Respiratory Physiology & Neurobiology 236 (2017) 69–77

Contents lists available at ScienceDirect

Respiratory Physiology & Neurobiology

journal homepage: www.elsevier.com/locate/resphysiol

Respiratory dynamics in phonation and breathing—A real-time MRI

study
Louisa Traser a,b,c,∗ , Ali Caglar Özen d , Fabian Burk a,c , Michael Burdumy d , Michael Bock d ,
Bernhard Richter a,c , Matthias Echternach a,c
a
Institute of Musicians’ Medicine, Medical Center – University of Freiburg, Germany
b
Department of Otolaryngology, Medical Center – University of Freiburg, Germany
c
Faculty of Medicine, University of Freiburg, Germany
d
Dept. of Radiology, Medical Physics, Medical Center – University of Freiburg, Germany

a r t i c l e i n f o a b s t r a c t

Article history: The respiratory system is a central part of voice production, but for phonation neither the underlying
Received 18 September 2016 functional relations between diaphragm (DPH) and rib cage (RC), nor differences to normal breathing
Received in revised form 8 November 2016 are yet understood. This study aims to compare respiratory dynamics in phonation and breathing via
Accepted 13 November 2016
dynamic MRI of the lung. Images of the breathing apparatus of 6 professional singers were captured
Available online 18 November 2016
in a 1.5 T MRI system in supine position during vital capacity breathing and maximal long sustained
phonation at 3 different pitches and loudness conditions. In a dynamic series of cross-sectional images of
Keywords:
the lung, distances between characteristic anatomical landmarks were measured. During exhalation in
Real-time magnetic resonance imaging
Respiration
normal breathing the diaphragm and rib cage moved synchronously to reduce lung volume, but during
Phonation phonation different functional units could be identified, which support phonation by facilitating the
Singer Imaging control of subglottic pressure.
Diaphragm © 2016 Elsevier B.V. All rights reserved.

1. Introduction Leanderson, 1987; Leanderson et al., 1987; Sundberg et al., 1989),

Respiratory dynamics have been the focus of pulmonary
research for decades, but studies have previously been focused on
breathing mechanics. Nevertheless, the respiratory system is also a
central part of voice production: the breathing apparatus regulates
subglottic pressure (psub ) which arises when the vocal folds are
closed for phonation (Sundberg, 1987). Careful adjustment of psub
is essential for voice production as it has a considerable influence on
sound pressure level, fundamental frequency and vocal tract res-
onances (Leanderson et al., 1987; McAllister and Sundberg, 1998;
Sundberg et al., 1993).
Respiratory patters in phonation have so far been mainly evalu-
ated in singers, since professional musicians use a very consistent
breathing strategy (Bouhuys et al., 1966; Leanderson and Sundberg,
1988; Pettersen and Eggebø, 2010; Thomasson and Sundberg, 2001,
1999; Watson et al., 1990; Watson and Hixon, 1985) which is inten-
sively trained during vocal education. In previous studies methods
such as bodyplethysmography (Bouhuys et al., 1966), transdi-
aphragmmatic pressure measurements (Bouhuys et al., 1966;
∗ Corresponding author at: Breisacher Str. 60, 79106 Freiburg, Germany.
E-mail address: Louisa.Traser@uniklinik-freiburg.de (L. Traser).
magnetometry (Collyer et al., 2008) and respitrace (Thomasson,
2003; Thomasson and Sundberg, 2001; Watson et al., 1990) have
been applied, but these techniques either only portray the impact of
the respiratory system on the body’s surface, or they measure solely
cumulative effects of different parts of the respiratory apparatus.
Different approaches have been chosen for imaging the dynamic
system of the lung in respiration, including x-ray, computed tomog-
raphy (CT) and sonography (Hoffman, 1985; Macklin, 1922; Nason
et al., 2012; Pettiaux et al., 1997). Imaging methods using haz-
ardous ionizing radiation are limited in their application for studies
in healthy subjects due to ethical concerns, and ultrasound suffers
from limited penetration depths in the lung. Thus, implementa-
tion of dynamic lung magnetic resonance imaging (MRI) has been
an important step in the study of functional aspects of respiratory
anatomy in physiological and pathological situations (Lee et al.,
2014). Recent improvements of imaging hardware and software
have led to the possibility of dynamic imaging of the diaphragm
(DPH) and the rib cage (RC) (Cluzel et al., 2000; Gierada et al., 1995;
Hatabu et al., 1999; Kolar et al., 2010; Kondo et al., 2000; Takazakura
et al., 2004), regional ventilation (Eichinger et al., 2007b; Molinari
et al., 2008; Simon et al., 2012) and even of 3D reconstruction of
lung movement (Biederer et al., 2009; Plathow et al., 2005b; Tokuda

http://dx.doi.org/10.1016/j.resp.2016.11.007
1569-9048/© 2016 Elsevier B.V. All rights reserved.
70 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77
Fig. 1. Exemplary subject position with supine body orientation on the patient table
of the clinical 1.5 T MRI system. Image data were acquired with the posterior inte-
grated spine array coil and the anterior body matrix coil, which was loosely arranged
at the thorax so as not to interfere with the motion of the rib cage during phona-
tion. A modified electroglottographic system was worn on the neck of the subject
for simultaneous monitoring of the glottal cycle.
et al., 2009) including MRI based spirometry (Eichinger et al., 2007a;
Kolar et al., 2010; Plathow et al., 2005a).
To the best of the author’s knowledge the only imaging study
to assess the dynamics of the respiratory system during phonation
was performed by Pettersen et al. using ultrasound, but only small
parts of the respiratory system were able to be visualized due to the
low penetration depth of ultrasound into lung tissue (Pettersen and
Eggebø, 2010). As a consequence, for breathing during phonation,
neither the underlying functional relationships between different
parts of the breathing apparatus, nor any possible differences in
contrast to normal breathing are yet understood in detail. It is
hypothesized that (1) RC and DPH-motion differ between normal
breathing and phonation and that (2) movement patterns of the
breathing apparatus differ for different pitches and loudness con-
ditions.
This study aims, therefore, to analyse respiratory dynamics
in phonation and breathing using dynamic MRI of the lungs. In
keeping with previous studies (Echternach et al., 2014, 2010), pro-
fessional singers were chosen as subjects, giving the additional
benefit that they are, through education, less vulnerable to distract-
ing noises during the MR imaging, and that they also use a very
consistent breathing strategy (Bouhuys et al., 1966; Leanderson
and Sundberg, 1988; Pettersen and Eggebø, 2010; Thomasson and
Sundberg, 2001, 1999; Watson et al., 1990; Watson and Hixon,
1985).
2. Methods
2.1. Magnetic resonance imaging
In this study the breathing apparatus of professional singers
was dynamically imaged using a clinical 1.5 T MRI system (Tim
Symphony, Siemens, Erlangen, Germany) (Fig. 1). For dynamic
imaging, a 2D trueFISP imaging sequence (TR/TE = 3/1.5 ms, ␣ = 6◦ ,
BW = 977 Hz/px, slice thickness = 10 mm, acquisition matrix = 256,
FOV = 420 mm) was applied with a temporal resolution of approx-
imately 3 frames per second (fps). To achieve this high temporal
resolution a reduced field-of-view of 65% and parallel imaging
acceleration by a factor of 1.3 was used. Images were obtained via
a combination of a posterior spine matrix coil and an anterior 6-
channel body matrix coil (Fig. 1). Dynamic imaging was performed
over a time span of 30–60s, depending on the task. Sagittal images
of the right lung (to avoid motion artefacts from the beating heart,
which would complicate the subsequent image analysis) and coro-
nal images were acquired.
For the sagittal images a slice through the right lung was cho-
sen since the beating heart would cause motion artefacts in the left
lung which would complicate the subsequent analysis of the image.
Initially a 3D localizer data set was acquired in order to define the
image plane. In this data, the sagittal plane was placed in such a
way that the vertex of the diaphragmatic coupola and the apex
of the lung could be identified. The coronal plane was placed simi-
larly encompassing both vertices of the left and right diaphragmatic
coupolae and the apices of the left and right lung. All measurements
for each subject were taken in the supine position during one single
session. The subjects wore headphones for hearing protection and
communication.
2.2. Electroglottography and audio recording
The monitoring of glottal resistance is of fundamental impor-
tance during phonation, since it is closely associated with Psub and
airflow through the glottis (Sundberg, 1987). As glottal resistance
is in turn directly dependent on the glottal area, a simultaneous
electroglottographic (EGG) recording was taken using a modified
MR-safe EGG device (Laryngograph Ltd. London, UK) as described
previously (Özen et al., 2015).
EGG measurements indirectly measure the contact area of the
vocal folds during phonation: a weak alternating current passes
between two electrodes placed on the surface of the neck at the
level of the thyroid cartilage, and changes to tissue impedance
caused by the vibration of the vocal folds are recorded at a millisec-
ond frame rate. From this signal it is possible to calculate the open
quotient (OQ), i.e. the ratio between the time the vocal folds are
not in contact with each other and the vibratory cycle of vocal folds.
This procedure is based on the work of Howard et al. (Howard, 1995;
Howard et al., 1990) using a combination of an EGG based threshold
method for detection of glottal opening (at 3/7) with detection of
glottal closing instants on the dEGG (derivative of EGG) signal. This
method should give more accurate results because of the reliable
detection of glottal closing instants. Additionally the EGG allows
the fundamental frequency (fo ) to be effectively determined.
The EGG signal was then analyzed with regards to OQ and fo in
time intervals of 100 ms over maximum phonation time (MPT) and
the results pooled in 20% time steps. Deviations from the expected
fo were calculated in cents due to its logarithmic scale. The audio
signal was simultaneously recorded using a microphone system
(Pre-polarized Free-field 1/2 Microphone, Type 4189, Brüel&Kjær,
Nærum, Denmark) adapted for use in the MR environment.
2.3. Image analysis
Distances between anatomical landmarks were measured in
each acquired image frame to characterize the motion of the lungs.
In the sagittal Images 4 distances were determined, and in the
coronal Images 2 distances were measured (Fig. 2 and Table 1)
2.4. Spirometry
Spirometry was performed for all subjects (Table 2) separate
to the MRI in order to independently analyze pulmonary function
 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77 71

Fig. 2. Measured distances in the sagittal (A) and coronal (B) plane and their definition according to anatomical landmarks, see also Table 1. DPHant = diaphragm anterior,
DPHhighestP = highest point of diaphragm, DPHpost = diaphragm posterior, a-pD = anterior-posterior diameter, DPHright = diaphragm right, DPHleft = diaphragm left.

Table 1
Measured distances and sagittal and coronal plane and their definitions according to anatomical landmarks. DPH = diaphragm, RC = rib cage. DPHant = diaphragm anterior,
DPHhighestP = highest point of diaphragm, DPHpost = diaphragm posterior, a-pD = anterior-posterior diameter, DPHright = diaphragm right, DPHleft = diaphragm left.

Sagittal plane

DPHant Craniocaudal lung height from the angle of the anterior DPH and the RC to the apex of the lung
DPHhighestP Craniocaudal lung height from highest point of DPH to the apex of the lung
DPHpost Craniocaudal lung height from the angle of the posterior DPH and the RC to the apex of the lung
a-pD the anterior-posterior lung diameter at the height of the 5th rip

Coronal plane

DPHright & DPHleft Craniocaudal lung height from highest point of DPH to the apex of the right and left lung

Table 2
Age, gender, voice classification, classification according to the Bunch and Chapman taxonomy, forced expiratory volume in one second (=FEV1), vital capacity (=VC), height
and weight of subjects.

Subject Age Gender Voice Bunch and Chapman VC in l FEV1 in l/s Height in cm Weight in kg
Classification taxonomy

1 25 female Soprano 3.15b1 3.87 3.46 167 55

2 28 female Soprano 4.5 4.72 3.88 165 65
3 25 female Soprano 3.15b1 3.71 3.28 158 47
4 42 male Tenor 4.5 6.32 5.31 192 93
5 30 male Baritone 3.1a 6.26 5.29 190 100
6 27 male Tenor 7.1 5.28 5.15 172 80

using a spirometer (ZAN, Messgeräte GmbH, Oberthulba, Germany)
according to current clinical guidelines (Sorichter, 2009).
2.5. Subjects and tasks
The Medical Ethics Committee of University of Freiburg,
Germany approved this study (Nr.273/14). Six professional, singers,
trained in the Western classical tradition were included. Table 2
shows their age, gender, voice classification, classification accord-
ing to the Bunch and Chapman taxonomy (Bunch and Chapman,
2000) and subject characteristics.
At the time of the recording, none of the participants complained
of any vocal symptoms or suffered from pulmonary disease.
The phonation tasks were chosen according to the voice classifi-
cation of the singer and represent a low (P1 ), medium (P2 ) and high
pitch (P3 ) in the tessitura of the respective repertoire of the singer.
The singers were asked to phonate a pitch for as long as possible
(maximum phonation time, MPT) in different loudness conditions
(pianissimo = pp, mezzo forte = mf, fortissimo = ff). Additionally the
singers were asked to breathe in and out to the greatest extent in
order to assess their vital capacity. The duration of exhalation was
later referred to as ‘exhaling time’ (=ET). The protocol is given in
Table 3.
The vowel/a/was chosen for phonation to avoid possible artic-
ulatory effects which might be expected when the fundamental
frequency reaches the first vocal tract resonance (Echternach et al.,
2010). Images were acquired for each task both in sagittal and coro-
nal orientation resulting in a total of 10 dynamic imaging series per
subject.
2.6. Statistical analysis and normalisation
For all 6 subjects 5 phonation tasks were recorded in 2 planes.
To compare breathing and phonation measures within (intra)
and between (inter) singers with different durations statistically,
the time axis was re-scaled (tnorm ) starting at the beginning
72 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77

Table 3
Description of the tasks performed by subjects according to different voice classifications including pitch and loudness. mf = mezzo forte, ff = fortissimo, pp = pianissimo.

Female Male

Voice Soprano Tenor Baritone

classification:

Task: vital capacity breathing

pitch loudness pitch loudness pitch loudness

P1mf: A3 mf A2 mf F2 mf
(220 Hz) (110 Hz) (87 Hz)
P2mf: A4 mf A3 mf F3 mf
(440 Hz) (220 Hz) (175 Hz)
P3mf: A5 mf A4 mf F4 mf
(880 Hz) (440 Hz) (349 Hz)
P2pp: A4 (440 Hz) pp A3 pp F3 pp
(220 Hz) (175 Hz)
P2ff: A4 ff A3 ff F3 ff
(440 Hz) (220 Hz) (175 Hz)

of phonation/exhalation (tstart ) and ending with end of phona- Table 4

Mean MPT (maximum phonation time) and OQ (open quotient) with SD (standard
tion/exhalation (tend ). The measured distances A at different
deviation) of all subjects for the different tasks.
locations were normalized (Anorm ) to the distance at Tstart and Tend
according to Task Mean MPT in s SD MPT Mean OQ SD OQ

P1mf 29.8 16.7 0.49 0.039

A (t) − A (tend ) P2mf 25.3 14.7 0.57 0.072
Anorm (t) = · 100% P3mf 14.4 5.7 0.61 0.127
A (tstart ) − A (tend ) P2pp 25.7 11.2 0.56 0.073
P2ff 19.9 9.8 0.52 0.058
In addition the slope (m) of all graphs was then calculated
in steps of 20% in the normalized time (m1 − m5 ) as the ratio of
changes in measured distances over time.

Anorm (tn ) − Anorm (tn−1 )

mn =
tn − t

Next, a repeated measures ANOVA was performed to analyze

differences in slope with regards to all tasks and locations. For fur-
ther evaluation of statistically significant differences in the ANOVA
measurements a univariate t-test was performed.
The intra individual maximal and minimal DPH height and
thorax width was also identified for each subject. DPH and RC devi-
ations in phonation were analyzed in relation to measures during
VC breathing. An additional normalization was consequently per-
formed wherethe amplitude at maximal inspiration was set to 100%
(AVCmax ) and the amplitude at maximal expiration was set to 0%
(AVCmin ). Time was scaled as described above. Differences between
start and end points for different locations and tasks were again
evaluated using an ANOVA and univariate t-test.
OQ and fo derived from the EGG signal were analyzed in steps of
100 ms over MPT and mean values were pooled in steps of 20% (t1-5 ).
A paired student’s t-test was performed to determine whether sta-
tistically significant changes occurred over time.
For all statistical analyses SPSS 23.0 software (SPSS, Inc., Chicago,
IL) was used. The level of significance was set to p < 0.05.
Fig. 3. Boxplots of the ranges of diaphragm (DPH) and rib cage (RC) movement at
different locations during phonation (light grey) and VC breathing (dark grey). Sig-
nificant differences of DPH movement ranges are marked with *. DPHant = diaphragm
3. Results anterior, DPHhighestP = highest point of diaphragm, DPHpost = diaphragm posterior,
a-pD = anterior-posterior diameter, DPHright = diaphragm right, DPHleft = diaphragm
left.
3.1. OC, fo and MPT evaluation

The mean deviation from the requested fo was 27.53 cent
(standard deviation, SD = 30.04, 100 cent represents one halftone).
The longest MPT was reached for P1mf while P3mf was sustained
for a significantly shorter duration compared to the lowest pitch
(p = 0.026); all other phonation times did not differ significantly. No
significant difference between t1-5 was found for OQ and fo . Mean
values of MPT and OQ with SD are presented in Table 4.
3.2. Range of DPH and RC movement in breathing and phonation
The ranges of DPH and RC movement were compared at different
locations and between breathing and phonation and did not differ
significantly (p = 0.38, Fig. 3). During both breathing and phonation
the statistically significant greatest extent of DPH movement was
found in the posterior part of the lung, and a smaller DPH movement
 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77 73

Fig. 4. First row: diaphragm (DPH) and rib cage (RC) movement at different locations during breathing, phonation and delta of phonation and breathing in cm including
standard error, elapsed time (tnorm ) on abscissa. Second row: Normalized DPH and RC movement (Anorm ) at different locations during breathing, phonation and delta of
phonation and breathing including standard error, elapsed time (tnorm ) on abscissa. Individual curves for all subjects are presented as supplemental material.

was seen in the middle section and the anterior DPH (Supplemen-
tal Table S1). No significant differences were found for the ranges
between the left and right DPH (p = 0.3).
Comparing different phonation tasks, no significant difference
for the starting positions (p = 0.99), ending positions (p = 0.99) or
their respective movement ranges (p = 0.94) could be detected.
3.3. Mean curves of normalized data and slope comparison
A normalized pooled data curve of all subjects and tasks includ-
ing SDs is presented in Fig. 4 for phonation and respiration. The
configuration of the curves is different between breathing and
phonation: in phonation the slope of the DPHant and A-pD graphs
increases over time, while for all other parameters it decreases. A
statistically significant difference in slope was found at different
locations in t1,2,4,5 (p < 0.001) but not in t3 (all p-values are pre-
sented in Supplemental Table S2). For different phonatory tasks no
statistically different behaviour was detected (p = 0.3).
In contrast, the respiration graphs show a parallel asymptotical
course (Fig. 4). Differences in slope for t1-5 for the factor ‘location’
revealed no statistical difference (p = 0.63).
The single curves of each subjects are presented in the supple-
mental material (Supplemental Figs. S1 and S2) and show an inter-
and intra-individual very consistent curve progression.
Differences in curve progression can also be observed in Fig. 5,
which compares a kymogram of DPH movement in breathing and
phonation for the anterior, middle and posterior DPH part with a
normalized time axis. A video is also included in the supplemental
material (Supplemental Video S1), which shows the phonation and
exhalation movement parallel to normalized time.
The differential curves of breathing and phonation illustrate that
the difference between both curves is highest in the first 65% of
74 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77
Fig. 5. Kymogram of diaphragm (DPH) movement for subject 6 in breathing (upper row) and phonation (lower row) for the anterior, middle and posterior part with normalized
time axis.
MPT and ET becoming smaller at the end. The curves start with
negative values (breathing values are greater than phonation val-
ues) increasing to positive values (i.e. phonation > breathing) after
20% of the elapsed time. Differences are highest in anterior DPH
and ap-D.
3.4. Evaluation of DPH and RC movement in phonation with
regards to VC
A pooled data curve for normalized data in regard to VC breath-
ing of all subjects and tasks including standard error is presented
in Fig. 6. For the posterior part of the diaphragm negative values
were reached at the very end of phonation. The relative percentage
values in regard to VC breathing at start- (p = 0.23) and endpoint
(p = 0.90) did not differ significantly in different tasks, but at differ-
ent locations (p = 0.002) (Fig. 7). Here, DPHant showed statistically
significant higher values compared to the other parameters. All
p-values are presented in Supplemental Table S3.
3.5. Analysis of agreement of DPHhighestP and DPHright
As DPHhighestP and DPHright represent the same morphometric
distance, and they are recorded directly on after the other (sagittal
and coronal image acquisition), the agreement of both parameters
was evaluated as a measure of consistency. The two-tailed Pearson
correlation analysis of m1-5 as well as the measured distances at
start/endpoint of phonation in DPHhighestP and DPHright showed a
strong significant correlation, p < 0.001, r2 = 0.45 (slope in t1-5 ) and
p < 0.001, r2 = 0.42 (start point) and p < 0.001, r2 = 0.5 (end point).
4. Discussion
This study compares breathing and phonation via dynamic MRI
and subsequent data analysis based on distance changes in the
lung. In general the results show a divergent kinematic behaviour
between both conditions.
 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77
Fig. 6. Diaphragm (DPH) and rib cage (RC) movement at different locations during
phonation given as percentage of maximal movement range in breathing, including
standard errors. Elapsed time on abscissa.
Fig. 7. Boxplots of the percentage start and end points of phonation with regard
to vital capacity (VC) breathing at different locations during phonation. Significant
differences are marked with *.
75
The range of DPH and RC movement was similar in VC breath-
ing and phonation and the degree of DPH excursion decreased
significantly from posterior to middle to anterior during exhala-
tion and phonation. Similar differences in range of DPH movement
at different locations have been found previously (Gierada et al.,
1995; Takazakura et al., 2004), suggesting that, analogous to nor-
mal breathing, during phonation the different parts of the DPH do
not behave uniformly. This motion heterogeneity may be related to
the different attachment parts of the DPH: the anterior portions of
the DPH are fixed to the lower six ribs and the sternum. The line of
attachment of each hemidiaphragm begins at the anterior axillary
line and extends cranially and medially to meet at the xiphoid pro-
cess (Kleinman and Raptopoulos, 1985). Two diaphragmatic crurae
join the diaphragm posteriorly to the upper lumbar vertebral bodies
and disks and they are joined by a fibrous median arcuate liga-
ment (Nason et al., 2012). 43% of the muscular part is located in
the dome of DPH (Tetzlaff and Eichinger, 2009), and hence mus-
cle contraction leads to the reduction, up to the disappearance, of
the appositional zone during maximum inspiration (Cluzel et al.,
2000). It can be speculated that the more cranial attachment of the
anterior diaphragm and the distribution of the muscular parts and
appositional zone lead to the differences in movement range. In
our data the range of movement of the posterior DPH was approx-
imately two times greater than that of the anterior part, in both
singing and breathing, which is in agreement with previous reports
(Gierada et al., 1995; Kondo et al., 2000).
Our findings are however in strong contrast to Pettersen et al.
(Pettersen and Eggebø, 2010) who used ultrasound imaging of
the anterior DPH and a kidney as representative for the posterior
DPH during phonation. For 3 professional singers singing sustained
tones it was found that 2 singers had an ascendance of the anterior
DPH twice as large as that of the posterior DPH. In one singer the
magnitude of DPH ascendance was about the same in both regions.
It could be deduced that these differences are a postural effect
since this study was performed upright. However, an MRI study
by Takazakura revealed subsequently that, in both the upright and
supine position, movement in the posterior DPH part was greater
than that of the anterior DHP (Takazakura et al., 2004). As ultra-
sound is limited to imaging only a small window and indirect
approaches for the posterior part of DPH were used, these con-
flicting results could well be a consequence of technical limitations
of ultrasound as well as of the small sample size.
No statistically significant difference was found in DPH and RC
movement range between different pitch and loudness conditions.
Thus, the pitch had no influence on the extent of DPH ascendance
and reduction of a-p diameter. Only the time in which the move-
ment took place was significantly shorter for the highest compared
to the lowest pitch. This was in contrast to our second hypothesis.
In contrast to the general accordance between the DPH and RC
movement ranges in breathing and phonation, significantly dif-
ferent behaviour could be detected over time: in phonation the
reduction of intrapulmonary volume was predominantly gener-
ated by the elevation of the posterior and middle part of the DPH
at the beginning, while the anterior DPH and the RC remained in a
more inspiratory position. This movement slowed down at approx-
imately 50% of MPT. Simultaneously the elevation of the anterior
part of DPH and the lowering of RC increased its movement velocity.
In contrast, in normal VC breathing all measured parts of the
breathing apparatus contributed synchronously to expiration in
equal shares with higher movement velocity at the start. The differ-
ence between singing and breathing was predominant in the first
65% of the elapsed time.
As the OQ and fo did not change statistically significant over
time, the impact of potential changes of the glottal resistance with
consequent alterations of the breathing apparatus could be ruled
out in the presented data. Furthermore no statistically significant
76 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77
difference could be observed between right and left lung suggest-
ing, that anatomical differences are not the cause.
The different behavior of the respiratory system during phona-
tion and normal breathing could be related to the requirement
to keep the psub constant for phonation, while the effect of recoil
forces changes with lung volume. Changes in psub are unwanted
for a constant phonation as they lead to changes in pitch, loud-
ness and airflow over the glottis. In voice education, singers train
to control psub in spite of the fact, that the direction and intensity
of the recoil forces of the lung changes with its volume (Sundberg,
1987). The continuous regulation of respiratory muscles in chang-
ing lung volumes is what has been referred to as “breath support”
in singing (Proctor, 1980). When at high lung volume recoil forces
exceed the aspired psub , it could be assumed that the RC and the
anterior DPH is stabilized in the inspiratory position to reduce the
psub . In contrast, at the end of a MPT at low lung volume, the expi-
ratory muscles, including the abdominal wall and expiratory RC
muscles, are activated to squeeze out the lung via compression of
the RC and abdominal organs (Leanderson et al., 1987). Hence, for
the phonation, the anterior DPH and RC could be regarded as one
functional unit and the middle and posterior DPH as another. The
alternating stabilization of one part of the respiratory system while
the other is moved could be an advantage in phonation in order to
control psub .
The range of DPH and RC movement in singing was not
exhausted equally in all regions with respect to the individual
maximum and minimum position in VC breathing: Phonation was
stopped on average near 0% of VC, but posterior DPH part was
even more elevated at the end of phonation compared to max-
imal exhaling. This also is in accordance with Gould et al., who
found that singers had a significant higher vital capacity compared
to untrained individuals due to reduction of residual lung volume.
Thus, part of voice training could be the more effective squeezing of
the lung at the end of phonation, which was predominantly done
in the posterior part. Singers started to phonate at 80–90% of VC
without significant difference in task or location. Even if a MPT was
asked, 100% of VC lung extension was not reached, maybe because
the fine control of psub is not optimal at maximal lung expan-
sion. This is in accordance with magnetometer studies (Watson
and Hixon, 1985) and bodypletysmography (Bouhuys et al., 1966)
in singers.
One major limitation of the provided data in this study is that the
measurements were taken in the supine body position due to the
use of a clinical horizontal-bore MRI system. In contrast to stud-
ies on vocal tract morphology, were only minor posture related
differences could be revealed in professional singers (Traser et al.,
2012), studies on posture related differences in normal breathing
showed that the DPH- motion in the supine position was signifi-
cantly greater than that in the upright position (Takazakura et al.,
2004). Functional residual capacity increased (Ibañez and Raurich,
1982; Navajas et al., 1988), vital capacity and forced VC (Vilke
et al., 2000) decreases when the subject changes from the sitting to
supine position. Using ultrasound, Houston et al. showed that in a
supine position the anterior DPH moved more, but the inspirated
volume was less than in the sitting position (Houston et al., 1992).
The exact influence of position on DPH motion is not understood in
detail; nevertheless professional singers today are used to singing
in different body positions especially in director-led theatre. Due
to the potential distractions in the noisy MR environment only pro-
fessionally trained singers could be included in the study but even
so, the possible influence of the Lombard effect (Lombard, 1911) on
phonation in the scanner cannot be excluded. The total number of
only 6 subjects is small, but the availability of professional trained
singers is limited.
5. Conclusion
In accordance with our first hypothesis different kinematic
behaviour could be detected between breathing and phonation.
During exhalation DPH and RC moved synchronously to reduce lung
volume, whereas during phonation the anterior DPH and RC as well
as middle and posterior diaphragm act as different functional units.
The range of DPH and RC movement in singing is compared to VC
breathing not maximally used in the beginning of phonation, but
even overstepped at the end, especially in the posterior DPH. It
seems possible that the alternating stabilization of one part of the
respiratory system could support singing by facilitating control of
psub . Interestingly a dependency of respiratory mechanics on pitch
or loudness in MPT was not observed, and only the duration of MPT
was significantly shorter for the highest pitch.
Author contribution
Conception and design: LT, AÖ, FB, MBu, MBo, BR, ME.
Acquisition of data: LT, AÖ, FB, MBu.
Analysis and interpretation: LT, AÖ, FB, MBu, MBo, BR, ME.
Drafting the manuscript for important intellectual content: LT,
AÖ, FB, MBu, MBo, BR, ME.
Final approval of the version to be published: LT, AÖ, FB, MBu,
MBo, BR, ME.
Acknowledgements
The authors thank Manfred Nusseck, PhD for his help with sta-
tistical evaluation and Dr Jude Brereton, PhD, for native correction.
The authors would also like to thank the subjects for their willing-
ness to take part in this study. Part of the material was presented
at the Pan European Voice Conference 2015 in Florence.
Matthias Echternach’s (grant no.: EC 409/1-1) Bernhard
Richter‘s (grant no.: RI 1050/4-1) and Michael Bock’s (grant no.:
MB 3025/2-2 and HA 7006/1-1) work is supported by the Deutsche
Forschungsgemeinschaft (DFG).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.resp.2016.11.007.
References
Biederer, J., Dinkel, J., Remmert, G., Jetter, S., Nill, S., Moser, T., Bendl, R., Thierfelder,
C., Fabel, M., Oelfke, U., Bock, M., Plathow, C., Bolte, H., Welzel, T., Hoffmann, B.,
Hartmann, G., Schlegel, W., Debus, J., Heller, M., Kauczor, H.-U., 2009.
4D-Imaging of the lung: reproducibility of lesion size and displacement on
helical CT MRI, and cone beam CT in a ventilated ex vivo system. Int. J. Radiat.
Oncol. Biol. Phys. 73, 919–926.
Bouhuys, A., Proctor, D.F., Mead, J., 1966. Kinetic aspects of singing. J. Appl. Physiol.
21, 483–496.
Bunch, M., Chapman, J., 2000. Taxonomy of singers used as subjects in scientific
research. J. Voice 14, 363–369.
Cluzel, P., Similowski, T., Chartrand-Lefebvre, C., Zelter, M., Derenne, J.P., Grenier,
P.A., 2000. Diaphragm and chest wall: assessment of the inspiratory pump
with MR imaging-preliminary observations. Radiology 215, 574–583.
Collyer, S., Thorpe, C.W., Callaghan, J., Davis, P.J., 2008. The influence of
fundamental frequency and sound pressure level range on breathing patterns
in female classical singing. J. Speech Lang. Hear. Res. 51, 612–628.
Echternach, M., Sundberg, J., Arndt, S., Markl, M., Schumacher, M., Richter, B., 2010.
Vocal tract in female registers–a dynamic real-time MRI study. J. Voice 24,
133–139.
Echternach, M., Traser, L., Richter, B., 2014. Vocal tract configurations in tenors’
passaggio in different vowel conditions-A real-time magnetic resonance
imaging study. J. Voice 28, 262e1–262e8.
Eichinger, M., Puderbach, M., Smith, H.-J., Tetzlaff, R., Kopp-Schneider, A., Bock, M.,
Biederer, J., Kauczor, H.-U., 2007a. Magnetic
resonance-compatible-spirometry: principle, technical evaluation and
application. Eur. Respir. J. 30, 972–979.
 L. Traser et al. / Respiratory Physiology & Neurobiology 236 (2017) 69–77
Eichinger, M., Tetzlaff, R., Puderbach, M., Woodhouse, N., Kauczor, H.-U., 2007b.
Proton magnetic resonance imaging for assessment of lung function and
respiratory dynamics. Eur. J. Radiol. 64, 329–334.
Gierada, D.S., Curtin, J.J., Erickson, S.J., Prost, R.W., Strandt, J.A., Goodman, L.R.,
1995. Diaphragmatic motion: fast gradient-recalled-echo MR imaging in
healthy subjects. Radiology 194, 879–884.
Hatabu, H., Chen, Q., Stock, K.W., Gefter, W.B., Itoh, H., 1999. Fast magnetic
resonance imaging of the lung. Eur. J. Radiol. 29, 114–132.
Hoffman, E.A., 1985. Effect of body orientation on regional lung expansion: a
computed tomographic approach. J. Appl. Physiol. 59, 468–480.
Houston, J.G., Morris, A.D., Howie, C.A., Reid, J.L., McMillan, N., 1992. Technical
report: quantitative assessment of diaphragmatic movement–a reproducible
method using ultrasound. Clin. Radiol. 46, 405–407.
Howard, D.M., Lindsey, G.A., Allen, B., 1990. Toward the quantification of vocal
efficiency. J. Voice 4, 205–212.
Howard, D.M., 1995. Variation of electrolaryngographically derived closed
quotient for trained and untrained adult female singers. J. Voice 9, 163–172.
Ibañez, J., Raurich, J.M., 1982. Normal values of functional residual capacity in the
sitting and supine positions. Intensive Care Med. 8, 173–177.
Kleinman, P.K., Raptopoulos, V., 1985. The anterior diaphragmatic attachments: an
anatomic and radiologic study with clinical correlates. Radiology 155, 289–293.
Kolar, P., Sulc, J., Kyncl, M., Sanda, J., Neuwirth, J., Bokarius, A.V., Kriz, J., Kobesova,
A., 2010. Stabilizing function of the diaphragm: dynamic MRI and
synchronized spirometric assessment. J. Appl. Physiol. 109, 1064–1071.
Kondo, T., Kobayashi, I., Taguchi, Y., Ohta, Y., Yanagimachi, N., 2000. A dynamic
analysis of chest wall motions with MRI in healthy young subjects. Respirology
5, 19–25.
Leanderson, R., Sundberg, J., 1988. Breathing for singing. J. Voice 2, 2–12.
Leanderson, R., Sundberg, J., von Euler, C., 1987. Breathing muscle activity and
subglottal pressure dynamics in singing and speech. J. Voice 1, 258–261.
Leanderson, R., 1987. Role of diaphragmatic activity during singing: a study of
transdiaphragmatic pressures. J. Appl. Physiol. 62, 259–270.
Lee, C.U., White, D.B., Sykes, A.-M.G., 2014. Establishing a chest MRI practice and its
clinical applications: our insight and protocols. J. Clin. Imaging Sci. 4, 17.
Lombard, É., 1911. Le signe de l’élévation de la voix. Ann. des Mal. l’oreille, du
larynx, du nez du pharynx 37, 101–119.
Macklin, C.C., 1922. X-ray studies on bronchial movemnets. Am. J. Anat. 35,
303–320.
McAllister, A., Sundberg, J., 1998. Data on subglottal pressure and SPL at varied
vocal loudness and pitch in 8- to 11-year-old children. J. Voice 12, 166–174.
Molinari, F., Puderbach, M., Eichinger, M., Ley, S., Fink, C., Bonomo, L., Kauczor,
H.-U., Bock, M., 2008. Oxygen-enhanced magnetic resonance imaging:
influence of different gas delivery methods on the T1-changes of the lungs.
Invest. Radiol. 43, 427–432.
Nason, L.K., Walker, C.M., McNeeley, M.F., Burivong, W., Fligner, C.L., Godwin, J.D.,
2012. Imaging of the diaphragm: anatomy and function. Radiographics 32,
E51–70.
Navajas, D., Farre, R., Rotger, M., Milic-Emili, J., Sanchis, J., 1988. Effect of body
posture on respiratory impedance. J. Appl. Physiol. 64, 194–199.
Özen, A.C., Traser, L., Echternach, M., Dadakova, T., Burdumy, M., Richter, B., Bock,
M., 2015. Ensuring safety and functionality of electroglottography
measurements during dynamic pulmonary MRI. Magn. Reson. Med. 76,
1629–1635.
77
Pettersen, V., Eggebø, T.M., 2010. The movement of the diaphragm monitored by
ultrasound imaging: preliminary findings of diaphragm movements in
classical singing. Logoped. Phoniatr. Vocol. 35, 105–112.
Pettiaux, N., Cassart, M., Paiva, M., Estenne, M., 1997. Three-dimensional
reconstruction of human diaphragm with the use of spiral computed
tomography. J. Appl. Physiol. 82, 998–1002.
Plathow, C., Schoebinger, M., Fink, C., Ley, S., Puderbach, M., Eichinger, M., Bock, M.,
Meinzer, H.-P., Kauczor, H.-U., 2005a. Evaluation of lung volumetry using
dynamic three-dimensional magnetic resonance imaging. Invest. Radiol. 40,
173–179.
Plathow, C., Zimmermann, H., Fink, C., Umathum, R., Schöbinger, M., Huber, P.,
Zuna, I., Debus, J., Schlegel, W., Meinzer, H.-P., Semmler, W., Kauczor, H.-U.,
Bock, M., 2005b. Influence of different breathing maneuvers on internal and
external organ motion: use of fiducial markers in dynamic MRI. Int. J. Radiat.
Oncol. Biol. Phys. 62, 238–245.
Proctor, D.F., 1980. Breathing, Speech and Sond. Springer, New York.
Simon, B.A., Kaczka, D.W., Bankier, A.A., Parraga, G., 2012. What can computed
tomography and magnetic resonance imaging tell us about ventilation? J. Appl.
Physiol. 113, 647–657.
Sorichter, S., 2009. Lung function. Radiologe 49, 676–686.
Sundberg, J., Leanderson, R., von Euler, C., 1989. Activity relationship between
diaphragm and cricothyroid muscles. J. Voice 3, 225–232.
Sundberg, J., Elliot, N., Gramming, P., Nord, L., 1993. Short-term variation of
subglottal pressure for expressive purposes in singing and stage speech: a
preliminary investigation. J. Voice 7, 227–234.
Sundberg, J., 1987. The Science of the Singing Voice. Northern Illinois University
Press.
Takazakura, R., Takahashi, M., Nitta, N., Murata, K., 2004. Diaphragmatic motion in
the sitting and supine positions: healthy subject study using a vertically open
magnetic resonance system. J. Magn. Reson. Imaging 19, 605–609.
Tetzlaff, R., Eichinger, M., 2009. Magnetic resonance imaging of respiratory
movement and lung function. Radiologe 49, 712–719.
Thomasson, M., Sundberg, J., 1999. Consistency of phonatory breathing patterns in
professional operatic singers. J. Voice 13, 529–541.
Thomasson, M., Sundberg, J., 2001. Consistency of inhalatory breathing patterns in
professional operatic singers. J. Voice 15, 373–383.
Thomasson, M., 2003. Belly-in or belly-out? Effects of inhalatory behaviour and
lung volume on voice function in male opera singers. Speech Transm. Lab. Q.
Status Prog. Rep. 45, 61–73.
Tokuda, J., Schmitt, M., Sun, Y., Patz, S., 2009. Lung motion and volume
measurement by dynamic 3D MRI using a 128-channel receiver coil. Acad.
Radiol. 16, 22–27.
Traser, L., Burdumy, M., Richter, B., Vicari, M., Echternach, M., 2012. The effect of
supine and upright position on vocal tract configurations during singing—a
comparative study in professional tenors. J. Voice, 1–8.
Vilke, G.M., Chan, T.C., Neuman, T., Clausen, J.L., 2000. Spirometry in normal
subjects in sitting, prone, and supine positions. Respir. Care 45, 407–410.
Watson, P.J., Hixon, T.J., 1985. Respiratory kinematics in classical (opera) singers. J.
Speech Hear. Res. 28, 104–122.
Watson, P.J., Hixon, T.J., Stathopoulos, E.T., Sullivan, D.R., 1990. Respiratory
kinematics in female classical singers. J. Voice 4, 120–128.