Mark S Rea*, Mariana G Figueiro, Andrew Bierman, John D Bullough
The present paper reflects a work in progress toward a definition of circadian light, one that should be informed
by the thoughtful, century-old evolution of our present definition of light as a stimulus for the human visual sys-
tem. This work in progress is based upon the functional relationship between optical radiation and its effects on
nocturnal melatonin suppression, in large part because the basic data are available in the literature. Discussed here
are the fundamental differences between responses by the visual and circadian systems to optical radiation. Brief
reviews of photometry, colorimetry, and brightness perception are presented as a foundation for the discussion of
circadian light. Finally, circadian light (CLA) and circadian stimulus (CS) calculation procedures based on a published
mathematical model of human circadian phototransduction are presented with an example.
The suprachiasmatic nuclei (SCN) in the hypothalamus
host the master circadian clock that organizes and
orchestrates the timing of all daily biological functions,
from complicated physiological systems to single cells.
The SCN in humans have, on average, an intrinsic per-
iod slightly greater than 24 hours  that is modulated
by the temporal pattern of light and dark on the retina.
As a result of the earth’s rotation on its axis, the tem-
poral pattern of light and dark on the retina synchro-
nizes the SCN to a matching 24-h period. Recent
research has demonstrated that disruption of the nat-
ural, 24-h pattern of light and dark from rapid flight
across time zones or from rotating shift work can lead
to a wide variety of maladies, from poor performance to
sleep loss, weight gain, and even breast cancer [2-9].
Because it is increasingly evident that retinal light and
dark exposures can profoundly affect human health and
well-being, it is increasingly important to be able to
quantify both light and dark as stimuli to the human
The present paper deals with the evolving definition of
circadian light. Technically, the adjective circadian must
be used to modify the noun light because light is
defined specifically in terms of optical radiation capable
of producing a visual sensation in humans [10,11].
Strictly speaking then, light cannot be used synony-
mously with optical radiation capable of producing a
non-visual, circadian response in humans or with optical
radiation producing a visual response in another species.
Nevertheless, in the vernacular, light is used as a term
to describe optical radiation with a spectral power distri-
bution anywhere within the “visible region” of the elec-
tromagnetic spectrum (approximately 380 nm to
730 nm), irrespective of its biological consequences.
Moreover, the term light is always used, with or without
strict regard for its ability to stimulate human vision, as
a noun to describe the stimulus to rather than the
response from a biological system. This is an important
point because light is circularly defined; light as a stimu-
lus to the human visual system was derived from
responses by the human visual system. Thus, any formal
definition of circadian light as a stimulus to the circa-
dian system must also be dependent on the measured
response from the circadian system. Fundamentally
then, it is necessary to be able to measure a reliable
response of the human circadian system to optical radia-
tion incident on the retina to define the stimulus to the
human circadian system. This inherent, and potentially
confusing, circularity always must be considered as a
formal definition of circadian light develops.
Notwithstanding this potentially confusing circularity,
it will be difficult to develop a definition of light for the
circadian system that is strictly homologous with the
formal definition of light for the visual system because,
for reasons discussed in this paper, the responses by
these two systems to optical radiation on the retina are
fundamentally different. The biophysical mechanisms
underlying phototransduction for the two systems are
* Correspondence: email@example.com
Lighting Research Center, Rensselaer Polytechnic Institute, 21 Union Street,
Troy, NY 12180, USA
Rea et al. Journal of Circadian Rhythms 2010, 8:2
© 2010 Rea et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
similar but different enough to require thoughtful delib-
eration as a definition of circadian light evolves. With-
out a clear understanding of these differences,
experimental results from studies of the impact of opti-
cal radiation on circadian physiology can be easily mis-
interpreted. Since, however, so much history and
thought underlie our concept of light based upon the
human visual system, these insights make the discussion
of circadian light more readily explained and more easily
understood. For this reason brief reviews of photometry,
colorimetry and brightness perception are presented as
a foundation for the discussion of circadian light.
The photopic luminous efficiency function
Psychophysical experiments were conducted by several
laboratories nearly a century ago to develop “the spec-
tral sensitivity of human vision.” Following a consensus
process, the data from these experiments were com-
bined to form Vl, the photopic luminous efficiency
function formally defining light , shown in Figure 1.
Vlthen is the bridge between radiometry, the measure-
ment of radiant energy, and photometry, the measure-
ment of light. Depending upon the geometric properties
of interest, radiant flux (radiant energy per unit time) is
weighted by Vlin the fundamental definitions of lumi-
nous intensity (Vl-weighted radiant intensity, or radiant
flux within a solid angle), illuminance (Vl-weighted irra-
diance, or radiant flux incident on a surface area), and
luminance (Vl-weighted radiance, or luminous intensity
per unit area of a surface) . Circadian light could
then similarly bridge radiometry to circadian photome-
try and would have parallel definitions with those used
for light based upon geometrical considerations.
The photopic luminous efficiency function is actually
only one of a wide variety of functions that can be used
to characterize the spectral sensitivity of the human
visual system. Figure 1 also shows a small sample of
human spectral sensitivity functions published in the lit-
erature [12-15]. In fact, depending upon the experimen-
tal conditions, many spectral response functions can be
obtained from the human visual system. Vlis quite spe-
cial, however, because, in addition to its metrological
seniority, it has the important practical feature of exhi-
biting additivity. Additivity means that when two lights
(A and B) of different spectral power distributions but
of equal luminance (A = B) are combined by unit frac-
tional amounts, they will continue to have the same
luminance . That is,
if A B then pA, qBAB
where p and q are unit fractional amounts, such that
p + q = 1
Additivity as defined by Equation 1 significantly
restricts the relevance of Vlfor representing the spectral
sensitivity of the human visual system to a small handful
of visual task conditions . Because of additivity, how-
ever, Vlhas become the universal “visual response func-
tion” for commerce and for government regulations
Despite its assumed universality, the psychophysical
techniques used to develop Vlonly functionally charac-
terize the spectral sensitivity of the achromatic (lumi-
nance) channel for the human fovea which is dominated
by input from only two of the three cone types. The
fovea constitutes only about 2% of the retina and pro-
vides humans with high spatial resolution. Only densely
packed long-wavelength (L) and middle-wavelength (M)
sensitive cones are found in the center of the fovea;
although all three cone types are found throughout the
remainder of the retina, the short-wavelength (S) sensi-
tive cones are much rarer and, like the rods, absent
from the center of the fovea. The S cone is also slower
to respond to rapid modulations of light level than the
L and M cones . Vlis largely (but not exclusively)
based upon a psychophysical technique known as flicker
photometry. A small disc presented to an observer at
the center of the fovea oscillates in time (flickers)
between two lights of different spectral power distribu-
tions (perceived colors). By gradually adjusting the radi-
ance of one light and the flicker rate, the two lights
eventually appear as a steady light of a single hue. At
this point where the two oscillating lights just fuse into
what appears to be a fixed luminous disc, the two lights
are defined as having the same luminance. Vlis deter-
mined by taking the reciprocal of the radiance at each
wavelength needed to reach this constant-luminance
flicker criterion and normalizing these values to the
reciprocal of the radiance associated with the wave-
length requiring the least amount of optical radiation
needed to elicit the criterion response (lmax= 555 nm).
Figure 1 Photopic and scotopic luminous efficiency functions
 and other spectral sensitivity functions measured with
humans (mesopic: Rea et al. , glare: Bullough ,
brightness, central: CIE , brightness, peripheral: Weale ).
Peak wavelengths for each function are noted in the legend.
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 2 of 10
By utilizing rapidly oscillating lights in the fovea, the S
cone is functionally excluded from the definition of light
even though this photoreceptor plays an extremely
important role in our perception of brightness .
Nevertheless, Vl has gained ubiquity in metrology
because additivity is essential for any system of photo-
metry supporting commerce and government.
To be useful, a system of photometry first must, more
or less, describe the relative brightness of a light source.
The current system of photometry based on Vldoes so,
more or less. As a “white” light source (e.g., daylight,
incandescent, fluorescent) generates greater radiance,
the light source should appear brighter and the photo-
metric quantity should increase. Further, as the spectral
power distribution of the light emitted by a source shifts
to the middle of the “visible spectrum” (i.e., near 555
nm) the source should also appear brighter and the
photometric quantity should also increase. In general,
photometry based upon Vlprovides quantities consis-
tent with these expectations. Ironically perhaps, a photo-
metric system based on apparent brightness will not
conform to these prima facie expectations. As will be
discussed in more detail in the next section, it is possi-
ble to show that when two lights of equal brightness are
added together their sum can actually appear less bright
than either light alone. Commerce and government sim-
ply could not employ a non-additive system of photo-
metry where summing more optical radiation produced
The significance of additivity in the definition, and
thereby, the sale and regulation of light should not be
underestimated. Not only does additivity ensure that the
combination of optical radiation always increases the
amount of measured light, additivity also provides for
inexpensive and practical means of measuring that light.
Additivity ensures that, at any level of optical radiation,
a linear detector-filter combination matching the spec-
tral response of Vlwill provide photometric quantities
identical to the sum of the spectral power obtained at
each wavelength by a much more expensive and compli-
cated spectroradiometer. Additivity is the dominant and
perhaps only reason Vlhas not been displaced by com-
merce and government after nearly a century of research
showing the inherent limitations of Vlfor characterizing
the visual stimulus [16,18].
Spectral sensitivity of brightness perception
In addition to the visual system’s achromatic luminance
channel, the spectral sensitivity of which is well charac-
terized by Vl, two spectrally opponent color channels
simultaneously contribute to our perceptions of bright-
ness. The three visual channels leading to brightness
perceptions are formed in the retina from the three
cone photoreceptor types (L, M, and S cones) but,
depending upon the subsequent neural connections,
they are combined in different ways to provide bright-
ness information to the visual cortex. As previously dis-
cussed, the spectral sensitivity of the luminance channel
is dominated by the summed input from the L and M
cones. The two color channels, red versus green (r-g)
and blue versus yellow (b-y) are termed spectrally oppo-
nent channels because each provides opposing color
information to the brain .
For one type of r-g channel, excitatory input is pro-
vided by the L cones and inhibitory input is provided by
the M cones. For example, when the L cone provides
relatively more input to the r-g channel than the M
cone, the r-g channel signals “red” to the brain. Simi-
larly, for one type of b-y channel, excitatory input is
provided by the S cones and inhibitory input is provided
by both the L cones and the M cones. When a light sti-
mulates the S cones more than the combined input
from the L and M cones, the b-y channel signals “blue”
to the brain.
As spectrally opponent systems, these channels can
signal either “red” or “green” and either “blue” or “yel-
low” to the brain. Moreover, a spectrally opponent sys-
tem is inherently a subadditive system because the
addition of, say, a “green” light to a “red” light can
decrease the response of the r-g system cell. Since the
two spectral opponent channels contribute to brightness
perception, two lights added together can actually
appear less bright than either light alone.
Much of the research attempting to understand
human brightness perception has utilized both photo-
metry and colorimetry as indirect methods of measuring
the apparent brightness of lights of different spectral
compositions. Colorimetry originated from controlled
observations in the 19thcentury showing that with
three, and only three, so called primary lights humans
can match the appearance of any other test light . In
other words, by adjusting the radiances of the three pri-
mary lights, it was possible to create an additive mixture
of these primary lights that was completely indistin-
guishable from the test light. (Some very saturated color
stimuli cannot be matched using a set of three physical
primaries without slightly changing the color of the test
light by adding one of the physical primaries to it.
Mathematically, this is equivalent to using a negative
amount of primary to make the match. To avoid the use
of negative amounts of primaries, the CIE color system
 makes use of imaginary primaries that are mathe-
matically defined, but not physically realizable.) As
shown in subsequent research, the mixture matches the
test light because the photon absorptions by the three
cone photoreceptors are exactly the same for the mix-
ture of primaries and for the test. Thus, the color of any
test light can be quantified in terms of the relative
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 3 of 10
amounts of the primary lights needed to match its
appearance. The radiant powers in the three primaries
are typically normalized with a linear transformation so
that their sum is unity. By knowing two of the normal-
ized values, the third value is also known. In this way
the color, or more precisely the chromaticity, of the test
light can be illustrated graphically in two dimensions.
Figure 2 illustrates the results of colorimetric calcula-
tions based on the spectral power distribution of the
light source and the three color matching functions in
the (x’, y’) color system. (This color system is nearly
identical to the Commission Internationale de l’Eclairage
(x, y) 1931 color system presently in common use ,
but with small differences for short [<460 nm] wave-
lengths .) The physical specification of the chromati-
city of any light, natural or fabricated, can be defined as
a single point within the area enclosed by the outermost
Of course, quantifying the relative amounts of three
primaries needed to match the test light does not fully
characterize its color appearance. A stimulus that is
recognized as orange and another that is recognized as
brown can have the same chromaticities, but factors
such as the objects’ relative luminances against their
surrounding luminances will influence their perceived
colors. Unlike photometry, where Vlclosely charac-
terizes the spectral sensitivity of the human fovea for
some types of visual tasks , colorimetric specification
of a test light does not represent the response of a single
visual channel or even the combination of multiple
visual channels. Since brightness is influenced by both
chromatic as well as achromatic visual channels,
Figure 2 Chromaticity diagram based on the Judd correction , showing contours of equal B/L value . Also shown are the
chromaticity coordinates for a red light (RED) and for a green light (GRN) that, when added, produce a yellow light (YEL) with the illustrated
chromaticity coordinates as described in the text.
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 4 of 10
additional techniques must be used to characterize and
measure apparent brightness.
One common method of quantifying the brightness of
lights is combining photometry and colorimetry by uti-
lizing the ratio of the luminance needed by a test light
of a given chromaticity to match the brightness of a
reference light of another chromaticity and of a known
luminance. The luminance of the reference light is
designated B and the luminance of the test light that
matches its brightness is designated L. Therefore, the
brightness of any light in chromaticity space relative to
a given reference light source can be described in terms
of a unitless ratio, its B/L value. Figure 2 also shows
constant B/L contours in the (x’, y’) color space 
and, as described next, illustrates the subadditive nature
of brightness perception.
Consider a red light (a 630-nm spectral light) and a
green light (a 520-nm spectral light), with (x’, y’) chro-
maticity coordinates of (0.70, 0.30) and (0.07, 0.83),
respectively, as shown in Figure 2. Suppose the lumi-
nance of the red light is 10 cd/m2and that of the green
light is 15 cd/m2. Using the B/L values from Guth et al.
 shown in Figure 2 (2.93 for red and 2.15 for green),
their apparent brightnesses can be calculated from the
product of their luminance (L, in cd/m2) and their B/L
• red: 10 cd/m2× 2.93 = 29.3
• green: 15 cd/m2× 2.15 = 32.3
If the red and the green lights are superimposed onto
each other, the luminance of the resulting yellow light
would, of course, be 25 cd/m2(10 + 15 cd/m2). The (x’,
y’) chromaticity coordinates of this yellow light are
(0.48, 0.49), corresponding to a B/L value of 1.07 using
the B/L contours from Guth et al.  in Figure 2.
Therefore, the apparent brightness of this yellow light
can be calculated as it was for the red and green lights:
• yellow: 25 cd/m2× 1.07 = 26.8
Remarkably, the brightness of the yellow light created
by combining the original red and green lights appears
less bright than either the red or the green light alone,
despite the yellow light being created from the superim-
position of the red and green lights.
A “photodian” luminous efficiency function
It seems natural that as more research is conducted on
the impact of optical radiation on the circadian system,
particularly as it might affect human health, attempts
would be made to develop a spectral sensitivity function
for the circadian system. It also seems natural that
attempts would be made to develop an additive sprectral
efficiency function comparable to Vlfor the circadian
system, a Cl[24,25]. Certainly it is possible to develop
such a function from the available data (e.g., [26,27])
through international consensus to support commerce
and government, but it is important to point out why
an additive function like Clcould probably never be
exactly comparable to Vl.
Intrinsically photosensitive retinal ganglion cells
(ipRGCs) have been shown to provide direct input to
the SCN [28,29]. Figueiro et al.  were the first to
suggest that multiple photoreceptors contributed to
human circadian phototransduction via color opponent
processes distal to the ipRGCs in the retina. Spectral
opponency is an inherent attribute of the human retina,
initiated distal to the ipRGCs in the outer plexiform
layer of the retina, and underlies both color perception
and the subadditive nature of apparent brightness per-
ception previously described. Demonstrations of subad-
ditivity in human circadian phototransduction have been
performed by Figueiro et al.  specifically designed to
test the conclusions by Figueiro et al. . More
recently, Figueiro and colleagues demonstrated, as pre-
dicted, that the subadditive response to light by the cir-
cadian system is formed from spectral opponent
mechanisms in the retina . It is interesting in this
regard that additivity has been demonstrated in mouse
circadian phototransduction [33,34]. This species does
not exhibit subadditivity presumably because, quite
unlike humans, mice have very little neural apparatus to
support color vision .
Vl, as previously discussed, is based upon a specific
experimental paradigm isolating the achromatic visual
channel in the fovea. This channel has been shown to
be additive in response to optical radiation for a given
criterion response (i.e., a constant-luminance flicker cri-
terion). Our current understanding of the circadian sys-
tem indicates that there is only one functional channel
leading to the SCN from the retina, and that in humans,
this channel exhibits subadditivity to certain combina-
tions of wavelengths . Clearly, a more detailed
understanding of input to the SCN may emerge follow-
ing additional research. For example, Aggelopolous and
Meissl  suggest that there are multiple neural path-
ways providing input to the SCN in rats. Whether these
different neuron pathways exist in humans or constitute
different functional channels for the SCN has yet to be
determined. Since there is no evidence to date that the
human circadian system exhibits additivity, an additive
“photodian” luminous efficiency function for measuring
circadian light (i.e., a Cl) would only serve as a conveni-
ence to commerce and government. In other words,
unlike Vl, there would be no physiological foundation
for a system of metrology based upon Cl. This lack of
homology between physiology and metrology may or
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 5 of 10
may not be an important aspect in the deliberations for
developing a system of circadian photometry, but it is
certainly important to draw attention to this difference
for scientific purposes, much as it is important to draw
attention to the difference between luminance and
Spectral sensitivity of the circadian system
The retino-hypothalamic tract (RHT) is comprised of
ipRGC axons and carries photic information from the
retina to the SCN. In addition to the direct conversion
of optical radiation to neural signal input to the master
clock, the ipRGCs also carry spectrally-opponent infor-
mation originating from the classical photoreceptors
and processed by vertical (bipolar cells) and lateral
(horizontal and amacrine cells) neurons, to the SCN.
Of particular interest with regard to developing a defi-
nition of circadian light are the spectrally-opponent
(color) mechanisms in the distal retina that provide
synaptic connections to the ipRGCs . In addition,
amacrine cells that control the transition from scotopic
(rod) to photopic (cone) responses in retinal ganglion
cells also appear to provide synaptic threshold control
of the ipRGC responses. These complicated neural
connections have been mathematically modeled to
develop a definition of the circadian (light) stimulus
. The mathematical model of human circadian
phototransduction developed by Rea et al.  is based
on the neuroanatomy and neurophysiology of the
retina and on published psychophysical studies of noc-
turnal melatonin suppression using lights of different
spectral power distributions. The model generates
values of circadian light (CL) for any spectral power
distribution (i.e., for any light source, real or imagined,
at any irradiance). CL is characterized by a high abso-
lute threshold to optical radiation with a peak spectral
response at short wavelengths. The model accounts for
participation of ipRCGs as well as rods and cones in
circadian phototransduction via neural connections,
including spectral opponency, in the outer plexiform
layer of the retina. Additional file 1 describes the com-
putation procedure for calculating CL. It should be
noted that the term CL is used in this paper to replace
the term circadian stimulus (CS), used in the paper
that describes the model of circadian phototransduc-
tion . Notwithstanding the nonlinearities inherent
in the circadian phototransduction model, CL is spec-
trally weighted irradiance for the human circadian sys-
tem, a term more comparable to illuminance, which is
spectrally weighted irradiance for the human visual
system. As described in more detail below, the term
CS will be henceforth used to describe the effective
photic stimulus for the circadian system as measured
by acute nocturnal melatonin suppression.
Figure 3 shows the modeled spectral sensitivity of
the circadian system for both narrowband and poly-
chromatic light stimuli . Because the model
includes spectral opponency, responses from light sti-
muli created by a combination of narrowband sources
cannot be predicted from the spectral sensitivity
derived from narrowband light stimuli alone. In fact,
for light stimuli with a particular balance of short-
wavelength (e.g., around 450 nm) and long-wavelength
(e.g., longer than about 510 nm) components, the
response of the human circadian system to light is less
than what would be predicted by an additive spectral
efficiency function derived from responses to narrow-
band stimuli [30,31]. Emphasis for modeling was
placed upon studies measuring nocturnal melatonin
suppression because, in fact, there are presently no
comparable spectral sensitivity functions for the circa-
dian system using any other outcome measure (e.g.,
phase shifting). The values of the coefficients in Addi-
tional file 1 relating the opponent channels were opti-
mized to align with published nocturnal human
melatonin suppression data using narrowband spectra
[26,27]. This resulted in a good fit (r2= 0.82) between
all comparable suppression data using both narrow-
band and broadband spectra [24,26,27,30,38,39] and a
four-parameter logistic function  characterizing the
melatonin suppression response as the light stimulus
increases from threshold to saturation (Figure 4). CL is
defined in terms of irradiance, not radiance as with
brightness because image formation on the retina is
not believed to be important to the circadian system.
Rather, CL is geometrically described in terms of radi-
ant flux density on the cornea and therefore is geome-
trically comparable to illuminance at the eye.
Previously published studies have been conducted to
test the utility of the model of human circadian photo-
transduction [32,41]. Nocturnal melatonin suppression
by light was assessed by Figueiro et al.  for two light
source spectra and four light levels. A priori predictions
of melatonin suppression were made based on calcula-
tions of CL given by each light level and spectra. Results
demonstrated that the model predictions were consis-
tent with melatonin suppression obtained at all four
light levels, although uncertainty was greater at the low-
est light level, which was close to threshold response.
Also, Figueiro et al.  measured nocturnal melatonin
suppression following exposure to lights presented
monocularly and binocularly to demonstrate that the
subadditive response to light by the circadian system
originated in the retina as predicted by Rea et al. .
Figueiro et al.  measured nocturnal melatonin sup-
pression from short-wavelength light stimuli; these data
were consistent with predictions made using the model
by Rea and colleagues . Similarly, noctural melatonin
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 6 of 10
suppression measurements reported by Revell and Skene
 in response to narrowband and broadband light sti-
muli varying in intensity were shown  to be consis-
tent with predictions using this model .
Utilization of circadian light
As previously described, Rea et al.  proposed a
mathematical model for quantifying circadian light for
any spectral irradiance distribution. Two changes in the
circadian light nomenclature from that paper have been
made for metrological clarity. First, because the units of
CL (spectrally weighted irradiance in W/m2) are new
and, therefore, are not particularly intuitive to a user, a
normalized quantity, CLA, was derived to more easily
compare CL values with values of photopic illuminance,
in lux (lx). A value of CL can be determined, measured
Figure 3 Nocturnal human melatonin suppression data from Brainard et al.  and Thapan et al.  for narrowband spectra
(symbols), and a spectral sensitivity function resulting from exposure to narrowband illumination (solid curve). Also shown is the
spectral sensitivity for broadband illumination when spectral opponency is exhibited .
Figure 4 Nocturnal human melatonin suppression data [24,26,27,30,38,39] (left ordinate), plotted as a function of CL quantities
(abscissa) predicted by the model of Rea et al. . Also shown is the best-fitting four-parameter logistic function  to all of the data from
threshold to saturation. The circadian light quantity CL was formerly named circadian stimulus (CS) ; CS (right ordinate) now refers to the
effective stimulus based on nocturnal melatonin suppression.
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 7 of 10
or calculated, for 1000 lx of CIE standard illuminant A
, a blackbody radiator at a color temperature of
2856 K similar in relative spectral power distribution to
an incandescent lamp, and a scalar multiplier deter-
mined to make the product of CL and the multiplier
equal 1000. The product of CL and this multiplier
defines the quantity CLA. Any value of CL can then be
normalized in terms of a reference illuminance of 1000
lx from the standard illuminant A equaling 1000 CLA
units. CLAis numerically identical to orthodox photopic
illuminance when illuminant A produces 1000 lx, but
can differ, sometimes significantly, for other spectral
power distributions and illuminance levels due to non-
linear operations in the CL formulation (see Additional
file 1). Nevertheless, for many common white light
sources values of CLAare similar in magnitude to illu-
minance values (in lx) at any level.
Second, circadian stimulus (CS; [3,36,44,45]) in the
original formulation is now defined as CL and, after
normalization, as CLA. To understand why, consider
two light sources producing very different irradiance
and spectral quantities, resulting in CLA values of
10,000 and of 20,000 units. Despite a large difference in
the values of CLA, the two sources would not be
expected to produce different outcomes from the circa-
dian system, at least in terms of nocturnal melatonin
suppression. Both would produce saturating levels of
suppression of about 75% percent after an hour of expo-
sure. Thus, while the two sources would be character-
ized as being very different in terms of CLA, their
effectiveness as a circadian stimulus in terms of noctur-
nal melatonin suppression would be identical. The term
CS for a given light source is therefore now defined in
terms of the relative effectiveness of CL, or CLA, for pro-
ducing a meaningful circadian response. The logistic
function in Figure 4 is used to relate a given CL, or
CLA, value to its corresponding CS value, between 0 (or
0%) and 0.75 (or 75%), characterizing the relative effec-
tiveness of the source as a stimulus to the circadian
The implications for establishing quantitative mea-
sures of CL, CLAand CS are key to developing an
understanding of how temporal patterns of light and
dark affect human health and well-being. Without a
quantitative understanding of the circadian light stimu-
lus it will be difficult or impossible to make significant
progress in unraveling the role that circadian disruption
has on diseases such as breast cancer [2,3], cardiovascu-
lar disease [4-6], diabetes [7,8] and sleep disorders .
Toward this end a circadian light dosimeter, the Day-
simeter (Figure 5), was developed to quantify circadian
light exposures in these vulnerable populations. The
Daysimeter, previously described  and subsequently
refined , is a personal head-worn device that
measures CLAand photopic illuminance near the plane
of the wearer’s cornea. The Daysimeter also includes cali-
brated accelerometers to measure rest and activity. Data
from the Daysimeter are recorded for as long as one
month of wear and retrieved for post-processing. Each
Daysimeter has its own spectral, spatial, and absolute
light calibration so, following post-processing, it is possi-
ble to quantify individual CLAexposures in real life over
extended periods. These data have great potential for
understanding the impact of circadian disruption on
human health because, for the first time, researchers and
clinicians can actually measure circadian disruption
among individuals in these vulnerable populations.
Although beyond the scope of this paper, the Daysimeter
has, in fact, recently been used to quantify and compare
circadian disruption in day-shift and rotating-shift nurses
[47,48]. Future research will undoubtedly utilize instru-
ments like the Daysimeter to develop, for example, new
shift-work schedules, new architectural practices and new
light sources, all of which will depend upon our collective
ability to measure and calculate circadian light.
Figure 5 Close-up photograph of the Daysimeter. Two light
sensors are located near the plane of the cornea, calibrated in
terms of their absolute, spatial and spectral response to provide
both photopic and circadian light readings. The rear housing
attached to the earpiece contains accelerometers for measuring
activity as well as memory and control circuitry, all powered by a
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 8 of 10
Circadian light as a work in progress
The definition of circadian light proposed here is based
on the current knowledge of the neuroanatomy and
neurophysiology of the human retina and on published
psychophysical studies of nocturnal melatonin suppres-
sion using lights of different spectral power distribu-
tions. CLAand CS are provisionally defined then in
terms of what is known about nocturnal melatonin sup-
pression in humans (for an hour-long exposure to light,
near the midpoint of the melatonin production curve,
and with naturally-constricted pupils). Of course, acute
melatonin suppression is not the only non-visual
response by the circadian system; other non-visual
responses include phase shifting and alertness. For at
least these two cases, however, light-induced phase-
shifting and light-induced nocturnal alertness appear to
have similar threshold-to-saturation response character-
istics [40,44]. A very recent study, however, has shown
that both red and blue lights can affect alertness  as
well as cortisol and alpha amylase production (Figueiro
and Rea., unpublished data) at night indicating that not
all light-induced, non-visual responses have the same
spectral sensitivity as nocturnal melatonin suppression.
The development of new response characteristics for
these non-visual systems, if they emerge, would be very
reminiscent of those that were developed in visual
science where multiple spectral sensitivity functions for
different visual channels were established (cf. Figure 1).
If it is shown that the relationships between CLAand
other non-visual responses, such as phase shifting, are
different than the one demonstrated for nocturnal mela-
tonin suppression, another CS function could be devel-
oped and designated with an appropriate subscript (such
as CSnmelfor nocturnal melatonin suppression and
CSpshiftfor phase shifting). Again, this development
would be quite similar to the evolution of different
visual spectral sensitivity functions.
As a final note, even the model of human circadian
phototransduction based upon nocturnal melatonin sup-
pression and used to calculate CLAand CS is probably
incomplete. It does not take into account possible parti-
cipation of different types of ipRGCs  and recent
evidence that the melanopsin photopigment in the
ipRGCs follows a very different regenerative process
than that employed by the classical photoreceptors .
These phenomena may have heretofore unknown effects
on the spectral and absolute sensitivities of the circadian
system that would demand consideration in a revised
model of phototransduction and therefore an evolving
definition of circadian light. Hopefully, however, the
information presented here is an important step toward
the precise application of light stimuli for the human
Light is formally defined as optical radiation capable of
providing visual sensation in humans. The current defi-
nition of light does not directly relate to its effects on
the human circadian system. Since temporal patterns of
retinal light (and dark) exposures regulate the human
circadian system and since disruption of the circadian
system has broad implications for health and well-being
[2-9,52,53], it is becoming increasingly important to
develop a new definition of circadian light.
Toward that end, the present paper has placed the
evolving development of a definition of circadian light
into the historical context of light as it has been
defined for metrology and as it affects human vision.
As described here, an additive “photodian” luminous
efficiency function for circadian light will probably
never be exactly comparable to the photopic luminous
efficiency function used in conventional photometry
based upon the human visual system. Nevertheless, it
is increasingly important that a measurement system,
such as CL, CLA, and CS as presented here, be devel-
oped for quantifying the photic stimulus for the human
Additional file 1: Circadian light (CL, CLA) and circadian stimulus
(CS) calculation procedure .
Click here for file
This research was supported in part by the Trans-NIH Genes, Environment
and Health Initiative Grant U01 DA023822 to the first author and by CDC
Grant R01 OH008171 to Dr. Eva Schernhammer at Harvard Public Health. The
New York Energy Research and Development Authority through the
National Science Foundation (NSF) Smart Lighting Engineering Research
Center (EEC-0812056) provided support for improvements to the Daysimeter.
The authors thank Dennis Guyon for manuscript editing and preparation of
the final graphics.
The outline of the article was developed by all co-authors. MSR led the
effort and wrote a partial draft of the manuscript with MGF. AB and JDB
wrote specific sections of the text and prepared the figures. All co-authors
reviewed and approved the final manuscript.
The authors declare that they have no competing interests.
Received: 11 December 2009
Accepted: 13 February 2010 Published: 13 February 2010
1. Moore-Ede MC, Sulzman FM, Fuller CA: The Clocks That Time Us Cambridge:
Harvard University Press 1982.
2. Stevens RG, Rea MS: Light in the built environment: potential role of
circadian disruption in endocrine disruption and breast cancer. Cancer
Causes Control 2001, 12:279-87.
3. Figueiro MG, Rea MS, Bullough JD: Does architectural lighting contribute
to breast cancer?. J Carcinog 2006, 5:20.
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 9 of 10
4. Lemmer B: Importance of circadian rhythms for regulation of the Download full-text
cardiovascular system: studies in animal and man. Conf Proc IEEE Eng
Med Biol Soc 2006, 1:168-70.
Maemura K, Takeda N, Nagai R: Circadian rhythms in the CNS and
peripheral clock disorders: role of the biological clock in cardiovascular
diseases. J Pharmacol Sci 2007, 103:134-8.
Young ME, Bray MS: Potential role for peripheral circadian clock
dyssynchrony in the pathogenesis of cardiovascular dysfunction. Sleep
Med 2007, 8:656-67.
Kreier F, Kalsbeek A, Sauerwein HP, Fliers E, Romijn JA, Buijs RM: “Diabetes
of the elderly” and type 2 diabetes in younger patients: possible role of
the biological clock. Exp Gerontol 2007, 42:22-7.
Laposky AD, Bass J, Kohsaka A, Turek FW: Sleep and circadian rhythms:
key components in the regulation of energy metabolism. FEBS Lett 2008,
Knutson KL, Spiegel K, Penev P, Van Cauter E: The metabolic
consequences of sleep deprivation. Sleep Med Rev 2007, 11:163-78.
Commission Internationale de l’Eclairage: Light as a True Visual Stimulus
Paris: Commission Internationale de l’Eclairage 1978.
Rea MS, ed: IESNA Lighting Handbook: Reference and Application New York:
Illuminating Engineering Society of North America 2000.
Commission Internationale de l’Eclairage: CIE Proceedings 1924 Cambridge:
University of Cambridge 1926.
Rea MS, Bullough JD, Freyssinier-Nova JP, Bierman A: A proposed unified
system of photometry. Lighting Res Technol 2004, 36:85-111.
Bullough JD: Spectral sensitivity for extrafoveal discomfort glare. J Mod
Opt 2009, 56:1518-22.
Weale RA: Spectral sensitivity and wave-length discrimination of the
peripheral retina. J Physiol 1953, 119:170-90.
Lennie P, Pokorny J, Smith VC: Luminance. J Opt Soc Am A 1993,
Brindley GS, Du Croz JJ, Rushton WA: The flicker fusion frequency of the
blue-sensitive mechanism of color vision. J Physiol 1966, 183:497-500.
Wyszecki G, Stiles WS: Color Science New York: Wiley 1982.
De Valois RL, Abramov I, Jacobs GH: Analysis of response patterns of LGN
cells. J Opt Soc Am 1966, 56:966-77.
Helmholtz H: Handbuch der Physiologischen Optik Hamburg: Voss 1896.
Commission Internationale de l’Eclairage: Colorimetry Vienna: Commission
Internationale de l’Eclairage 2004.
Commission Internationale de l’Eclairage: CIE 1988 2° Spectral Luminous
Efficiency Function for Photopic Vision Vienna: Commission Internationale de
Guth SL, Massof RW, Benzschawel T: Vector model for normal and
dichromatic color vision. J Opt Soc Am 1980, 70:197-212.
Rea MS, Bullough JD, Figueiro MG: Phototransduction for human
melatonin suppression. J Pineal Res 2002, 32:209-13.
Gall D, Bieske K: Definition and measurement of circadian radiometric
quantities. Proceedings of the CIE Symposium ‘04 on Light and Health
Vienna: Commission Internationale de l’Éclairage 2004, 129-32.
Brainard GC, Hanifin JP, Greeson JM, Byrne B, Glickman G, Gerner E,
Rollag MD: Action spectrum for melatonin regulation in humans:
evidence for a novel circadian photoreceptor. J Neurosci 2001, 21:6405-12.
Thapan K, Arendt J, Skene DJ: An action spectrum for melatonin
suppression: evidence for a novel non-rod, non-cone photoreceptor
system in humans. J Physiol 2001, 535:261-7.
Berson DM, Dunn FA, Takao M: Phototransduction by retinal ganglion
cells that set the circadian clock. Science 2002, 295:1070-3.
Hattar S, Liao HW, Takao M, Berson DM, Yau KW: Melanopsin-containing
retinal ganglion cells: architecture, projections, and intrinsic
photosensitivity. Science 2002, 295:1065-70.
Figueiro MG, Bullough JD, Parsons RH, Rea MS: Preliminary evidence for
spectral opponency in the suppression of melatonin by light in humans.
Neuroreport 2004, 15:313-6.
Figueiro MG, Bullough JD, Bierman A, Rea MS: Demonstration of additivity
failure in human circadian phototransduction. Neuroendocrinol Lett 2005,
Figueiro MG, Bierman A, Rea MS: Retinal mechanisms determine the
subadditive response to polychromatic light by the human circadian
system. Neurosci Lett 2008, 438:242-5.
Bullough JD, Figueiro MG, Possidente BP, Parsons RH, Rea MS: Additivity in
murine circadian phototransduction. Zoolog Sci 2005, 22:223-7.
34. Dkhissi-Benyahya O, Gronfier C, De Vanssay W, Flamant F, Cooper HM:
Modeling the role of mid-wavelength cones in circadian responses to
light. Neuron 2007, 53:677-87.
Jacobs GH, Williams GA, Fenwick JA: Influence of cone pigment
coexpression on spectral sensitivity and color vision in the mouse. Vision
Res 2004, 44:1615-22.
Rea MS, Figueiro MG, Bullough JD, Bierman A: A model of
phototransduction by the human circadian system. Brain Res Rev 2005,
Aggelopoulos NC, Meissl H: Responses of neurones of the rat
suprachiasmatic nucleus to retinal illumination under photopic and
scotopic conditions. J Physiol 2000, 523:211-22.
McIntyre IM, Norman TR, Burrows GD, Armstrong SM: Human melatonin
suppression by light is intensity dependent. J Pineal Res 1989, 6:149-56.
Rea MS, Bullough JD, Figueiro MG: Human melatonin suppression by
light: a case for scotopic efficiency. Neurosci Lett 2001, 299:45-8.
Zeitzer JM, Dijk DJ, Kronauer R, Brown E, Czeisler C: Sensitivity of the
human circadian pacemaker to nocturnal light: melatonin phase
resetting and suppression. J Physiol 2000, 526:695-702.
Figueiro MG, Rea MS, Bullough JD: Circadian effectiveness of two
polychromatic lights in suppressing human nocturnal melatonin.
Neurosci Lett 2006, 406:293-7.
Figueiro MG, Bierman A, Bullough JD, Rea MS: A personal light-treatment
device for improving sleep quality in the elderly: dynamics of nocturnal
melatonin suppression at two exposure levels. Chronobiol Int 2009,
Revell VL, Skene DJ: Light-induced melatonin suppression in humans
with polychromatic and monochromatic light. Chronobiol Int 2007,
Bullough JD, Bierman A, Figueiro MG, Rea MS: On melatonin suppression
from polychromatic and narrowband light. Chronobiol Int 2008, 25:653-5.
Figueiro MG, Bullough JD, Bierman A, Fay CR, Rea MS: On light as an
alerting stimulus at night. Acta Neurobiol Exp 2007, 67:171-8.
Bierman A, Klein TR, Rea MS: The Daysimeter: a device for measuring
optical radiation as a stimulus for the human circadian system. Meas Sci
Technol 2005, 16:2292-9.
Rea MS, Bierman A, Figueiro MG, Bullough JD: A new approach to
understanding the impact of circadian disruption on human health. J
Circadian Rhythms 2008, 6:7.
Miller D, Bierman A, Figueiro MG, Schernhammer ES, Rea MS: Ecological
measurements of light exposure, activity, and circadian disruption in
real-world environments. Proceedings of Experiencing Light 2009 Eindhoven:
Eindhoven University of Technology 2009, 53-61.
Figueiro MG, Bierman A, Plitnick B, Rea MS: Preliminary evidence that both
blue and red light can induce alertness at night. BMC Neurosci 2009,
Schmidt TM, Kofuji P: Functional and morphological differences among
intrinsically photosensitive retinal ganglion cells. J Neurosci 2009,
Lucas RJ: Chromophore regeneration: melanopsin does its own thing.
Proc Natl Acad Sci USA 2006, 103:10153-4.
Stevens RG, Blask DE, Brainard GC, Hansen J, Lockley SW, Provencio I,
Rea MS, Reinlib L: Meeting report: the role of environmental lighting and
circadian disruption in cancer and other diseases. Environ Health Perspect
Straif K, Baan R, Grosse Y, Secretan B, El Ghissassi F, Bouvard V, Altieri A,
Benbrahim-Tallaa L, Cogliano V: Carcinogenicity of shift-work, painting,
and fire-fighting. Lancet Oncol 2007, 8:1065-6.
Dacey DM, Liao HW, Peterson BB, Robinson FR, Smith VC, Pokorny J,
Yau KW, Gamlin PD: Melanopsin-expressing ganglion cells in primate
retina signal color and irradiance and project to the LGN. Nature 2005,
Commission Internationale de l’Eclairage: CIE 10 Degree Photopic
Photometric Observer Vienna: Commission Internationale de l’Eclairage 2005.
Smith VC, Pokorny J: Spectral sensitivity of the foveal cone
photopigments between 400 and 500 nm. Vision Res 1975, 15:161-71.
Cite this article as: Rea et al.: Circadian light. Journal of Circadian Rhythms
Rea et al. Journal of Circadian Rhythms 2010, 8:2
Page 10 of 10