See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/309744181

The Effectiveness of Low-Level Light Therapy in Attenuating Vocal Fatigue

Article · November 2016

DOI: 10.1016/j.jvoice.2016.09.004

CITATIONS READS
18 1,069

2 authors, including:

James Heaton
Harvard Medical School
133 PUBLICATIONS   2,971 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Development of improved voice prosthesis technology View project

Assessment of facial function in peripheral facial palsy View project

All content following this page was uploaded by James Heaton on 04 December 2017.

The user has requested enhancement of the downloaded file.

 ARTICLE IN PRESS
The Effectiveness of Low-Level Light Therapy in
Attenuating Vocal Fatigue
*Loraine Sydney Kagan and *,†James T. Heaton, *†Boston, Massachusetts

Summary: Objectives. Low-level light therapy (LLLT) is effective in reducing inflammation, promoting wound
healing, and preventing tissue damage, but has not yet been studied in the treatment of voice disorders. The objective
of this study was to investigate the possible effectiveness of LLLT in attenuating symptoms of vocal fatigue created
by a vocal loading task as measured by acoustic, aerodynamic, and self-reported vocal effort.
Methods. In a randomized, prospective study, 16 vocally healthy adults divided into four groups underwent a 1-hour
vocal loading procedure, followed by infrared wavelength LLLT (828 nm), red wavelength LLLT (628 nm), heat, or
no heat–light (control) treatment targeting the laryngeal region of the ventral neck surface. Phonation threshold pres-
sure (PTP), relative fundamental frequency (RFF), and the inability to produce soft voice (IPSV) self-perceptual rating
scale were recorded (1) at baseline, (2) immediately after vocal loading, (3) after treatment, and (4) 1 hour after treatment.
Results. Vocal loading significantly increased PTP and IPSV and decreased onset and offset RFFs, consistent with a
shift toward vocal dysfunction. Red light significantly normalized the combination of PTP, IPSV, and RFF measures
compared to other conditions.
Conclusions. RFF is sensitive to a vocal loading task in conjunction with PTP and IPSV, and red LLLT may have a
normalizing effect on objective and subjective measures of vocal fatigue. The results of this study lay the groundwork
and rationale for future research to optimize LLLT wavelength combinations and overall dose.
Key Words: Low-level light therapy–Vocal fatigue–Vocal hyperfunction–Relative fundamental frequency–
Dysphonia.

INTRODUCTION therapy for muscle tension dysphonia (T. Stadelman-Cohen, per-

Low-level laser therapy or low-level light therapy (LLLT) has
been proven to be a noninvasive therapeutic procedure for re-
ducing inflammation and pain, and for promoting faster wound
healing and nerve regeneration.1 Efficacious light wavelengths
include red (approximately 633 nm) and infrared (IR) (approx-
imately 830 nm), which can be emitted via laser or light-
emitting diode (LED) light sources. Although LLLT is typically
administered by licensed dermatologists and physical thera-
pists, LLLT devices are also available over the counter (OTC)
to rejuvenate aged skin or to treat inflammation or pain in pets
and people. For example, the Omnilux New-U is approved by
the Food and Drug Administration (without a prescription) to
treat periorbital skin wrinkles (ie, “crow’s-feet”) with both red
and IR LED lights in a handheld unit for self-administration to
the skin surface. Similarly, the Willow Curve (Physicians Tech-
nology, LLC) is an array of red and IR LED lights for at-home
treatment of pain and inflammation.
Speech–language pathologists (SLPs) at the Massachusetts
General Hospital Center for Laryngeal Surgery and Voice Re-
habilitation have anecdotally noted that the application of red
light from a New-U unit placed on the neck surface overlying
the larynx for 5–10 minutes can cause neck muscles to feel “more
relaxed” as self-reported by some individuals receiving voice
Accepted for publication September 8, 2016.
From the *MGH Institute of Health Professions, Charlestown Navy Yard, Boston, Mas-
sachusetts; and the †Massachusetts General Hospital Department of Surgery, Harvard Medical
School, Boston, Massachusetts.
Address correspondence and reprint requests to James T. Heaton, MGH Center for
Laryngeal Surgery and Voice Rehabilitation, 1 Bowdoin Square, 11th Floor, Boston, MA
02144. E-mail: James.Heaton@mgh.harvard.edu
Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■
0892-1997
© 2016 The Voice Foundation. Published by Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jvoice.2016.09.004
sonal communication, April 4, 2014). An SLP at the Center has
also noted that LLLT gives a similar feeling to having “warmed
up” vocally (perhaps as a result of the heat generated by the red
wavelength), although a placebo effect cannot be ruled out. Ra-
diographic testing conducted at the Center revealed that both red
and IR light wavelengths from the New-U LLLT device are
capable of at least partially penetrating the ventral neck surface
through to the laryngeal mucosa within the airway, showing the
potential for LLLT to affect laryngeal physiology. Specifically,
the vocal folds and paraglottic tissues glow brightly during IR
LLLT application when viewed with an IR sensitive camera
(Microsoft LifeCam VX-3000, Microsoft, Redmond Washing-
ton, USA) or during red LLLT when viewed with a rigid
endoscope (without fiberoptic illumination). In addition, a pho-
todiode radiometer (Nova II meter with PD300-TP probe, Ophir
Optronics, North Andover, Massachusetts, USA) positioned in
the canine glottis indicates that approximately 1.5% of the 633-
nm (red) light and 4.8% of the 830-nm (IR) light reach the medial
vocal fold surface from neck-placed New-U LLLT units (un-
published observation).
LLLT involves exposing cells or tissues to levels of light,
usually in the red and IR wavelengths, that are relatively low
in energy compared to what is used for purposes such as abla-
tion, cutting, and thermally coagulating tissue. LLLT was
discovered in 1967 by Mester et al,2 who noticed that applying
a low-level laser to the backs of shaven mice induced faster hair
regrowth. Since the initial observations of Mester et al, the ap-
plications of LLLT have grown tremendously and now include
wound healing, tissue repair, relief of inflammation and edema
of injuries or chronic diseases, and nerve regeneration.
The cellular mechanisms of LLLT are not completely under-
stood. The most promising theory is that LLLT acts on
 ARTICLE IN PRESS
2 Journal of Voice, Vol. ■■, No. ■■, 2016
mitochondria to increase adenosine triphosphate (ATP) produc-
tion and modulate reactive oxygen species, and creates
downstream effects resulting in increased transcription.1 Red and
IR wavelengths absorbed by chromophores in mitochondria cause
electrons within the chromophore to jump to a higher energy
level, causing a reaction in the transmembrane complex in the
electron transport chain. This photobiomodulation results in in-
creased electron transport and thus increased production of ATP.3,4
Upregulation of mitochondria causes increased generation of re-
active oxygen species, which activate more transcription factors,
leading to increased stimulatory and protective genes.5 This in-
duction of transcription factors causes downstream effects,
including increased cell proliferation, growth factors, inflam-
matory modulators, and increased tissue oxygenation.6
There has been a great deal of evidence demonstrating the pos-
itive effects of LLLT on wound healing, tissue repair, relief of
inflammation and edema of injuries or chronic diseases, and nerve
regeneration. To promote wound healing, it is theorized that LLLT
induces the local release of cytokines, chemokines, and other
biological response modifiers that increase the wound strength
and reduce the time to heal.7–9 LLLT has also been found to
promote collagen formation and induce neovascularization.10,11
In sports medicine, a field that shares similarities to vocal per-
formance, LLLT has been shown to increase endurance, strength,
and speed of recovery for athletes as well as aid in the relief of
inflammation and edema of injuries or chronic diseases.12–15 Ad-
ditionally, LLLT has been shown to improve neural regeneration
after a stroke or traumatic brain injury, in neurodegenerative dis-
eases, and after spinal cord and peripheral nerve damage.16
Although vocal fatigue is largely classified as a subjective phe-
nomenon, it has both physiological and biological bases and can
be measured accordingly. From a neuromuscular perspective,
when skeletal muscle fatigue and force required exceeds avail-
able output, additional muscles are recruited. While true of skeletal
muscles, this is not entirely the case for the laryngeal muscu-
lature. Although Boucher et al17 found that lateral cricoarytenoid
(LCA) muscles exhibited signs of EMG spectral compression
(a marker of muscular fatigue) in subjects performing a vocal
loading task over 12–14 hours, it is the prevailing opinion that
the laryngeal musculature is composed mainly of fatigue-
resistant muscle fibers and does not fatigue with normal speaking
activities.18 Reactive vocal hyperfunction of the intrinsic or ex-
trinsic laryngeal musculature may occur, however, negatively
impacting the voice. When an individual is vocally fatigued, re-
active vocal hyperfunction of the intrinsic or extrinsic laryngeal
musculature may occur, which negatively impacts the voice.19,20
Stepp et al19 found that the offset and onset relative fundamen-
tal frequencies (RFFs) surrounding a voiceless consonant were
decreased for individuals with vocal hyperfunction, presum-
ably because of excessive tension of the laryngeal muscles.
Another source of vocal fatigue may result from mechanical stress
placed on the vocal fold mucosa, which increases vocal fold
viscosity.21,22 Phonation threshold pressure (PTP), which is the
lowest subglottal pressure needed to initiate and sustain vocal
fold vibration, increases during vocally fatiguing tasks, largely
because of the vocal folds’ increased viscosity.23,24 A less inva-
sive measure of vocal fold viscosity, the inability to produce soft
voice (IPSV), asks a subject to perform a specific soft voice task
and to rate its difficulty on a 10-point scale.25 Elevated IPSV
values have been found for up to 24 hours following acute
phonotrauma, but ultimately resolve completely.26 According-
ly, RFF, PTP, and IPSV were used to objectively measure signs
of vocal fatigue in the present study.
There has only been one previous study of LLLT on the la-
ryngeal area. In a preliminary study of the application of LLLT
to treat intubation-induced laryngopharyngeal reflux in rats, rats
were intubated and given four irradiation sessions at 780 nm,
45-second durations (delivering 4.2 J) at 48-hour intervals. La-
rynges were then analyzed for myeloperoxidase (MPO) activity,
which is an indicator of neutrophil migration, a negative effect
triggered by intubation. The larynges of the group that had LLLT
displayed significantly increased levels of MPO compared to con-
trols that had not been intubated, but showed decreased levels
of MPO compared to controls that had been intubated but
untreated.27 Therefore, the use of LLLT may be beneficial in the
voice-disordered population, including for those whose patholo-
gies encompass some alteration of the lamina propria (eg, trauma
and fibrovascular changes).
Vocal fatigue has traditionally been classified as a symptom
reported subjectively by individuals. Typical complaints asso-
ciated with reported vocal fatigue include increased effort and
discomfort, reduced pitch range and flexibility, reduced loud-
ness or impaired quality, increased fatigue over the course of a
day, and improvement after vocal rest.28–31 Clinically, the treat-
ment of vocal fatigue involves reducing muscular tension (if
present) through exercises or palpation; preventing or limiting
the inflammatory response through vocal hygiene, hydration, or
warm-ups; or promoting wound healing through vocal rest.32 In-
flammation reduction and wound healing are both involved in
treating vocal fatigue,32–34 which is a typically temporary yet de-
bilitating voice condition. Therefore, LLLT is an ideal candidate
for a pilot study of a new therapy for the treatment of vocal fatigue.
Although LLLT has been widely used in other fields relating
to inflammation and wound healing, no scientific study of the
effectiveness of LLLT on the treatment of voice disorders has
yet been undertaken. The purpose of the present study was to
determine the effectiveness of LLLT in attenuating symptoms
of vocal fatigue created by a vocal loading task as measured by
acoustic, aerodynamic, and self-reported vocal effort. It was hy-
pothesized that two common wavelengths of LLLT (red and IR)
would each have a positive effect on measures associated with
muscular fatigue and inflammation. If LLLT has a beneficial impact
on vocal musculature and tissues, it could have wide clinical rel-
evance to populations with voice disorders. Not only might LLLT
speed the recovery of certain patients, but also it might relieve
patients’ discomfort during the healing process as well.
METHODS
Participants
The study was approved by the Massachusetts General Hospi-
tal Institutional Review Board. Sixteen vocally healthy adults
participated in the present study (ages 22–35, M = 26, standard
deviation [SD] = 3.7; 5 males, 11 females). Specific inclusion
 ARTICLE IN PRESS
Loraine Sydney Kagan and James T. Heaton Effectiveness of LLLT in Vocal Fatigue 3

criteria were nonsinger males or females between the ages of

18 and 44 years who had no known history of voice disorders
and had never been patients of the MGH Voice Center to ensure
that the measures collected are representative of the vocally normal
adult population. Potential subjects were screened via the Con-
sensus Auditory-Perceptual Evaluation of Voice (CAPE-V) and
laryngeal endoscopic exam by a clinically certified SLP. Spe-
cifically, all participants passed screening for normal laryngeal
appearance (color, symmetry of movement, closure pattern, and
absence of surface pathologies) and CAPE-V rating. The par-
ticipants were also screened for light sensitivity to the red and
IR light wavelengths with a 20-minute application to the forearm
(per the product instructions) and were asked to report any per-
sistent redness beyond 24 hours following application. For the
day of the study participation, the participants were asked to avoid
vocally fatiguing behaviors (screaming or yelling) before coming
to the Center.

Materials and protocols

One popular handheld LED device commercially available OTC
for LLLT is the Omnilux New-U, which emits both red (628 nm)
and IR (828 nm) lights, and has been shown to reduce the ap-
pearance of periorbital wrinkles.35 This device was selected
because of its portability and commercial availability, as well
as its ability to produce both red and IR wavelengths. Two
Omnilux New-U devices were yoked together into a collar device
so that the lights emit directly over the thyroid laminae
(Figure 1A–D).
The total power output of the New-U, which has a 4.7 × 6.1 cm
array of 60 red LEDs (628 nm) and 28 IR LEDs (828 nm), is
70 mW/cm2 for the red setting and 55 mW/cm2 for the IR setting.
The recommended daily dose for the New-U in treating perior-
bital wrinkles is 20 minutes for either wavelength, which was used
as the dosage in this experiment. The light energy delivered to
the skin and deeper tissues under the LED array in the present
study was well below international safety limits for these wave-
lengths (ie, approximately 1/20th of the limit) according to the
device’s operating manual. The red setting heats the neck surface
to approximately 43.3°C, whereas the IR setting heats the neck
surface to 35.5°C based on a thermometer (Omega HH506R,
Omega, Norwalk, Connecticut, USA) attached with surgical tape FIGURE 1. All light/heat conditions were administered with the above
to the skin over the thyroid lamina. For the heat control (no light), therapy room setup (A—top view, B—side view). Two LLLT devices
sham devices were created by covering the Plexiglas cover of the were yoked together for light or heat delivery to both sides of the neck.
LLLT units with optically opaque polymer tape and cloth tape Cloth was taped to the upper face to block potential awareness of the
(Figure 1C). The heat controls created a sustained temperature experimental condition. C. The heat control device was made by placing
between 42.2°C and 43.5°C when the red light setting was active optically opaque polymer tape and cloth tape on the LLLT units. D.
(with light transmission blocked). For the control group, LLLT The nonheated control group had neck-placed LLLT units that were
units were placed on the ventral neck surface identically to the not turned on, with additional light-blocked units placed on each side
other condition, but additional units were placed next to partici- of the head to provide the LLLT unit fan noise. LLLT, low-level light
pants’ ears and turned on to create the same sound that the other therapy.
therapies produced (Figure 1D). For all therapies, the patients were
asked to lie down on a massage table in a therapy room and were
blindfolded to prevent response bias to the intervention (red light, Germany) positioned approximately 10 cm from the partici-
IR light, heat control, or control; Figure 1A and B). pant’s lips. The microphone preamplifier (Model 302 Dual
Vocal recordings were conducted in a sound-treated booth at Microphone Preamplifier; Symetrix, Inc., Mountlake Terrace, WA)
the MGH Voice Center using a head-mounted condenser mi- output was band-pass filtered from 10 to 10,000 Hz (CyberAmp
crophone (MKE104; Sennheiser Electronic GmbH, Wennebostel, Model 380; Axon Instruments, Inc., Union City, CA) and
 ARTICLE IN PRESS
4 Journal of Voice, Vol. ■■, No. ■■, 2016
digitized at 20 kS/s, 16-bit quantization using Axoscope soft-
ware (Version 10.4, Molecular Devices LLC, Sunnyvale
California, USA). The microphone was calibrated using an in-
traoral electrolarynx (Cooper Rand 9V, Luminaud Inc, Mentor,
Ohio USA) and a sound level meter (Rion NL-20, Rion Co.,
Tokyo, Japan).
Whereas the vocal loading procedures from both Verdolini
Abbott et al34 and Hunter and Titze26 have been shown to in-
crease measures associated with vocal fatigue, the vocal loading
procedure for this experiment was taken from Verdolini Abbott
et al34 because of the shorter length of the procedure (1 hour vs
2 hours) and because the procedure was more clearly outlined
and not specifically tailored to teachers. Vocal loading trials were
audio recorded with a headset microphone and the sound wave-
form was displayed in real time on a computer monitor. A sound
level meter was held at 30 cm from the participants’ mouths and
the participants were asked to repeat a familiar phrase while grad-
ually increasing the volume until the examiner noted consistent
peaking above 75 dB on the sound level meter and marked that
threshold on the waveform graphic display. The participants were
shown the 75-dB threshold on the waveform graphic display and
were instructed to maintain a vocal amplitude exceeding that thresh-
old (with the exception of pauses between words and phrases).
The participants were cued to read out loud from a book as if
they were in a large room, without amplification, for 15 minutes,
followed by 5 minutes of rest. The reading task was repeated
three times. The examiner monitored the screen output nearly
constantly and cued the subjects to maintain a target intensity
range of 75–90 dB during phonatory loading. Because of the length
of the experiment, the patients were provided with a 6-oz bottle
of water to maintain comfort levels during the vocal loading task.
IPSV collection was based on the protocol from Hunter and
Titze.26 The participants were provided written and oral instruc-
tions and were asked to perform the following soft-voice tasks
and to rate them on a scale of 1–10 for difficulty: sustaining a steady
/i/, gliding from a low pitch to a high pitch on /i/, repeating a stac-
cato /i-i-i-i-i/, and singing the first two phrases of “Happy Birthday.”
The participants were asked to keep in mind “undesirables” such
as roughness, breathiness, phonation breaks, and louder than ex-
pected phonation as signs that a task was more difficult. IPSV was
calculated by averaging the difficulty rating for the four tasks at
each time point. Although not specifically asked, the majority of
the participants spontaneously commented that they felt vocally
fatigued or affected following the vocal loading task.
PTP collection was based on the protocol from Verdolini Abbott
et al,34 in which PTP was shown to increase post loading.
Subglottal pressure was indirectly measured through intraoral
pressure (Glottal Enterprises oral pressure transducer and adapter,
Glottal Enterprises, Syracuse NY, USA; Axoscope 10.4 soft-
ware; Molecular Devices LLC, Sunnyvale California, USA)
during bilabial voiceless consonants. Subjects produced repeat-
ed sets of /pi pi pi pi pi/ utterances as quietly as possible at C4
(262 Hz, males) or C5 (523 Hz, females), indicated with an elec-
tronic piano keyboard, and at a rate of 90 beats/minute, indicated
by a metronome. Pitches were verified perceptually by a trained
examiner to an accuracy of about one-quarter tone for each syl-
lable. The examiner also monitored the subjects’ loudness for
PTP trials, empirically, continually encouraging them to phonate
as quietly as possible. Training for the PTP task continued until
the examiner and the subject considered that the subject could
perform the task reliably according to criteria. Calibration files
were created with a pressure calibrator (PC-1H, Glottal Enter-
prises) with values at 0, 5, 10, 15, and 20 cm H2O to convert
from volts to centimeters of water during data analysis.
Measurements and analysis for RFF were made using the pro-
tocol developed by Lien and Stepp.36 The participants were
provided a printout of the stimuli and were asked to read the
following utterances: /afa afa afa/ /ifi ifi ifi/ /ufu ufu ufu/ in their
comfortable pitch and loudness, with nearly equal stress on both
syllables. The experimenter asked the participant to repeat any
utterances that were mispronounced or glottalized (eg, /afʔa/) to
ensure adequate duration of the second vowel phoneme for analysis.
Experimental design
The participants were randomly assigned and blinded to one of
four groups: IR LLLT, red LLLT, heat control, or control. The
participants were measured for baseline PTP, IPSV, and RFF.
The participants then performed a vocal loading task for 60
minutes. PTP, IPSV, and RFF were collected immediately fol-
lowing the loading task. These recordings took approximately
5–7 minutes. The participants then received assigned therapy.
Post-treatment PTP, IPSV, and RFF were measured immediate-
ly after treatment, as well as 1 hour after treatment to assess any
latent therapy effects. The patients were asked to maintain vocal
silence in between measurements to maintain as consistent a level
of vocal loading as possible.
Data analysis
PTP data were analyzed based on the protocol from Verdolini
Abbott et al.34 Using Axoscope software (Molecular Devices LLC,
Sunnyvale California, USA) for each set of three /pi pi pi pi pi/
strings, the examiner selected the set with the lowest, flattest pres-
sure signal peaks. Production was derived from syllables two to
five. The examiner identified the midpoints of the oral pressure
peaks for syllables 2–3, 3–4, and 4–5. Using calibration files made
during the experiment sessions, these three points were con-
verted from volts to centimeters of waters. The three oral pressures
during the /p/ occlusion were accepted as an approximation of
PTP and averaged. For one participant, the examiner was unable
to elicit five continuous flat peaks, and those results were excluded.
To extract RFF from recordings, the examiner was trained by
Stephanie Lien, the first author of a validated protocol for mea-
suring RFF.36 Using Praat (Praat Software, Amsterdam, The
Netherlands), 11 pulse timings before and after the voiceless con-
sonant in each vowel–voiceless consonant–vowel combination,
compromising 20 vocal cycles, were analyzed for the two RFF
variables: offset and onset RFFs. For each of those 20 vocal
cycles, RFF was calculated by taking the inverse of the period
and converting it to semitones (STs) relative to the reference fre-
quency using Equation 1, in which f is the frequency of the
measured cycle and fref is the reference frequency:
⎛ f ⎞
RFF (ST) = 39.86 × log10 ⎜⎜ ⎟⎟⎟. (1)
⎜⎝ fref ⎠
 ARTICLE IN PRESS
Loraine Sydney Kagan and James T. Heaton Effectiveness of LLLT in Vocal Fatigue
The reference frequency is the cycle furthest away from the
voiceless consonant (offset cycle 1 for all offset cycles or onset
cycle 10) and is the most likely to capture a steady-state portion
of the vowel. The reference frequency for each instance is used
to normalize values to allow for comparison across individuals
with different base fundamental frequencies. RFF instances were
excluded if there were glottalizations, pitch breaks, irregular or
W-shaped waveforms, or less than 10 vocal cycles. Addition-
ally, RFF instances were excluded if the RFF values for offset
cycle 2 and onset cycle 9 (the vocal cycles directly adjacent to
the reference cycle) had an absolute value >0.8 ST, indicating
that fref was not in a steady state.
Statistical analysis
Measurements for all variables were zeroed by the partici-
pants’ baseline to eliminate any difference at baseline and then
standardized and combined into one data set. Signs for RFF offset
and RFF onset were reversed to correspond to the same direc-
tion of change anticipated for PTP and IPSV measures from the
vocal loading task. A repeated measures analysis of variance was
conducted on these standardized scores to analyze the effect by
each individual treatment condition. Post hoc Tukey tests were
performed to assess both the effect of the loading task and any
treatment effects. Because of the small sample size, only spe-
cific post hoc contrasts were chosen (a priori) to analyze change
post treatment and 1 hour post treatment to examine the rate of
recovery by condition. For all statistical analyses, alpha was set
at 0.05.
RESULTS
Results of the analysis of variance examining the effects of con-
dition and time on aggregated standardized scores indicated
significant effects for both condition and time (Figure 2; con-
dition, P = 0.01; time, P < 0.0001; multiple R2 = 0.1667). There
was no interaction effect with time and condition (P = 0.58). Fol-
lowing up with post hoc Tukey tests, standardized scores for all
measurements were significantly higher (disordered) post loading
compared to baseline (P < 0.001) and were significantly lower
(normalized) 1 hour post treatment compared to baseline
(P < 0.0001). Values were marginally lower immediately post
treatment (P = 0.07).
On average, vocal loading decreased RFF offset cycle 10 (base-
line: M = 1.10 ST, SD = 0.53; post loading: M = −1.31 ST,
SD = 0.56) and RFF onset cycle 1 (baseline: M = 2.52 ST,
SD = 0.82; post loading: M = 2.21 ST, SD = 0.62; Figures 3 and
4). On average, vocal loading increased PTP (baseline:
M = 5.06 cm H2O, SD = 1.12; post loading: M = 5.90 cm H2O,
SD = 1.81; Figure 5) and IPSV (baseline: M = 2.19, SD = 1.47;
post loading: M = 4.89, SD = 1.96; Figure 6).
Treatment effects were then analyzed by group for change post
treatment and 1 hour post treatment compared to post loading.
Reported values below reflect change in actual measurement post
loading. Previous research has indicated that PTP and IPSV in-
crease in response to vocal loading and decrease throughout the
healing trajectory.23,26,34,37 Conversely, RFF has been shown to
respond in the opposite direction (ie, decreases) in the pres-
ence of vocal hyperfunction.19,38 Control group participants
5
FIGURE 2. Aggregated standardized scores for all variables (inabil-
ity to produce soft voice, phonation threshold pressure, and relative
fundamental frequency) are plotted as a function of the treatment group
(N = 4 per group) and time (post loading, post treatment, and 1 hour
post treatment). Group series were graphically normalized to a post
loading value of 1, and a reduction toward 0 reflects a return toward
the baseline (normal) condition. There were significant effects for both
treatment condition and time (analysis of variance P < 0.001), and the
red LLLT condition significantly decreased from the postloading to the
1-hour post-treatment time points (P = 0.003). LLLT, low-level light
therapy.
exhibited lowered IPSV scores both immediately (M = −1.13,
SD = 0.85) and 1 hour post treatment (M = −2.06, SD = 1.43),
lowered PTP both immediately (M = −0.21 cm H2O, SD = 1.76)
and 1 hour post treatment (M = −1.38 cm H2O, SD = 0.47), in-
creased offset cycle 10 immediately post treatment (M = 0.07
ST, SD = 0.48) but lowered offset cycle 10 one hour post treat-
ment (M = −0.152 ST, SD = 0.44), and lowered onset cycle 1
immediately post treatment (M = −0.18 ST, SD = 0.23) but in-
creased onset cycle 1 one hour post treatment (M = 0.15 ST,
SD = 0.23). Heat group participants exhibited lowered
IPSV both immediately (M = −0.88, SD = 1.36) and 1 hour
post treatment (M = −2.19, SD = 1.70), lowered PTP
immediately (M = −0.07 cm H2O, SD = 0.64) and 1 hour post
treatment (M = −0.14, SD = 1.14), increased offset cycle 10 im-
mediately (M = 0.45 ST, SD = 0.47) and 1 hour post treatment
(M = 0.37 ST, SD = 0.59), and lowered onset cycle 1 immedi-
ately post treatment (M = −0.43 ST, SD = 0.42) but increased onset
cycle 1 one hour post treatment (M = 0.46 ST, SD = 0.65). Red
group participants exhibited lowered IPSV scores both
immediately (M = −1.81, SD = 1.30) and 1 hour post
treatment (M = −2.81; SD = 2.09), lowered PTP immediately
(M = −0.90 cm H2O, SD = 1.11) and 1 hour post treatment
(M = −1.69 cm H2O, SD = 1.00), lowered offset cycle 10 im-
mediately post treatment (M = −0.03 ST, SD = 0.39) and increased
offset cycle 10 one hour post treatment (M = 0.24 ST, SD = 0.27),
and increased onset cycle 1 immediately (M = 0.43 ST, SD = 0.42)
and 1 hour post treatment (M = 0.41 ST, SD = 0.78). IR group
 ARTICLE IN PRESS
6 Journal of Voice, Vol. ■■, No. ■■, 2016
FIGURE 3. RFF offset cycle 10 (A) and onset cycle 1 (B) measured in STs as a function of treatment group and time. N = 4 per group. Error
bars indicate standard error. IR, infrared; RFF, relative fundamental frequency; ST, semitone.
participants exhibited lowered IPSV both immediately (M = −1.5,
SD = 0.94) and 1 hour post treatment (M = −2.00, SD = 1.5),
lowered PTP immediately (M = −2.11 cm H2O; SD = 1.89) and
1 hour post treatment (M = −1.24 cm H2O, SD = 1.19), in-
creased offset cycle 10 immediately (M = 0.10 ST, SD = 0.35)
and 1 hour post treatment (M = 0.20 ST, SD = 0.42), and lowered
onset cycle 1 immediately (M = −0.50 ST, SD = 0.43) and 1 hour
post treatment (M = −0.42 ST, SD = 0.41).
FIGURE 4. Average RFF post treatment in STs as a function exper-
imental group and vocal cycle (offset 1–10 and onset 1–10). N = 4 per
group. Error bars are omitted for clarity. IR, infrared; RFF, relative fun-
damental frequency; ST, semitone.
Post hoc Tukey tests for user-defined contrasts analyzed change
post treatment and 1 hour post treatment to post loading for each
condition. The only significant comparison was that standard-
ized scores for the red condition significantly decreased post
loading to 1 hour post treatment (P = 0.003). No other com-
parisons for the other conditions were significant.
DISCUSSION
The goal of the present study was to examine the potential
effectiveness of red and IR LLLTs vs heat and a nonheated
(sham) control on reducing the impact of vocal fatigue created
by a vocal loading task. Vocal fatigue was assessed using
acoustic (RFF), aerodynamic (PTP), and self-perceptual mea-
sures (IPSV) immediately following vocal loading and 1 hour
post treatment.
Standardized scores of PTP, RFF, and IPSV were signifi-
cantly higher (disordered) post loading compared to baseline and
were significantly lower (normalized) 1 hour post treatment, in-
dicating that, when used in combination, PTP, RFF, and IPSV
are sensitive to a vocal loading task and the recovery trajectory
following vocal loading. ISPV increased following a loading task
and decreased an hour after loading had ended, which is con-
sistent with previous findings.26 In Hunter and Titze’s study,26
participants’ IPSV showed 50% recovery within 4–6 hours of
a 2-hour vocal loading task. Similarly, in this study, at 1 hour
post treatment, participants’ ISPV showed complete recovery,
and in some cases improvement, compared to baseline. Like-
wise, PTP increased following loading and eventually decreased
 ARTICLE IN PRESS
Loraine Sydney Kagan and James T. Heaton Effectiveness of LLLT in Vocal Fatigue 7

FIGURE 5. Phonation threshold pressure measured in centimeters of water as a function of the treatment group (N = 4 per group) and time (base-
line, post loading, post treatment, and 1 hour post treatment). Error bars indicate standard error. IR, infrared.

for all treatment groups, which is consistent with previous
studies.23,34,37
As described in the Methods section, offset cycle 10 and onset
cycle 1 of the RFF are theorized to be indicative of the changes
in tension of the laryngeal system during the laryngeal devoicing
gesture during the approach to and from a voiceless consonant.19
Increased rigidity due to fatigue-related hyperfunction might be
expected to lower the offset cycle closest to the voiceless con-
sonant due to the change in rigidity needed to devoice, and
likewise lower the first onset cycle closest to the voiceless con-
sonant due to the elevated baseline tension in the system.19
Additional studies have found that lowered onset cycle 1 is a
marker of acute vocal trauma.20 The findings from the present
study are consistent with an effect of vocal loading on RFF re-
sembling hyperfunction or acute vocal trauma. On average, offset
cycle 10 and onset cycle 1 both dropped in value following vocal
loading. Additionally, standardized values of RFF (in which signs
have been reversed), with PTP and IPSV, increased signifi-
cantly following loading, indicating that RFF, in combination
with PTP and IPSV, is sensitive to changes following a vocal
loading task.
Standardized values for the aggregated data set of RFF, PTP,
and IPSV were significantly lower 1 hour post treatment, indi-
cating that, on average, signs of vocal fatigue returned to baseline
or below baseline for all participants, regardless of condition.
However, when analyzing change post treatment and 1 hour post
treatment by condition, only participants in the red group ex-
hibited significantly lower values 1 hour post treatment compared
to post loading, indicating that the significant drop in the red
group may account for most of the improvement in the partici-
pants as a whole. No significant treatment effects were found
immediately post treatment, indicating that the mechanisms of
LLLT may work on a longer time frame than the 20-minute ap-
plication and possibly even beyond the 1-hour post treatment
mark. Future work should examine the potential usefulness of
LLLT for alleviating even longer-lasting symptoms of vocal
fatigue, better reflecting a real-world therapeutic application.
Red light might have been more effective than heat or IR
because of its activation of transcription factors (a property shared
with IR light, but not heat) and shorter wavelength than IR light.
As mentioned previously, this induction of transcription factors
causes downstream effects, including increased cell prolifera-
tion, growth factors, inflammatory modulators, and increased
tissue oxygenation.4 These downstream effects are not present
in heat treatment, which may partially explain the more bene-
ficial impact of red light. Additionally, although IR light penetrates
tissue more deeply than red light wavelengths, it is possible that
the fatigue induced by vocal loading was more efficiently re-
solved by addressing superficial structures more so than deeper
structures. It is also possible that either chosen light exposure
was insufficient, or perhaps even excessive, for achieving maximal
benefit toward relieving this mild form of phonotrauma.

FIGURE 6. Average task difficulty (rated on a scale from 1 to 10) for inability to produce soft voice as a function of the treatment group (N = 4
per group) and time (baseline, post loading, post treatment, and 1 hour post treatment). Error bars indicate standard error. IR, infrared.
 ARTICLE IN PRESS
8 Journal of Voice, Vol. ■■, No. ■■, 2016
LIMITATIONS AND DIRECTIONS FOR
FUTURE RESEARCH
Because of the need to maintain statistical power, all variables
were standardized and aggregated, which eliminated the ability
to analyze treatment effects on specific variables. Future studies
with greater sample sizes would enable a closer look at the in-
dividual impact of LLLT on each specific variable.
Additionally, the use of a more impactful fatiguing task, gen-
erating longer-lasting symptoms for the control groups, may have
revealed unequivocal LLLT treatment effects immediately post
treatment and 1 hour post treatment. While there is still some
debate as to whether or not the use of a vocal loading proce-
dure is truly representative of vocally fatiguing behavior, the
majority of research shows that vocal loading does impact acous-
tic, aerodynamic, and perceptual parameters in a manner
associated with vocal fatigue.23,26,37,39 Both duration and inten-
sity are also of concern when attempting to induce vocal fatigue
with a loading procedure. Loading procedures range in the lit-
erature from a few seconds40 to 14 hours,17 with the majority
between 1 and 2 hours.34,39 To balance the need to induce signs
and symptoms of vocal fatigue while maximizing participants’
short-term recovery, a loading procedure of only 1 hour (three
cycles of 15 minutes of reading followed by 5 minutes of rest)
was used in the present study. Further examination of the effects
of LLLT on alleviating vocal fatigue could use longer and louder
fatiguing tasks, or study individuals prone to vocal fatigue (eg,
those with glottic insufficiency or muscle tension dysphonia) as
a more clinically relevant test of this potential intervention.
Another possible reason for the lack of treatment effect for
both LLLT wavelengths is that the participants were allowed to
drink 6 oz of water because of the length of the experiment.
Results show that providing water might affect the loading task
impact on indicators of fatigue such as increases in PTP, but par-
ticipant comfort during this effortful hour of loud reading was
also a priority. Limiting participants’ consumption to 6 oz was
a compromise in design. Although previous studies have shown
that increased systemic hydration delays the onset of vocal fatigue,
allowing even this limited hydration might have mitigated the
effects on PTP on loading and treatment.23
In general, these results are promising for an initial study of
a new therapeutic technique. Red LLLT was shown to have a
significant treatment effect on RFF, PTP, and IPSV measures 1
hour following treatment. As has been shown in a review by
Huang et al,5 the dose of LLLT greatly impacts the treatment
effect elicited. Repeated applications have often been used in
pain reduction studies.15,41 A single 20-minute application was
chosen based on the recommendation of the user manual for the
device used to apply LLLT. A 20-minute dose was assumed to
be beneficial or at least not harmful based on the recommen-
dations from the device’s manufacturer. Red and IR LLLT in
doses of 10 minutes each has been applied in the MGH Voice
Center in individuals with muscle tension dysphonia, with seem-
ingly beneficial results per patient report. This anecdotal
experience suggests that future research should not only eval-
uate different light exposure periods but also examine the potential
benefit of combining wavelengths in a single treatment period.
Another previously mentioned OTC device, the Willow Curve,
produces both wavelengths simultaneously (unlike the New-U
used in the present study, which only produces one wave-
length at a time), further indicating a possible advantage of
combining light wavelengths.
Future studies are needed to identify optimal doses per wave-
length, delivered independently or combined, for alleviating
symptoms of vocal fatigue and other laryngeal dysfunctions. In
addition, future work should include not only acoustic, aerody-
namic, and self-perceptual measures of laryngeal function (as
reported here) but also examination of the effects of LLLT on
a cellular level through inflammatory markers associated with
vocal fatigue34,42 and other physiological measures of laryn-
geal hyperfunction including electromyography43,44 and vocal fold
kinematics.38
In addition to establishing an optimal laryngeal LLLT dose,
future work is need to determine when therapy should be applied
in relation to behaviors or events evoking vocal dysfunction. In
the present study, we chose to apply LLLT as a treatment fol-
lowing a fatiguing vocal task, rather than pre-expose participants
before the fatiguing task. In the LLLT literature, LLLT has been
studied both before and after a fatiguing task.12,13 Based on these
studies, pre-exposure may also be beneficial, and perhaps an ideal
approach will involve a combination of pre- and postphonotrauma
application.
CONCLUSIONS
The present study was the first to investigate the potential ef-
fectiveness of LLLT in treating vocal fatigue. Red and IR lights
were administered to the ventral neck surface for 20 minutes im-
mediately following a vocal loading task (1 hour of >75 dB
reading aloud) compared with the same duration of heat or sham
device application. Results indicate that red light may be effec-
tive in improving acoustic, aerodynamic, and self-perceptual
markers of vocal fatigue. Although these findings indicate that
LLLT is a promising treatment for vocal fatigue, future work
is needed to determine optimal light doses, whether wave-
lengths are more efficacious in combination vs isolation, and when
the doses should be applied relative to phonotrauma (before, after,
or both) for providing fatigue resistance or hastening recovery.
LLLT is a simple, potentially efficacious treatment for vocal
fatigue that could have wide clinical relevance to populations
with voice disorders or occupations with high voice use demands.
Acknowledgments
We would like to thank Tara Stadelman-Cohen, CCC-SLP; Carol
Krusemark, CCC-SLP; and Jason Muise, CCC-SLP, for per-
forming laryngeal endoscopy on study participants. We would
also like to thank Eric Stone, MS; Jarrad Van Stan, CCC-SLP;
Stephanie Lien, PhD; and Amanda Fryd, CCC-SLP, for their as-
sistance with laboratory equipment, measurements, and analysis.
This project was supported by a grant from the Christopher
Norman Education Fund at the MGH Institute of Health
Professions.
REFERENCES
1. Chung H, Dai T, Sharma SK, et al. The nuts and bolts of low-level laser
(light) therapy. Ann Biomed Eng. 2012;40:516–533.
 ARTICLE IN PRESS
Loraine Sydney Kagan and James T. Heaton Effectiveness of LLLT in Vocal Fatigue
2. Mester E, Szende B, Tota J. Effect of laser on hair growth of mice. Kiserl
Orvostud. 1967;19:628–631.
3. Karu T. Photobiological fundamentals of low-power laser therapy. IEEE J
Quantum Electron. 1987;23:1703–1717.
4. Karu T, Pyatibrat L, Kalendo G. Irradiation with HeNe laser increases ATP
level in cells cultivated in vitro. J Photochem Photobiol B. 1995;27:219–223.
5. Huang Y-Y, Chen AC-H, Carroll JD, et al. Biphasic dose response in low
level light therapy. Dose Response. 2009;7:358–383. doi:10.2203/dose-
response.09-027.Hamblin.
6. Karu T, Kolyakov S. Exact action spectra for cellular responses relevant
to phototherapy. Photomed Laser Surg. 2005;23:355–361.
7. Meyers AD. Lasers and wound healing. Arch Otolaryngol Head Neck Surg.
1990;116:1128.
8. Bisht D, Mehrotra R, Singh P, et al. Effect of helium-neon laser on wound
healing. Indian J Exp Biol. 1999;37:187–189.
9. Hawkins D, Houreld N, Abrahamse H. Low level laser therapy (LLLT) as
an effective therapeutic modality for delayed wound healing. Ann N Y Acad
Sci. 2005;1056:486–493.
10. Noble PB, Shields ED, Blecher PD, et al. Locomotory characteristics of
fibroblasts within a three-dimensional collagen lattice: modulation by a
helium/neon soft laser. Lasers Surg Med. 1992;12:669–674.
11. Salate AC, Barbosa G, Gaspar P, et al. Effect of In-Ga-Al-P diode laser
irradiation on angiogenesis in partial ruptures of Achilles tendon in rats.
Photomed Laser Surg. 2005;23:470–475.
12. De Marchi T, Junior ECPL, Bortoli C, et al. Low-level laser therapy (LLLT)
in human progressive-intensity running: effects on exercise performance,
skeletal muscle status, and oxidative stress. Lasers Med Sci. 2012;27:231–
236.
13. de Brito Vieira WH, Ferraresi C, de Andrade Perez SE, et al. Effects of
low-level laser therapy (808 nm) on isokinetic muscle performance of young
women submitted to endurance training: a randomized controlled clinical
trial. Lasers Med Sci. 2012;27:497–504.
14. Leal ECP, Lopes-Martins RÁB, Frigo L, et al. Effects of low-level laser
therapy (LLLT) in the development of exercise-induced skeletal muscle
fatigue and changes in biochemical markers related to postexercise recovery.
J Orthop Sports Phys Ther. 2010;40:524–532.
15. Bjordal JM, Lopes-Martins RA, Joensen J, et al. A systematic review with
procedural assessments and meta-analysis of low level laser therapy in lateral
elbow tendinopathy (tennis elbow). BMC Musculoskelet Disord. 2008;9:1.
16. Hashmi JT, Huang Y-Y, Osmani BZ, et al. Role of low-level laser therapy
in neurorehabilitation. PM R. 2010;2:S292–S305.
17. Boucher VJ, Ahmarani C, Ayad T. Physiologic features of vocal fatigue:
electromyographic spectral-compression in laryngeal muscles. Laryngoscope.
2006;116:959–965.
18. Claassen H, Werner JA. Fiber differentiation of the human laryngeal muscles
using the inhibition reactivation myofibrillar ATPase technique. Anat Embryol
(Berl). 1992;186:341–346.
19. Stepp CE, Hillman RE, Heaton JT. The impact of vocal hyperfunction on
relative fundamental frequency during voicing offset and onset. J Speech
Lang Hear Res. 2010;53:1220–1226.
20. Murray ESH, Hands GL, Calabrese CR, et al. Effects of adventitious acute
vocal trauma: relative fundamental frequency and listener perception. J Voice.
2016;30:177–185.
21. Cooper DS, Titze IR. Generation and dissipation of heat in vocal fold tissue.
J Speech Lang Hear Res. 1985;28:207–215.
22. Titze IR, Martin DW. Principles of voice production. J Acoust Soc Am.
1998;104:1148.
23. Solomon NP, DiMattia MS. Effects of a vocally fatiguing task and systemic
hydration on phonation threshold pressure. J Voice. 2000;14:341–362.
View publication stats
9
24. Milbrath RL, Solomon NP. Do vocal warm-up exercises alleviate vocal
fatigue? J Speech Lang Hear Res. 2003;46:422–436.
25. Bastian RW, Keidar A, Verdolini-Marston K. Simple vocal tasks for detecting
vocal fold swelling. J Voice. 1990;4:172–183.
26. Hunter EJ, Titze IR. Quantifying vocal fatigue recovery: dynamic vocal
recovery trajectories after a vocal loading exercise. Ann Otol Rhinol Laryngol.
2009;118:449.
27. Marinho RR, Matos RM, Santos JS, et al. Potentiated anti-inflammatory
effect of combined 780 nm and 660 nm low level laser therapy on the
experimental laryngitis. J Photochem Photobiol B. 2013;121:86–93.
28. Stemple JC, Glaze LE, Gerdeman BK. Clinical Voice Pathology: Theory
and Management. Cengage Learning; 2000.
29. Kitch JA, Oates J. The perceptual features of vocal fatigue as self-reported
by a group of actors and singers. J Voice. 1994;8:207–214.
30. Gotaas C, Starr CD. Vocal fatigue among teachers. Folia Phoniatr Logop.
1993;45:120–129.
31. Casper JK, Leonard R. Understanding Voice Problems: A Physiological
Perspective for Diagnosis and Treatment. Lippincott Williams & Wilkins;
2006.
32. Solomon NP. Vocal fatigue and its relation to vocal hyperfunction. Int J
Speech Lang Pathol. 2008;10:254–266.
33. Ingle JW, Helou LB, Li NY, et al. Role of steroids in acute phonotrauma:
a basic science investigation. Laryngoscope. 2014;124:921–927.
34. Verdolini Abbott K, Li NY, Branski RC, et al. Vocal exercise may attenuate
acute vocal fold inflammation. J Voice. 2012;26:814, e1-e13.
35. Russell B, Kellett N, Reilly L. A study to determine the efficacy of
combination LED light therapy (633 nm and 830 nm) in facial skin
rejuvenation. J Cosmet Laser Ther. 2005;7:196–200.
36. Lien Y-AS, Stepp CE. Comparison of voice relative fundamental frequency
estimates derived from an accelerometer signal and low-pass filtered and
unprocessed microphone signals. J Acoust Soc Am. 2014;135:2977–
2985.
37. Enflo L, Sundberg J, McAllister A. Collision and phonation threshold
pressures before and after loud, prolonged vocalization in trained and
untrained voices. J Voice. 2013;27:527–530.
38. Stepp CE, Hillman RE, Heaton JT. A virtual trajectory model predicts
differences in vocal fold kinematics in individuals with vocal hyperfunction.
J Acoust Soc Am. 2010;127:3166–3176.
39. Remacle A, Finck C, Roche A, et al. Vocal impact of a prolonged reading
task at two intensity levels: objective measurements and subjective self-
ratings. J Voice. 2012;26:e177–e186.
40. Södersten M, Ternström S, Bohman M. Loud speech in realistic
environmental noise: phonetogram data, perceptual voice quality, subjective
ratings, and gender differences in healthy speakers. J Voice. 2005;19:29–
46.
41. Chow RT, Johnson MI, Lopes-Martins RA, et al. Efficacy of low-level laser
therapy in the management of neck pain: a systematic review and meta-
analysis of randomised placebo or active-treatment controlled trials. Lancet.
2009;374:1897–1908.
42. Verdolini K, Branski RC, Rosen CA, et al. Shifts in biochemical markers
associated with wound healing in laryngeal secretions following
phonotrauma: a preliminary study. Ann Otol Rhinol Laryngol.
2003;112:1021–1025.
43. Stepp CE, Hillman RE, Heaton JT. Use of neck strap muscle intermuscular
coherence as an indicator of vocal hyperfunction. IEEE Trans Neural Syst
Rehabil Eng. 2010;18:329–335.
44. Stepp CE, Heaton JT, Braden MN, et al. Comparison of neck tension
palpation rating systems with surface electromyographic and acoustic
measures in vocal hyperfunction. J Voice. 2011;25:67–75.