A Novel Treatment for Hypophonic Voice:

Twang Therapy

*†Lori E. Lombard and ‡Kimberly M. Steinhauer

*Indiana and †‡Pittsburgh, Pennsylvania

Summary: A hypophonic voice, characterized perceptually as weak and

breathy, is associated with voice disorders such as vocal fold atrophy and uni-
lateral vocal fold paralysis. Although voice therapy programs for hypophonia
typically address the vocal folds or the sound source, twang voice quality was
examined in this study as an alternative technique for increasing vocal power
by altering the epilarynx or the sound filter. Objective: This study investigated
the effect of twang production on physiologic, acoustic, and perceived voice
handicap measures in speakers with hypophonia. Design/Methods: This pro-
spective pilot study compared the vocal outcomes of six participants with hy-
pophonia at pre- and posttreatment time points. Outcome measures included
mean airflow rate, intensity in dB sound pressure level (SPL), maximum pho-
nation time, and self-report of voice handicap. Results: All subjects improved
in at least three of the four vocal outcome measures. Wilcoxon signed-rank
test of paired differences revealed significant differences between pre- and
posttherapy group means for airflow rate, SPL, and Voice Handicap Index
scores. Conclusion: The twang voice quality as a manipulation of the sound
filter offers a clinical complement to traditional voice therapies that primarily
address the sound source.
Key Words: Twang—Hypophonia—Voice therapy—Epilarynx.

INTRODUCTION most prevalent voice disorders among older

Hypophonic voice secondary to unilateral vocal
fold paralysis (UVFP) or atrophy is one of the
Accepted for publication December 20, 2005.
Presented at the Voice Foundation 29th Annual Symposium:
Care of the Professional Voice, June 2000, Philadelphia, PA.
From the *Indiana University of Pennsylvania, Indiana,
Pennsylvania; †University of Pittsburgh Voice Center, Pitts-
burgh, Pennsylvania; and the ‡VA Pittsburgh Healthcare Sys-
tem, University of Pittsburgh, Department of Communication
Science and Disorders, Pittsburgh, Pennsylvania.
Address correspondence and reprint requests to Lori Lom-
bard, 203 Davis Hall, Indiana University of Pennsylvania,
Indiana, PA 15705. E-mail: llombard@iup.edu
Journal of Voice, Vol. 21, No. 3, pp. 294–299
0892-1997/$32.00
Ó 2007 The Voice Foundation
doi:10.1016/j.jvoice.2005.12.006
adults.1,2 UVFP is caused typically by peripheral
involvement of the recurrent laryngeal nerve,
whereas vocal fold atrophy is associated with de-
generative neurologic disease or aging. Perceptual
characteristics of hypophonic voice include low
intensity, breathiness, and diplophonia.
Although the quantity of systematic assessments
of these disorders far outnumbers the quantity of
systematic treatment programs, current interven-
tions for hypophonia include both surgical and be-
havioral approaches. Surgical options include vocal
fold augmentation for mild glottal incompetence
and medialization for more significant glottal
gaps.3–5 Concurrent voice therapy often is advo-
cated to enhance postsurgical outcomes; yet, over
50% of patients are reluctant to select surgical

294
 HYPOPHONIC VOICE
intervention when offered voice therapy as a less
invasive alternative.4 However, there are little data
on the effectiveness of voice therapy alone.6
Thus far, behavioral treatment of hypophonic
voice has addressed the pathophysiology of the vo-
cal folds, or sound source, through approaches such
as forceful adduction exercises and the Vocal Func-
tion Exercise program (VFE).6,7 The therapy tech-
nique presented in this article is novel because it
approaches the hypophonic voice via constriction
of the epilarynx, or filter, using production of
a ‘‘twang’’ voice quality. The authors postulated
that the same twang quality used to boost vocal
power of the expert speaker or singer8 might pro-
vide a much-needed boost in vocal power to pa-
tients with a weak, breathy voice.
Twang is the bright, brassy, ringing voice quality
commonly heard in country-western singing, witch
cackling, a child’s ‘‘nya, nya’’ taunt, and is equated
often with duck quacking. Physiologically, twang is
produced by narrowing the epilarynx via constric-
tion of the aryepiglottic sphincter.8,9 The acoustic
outcome of narrowing this laryngeal vestibule is
a clustering of the third, fourth, and fifth formants
resulting in a higher amplitude of all vowel spectra
in the vicinity of 3.0 kHz.10 Perceptually, the voice
is amplified with piercing intensity because this
increase in energy between 2.8 and 4.3 kHz
corresponds with the resonant frequency of the ex-
ternal auditory meatus; as a result, fundamental fre-
quencies below 1.0 kHz receive a boost of 15–20
dB in sound transmitted to the middle ear without
an increase in vocal effort by the speaker.8 This pro-
duction has been exploited by expert speakers and
singers to boost vocal resonance or ‘‘squillo’’ and
is referred to as the speaker’s ring or singer’s for-
mant.8,10–12
A nonlinear theory of voice production9,13 attri-
butes this brilliant increase in acoustic power cou-
pled with a relative ease in physical production to
the properties of a vocal tract that is narrowed at
the epilarynx and free above—a configuration that
avoids interference with articulators involved in
vowel production (ie, tongue, lips, velum). Accord-
ing to this theory, the inertive vocal tract enhances
vocal fold vibration because the supraglottal pres-
sure driving the airflow is synchronous with the
velocity of the vocal folds. By constricting the
295
aryepiglottic sphincter and forming an inertive vo-
cal tract, the speaker can improve vocal fold oscil-
lation, increase amplitude of the speaker’s/singer’s
formant, and capitalize on the vocal ring necessary
for resonant voice projection. In addition, the glot-
tal/epilaryngeal ratio for twang quality is predicted
to approximate 1 due to the small glottis paired
with the small epilarynx; for this reason, theories
of impedance matching also predict that twang
should be an acoustically powerful voice quality.14
If optimizing resonance via twang is possible by
speakers with a healthy larynx, then epilaryngeal
constriction by speakers with atrophy or UVFP
may increase power in otherwise breathy, weak
voices. For this reason, Twang was explored as
a method to improve the voice quality of six pa-
tients exhibiting hypophonic voice. The primary
aim was to determine the effect of twang therapy
on aerodynamic, acoustic, and perceptual voice
outcomes in speakers with hypophonic voice as
measured at pre- and posttherapy. The authors
hypothesized that the twang speaking technique
produced by adults with atrophy or UVFP will sig-
nificantly decrease average airflow, increase mean
sound pressure level (SPL), increase maximum
phonation time (MPT), and decrease patient report
on the Voice Handicap Index (VHI).15
METHOD
Subjects
Six women between 52 and 73 years of age seek-
ing treatment for hypophonic voice at The Univer-
sity of Pittsburgh Voice Center participated in this
study. Consent was obtained before initiation of
the study. Four participants were identified with vo-
cal fold atrophy and two with UVFP. Participants
were selected after diagnosis of atrophy or UVFP
via endoscopic examination by a board-certified
otolaryngologist. Exclusion criteria included dys-
phagia. None of the participants had prior exposure
to the experimental voice task; yet, two subjects
previously sought voice therapy at another clinic.
Apparatus
Data collection for pre- and posttherapy evalua-
tions included an evaluation by an otolaryngologist
using flexible endoscopy and/or rigid endoscopy
Journal of Voice, Vol. 21, No. 3, 2007
296 LORI E. LOMBARD AND KIMBERLY M. STEINHAUER
and an evaluation by a certified speech-language
pathologist. The baseline voice analysis data were
collected in a sound-treated room. Subjects were
seated, and the mask from the Kay Elemetrics
Aerophone (Kay Elemetrics Corporation, Lincoln
Park, NJ) with pressure-flow sensors and a micro-
phone was pressed securely over their mouth and
nose. Subjects produced the stimulus /pI/ five times
in succession. The middle three syllable produc-
tions were analyzed to obtain the measures of
mean airflow rate and sound pressure level. Patients
were then asked to take as deep of a breath as pos-
sible and sustain /a/ as long as possible. Repeated
measures in seconds were obtained with a standard
stopwatch. The VHI, a patient self-report instru-
ment that quantifies perceptions of disability sec-
ondary to voice use, was completed by each
subject before and after treatment.15
General procedures
Voice tasks for twang therapy
After initial evaluation, subjects received 30-
minute, individual, weekly voice therapy sessions
with a certified speech-language pathologist. A se-
ries of tasks was facilitated by the clinician to elicit
the target behavior of twang resonance. The first
task was imitation without instruction. The clini-
cian produced rote speech (eg, days of the week)
using twang resonance. The subject was asked to
imitate the same voice quality and was provided
no behavioral instructions. The second task was
imitation of the clinician using the twang with the
most facilitative sounds and words (eg, /ae/,
‘‘maybe,’’ ‘‘back’’). To prevent excess vocal effort
and increase naturalness, task three included more
imitation of the same targets as task two except
with the instruction to minimize articulatory effort.
Subjects were frequently probed to evaluate and re-
duce vocal and articulatory effort. Once consistency
of the twang targets was established, task four, neg-
ative practice, was introduced to establish a propri-
oceptive understanding. Subjects were asked to
repeat the twang target (ie, the sound or word
with which they had the most accuracy and consis-
tency). Then they were asked to alternate the target
with counting to three using their baseline voice
quality, which was typically breathy and weak. Af-
ter the task, subjects were asked to identify and
Journal of Voice, Vol. 21, No. 3, 2007
verbalize their production strategy. Once a strategy
was established, task five of extended phrase length
was introduced to facilitate the subject’s ability to
implement the technique during speech phrases
that were more complex. After consistency and per-
ceptual voice quality naturalness was achieved,
subjects were asked to rate their vocal effort and/
or fatigue. If self-perceptions included a sense of
physical strain, task two was repeated to establish
the target without effort. With task five completed,
participants were asked to implement the technique
in conversational speech for a week. If participants
reported success without perceptions of physical ef-
fort or strain, they were discharged from therapy.
Total number of therapy sessions per subject ranged
from 2 to 8. All subjects were scheduled for a re-
evaluation for data collection 1 month after the
end of therapy. Due to scheduling conflicts, data
collection after therapy ranged from 3 weeks to 9
weeks.
Measurements
Three types of measurements were taken from
each participant: physiologic, acoustic, and quality
of life. The physiologic measures were mean air-
flow rate (MFR) and MPT. These measurements
represented coordination between laryngeal and re-
spiratory systems. MFR was collected while sub-
jects produced CV string /pI/ five times using the
Aerophone apparatus described above. Data were
analyzed on the middle three productions. MPT
was collected by having the subjects sustain /a/ as
long as possible.16 The acoustic measurement was
mean SPL as an indication of power. Mean SPL
was collected while subjects produced CV string /
pI/ five times using the Aerophone apparatus de-
scribed above. The VHI,17 the perceptual measure,
was used as a quality-of-life tool that represents
patient self-assessment of the emotional, physical,
and functional aspects of their voice. The 30-item
VHI asks the patient to answer each voice-related
question on a 5-point Likert scale ranging from
‘‘0,’’ ‘‘never’’ felt this way about the voice prob-
lem, to ‘‘4,’’ ‘‘always’’ felt this to be an issue.
Statistical analysis
Due to the exploratory nature of the study and
small sample size, the Wilcoxon signed-rank test
 HYPOPHONIC VOICE 297

of paired differences was used to analyze differ- TABLE 2. Means and Standard Deviations () for
ences between the pre- and postscores for each Pre- and Posttwang Therapy Time Points
dependent measure. Actual data (Table 1) and de- Outcome Measure Pre Post
scriptive statistics (Table 2) for each subject will
be reported. MFR (mL/sc)* 579.5 (195) 278.3 (130.3)
MPT (sc) 12.8 (4.3) 15.5 (2.1)
Intensity (SPL)* 71.2 (3.3) 75.3 (2.1)
RESULTS VHI* 51.8 (13.1) 33.0 (12.2)
Results are presented according to type of depen- *P ! 0.05.
dent measure and data for each subject are offered
in Table 1.
between the combined pre–post-therapy group
Physiologic measures means (Table 2).
MFR
From pre to posttherapy time points, MFR de- Perceptual measure
creased substantially for every subject indicating VHI
improvement in glottal-respiratory efficiency for From pre- to posttherapy time points, VHI total
these hypophonic subjects. The Wilcoxon signed- score decreased for all subjects indicating a per-
rank test of paired differences revealed a significant ceived decrease in the negative impact of their
difference between the combined pre–post-therapy voice disorder on daily living. Paired Wilcoxon
group means (Table 2). signed-rank tests revealed a significant difference
between the combined pre- to posttherapy group
MPT means (Table 2).
From pre- to posttherapy time points, MPT in-
creased for four subjects, decreased for one subject,
DISCUSSION
and remained unchanged for one subject. Although
the results are encouraging, paired Wilcoxon Results from this pilot study revealed that the use
signed-rank tests revealed no significant pre–post- of twang quality as an intervention for hypophonia
therapy differences. was a successful alternative for rehabilitation of the
weak voice. All subjects improved on one or more
Acoustic measure of the voice outcome measures indicating that
SPL manipulations of the epilarynx, or filter, can influ-
From pre- to posttherapy time points, SPL in- ence vocal output power even when vocal fold
creased for five subjects and decreased minimally oscillation, or sound source, is compromised. An
for one subject indicating an increase in vocal out- audibly stronger voice was verified by improvements
put power due to twang quality. Paired Wilcoxon in acoustic, physiologic, and voice-related self-
signed-rank tests revealed a significant difference assessment measures, ie, decreased MFR, increased
SPL, increased MPT, and decreased VHI score.
TABLE 1. Data for Subjects at Pre- and Posttwang The dominating theory underlying traditional hy-
Therapy Time Points pophonic voice therapy is a linear model that views
MFR MPT SPL VHI the vocal folds as the source, the vocal tract above
the vocal folds as the filter, and the respiratory sys-
Subject No. Pre Post Pre Post Pre Post Pre Post tem as driving power force.18 Therefore, popular
1 641 305 21 19 68.6 78.9 55 43 interventions exploit breath support and phonatory
2 655 322 14 14 68.2 75.9 67 28 exercises such as sustained phonation, pitch glides,
3 306 215 9 16 68.8 72.6 64 47 and forceful adduction in a bottom-up fashion to in-
4 550 499 11 16 72.4 75.2 37 21 crease vocal output power.
5 721 122 13 13 77.1 74.4 29 18
A recent, contrasting theory of vocal out-
6 466 207 11 15 72.2 74.7 76 41
put power9 presents source-filter interactions as

Journal of Voice, Vol. 21, No. 3, 2007

298 LORI E. LOMBARD AND KIMBERLY M. STEINHAUER
a nonlinear dynamic model in which the shape of
the filter, in this case, epilaryngeal space right
above the vocal folds, can optimize the oscillatory
pattern of the vocal folds below (the source). In ad-
dition, Titze14 used computer simulations to exam-
ine the effect of glottal-to-epilaryngeal cross-
sectional area ratios on radiated oral vocal output
power. Predictions were based on the Maximum
Power Transfer Theorem for simple electrical cir-
cuits that states output power should be maximized
as the ratio of source and load impedance ap-
proaches 1.0. The results of the simulation sup-
ported the hypotheses and revealed that voice
output power was maximized as the ratio of glottal
source impedance and vocal tract (epilarynx) load
impedance approached 1.0.
According to Titze’s theory, two voice qualities
were predicted to meet the optimal impedance
matching configurations: yawn sigh, with a large
glottis and large epilarynx, and twang quality,
with a small glottis and small epilarynx. For this
reason, the authors hypothesized that the epilaryng-
eal narrowing in twang quality first facilitated vocal
fold oscillation, thus leading to a longer closed
phase. In turn, this small glottis balanced with
a small epilarynx may have generated an efficient
source/filter ratio that shifted closer to one. An ad-
vantage of this new glottal/epilaryngeal relationship
was a boost in vocal power paired with an ease of
vocal production, as confirmed by lower VHI
scores and patient testimony.
All subjects commented on how pleased they
were to finally achieve loudness without the high
degree of fatigue and effort they previously experi-
enced. They also were relieved that twang added
loudness in their voices, yet sounded more natural
than the piercing twang modeled by the clinician.
Perhaps, this phenomenon occurred because of
the dynamic interaction between a functioning epi-
larynx and a compromised glottal source, thus
dampening the ‘‘annoying’’ piercing quality inher-
ent in a healthy twang quality. One subject with
an introverted personality objected to making
such a brazen twang quality; however, she elected
to use it socially, for speaking in noisy environ-
ments and professionally, and for presentation
demands. Her ability to move between speaking
techniques was facilitated due to the ease of
Journal of Voice, Vol. 21, No. 3, 2007
acquiring the twang technique. Remarkably, five
of the six subjects acquired the twang quality
within two sessions.
The encouraging results from this pilot study
provide a starting point for future research explor-
ing the mechanisms and effectiveness of twang
quality as an intervention for hypophonic voice. Al-
though caution must be taken when interpreting the
results of studies with a small sample size, the clin-
ical evidence for twang as an impetus for increasing
vocal power from these six subjects is powerful as
observed in the selected measures. Future studies
should compare the relationship between twang
technique and additional voice outcomes, such as
intensity between 2.0- and 3.5-kHz bandwidth.
Increases of acoustic energy in this region were
observed in this pilot study but not quantified. In
addition, effectiveness of the twang technique
should be tested using a larger sample size within
a prospective, randomized, controlled design.
CONCLUSION
The authors explored twang technique as a strat-
egy for treating the weak, breathy, hypophonic
voice. Preliminary results showed that speakers
with hypophonia who produced twang quality de-
creased MFR, increased MPT, increased intensity,
and decreased self-perception of voice handicap.
These results support a nonlinear theory of vocal
output power that exploits glottal epilaryngeal rela-
tionships as the fundamental premise.
This new voice therapy technique promises ex-
citing clinical potential. The strategy is both easy
for clinicians to teach and for patients to learn.
For this reason, twang may be used as a complement
to traditional therapies.
REFERENCES
1. Coyle SM, Weinrich BD, Stemple JC. Shifts in relative
prevalence of laryngeal pathology in a treatment-seeking
population. J Voice. 2001;15:424–440.
2. Kelchner LN, Stemple JC, Gerdeman B, Le Borgne,
Adam S. Etiology, pathophysiology, treatment choices,
and voice results for unilateral adductor vocal fold
paralysis: a 3-year retrospective. J Voice. 1999;13:592–
601.
 HYPOPHONIC VOICE
3. Adams SG, Irish JC, Durkin LC, Wong D, Brown DH.
Evaluation of vocal function in unilateral vocal fold paral-
ysis following thyroplastic surgery. J Otolaryngol. 1996;
25:165–170.
4. Heuer RJ, Sataloff RT, Emerich K, Rulnick R, Baroody M,
Spiegel JR, et al. Unilateral recurrent laryngeal nerve
paralysis: the importance of preoperative voice therapy.
J Voice. 1997;11:88–94.
5. Verdolini K. In: National Center for Voice and Speech’s
Guide to Vocology. Denver, CO: National Center for Voice
and Speech; 1998.
6. Miller S. Voice therapy for vocal fold paralysis. Otolaryngol
Clin North Am. 2004;37:105–119.
7. Stemple JC, Glaze LE, Gerdeman BK. Clinical Voice
Pathology: Theory and Management. 2nd ed. San Diego,
CA: Singular Publishing Group; 1995.
8. Yanagisawa R, Estill J, Kmucha ST, Leder SB. The contri-
bution of aryepiglottic constriction to ‘‘ringing’’ voice
quality—a videolaryngoscopic study. J Voice. 1989;3:
342–350.
9. Titze IR. Acoustic interpretation of resonant voice. J Voice.
2001;15:519–528.
10. Sundberg J. The Science of the Singing Voice. DeKalb, IL:
Northern Illinois University Press; 1987.
299
11. Barrichelo VMO, Heuer RJ, Dean CM, Sataloff RT.
Comparison of singer’s formant, speaker’s ring, and LTA
spectrum among classical singers and untrained normal
speakers. J Voice. 2001;15:344–350.
12. Vennard W. Singing.the Mechanism and the Technic.
New York: Carl Fischer, Inc; 1967.
13. Titze IR, Story BH. Acoustic interactions of the voice
source with the lower vocal tract. J Acoust Soc Am.
1996;101:2234–2243.
14. Titze IR. Regulating glottal airflow in phonation: applica-
tion of the maximum power transfer theorem to a low-
dimensional phonation model. J Acoust Soc Am. 2002;
111:367–376.
15. Jacobsen BH, Johnson A, Grywalski C, Silbergleit A,
Jacobsen G, Benninger MS, et al. The Voice Handicap
Index (VHI): development and validation. Am J Speech
Lang Pathol. 1997;6:66–70.
16. Baken RJ, Orlikoff R. Clinical Measurement of Speech and
Voice. 2nd ed. San Diego, CA: Singular Publishing; 2000.
17. Jacobsen BH, Johnson A, Grywalsky C. The Voice Hand-
icap Index (VHI): development and validation. J Voice.
1998;12:540–550.
18. Fant G. Acoustic Theory of Speech Production. The
Hague, The Netherlands: Mouton; 1960.
Journal of Voice, Vol. 21, No. 3, 2007