Short Blue Light Pulses (30 Min) in the Morning Support a Sleep-Advancing Protocol in a Home Setting

Abstract

Many people in our modern civilized society sleep later on free days compared to work days. This discrepancy in sleep timing will lead to so-called ‘social jetlag’ on work days with negative consequences for performance and health. Light therapy in the morning is often proposed as the most effective method to advance the circadian rhythm and sleep phase. However, most studies focus on direct effects on the circadian system and not on posttreatment effects on sleep phase and sleep integrity. In this placebo-controlled home study we investigated if blue light, rather than amber light therapy, can phase shift the sleep phase along with the circadian rhythm with preservation of sleep integrity and performance. We selected 42 participants who suffered from ‘social jetlag’ on workdays. Participants were randomly assigned to either high-intensity blue light exposure or amber light exposure (placebo) with similar photopic illuminance. The protocol consisted of 14 baseline days without sleep restrictions, 9 treatment days with either 30-min blue light pulses or 30-min amber light pulses in the morning along with a sleep advancing scheme and 7 posttreatment days without sleep restrictions. Melatonin samples were taken at days 1, 7, 14 (baseline), day 23 (effect treatment), and day 30 (posttreatment). Light exposure was recorded continuously. Sleep was monitored through actigraphy. Performance was measured with a reaction time task. As expected, the phase advance of the melatonin rhythm from day 14 to day 23 was significantly larger in the blue light exposure group, compared to the amber light group (84 min ± 51 (SD) and 48 min ± 47 (SD) respectively; t₃₆ = 2.23, p < 0.05). Wake-up time during the posttreatment days was slightly earlier compared to baseline in the blue light group compared to slightly later in the amber light group (–21 min ± 33 (SD) and +12 min ± 33 (SD) respectively; F_1,35 = 9.20, p < 0.01). The number of sleep bouts was significantly higher in the amber light group compared to the blue light group during sleep in the treatment period (F_1,32= 4.40, p < 0.05). Performance was significantly worse compared to baseline at all times during (F_1,13 = 10.1, p < 0.01) and after amber light treatment (F_1,13 = 17.1, p < 0.01), while only in the morning during posttreatment in the blue light condition (F_1,10= 9.8, p < 0.05). The data support the conclusion that blue light was able to compensate for the sleep integrity reduction and to a large extent for the performance decrement that was observed in the amber light condition, both probably as a consequence of the advancing sleep schedule. This study shows that blue light therapy in the morning, applied in a home setting, supports a sleep advancing protocol by phase advancing the circadian rhythm as well as sleep timing.

Keywords

blue light light therapy sleep performance circadian system melatonin human

Sleeping out of phase with the external light-dark cycle, especially sleeping later, has become a common habit in our civilized society. Despite the fact that workdays and school days typically start early in the day, many people go to bed late and try to compensate for the sleep deficit by extending their sleep on free days. From a large database of 55,000 people, it was indeed calculated that the average wake-up time on free days was 2 hours later than on workdays (Roenneberg et al. 2007). In addition, a recent analysis of a Dutch survey on sleep habits found that 74% of the adult population sleeps later on free days than on workdays (n = 8074; Gordijn, unpublished data, 2006). As a result of the discrepancy in sleep timing on workdays and free days, many adults experience a so-called weekly “social jetlag” (Wittman et al., 2006).

As with a “normal” jetlag resulting from traveling to different time zones, sleeping at the wrong internal phase can disturb bodily rhythms (Rajaratnam & Arendt, 2001; Rüger & Scheer, 2009; Baron & Reid, 2014). This may lead to stress and health problems. Indeed, correlational studies have indicated that if the amount of social jetlag is higher, people have a higher risk of living an unhealthy lifestyle, being overweight, and experiencing stress-related problems and even depression (Wittman et al., 2006, 2010; Roenneberg et al., 2012; Merikanto et al., 2013; Touitou, 2013). Furthermore, people with a late sleep phase are highly represented in the population (for 60%, the midpoint of sleep on free days is at 0430 h or later; Roenneberg et al., 2007), and they have a higher chance for a larger social jetlag (Wittman et al., 2006). A large number of people in the population may therefore benefit from a correction of their timing of sleep.

Light therapy in the morning is often proposed as the most effective method to advance a late sleep phase (Barion & Zee, 2007; Morgenthaler et al., 2007; Bjorvatn & Pallesen, 2009). Indeed, in a clinical setting, light exposure in the morning has been shown to be effective in advancing circadian rhythms of patients with delayed sleep-wake phase disorder (DSPD). Home therapy studies for the treatment of DSPD used light exposure of 2 h in the very early morning (Rosenthal et al., 1990) or directly after habitual wake up (Lack et al., 2007). More recent DSPD therapy studies examined the effect of shorter light pulses (30 min to 2 h, Gradisar et al., 2011; 30 to 45 min, Saxvig et al., 2014), starting directly after waking up, along with a sleep-advancing protocol with strict sleep schedules. Light therapy itself does not always seem to be crucial for the therapeutic effect: a recent study found that even a sleep-advancing protocol alone can phase advance the circadian rhythm of late sleepers (Sharkey et al., 2011).

The experimental setup to test morning light therapy is highly variable in the aforementioned DSPD studies and not well studied in late chronotypes in general. Furthermore, there is no generally accepted optimal protocol that can be applied at home (reviewed in Schroeder & Colwell, 2013). Another issue is the lack of studies with good control conditions, since placebo lamps do not really exist. Last but not least, although the goal of light therapy in late chronotypes is to shift sleep and improve daytime functioning, the focus of most studies has been on shifting melatonin as an output marker. The posttreatment effect on sleep phase itself has hardly been studied, and daytime functioning after light treatment has hardly been reported.

The rationale of using light therapy to shift sleep is based on the theory that it can influence the synchronization of sleep by adjusting process C of the “2-process model” of sleep regulation (Borbély, 1982; Daan et al., 1984). In humans and many other organisms, process C is highly controlled by the central master clock in the brain, the suprachiasmatic nucleus (SCN; reviewed in Moore, 2007). There is accumulating evidence that light at least influences process C as the SCN is predominately synchronized by light (reviewed in Meijer and Schwartz, 2003, and in Moore, 2007).

Phase shifts of the SCN can be approached by measuring shifts of physiological circadian rhythms, of which the shift of melatonin onset measured in very low light intensities <10 lux (dim light melatonin onset [DLMO]) is currently known as the most reliable marker (Lewy and Sack, 1989; Lewy et al., 1999; Klerman et al., 2002; Revell et al., 2005b; Arendt, 2005; Van Someren & Nagtegaal, 2007; Benloucif et al., 2008). Many studies have found that light in the morning induces a phase advance of DLMO (Dijk et al., 1989; Buresova et al., 1991; Samková et al., 1997; Gordijn et al., 1999; Danilenko et al., 2000; Revell et al., 2005a) whereas light in the evening induces a phase delay (Kräuchi et al., 1997; Gordijn et al., 1999; Zeitzer et al., 2000; Gronfier et al., 2004). These time-dependent effects of light are summarized in a phase-response curve (PRC) as extensively analyzed in rodents by Daan and Pittendrigh (1976). PRC studies have found an optimum for the phase-advancing effects of long and short (blue) light pulses when starting the light exposure 9 h after DLMO (Khalsa et al., 2003; St. Hilaire et al., 2012; Rüger et al., 2013). It has also been shown that the beginning of a light pulse is most important for the phase-advancing effect (Beersma et al., 2009), which supports the idea that light pulses given approximately 9 h after DLMO will be effective in phase-advancing circadian rhythms.

Besides the timing of light exposure, there are other important factors that influence the effectiveness of light on the SCN. Important factors are intensity (Zeitzer et al., 2005), spectral composition (Berson et al., 2002; Lockley et al., 2003), and duration (Chang et al., 2012) of the light exposure. To optimize the final treatment success (and thereby the compliance to light therapy), it is important to find the optimal combination of these factors.

We tested the effect of duration in a pilot study (Geerdink et al., 2012), and we found that high-intensity short blue light pulses of 30 min timed 9 h after DLMO on 3 consecutive days were able to advance the melatonin rhythm by 50 min. Importantly, we found that the effect on the melatonin rhythm of those three 30-min blue light pulses was not significantly different from the effect of three 60-min blue light pulses. Therefore, we concluded that we could use these short 30-min high-intensity blue light pulses to phase advance the circadian rhythm in a home-setting protocol.

To test if short 30-min high-intensity blue light pulses cannot only be used for phase advancing the melatonin rhythm but also for treating a late sleep phase, the main question of the present study was whether these light pulses were able to advance sleep timing with preservation of sleep integrity and daily performance. We tested the effects of blue morning light pulses against amber morning light pulses, which we considered to serve as a “placebo” condition.

Materials and Methods

Participants

Participants were recruited from the northern population of the Netherlands, via newspaper advertisements, posters at the faculties of the University of Groningen, and Internet advertisements. People were invited to respond to the advertisement only if they felt they were suffering from sleeping later on free days compared with workdays and if they really wanted to change this. Therefore, 2 of the selection criteria were a calculated social jetlag (Wittman et al., 2006) of more than 1 h and a calculated shorter sleep duration during workdays compared with free days (calculated from the Munich Chronotype Questionnaire; Roenneberg et al., 2003). Other selection criteria were being healthy, between 18 and 65 years old, not suffering from sleep problems other than late sleep, not color blind or visually impaired, not consuming too many cigarettes (more than 10 per week), not consuming too many drinks with caffeine (more than 8 per day) or alcohol (more than 10 per week), not working in night shifts, not having been in other time zones in the preceding 3 months, and not being severe dysphoric or depressed (Beck Depression Inventory II rating <16; Beck et al., 1996, 2002). Female subjects were included only if they were using hormonal contraceptives. All subjects were Dutch citizens from birth and fluent in Dutch. All subjects gave written informed consent and were compensated for their participation. The experimental protocol was approved by the Medical Ethics Committee of the University Medical Center Groningen. We selected 42 participants for this study (23 women, 19 men) with a mean age of 21.4 years (SD ±6.5). For a detailed description of subject characteristics, see Table 1.

Table 1.

General group statistics.

Variable	Condition	n	Mean	SD	SEM	p Value
Gender: 1 = female, 0 = male	Blue	18	0.5	0.5	0.1	0.67
Gender: 1 = female, 0 = male	Amber	21	0.6	0.5	0.1	0.67
Age	Blue	18	23.1	8.6	2	0.31
Age	Amber	21	21	3.4	0.7	0.31
Length	Blue	18	1.8	0.1	0	0.58
Length	Amber	21	1.8	0.1	0	0.58
Weight	Blue	18	67.1	11.3	2.7	0.53
Weight	Amber	21	69.2	9.3	2	0.53
Cigarettes per day	Blue	18	0.7	1.8	0.4	0.45
Cigarettes per day	Amber	21	1.6	4.4	1	0.45
Alcohol consumption per week	Blue	18	4.5	3.4	0.8	0.56
Alcohol consumption per week	Amber	21	5.2	4.2	0.9	0.56
Coffee per day	Blue	18	1.5	1.4	0.3	0.78
Coffee per day	Amber	21	1.3	1.5	0.3	0.78
Back Depression Inventory score	Blue	18	3.8	3.8	0.9	0.58
Back Depression Inventory score	Amber	21	4.6	4.4	1	0.58
Mid-sleep workdays	Blue	18	4.3	0.9	0.2	0.56
Mid-sleep workdays	Amber	21	4.1	0.7	0.2	0.56
Mid-sleep free days	Blue	18	6.5	0.9	0.2	0.88
Mid-sleep free days	Amber	21	6.5	1	0.2	0.88
Dim light melatonin onset, baseline	Blue	18	23.1	1	0.2	0.89
Dim light melatonin onset, baseline	Amber	21	23	1.1	0.2	0.89

Experimental Design

All participants were quasi-randomly assigned (balancing for age, gender, chronotype) to a sleep-advancing protocol supported by either high-intensity blue light or amber light (control) with similar photopic illuminance (see the “Light Treatment” section for details). None of the subject characteristics was significantly different between groups (see Table 1).

The protocol consisted of 30 days in total and was completely performed at home. All participants started with 14 baseline days with no sleep instructions. This was followed by 9 intervention days with either 30-min blue light pulses or 30-min amber light pulses in the morning along with a sleep-advancing scheme. The scheme was based on the habitual sleep offset of days 4 to 10 in the baseline period and consisted of 1 h advance every 3 days. This scheme was followed by 7 posttreatment days without sleeping instructions and no use of light treatment devices (see Figure 1). The study was carried out from June 2011 to April 2012. Participants were evenly divided over conditions per season.

Figure 1.

Scheme of the experimental design. Consecutive days are indicated by rows.

To quantify effects on the melatonin rhythm, saliva samples were taken on the evenings of days 1 and 7 (prebaseline), day 14 (baseline), day 23 (effect intervention), and day 30 (post treatment effect). Saliva collection started 5 h prior to habitual bedtime. Subjects were asked to stay in dim light (intensity at which they can just read, <10 lux), starting 1 h prior to the first sample. In addition, they were asked to wear blue light–blocking glasses (Uvex S1933X; Skyper Safety Eyewear). On days 14, 23, and 30, participants were instructed to not only collect saliva during the evening but also from habitual sleep onset until waking up (1 sample per 2 h). In addition, 3 additional samples (1 per hour) were taken after habitual sleep offset in the morning on days 14, 23, and 30. This resulted in 13 samples in total on these days. Saliva was collected with cotton swabs (Salivette, Sarstedt BV, Etten-Leur, the Netherlands). Melatonin concentration was assessed by radioimmunoassay (RK-DSM; Bühlmann Laboratories, Alere Health, Tilburg, the Netherlands). The limit of detection was 0.3 pg/mL. Intra-assay variation was 15.9% at a low melatonin concentration (mean = 2.0 pg/mL, n = 17) and 13.1% at a high melatonin concentration (mean = 24.5 pg/mL, n = 15). Interassay variation was 13.1% at low melatonin concentration (mean = 2.0 pg/mL, n = 16) and 15.0% at high melatonin concentration (mean = 21.4 pg/mL, n = 16). A bimodal skewed baseline cosine function was fitted through the individual data points of the baseline curve on day 14 (Van Someren & Nagtegaal, 2007). The maximum value of the fitted baseline curve of each individual was set at 100%. DLMO was calculated as the moment at which the melatonin rhythm crosses the 25% value of this fitted maximum by linearly interpolating the raw values preceding and following the 25% value.

During the whole protocol, sleep was monitored with the Actiwatch Spectrum monitor (Philips Respironics Inc., Murrysville, PA). Participants completed sleep diaries including sleepiness measurements 5 and 30 min after waking up (Karolinska Sleepiness Scale [KSS]). Cognitive performance was measured with a 3-min reaction time task (psychomotor vigilance task) on a handheld minicomputer (HP Ipaq114) on the last 3 baseline days (days 12, 13, and 14), on the middle treatment days (days 19, 20, and 21), and on 3 posttreatment days (days 27, 28, and 29). Performance was measured at 4 time points during the day, namely, during the morning (30 min after wake up), at noon (1200 h), at late afternoon (1800 h), and during the evening (2300 h). Because of large differences in sleep times between free days and workdays, which is typical for this population, comparisons of sleepiness and performance between light conditions were limited to workdays only.

Participants received an instruction book with a detailed scheme regarding what they had to do on specific days. In addition, participants received text messages (or phone calls) when they needed to make appointments with the researchers (day 1, day 10, day 23, day 30), and they received a text message reminder the day before starting with light therapy (day 14).

Compliance

Before analyzing the data, we carefully checked compliance of the subjects by inspection of the light data and the actigraphy data during the treatment days. Subjects wore glasses or a hairband with a light sensor, Daysimeter (Lumen Tech Innovations, New York, NY), which measured spectral characteristics of light at the eye level. The Daysimeter employs 2 photosensors that separately measure a photopic (visual) and a short-wavelength (blue) response to optical radiation. This blue light sensor has peak sensitivity at short wavelengths (460 nm, 80 nm full-width-half-maximum bandpass), near the peak of the spectral sensitivity of the human circadian system (Rea et al., 2005, 2010). The device also records activity (movement). These Daysimeters were used to monitor the compliance to the instructions of when to use the experimental light and were also used to measure the exposure to natural and artificial light sources during the day. A program in Matlab (Mathworks, version R2014b) was developed to detect if there were differences with respect to light intensity during the mornings of every experimental day. Compliance of the participants was checked according to the timing of light exposure with the lamps. Subjects were excluded from the database if they started the light exposure more than 30 min too early or too late on average (over the 9 treatment days). We excluded 2 participants (blue light condition) because of this criterion. In 5 of the 42 subjects, technical failures prohibited the use of the light exposure data for this purpose. In those cases, we used the actigraphy data instead, by inspecting the sleep offset times during the treatment days. We excluded subjects who did not show any activity for 3 or more days within 30 min after the time point at which they were instructed to wake up. We had to exclude 1 extra subject (amber light condition) because of this criterion. Eventually the data of 39 participants (blue n = 18, amber n = 21) have been used for analyses. A check for sleep timing compliance during the 9 treatment days revealed that subjects indeed shifted their sleep offset by 2 h in both groups: habitual sleep offset was on average 0912 h (0:59) in the blue group and 0900 h (1:18) in the amber group during days 4 to 10 at baseline. Average wake-up time over the last 3 days of the treatment period was 0708 h (SEM 0:59) in the blue group and 0706 h (SEM 1:00) in the amber group.

Light Treatment

In our previous pilot study, we found that 3 optimally timed light pulses of 30-min high-intensity blue light were able to shift the melatonin rhythm for about 50 min (Geerdink et al., 2012). Therefore, in this study, we used 30 min of light therapy, which is easy to comply within a home situation. Subjects were instructed to expose themselves to either 30 min of high-intensity blue light (Philips GoLite BLU HF3320, peak transmission at 470 nm, intensity at the cornea 2306 melanopic-lux [Lucas et al., 2014], 300 photopic lux, 3.0 W/m²) at a distance of 40 cm and angle of exposure of 45°) or to control 30 min of amber light with similar photopic illuminance (adapted Philips GoLite HF3320, peak transmission at 590 nm, intensity at the cornea 70 melanopic-lux, 250 photopic lux, 0.5 W/m²). Timing of the light pulse was based on the subjects’ habitual sleep schedule during baseline days 4 to 10 measured by actigraphy and on the subjects’ preferred timing of sleep after the treatment. The last 3 light pulses were set at the preferred sleep offset, and the timing of the light exposure was set 1 h earlier every 3 days. To get sufficient sleep, sleep timing was scheduled to shift along with the light therapy (see also Figure 1). On average, subjects started with the first light pulse 7.9 h (±0.6 h) after habitual sleep onset. At the end of the study, the actual timing according to DLMO on the night prior to the first light pulse could be calculated. Light was administered on average 10.4 h, 9.4 h, and 8.4 h (±1.2 h) after baseline DLMO on, respectively, the first days (days 15-17), middle days (days 18-20), and last days (days 21-23) of the treatment period (which is close to the optimal timing for phase advances, see the opening section of the article). Based on the middle 3 light pulses, light exposure timing with respect to baseline DLMO was not significantly different between the 2 groups (t₃₆ = 0.22, p = 0.83), namely, on average 9.5 h (±0.8 h) in the blue condition and on average 9.3 h (±0.9 h) in the amber condition.

Environmental Light Measurements

Before analyzing any result, we first checked if the daily environmental lighting conditions were not significantly different between conditions. All light measurements were monitored with the Daysimeter light sensor (see the “Compliance” section for more information). This was done since not only morning light is of influence on our circadian system; it is the complete pattern of light exposure during the whole day that is responsible for phase-shifting effects (Appleman et al., 2013; Burgess, 2013b). Especially low light levels in the evening are essential for an optimal phase advance.

Light exposure levels of the subjects in the 2 groups were compared to check for possible behavioral differences that could result in differences in light exposure other than the actual light treatment. Light levels were calculated on the treatment days during the morning (0600-1200 h), afternoon (1200-1800 h), and evening (1800-2400 h). Due to technical failures, light data of 27 subjects (13 in the blue group, 14 in the amber group) were analyzed.

Sleep Analyses

During the whole protocol, sleep was monitored with the Actiwatch Spectrum monitor (Philips Respironics Inc.) and analyzed afterward with the Respironics Actiware 5 software. Objective estimates of sleep quantity and sleep integrity were defined by the following parameters: sleep onset latency (SOL), sleep duration, minutes awake after sleep onset (WASO), number of sleep bouts (as a measure for sleep consolidation), and sleep efficiency (defined as [sleep time – SOL- WASO]/sleep time × 100%). It should be noted that sleep integrity as measured with actimetry is different and never as accurate as sleep integrity measured with polysomnography (reviewed in Ancoli-Israel et al., 2003). Especially SOL is often underestimated with actigraphy (Chae et al., 2009; O’Hare et al., 2015). Still, as differences within individuals are compared, it is possible to evaluate the changes in relation to these sleep parameters over the different periods.

Statistical Analyses

For all variables, group means (described as mean ± standard deviation) were tested against each other. For differences between conditions concerning melatonin concentrations (DLMO, area under the curve) independent sample t tests were used. Paired sample t tests were used to test the differences of the melatonin concentrations between baseline and the end of treatment and between baseline and after posttreatment per condition. If more parameters needed to be tested, like with the sleep analysis, a repeated-measures analysis of variance (ANOVA) for each parameter was used with 1 within-subjects factor for period (3 levels; baseline, treatment, posttreatment) and 1 between-subjects factor for condition. If the interaction of condition × period was significant, a simple contrasts test (equivalent to a paired sample t test) was included to the ANOVA to see which period was different from the baseline period. For the performance analysis, an extra factor for time of day (4 levels; morning, noon, afternoon, evening) was included. The significance level was defined as p = 0.05. All parameters were tested 2-tailed.

Results

Environmental Light Levels during the Morning, Afternoon, and Evening

No significant differences were found for the average photopic illuminance levels between conditions during the morning (blue group; 519 ± 324 lux/min, amber group; 548 ± 307 lux/min; t₂₅ = −0.25, p = 0.81), nor during the afternoon (blue group; 1664 ± 977 lux/min, amber group; 1425 ± 872 lux/min; t₂₅ = 0.67, p = 0.51), nor during the evening (blue group; 96 ± 113 lux/min, amber group; 132 ± 110 lux/min; t₂₅ = −0.85, p = 0.40).

With respect to the blue light illuminance levels, a significant difference between conditions was found during the morning. The blue morning light exposure measured by the Daysimeters (see the “Materials and Methods” section) was found to be significantly higher in the blue light therapy group (blue group, 1485 ± 697 blue lux/min; amber group, 965 ± 539 blue lux/min; t₂₅ = 2.2, p < 0.05), while not significantly different between the 2 conditions during the afternoon (blue group, 3208 ± 1854 blue lux/min; amber group, 2581 ± 1626 blue lux/min; t₂₅ = 0.94, p = 0.36) nor during the evening (blue group, 163 ± 224 blue lux/min; amber group, 226 ± 215 blue lux/min; t₂₅ = −0.74, p = 0.47). This difference in blue light exposure in the morning between groups is explained by the light treatment itself.

Effects on the Melatonin Rhythm: DLMO

The average DLMO times on the days that the saliva samples were taken are shown in Figure 2A. Day 14 was the day before the light exposure period started and was considered as the baseline day. Although, on average, a slight delay for both groups from days 1 to 7 to 14 seems to occur (10 ± 51 min in the blue light condition and 17 ± 60 min in the amber light condition), the DLMO times on day 14 were not significantly different from day 1 and day 7 for both conditions (blue, F_1,2 = 0.18, p = 0.83; amber, F_1,2 = 1.16, p = 0.33).

Figure 2.

(A) The average (–SD) dim light melatonin onset (DLMO) times on different days of the protocol. Average DLMO times on day 23 differed significantly from the average DLMO times on day 14, *p < 0.001. (B) The average (+SD) DLMO shifts from baseline. Treat-Blue = DLMO shift from baseline after treatment in the blue light condition; Treat-Amber = DLMO shift from baseline after treatment in the amber light condition (*p < 0.05); Post-Blue= DLMO shift from baseline after the posttreatment period in the blue light condition; Post-Amber= DLMO shift from baseline after the posttreatment period in the amber light condition.

Compared with day 14, a significant phase advance was observed on day 23 (after 9 days of light exposure) in both groups (Figure 2A and B). DLMO shifted on average by 48 min (±47; t₁₉ = 4.56, p < 0.001) in the group that received amber light and by 84 min (±51; t₁₇ = 6.95, p < 0.001) in the group that received blue light. The shift in the blue light group was significantly larger than the shift in the amber light group (t₃₆ = 2.23, p < 0.05; Figure 2B).

At day 30, DLMO was delayed (shifted back) in both groups compared with day 23 (Figure 2A and B; 70 min [±68] in the blue light condition, t₁₇ = −4.26, p < 0.01; 48 min [±37] in the amber light condition, t₁₉ = −6.01, p < 0.001). The DLMO shifts from day 23 to day 30 were not significantly different from each other (t₃₆ = −1.26, p = 0.22). DLMO at day 30 was not significantly different from baseline, neither for the blue group (t₁₇ = 0.90, p = 0.38) nor for the amber group (t₁₉ = 0.06, p = 0.95; Figure 2B).

Effects on the Melatonin Rhythm: Area under the Curve

The levels of the melatonin concentrations were expressed as the percentage of the fitted maximum on day 14 for each time point per individual on days 14, 23, and 30. Then the average melatonin profiles were calculated for each condition. It was not possible to estimate maxima for all participants, because some participants missed 2 samples or more at the time points at which the maximum melatonin concentration was expected. Therefore, for this analysis, 16 participants have been included in the blue light condition and 17 in the amber light condition. The average pattern of the melatonin curves for day 14 and day 23 is depicted in Figure 3A for the blue condition and in Figure 3B for the amber condition.

Figure 3.

(A) The average melatonin profiles on day 14 (–SD) and day 23 (+SD) in the blue condition in percentage relative to the fitted maximum of day 14 (for more details, see the “Materials and Methods” section). The area under the curve was significantly larger on day 23 compared with day 14 (*p < 0.05). (B) The average melatonin profiles on day 14 (+SD) and day 23 (–SD) in the amber condition. No significant differences were found. The level at 25% of the maximum melatonin concentration is indicated by the horizontal black lines.

The average area under the curve of the melatonin pattern at night in the blue light condition was significantly higher, on average +15% (±20%) after light therapy compared with the area under the curve during the baseline night (t₁₅ = −2.24, p < 0.05; Figure 3A). This was not the case for the amber light condition. There was even a trend for a decrease in the area under the curve after amber light therapy; on average −6% (±33%) lower compared with baseline (t₁₆ = 1.70, p = 0.10; Figure 3B). The change in the area under the curve between day 14 and day 23 was significantly different between both conditions (t₃₁ = 2.74, p < 0.05). Seven days after treatment (on day 30), the areas under the curve were no longer significantly different from day 14, neither for the amber condition, on average +4% (±33%; t₁₆ = −0.19, p = 0.85), nor for the blue condition, on average +8% (±44%; t₁₅ = −0.60, p = 0.56; data not shown). There was also no significant difference between conditions for the area under the curve on day 30 (t₃₁ = 0.43, p = 0.67).

Effects on Sleep Timing

Estimates of sleep timing and integrity were calculated as the average over the first 14 days in the baseline period, over the 9 days of the treatment period, and over the 7 days in the posttreatment period. The differences between the treatment period and baseline period as well as the differences between the posttreatment period and baseline period were tested between the 2 conditions (Table 2).

Table 2.

Sleep parameters with either amber or blue light pulses (mean ± SD).

Sleep Parameter	Condition	n	Baseline (14 Days)	Treatment (9 Days)	p within Different Base	Posttreatment(7 Days)	p withinDifferent Base
Sleep onset (h:min)	Blue	16	0:58 (1:00)	0:15 (0:54)	<0.001	0:18 (1:06)	<0.001
Sleep offset (h:min)^b	Blue	16	8:24 (0:54)	7:45 (0:54)	<0.001	8:03 (0:57)	<0.05
SOL (min)	Blue	15	8.7 (4.9)	10.7 (5.1)	0.13	10.8 (6.8)	0.16
Sl. duration (min)^b	Blue	15	394.0 (27.3)	383.1 (23.4)	0.35	397.0 (43.0)	0.99
WASO (min)	Blue	15	53.0 (16.0)	58.7 (30.2)	0.34	54.6 (26.3)	0.73
Sleep bouts (nr)^a	Blue	15	25.1 (6.4)	24.3 (5.6)	0.36	24.7 (7.6)	0.59
Sleep efficiency (%)	Blue	15	86.5 (6.8)	84.8 (6.0)	0.13	85.9 (5.6)	0.34
Sleep onset (h:min)	Amber	21	1:12 (1:10)	0:24 (1:17)	<0.001	0:47 (1:11)	<0.05
Sleep offset (h:min)^b	Amber	21	8:11 (1:05)	7:52 (1:05)	<0.05	8:22 (0:48)	0.12
SOL (min)	Amber	19	13.3 (8.8)	13.6 (10.8)	0.09	9.5 (6.8)	0.38
Sl. duration (min)^b	Amber	19	364.2 (46.2)	382.4 (44.5)	<0.05	386.8 (54.1)	<0.05
WASO (min)	Amber	19	51.6 (36.4)	61.3 (41.8)	<0.05	61.3 (44.6)	0.28
Sleep bouts (nr)^a	Amber	19	24.6 (5.2)	25.9 (5.7)	<0.05	26.8 (8.28)	0.16
Sleep efficiency (%)	Amber	19	85.5 (8.28)	83.4 (8.7)	<0.05	85.0 (8.7)	0.85

WASO = minutes awake after sleep onset; SOL = sleep-onset latency.

Sleep parameter is significantly different between conditions during treatment.

Sleep parameter is significantly different between conditions during posttreatment.

During the treatment period, sleep-onset time and wake-up time in both the blue light condition and the amber light condition significantly advanced compared with the baseline workdays period (see Table 2). During treatment, no significant differences between conditions in sleep time advances were found, although there was a trend for a slightly earlier sleep offset in the blue light condition (sleep onset, F_1,35 = 0.09, p = 0.76; sleep offset, F_1,35 = 3.47, p = 0.07). No differences between conditions were expected during the treatment period as the effects were mainly due to the instructions of the sleep phase–advancing protocol and the instructions on the timing of light exposure. In contrast, during the posttreatment period (wherein the participants could freely choose their sleep timing), wake-up time was significantly different between both conditions (sleep offset, F_1,35 = 9.20, p < 0.01). The average wake-up time was slightly delayed in the amber light group and was still significantly advanced in the blue light group compared with baseline (see Table 2).

Effects on Sleep Integrity

Effects on the estimates of sleep are shown in Table 2. Sleep duration was longer for participants in the amber group during both the treatment and posttreatment period compared with the baseline period, whereas it was shorter during treatment and not different during posttreatment compared with baseline for participants in the blue group. During treatment, WASO was also significantly increased compared with baseline and sleep efficiency was significantly lower for participants in the amber group. This was not the case in the blue light group. The number of sleep bouts was significantly higher during sleep in the treatment period for the participants in the amber light group compared with the number of sleep bouts during sleep in the participants in the blue light group (F_1,32 = 4.40, p < 0.05). These results may indicate that participants in the amber group had less consolidated sleep during the treatment period compared with subjects in the blue light group.

Sleepiness Scores

Sleepiness scores, as measured with the KSS, were measured every day of the protocol, 5 min after waking up and 30 min after waking up. Compared with the first 14 baseline days, the sleepiness scores (Table 3) were not different between conditions during treatment as well as during posttreatment, neither at 5 minutes after waking up (treatment, F_1,36 = 0.06, p = 0.80; posttreatment, F_1,36 = 0.06, p = 0.81) nor at 30 min after waking up (treatment, F_1,36 = 1.46, p = 0.24; posttreatment, F_1,36 = 0.88, p = 0.57). During posttreatment, participants felt significantly less sleepy 5 min as well as 30 min after waking up in both groups compared with the same time points in the baseline period (Table 3; blue 5 min, F_1,17 = 7.93, p < 0.05; blue 30 min, F_1,17 = 8.24, p < 0.05; amber 5 min, F_1,17 = 5.19, p <0.05; amber 30 min, F_1,17 = 4.16, p < 0.05), although it should be noted that the time points at which KSS was measured during the posttreatment period were later in the amber group than in the blue group, because participants in the amber group woke up later in this period (see the “Effects on Sleep Timing” section).

Table 3.

Average Sleepiness Scores (SD).

KSS baseline	Blue	Amber
5 min	6.2 (1.2)	5.9 (0.8)
30 min	4.8 (1.1)	4.5 (0.8)
Sleepiness decline	1.4 (0.7)	1.5 (0.7)
KSS treatment
5 min	6.4 (1.5)	6.2 (1.2)
30 min	4.5 (1.2)	4.5 (0.9)
Sleepiness decline	1.9 (0.7)*	1.7 (1.0)
KSS posttreatment
5 min	5.7 (1.1)*	5.5 (1.2)*
30 min	4.2 (0.9)*	4.1 (0.9)*
Sleepiness decline	1.6 (0.7)	1.4 (0.8)

KSS = Karolinska Sleepiness Scale.

Significantly different from baseline, p < 0.05.

The decline of the sleepiness score in the morning, which is related to the phenomenon of sleep inertia, was measured as the difference between the KSS score 30 min after waking up and the KSS score 5 min after waking up (Table 3). Compared with baseline, participants in the blue group had a significantly larger decline in their KSS score during the treatment period (F_1,17 = 7.36, p < 0.05), while there was no significant difference with baseline for the decline in KSS in the amber group (F_1,17 = 1.98, p = 0.18). In other words, sleep inertia dissipation was larger during blue light treatment than during amber light treatment.

Effects on Performance

Performance was measured by analyzing the average reaction time of a 5-min psychomotor vigilance task psychomotor vigilance task. First, the average baseline reaction times of each participant were calculated for each of the 4 time points during the day. The reaction times during treatment and posttreatment days were expressed as a percentage of the baseline reaction time. The average percentages of the reaction times for each condition are shown in Figures 4A (treatment) and 4B (posttreatment). For the statistics, the raw reaction times were used. In the amber light group, daytime performance was significantly worse during both the treatment period (F_1,13 = 10.1, p < 0.01; see Figure 4A) and during the posttreatment period (F_1,13 = 17.1, p < 0.01; see Figure 4B) compared with baseline. During treatment in the blue light group, there was a trend that performance was worse in the afternoon (F_1,10 = 4.0, p = 0.07; see Figure 4A) compared with the afternoon in the baseline period. During the posttreatment period, performance in the blue group was worse only in the morning (F_1,10 = 9.8, p < 0.05; see Figure 4B) compared with performance in the morning in the baseline period.

Figure 4.

(A) The relative performance values (mean relative reaction time ± SD) compared with baseline (baseline was set at 100%) during light therapy treatment. Performance was significantly worse at all times in the amber condition (**p < 0.01). (B) The relative performance values (reaction time) compared with baseline after light therapy treatment. Performance was significantly worse at all times in the amber condition (**p < 0.01). Performance was significantly worse during the morning in the blue condition (*p < 0.05).

Discussion

The focus of this study was to examine the (post-) treatment effects of additional blue light (relative to amber light) therapy on top of a sleep phase–advancing protocol on both the phase of the melatonin rhythm and on sleep and daytime performance. Recent studies have found that optimally timed short pulses (30 min) of blue light during the morning are effective in phase advancing the phase of melatonin (Geerdink et al., 2012; Crowley and Eastman, 2014). The present study examined whether such 30-min light pulses are also effective in shifting sleep with preservation of sleep integrity and performance in a home-setting light therapy study.

Phase Advances of DLMO

After both the blue and the amber light therapy protocols, a significant phase advance of DLMO was observed. The observed phase advances in both groups were not unexpected. It has been found that sleep schedules alone can phase advance the melatonin rhythm (Gordijn et al., 1999; Danilenko et al., 2003; Sharkey et al., 2011). In a home setting, earlier sleep and wake times will also automatically result in an earlier light exposure pattern, which in turn will induce a phase advance. Yet, despite all the environmental influences of a home setting in the current study, additional blue light on top of the phase-advancing sleep protocol turned out to be more effective in advancing DLMO than additional amber light. This can be explained only by the significant difference in the amount of short blue wavelengths in the morning in the blue light protocol (2306 melanopic lux from the device and 1485 blue lux measured by the Daysimeter between 0600 h and 1200 h) compared with the amber light protocol (70 melanopic lux from the device and 965 blue lux measured by the Daysimeter between 0600 h and 1200 h), since exposure to all other white light and blue light intensities during the day were not significantly different between the 2 groups.

Melatonin Production

Not only the phase advance of the melatonin rhythm was different between the conditions, but also the total production of nighttime melatonin was found to be different the first night after treatment. In the blue light condition, the total area under the curve of the nighttime melatonin pattern was larger after treatment compared with baseline, while this was not the case in the amber light condition. The possible influence of daytime light on nighttime melatonin production has been investigated before. Melatonin concentration has been reported to be higher in summer compared with winter (Tarquini et al., 1997). Applying 4 weeks of bright light for 4 h a day (1000-1200 h and 1400-1600 h) to Japanese elderly persons in a nursery home yielded a larger area under the curve of melatonin concentration compared with baseline (Mishima et al., 2001). Three other studies have found a higher amplitude of the melatonin rhythm after 3 (Hashimoto et al., 1997; Park & Tokura, 1999) or 7 days (Takasu et al., 2006) with bright light exposure during the whole day. The present study is the first study that shows that even short blue light pulses of 30 min (applied for 9 days) in the morning can have an effect on melatonin production at night.

There are several processes that may explain the influence of light during the day on melatonin concentration during the night. Light has dual effects on the output of melatonin. It regulates the phase of melatonin by resetting the SCN (Khalsa et al., 2003; Rüger et al., 2006), and light suppresses melatonin concentrations when it is produced (Lewy et al., 1980; Zeitzer et al., 2000), probably via inhibition of the activity of the B-adrenergic receptors (Cagnacci et al., 1992). Therefore, the melatonin production could have been increased by 2 processes: either by an increase in the synchronization of single-cell firings in the SCN (Meijer et al., 1998, Masubuchi et al., 1999; Harb et al., 2014), and therefore in the amplitude of the underlying circadian oscillation, or by an increase in the sensitivity of the pineal gland to light signals, which may result in an upregulation of the B-adrenergic receptors.

Sleep Timing

Even though a phase advance of the circadian system is important in a successful phase-advancing treatment, the primary result of a therapy aiming at advancing sleep should be the effect on sleep timing. In this study, during the sleep phase–advancing protocol and light treatment, no significant differences were found in sleep timing between the 2 groups. This was expected as the participants got clear instructions about their sleep timing. It also means that differences between groups on other parameters can be attributed only to the additional blue light exposure and not to sleep timing differences per se. Despite the fact that the DLMO returned to baseline values at the last day of the protocol posttreatment, participants in the blue light therapy group still woke up about 30 min earlier compared with participants in the amber light group during the posttreatment period. It has been shown before that voluntary sleep timing can be advanced by morning light therapy (Corbett et al., 2012). However, in that study, voluntary sleep timing was measured during the treatment period only and not after treatment. Another difference between that study and our study was that the treatment period was much longer (3 months). A recent study of Saxvig et al. (2014) tested the follow-up effects after 3 months of continuous use of bright light and melatonin capsules, which resulted in earlier voluntary sleep times for the group that continued treatment. Still, a follow-up with continuous light therapy is different than measuring posttreatment effects directly. To our knowledge, this is the first study that shows that short blue light pulses in the morning can have a posttreatment effect on sleep timing during 1 week.

Sleep Integrity

Low daytime light levels are correlated with poor sleep quality scores (Leger et al., 2011; Boubekri et al., 2014), whereas increased light levels during the day are correlated with an improvement of sleep quality (Viola et al., 2008; Paul et al., 2015). It should be noted that most of these studies included only subjective sleep quality scores and that studies were either only correlational or, in case of an experimental setup, light intensity during the day was increased without paying attention to its timing. Because of these limitations, it is hard to say what the mechanisms behind these results were. In the present study, also different effects between the 2 light therapy groups considering sleep quantity and sleep integrity (measured with actigraphy) were found. During treatment, WASO and the number of sleep bouts were significantly increased and sleep efficiency was significantly decreased for participants in the amber group compared with baseline whereas they were not different in participants of the blue group. At the same time, sleep duration was increased in the amber group and not in the blue group. We speculate that the longer sleep duration in the amber group is a consequence of the increased sleep fragmentation and that the increased sleep fragmentation resulted from the induced mismatch between sleep timing and melatonin secretion: both groups were given the instructions to sleep 2 h earlier by the end of the treatment period, while the DLMO of the blue and amber group shifted for 84 and 48 min, respectively. In a study by Lazar and colleagues (2013), it was found that if the phase angle between the DLMO and habitual sleep onset is shorter, people have a higher insomnia score. In the present study, higher concentrations of melatonin were reached later in the amber group compared with the blue group, and the phase angle between DLMO and sleep was shorter.

Another explanation, which is more speculative, for more sleep fragmentation in the amber group may be found in the effects of the intensity and color of light on sleep homeostasis itself through the melanopsin pathway. This effect has not been extensively studied and actually only in mice. Nevertheless, the results that have been found are clear. If mice are lacking the OPN4 gene, which is responsible for the expression of melanopsin, the homeostatic sleep pressure is much lower and sleep is more fragmented, while the circadian rhythm is still present (Altimus et al., 2008; Tsai et al., 2009). Because the melanopsin photopigment is responsive to short-wavelength light, participants in the amber light group had less and participants in the blue group had more melanopsin stimulation, which may have resulted in the observed difference in sleep consolidation.

Sleepiness and Performance

Blue light was more efficient in decreasing sleep inertia after waking up than amber light was. It is known that blue-enriched light is efficient in keeping humans alert (Hommes & Giménez, 2015), especially during the night when sleepiness scores are high (Lockley et al., 2006; Rüger et al., 2006; Phipps-Nelson et al., 2009; Chellappa et al., 2011; Rahman et al., 2014). Because the effect is strongest at high sleepiness scores, one would expect that the effectiveness of blue light during the day would be highest directly after waking up. The results of the present study are indeed in line with this theory. They are also consistent with an earlier study of Revell and colleagues (2006), wherein they used light of different wavelengths and found that the alerting effect of light with 470 nm (blue light) is strongest directly after waking up.

In contrast, previous studies found no clear effects of blue light during the morning on performance tasks. A study by Santhi et al. (2013) found no improvement of reaction times with various blue light conditions directly after waking up during the first hours of the light pulse. They only found a faster reaction time after 4 h in the condition that contained the highest blue light intensity. Gabel et al. (2013) used 20 min of similar blue light in the morning and found no daily improvement of performance at all. It should be noted that in these studies, the light exposure either started very early (Santhi et al., 2013) or started 2 h after wake up and after sleep restriction (Gabel et al., 2013). In the present study, we also did not find improved performance in the morning under blue light exposure, even though the sleepiness scores were lower. Still, in the blue light group, morning performance did also not significantly worsen during treatment, even though the subjects had to wake up 2 h earlier than during the baseline period. In the amber light group, performance was worsened during the whole day in the treatment as well as in the posttreatment period. This decrease in performance may be linked to a decrease in sleep integrity, as poor sleep integrity can decrease performance during the day (Bonnet, 1989, Miyata et al., 2013). It may also be linked to the change in the phase of melatonin in relation to sleep timing (Burgess et al., 2013a; Sletten et al., 2015). It was expected that a possible decrease in performance, if present, would have been largest during the postlunch dip (reviewed in Carrier and Monk, 2000). While the amber group had a worsened performance during the whole day, the blue group showed a trend for a worsened performance only in the late afternoon during the treatment period. This may be attributed to the postlunch dip. The postlunch dip is dependent on an interaction between the duration of prior waking and the circadian clock (Bes et al., 2009). In this study, both of these components are changed.

Performance was significantly worse in the morning in the posttreatment period, also for the blue light group. This may be explained by the changed phase angle between the melatonin rhythm and sleep phase after light therapy. The participants in the blue group still woke up earlier compared with baseline, while the timing of their melatonin rhythm was shifted back to the timing during the baseline period. This especially may have had consequences for performance in the morning, as the participants woke up too early in relation to their internal clock.

Conclusion

The results of this placebo-controlled home study support the conclusion that blue light therapy was able to support a sleep phase advancing protocol by increasing the shift in the melatonin rhythm and posttreatment sleep phase with (to a large extent) preservation of sleep integrity and performance. After treatment, the phase advance of the melatonin rhythm shifted back to the original values, while sleep timing was still advanced after the use of blue light therapy. This confirms that blue light therapy was supportive in shifting the sleep phase more than amber light was. Whether the discrepancy between the melatonin rhythm and sleep phase after light therapy is desirable is the question, as performance was worse during the morning after treatment in the blue light therapy group. To investigate if performance can be stabilized and if the effects on sleep are long-lasting, more research should be done on longer (blue) light therapy protocols with voluntary sleep timing. Also the effect of light on sleep homeostasis in humans has to be investigated to acquire more knowledge on optimizing light therapy protocols.

Footnotes

Acknowledgements

The work at the University of Groningen was financially supported by an unrestricted research grant of Philips Consumer Lifestyle B.V. Drachten, the Netherlands, developed by the legal authorities of both parties. The scientific advisor of the sponsor and coauthor of this paper, Dr. V. Hommes, had no other input in the project, or in the analysis and writing of the manuscript, as is expected from a scientific coauthor. The whole responsibility of the analyses and the article lies with the researchers of the University of Groningen.

References

Altimus

Güler

Villa

McNeill

Legates

Hattar

(2008) Rods-cones and melanopsin detect light and dark to modulate sleep independent of image formation. Proc Natl Acad Sci USA 105:19998-20003.

Ancoli-Israel

Cole

Alessi

Chambers

Moorcroft

Pollak

(2003) The role of actigraphy in the study of sleep and circadian rhythms. Sleep 26(3):342-392.

Appleman

Figueiro

Rea

(2013) Controlling light-dark exposure patterns rather than sleep schedules determines circadian phase. Sleep Med 14:456-461.

Arendt

(2005) Melatonin: characteristics, concerns, and prospects. J Biol Rhythms 20:291-303.

Barion

Zee

(2007) A clinical approach to circadian rhythm sleep disorders. Sleep Med 8:566-577.

Baron

Reid

(2014) Circadian misalignment and health. Int Rev Psychiatry 26:139-154.

Beck

Steer

Ball

(1996) Comparison of Beck Depression Inventories–IA and –II in psychiatric outpatients. J Pers Asses 67:588-597.

Beck

Steer

Brown

(1996) Manual for the Back Depression Inventory (Dutch version). Van der Does

A. J. W.

editor. Lisse, the Netherlands: Swets Test.

Beersma

DGM

Comas

Hut

Gordijn

MCM

Rueger

Daan

(2009) The progression of circadian phase during light exposure in animals and humans. J Biol Rhythms 24:153-160.

10.

Benloucif

Burgess

Klerman

Lewy

Middleton

Murphy

Parry

Revell

(2008) Measuring melatonin in humans. J Clin Sleep Med 4:66-69.

11.

Berson

Dunn

Takao

(2002) Phototransduction by retinal ganglion cells that set the circadian clock. Science 295:1070-1073.

12.

Bes

Jobert

Schulz

(2009) Modeling napping, post-lunch dip, and other variations in human sleep propensity. Sleep 32:392-398.

13.

Bjorvatn

Pallesen

(2009) A practical approach to circadian rhythm sleep disorders. Sleep Med Rev 13:47-60.

14.

Bonnet

(1989) The effect of sleep fragmentation on sleep and performance in younger and older subjects. Neurobiol Aging 10:21-25.

15.

Borbély (1982) A two process model of sleep regulation. Hum Neurobiol 1(3):195-204.

16.

Boubekri

Cheung

Reid

Wang

Zee

(2014) Impact of windows and daylight exposure on overall health and sleep quality of office workers: a case-control pilot study. J Clin Sleep Med 10:603-611.

17.

Buresova

Dvorakova

Zvolsky

Illnerova

(1991) Early morning bright light phase advances the human circadian pacemaker within one day. Neurosci Lett 121:47-50.

18.

Burgess

(2013b) Evening ambient light exposure can reduce circadian phase advances to morning light independent of sleep deprivation. J Sleep Res 22:83-88.

19.

Burgess

Legasto

Fogg

Smith

(2013a) Can small shifts in circadian phase affect performance? Appl Ergon 44:109-111.

20.

Cagnacci

Elliot

Yen

(1992) Melatonin a major regulator of the circadian rhythm of core body temperature in humans. J Clin Endocrinol Metab 75(2):447-452.

21.

Carrier

Monk

(2000) Circadian rhythms of performance: new trends. Chronobiol Int 17:719-732.

22.

Chae

Kripke

Poceta

Shadan

Jamil

Cronin

Kline

(2009) Evaluation of immobility time for sleep latency in actigraphy. Sleep Med 10:621-625.

23.

Chang

A-M

Santhi

St. Hilaire

Gronfier

Bradstreet

Duffy

Lockley

Kronauer

Czeisler

(2012) Human responses to bright light of different durations. J Physiol 590:3103-3112.

24.

Chellappa

Steiner

Blattner

Oelhafen

Götz

Cajochen

(2011) Non-visual effects of light on melatonin, alertness and cognitive performance: can blue-enriched light keep us alert? PLoS One 6(1):e16429.

25.

Corbett

Middleton

Arendt

(2012) An hour of bright white light in the early morning improves performance and advances sleep and circadian phase during the Antarctic winter. Neurosci Lett 525:146-151.

26.

Crowley

Eastman

(2014) Phase advancing human circadian rhythms with morning bright light, afternoon melatonin, and gradually shifted sleep: can we reduce morning bright-light duration? Sleep Med 16:288-297.

27.

Daan

Beersma

DGM

Borbély

(1984) Timing of human sleep: recovery process gated by a circadian pacemaker. Am J Physiol 2:R161-R178.

28.

Daan

Pittendrigh

(1976) A functional analysis of circadian pacemakers in nocturnal rodents. II. The variability of phase response curves. J Comp Physiol 106:253-266.

29.

Danilenko

Cajochen

Wirz-Justice

(2003) Is sleep per se a zeitgeber in humans? J Biol Rhythms 18:170-178.

30.

Danilenko

Wirz-Justice

Kräuchi

Cajochen

Weber

Fairhurst

Terman

(2000) Phase advance after one or three simulated dawns in humans. Chronobiol Int 17:659-668.

31.

Dijk

Beersma

Daan

Lewy

(1989) Bright morning light advances the human circadian system without affecting NREM sleep homeostasis. Am J Physiol 256:R106–R111.

32.

Gabel

Maire

Reichert

Chellappa

Schmidt

Hommes

Viola

Cajochen

(2013) Effects of artificial dawn and morning blue light on daytime cognitive performance, well-being, cortisol and melatonin levels. Chronobiol Int 30:988-997.

33.

Geerdink

Beersma

DGM

Hommes

Gordijn

MCM

(2012) Blue light in the morning phase advances the rhythm of melatonin and reduces sleepiness at waking up. Abstract presented at the Society for Research on Biological Rhythms Congress, Destin, Florida.

34.

Gordijn

MCM

Beersma

DGM

Korte

Van Den Hoofdakker

(1999) Effects of light exposure and sleep displacement on dim light melatonin onset. J Sleep Res 8:163-174.

35.

Gradisar

Dohnt

Gardner

Paine

Starkey

Menne

Slater

Wright

Hudson

Weaver

. (2011) A randomized controlled trial of cognitive-behavior therapy plus bright light therapy for adolescent delayed sleep phase disorder. Sleep 1671-1680.

36.

Gronfier

Wright

Kronauer

Jewett

Czeisler

(2004) Efficacy of a single sequence of intermittent bright light pulses for delaying circadian phase in humans. Am J Physiol Endocrinol Metab 287:E174-E181.

37.

Harb

Hidalgo

Martau

(2014) Lack of exposure to natural light in the workspace is associated with physiological, sleep and depressive symptoms. Chronobiol Int 32:368-375.

38.

Hashimoto

Kohsaka

Nakamura

Honma

(1997) Midday exposure to bright light changes the circadian organization of plasma melatonin rhythm in humans. Neurosci Lett 221:89-92.

39.

Hommes

Giménez

(2015) A revision of existing Karolinska Sleepiness Scale responses to light: a melanopic perspective. Chronobiol Int 32:750-756.

40.

Khalsa

SBS

Jewett

Cajochen

Czeisler

(2003) A phase response curve to single bright light pulses in human subjects. J Physiol 549:945-952.

41.

Klerman

Gershengorn

Duffy

abnd Kronauer

(2002) Comparisons of the variability of three markers of the human circadian pacemaker. J Biol Rhythms 17:181-193.

42.

Kräuchi

Cajochen

Danilenko

Wirz-Justice

(1997) The hypothermic effect of late evening melatonin does not block the phase delay induced by concurrent bright light in human subjects. Neurosci Lett 232:57-61.

43.

Lack

Bramwell

Wright

Kemp

(2007) Morning blue light can advance the melatonin rhythm in mild delayed sleep phase syndrome. Sleep Biol Rhythms 5:78-80.

44.

Lazar

Santhi

Hasan

JCY

Johnston

Schantz

Von, Archer

Dijk

(2013) Circadian period and the timing of melatonin onset in men and women: predictors of sleep during the weekend and in the laboratory. J Sleep Res 22:155-159.

45.

Leger

Bayon

Elbaz

Philip

Choudat

(2011) Underexposure to light at work and its association to insomnia and sleepiness. J Psychosom Res 70:29-36.

46.

Lewy

Cutler

Sack

(1999) The endogenous melatonin profile as a marker for circadian phase position. J Biol Rhythms 14:227-236.

47.

Lewy

Sack

(1989) The dim light melatonin onset as a marker for circadian phase position. Chronobiol Int 6:93-102.

48.

Lewy

Wehr

Goodwin

Newsome

Markey

(1980) Light suppresses melatonin secretion in humans. Science 210:1267-1269.

49.

Lockley

Brainard

Czeisler

(2003) High sensitivity of the human circadian melatonin rhythm to resetting by short wavelength light. J Clin Endocrinol Metab 88:4502-4505.

50.

Lockley

Evans

Scheer

Brainard

Czeisler

Aeschbach

(2006) Short-wavelength sensitivity for the direct effects of light on alertness, vigilance, and the waking electroencephalogram in humans. Sleep 29:161-168.

51.

Lucas

Peirson

Berson

Brown

Cooper

Czeisler

Figueiro

Gamlin

Lockley

O’Hagan

. (2014) Measuring and using light in the melanopsin age. Trends Neurosci 37:1-9.

52.

Masubuchi

Hashimoto

Endo

Honma

(1999) Amplitude reduction of plasma melatonin rhythm in association with an internal desynchronization in a subject with non-24-hour sleep-wake syndrome. Psychiatry Clin Neurosci 53:249-251.

53.

Meijer

Schwartz

(2003) In search of the pathways for light-induced pacemaker resetting in the suprachiasmatic nucleus. J Biol Rhythms 18:235-249.

54.

Meijer

Watanabe

Schaap

Albus

Détári

(1998) Light responsiveness of the suprachiasmatic nucleus: long-term multiunit and single-unit recordings in freely moving rats. J Neurosci 18:9078-9087.

55.

Merikanto

Lahti

Puusniekka

Partonen

(2013) Late bedtimes weaken school performance and predispose adolescents to health hazards. Sleep Med 14:1105-1111.

56.

Mishima

Okawa

Shimizu

Hishikawa

(2001) Diminished melatonin secretion in the elderly caused by insufficient environmental illumination. J Clin Endocrinol Metab 86:129-134.

57.

Miyata

Noda

Iwamoto

Kawano

Okuda

Ozaki

(2013) Poor sleep quality impairs cognitive performance in older adults. J Sleep Res 22:535-541.

58.

Moore

(2007) Suprachiasmatic nucleus in sleep-wake regulation. Sleep Med 8:27-33.

59.

Morgenthaler

Lee-Chiong

Alessi

Friedman

Aurora

Boehlecke

Brown

Chesson

Kapur

Maganti

. (2007) Practice parameters for the clinical evaluation and treatment of circadian rhythm sleep disorders. An American Academy of Sleep Medicine report. Sleep 30:1445-1459.

60.

O’Hare

Flanagan

Penzel

Garcia

Frohberg

Heneghan

(2015) A comparison of radio-frequency biomotion sensors and actigraphy versus polysomnography for the assessment of sleep in normal subjects. Sleep Breath 19:91-98.

61.

Park

Tokura

(1999) Bright light exposure during the daytime affects circadian rhythms of urinary melatonin and salivary immunoglobulin A. Chronobiol Int 16:359-371.

62.

Paul

Love

Hawton

Brett

McCreary

Arendt

(2015) Light treatment improves sleep quality and negative affectiveness in high arctic residents during winter. Photochem Photobiol 91:567-573.

63.

Phipps-Nelson

Redman

Schlangen

LJM

Rajaratnam

SMW

(2009) Blue light exposure reduces objective measures of sleepiness during prolonged nighttime performance testing. Chronobiol Int 26:891-912.

64.

Rahman

Flynn-Evans

Aeschbach

Brainard

Czeisler

Lockley

(2014) Diurnal spectral sensitivity of the acute alerting effects of light. Sleep 37:271-281.

65.

Rajaratnam

Arendt

(2001) Health in a 24-h society. Lancet 358:999-1005.

66.

Rea

Figueiro

Bierman

Bullough

(2010) Circadian light. J Circadian Rhythms 8:2.

67.

Rea

Figueiro

Bullough

Bierman

(2005) A model of phototransduction by the human circadian system. Brain Res Rev 50:213-228.

68.

Revell

Arendt

Fogg

Skene

(2006) Alerting effects of light are sensitive to very short wavelengths. Neurosci Lett 399:96-100.

69.

Revell

Arendt

Terman

Skene

(2005a) Short-wavelength sensitivity of the human circadian system to phase-advancing light. J Biol Rhythms 20:270-272.

70.

Revell

Kim

Tseng

Crowley

Eastman

(2005b) Circadian phase determined from melatonin profiles is reproducible after 1 wk in subjects who sleep later on weekends. J Pineal Res 39:195-200.

71.

Roenneberg

Allebrandt

Merrow

Vetter

(2012) Social jetlag and obesity. Curr Biol 22:939-943.

72.

Roenneberg

Kuehnle

Juda

Kantermann

Allebrandt

Gordijn

Merrow

(2007) Epidemiology of the human circadian clock. Sleep Med Rev 11:429-438.

73.

Roenneberg

Wirz-Justice

Merrow

(2003) Life between clocks: daily temporal patterns of human chronotypes. J Biol Rhythms 18:80-90.

74.

Rosenthal

Joseph-Vanderpool

Levendosky

Johnston

Allen

Kelly

Souetre

Schultz

Starz

(1990) Phase-shifting effects of bright morning light as treatment for delayed sleep phase syndrome. Sleep 13:354-361.

75.

Rüger

Gordijn

MCM

Beersma

DGM

de Vries

Daan

(2006) Time-of-day-dependent effects of bright light exposure on human psychophysiology: comparison of daytime and nighttime exposure. Am J Physiol Regul Integr Comp Physiol 290:R1413-R1420.

76.

Rüger

Scheer

(2009) Effects of circadian disruption on the cardiometabolic system. Rev Endocr Metab Disord 10:245-260.

77.

Rüger

St. Hilaire

Brainard

Khalsa

S-BS

Kronauer

Czeisler

Lockley

(2013) Human phase response curve to a single 6.5 h pulse of short-wavelength light. J Physiol 591:353-363.

78.

Samková

Vondrašová

Hájek

Illnerová

(1997) A fixed morning awakening coupled with a low intensity light maintains a phase advance of the human circadian system. Neurosci Lett 224:21-24.

79.

Santhi

Groeger

Archer

Gimenez

Schlangen

LJM

Dijk

(2013) Morning sleep inertia in alertness and performance: effect of cognitive domain and white light conditions. PLoS One 8:1-13.

80.

Saxvig

Wilhelmsen-Langeland

Pallesen

Nordhus

I-H

Vedaa

Lundervold

Bjorvatn

(2014) A randomized controlled trial with bright light and melatonin for the treatment of delayed sleep phase disorder: effects on subjective and objective sleepiness and cognitive function. J Biol Rhythms 28:306-321.

81.

Schroeder

Colwell

(2013) How to fix a broken clock. Trends Pharmacol Sci 34:605-619.

82.

Sharkey

Carskadon

Figueiro

Zhu

Rea

(2011) Effects of an advanced sleep schedule and morning short wavelength light exposure on circadian phase in young adults with late sleep schedules. Sleep Med 12:685-692.

83.

Sletten

Segal

Flynn-Evans

Lockley

Rajaratnam

SMW

(2015) Inter-individual differences in neurobehavioural impairment following sleep restriction are associated with circadian rhythm phase. PLoS One 10:e0128273.

84.

St. Hilaire

Gooley

Khalsa

SBS

Kronauer

Czeisler

Lockley

(2012) Human phase response curve to a 1h pulse of bright white light. J Physiol 590.13:3035-3045.

85.

Takasu

Hashimoto

Yamanaka

Tanahashi

Yamazaki

Honma

(2006) Repeated exposures to daytime bright light increase nocturnal melatonin rise and maintain circadian phase in young subjects under fixed sleep schedule. Am J Physiol Regul Integr Comp Physiol 291:R1799-R1807.

86.

Tarquini

Cornélissen

Perfetto

Tarquini

Halberg

(1997) Chronome assessment of circulating melatonin in humans. In Vivo 11:473-484.

87.

Touitou

(2013) Adolescent sleep misalignment: a chronic jet lag and a matter of public health. J Physiol Paris 107:323-326.

88.

Tsai

Hannibal

Hagiwara

Colas

Ruppert

Ruby

Heller

Franken

Bourgin

(2009) Melanopsin as a sleep modulator: circadian gating of the direct effects of light on sleep and altered sleep homeostasis in Opn4-/- mice. PLoS Biol 7(6):e1000125.

89.

Van Someren

EJW

Nagtegaal

(2007) Improving melatonin phase estimates. Sleep Med 8:590-561.

90.

Viola

James

Schlangen

LJM

Dijk

(2008) Blue-enriched white light in the workplace improves self-reported alertness, performance and sleep quality. Clin Res 34:297-306.

91.

Wittmann

Dinich

Merrow

Roenneberg

(2006) Social jetlag: misalignment of biological and social time. Chronobiol Int 23:497-509.

92.

Wittmann

Paulus

Roenneberg

(2010) Decreased psychological well-being in late ‘chronotypes’ is mediated by smoking and alcohol consumption. Subst Use Misuse 45:15-30.

93.

Zeitzer

Dijk

Kronauer

Brown

Czeisler

(2000) Sensitivity of the human circadian pacemaker to nocturnal light: melatonin phase resetting and suppression. J Physiol 526(pt 3):695-702.

94.

Zeitzer

Khalsa

SBS

Boivin

Duffy

Shanahan

Kronauer

Czeisler

(2005) Temporal dynamics of late night photic stimulation of the human circadian timing system. Am J Physiol Regul Integr Comp Physiol 289:R839-R844.