Light level and duration of exposure determine the impact of self-luminous tablets on melatonin suppression

Abstract

Exposure to light from self-luminous displays may be linked to increased risk for sleep disorders because these devices emit optical radiation at short wavelengths, close to the peak sensitivity of melatonin suppression. Thirteen participants experienced three experimental conditions in a within-subjects design to investigate the impact of self-luminous tablet displays on nocturnal melatonin suppression: 1) tablets-only set to the highest brightness, 2) tablets viewed through clear-lens goggles equipped with blue light-emitting diodes that provided 40 lux of 470-nm light at the cornea, and 3) tablets viewed through orange-tinted glasses (dark control; optical radiation <525 nm ≈ 0). Melatonin suppressions after 1-h and 2-h exposures to tablets viewed with the blue light were significantly greater than zero. Suppression levels after 1-h exposure to the tablets-only were not statistically different than zero; however, this difference reached significance after 2 h. Based on these results, display manufacturers can determine how their products will affect melatonin levels and use model predictions to tune the spectral power distribution of self-luminous devices to increase or to decrease stimulation to the circadian system.

Full text

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Author's personal copy
Light level and duration of exposure determine the impact of self-luminous
tablets on melatonin suppression
Brittany Wood, Mark S. Rea, Barbara Plitnick, Mariana G. Figueiro
*
Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union Street, Troy, NY 12180, USA
article info
Article history:
Received 1 February 2012
Accepted 12 July 2012
Keywords:
Melatonin
Electronic displays
Tablets
abstract
Exposure to light from self-luminous displays may be linked to increased risk for sleep disorders because
these devices emit optical radiation at short wavelengths, close to the peak sensitivity of melatonin
suppression. Thirteen participants experienced three experimental conditions in a within-subjects design
to investigate the impact of self-luminous tablet displays on nocturnal melatonin suppression: 1) tablets-
only set to the highest brightness, 2) tablets viewed through clear-lens goggles equipped with blue light-
emitting diodes that provided 40 lux of 470-nm light at the cornea, and 3) tablets viewed through
orange-tinted glasses (dark control; optical radiation <525 nm z0). Melatonin suppressions af ter 1-h and
2-h exposures to tablets viewed with the blue light were significantlygreater than zero. Suppression levels
after 1-h exposure to the tablets-only were not statistically different than zero; however, this difference
reached significance after 2 h. Based on these results, display manufacturers can determine how their
products will affect melatonin levels and use model predictions to tune the spectral power distribution of
self-luminous devices to increase or to decrease stimulation to the circadian system.
Ó2012 Elsevier Ltd and The Ergonomics Society. All rights reserved.
1. Background
Melatonin is a hormone produced by the pineal gland at night
and under conditions of darkness in both diurnal and nocturnal
species. It is a timing messenger, signaling nighttime information
throughout the body. Exposure to light at night can retard or even
cease nocturnal melatonin production. Short-wavelength light is
maximally effective at suppressing melatonin (peak
sensitivity z460 nm). Suppression of melatonin by light at night
has been implicated in disruption of sleep, increased risk for
obesity, as well as increased risk for more serious diseases, such as
breast cancer (Blask et al., 1999).
Technological developments have led to bigger and brighter
self-luminous electronic devices, such as televisions, computer
screens, and cell phones. Some have suggested that light at night
from electronic devices can suppress nocturnal melatonin (Figueiro
et al., 2011; Cajochen et al., 2011), which may disrupt sleep or pose
a health risk. To produce white light, these electronic devices must
emit light at short wavelengths, which makes them potential
sources for suppressing melatonin at night or for delaying the onset
of melatonin in the evening, thereby possibly reducing sleep
duration and disrupting sleep. This is particularly worrisome in
populations such a young adults and adolescents, who already tend
to be more “night owls”.
For example, Cajochen et al. (2011) showed that a 5-h exposure to
a (white) light-emitting diode (LED) backlit computer screen signifi-
cantly suppressed melatonin and enhanced performance compared
to a non-LED backlit screen. Their results showed that although
melatoninlevels were still rising over thecourse of the night, they did
not rise as steeply as when subjects remained in darkness.
Using a similar protocol as the one employed inthe present study,
Figueiroet al. (2011) showed thata 2-h exposureto light from cathode
ray tube computer screens induced a slight, but not statistically
significantreduction in melatonin concentrations in college students.
The present study extends the findings from this previous study by
investigating the impact of self-luminous tablets on melatonin
suppression. In order tosimulate typicalusage of these devices(Apple
iPads), participants were allowed to choose their preferred tasks on
the tablets (e.g., games, on-line shopping, reading, etc.). The
Dimesimeter (Figueiro et al., in press), a circadian light meter devel-
oped to measure photopic and circadian light, was used to record
personal light exposures during the experiment. These data were
used in conjunction with the model of human circadian photo-
transduction proposed by Rea and colleagues to calculate photopic
illuminance (lux), circadian light (CL
A
), and circadian stimulus (CS)
levels (Rea et al., 2005,2010,2011). Comparisons of actual and pre-
dicted melatonin suppression were performed.
*Corresponding author. Tel.: þ1 518 687 7100; fax: þ1 518 687 7120.
E-mail addresses: woodb5@rpi.edu (B. Wood), ream@rpi.edu (M.S. Rea), plitnb@
rpi.edu (B. Plitnick), figuem@rpi.edu (M.G. Figueiro).
Contents lists available at SciVerse ScienceDirect
Applied Ergonomics
journal homepage: www.elsevier.com/locate/apergo
0003-6870/$ esee front matter Ó2012 Elsevier Ltd and The Ergonomics Society. All rights reserved.
http://dx.doi.org/10.1016/j.apergo.2012.07.008
Applied Ergonomics 44 (2013) 237e240

Author's personal copy
2. Methods
2.1. Subjects
Thirteen subjects were recruited to participate in the study bye-
mail, web posting, and word-of-mouth. The mean standard
deviation (SD) age of the subjects was 18.9 5.2 years. Individuals
were excluded from participation if they smoked or had a major
health problem such as heart disease, diabetes, and high blood
pressure. Individuals were also excluded if they were taking over-
the-counter melatonin or prescription medication such as blood
pressure medication, antidepressants, beta-blockers, or sleep
medicine. Women taking oral contraception were allowed to
participate. To ensure they were not extreme early or extreme late
types, potential subjects were asked to complete a Munich Chro-
notype Questionnaire (MCTQ) (Roenneberg et al., 2003). This
screening method helped ensure that the subjects would produce
melatonin between the hours of 23:00 and 01:00, the period of data
collection. The mean SD MCTQ score of the subjects was 3.5 1.3.
2.2. Lighting conditions
During the experiment, the three lighting conditions were
provided in the same test room. In one condition, subjects viewed
their tablets through a pair of clear goggles fitted with short-
wavelength (blue) LEDs. The goggles were calibrated to deliver
40 lux (40
m
W/cm
2
) of blue light (
l
max
z470-nm) at the cornea of
each participant (Fig. 1). Two LEDs were mounted to each lens; one
was located above and one was located below a line of sight
through the center of each goggle lens. To minimize discomfort
glare and blue-light hazard (Bullough, 2000), polycarbonate
translucent tape diffused the light emitted by the LEDs. Before each
experimental session, the goggles were calibrated using an optical
fiber with a Lambertian diffuser on one end. The irradiances of the
left and right lens were measured independently and, if necessary,
the voltage from a 9 V battery was adjusted so that a mean illu-
minance of 40 lux at the corneas was achieved. This light level was
selected because in previous studies, it has been shown to deliver
a light stimulus that is predicted to be above threshold and below
saturation (Rea et al., 2005). Based upon these previous findings,
the tablet with the blue LEDs condition served as a “true-positive”
condition for the present experiment.
The second experimental condition involved viewing the tablets
through orange-tinted glasses. These glasses (SAF-T-CURE
Ò
Orange
UV Filter, Chicago IL) filtered out optical radiation below 525 nm
(Fig.1). The tablet with the orange-tinted glasses served as the “dark”
control condition because the short-wavelength radiation capable of
suppressing melatonin was filtered out (Rea et al., 2005,2011).
The third condition was the tablet-only experimental condition;
the relative spectral power distribution (SPD) of an exemplar tablet
is shown in Fig. 1 as measured with an Ocean Optic USB 650
spectroradiometer (Dunedin FL). During all three conditions, the
tablets were set to full brightness. The only additional lighting in
the room was from two red lights located behind the subjects (<1
lux at the cornea). No measurable stray light from the blue-light
goggles reached the other subjects’eyes. All the participants were
seated facing the same direction and the subjects wearing the
goggles with the blue LEDs sat in front of all the other participants.
2.2.1. Tablet displays
Each participant used an Apple iPad; two subjects used the first
generation (iPad 1) and the others used the second generation (iPad
2). The display of the iPad is approximately 9.7 inches diagonally and
is backlit with LEDs. The screen resolution is 1024 728 pixels at 132
pixels per inch (Apple Inc., 2012). Each iPad was viewed while set to
full brightness. Measurements from a calibrated illuminance meter
(Gigahertz-Optik, X91, Turkenfeld, Germany) obtained prior to the
experiment revealed that, with an all-white background at a 10-inch
viewing distance, the iPads could deliver 40 lux at the corneas.
2.2.2. Dimesimeter measurements
In order to accurately record personal lightexposures during the
experiment, each subject wore a Dimesimeter close to the plane of
the cornea. The Dimesimeter is a small (w2 cm diameter)
measurement instrument that continuously records light [via red
(R), green (G), and blue (B) sensors] and activity levels (Figueiro
et al., in press). A Dimesimeter was worn on a headband by
subjects during the tablet-only condition. Similarly, it was clipped
to the temple of the goggles for the condition where the blue LEDs
were energized. The Dimesimeter was mounted behind the bridge
of the orange-tinted glasses, just above the nosepiece, for the third,
control condition. During the experiment, data were logged every
30 s from 23:00 to 01:00. The RGB light data from the Dimesimeters
were downloaded and stored on a computer following each
experimental session. After downloading, the raw measurements
from the RGB sensors were converted into photopic illuminance
(lux), circadian light (CL
A
), and circadian stimulus (CS) levels. (Rea
et al., 2005,2010,2011). Concisely, photopic illuminance is irradi-
ance transformed by the photopic luminous efficiency function
(V(
l
)), providing an orthodox measure of the spectral sensitivity of
the human fovea. CL
A
, based on nocturnal melatonin suppression, is
irradiance weighted by the spectral sensitivity of the retinal pho-
totransduction mechanisms stimulating the biological clock. CS is
a transform of CL
A
into relative units from 0, the threshold for
circadian system activation, to 0.7, response saturation, and corre-
sponds to relative suppression of nocturnal melatonin after one
hour of light exposure for a 2.3 mm diameter pupil during the
midpoint of melatonin production.
2.2.3. Protocol
All thirteen participants experienced the three experimental
sessions, each one week apart. They were asked to maintain
a regular sleep schedule for the week prior to each session,
requiring them to go to bed no later than 23:00 and wake up no
later than 07:30. In order to verify compliance with the sleep
schedule, all thirteen subjects were asked to keep sleep logs the
week of the experiment. Seven of the subjects were also asked to
call into the lab at 07:30 and 08:30 each day to ensure they were
awake. Six of the subjects attend high school at the same time each
day, so they were not expected to call in. On the day of the
Fig. 1. Spectral transmittance of the orange-tinted glasses, the relative spectral power
distribution (SPD) of a 470-nm (blue) LED, and the relative SPD of an iPad 1 (white
screen, full brightness) used in the experiment.
B. Wood et al. / Applied Ergonomics 44 (2013) 237e240238

Author's personal copy
experiment, all subjects were asked to refrain from napping and
consuming caffeinated products. In order to counterbalance the
lighting conditions, subjects were randomly assigned to one of
three groups. During the experiment, all three groups were seated
in the same room and by the end of the three weeks all subjects had
been exposed to all three lighting conditions.
On the day of the experiment (Friday nights), subjects arrived at
the laboratory at 22:30 and spent the first 30 min in dim red light
(less than 1 lux at the cornea from two red LED traffic lights,
l
max
¼630 nm). At 23:00, the first saliva sample was collected
while the subjects remained in the dim light. Saliva samples were
collected using the salivette system (SciMart, Saint Louis MO). To
provide a saliva sample, subjects were asked to remove a cotton
cylinder from a plastic test-tube and chew it until saturated. When
saturated, the subjects were asked to return the cotton to the tube
without touching it with their hands. The experimenter then
collected the sample and centrifuged it for 5 min at 3000 g to
remove the saliva from the cotton. The cotton was discarded and
saliva immediately frozen (20

C).
After the subjects were done with their first saliva sample
collection, they were asked to turn on their tablets and start using
them. From 23:00 to 01:00 the subjects were allowed to engage in
whatever task they chose on their tablets; viewing position and
distance were not controlled. Saliva samples were collected after
1 h (at 00:00) and 2 h (at 01:00) of tablet use.
3. Data Analyses
Saliva samples were later assayed by radioimmunoassay using
a commercially available kit from Labor Diagnostika Nord (Nord-
horn, Germany). The limit of detection was 0.9 pg/mL and the intra-
and inter-assay coefficients of variability were determined to be
11.4 and 12.7%, respectively.
Adjusted dark values (A) were calculated for each subject to
account for their natural rise in melatonin level concentrations (C)
while in the dark and for differences in the initial melatonin
concentrations week-to-week. Melatonin concentrations obtained
from each subject during the tablet with the orange-tinted glasses
condition (D) were used for the adjusted dark values. Melatonin
suppression was calculated for both lighting conditions(L
1
¼tablet-
only and L
2
¼tablet with blue LEDs) using the adjusted dark value at
the same sampling time (T
n
). The adjusted dark value (A) for a given
lighting condition (L
m
) and sampling time (T
n
) is given by:
A¼C
T1;Lm
=C
T1;D
C
Tn;D
(1)
Where: C¼melatonin concentration (pg/ml)
T¼sampling time, n¼1,2,3
1¼23:00; 2 ¼00:00 and 3 ¼01:00
L¼lighting condition, m¼1,2
1¼tablet-only; 2 ¼tablet with blue LEDs
D¼tablet with orange-tinted glasses
Melatonin suppressionS¼1C
Tn;Lm
=A(2)
4. Results
4.1. Light measurements
Table 1 shows the mean standard error of the means (SEM)
light measurement, CL
A
, and CS values from the Dimesimeter. In
addition, luminance measurements were made during the
experiment. Luminance values from the subjects’devices ranged
from 1.4 to 184 cd/m
2
. Mean [median] SEM luminance values
were 77 [73] 66 cd/m
2
.
Based on the CL
A
measurements with the Dimesimeter and
assuming a reference pupil diameter of 2.3 mm, the calculated
mean SEM CS values after 1-h exposures were 0.46 0.0013 for
the tablet with blue LEDs condition and 0.03 0.0066 for the
tablet-only condition. These CS values translate into a predicted
suppression of approximately 46 and 3% for the tablet with the blue
LEDs and the tablet only conditions, respectively. No predictions
were made for the 2-h exposures data because the model by Rea
and colleagues assumes a 1-h exposure (Rea et al., 2005). It had
been assumed that the tablet with the orange-tinted glasses would
not produce exposures necessary for suppression. Indeed, using the
measured CL
A
values and a 2.3 mm diameter pupil, the calculated
mean SEM CS value was 0.0017 0.0004.
4.2. Melatonin
Although thirteen subjects completed the experiment, two
subjects did not provide a sufficient quantity of saliva for assay at
the 00:00 sampling time and one subject did not provide a sample
at 01:00. Thus, two subjects were omitted from the 1-h analysis
(n¼11) and one subject was omitted from the 2-h analysis (n¼12).
Table 1 and Fig. 2 show the mean SEM suppression values for
the tablet-only and for the tablet with blue LEDs conditions at
00:00 and at 01:00. Two-tailed, One-Sample t-tests were used to
determine statistical reliability. For the tablet with blue LEDs
condition, suppression values were significantly different than zero
at 00:00 (t(10) ¼15.0, p<0.001) and at 01:00 (t(11) ¼16.1,
p<0.001). The mean SEM suppression after 1-hr exposure to the
tablet with blue LEDs condition was 48% 4%, very close to the
predicted values, which was 46%. For the tablet-only condition,
suppression was not significantly different than zero after 1-h
exposure (t(10) ¼1.80, p¼0.103) to the tablet, but was signifi-
cantly greater than zero after 2 h of exposure (t(11) ¼3.39,
p¼0.006). Suppression after 1-h exposure to the tablet only
condition was 7% 4%, again, very close to the model predictions,
which was 3%.
5. Discussions
The present study extends results from Figueiro et al. (2011)
showing that a 2-h exposure to self-luminous tablets can result in
a measurable, statistically reliable suppression of melatonin in
Table 1
Lighting conditions (photopic illuminance in lux and CL
A
measured with the
Dimesimeter), predicted melatonin suppression (CS) and measured melatonin
suppression after 1-h and 2-h exposures. Mean standard error of the mean (SEM)
values are shown.
Photopic
illuminance
(lux)
CL
A
CS
b
Measured
suppression
(%)
1 h Tablet þblue LEDs 59 5.0 648 4.9 0.46 0.0013 48 4
Tablet þorange-
tinted glasses
a
9.8 1.9 1.5 0.31 0.0017 0.0004 NA
Tablet-only 18 3. 8 19 4.6 0.03 0.0066 7.0 4
2 h Tablet þblue LEDs 57 3.8 645 3.4 NA 66 4
Tablet þorange-
tinted glasses
a
9.9 1.6 1.5 0.29 NA NA
Tablet-only 16 2.7 17 3.51 NA 23 6
NA: not applicable.
a
The tablet with the orange-tinted glasses condition was used as the dark control.
b
Based upon a 1-hr duration of light exposure and a 2.3 mm pupil diameter.
B. Wood et al. / Applied Ergonomics 44 (2013) 237e240 239

Author's personal copy
a population of young adults. It should be pointed out that, as
predicted by the model of human circadian phototransduction, 1-h
exposures to the tablet-only condition at full brightness resulted in
a level of melatonin suppression close to that which was predicted
(i.e., the CS values); however, this level of suppression was not
statistically different than zero. After 2 h of exposure to the tablet-
only condition, melatonin suppression was statistically different
than zero. Therefore, it is important to have quantitative estimates
of both the SPD (level and spectrum) and the duration of exposure
before drawing reliable inferences about the ability of self-
luminous tablets to induce nocturnal melatonin suppression.
Moreover, the type of task being performed on the tablets will also
determine how much light the self-luminous devices are delivering
at the cornea and, therefore, its impact on evening melatonin levels.
As shown by our Dimesimeter measurements, the range of phot-
opic illuminance levels at the cornea from the tablets alone varied
from 5 lux, which is likely not affecting melatonin levels, to over
50 lux, which as shown in the present study, will result in
measurable melatonin suppression after a 2-h exposure.
Since the predictions of acute melatonin suppression based
upon the measured circadian light levels and the model by Rea and
colleagues (Rea et al., 2005,2010,2011) were very close to the
observed suppression levels after 1 h, manufacturers may now be
able to design self-luminous display screens that can either
increase (e.g., desirable during morning hours) or decrease (e.g.,
desirable during evening hours) circadian stimulation. Further, it
might be possible to develop software to control circadian light
exposures based upon the SPD of the display together with the time
and the hours of operation, with the purpose of limiting melatonin
suppression in the evening. The present results may be a positive
step toward the development of more “circadian-friendly”elec-
tronic devices. Even if new technologies (e.g., organic light emitting
diodes, OLEDs) are used in the development of self-luminous
electronic devices, these results are still relevant because we
showed that the model is useful in predicting the effectiveness of
these devices on melatonin suppression after 1-h exposure. In
other words, as long as the spectral irradiance distribution at the
cornea from a self-luminous technology is known, one can predict
its impact on melatonin suppression after 1 h of viewing (Rea et al.,
2010). Since a large portion of the population spends most of their
waking hours in front of a self-luminous display, it is important that
manufactures and users have a tool to increase or to decrease
circadian stimulation delivered by their self-luminous displays.
However, it is also important to consider how and how long
these devices are used. Large self-luminous displays (e.g., flat-
screen televisions) or ones that are operated close to the eyes
(e.g., cell phones) would be expected to provide relatively high
circadian stimulation. For example, Shieh and Lee (2007) showed
that the preferred viewing distance for E-paper was 500 mm,
which is similar to the preferred viewing distance for visual display
terminals. These distances are certainly much closer than those of
television viewing. Currently, the model of human circadian pho-
totransduction by Rea and colleagues only takes into account the
absolute SPD of the light source. Variables such as duration of
exposure and spatial distribution also need to be added to the
model to improve its ability to predict the impact of self-luminous
displays on acute melatonin suppression.
Finally, it is important to acknowledge that usage of self-
luminous electronic devices before sleep may disrupt sleep even
if melatonin is not suppressed. Clearly, the tasks themselves may be
alerting or stressful stimuli that can lead to sleep disruption. For
now, however, it is recommended that these devices be dimmed at
night as much as possible in order to minimize melatonin
suppression, and that the duration of use be limited prior to
bedtimes.
Acknowledgments
The authors would like to acknowledge Sharp Laboratories of
America for providing funding for the study. Gary Feather, Xiao-Fan
Feng, and Ibrahim Sezan are acknowledged for their support.
Robert Hamner, Aaron Smith, Anna Lok, Howard Ohlhous, Dennis
Guyon and Ines Martinovic are acknowledged for their technical
and editorial support.
References
Apple Inc., 2012. iPad: Technical Specifications. Available from: http://www.apple.
com/ipad/specs/ (accessed 06.01.12.).
Blask, D., Sauer, L., Dauchy, R., Holowachuk, E., Ruhoff, M., 1999. New insights into
melatonin regulation of cancer growth. Advances in Experimental Medicine
and Biology 460, 337e343.
Bullough, J.D., 2000. The blue-light hazard: a review. Journal of the Illuminating
Engineering Society 29, 6e14.
Cajochen, C., Frey, S., Anders, D., Späti, J., Bues, M., Pross, A., Mager, R., Wirz-
Justice, A., Stefani, O., 2011. Evening exposure to a light-emitting diodes (LED)-
backlit computer screen affects circadian physiology and cognitive perfor-
mance. Journal of Applied Physiology 110, 1432e1438.
Figueiro, M.G., Hamner, R., Bierman, A., Rea, M.S. Comparisons of three practical
field devices used to measure personal light exposures and activity levels,
Lighting Research and Technology, in press.
Figueiro, M.G., Plitnick, B., Wood, B., Rea, M.S., 2011. The impact of light from
computer monitors on melatonin levels in college students. Neuro Endocri-
nology Letters 32, 158e163.
Rea, M.S., Figueiro, M.G., Bierman, A., Bullough, J.D., 2010. Circadian light. Journal of
Circadian Rhythms 8, 2.
Rea, M.S., Figueiro, M.G., Bierman, A., Hamner, R. Modeling the spectral sensitivity
of the human circadian system, Lighting Research and Technology, 1–12
(published online 14 December 2011).
Rea, M.S., Figueiro, M.G., Bullough, J.D., Bierman, A., 2005. A model of photo-
transduction by the human circadian system. Brain Research Reviews 50,
213e228.
Roenneberg, T., Wirz-Justice, A., Merrow, M., 2003. Life between clocks: daily
temporal patterns of human chronotypes. Journal of Biological Rhythms 18,
80e90.
Shieh, K.-K., Lee, D.-S., 2007. Preferred viewing distance and screen angle of elec-
tronic paper displays. Applied Ergonomics 38, 601e608.
Fig. 2. Mean SEM suppression values for the tablet-only and for the tablet with blue
LEDs at 00:00 and at 01:00. Suppression after 1-h exposure to the tablet-only condi-
tion was not significantly different than zero; all other suppression values (marked
with asterisks) were significantly greater than zero (p<0.05). The orange-tinted
glasses condition served as the “dark”control condition.
B. Wood et al. / Applied Ergonomics 44 (2013) 237e240240