Restricting short-wavelength light in the evening to improve sleep in recreational athletes – A pilot study

Abstract

Sleep is crucial for recovery and skill acquisition in athletes. Paradoxically, athletes often encounter difficulties initiating and maintaining sleep, while having sufficient sleep opportunity. Blue (short-wavelength) light as emitted by electronic screens is considered a potential sleep thief, as it suppresses habitual melatonin secretion. The current study sought to investigate whether blocking short-wavelength light in the evening can improve sleep onset latency and potentially other sleep parameters among recreational athletes. The study had a within-subject crossover design. Fifteen recreational athletes, aged between 18 and 32 years (12 females, 3 males), were randomly assigned to start the intervention period with either the light restriction condition (LR; amber-lens glasses), or the no-light restriction condition (nLR; transparent glasses). Sleep hygiene practices, actigraphy and diary-based sleep estimates were monitored during four consecutive nights within each condition. Sleep hygiene practices did not significantly differ between conditions. Results indicate that blocking short-wavelength light in the evening, as compared to habitual light exposure, significantly shortened subjective sleep onset latency (Δ = 7 min), improved sleep quality (Δ = 0.6; scale 1–10), and increased alertness the following morning. Actigraphy-based sleep estimates showed no significant differences between conditions. Blocking short-wavelength light in the evening by means of amber-lens glasses is a cost-efficient and promising means to improve subjective sleep estimates among recreational athletes in their habitual home environment. The relatively small effects of the current study may be strengthened by additionally increasing morning- and daytime light exposure and, potentially, by reducing the alerting effects of media use before bedtime.

Keywords:

• Athletes
• sleep
• light restriction
• short-wavelength light
• blue-blockers

Highlights

• Athletes can greatly benefit from sleep optimization.

• Reduced exposure to blue-enriched light in the evening improves perceived sleep onset latency, sleep quality and alertness in the morning.

Introduction

Sleep is considered indispensable for recovery and skill acquisition in athletes (Fullagar et al., 2015; Halson, 2008). Paradoxically, athletes show markers of poor sleep quality and sleep efficacy despite having sufficient opportunity to sleep (i.e. based on an approximation of 8:30 h of time spent in bed (TIB); Knufinke, Nieuwenhuys, Geurts, Møst, et al., 2018; Leeder, Glaister, Pizzoferro, Dawson, & Pedlar, 2012). Modern-technologies, such as smartphones and laptops, are often blamed for high-jacking sleep, particularly in adolescents and athletes who are susceptible to the ‘fear of missing out’ (Romyn, Robey, Dimmock, Halson, & Peeling, 2016). The effect is twofold; The psychological sleep threatening components concern the stimulating effect of media content associated with higher bedtime arousal and delayed bedtimes (Halson, 2016). Another, physiological, component which will be focused on in this study, concerns evening exposure to artificial light, especially short wavelength light, which is thought to delay circadian phase by suppressing habitual melatonin synthesis (Heath et al., 2014; West et al., 2011; Wood, Rea, Plitnick, & Figueiro, 2013). Melatonin is crucial for sleep initiation and maintenance in humans (Brainard et al., 2001). Given the significance of sleep as a means for recovery and skill acquisition in athletes (Bonnar, Bartel, Kakoschke, & Lang, 2018), the current study sought to determine whether reducing evening exposure to short-wavelength light can improve sleep among recreational athletes.

Sleep and wakefulness are regulated by two distinct process: a homeostatic process (Process S) that depicts increasing sleep pressure following sustained wakefulness, and a circadian process (Process C)(Borbely, 1982), which is regulated by the circadian system and requires periodic light–dark exposure for stable entrainment to the geographical day (Czeisler et al., 1986; Duffy & Wright, 2005). Specifically, information on environmental light is received by photoreceptors in the retina and, via non-image forming intrinsically photoreceptive retinal ganglion cells (ipRGC) (Brainard et al., 2001; Thapan, Arendt, & Skene, 2001), directly transmitted to the suprachiasmatic nucleus (SCN), the site of the circadian ‘master clock’. The SCN, in turn, sends information on circadian time to, for example, the pineal gland where melatonin is secreted in the evening, or suppressed in case of evening-light exposure (Brainard et al., 2001). Hence, the timing of light exposure is crucial: daytime light exposure facilitates the process of waking up and staying alert during the early day, while in the evening, the absence of light facilitates sleepiness.

Due to artificial lightning and the advent of light-emitting hand-held screens, however, evening light is often abundant, tricking our circadian system to think it is daytime (Roenneberg, Wirz-Justice, & Merrow, 2003). Part of the environmental lighting and almost all electronic screens are rich in short-wavelength light (∼460 nm), which lies in the visual spectrum of blue–green. Importantly, a specific type of photoreceptive cells is particularly sensitive to light of this (short) wavelength (446–477 nm; Brainard et al., 2001) and modulates activation of the suprachiasmatic nucleus (SCN) such that melatonin secretion becomes actively suppressed (Thapan et al., 2001; West et al., 2011). A full suppression or delay in melatonin secretion is associated with reduced sleepiness, increased alertness, delayed sleep onset and reduced sleep efficiency (Chang, Aeschbach, Duffy, & Czeisler, 2015; Fossum, Nordnes, Storemark, Bjorvatn, & Pallesen, 2014). Accordingly, exposure to especially short-wavelength light in the evening may delay the circadian phase and disrupt sleep in athletes.

In line with the hypothesis that evening light exposure may negatively affect sleep, previous studies have shown that light-restriction can help to maintain a habitual melatonin secretion among individuals with sleep difficulties (Burkhart & Phelps, 2009). While these results are certainly promising, it should be noted that subsequent effects on sleep have not often been assessed, that many studies have been conducted among shift-workers (Sasseville, Paquet, Sévigny, & Hébert, 2006) or problematic sleepers (Burkhart & Phelps, 2009), and that studies have often lacked a (sufficiently neutral) control condition, or reported differences at baseline in a between subject design (Burkhart & Phelps, 2009).

Considering these limitations and gaps in the current literature, the present study investigated the effectiveness of blocking short-wavelength light in the evening on sleep under natural conditions in a non-sleep disordered and physically active population. Using a within-subject crossover design, recreational athletes were instructed to wear amber-lens glasses before bedtime in the experimental condition, which were substituted by non-vision adjusting transparent glasses in the control condition. In both conditions, sleep was monitored using wrist-worn actigraphy and daily sleep diaries for a period of nine nights. Considering the recent literature and the effects of short-wavelength light on the melatonin synthesis (Lockley, Brainard, & Czeisler, 2003), it was hypothesized that using blue-light blocking amber-lens glasses in the evening will improve actigraphy- and diary-based sleep onset latency, and potentially secondary measures such as total sleep time, sleep efficiency, and subjectively rated sleep quality.

Method

Fifteen recreational athletes, aged between 18 and 32 years (mean ± standard deviation (M ± SD); 23.27 ± 3.63 yrs) of whom 12 were females (3 males), participated with written informed consent. Inclusion criteria were (1) exercising one or more hours a week (endurance and/or weight training), (2) moderate to good subjective sleep quality based on the Pittsburgh Sleep Quality Index (Buysse, Reynolds, Monk, Berman, & Kupfer, 1989) (PSQI: all < 7; M ± SD; 3.87 ± 1.55), (3) no severe subjective sleep complaints based on the Holland Sleep Disorder Questionnaire (Kerkhof et al., 2013) (HSDQ: all < 2.06; M ± SD; 1.57 ± 0.26), (4) being free of sleep medication, (5) consuming < 500 mg caffeine a day (∼ 5 espressos) and < 5 standard units alcohol, (6) no current use of psychoactive medication, (7) absence of psychiatric and mood disorders, (8) no serious or unstable medical illness, (9) no diagnosed sleep disorders, (10) no time-zone crossing travel during the assessment period, (11) no pregnancy, and (12) no shift work. Participants were recruited among active members of the University Sports Centre. The study was approved by the faculty’s ethical committee [ECSW2016-1403-376], and participation was financially reimbursed. Data were collected in April 2016.

Design and procedure

The study had a within-subject crossover design. Participants were randomly assigned to start the intervention period with either the light restriction condition (LR), or the no-light restriction condition (nLR). Each condition started with two days of habituation, followed by seven days of intervention, that were scheduled four days apart to allow all participants to always start on the same weekday (Monday). The order of conditions was counterbalanced between participants. In the light-restriction condition, participants were instructed to wear amber-lens glasses three hours before bedtime (Eye shield soft red Safety Glasses, Königswinter, Germany), which were substituted by non-vision adjusting transparent glasses in the no-light restriction condition (clear non-prescription lenses black, by Oramics). To prevent explicit outcome expectancies from influencing our findings, participants were informed that the study was designed to assess the effects of light regulation on mood and alertness. Across both conditions, sleep was monitored for nine consecutive nights by means of wrist-worn actigraphy and paper-based morning- and evening diaries. To allow for a fair comparison between the experimental conditions, participants were instructed to follow a set of behavioural guidelines throughout both conditions (see below).

Light restriction

In the LR condition, participants were instructed to wear amber-lens glasses (Eye shield soft red Safety Glasses, Königswinter, Germany) during the last three hours before bedtime and at the latest at PM 9.00 (Sasseville et al., 2006). The amber-lens glasses filter 100% of wavelength up to 400 nm, and 89–99.9% of wavelength between 400 and 500 nm. The effectiveness of these glasses in preserving normal evening melatonin production during bright-light exposure has been reported elsewhere (Sasseville et al., 2006). Therefore, most of the blue (wavelength of 490–450 nm), and parts of the green light (560–520 nm) was effectively restricted.

To standardize the experimental protocol across conditions, in the nLR condition, participants were instructed to wear non-vision adjusting transparent glasses (clear non-prescription lenses, by Oramics). Hence, the short-wavelength light was not restricted in the evening.

Sleep estimates

Sleep data were collected by means of wrist-actigraphy and paper-based sleep diaries, which are less sensitive to detecting small changes in sleep onset than the gold-standard polysomnography (Chae et al., 2009; Rogers, Caruso, & Aldrich, 1993), but generally well accepted in any field-based measurement of sleep (Knufinke, Nieuwenhuys, Geurts, Møst, et al., 2018; Leeder et al., 2012).

Objective sleep estimates were collected using an actigraph (Actiwatch 2, Philips Respironics, Murrysville, USA), that was continuously worn around the non-dominant wrist and only detached during training or when being in contact with water. Activity and photonic light was sampled in 60 s bins. The primary measure of interest was sleep onset latency (min), and secondary measures were wake after sleep onset (min), fragmentation index (%), total sleep time (h:min), and sleep efficiency (%). Actigraphy data were analysed using Respironics Actiware 5 (Philips Respironics, Murrysville, USA) and processed in accordance with the guidelines formulated by the Society of Behavioural Sleep Medicine (SBSM) (Ancoli-Israel et al., 2015). Data were visually inspected and excluded when activity counts and light values indicated detachment of the sensor. In all other cases, rest intervals were manually set when (i) event markers identified bed- and rise time, or – in case event markers were missing – when (ii) light and activity was absent. If light and activity values were ambiguous, (iii) diary entries were used to set rest intervals. The default setting (10-minutes immobility parameter) was used to identify sleep onset and sleep offset. Epochs were scored as wake if activity counts were above 40 (medium sleep-wake threshold).

Subjective sleep estimates were assessed using the Consensus Sleep Diary-E (Carney et al., 2012). In the morning, the sleep diary was filled in immediately following awakening, and in the evening shortly before switching the lights off. Primary measures of interest were sleep onset latency (min), and secondary measures were wake after sleep onset (min), number of awakenings (#), total sleep time (h:min), subjective sleep quality (scale 1–10), and the feeling of being refreshed (scale 1–10). Lastly, alertness/sleepiness was assessed upon awakening and before bedtime using the Karolinska Sleepiness Scale (KSS) (Akerstedt & Gillberg, 1990). Scores ranging from 1–9, with higher scores indicating higher sleepiness.

Behavioural guidelines and evaluation of the glasses

To allow for a fair comparison between conditions, participants were instructed to follow a list of behavioural guidelines, including: (1) a regular sleep-wake pattern that was standardized within individuals, but could differ between individuals; (2) restricted alcohol and caffeine consumption to 300 mg caffeine and a maximum of two alcohol consumptions a day, with a weekly maximum of five alcohol containing beverages being tolerated.

During both conditions, sleep hygiene practices and compliance with the behavioural guidelines were monitored on a daily basis, by means of the evening section of the Consensus Sleep Diary-E (Carney et al., 2012), and an adapted version of the Sleep Hygiene Index (SHI; Mastin, Bryson, & Corwyn, 2006) (yes/no answer format). The SHI asks participants to report the presence of several environmental characteristics and engagement in broad categories of physiologically and psychologically activating evening behaviour that may potentially disturb sleep. Sleep hygiene items on staying in bed longer (item #5) and on sleep environment (item #10, 11) were removed, while items on sleep location (home/away), and on having a bed- or room partner (yes/no) were added.

In order to evaluate the convenience of the glasses, each evening, participants had to rate the usability and comfort of the respective evening glasses on a scale from 1 to 10, on which 10 indicates high usability/comfort.

Data processing and statistical analysis

The first two habituation nights and, due to technical issues¹, the first three nights of each condition were omitted from analysis, leaving four nights per individual and per condition. Actigraphy- and diary-based sleep estimates, sleep hygiene scores, and ratings on usability/comfort were averaged across participants and conditions. Actigraphy and diary-based sleep onset latencies followed a non-normal distribution and were log10-transformed. Compliance to the behavioural guidelines and to sleep hygiene practices was compared between conditions using a paired t-test. The effectiveness of light-restriction as a means to improve sleep estimates was assessed using one-way repeated measure ANOVA’s. Based on our within-subject design, intervention effects directly follow from comparing the LR and nLR conditions (main effect of condition).

Results

Sleep hygiene practices and protocol compliance

Preliminary analysis showed no significant differences between conditions for the respective sleep hygiene items (all p’s > .068, Table I), indicating that the light-restriction and the no-light-restriction conditions were performed under similar environmental and behavioural circumstances. Table I also displays the frequency of various sleep hygiene practices across all measurement days within each condition (i.e. showing the percentage of days on which specific practices occurred). Compliance with regular sleep-wake patterns was fair, as displayed by the lights-off and lights-on time shown in Table II. For both actigraphy- and diary-based estimates, lights-off and lights-on times did not differ across conditions, with all p’s > .560.

Table I. Sleep hygiene practices.

	No-light restriction			Light restriction			Results of statistical analysis
	M	SD	%	M	SD	%	Df	t	p
Daytime naps lasting two or more hours, or past 3 PM	.07	.21	6.90	.08	.18	8.33	14	.231	.820
Went to bed 30 min earlier or later than yesterday	.33	.35	32.76	.30	.33	30.00	14	−.280	.784
Got-up 30 min earlier or later than yesterday	.25	.35	24.14	.30	.24	30.00	14	.459	.654
Exercised within the last hour before bedtime	.12	.28	12.07	.08	.22	8.33	14	−.619	.546
Used alcohol, tobacco, caffeine within the last four hours before bedtime	.28	.34	30.77	.23	.35	23.33	14	−.676	.510
Did something that may wake me up before bedtime	.66	.26	65.52	.60	.41	60.00	14	−.603	.556
Went to bed feeling stressed, angry, upset, or nervous	.10	.21	10.34	.15	.23	15.00	14	.899	.384
Use my bed for things other than sleeping or sex	.41	.41	39.66	.47	.38	46.67	14	.577	.573
I did important work before bedtime	.18	.29	18.87	.05	.10	5.00	14	−1.586	.135
I thought, planned, or worried when I was in bed	.02	.06	1.72	.13	.28	13.33	14	1.974	.068
Slept at home	.98	.15	98.28	.92	.15	91.67	14	−1.262	.228
Had a bed or room partner	.04	.10	3.45	.08	.18	8.33	14	.953	.357

Note. The answer format of all questions was Yes (1) or No (0). Statistics are based on mean values rather than percentages. The percentages indicate how often a behaviour occurred across all nights within each condition.

Table II. Descriptive statistics and results of the statistical testing for actigraphy-based and diary-based sleep estimates (self-report).

	No-light restriction		Light restriction		Results of statistical analysis
	M	SD	M	SD	Df2	F	p	η2p
Actigraphy-based
Lights-off (hh:min)	00:03	01:02	23:57	00:54	13	.358	.560	.027
Lights-on (hh:min)	08:11	00:37	8:12	00:49	13	.003	.956	.000
SOL (hh:min)	00:11	00:12	00:06	00:03	13	1.117	.310	.079
WASO (hh:min)	00:45	00:18	00:47	00:17	13	.139	.715	.011
Fragmentation (%)	27.89	9.21	28.65	8.36	13	.255	.622	.019
TST (hh:min)	07:10	01:02	07:17	00:38	13	.608	.450	.045
SE (%)	85.58	5.76	85.57	5.12	13	.000	.998	.000
Diary based (self-report)
Lights-off	23:55	00:55	23:46	00:45	14	1.022	.329	.068
Lights-on	08:09	00:32	08:14	00:40	14	.242	.630	.017
SOL (hh:min)	00:19	00:11	00:12	00:07	12	11.607	.005	.492
WASO (hh:min)	00:13	00:09	00:11	00:08	9	.564	.472	.059
Awakenings (#)	1.74	.67	1.23	.42	9	3.482	.095	.279
TST (hh:min)	07:30	00:36	07:49	00:40	13	2.740	.122	.174
Sleep Quality
Sleep Quality (1–10)	6.91	1.02	7.48	.76	14	6.106	.027	.304
Refreshed (1–10)	6.43	1.25	6.94	1.14	14	4.45	.053	.241
KSS evening (1–9)	6.40	1.03	5.95	1.15	14	2.878	.112	.171
KSS morning (1–9)	4.23	1.32	3.62	1.02	14	4.634	.049	.249

Evaluation of the comfort of the glasses revealed that the amber-lens glasses (light-restriction condition; M ± SD; 6.72 ± 1.22) were rated about equally comfortable as the transparent glasses (no-light restriction condition; M ± SD; 6.96 ± 1.11). Also in terms of usability, the glasses were rated equally, with M ± SD; 7.33 ± 1.13 for the transparent glasses and M ± SD; 7.07 ± 1.32 for the amber-lens glasses. Taken together, these data indicate that it is unlikely that the comfort and usability of the different types of glasses impacted the current results.

Sleep estimates

A full overview of sleep estimates (means, standard deviations and test outcomes) is displayed in Table II. None of the actigraphy-based sleep estimates differed between the conditions (all p’s > .310). Diary-based sleep estimates, however, revealed shorter subjective sleep onset latencies and better subjective sleep quality in the light-restriction condition. Specifically, subjectively reported sleep onset latency was 7 minutes shorter in the light-restriction condition, with F(1,12) = 11.607, p = .005, η2p = .492. In addition, self-reported sleep quality was 0.6 points higher in the light-restriction condition compared to the no-light restriction condition, with F(1,14) = 6.106, p = .027, η2p = .304. Lastly, participants felt more alert in the morning following evenings on which the amber-lens glasses were worn, as compared to the transparent glasses (KSS; Δ = 0.61), with F(1,14) = 4.634, p = .049, η2p = .249. In the evening, ratings of sleepiness showed no significant difference between conditions (KSS; p = .112). None of the remaining sleep estimates differed significantly between conditions (all p’s > .053, Table II).

Discussion

The current pilot-study set out to investigate the effect of blocking short-wavelength light in the evening on improving sleep among recreational athletes. In line with the hypothesis, blocking short-wavelength light resulted in shorter subjective sleep onset latencies, better subjective sleep quality and higher subjective alertness in the morning, as compared to habitual evening light exposure (no light-restriction).

In line with previous studies suggesting that habitual melatonin secretion may be preserved by means of evening light restriction (Burkhart & Phelps, 2009), blocking-short wavelength light was mainly effective in shortening subjective sleep onset latency, but did not impact on secondary sleep estimates, such as total sleep time or wake after sleep onset. The observation that the sleep permissive effects of light-restriction were limited to subjective estimates is probably not surprising given the low wake-detection capacity of actigraphy (Chae et al., 2009; Paquet, Kawinska, & Carrier, 2007), and usually small correlations between subjective and objective measures of sleep (Lockley, Skene, & Arendt, 1999). Although future studies are thus required to examine effects on objective sleep estimates in more detail (preferably using polysomnography), the current results are in line with findings of several studies which show that exposure to short-wavelength light in the evening (e.g. through the use of electronic devices; Grønli et al., 2016; Heath et al., 2014; Jones et al., 2018; Rångtell et al., 2016) is not always reflected in objective measures of sleep.

Importantly, the observed decrease in subjective sleep onset was accompanied with an increase in subjective sleep quality and morning alertness (see Table II). Subjective sleep quality and alertness in the morning are considered primary outcomes in evaluating insomnia complaints (Morin & Benca, 2012), and improvements bear relevance for recreational athletes, as good subjective sleep quality and morning alertness can increase motivation and decrease the perception of effort (Hull, Wright, & Czeisler, 2003).

From a practical perspective, the current results indicate that selective evening short-wavelength light restriction may prove to be a cost-efficient and accessible way to improve subjective markers of sleep in recreational athletes, requiring only small behavioural adjustments and no prior training. The high ease of use of amber-lens glasses facilitates high ‘therapy compliance’, especially when their design allows to wear them in public without attracting much attention. Given the promising results and the potential high therapy compliance, future research should further investigate and attempt to increase the effectiveness of short-wavelength light-restriction for optimizing sleep in recreational athletes. The current effects may be strengthened by additionally increasing morning- and daytime light exposure, which has already been shown to facilitate circadian entrainment and improve sleep estimates among populations with circadian rhythm- and neurological disorders (Khalsa, Jewett, Cajochen, & Czeisler, 2003; Van Someren, Kessler, Mirmiran, & Swaab, 1997; Wams et al., 2017). Moreover, because many devices that contribute to evening light exposure (e.g. smartphone, tablets) are associated with increased anxiety, stress, arousal and delayed bedtimes (Chang et al., 2015; Romyn et al., 2016), results may be further strengthened if the stimulating effect of media use can be limited by ‘unplugging’ within the last hour before bedtime, and by keeping technologies out of the bedroom (e.g. Halson, 2016; West et al., 2011).

Limitations encountered in the current study should be addressed in future research. As such, employing a measure of compliance and monitoring the duration during which the glasses were worn before bedtime (protocol compliance), obtaining a more detailed measure of daytime light exposure (e.g. electronic device use), and utilizing markers of circadian phase, such as melatonin or core body temperature, may prove useful in determining the optimal duration and timing of light restriction. Furthermore, in line with previous studies, effects of light manipulation were visible in subjective but not in actigraphy-based measures of sleep (see also: Grønli et al., 2016; Heath et al., 2014; Jones et al., 2018; Rångtell et al., 2016). Although participant instructions were designed to prevent outcome expectancies from influencing our findings, it is therefore important that the current findings be replicated and extended with objective measurements. Especially since effects are likely to concentrate around sleep onset, employing sensitive objective measures of sleep, such as the gold standard Polysomnography, is highly recommended. Finally, future studies may attempt to extend the current findings to an elite athlete population (among whom evening exposure to short-wavelength light has been found to be particularly prevalent; Knufinke, Nieuwenhuys, Geurts, Coenen, & Kompier, 2018; Romyn et al., 2016) and examine whether positive effects on sleep also bear implications for recovery and performance.

Conclusion

In addition to current approaches aiming to improving sleep in athletes (see Bonnar et al., 2018 for a recent review), the current pilot-study suggests that restriction of short-wavelength light in the evening can be an effective means to improve subjective sleep onset latency, sleep quality and alertness in the morning among non-sleep disordered, recreational athletes in their habitual home-environment.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research was funded by the STW Technology Foundation, The Netherlands under Grant number [STW 12865].

Notes

1. During the first days of the protocol, participants pilot-tested a prototype of a light-emitting morning goggle to be used in later (follow-up) studies. Due to technical issues in the manufacturing of the morning goggles, this part of the study was aborted after night 2. To prevent any potential effect on our results – and before processing or analysing the data – we conservatively excluded night 1–3 in both conditions for all participants.