Endocrinology and Metabolism

Spectral modulation attenuates molecular, endocrine, and neurobehavioral disruption induced by nocturnal light exposure

Shadab A. Rahman, Shai Marcu, Colin M. Shapiro, Theodore J. Brown, Robert F. Casper


The human eye serves distinctly dual roles in image forming (IF) and non-image-forming (NIF) responses when exposed to light. Whereas IF responses mediate vision, the NIF responses affect various molecular, neuroendocrine, and neurobehavioral variables. NIF responses can have acute and circadian phase-shifting effects on physiological variables. Both the acute and phase-shifting effects induced by photic stimuli demonstrate short-wavelength sensitivity peaking ≈450–480 nm. In the current study, we examined the molecular, neuroendocrine, and neurobehavioral effects of completely filtering (0% transmission) all short wavelengths <480 nm and all short wavelengths <460 nm or partially filtering (∼30% transmission) <480 nm from polychromatic white light exposure between 2000 and 0800 in healthy individuals. Filtering short wavelengths <480 nm prevented nocturnal light-induced suppression of melatonin secretion, increased cortisol secretion, and disrupted peripheral clock gene expression. Furthermore, subjective alertness, mood, and errors on an objective vigilance task were significantly less impaired at 0800 by filtering wavelengths <480 nm compared with unfiltered nocturnal light exposure. These changes were not associated with significantly increased sleepiness or fatigue compared with unfiltered light exposure. The changes in molecular, endocrine, and neurobehavioral processes were not significantly improved by completely filtering <460 nm or partially filtering <480 nm compared with unfiltered nocturnal light exposure. Repeated light-dark cycle alterations as in rotating nightshifts can disrupt circadian rhythms and induce health disorders. The current data suggest that spectral modulation may provide an effective method of regulating the effects of light on physiological processes.

  • circadian rhythms
  • short-wavelength light
  • melatonin
  • cortisol
  • alertness
  • mood
  • sleepiness

various physiological processes follow circadian (near-24-h) rhythms regulated by an endogenous pacemaker located in the suprachiasmatic nuclei (SCN) of the anterior hypothalamus. The SCN integrate endogenous and exogenous cues and synchronize the timing of physiological processes to environmental cycles (10). In most species, light is the strongest environmental time cue that resets circadian rhythms. Repeated changes in light-dark cycles, as in the case of shift work, can disrupt the synchronization between individual physiological processes and the synchronization between endogenous and external environmental rhythms, consequently inducing a wide range of health disorders (24, 54).

In addition to phase-resetting effects, light exposure can also induce acute effects such as altered sympathetic regulation (50), rapid suppression of melatonin secretion (65), altered gene expression, and neurobehavioral performance (8, 14). Since a shared neural and molecular pathway mediates both acute and phase-shifting responses to light (41), the same photic stimulus can induce both acute and phase-shifting effects (65).

The effects of light on both acute physiological changes and circadian phase shifts demonstrate a spectral sensitivity such that suppression of melatonin secretion (5, 61), alterations in clock gene expression (46), and changes in sleepiness and alertness (38) are all more sensitive to short-wavelength light (blue) than mid- (green) or long-wavelength (red) light. However, physiological responses to different wavelengths of light may also be dependent on their relative intensities. Light with high correlated color temperature (short-wavelength/blue-enriched light) was more effective in suppressing melatonin at 200 lux intensity compared with light with low correlated color temperature (red-enriched light) (33), whereas at much higher intensities (4,000–5,000 lux), differences between blue-enriched and regular white light on melatonin suppression have not been demonstrated (57, 59).

Recent studies have demonstrated that filtering all wavelengths <530 nm attenuated nocturnal light-induced suppression of melatonin secretion (28, 52). Furthermore, we have demonstrated in rats that filtering a narrow, 10-nm bandwidth of wavelengths between 470 and 480 nm from polychromatic white light prevents melatonin suppression, corticosterone elevation, and central and peripheral clock gene expression alterations induced by 12-h nocturnal light exposure following 12-h daytime light exposure (46). We also showed that filtering the bandwidth between 450 and 460 nm attenuated corticosterone elevation but not melatonin suppression and clock gene expression alterations, suggesting that the range of wavelengths between 470 and 480 nm may induce maximal physiological disruption. Therefore, based on the most effective spectral range (470–480 nm) determined in our animal study, the objective of the present study was to determine whether completely filtering (0% transmission) all short wavelengths <480 nm from polychromatic white light at night could attenuate the acute disruptive effects of light on molecular and endocrine variables and neurobehavioral performance in humans.


Study population.

Healthy individuals (5 females and 7 males, mean age 25.8 ± 0.5 yr) were recruited through local advertisement to participate in the study. All female participants were on oral contraceptives to prevent menstrual cycle variability as they progressed through the 5 consecutive weeks of the study. Exclusion criteria included prior history of shift work, history of sleep disorders (as determined by prior visits to a physician with sleep complaints and/or a sleep clinic visit), history of ocular/vision diseases (as determined by prior visits to an optometrist), color blindness [as determined by the Ishihara Test for Color Blindness (42)], a score greater than 16 on the Centre for Epidemiologic Studies Depression Scale, suggestive of depression (45), and being on any form of medication. The study protocol was approved by the Human Research Ethics Committee of the University Health Network (Toronto, ON, Canada), and all participants provided informed consent.

Overall study design.

The study design was a within-subject design, with all subjects completing each of the five lighting conditions over 5 consecutive weeks. The study was conducted in one season between the months of December and January. Individuals were instructed to maintain a preset 2300–0700 (± 1 h) sleep schedule starting 2 wk before the study and during the 5 wk of the study. Mean bedtime and wake times for the 2 wk before the study were (mean ± SD) 2343 ± 78 min and 0800 ± 90 min, respectively. On the day of the testing, individuals woke up at their habitual times and arrived at the laboratory by 1800 (± 30 min). Upon arrival to the laboratory, all individuals completed their dinner by 1930. Light exposure prior to arrival at the laboratory was not controlled, and it was left up to the participants to maintain their habitual routines. However, the duration of light exposure was fixed for each individual by having their sleep/wake cycles regulated. In addition, all individuals were maintained in the same room for ≥1 h and exposed to the same overhead lighting before the various filtered or nonfiltered light exposure was begun at 2000. Therefore, there was no dark adaptation between wake time and the start of the nocturnal light exposure, in keeping with a usual night shiftwork rotation. Each individual was randomly assigned to complete the five different lighting conditions described below between 2000 and 0800. All individuals received 12 h of continuous filtered or unfiltered light between 2000 and 0800. During each overnight testing session, objective and subjective neuropsychometric tests and saliva samples were collected every 2 h, and buccal swab cell collections were completed every 4 h. Neuropsychometric testing was not conducted under darkness, although all individuals were kept awake between 2000 and 0800 to minimize the confounding effects of sleep deprivation on molecular and endocrine variables studied. No electronic devices such as personal computers, cell phones, or clocks were allowed, and individuals were not exposed to any other light besides the overhead fluorescent lamps. During the overnight session, individuals played board games as a group when they were not testing, and the sequence of games was played in the same order between successive overnight sessions. Individuals received a single isocaloric snack at 0405 after sample collection at 0400. Subjects rinsed their mouths with water immediately after completing the snack. No other food or beverages were permitted between 2000 and 0800 except water, which was consumed only after sample collection and not during the 30 min immediately prior to saliva collection. After completing overnight testing, all individuals returned home by public transportation and were instructed to nap for 4 h between 0900 and 1300 and return to bed by 2300 ± 1 h the same night.

Light exposure.

The lighting conditions included 1) complete darkness, 2) unfiltered fluorescent white light exposure (range 380–730 nm), 3) fluorescent white light as used during unfiltered fluorescent white light exposure, from which all wavelengths <480 nm were completely filtered (0% transmission <480 nm), 4) fluorescent white light as used during unfiltered fluorescent white light exposure, from which all wavelengths <460 nm were completely filtered (0% transmission <460 nm), and 5) fluorescent white light as used during unfiltered fluorescent white light exposure, from which all wavelengths <480 were partially filtered (∼30% transmission). Intensity measurements are shown in Tables 1 and 2 and spectral measurements in Fig. 1. The photon flux and irradiance measurements between 420 and 540 nm integrated over 10-nm bandwidths are provided in Table 3. Unfiltered light was generated using overhead ceiling-mounted fluorescent lamps (T-8 48-in., 32-W, F032/850 5000K Octron Eco Fluorescent Bulb; Osram Sylvania), and the angle of gaze was not held constant to mimic regular work conditions when individuals moved freely. However, individuals were instructed not to look directly at the lights or look directly down. Furthermore, individuals were instructed to look through the center of the lens when using the filter devices. At least two investigators were in the room at all times to ensure compliance.

View this table:
Table 1.

Spectral characteristics of light source

View this table:
Table 2.

Chromaticity characteristics of light source

Fig. 1.

Spectral characteristics of the light source before and after filtering. A: spectral irradiance of the unfiltered light source (dashed line) and transmission curves of the 3 different filters (solid lines) as measured directly along the line of the light source (0° from vertical) at the level of the participant's eye. B: magnification of the spectral irradiance of the unfiltered light source (dashed line) and transmission curves of the 3 different filters (solid lines) in the 420- to 540-nm effective range.

View this table:
Table 3.

Photon flux density and irradiance results of light source in the 420- to 540-nm range

Light was filtered at the level of the eye instead of directly at the source. Identical wrap-around spectacle frames were fitted with polycarbonate lenses that were coated with thin-film Fabry-Perot interference filters (<480- and <460-nm filters were provided by Zircadium, and partial filters were provided by Hoya Lens Canada). The direct application of the thin-film filters onto the lenses and the use of wrap-around frames ensured a single unit of filtering device minimizing stray light incident on the eyes. The partial filter (∼30% transmission <480 nm) was used as a placebo filter for the two active filters. The partial filter was designed to have as similar a color as the active filters to keep participants blinded from knowing when they used the active filters. In addition, the ∼50% reduction in intensity in the 420- to 540-nm range using the partial filter compared with unfiltered light was used as a control for intensity differences between unfiltered light and the <460- and <480-nm complete filters. The 1931 CIE xy chromaticity values for the filtered and unfiltered conditions are given in Table 2. Light intensity and spectral characteristics were measured using a National Institute of Standards and Technology-calibrated spectroradiometer (PS-100; Apogee Instruments).

Hormone analysis.

Melatonin and cortisol were assayed in saliva. All individuals were seated during saliva collection, and eating or drinking was not permitted for 30 min prior to sample collection. Saliva was collected every 2 h in 5-ml tubes (saliva collection device; Alpco Diagnostics) and immediately stored at −20°C until further processing. Two aliquots of each saliva sample were frozen separately for melatonin and cortisol assays, respectively, to minimize repeated freeze-thaw cycles. All frozen saliva samples were defrosted on ice and centrifuged again at 1,500 g for 15 min at 4°C prior to being assayed. Salivary melatonin and cortisol (Alpco Diagnostics and Cayman Chemical, respectively) were measured by enzyme-linked immunosorbent assays per the manufacturers' instructions. All samples were batch processed to cover an equal number of samples from each of the five conditions together in each run of the assay. Melatonin samples were pretreated before assaying per the manufacturer's instructions. The intra-assay precision was 12.6% and the interassay precision 22.9% for the melatonin assay. Since the range of detection was between 1.5 and 25 pg/ml of melatonin (limit of detection was 0.5 pg/ml), saliva samples were diluted 1:4 (within the range of linearity as instructed in the manufacturer's protocol) with autoclaved double-distilled water to detect sample concentrations >25 pg/ml (7). Both low and high controls provided with the kit were assayed with and without dilution during each run of the assay. For the cortisol assay, the intra-assay variation was 1.1% at 4,000 pg/ml and 13.4% at 41.0 pg/ml, and the interassay variation was 6.7% at 4,000 pg/ml and 25.8% at 41.0 pg/ml. The limit of detection was 6.6 pg/ml. Cortisol samples were diluted 1:4 to cover the interindividual variation in cortisol levels during the morning rise in cortisol (35). Both the melatonin and cortisol assays required overnight incubation, and all assay runs were incubated at 4°C for 18 h. Melatonin and cortisol concentrations were determined using a standard curve method, and values that were outside of the linear range (80–20% of %B/B0) of the standard curve were reanalyzed once more with appropriate dilutions and discarded if they were outside of the linear range on the second run as well.


Neuropsychometric tests were carried out every 2 h, except for the vigilance test, which was carried out every 4 h. The neuropsychometric parameters tested were alertness, vigilance, mood, sleepiness, and fatigue. Subjective alertness was measured using a state derivative of the Toronto Hospital Alertness Test (THAT) self-report questionnaire (55). The THAT is a 10-item alertness rating scale that assesses a range of activities such as ability to concentrate, think of new ideas, or focus on a task. Vigilance was measured objectively using the Digit Vigilance Test (DVT) (30). The DVT is designed to measure speed of information processing and errors of omission (failing to respond to a stimulus) and commission (responding to the incorrect stimulus) during rapid visual tracking and accurate selection of target stimuli. Subjective sleepiness was measured using the Stanford Sleepiness Scale (SSS) self-report questionnaire (22). The SSS is a seven-item scale that rates an individual's subjective perception of sleepiness, unlike the THAT scale, which evaluates primarily an individual's subjective perception of task processing. Subjective fatigue was measured using the seven-item Fatigue Scale self-report questionnaire, which is designed to rate an individual's subjective energy levels to complete a task (27). Subjective mood was evaluated using a visual analog scale, with sad/depressed and happy/spirited as the two poles of the spectrum.

Gene expression analysis by quantitative real-time RT-PCR.

PCR primers used (forward 5′ to 3′, reverse 5′ to 3′) were hPer2 (NM_022817: forward, GTC CAC CTC CCT GCA GAC AA; reverse, CTG GTA ATA CTC TTC ATT GGC TTT CA), hBmal1 (NM_001178.4: forward, GAA ATC ATG GAA ATC CAC AGG ATA A; reverse, GAG GCG TAC TCG TGA TGT TCA AT), and hβ-actin (NM_01101.2: forward, GCA TTG TTA CAG GAA GTC CCT TG; reverse, CTA TCA CCT CCC CTG TGT GGA).

Buccal swabs were collected every 4 h using sterile cytology brushes (Puritan Medical Products). The brush tip was separated from the handle and placed in Trizol (Invitrogen Life Technologies) and stored at −80°C for processing at a later time. Total RNA was extracted using Trizol according to the manufacturer's instructions. RNA was quantified by measuring absorbance at 260 nm (μQuant microplate spectrophotometer; Bio-Tek Instruments), and samples were treated with DNaseI (Ambion) according to the manufacturer's instructions. RNA quality was determined by evaluating the absorbance ratio at 260:280 nm, and only samples >1.8 were used for subsequent gene expression analysis. Quantitative real-time RT-PCR was carried out on total RNA (100 ng) using a SuperScript III Platinum SYBR Green One-Step qRT-PCR Kit with 20-μl reaction volumes and 1 μM forward and reverse primer concentrations according to the manufacturer's instructions. The PCR reaction product was generated using the Applied Biosystems 7900 HT Thermal Cycler with 50-μl reaction volumes (Applied Biosystems). Quantitative RT-PCR conditions were as follows: 50°C for 3 min, 95°C for 5 min, and then 40 cycles of 95°C for 15 s, 60°C for 1 min, and 40°C for 1 min. Expression was quantified, using β-actin as a reference, by the comparative threshold cycle (CT) method per the manufacturer's instructions (Applied Biosystems). Dissociation curves for all reactions and gel analysis and sequencing of certain PCR products confirmed gene-specific product amplification. Validation assay to compare PCR efficiency using each of the three primer sets was conducted following the CT slope method per the manufacturer's instructions (Applied Biosystems). A six-log dilution range (100–0.001 ng) of target template was used to determine the CT value for each dilution. A plot of CT vs. log cDNA concentration was constructed, and efficiency was evaluated as the slope of the linear regression curve (hβ-actin: 0.033; hBmal1: 0.042; hPer2: 0.035).

Data analysis.

Data obtained in the hormone assays, gene expression, and neuropsychometric test scores were expressed as means ± SE, unless otherwise stated as means ± SD. Because of large interindividual differences in the ranges over which participants rated themselves on the self-report questionnaires, neuropsychometric data from each lighting condition were transformed to percentage changes from levels observed at 2000 for each individual. Individual percentage changes were then averaged and expressed as a mean value for each of the four lighting conditions. All data were subjected to two-factor (time × spectral condition) analysis of variance followed by Bonferroni post hoc analysis. A two-sided P value of <0.05 was considered to indicate statistical significance. GraphPad Prism Software version 5.0 was used for all statistical analyses.


Effects of spectral modulation on acute endocrine disruption induced by nocturnal light exposure.

The main effects of time, spectral condition, and their interaction on nocturnal melatonin and cortisol levels are provided in Table 4. Consistent with previous reports, exposure to unfiltered light between 2000 and 0800 significantly suppressed melatonin secretion compared with no light exposure during the same period (Fig. 2A). Similar melatonin suppression was observed when the subjects were exposed to <460 nm completely filtered or ∼30% partially filtered <480 nm light (Fig. 2A). In contrast, <480 nm completely filtered light induced significantly less melatonin suppression than the other filtered or unfiltered light exposure conditions (Fig. 2A).

View this table:
Table 4.

Summary of statistical analysis of the effects of time, spectral condition, and their interaction on physiological and neuropsychometric variables

Fig. 2.

Normalization of melatonin and glucocorticoid secretion and clock gene expression under nocturnal light exposure. Nocturnal light exposure induced alteration in melatonin secretion (A), cortisol secretion (B), and Per2 (C) and Bmal1 (D) clock gene expression, which were differentially normalized in a wavelength-dependent manner to the levels observed under no-light exposure. Under each condition, starting from 2000 to 0800, every 2 h (n = 12/time point) saliva was collected for measuring hormone levels by ELISA. Buccal swabs were collected every 4 h, and clock gene expression was measured by quantitative real-time RT-PCR. Data represent means ± SE. Statistical differences between groups were analyzed using 2-way (time × spectral condition) ANOVA followed by Bonferroni post hoc analysis. A: asignificant difference between unfiltered light exposure condition and both dark control (CTRL) and <480-nm completely filtered condition; bsignificant difference between dark CTRL and levels obtained with <460-nm filtered light exposure condition; csignificant difference between unfiltered light exposure condition and <460-nm filtered light exposure condition; dsignificant difference between <480-nm filtered light exposure condition and levels obtained with <460-nm filtered light exposure condition; edifference between <480-nm filtered light exposure condition and levels obtained with placebo-filtered light exposure condition. B: fsignificant difference between dark CTRL and <480- and <460-nm filtered light exposure conditions and levels obtained with placebo and unfiltered light exposure conditions. C: gsignificant difference between dark CTRL and <480-nm filtered light exposure conditions and levels obtained with <460-nm and placebo filtered and unfiltered light exposure conditions. D: hsignificant difference between <480-nm filtered light exposure condition and levels obtained with <460-nm and placebo filtered light exposure conditions; isignificant difference between <480-nm filtered light exposure condition and levels obtained with unfiltered light exposure condition; jsignificant difference between dark CTRL and levels obtained with unfiltered light exposure condition; ksignificant difference between dark CTRL and levels obtained with placebo-filtered light exposure condition; lsignificant difference between dark CTRL and levels obtained with <460-nm filtered light exposure condition.

Unfiltered light exposure significantly increased cortisol secretion at 0200 and 0400 compared with no light exposure (Fig. 2B). Similar to the melatonin response, alterations in cortisol secretion induced by unfiltered light exposure were prevented by filtering wavelengths <480 nm (Fig. 2B). In keeping with our observations in the rat, filtering wavelengths <460 nm also prevented the elevation in cortisol secretion (Fig. 2B). In contrast, the placebo filters did not prevent the disruption in cortisol secretion induced by nocturnal light exposure (Fig. 2, A and B).

Effects of spectral modulation on acute alterations in peripheral clock gene expression induced by nocturnal light exposure.

The main effects of time, spectral condition, and their interaction on Per2 and Bmal1 gene expression are provided in Table 4. Exposure to unfiltered nocturnal light significantly suppressed Per2 (Fig. 2C) and increased Bmal1 gene expression (Fig. 2D) in peripheral cells compared with no light exposure between 2000 and 0800. Filtering wavelengths <480 nm restored peripheral clock gene expression to the dark control levels (Fig. 2, C and D). Filtering wavelengths <460 nm or the use of placebo filters did not prevent the disruption in peripheral clock gene expression (Fig. 2, C and D), producing results similar to exposure to unfiltered nocturnal light.

Effects of spectral modulation on nocturnal neuropsychometric performance.

The main effects of time, spectral condition, and their interaction on neuropsychometric variables are presented in Table 4. Under both unfiltered and filtered lighting conditions, subjective alertness and mood significantly decreased between 2000 and 0800 (Fig. 3A), and total errors on the vigilance test increased (Fig. 3B). However, filtering wavelengths <480 nm were associated with the least reduction in alertness (Fig. 3A) and a significant increase in subjective mood ratings at 0800 (Fig. 3C) compared with all other lighting conditions. Furthermore, there was a significant decrease in total errors in the vigilance task at 0800 (Fig. 3B) by filtering wavelengths <480 nm compared with the other lighting conditions. Subjective sleepiness and fatigue increased between 2000 and 0800 under all lighting conditions (Fig. 3, D and E), and none of the spectral modulation conditions was associated with significant changes in sleepiness and fatigue compared with unfiltered light exposure (Fig. 3, D and E).

Fig. 3.

Effects of filtering short wavelengths on neuropsychometric performance under nocturnal light exposure. Data are presented as the mean of percentage changes from the 2000 baseline score of each individual for each neuropsychometric test under each of the light exposure conditions. A: although exposure to nocturnal light significantly reduced alertness under all 4 lighting conditions, filtering wavelengths <480 nm minimized impairment of subjective alertness at 0800. Similar improvements were observed in vigilance (B) and mood (C). Filtering wavelengths <480 nm did not increase sleepiness (D) or fatigue (E) compared with unfiltered light exposure condition. Starting from 2000 to 0800, every 2 h (n = 12/time point) during unfiltered and filtered light exposure, subjects completed self-report questionnaires assessing alertness, mood, sleepiness, and fatigue, and vigilance was tested every 4 h. Statistical differences between groups were analyzed using 2-way (time × spectral condition) ANOVA followed by Bonferroni post hoc analysis. (asignificant difference under <480-nm filtered light exposure condition compared with unfiltered, <460-nm, and placebo filtered light exposure conditions).


The sensitivity of molecular, endocrine, and neurobehavioral performance to short-wavelength photic stimuli at night has been demonstrated by several groups (5, 8, 38, 48, 61). We and others have demonstrated previously that exposure at night to polychromatic white light from which short-wavelength light (<530 nm) has been filtered prevents melatonin suppression in humans (28, 29, 52). The present study extended these findings, demonstrating that completely filtering short wavelengths <480 nm from nocturnal light not only attenuates the suppression of melatonin but also prevents alterations in cortisol secretion and clock gene expression in humans. Furthermore, completely filtering short wavelengths <460 nm or partially filtering short wavelengths <480 nm by ≤70% did not attenuate the physiological changes induced by nocturnal light exposure. These findings are consistent with our observations in rats (46) that a narrow band of light wavelengths between 470 and 480 nm is responsible for the circadian disruption induced by light exposure at night. The spectral modulation changed both the intensity and spectral characteristics of the light, both of which may independently affect physiological variables. In the present study, there was a 1.6-fold difference in irradiance between the unfiltered condition and the <480-nm completely filtered condition. However, it is unlikely that this difference in intensity can account for the normalization in physiological variables, as observed in the current study. The <460-nm completely filtered condition induced a similar 1.6-fold reduction in irradiance relative to the unfiltered condition but was ineffective in normalizing melatonin suppression and clock gene expression alteration induced by unfiltered light exposure. This suggests that the normalizing effects of the <480-nm completely filtered light are wavelength specific and not due to reduction in intensity caused by spectral modulation. The placebo partial filters reduced transmission by ∼70% below 480 nm and also reduced overall light intensity but did not normalize endocrine, molecular, or neurobehavioral changes associated with unfiltered light exposure, further supporting our conclusion that attenuation of nocturnal light-induced alterations in physiological variables is wavelength specific instead of intensity dependent.

In humans, circadian phase shifting, thermoregulation, peripheral clock gene expression, heart rate, and objective measurements of alertness all demonstrate short-wavelength (blue light) sensitivity similar to melatonin suppression (9, 37). Melatonin suppression is most sensitive to short wavelengths in the 420- to 520-nm range (5, 61), with maximal suppression thought to occur between 446 and 477 nm (5). However, there is some evidence that longer wavelengths (red light) at very high intensity, much higher than the levels used in the present study, can also suppress nocturnal melatonin levels (19). Therefore, filtering a major portion (90%) of this effective spectrum between 420 and 520 nm using <480-nm filters may have attenuated changes in endocrine and molecular variables. However, the <460-nm filters also removed ∼88% of the effective spectrum but did not attenuate the physiological changes, suggesting that the magnitude of physiological response to specific wavelengths within the 420- to 520-nm range may be differentially weighted. Furthermore, the selective range of wavelengths between 460 and 480 nm may have the strongest effect. We have shown previously that filtering the range of wavelengths between 470 and 480 nm from polychromatic white light prevented endocrine and molecular changes in rats (46), supporting our current findings and the hypothesis of differential weighting of physiological response to selective wavelength ranges within the larger 420- to 520-nm range.

Furthermore, the 420- to 520-nm effective range is based on studies using monochromatic pulses administered at night. Several studies have suggested that response to monochromatic and polychromatic light may be different (15, 47). One group has reported that there is a subadditive affect of polychromatic light on melatonin suppression such that a combination of monochromatic pulses induces less melatonin suppression than expected from each of the monochromatic pulses administered individually (15). Another group has shown that exposure to polychromatic light from a high-pressure mercury lamp induced greater melatonin suppression than monochromatic short-wavelength light (49). However, the differential effects in the latter study were dependent on intensity so that the differences between polychromatic and monochromatic light on melatonin suppression were observed at very high and low intensities but not at mid-level intensities. Taken together, the studies suggest that there are differences between physiological responses to monochromatic and polychromatic light, and these differences are likely dependent on the intensity and unique spectral properties of the light. Our current study used fluorescent light, commonly found in homes and offices/workplaces, and also extends the findings from our previous study using halogen lighting, another common source of household lighting.

The physiological response to nocturnal light exposure with or without dark adaptation may not be directly comparable. Prior light exposure history influences the magnitude of the physiological response. Exposure to dim light or darkness prior to nocturnal light exposure increases melatonin suppression, whereas exposure to moderately bright light prior to nocturnal light exposure reduces the magnitude of melatonin suppression (21, 56). The studies used to develop the 420- to 520-nm effective range exposed individuals to light after dark adaptation, whereas in the present study subjects did not have any dark adaptation from wake time in the morning prior to continuous light exposure until the end of the overnight session. This may have reduced sensitivity to the wavelengths remaining in the range of 420–520 nm after <480-nm filtering, and therefore, filtering a major portion of the effective spectrum attenuated physiological changes associated with nocturnal light exposure.

Exposure to unfiltered or placebo-filtered nocturnal light increased cortisol levels compared with no light exposure and <480- and <460-nm filtered light. However, effects of light exposure on cortisol in previous studies are inconsistent, including decreased (32), increased (53), and no change (38) in cortisol levels. Differences in intensity, time of light exposure, and duration of light exposure may all have contributed to the differences between the present study and prior studies. In a more recent study, it was demonstrated that bright light exposure during the rising or descending phase of cortisol secretion (that is, when cortisol levels are high) caused a significant reduction in cortisol levels (26). In our study, cortisol levels were elevated by light exposure during the quiescent period of cortisol secretion (when cortisol levels are low); however, light exposure also started at a different phase compared with previous studies, and the difference in the relative phase of light exposure may have played an important role in mediating the changes in cortisol levels observed in the current study. Nonetheless, the current and prior studies suggest that cortisol secretion may be acutely altered by nocturnal light exposure.

The present study suggests that the central pacemaker likely mediates these changes in cortisol levels, since the adrenal gland is not directly light sensitive in humans and, therefore, was not likely directly affected by spectral modulation. A direct sympathetic connection via the splanchic nerve between the SCN and the adrenal gland has been demonstrated in animals (6, 23). Furthermore, a circadian gating mechanism regulated by an adrenal-cortical molecular clock has been shown to play a role in the circadian rhythmicity of glucocorticoid secretion (43). We and others have demonstrated previously that exposure to nocturnal light alters clock gene expression in the adrenal gland in rodents (23, 46), suggesting that the gating mechanism may be altered by light exposure and affect glucocorticoid levels. In addition, in our previous study, light-induced elevation of corticosterone was independent of changes in circulating ACTH (46). Therefore, light exposure and spectral modulation likely affect glucocorticoid levels via a neural pathway, and the molecular gating mechanism mediates the phase-dependent effects of light exposure on circulating glucocorticoid levels.

Melatonin, cortisol, and clock genes regulate critical physiological processes. Melatonin has been demonstrated to be an endocrine time cue, an oncostatic agent, and an antioxidant (64). Nocturnal suppression of melatonin secretion by light has been proposed to be one of the principal mechanisms for the increased risk of breast cancer in shiftworkers (3). Glucocorticoids regulate a spectrum of physiological functions ranging from stress responses to reproduction (44). Elevated glucocorticoid levels are associated with numerous health issues, including cardiovascular disease, insulin resistance and metabolic syndrome, developmental disorders, and even deleterious fetal effects during pregnancy (62). More than 1,500 genes for more than 1,400 physiological functions have a circadian expression pattern in murine peripheral organs (60). The circadian clock has been shown to function as a tumor suppressor at the systemic, cellular, and molecular levels in vivo (16). Exposure to continuous nocturnal lighting increases the proliferation rate of human breast cancer xenografts in nude rats (4). Recently, several studies have demonstrated an abnormality of clock gene expression in breast cancer, particularly a downregulation of Per2 expression (11). Prostate cancer has also been associated with shift work and circadian disruption induced by shift work (34). A significant association between single nucleotide polymorphisms in core clock genes, including Per1/2/3 and Bmal1, with overall risk or risk of aggressive disease has been reported (66). Furthermore, a downregulation of Per2 and upregulation of Bmal1 was seen in prostate cancer cells compared with nonmalignant prostate epithelial cells (25). Interestingly, prostate cancer cell lines treated with melatonin had a reversal in this clock gene expression rhythm similar to that observed under darkness or under <480-nm filtered light conditions in our current study, and a resultant inhibition of cell growth and viability was observed (34). Filtering wavelengths <480 nm from nocturnal lighting normalizes melatonin and glucocorticoid levels as well as peripheral clock gene expression and may be a simple and effective method of preventing physiological disruption induced by nocturnal light exposure. However, future studies are warranted to examine whether such an approach can be used to manipulate circadian rhythms in shift workers so that sleep and work episodes do not occur at incorrect circadian phases and to study the effects of such manipulation on performance and worker well-being. Appropriately timing light and darkness episodes has previously been proposed as a countermeasure for shiftwork-induced circadian misalignment (12, 58). Reducing the intensity of light using dark glasses during the morning commute has been shown to be beneficial in some studies. The use of spectral modulation provides an additional method of modulating the effects of light on physiological variables, and this approach may facilitate the use of spectral modulation instead of intensity modulation in conditions when light intensity cannot be decreased.

In addition to endocrine and molecular changes, nocturnal light exposure also induces behavioral alterations in alertness, vigilance, and mood. Light stimulates the ascending arousal system, and eventually the cerebral cortex, to enhance alertness (51). Exposure to short-wavelength (<500 nm) monochromatic pulses increased alertness more than longer wavelengths (>500 nm) (9, 38). In contrast, in the present study, completely filtering wavelengths <480 nm did not reduce alertness or vigilance compared with unfiltered light exposure. Furthermore, completely filtering wavelengths <480 nm from nocturnal light was associated with significantly less reduction in subjective alertness and with reduced errors in an objective test of vigilance. In addition, mood was improved at 0800 compared with the other lighting conditions. Light exposure can acutely increase alertness levels. In the present study, the reduced impairment in neurobehavioral performance under the <480-nm completely filtered condition was observed only at 0800 and not at the other times. This observation suggests that the improvement in performance was not an acute effect of the <480-nm filtered light but rather that time-dependent physiological changes mediated this neurobehavioral response.

Furthermore, neurobehavioral performance is inversely related to durations of wakefulness, and the extended wake prior to nocturnal assessment of performance may have contributed to the lack of difference in alertness and vigilance observed between the filtered and unfiltered conditions. Neurobehavioral performance was not measured until 2000, whereas the subjects were awake since ∼0900, and as a result the extended wake period prior to nocturnal light exposure being started may have already decreased baseline (2000) alertness levels so that filtering short wavelengths was not associated with further impairment in neurobehavioral performance. In addition, it is possible that the tests used in the current study were not sufficiently sensitive to detect subtle changes in neurobehavioral performance. However, we believe this suggestion is unlikely because similar subjective sleepiness scales and vigilance tasks have been used previously to measure changes in neurobehavioral performance associated with light exposure (9, 38).

Subjective alertness and mood are under circadian regulation, and different magnitudes of circadian phase resetting under the different lighting conditions may have also contributed to the differences in neurobehavioral performance. Under prolonged/overnight wakefulness, mood and performance decrease to a minimum close to the nadir of core body temperature and increase thereafter (13). In entrained individuals, the core body temperature nadir normally corresponds to early morning (0400–0600), and mood improvements mediated by the circadian system can be observed thereafter (39, 40). These times correspond to the time when alertness, mood, and vigilance were less impaired under the <480-nm completely filtered condition. The current findings suggest that completely filtering short wavelengths <480 nm may have attenuated circadian disruption and may have induced a more robust improvement in neurobehavioral performance closer to habitual wake time, as described in previous studies (39, 40).

Although endogenous phase was not directly measured in the present study, the endocrine and molecular parameters examined in this study have been shown to be reliable circadian phase markers. Melatonin levels are directly regulated by the SCN, and several studies have validated the use of peripheral melatonin levels to reliably assess circadian phase (36). In addition to melatonin, glucocorticoid rhythms and peripheral clock gene expression rhythms are also under circadian regulation and serve as additional markers for circadian phase (31). Results from the present study, demonstrating that completely filtering wavelengths <480 nm prevented shifting of all three circadian phase markers, suggest that this approach may attenuate circadian phase alterations. However, future studies designed to directly assess circadian resetting are required to investigate whether spectral modulation can attenuate circadian phase resetting.

In mammals, the photic stimulus for circadian response is captured exclusively by the eyes and transduced directly to the SCN via a dedicated neural pathway, the retinohypothalamic tract (2). Circadian photoreception is mediated by the classical visual photoreceptors rods and cones and intrinsically photosensitive retinal ganglion cells (ipRGCs) (2, 18). The relative role of each of the photoreceptors in mediating circadian photic responses is most likely dependent on several parameters, including intensity, spectral characteristics, and duration of exposure as well as prior light exposure history (17, 56, 63). Both rod and cone photoreceptors connect to ipRGCs (2, 20). The intensity and duration of light exposure likely influence the degree of input from each of the photoreceptor systems. A recent study demonstrated that cone photoreceptors play a substantial role in nonvisual responses at the beginning of a light exposure and at low irradiances, whereas melanopsin appears to be the primary circadian photopigment in response to long-duration light exposure and at high irradiances (17). It is likely that by the time the individuals started their nocturnal light exposure episode in the present study, the cone photoreceptor-mediated nonvisual response was desensitized, which may explain the efficacy of <480-nm completely filtered light in attenuating endocrine and molecular changes. The effective range filtered by the <480-nm optical filter lenses corresponds to the absorption spectrum of melanopsin, suggesting that filtering short wavelengths attenuates ipRGC activation and thereby attenuates the signal to SCN and other neural centers responsible for mediating acute and circadian responses to light.

A recent study in mice suggested that, even under high irradiances and prolonged light exposure, rods can also signal to ipRGCs via cone bipolar cells and mediate non-image-forming photic responses (1). These observations suggest that under the conditions of the present study, even if filtering short wavelengths <480 nm attenuated direct activation of ipRGCs, the rods may still continue signaling to the ipRGCs. Further work is required to confirm whether ipRGCs are active under the <480-nm condition due to differential activation of cones and cone bipolar cells and the signaling pathway between rods and cones via gap junction-mediated electric coupling. The underlying mechanism for attenuating nocturnal light exposure-induced alteration in physiological variables by spectral modulation requires further investigation. However, this line of investigation may reveal new approaches to counteract some of the adverse affects of aberrant light exposure, which is becoming alarmingly more prevalent in our modern 24-h society. These data also highlight the need to conduct observations on the effects of light in field studies under naturalistic settings in parallel with controlled laboratory studies since the outcomes may be quite different and not directly comparable due to underlying physiological changes associated with each condition.


This study was funded by Ontario Centers of Excellence Grant No. OCE-IA90387. S. A. Rahman was funded by a Government of Ontario/Pharmacia Canada/Genesis Research Foundation/OBGYN Graduate Scholarship in Science and Technology at the University of Toronto, Faculty of Medicine, and the Frederick Banting and Charles Best Canada Graduate Scholarships Doctoral Award from the Canadian Institutes of Health Research.


S. A. Rahman and R. F. Casper have filed for a patent of prevention of circadian rhythm disruption by using optical filters. S. A. Rahman, C. M. Shapiro, and R. F. Casper own shares in Zircadium. S. Marcu and T. J. Brown have nothing to declare.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
View Abstract