Electric Lighting, Adolescent Sleep and Circadian Outcomes, and Recommendations for Improving Light Health

Light Exposure in School

School comprises a substantial portion of adolescents’ schedules, ~33 hours per week in the U.S. (49). Educational lighting standards (e.g. IES RP-3–20) were developed to support perceptual vision (e.g., minimizing glare, improving visibility, providing sufficient horizontal and vertical illuminance for specific settings and tasks, incorporating and accounting for daylighting), but are not optimized for non-image forming vision and circadian needs (50, 51). Reduced short-wavelength light exposure (using orange-tinted glasses to block short-wavelength light) from rise time until the end of the school day produced delays in dim light melatonin onset (DLMO) (52, 53), demonstrating the importance of obtaining sufficient short-wavelength light during the school day.

This has implications for adolescents residing in highly structured or confined settings, including boarding schools, juvenile detention facilities, immigration detention centers, and residential treatment facilities. Juvenile detention centers were shown to have lower illuminance (124 lux v. 297 lux), fewer windows (2 v. 7), and earlier lights-on (6:07 am v. 6:54 am) and lights-off (8:42 pm v. 9:06 pm) times relative to treatment centers (54). Additionally, adolescents from juvenile detention centers reported bedtimes approximately 50 minutes earlier than their sleep onset times (55). This has implications for misalignment between the delayed circadian timing present in many adolescents and demands on sleep-wake schedule (i.e., early bed and rise times) for youth in juvenile detention centers, and resulting insomnia (54). Future experimental investigation is warranted to examine objective effects of electric light exposure patterns in structured adolescent institutional residential settings on sleep and circadian outcomes to establish the degree to which lighting interventions are needed.

Outdoor Electric Light at Night

Light at night (LAN) is a concern (56) for sleep-wake health, especially where efficient electrical grids and lighting promote year-round street lighting throughout the night in urbanized parts of the world (57). While high pressure sodium lighting is the most prevalent form of street lighting, LED street lighting use is increasing (56, 58) and may produce greater circadian or sleep disruptions due to the short-wavelength spectral peak inherent to LED technology (59): One hour of 18 lux or 27 lux white 6900 K LED light was predicted to stimulate the circadian system to a small extent (i.e., predicted 12% – 15% melatonin suppression), while metal halide (4000 K), high-pressure sodium (2050 K), and two cool white LED light sources (5200 K; 6900 K) did not suppress melatonin (60). Satellite measures of outdoor LAN were associated with increased subjective eveningness and later midpoint of sleep, after accounting for age and sex (61), and a later subjective weeknight bedtime, after controlling for sociodemographic factors in adolescents (62). These studies are limited by a lack of individual light-level readings, as nocturnal satellite readings do not correlate with individual indoor light exposure levels in adults or 11-to-13-year-old children (63, 64), except during the darkest period of night and only in youth whose parents reported that outdoor light occasionally or often influences indoor light levels (weak correlation). (63). However, satellite measures may still add value as they reflect community-level differences in LAN exposure (e.g., satellite LAN readings were correlated with increased air pollution, urbanicity, population density, reduced greenery, and lower socioeconomic status in youth) (63, 65). An expert panel recommended measuring LAN with calibrated devices whenever possible and using these data to validate questionnaires to serve as a reliable proxy for objective LAN measures in epidemiological studies (64).

Light-emitting Electronic Devices

A number of studies have evaluated the relationship between light-emitting electronic devices and adolescent sleep. Cross-sectional studies show device usage is associated with later bedtimes, longer sleep onset latency, reduced sleep duration, and increased sleep problems (66). Mobile phones may account for much of the impairment, with increased duration and later timing of use associated with poorer sleep quality in adolescents (67). However, this relationship may be mediated by several mechanisms: Using electronic devices before bed or in bed may directly displace sleep with the user remaining awake longer than without the device (68, 69). Some adolescents may engage with the device due to difficulty initiating sleep (68), and the device itself may have an alerting or physiologically arousing effect, delaying the settling process that promotes sleep onset (69). The LED screens in many electronic devices emit short-wavelength light that is known to delay circadian timing with evening exposure, and there have long-been concerns that light-emitting devices contribute to suppression of slow wave sleep in the first non-rapid eye movement cycle and delays in sleep and circadian timing (69–71). While light intensity and spectrum can both be of concern with these devices, it is likely that light intensity, rather than spectrum, is the dominant factor contributing to disruption (72) under real-world device usage.

Laboratory-controlled studies have evaluated the impact of light exposure from various electronic devices. A comparison of 1 h of evening exposure to bright (80 lux), dimmed (1 lux), and short-wavelength filtered (50 lux) tablet light exposure in adolescents showed no differences in a variety of sleep measures (73). In a separate study, exposure to 2 h (but not 1 h) of self-luminous tablet light (40 lux) at maximum brightness among adolescents and young adults was associated with increased melatonin suppression (74). These findings suggest that extended evening light exposures can inhibit the evening rise in melatonin, but the impacts on adolescent sleep are unknown.

Addressing the lack of ecological validity in laboratory studies, a home-based study in adolescents showed that viewing various self-luminous devices (average 87 lux) without orange-tinted (blue-light-blocking) glasses for 2 h on one evening was associated with increased melatonin suppression relative to viewing them through orange-tinted glasses for 3 h (75). Further, viewing self-luminous devices without orange-tinted glasses was associated with melatonin suppression following exposure for 1 and 2 h. Adolescents appear sensitive to light from electronic devices at 1 and 2 h of exposure in the home environment. However, it is unclear how device type, style of device engagement, and home lighting influenced outcomes (75).

A further home-based study in undergraduate students showed no main effects for tablet light exposure (unfiltered [< 200 lux] or short-wavelength filtered light [200 lux]) or stimulation level (participant’s Facebook account versus a sham Facebook account designed to be unstimulating), but found that the short-wavelength filtered and low stimulation condition was associated with significantly better self-reported sleep quality, sleep duration, sleep onset latency, and daytime function relative to the other conditions (76). The pattern of findings suggests that engaging in stimulating tasks while using electronic devices before bed disrupts sleep regardless of spectrum, but only the modification of stimulation level and light spectrum in tandem is associated with improved subjective sleep and daytime functioning.

Home Lighting

Although there has been minimal investigation of home lighting exposure among adolescents, studies in rural populations with and without electric home lighting are informative. Adolescents in rural Brazil with electric home lighting showed reduced sleep duration on school days relative to weekends and across the week, later sleep onset on school days relative to adolescents without electric home lighting (77, 78). However, the duration of evening light exposure is an important consideration. As a benchmark, 30 minutes of exposure to 30 lux is suggested as the threshold at which light begins to stimulate the circadian system (predicting 15% melatonin suppression) for incandescent lighting exposure in the home among women (79). In young adults, brighter home lighting was associated with later DLMO relative to dim home lighting, indicating the circadian phase-delaying influence of increased light intensity (80). Laboratory-controlled research corroborates these findings, indicating exposure to 3 h of fluorescent evening bright light (2000 lux) suppressed the rise in melatonin, whereas exposure to 3 h of dim light (60 lux) did not (81).

Wearable spectrometers indicate melanopic illuminance in the 3 h prior to bedtime can vary more than 10-fold (3.9–77.4 melanopic lux) across homes (82). The melanopic illuminance required for 50% melatonin suppression ranged from 3.1 to 181 melanopic lux, suggesting a high degree of inter-individual variability in sensitivity to any single home lighting setting (83). Relative to sunset, which showed rapid decreases from 181 to 3.1 melanopic lux within 22 minutes, the melanopic content of home lighting persisted through the evening, with an average of 13 melanopic lux observed at 10:00 pm, suggesting home lighting extends sunset in terms of its impact on the circadian system. Higher melanopic illuminance within the 3 h before bedtime was associated with increased wakefulness in the 90 min following bedtime. Higher melanopic illuminance was measured in homes with fluorescent and LED lighting, relative to incandescent lighting, suggesting that newer lighting technologies can negatively impact sleep and circadian health (83).

This is also relevant for sleeping environments within institutional residential facilities. Fifty-eight percent of youth across multiple juvenile detention facilities reported partial or overhead lights were turned on in spaces designated for sleep, and 23% endorsed the presence of “other” types of lighting, with the majority of those “other” sources being ‘blue’ light. Youth reported night awakenings lasting 16.8 minutes, and sleep quality and difficulty rising in the average range (55). Although descriptive, findings suggest the presence of sleep- and circadian-disrupting lighting practices in juvenile detention facilities; and these may extend to other types of institutional residential facilities.

Light Exposure across the Day and Evening

Beyond the immediate impact of light exposure delivered in specific contexts, cumulative light exposure history is important: increased daytime light exposure is associated with increased evening melatonin production and advanced circadian phase (84, 85). This may in part be due to daytime light reducing subsequent sensitivity to evening light (86, 87). In a small (n = 15) sample of adolescents, higher prior-day mean light intensity predicted earlier sleep onset but not offset, and longer sleep duration (88). A bidirectional association was found, with later prior-night sleep offset predicting lower mean light intensity. Similarly, evening chronotype adolescents with lower intensity light exposures across the preceding 24 h had later sleep onset and offset. Later last-exposure >10 lux of light associated with later sleep onset and bedtime and a reduction in total sleep time (89). Reduced morning light was associated with later sleep onset, next-day offset, and rise time (see Figure 1). Decreased afternoon-to-early-evening light was associated with later sleep onset, and increased evening-to-late-night light was associated with decreased total sleep time (89).

In a case-control study over 2 weeks, adolescents with delayed sleep-wake phase disorder (DSWPD) received less morning light and more evening light relative to controls. Interestingly, evening light levels (8:00 pm to 5:00 am) were relatively dim (<48 lux) across both groups and all week days (school and non-school), with less morning light on school relative to non-school days. As adolescents slept in dim light, above findings suggest light exposure patterns among groups were well aligned with group differences in sleep schedule, and school versus non-school days (90). When data were reanalyzed relative to sleep onset rather than clock time, adolescents with DSWPD instead had less pre-sleep light exposure in the 9 h prior to sleep onset than controls, and showed no difference in light exposure in the 9 h post-sleep offset. The authors suggest increased evening light may not drive phase delays in adolescents with DSWPD, and instead reduced sensitivity to the advancing effects of morning light exposure may contribute (90). As individuals with DSWPD have longer circadian periods, resulting in a greater tendency to delay circadian rhythms relative to controls, the combination of decreased evening light and comparable morning light may serve to stabilize their rhythm, although at a later clock time (91).

These studies show that, especially for circadian disorders, individual differences in circadian physiology and activity preferences can reveal information otherwise obscured when data are referenced only to solar time. Further, it is important to understand the compounding effects of photic history over time. While strong daytime light exposures are protective against night time light exposures, day-to-day reinforcement of circadian amplitude would result in improved circadian amplitudes and robustness to perturbation by evening light. The temporal kinetics of these processes have yet to be determined in health or disease, but will better explain how photic history can attenuate (or facilitate) the acute effects of lighting.

Supplementing Light Therapy

Several studies have examined the influence of supplementing morning LT with adjunctive interventions. Morning physical activity is known to advance circadian phase, but its potential for optimizing LT is still unclear (104). Morning short-wavelength (‘blueish-greenish’) or long-wavelength (‘redish’) LT with morning physical activity or sedentary activity in adolescents and young adults with DSWPD produced no significant group differences in improvements in sleep timing, sleep onset latency, and daytime functioning (104). However, the lack of objective increase in physical activity in the expected groups, and inclusion of three weekly sessions centered on advancing sleep-wake schedule across groups limits identification of the drivers of noted improvements (104).

Exogenous melatonin, which has soporific effects in some when administered 30–60 min prior to sleep, and circadian phase advancing or delaying effects when administered in the early evening or early morning hours, respectively, has also been investigated as an adjunctive treatment (105). However, 2 weeks of bright or dim LT with melatonin or placebo, along with gradual advancement of rise times across these four groups in adolescents and young adults yielded no significant group differences in improvements in DLMO, sleep disturbance, sleepiness and fatigue in all groups (106, 107). Three additional months of combined bright LT and melatonin were associated with greater reductions in sleep disturbance, sleepiness, and fatigue relative to no additional treatment (106, 107), suggesting the benefit of longer treatment duration.

A few studies have investigated LT with CBT (i.e., psychoeducation, cognitive restructuring, motivational interviewing, management of daytime effects of sleep loss, and relapse prevention) for youth with DSWPD or delayed sleep schedule (103, 108, 109). When compared to waitlist-control, LT with CBT showed preliminary efficacy in improving sleep onset latency, sleep onset time, total sleep time on school nights, wake after sleep onset, rise time, daytime sleepiness, and fatigue (108). However, relative to LT alone, LT with CBT was associated only with greater reductions in severity of sleep disturbance (109). The authors postulated that basic sleep education provided to both groups at treatment outset may have washed out CBT effects (109). A two-phase trial in adolescents addressed this research question from multiple perspectives. Results showed that pulsed LT produced greater improvements in sleep efficiency relative to sham-control, whereas pulsed LT with CBT was associated with improvements in a number of sleep measures relative to sham-control with CBT (103). These studies suggest LT alone may not be sufficient to produce significant change in sleep and circadian timing in adolescents. In contrast, LT with CBT produces more robust effects on sleep timing and measures of sleep disturbance in adolescents relative to waitlist-control and CBT alone, but the degree to which this combined intervention improves upon LT alone may require direct comparison in future studies. CBT provides additional structure and motivational enhancement to LT, which have utility for adolescent behavioral change.

Indeed, motivational components of increased desire and commitment to change are associated with greater adherence to advancing rise time schedule among adolescents and young adults receiving wearable, short-wavelength LT (110). Beyond individually-delivered CBT, school-based sleep education programs (SEP) provide wider treatment dissemination with limited provider involvement. Such SEP could also be administered to adolescents as part of institutional residential facility (e.g., boarding schools, treatment center, juvenile detention center) programming. Employing a motivational framework within SEP increased knowledge and motivation, but not sleep or daytime sleepiness relative to a control intervention (82). Efforts to supplement SEP with one weekend of wearable, short-wavelength LT did not bolster treatment effects on sleep knowledge, sleep outcomes, or motivation to engage in treatment components relative to SEP with parental involvement, SEP with parental involvement and LT, and class-as-usual; although a low dose of the adjunctive interventions may have contributed to the lack of group differences (111).

Interventions and Corrective Agents to Minimize Evening Light Exposure

An additional consideration is lighting interventions and corrective agents minimizing evening light exposure. University students exposed to one month of evening incandescent light exposure with a tryptophan and vitamin B6-rich breakfast (to stimulate synthesis of serotonin in the day, a melatonin precursor, and subsequent melatonin synthesis in the evening) followed by morning sunlight, had higher melatonin concentration relative to those receiving breakfast with sunlight alone, and no intervention (112). Additionally, 10 nights of 1-hour of pre-bedtime exposure to a low CCT (2000 K) 160 lux white LED lighting intervention administered to adolescents within the natural bedroom environment was associated with increased subjective sleep quality and decreased morning sleepiness and fatigue in response to 15 minutes of exercise relative to a high CCT (6000 K) 160 lux fluorescent lamp (113). Together, these findings suggest the importance of exposure to evening light of low CCT when other lighting factors are equal. Blue-light-blocking glasses, with amber or orange lenses, are another option (114). Two weeks of evening use from 9:00 pm until bedtime was associated with earlier sleep onset time, but no significant advance in DLMO in adolescents and young adults with DSWPD (115). Adolescents and young adults with sleep disturbance or irregularity showed no significant differences in sleep duration or night awakenings with a one-week course of blue light blocking glasses relative to non-blue-light-blocking glasses (116). In adolescent males, one week of evening use of blue-light-blocking glasses while engaged in device use was associated with significantly less evening melatonin suppression, and reduced vigilant attention and self-reported alertness before bedtime, but no differences in EEG-based or subjective sleep relative to clear glasses (117). Given the lack of associated adverse effects and ease of use over the short-term, blue-light-blocking glasses warrant further investigation to clarify the extent of their benefits, and establish optimal wear time, and long-term feasibility.

Shifting screens to warmer colors (e.g., “night shift mode”) to reduce short-wavelength light exposure is becoming a common feature on personal devices. In young adults, cool (5997 K) and warm (2837 K) CCT modes viewed through a commercial tablet on full brightness (97.8 lux and 54.3 lux, respectively) from a distance of 12 in. both suppressed melatonin less than short-wavelength light-emitting (40 lux of 470 nm light) goggles, but with no effect of CCT mode. This suggests a limited benefit of shifting CCT for preserving evening melatonin secretion without also decreasing illuminance (118). With respect to sleep effects, there were no differences in sleep patterns among 18–24-year-old young adults using phones set to the lowest brightness, with night shift mode enabled on the warmest color setting, no night shift mode, or no phone use for 1 hour prior to bed over 7 days in their home setting. However, for young adults with ≥ 6.8 hours of sleep per night over the week, no phone use was associated with higher sleep efficiency and decreased wake after sleep onset relative to night shift use. Surprisingly, phone use without night shift mode was associated with decreased wake after sleep onset relative to night shift mode use; suggesting that using night shift mode is not advantageous for sleep relative to no night shift mode, nor comparable to no phone use (119). Together, these studies suggest that reducing screen brightness or refraining from light-emitting device use in the hour before bed outweigh the effects of warm color-shifted devices.

Circadian-adapted School Lighting Interventions

The school environment provides a significant opportunity to stimulate the circadian system in adolescents. However, only a few studies have investigated the effects of LED or short-wavelength-enriched lighting in school settings or laboratory-based settings which mirror the classroom. There was no effect of LED (4000 K) relative to fluorescent (4000 K) school lighting type on afternoon cortisol suppression over six months (November to April; winter to spring) in 16–17-year-old students. However, the LED group showed increased afternoon cortisol suppression during darker months (i.e., November). Nevertheless, from the lack of lighting differences in cortisol suppression during brighter months, authors surmised that natural daylight conditions may have more strongly influenced cortisol suppression than the light sources (120). Additionally, morning exposure to short-wavelength-enriched white LED classroom lighting (5500 K, with additional LED modules 14000 K; 300 lux; direct/indirect luminaire) was associated with improvements in processing speed and concentration relative to fluorescent lighting (3000 K and 3500 K; ~300 lux; direct luminaire) in high school students. Unfortunately, only a subsample completed sleep diaries, and among them, there was no effect of lighting system (121).

Despite the potential of circadian-adapted (short-wavelength-enriched) lighting systems to improve daytime functioning among students, there are concerns regarding possible adverse effects (e.g., overstimulation, melatonin suppression in the afternoon hours). This has spurred interest in dynamic lighting systems which alter CCT across the day (122). In a laboratory-controlled experiment, exposure to 10 min of dynamic LED lighting (6500 K short-wavelength-enriched white light to 5000 K neutral white light; illuminance at 800–1000 lux) was associated with increased objective focus and arousal relative to 10 min of constant neutral white light (5000 K at 800–1000 lux) and 10 min of lights-off (< 1 lux) in high school students (122). While some LED lighting produces alerting effects, this is dependent on the luminaire’s design, and laboratory-controlled research needs to determine how the physical parameters of the light (e.g., spectrum, intensity) can be optimized by time of day to enhance circadian entrainment and improve sleep patterns.

A separate laboratory-controlled study showed that morning short-wavelength (‘blue-ish’)-enriched fluorescent lighting (peak at 458 nm; 1063 lux) was associated with enhanced attention on school-like tasks relative to morning long-wavelength (‘red-ish’)-enriched fluorescent lighting (peak at 611 nm; 876 lux); and circadian-consistent lighting (i.e., 1 h of short-wavelength-enriched morning light and 1 h of long-wavelength-enriched evening light) was associated with non-significant trends toward decreased nighttime awakenings. A subsample who received later evening light exposure trended towards reduced sleep onset latency relative to circadian-contrasting lighting (i.e., morning long-wavelength-enriched light and evening short-wavelength-enriched light). Thus, lighting environments during the school day and at home in the evening might improve sleep when they mimic the naturalistic changes in light that occur across the solar day and evening (123). Aforementioned adolescent institutional residential facilities could provide a structured environment through which to test the effects of such lighting regimens on sleep and circadian timing.

Some studies are limited by the measurements reported, as spectral power distribution may differ for light of the same CCT, impacting the light source’s effects on the circadian system. Further, classroom lighting simulations indicate that luminaires with a high vertical illuminance to horizontal illuminance ratio are recommended for overhead classroom lighting to better impact the circadian system, and illuminance has a greater impact on circadian stimulus than CCT (K). Guidelines recommend maximizing energy-efficiency of circadian-adapted classroom lighting through use of supplemental short-wavelength light, luminaires with high vertical-to-horizontal illuminance, and the strategic use of direct and indirect luminaires (124).