left vocal fold at the
computational time, we assumed left-right symmetry of vocal fold
resting position. The part of the medial surface where effective
vibration in the current study so that only the left vocal fold was
vibration occurred was shaded in green. Figure 3 shows the
used in the FSAI simulations. A contact plane was placed at the
shape change of this part of the medial surface under CT and TA
glottis midline to mimic vocal fold contact.
activations. Specifically, each colored subfigure shows the
In this study, the laryngeal model was discretized into 123 333 projection of this part of the medial surface on the Y-Z plane at
tetrahedral elements with each vocal fold having about 30 000 different aCT and aTA combinations. The color represents the
elements. A mesh-independence investigation was performed in contour of the lateral displacement (in the X direction). Note that
the previous study36 and the same mesh was used in the current only the selected cases with aCT and aTA equal to 0, 0.2, 0.5, 0.8,
study. The flow domain was discretized into 100 sections inside and 1 are plotted. Other cases showed similar trends and are not
the glottis. The vocal fold vibration was simulated for 0.2 s with a plotted. Each column corresponds to a fixed TA activation, and
time step of 2.27 × 10−5 s (which corresponds to a sampling rate of each row corresponds to a fixed CT activation. The sub-figures in
44.1 kHz). the last right column show the X-Y profiles of the medial surface
at a Z-cross section (indicated by the black lines in the Y-Z
The activation of the CT/TA was systematically varied from 0 (no
projection plots) at different TA activations and a fixed CT
activation) to 1.0 (full activation) with an increment of 0.1. It
activation. The subfigures in the top row depicts the same profiles
resulted in a total of 121 simulation cases. Previous studies
at different CT activations and a fixed TA activation.
showed that the phonation onset pressure was higher when the
13
CT/TA activation was higher, therefore, the minimum subglottal
pressure required for sustained vibrations varies by muscle
activation levels. To ensure sustained vibrations under most
activation conditions, we used a relatively high subglottal
pressure of 2.0 kPa. Among the 121 simulation cases, 112 of
them (92%) showed sustained periodic vibrations. The 9
unsuccessful cases were all at the highest CT and TA
activations.
FIG. 2. FIG. 3.
(Color online) (a) 3D shape of the vocal fold in the resting position. The
green shade indicates the part of the medial surface that effectively
vibrated in FSAI. The two cylinders indicate the location and size of
numerical indentation. (b) The profile of the mid-coronal cross section of
the vocal fold. The profiles of the internal muscles are also shown. w and
h measure the width and vertical height of the vocal fold, respectively.
The two rectangle bars indicate the location and size of numerical
indentation in the plane.
(Color online) Variation of the medial-surface shape of the vocal fold at
the selected aCT and aTA equal to 0, 0.2, 0.5, 0.8, and 1. Each colored
subfigure shows the projection of the medial surface on the Y-Z plane at
a certain CT/TA combination. The color represents the contour of the
lateral displacement (in the X direction). The sub-figures in the last right
column show the X-Y profiles of the medial surface at a Z-cross section
(indicated by the black lines in the Y-Z projection plots) at different TA
activations with a fixed CT activation. The subfigures in the top row
depicts the same profiles at different CT activations with a fixed TA
activation.
It can be seen that CT and TA activations had opposite effects on

the anterior-posterior (AP) length (Z direction) and vertical height
(Y direction) of the vocal fold: CT increased the AP length and
decreased the vertical height, while TA decreased the AP length
and increased the vertical height. These effects are consistent
with in vivo observations and other computer model
predictions.20,38 The figure also reflects that the change of the
vertical height primarily occurred at the superior edge through the
rising/falling of the edge. The movement at the inferior edge is

due to its connection to the cricoid cartilage. This observation is Figure 4 shows the contour of each parameter in the full CT/TA
also consistent with in vivo observation.61 Furthermore, the line activation space. Figure 4(a) shows that the vocal fold width was
plots in the revealed that the effectiveness of both CT and TA in predominantly affected by TA activation, confirming the previous
changing the medial surface shape reduced with activation level: observation in Fig. 3. Quantitatively, increasing the TA activation
they were most effective when the activation level was low. from aTA = 0 to 1 increased vocal fold width by about 14% at all
CT levels. Increasing the CT activation had little effect on the
Figure 3 also shows that TA activation significantly increased the
width; the largest effect is seen at zero activation of TA that
bulging of the vocal fold at all CT levels, while CT activation
increasing CT from aCT = 0 to 1 decreased the width by about 6%.
slightly decreased the bulging only when TA activation level was
Figure 4(b) shows that TA shortened the vocal fold in the
very low [cases (*,0 ∼ 0.2)]; at medium to high TA activations, CT
longitudinal direction while and CT elongated it. It also shows the
showed nearly no effect on the vocal fold bulging. These effects
contour line of stretch of 1 being slightly below the diagonal,
20
are consistent with in vivo observations. In addition, the contour
indicating that CT was slightly more effective than TA in
plots reflect that the increase in the vocal fold bulging induced by
regulating the stretch. This trend agrees with what was reported
TA activation was non-uniform on the vocal fold surface. In
in numerical38 and in vivo59 studies. The range of the stretch in
general, the increase was more at the posterior than at the
our simulations was from 0.9 to 1.1, smaller than one previous in
anterior and more at the inferior than at the superior; the increase
vivo canine measurement.59 Nevertheless, the current range is
was most prominent at an inferior-posterior region. One possible
still physiologically reasonable since vocal fold stretch in the
reason for more bulging near the posterior might be that the TAv
ranges between 0.92 to 1.14 has also been reported in in vivo
muscle is thicker near the posterior. As for more bulging near the
canine models.11 Figure 4(c) shows the antagonistic effect of CT
inferior, it might be associated with the connection of the inferior
and TA on changing the vocal fold height: CT decreased the
vocal fold to the cricoid cartilage which restricts the tissue from
height and TA increased the height. The contour line of zero (no
moving in the vertical direction but not in the lateral direction.
height change) was very below the diagonal line, indicating that
A group of parameters are defined to quantify the changes of the CT was more effective than TA in adjusting the vertical height.
vocal fold shape, including vocal fold width (w), stretch (λ), height The change of the vertical height ranged from −9% to 5% in our
(h), and contact area (AC). Figure 2(b) illustrates the definition of simulations. Figure 4(d) shows that TA monotonically increased
the vocal fold width and height. The width is defined as the the contact area at all CT levels, and CT increased the contract
horizontal distance between the most inferior and lateral points area only when aTA >= 0.2. When aTA < 0.2, CT showed no effect
on the mid-coronal plane of the vocal fold, and it is used to on the contact area. It is further shown that the increase in the
quantify the bulging of the vocal fold. The height is defined as the contact area with TA activation was due to the increased vocal
vertical distance between the most inferior and superior points on fold bulging and height, and the increase in the contact area with
the mid-coronal plane of the vocal fold. The stretch is defined as CT activation was due to the elongation of the vocal fold. At low
the ratio of the vocal fold length to its neutral length in the TA activations, CT slightly abducted vocal folds which reduced
longitudinal direction. Note that the vocal fold width, height, and its effect on the contact area.
stretch were measured in the absence of the contact plate. The
contact area is defined as the area of the medial surface of the
vocal fold in contact with the contact plate.
FIG. 4. Negative and positive stresses mean compression and tension,

respectively.
FIG. 5.
(Color online) Contours of geometric parameters of the vocal fold

posturing: variation of (a) vocal fold width (w), (b) stretch (λ), (c) height
(h), and (d) contact area (AC). Vertical axis: CT activation level.
Horizontal axis: TA activation level. w shows the horizontal distance
between the most inferior and most lateral points on the mid-coronal
plane of the vocal fold. h shows the vertical distance between the most
inferior and most superior points on the mid-coronal plane of the vocal
fold. λ shows the ratio of the vocal fold length to its neutral length in the
longitudinal direction. AC shows the area of the medial surface of the
vocal fold in contact with the contact plate. w and h are presented in
percentage of variation to the neutral position.
B. Effects of CT/TA activation on vocal fold stress

and stiffness
CT and TA regulate the fundamental frequency and vocal register

by adjusting the stress and stiffness of the vocal fold tissues. In
general, CT always stiffens the tissues by elongating them, while
TA has a complex effect: it, on one hand, slackens the tissues by
shortening them, and on the other hand, stiffens the muscle (Color online) Contour of the volume-averaged AP stress in the (a) cover
layer, (b) TA, (c) TAv, and (d) TAm. Vertical axis: CT activation level.
tissue, and its overall effect on the vocal fold stiffness is the
Horizontal axis: TA activation level. Contour levels in each layer are
combination of the two mechanisms. While these effects are
different due to the very different stress ranges. Negative and positive
widely known, it remains unknown how they interact to determine stresses mean compression and tension, respectively.
the overall stress and stiffness of the vocal fold. The computer
model in the current study separated the TAm, TAv and cover It is observed that the AP stress in the TAv, TAm, and TA was one
layer of the vocal fold, allowing quantification of the stress in to two orders of magnitude higher than that in the cover layer.
each individual layer and their combined effect on the stiffness of Because the passive material properties of the cover and TA
the vocal fold. tissues are very close, the large difference in the stresses
between them indicates that the active muscle stress outweighed
Figure 5 shows the contour of the volume-averaged AP stress in the passive stress. It was also noticed that the TA tissues had
the TAm, TAv, and cover layer as well as the entire TA by tensile stresses even when the vocal fold was shortened (for
combining the TAm and TAv. Please note that the contour levels example, at aCT = 0), which was also because the tensile active
in each layer are different due to the very different stress ranges. stress outweighed the compressive passive stress. The results
thus suggested that the vocal fold would always has a large

stress ratio between the muscle and cover tissues when TA was displacement occurs at the center of the loading area. The local
activated and the ratio would increase with TA activation; the only stiffness is estimated as the ratio of the total indentation force to
exception would be at zero TA activation when there is no active the maximum displacement. Note that superior and inferior
stress in the muscle tissues. loading were applied separately.
Due to the dominance of the active stress, the AP stress in the TA Figure 6 shows that the stiffness of the vocal fold at the superior
tissues was predominantly regulated by the TA activation; and inferior, as well as their mean (the average of the inferior and
however, the effectiveness of TA was affected by CT. Specifically, superior stiffness) and vertical gradient (the difference of inferior
TA was more effective in increasing the AP stress when the CT and superior stiffness divided by the vertical distance between
activation was higher. For example, increasing TA activation from the two indentation locations). It was found that the stiffness at the
zero to full increased the AP stress in TA from 0 to 7.9 kPa at zero inferior was significantly higher than that at the superior. A
CT activation, and from 1.2 to 11.4 kPa at full CT activation. From previous computational study62 revealed that the inferior-superior
another perspective, increasing CT activity helped achieving geometry asymmetry of the vocal fold could cause a higher
higher stress at the same TA activation. However, the effect of CT stiffness at the inferior than the superior. As a result, the mean
was pronounced only when the TA activation was medium to and vertical gradient of the stiffness were predominantly
high (0.3 ∼ 1) and CT activation was low to medium (0 ∼ 0.5). At determined by the inferior stiffness. It was also noticed that the
low TA activation (0 ∼ 0.2) and medium to high CT activation (0.4 variation pattern of the mean stiffness and stiffness gradient was
∼ 1), CT had little effect on the muscle stress [cases (0 ∼ 0.2, 0.4 well correlated with that of the AP stress in the TA, indicating that
∼ 1)]. The maximum stress in the TA tissues was achieved when the stiffness was primarily affected by the muscle stress.
both CT and TA at full activations. The results suggested that Therefore, both CT and TA activations increased the mean and
while TA was more effective in adjusting the muscle stress, CT vertical gradient of the stiffness, and more importantly, one
also played an important role in achieving higher stress by muscle was more effective when the other muscle activation was
increasing the effectiveness of TA regulation. In the cover layer, higher. For example, at zero CT activation, increasing TA
however, both CT and TA effectively controlled the stress. Due to activation from zero to full increased the mean stiffness from
the lack of active stress, the change of the stress was purely a 0.085 to 0.12 N/cm and the stiffness gradient from 0 to 0.12
response to the stretch variation. The maximum tension was N/cm2. At full CT activation, increasing TA activation from zero to
achieved when the CT activation was full and TA activation was full increases the mean stiffness from 0.086 to 0.17 N/cm and the
zero. stiffness gradient from 0.0 to 0.3 N/cm2. The effectiveness of CT
on regulating the stiffness was also the highest when TA was at
The tissue stresses affect vocal fold vibration by affecting the
full activation. As a result, the maximum stiffness and stiffness
transverse stiffness of the vocal fold. To examine this effect, a
gradient were achieved when both CT and TA were at full
numerical indentation method was used to measure the
activations. The results indicated that the most effective way for
transverse stiffness of the vocal fold at the mid-coronal position of
adjusting the stiffness of the vocal fold was to simultaneously
the medial surface. The details of the method were reported in a
62
adjust the CT and TA muscles.
previous study. Briefly, a uniform pressure of 2 kPa was applied
on a circular area with a radius of 1 mm on the vocal fold medial
surface at the superior and inferior positions (shown in Fig. 2).
The pressure caused local displacements and the maximum
FIG. 6. phases and computing the averaged flow rate at the same phase
over all the steady cycles. Figure 7(b) shows the phase-averaged
flow rate of the case (0.5, 0.4). T is the time period of the vibration
and the fundamental frequency was computed as f 0 = 1/T. To and
Tc are the time duration when the glottis was open and closed,
respectively. Tp and Tn are the time duration when the glottis was
opening and closing, respectively. A few quantities were
computed to characterize the waveform of the flow rate, including
the mean flow rate (Q¯), speed quotient (Tp/Tn), and closure
quotient (Tc/T).
FIG. 7.
(a) The time history of the volumetric flow rate of a representative case
(0.5, 0.4) of sustained vibration, and (b) the phase-averaged flow rate of
case (0.5, 0.4). A few quantities were defined to characterize the
waveform of the flow rate, including T being the time period of the
vibration, T o and T c being the time duration of an open and closed
glottis, respectively, T p and T n being the time duration of glottis opening
and closing, respectively.
(Color online) Contour of the vocal fold stiffness at the (a) inferior and (b) The contours of the glottal flow parameters under the CT and TA
superior as well as their (c) mean, and (d) vertical gradient. Mean
activations are plotted in Fig. 8. Note that the cases that did not
contour shows the average of the inferior and superior stiffness, and the
converge to a sustained vibration were left blank.
gradient contour shows the difference of inferior and superior stiffness
divided by the vertical distance between the two indentation locations.
Vertical axis: CT activation level. Horizontal axis: TA activation level.
C. Effects of CT/TA activation on glottic dynamics
Figure 7(a) shows the glottal flow dynamics of a representative

case (0.5, 0.4) with sustained vibrations. The vibration went
through initial irregular cycles characterized by varied
frequencies and amplitudes and gradually transitioned to steady
cycles characterized by a steady frequency and amplitude. To
quantify the waveform of the glottal flow rate, the phase-averaged
flow rate was computed by dividing each steady cycle into 200
FIG. 8. with a much lower rate. As a result, high frequencies were

observed at regions where TA activation was medium and CT
activation was high. The highest frequency occurred at case (1,
0.6). Our results show that CT monotonically increased f 0, and
TA first increased and then decreased f 0, as well as that high
(Color online) Contours of glottal flow parameters: (a) fundamental frequencies occurred at medium TA activation and high CT
frequency (f 0), (b) speed quotient, (c) closure quotient, (d) flow rate Q¯, activation agreed with an in vivo study in canine larynx.13
and (e) voice intensity (I). f 0 is one over the time period of the vibration,
speed quotient is the time duration of glottis opening over the closing, The discontinued change of f 0 at certain TA and CT activations
closure quotient is the time duration of closed glottis over the time period suggested that a bifurcation occurred in the nonlinear dynamic
of the vibration, and voice intensity is calculated as the root mean square system of vocal fold vibration. Empirical eigenfunction analysis of
(RMS) value of the sound pressure wave. Vertical axis: CT activation
the current simulation results showed that the vocal fold vibration
level. Horizontal axis: TA activation level.
can be largely accounted for by two basic modes representing a
1. Fundamental frequency lateral motion and a rotational motion. The bifurcation was
associated with the change of the mode order. Specifically,
f 0 in our simulations ranged from 145 Hz to 190 Hz, with the
before the frequency drop, the mode representing the lateral
lowest at the zero activation of both CT and TA, and the highest
motion was the most dominant with more than 60% of the total
at (1, 0.6). This frequency range is lower than what was reported
energy and the mode representing the rotational motion had
from in vivo canine experiments by Chhetri et al.11,13 One of the
about 25% of the total energy. The energy percentage of the two
possible reasons could be that the simulations were using a fixed modes flipped after the frequency drop and the mode
subglottal pressure while the experiments varied the subglottal representing the rotational motion became the most dominant.
pressures for obtaining sustained vibrations. Due to the page limit as well as that the focus of the current paper
is parameter correlations, dedicated analysis of modal
Our results showed that f 0 monotonically increased with CT
entrainment is discussed in a separate paper.
activation, but with a discontinuity at a certain activation level:
when CT was activated from zero to full at a fixed TA activity, f 0
2. Speed quotient
first continuously increased and then abruptly jumped to a high
value and then continued to increase but with a lower increasing Figure 8(b) shows the contour of the speed quotient in the full CT
rate. The lower increasing rate of f 0 at high CT activation was and TA activation space. The speed quotient characterizes the
likely due to that CT became less effective in increasing the relative duration of glottis opening and closing: a speed quotient
muscle stress (Fig. 5) and stiffness (Fig. 6) at high activation. The of 1 indicated the same duration of glottis opening and closing,
effect of TA on f 0 shows a more complex pattern. When CT and a speed quotient lower than 1 indicated shorter opening than
activation was low (aCT < 0.2), TA activation monotonically closing and vice versa. Our simulations showed that in most
increased f 0, but to a very slight extent. When aCT > 0.2, f 0 first cases, except a few cases at low TA activations, the speed
quickly increased with the TA activation and then reached its quotient was greater than one, indicating that the glottis had
maximum when the TA activation was slightly less than the shorter closing than opening. Interestingly, the pattern of speed
corresponding CT activation; further increasing the TA activation quotient also showed discontinuities at the same muscle
caused an abrupt drop of f 0, after which f 0 increased again but combinations where the discontinuity of f 0 occurred, suggesting
again that the frequency discontinuity was associated with our results, the effects of these factors were secondary. The fact
vibration changes. Generally, the speed quotient was lower in the that the flow rate was stable across a wide range of muscle
region of low TA and high CT, and higher in the region of high TA activations helps maintain speech effort in a wide range of
and low CT. A distinct region with the highest speed quotient was frequency.
found very low CT activations, which also corresponded to the
lowest f 0. In this region, the vibration modes of the vocal fold 5. Voice intensity
were found to be different from other regions. Figure 8(e) shows the contour of voice intensity in the CT and TA
activation space. The voice intensity was computed as the root
3. Closure quotient
mean square (RMS) value of the sound pressure wave. Across
Figure 8(c) shows the contour of the closure quotient in the full the full range of CT and TA activations, the change of voice
CT and TA activation space. In general, CT and TA activations intensity was small. This agrees with other studies which found
increased the closure quotient, except in the low CT activation that voice intensity is primarily controlled by subglottal pressure,
region (aCT < 0.3) where the closure quotient was distinctly high. which was not changed in our simulations.32,64 Even though the
Figure 4(d) shows that both CT and TA activations increased the changes were small, TA showed a consistent effect on
contact area, and Fig. 8(d) shows that CT and TA activations decreasing the voice intensity while CT showed a complex
generally increased closure quotient, implying a positive pattern. It was further noticed that the voice intensity after the
correspondence between closure quotient and contact area. In sudden frequency drop (in the TA activation direction) was lower
other words, a larger contact area requires more effort to push the than other cases, suggesting that the vibration change tended to
glottis to open, thus leading to an increased closure quotient. decrease the voice intensity.
However, in the low CT activation region, the very high closure
D. Multiple-parameter correlation between vocal fold
quotient was due to a distinct laryngeal dynamics.
posture and frequency
4. Flow rate To understand how muscle controls the frequency through

posturing, we studied the multiple-parameter correlation (r)
Figure 8(d) shows the contour of the mean flow rate in the CT
among posture parameters and frequency using the following
and TA activation space. It was seen that the flow rate dropped
equation:
quickly when TA activation increased from zero to 20%; after that,
TA activation continued to decrease the flow rate, but to a small r2=[rx1y ⋯ rxNy]rx1x1⋯rx1xN⋮⋱⋮rxNx1⋯rxNxNrx1y⋮rxNy, (12)
extent. CT activation showed little effect on the flow rate. This where rxiy is the correlation factor between each posture
pattern was consistent with the effect of CT and TA on the parameter and frequency, rxixj is the correlation between two
bulging of the vocal fold [Fig. 4(a)] that TA activation increased posturing parameters, and N is the number of posture parameters
the bulging and CT had little effect, and moreover TA was most participating in multiple correlation. The correlation between
effective at low activations. This consistency suggested that the each two variables (α and β) can be calculated using the
vocal fold bulging played a primary role in controlling the flow following equation:
rate, which agrees with a past computational study.63 Note that
rαβ=∑i=1n(αi−α¯)(βi−β¯)∑i=1n(αi−α¯)2∑j=1n(βj−β¯)20.5, (13)
past research showed other factors affecting the flow rate,
including vocal stress/stiffness and vertical stiffness gradient. In

where n is the number of data in variables. It should be noted that FIG. 9.

rαα=1 and rαβ=rβα.
The correlations between the posture parameters (w, h, λ, AC, σ,

and F/dx) and f 0 were studied and the combination of stress,
stretch, and stiffness was found to have the strongest correlation
with f 0 with the factor of r = 0.89. Figure 9(a) shows a three-
dimensional view of the stress, stretch, and stiffness of each case
colored by f 0. The pink plate corresponds to the stretch of 1,
meaning the vocal fold length was the natural length. Figures
9(b) and 9(c) show the two-dimensional view of the relationship
by using the stretch as the horizontal axis, and the stress [Fig.
9(b)] and stiffness [Fig. 9(c)] as the vertical axes. The black dash
line represents the cases of the stretch of 1. Left and right to the
dash line correspond to vocal fold shortening and elongation,
respectively. The arrows indicate the paths of increasing CT and (Color online) (a) Stress, stretch, and stiffness of studied cases, colored
TA activations. Figures 9(b) and 9(c) reveal that CT activation by the fundamental frequency. The pink plate shows the location of the
increased stretch, stress, and stiffness, although at different rates, no-stretch condition. Studied cases in (b) stretch-stress and (c) stretch-
resulting in monotonic increase in the frequency. TA activation, stiffness planes, colored by the fundamental frequency. The dashed line
shows the location of the no-stretch condition. The left and right sides of
however, increased the stress and stiffness but decreased the
the dashed lines are the shortening and elongation regions. Directions of
stretch, suggesting a complex mechanism in controlling f 0. Very
CT and TA activations are shown by arrows.
interestingly, if the vocal fold shortening and elongation zones
were studied separately, we found that in each zone f 0 Figure 10 shows the range of the voice intensity and frequency
monotonically increased with TA activation, indicating that f 0 was with the contour of the CT and TA activations. It should be noted
primarily affected by the stress and stiffness. However, f 0 abruptly that because the subglottal pressure was kept constant in our
dropped when the vocal fold stretch changed from elongation to simulations, the voice intensity change was caused only by
shortening by TA activation. Note that vocal fold experiences muscle activations. The figures reflect that the voice intensity had
compression in shortening and tension in elongation, and the the widest range at medium frequencies. Toward low and high
discontinuity of f 0 suggested a bifurcation of glottic dynamics. Our frequencies, the range of voice intensity decreased quickly,
results, thus, suggested that vocal fold in compression or tension meaning that its adjustability was low only by muscle activations.
would cause different glottic dynamics. Specifically, in our results, The figure also shows that CT activation had large variations at
when the vocal fold changed from in compression to tension, the high frequencies and TA activation had large variations at low
glottic dynamics changed from a condition with lower frequencies frequencies. It suggests that the voice intensity was effectively
to another with higher frequencies. adjusted by CT activations at high frequencies and by TA
activations at low frequencies. Furthermore, the results indicate
that to achieve high voice intensity at high frequencies, CT and
TA need to be at high and low activation levels, respectively.
FIG. 10. ratio between the muscle and cover tissues when TA was
activated and the ratio would increase with TA activation; the only
exception would be at zero TA activation when there is no active
stress in the muscle tissues. Due to the fact that the active stress
outweighed the passive stress, the vocal fold stiffness was
primarily affected by the muscle stress. Therefore, TA was more
effective in controlling the stiffness, yet its effectiveness was
increased by increasing CT activation. The results indicated that
the most effective way for adjusting the stiffness of the vocal fold
(Color online) Studied cases in frequency-intensity plane colored by
was to simultaneously adjust the CT and TA muscles.
muscle activation levels: (a) CT and (b) TA.
The simulations on FSAI showed that both CT and TA activation
IV. DISCUSSION had small effects on the flow rate and voice intensity, suggesting
that CT and TA are not primary controllers of these two features.
The simulations on the vocal fold posturing showed that TA
Past studies indicated that the voice intensity and flow rate were
activation decreased the AP length and increased the bulging,
primarily controlled by respiratory parameters, including the
height, and contact area of the vocal fold, and CT activation
subglottal pressure and vocal fold abduction/adduction.3,63 Our
increased the AP length and contact area and decreased the
simulations have kept the subglottal pressure constant and the
height of the vocal fold. CT activation slightly decreased the
results showed a positive correlation between the variation of
bulging of the vocal fold when TA activation was low and had
flow rate and vocal fold bulging, therefore agreeing with past
nearly no effect on the bulging when TA activation was medium
studies. CT and TA activations were found to both increase the
to high. These observations were found to generally agree with
closure quotient of the vocal fold vibration due to their effects on
past in vivo studies on canine larynges.20,61 Our results also
increasing the pre-phonatory contact area which increased the
showed that TA was predominant in controlling the vocal fold
required aerodynamic effort to open the glottis. A computational
bulging, while the CT was more effective in controlling the vocal
study on the effect of the vertical thickness of vocal fold medial
fold height and stretch.
surface on voice production (Wu and Zhang65) found that closure
Both CT and TA activations increased the vocal fold stress, quotient was significantly affected by the vertical thickness with a
however, the mechanisms were different. CT increased the stress positive correspondence. Our results, however, showed that CT
by generating passive stress associated with stretching, and TA activation reduced vocal fold vertical height [Fig. 4(c)] but
increased the stress by generating active stresses. It was found increased closure quotient. The results may seem inconsistent
that the active stress was one to two orders of magnitudes higher with Wu and Zhang65 at first glance. However, a closer
than the passive stress, therefore the stress in the TA tissues was comparison between their results and ours revealed that the
predominantly controlled by the TA activation. However, the increase in closure quotient in the two studies were both due to
effectiveness of TA was affected by CT: TA was more effective in the increase in the contact area. In Wu and Zhang,65 a three-
increasing the stress when CT activation was higher. Due to the dimensional vocal fold model was employed and so an increase
large difference between the active and passive stresses, our in the vertical thickness of the medial surface implied an increase
results suggested that the vocal fold would have a large stress in the contact area. The positive correspondence of the closure

quotient and vertical thickness thus implied a positive The current study has some limitations. Because of the limited
correspondence between the closure quotient and contact area. range of the vocal fold stretch, the range of f 0 was limited. The
dependence of vocal fold posturing and glottal flow results to the
An analysis on the relationship among voice intensity, f 0, and
different muscle activation levels could be different in the
muscle activations showed that the adjustability of the voice
population. To the best of the authors' knowledge, although some
intensity was highest at medium frequencies. At high
studies investigated the effects of age, gender, height, etc., on the
frequencies, the voice intensity was effectively adjusted by CT
vocal fold properties such as vocal fold length66 and fundamental
activations, and at low frequencies, it was effectively adjusted by
frequency,67 the muscle activation has not been considered and
TA activations. To achieve high voice intensity at high
the corresponding data is not available. Moreover, although
frequencies, CT and TA needs to be at high and low activation
ligaments play an important role in vocal fold dynamics,68
levels, respectively.
ligaments have not been included in the current model due to the
CT and TA showed a complex control on f 0. CT activation lack of information in MR images. Furthermore, LCA was not
monotonically increased f 0, but with an abrupt jump at a medium activated since the model has been constructed based on MRI
CT activation level. TA activation, however, showed a more scans of an artificially adducted canine larynx. Therefore, the
complex effect. When CT activation was low (<0.2), TA activation effect of TA activation on the artificially closed vocal fold could be
monotonically increased f 0, but to a very slight extent. When CT different from the one that starts from an open glottis adducted by
activation was medium to high, increasing TA activation first the LCA activation. Another limitation of the current study is the
quickly increased f 0, and then caused an abrupt drop of f 0, and at compressional stress in the fiber direction. Experimental studies
last increased f 0 again but to a much smaller extent. As a result, reported that the muscle showed anisotropic behavior under
high frequencies were observed at regions where TA activation compression with the stiffness in the cross-fiber direction being
was medium and CT activation was high. A similar complex higher than that in the fiber direction. In the current study, this
pattern was also observed in an in vivo canine study.13 anisotropic behavior is not modeled by the Neo-Hookean
Discontinuities were also observed in speed quotient at the same component, possibly resulting in an overprediction of the
muscle activations. These discontinuities suggested dynamics compressional stress in the fiber direction. However, since the
changes in vocal fold vibrations. Further analysis revealed that active stress was always one to two orders larger than the
the frequency pattern was highly correlated with a combination of passive stress in our results, even when the passive stress might
the stress, stiffness, and stretch of the vocal fold. When the vocal have been overpredicted under compressions, it is expected that
fold was either in compression or tension, increasing the CT and the effect to be very small. The CT:TA ratio of the maximum active
TA activations increased the stress and stiffness and so stress is another important parameter that is considered equal to
increased f 0. Abrupt frequency drop occurred when the vocal fold 1 in this study. Further studies are required to investigate the
changed from being in tension to compression due to TA effects of changing the maximum stress ratio on the posturing
activation. Our results, thus, suggested that while f 0 was primarily and glottal flow parameters. Moreover, subglottal pressure, which
adjusted by the stress and stiffness of the vocal fold, the condition is an essential factor in vocal fold vibration dynamics, was fixed
if the vocal fold is in compression or tension was also important in this study. Another interesting subject for further study is to
in determining the f 0 range by causing different glottic dynamics. disentangle the origins of the vocal folds stiffening during the
muscle activation and evaluate their effectiveness.
V. CONCLUSION References
A 3D high-fidelity continuum mechanics laryngeal model that 1. I. R. Titze, Principles of Voice Production (Second
incorporated both active muscle mechanics and FSAIs was used Printing) (National Center for Voice and Speech, Iowa City, IA,
2000), pp. 87–127.
to investigate the effects of CT and TA activations on vocal fold
Google Scholar
postures, vocal fold vibrations, and voice outcomes. The
laryngeal model included major anatomic features including the 2. R. Mittal, B. D. Erath, and M. W. Plesniak, “Fluid dynamics
of human phonation and speech,” Annu. Rev. Fluid Mech. 45,
conus elasticus, the paraglottic space, and all the relevant
437–467 (2013).
cartilages and muscles, and the geometry of each component
https://doi.org/10.1146/annurev-fluid-011212-140636
was reconstructed from MRI scans of a canine larynx. Parametric Google Scholar Crossref
simulations were conducted by systematically varying the
3. Z. Zhang, “Mechanics of human voice production and
activation level of the CT and TA muscles from 0 to 1. The
control,” J. Acoust. Soc. Am. 140, 2614–2635 (2016).
obtained vocal fold pre-phonatory postures, vocal fold vibrations, https://doi.org/10.1121/1.4964509
and voice outcomes at each muscle condition were quantified. Google Scholar Crossref PubMed
The effect of CT and TA activation on these quantities was
4. J. J. Ohala, “Aspects of the control and production of
studied and the associated mechanisms were discussed. Pre- speech,” UCLA Working Papers in Phonetics No. 15 (1970).
phonatory postures results revealed that both TA and CT Google Scholar
activations increased the vocal fold contact area, stress, and
5. J. E. Atkinson, “Correlation analysis of the physiological
stiffness. While TA activation decreased the length and increased factors controlling fundamental voice frequency,” J. Acoust.
the height of the vocal fold, CT activation showed opposite Soc. Am. 63, 211–222 (1978).
effects by increasing and decreasing the vocal fold length and https://doi.org/10.1121/1.381716
height, respectively. Moreover, TA activation increased the Google Scholar Crossref PubMed
bulging. FSAI results revealed that both CT and TA activations 6. G. B. Kempster, C. R. Larson, and M. K. Kistler, “Effects of
increased the closure quotient and had small effects on the flow electrical stimulation of cricothyroid and thyroarytenoid
rate and voice intensity. Furthermore, CT and TA showed a muscles on voice fundamental frequency,” J. Voice 2, 221–
229 (1988). https://doi.org/10.1016/S0892-1997(88)80080-8
complex pattern on the fundamental frequency behavior. It was
Google Scholar Crossref
shown that the frequency pattern was highly correlated with a
combination of the stress, stiffness, and stretch of the vocal fold. 7. I. R. Titze, E. S. Luschei, and M. Hirano, “Role of the
thyroarytenoid muscle in regulation of fundamental
frequency,” J. Voice 3, 213–224 (1989).
ACKNOWLEDGMENTS https://doi.org/10.1016/S0892-1997(89)80003-7
This study is supported by the National Science Foundation Google Scholar Crossref
under Grant 245 No. 1652632. The numerical simulation was 8. H.-S. Choi, G. S. Berke, M. Ye, and J. Kreiman, “Function
supported by the Extreme Science and Engineering Discovery of the thyroarytenoid muscle in a canine laryngeal model,”
Environment (XSEDE) under Allocation Award No. TG- Ann. Otol. Rhinol. Laryngol. 102, 769–776 (1993).
https://doi.org/10.1177/000348949310201006
CTS180004. The open source finite element package CalculiX
Google Scholar Crossref PubMed
(Ref. 69) was used for the simulations in this study.

9. H.-S. Choi, G. S. Berke, M. Ye, and J. Kreiman, “Function graded neuromuscular stimulation,” J. Acoust. Soc. Am. 131,
of the posterior cricoarytenoid muscle in phonation: In vivo 3347 (2012). https://doi.org/10.1121/1.4708538
laryngeal model,” Otolaryngol. Neck Surg. 109, 1043–1051 Google Scholar Crossref
(1993). https://doi.org/10.1177/019459989310900612
17. D. Berry, D. K. Chhetri, and J. Neubauer, “The role of the
posterior cricoarytenoid muscle in phonation,” J. Acoust. Soc.
10. D. K. Chhetri, J. Neubauer, and D. A. Berry, “Graded Am. 134, 4203 (2013). https://doi.org/10.1121/1.4831428
activation of the intrinsic laryngeal muscles for vocal fold Google Scholar Crossref
posturing,” J. Acoust. Soc. Am. 127, EL127–EL133 (2010).
18. A. Mendelsohn, D. K. Chhetri, J. Neubauer, D. A. Berry,
https://doi.org/10.1121/1.3310274
and G. S. Berke, “Acoustic effects of asymmetric laryngeal
activation,” Otolaryngol. Neck Surg. 147, 72 (2012).
11. D. K. Chhetri, J. Neubauer, and D. A. Berry, https://doi.org/10.1177/0194599812451438a111
“Neuromuscular control of fundamental frequency and glottal Google Scholar Crossref
posture at phonation onset,” J. Acoust. Soc. Am. 131, 1401–
19. G. Luegmair, D. Chhetri, and Z. Zhang, “Effects of
1412 (2012). https://doi.org/10.1121/1.3672686
thyroarytenoid muscle activation on phonation in an in vivo
canine larynx model,” J. Acoust. Soc. Am. 136, 2294 (2014).
12. D. K. Chhetri, J. Neubauer, J. L. Bergeron, E. Sofer, K. https://doi.org/10.1121/1.4900296
A. Peng, and N. Jamal, “Effects of asymmetric superior Google Scholar Crossref
laryngeal nerve stimulation on glottic posture, acoustics,
20. A. M. Vahabzadeh-Hagh, Z. Zhang, and D. K. Chhetri,
vibration,” Laryngoscope 123, 3110–3116 (2013).
“Three-dimensional posture changes of the vocal fold from
https://doi.org/10.1002/lary.24209
paired intrinsic laryngeal muscles,” Laryngoscope 127, 656–
664 (2017). https://doi.org/10.1002/lary.26145
13. D. K. Chhetri, J. Neubauer, E. Sofer, and D. A. Berry, Google Scholar Crossref PubMed
“Influence and interactions of laryngeal adductors and
21. P. Bhattacharya and T. Siegmund, “A computational
cricothyroid muscles on fundamental frequency and glottal
study of systemic hydration in vocal fold collision,” Comput.
posture control,” J. Acoust. Soc. Am. 135, 2052–2064 (2014).
Methods Biomech. Biomed. Engin. 17, 1835–1852 (2014).
https://doi.org/10.1121/1.4865918
https://doi.org/10.1080/10255842.2013.772591
14. D. K. Chhetri, J. Neubauer, and E. Sofer, “Influence of
22. F. Alipour, D. A. Berry, and I. R. Titze, “A finite-element
asymmetric recurrent laryngeal nerve stimulation on vibration,
model of vocal-fold vibration,” J. Acoust. Soc. Am. 108, 3003–
acoustics, and aerodynamics,” Laryngoscope 124, 2544–
3012 (2000). https://doi.org/10.1121/1.1324678
2550 (2014). https://doi.org/10.1002/lary.24774
23. E. J. Hunter, I. R. Titze, and F. Alipour, “A three-
15. D. K. Chhetri and J. Neubauer, “Differential roles for the
dimensional model of vocal fold abduction/adduction,” J.
thyroarytenoid and lateral cricoarytenoid muscles in
Acoust. Soc. Am. 115, 1747–1759 (2004).
phonation,” Laryngoscope 125, 2772–2777 (2015).
https://doi.org/10.1121/1.1652033
https://doi.org/10.1002/lary.25480
24. R. Mittal, H. Dong, M. Bozkurttas, F. M. Najjar, A. Vargas,
16. D. A. Berry, D. K. Chhetri, and J. Neubauer,
and A. von Loebbecke, “A versatile sharp interface immersed
“Observations of phonation threshold pressure and
boundary method for incompressible flows with complex
fundamental frequency in the in vivo canine larynx using
boundaries,” J. Comput. Phys. 227, 4825–4852 (2008). 32. S. Y. Lowell and B. H. Story, “Simulated effects of
https://doi.org/10.1016/j.jcp.2008.01.028 cricothyroid and thyroarytenoid muscle activation on adult-
Google Scholar Crossref PubMed male vocal fold vibration,” J. Acoust. Soc. Am. 120, 386–397
(2006). https://doi.org/10.1121/1.2204442
25. X. Zheng, Q. Xue, R. Mittal, and S. Beilamowicz, “A
coupled sharp-interface immersed boundary-finite-element
method for flow-structure interaction with application to human 33. G. R. Farley, “A biomechanical laryngeal model of voice
phonation,” J. Biomech. Eng. Asme 132, 111003 (2010). F0 and glottal width control,” J. Acoust. Soc. Am. 100, 3794–
https://doi.org/10.1115/1.4002587 3812 (1996). https://doi.org/10.1121/1.417218
Google Scholar Crossref Google Scholar Crossref PubMed
26. D. J. Daily and S. L. Thomson, “Acoustically-coupled 34. J. Yin and Z. Zhang, “Interaction between the
flow-induced vibration of a computational vocal fold model,” thyroarytenoid and lateral cricoarytenoid muscles in the
Comput. Struct. 116, 50–58 (2013). control of vocal fold adduction and eigenfrequencies,” J.
https://doi.org/10.1016/j.compstruc.2012.10.022 Biomech. Eng. ASME 136, 111006 (2014).
Google Scholar Crossref PubMed https://doi.org/10.1115/1.4028428
27. Q. Xue, X. Zheng, R. Mittal, and S. Bielamowicz,
“Subject-specific computational modeling of human 35. J. Yin and Z. Zhang, “Laryngeal muscular control of vocal
phonation,” J. Acoust. Soc. Am. 135, 1445–1456 (2014). fold posturing: Numerical modeling and experimental
https://doi.org/10.1121/1.4864479 validation,” J. Acoust. Soc. Am. 140, EL280–EL284 (2016).
Google Scholar Crossref PubMed https://doi.org/10.1121/1.4962375
28. S. Chang, C. K. Novaleski, T. Kojima, M. Mizuta, H. Luo,
and B. Rousseau, “Subject-specific computational modeling 36. B. Geng, N. Pham, Q. Xue, and X. Zheng, “A three-
of evoked rabbit phonation,” J. Biomech. Eng. ASME 138, dimensional vocal fold posturing model based on muscle
0110051–0110056 (2016). https://doi.org/10.1115/1.4032057 mechanics and magnetic resonance imaging of a canine
Google Scholar Crossref larynx,” J. Acoust. Soc. Am. 147, 2597–2608 (2020).
https://doi.org/10.1121/10.0001093
29. Z. Zhang, “Cause-effect relationship between vocal fold
physiology and voice production in a three-dimensional
phonation model,” J. Acoust. Soc. Am. 139, 1493–1507 37. M. Hirano and Y. Kakita, " “Cover-body theory of vocal
(2016). https://doi.org/10.1121/1.4944754 fold vibration,” in Speech Science, edited by R. Daniloff
Google Scholar Crossref PubMed (College-Hill Press, San Diego, CA, 1985), pp. 1–46.
Google Scholar
30. W. Jiang, X. Zheng, and Q. Xue, “Computational
modeling of fluid–structure–acoustics interaction during voice 38. J. Yin and Z. Zhang, “The influence of thyroarytenoid and
production,” Front. Bioeng. Biotechnol. 5, 7 (2017). cricothyroid muscle activation on vocal fold stiffness and
https://doi.org/10.3389/fbioe.2017.00007 eigenfrequencies,” J. Acoust. Soc. Am. 133, 2972–2983
Google Scholar Crossref PubMed (2013). https://doi.org/10.1121/1.4799809
31. I. R. Titze and B. H. Story, “Rules for controlling low-
dimensional vocal fold models with muscle activation,” J. 39. S. Deguchi, Y. Kawahara, and S. Takahashi,
Acoust. Soc. Am. 112, 1064–1076 (2002). “Cooperative regulation of vocal fold morphology and stress
https://doi.org/10.1121/1.1496080 by the cricothyroid and thyroarytenoid muscles,” J. Voice 25,
Google Scholar Crossref PubMed e255–e263 (2011).
https://doi.org/10.1016/j.jvoice.2010.11.006
Google Scholar Crossref PubMed Rhinol. Laryngol. 109, 505–513 (2000).

https://doi.org/10.1177/000348940010900512
40. N. Pham, Q. Xue, and X. Zheng, “Coupling between a
fiber-reinforced model and a Hill-based contractile model for
passive and active tissue properties of laryngeal muscles: A 48. K. A. Cox, F. Alipour, and I. R. Titze, “Geometric structure
finite element study,” J. Acoust. Soc. Am. 144, EL248–EL253 of the human and canine cricothyroid and thyroarytenoid
(2018). https://doi.org/10.1121/1.5055564 muscles for biomechanical applications,” Ann. Otol. Rhinol.
Google Scholar Crossref PubMed Laryngol. 108, 1151–1158 (1999).
https://doi.org/10.1177/000348949910801210
41. S. L. Smith and E. J. Hunter, “A viscoelastic laryngeal
muscle model with active components,” J. Acoust. Soc. Am.
135, 2041–2051 (2014). https://doi.org/10.1121/1.4866173 49. E. J. Hunter and I. R. Titze, “Refinements in modeling the
Google Scholar Crossref PubMed passive properties of laryngeal soft tissue,” J. Appl. Physiol.
103, 206–219 (2007).
42. J. M. Winters and L. Stark, “Analysis of fundamental
https://doi.org/10.1152/japplphysiol.00892.2006
human movement patterns through the use of in-depth
antagonistic muscle models,” IEEE Trans. Biomed. Eng.
BME-32, 826–839 (1985). 50. F. Alipour-Haghighi, A. L. Perlman, and I. R. Titze,
https://doi.org/10.1109/TBME.1985.325498 “Tetanic response of the cricothyroid muscle,” Ann. Otol.
Google Scholar Crossref Rhinol. Laryngol. 100, 626–631 (1991).
https://doi.org/10.1177/000348949110000805
43. N. Kumar and V. V. Rao, “Hyperelastic Mooney-Rivlin
model: Determination and physical interpretation of material
constants,” Parameters 2, 1 (2016). 51. F. Alipour-Haghighi, I. R. Titze, and A. L. Perlman,
Google Scholar “Tetanic contraction in vocal fold muscle,” J. Speech Lang.
Hear. Res. 32, 226–231 (1989).
44. G. Dhondt, The Finite Element Method for Three-
https://doi.org/10.1044/jshr.3202.226
Dimensional Thermomechanical Applications (Wiley,
Chichester, UK, 2004), pp. 219–223.
Google Scholar Crossref 52. F. Alipour and I. Titze, “Active and passive characteristics
of the canine cricothyroid muscles,” J. Voice 13, 1–10 (1999).
45. G. Liu, B. Geng, X. Zheng, Q. Xue, H. Dong, and G. V.
https://doi.org/10.1016/S0892-1997(99)80056-3
Lauder, “An image-guided computational approach to
inversely determine in vivo material properties and model
flow-structure interactions of fish fins,” J. Comput. Phys. 392, 53. F. Alipour, I. R. Titze, E. Hunter, and N. Tayama, “Active
578–593 (2019). https://doi.org/10.1016/j.jcp.2019.04.062 and passive properties of canine abduction/adduction
Google Scholar Crossref laryngeal muscles,” J. Voice 19, 350–359 (2005).
46. B. H. Story, “Physiologically-based speech simulation
using an enhanced wave-reflection model of the vocal tract,
Doctoral dissertation,” University of Iowa, Iowa City, IA, 1996, 54. C. L. Jones, A. Achuthan, and B. D. Erath, “Modal
pp. 32–41. response of a computational vocal fold model with a substrate
Google Scholar layer of adipose tissue,” J. Acoust. Soc. Am. 137, EL158–
EL164 (2015). https://doi.org/10.1121/1.4905892
47. C. W. Mineck, R. Chan, N. Tayama, and I. R. Titze,
“Three-dimensional anatomic characterization of the canine
laryngeal abductor and adductor musculature,” Ann. Otol.
55. D. D. Cook, E. Nauman, and L. Mongeau, “Ranking 63. F. Alipour and R. C. Scherer, “Vocal fold bulging effects
vocal fold model parameters by their influence on modal on phonation using a biophysical computer model,” J. Voice
frequencies,” J. Acoust. Soc. Am. 126, 2002–2010 (2009). 14, 470–483 (2000).
https://doi.org/10.1121/1.3183592 https://doi.org/10.1016/S0892-1997(00)80004-1
Google Scholar Crossref PubMed Google Scholar Crossref PubMed
56. G. M. Adran and F. H. Kemp, “The mechanism of the 64. S. Tanaka and M. Tanabe, “Glottal adjustment for
larynx. I. The movements of the arytenoid and cricoid regulating vocal intensity an experimental study,” Acta
cartilage,” Br. J. Radiol. 39, 641–654 (1966). Otolaryngol. 102, 315–324 (1986).
https://doi.org/10.1259/0007-1285-39-465-641 https://doi.org/10.3109/00016488609108682
Google Scholar Crossref PubMed Google Scholar Crossref PubMed
57. H. von LEDEN and P. Moore, “The mechanics of the 65. L. Wu and Z. Zhang, “Voice production in a MRI-based
cricoarytenoid joint,” Arch. Otolaryngol. 73, 541–550 (1961). subject-specific vocal fold model with parametrically
https://doi.org/10.1001/archotol.1961.00740020553008 controlled medial surface shape,” J. Acoust. Soc. Am. 146,
Google Scholar Crossref PubMed 4190–4198 (2019). https://doi.org/10.1121/1.5134784
58. W. S. Selbie, L. Zhang, W. S. Levine, and C. L. Ludlow,
“Using joint geometry to determine the motion of the 66. J. A. X. Filho, E. C. M. de Melo, C. de Giacomo Carneiro,
cricoarytenoid joint,” J. Acoust. Soc. Am. 103, 1115–1127 D. H. Tsuji, and L. U. Sennes, “Length of the human vocal
(1998). https://doi.org/10.1121/1.421223 folds: Proposal of mathematical equations as a function of
Google Scholar Crossref PubMed gender and body height,” Ann. Otol. Rhinol. Laryngol. 114,
390–392 (2005).
59. I. R. Titze, J. J. Jiang, and E. Lin, “The dynamics of length
https://doi.org/10.1177/000348940511400510
change in canine vocal folds,” J. Voice 11, 267–276 (1997).
https://doi.org/10.1016/S0892-1997(97)80004-5
Google Scholar Crossref PubMed 67. K. Zhang, T. Siegmund, and R. W. Chan, “A constitutive
model of the human vocal fold cover for fundamental
60. B. H. Story, “A parametric model of the vocal tract area
frequency regulation,” J. Acoust. Soc. Am. 119, 1050–1062
function for vowel and consonant simulation,” J. Acoust. Soc.
(2006). https://doi.org/10.1121/1.2159433
Am. 117, 3231–3254 (2005).
https://doi.org/10.1121/1.1869752
Google Scholar Crossref PubMed 68. J. E. Kelleher, T. Siegmund, and R. W. Chan, “Collagen
microstructure in the vocal ligament: Initial results on the
61. A. M. Vahabzadeh-Hagh, Z. Zhang, and D. K. Chhetri,
potential effects of smoking,” Laryngoscope 124, E361–E367
“Quantitative evaluation of the in vivo vocal fold medial
(2014). https://doi.org/10.1002/lary.24626
surface shape,” J. Voice 31, 513.e15–513.e23 (2017).
Google Scholar Crossref PubMed 69. G. Dhondt and K. Wittig, “A free software three-
dimensional structural finite element program,” www.calculix.
62. B. Geng, Q. Xue, and X. Zheng, “A finite element study
de (1998) (Last viewed May 18, 2021).
on the cause of vocal fold vertical stiffness variation,” J.
Google Scholar
Acoust. Soc. Am. 141, EL351–EL356 (2017).
https://doi.org/10.1121/1.4978363 © 2021 Acoustical Society of America.

Effects of Cricothyroid and Thyroarytenoid Interaction On Voice Control: Muscle Activity, Vocal Fold Biomechanics, Flow, and Acoustics

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Effects of Cricothyroid and Thyroarytenoid Interaction On Voice Control: Muscle Activity, Vocal Fold Biomechanics, Flow, and Acoustics

Uploaded by

Copyright:

Available Formats

15/10/2023 22:31 Effects of cricothyroid and thyroarytenoid interaction on voice control: Muscle activity, vocal fold biomechanics, flow,

The human voice originates from complex neuromuscular

1. Laryngeal muscle and non-muscle material models

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 5/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 6/44

U=C10I¯1−3+C01I¯2−3+1D1(J−1)2, (4) ρ∂2 d →∂t2=σ¯¯. +ρf →, ∇ (8)

2. Vocal fold dynamics model f1=a1b1+QρaircA1, (11)

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 9/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 10/44

FIG. 1. IA 2.035 10 2.5 200

Tissue C10 [kPa] C01 [kPa]

Cover 0.117 0.295

PGS (adipose) 0.117 0.295

Inter-muscular tissue 0.117 0.295

To accurately predict vocal fold posturing with CT/TA activation,

PCA (all bellies) 2.7 2 5.5 100

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 11/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 12/44

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 13/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 14/44

It can be seen that CT and TA activations had opposite effects on

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 15/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 16/44

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 17/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 18/44

FIG. 4. Negative and positive stresses mean compression and tension,

(Color online) Contours of geometric parameters of the vocal fold

B. Effects of CT/TA activation on vocal fold stress

CT and TA regulate the fundamental frequency and vocal register

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 19/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 20/44

C. Effects of CT/TA activation on glottic dynamics

Figure 7(a) shows the glottal flow dynamics of a representative

FIG. 8. with a much lower rate. As a result, high frequencies were

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 25/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 26/44

4. Flow rate To understand how muscle controls the frequency through

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 27/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 28/44

where n is the number of data in variables. It should be noted that FIG. 9.

The correlations between the posture parameters (w, h, λ, AC, σ,

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 29/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 30/44

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 31/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 32/44

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 33/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 34/44

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 35/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 36/44

Google Scholar Crossref PubMed Rhinol. Laryngol. 109, 505–513 (2000).

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 41/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 42/44

https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 43/44 https://pubs.aip.org/asa/jasa/article/150/1/29/606514/Effects-of-cricothyroid-and-thyroarytenoid 44/44

You might also like