Melanopic irradiance defines the impact of evening display light on sleep latency, melatonin and alertness

Abstract

Evening light-emitting visual displays may disrupt sleep, suppress melatonin and increase alertness. Here, we control melanopic irradiance independent of display luminance and colour, in 72 healthy males 4 h before habitual bedtime and expose each of them to one of four luminance levels (i.e., dim light, smartphone, tablet or computer screen illuminance) at a low and a high melanopic irradiance setting. Low melanopic light shortens the time to fall asleep, attenuates evening melatonin suppression, reduces morning melatonin, advances evening melatonin onset and decreases alertness compared to high melanopic light. In addition, we observe dose-dependent increases in sleep latency, reductions in melatonin concentration and delays in melatonin onset as a function of melanopic irradiance-not so for subjective alertness. We identify melanopic irradiance as an appropriate parameter to mitigate the unwanted effects of screen use at night. Our results may help the many people who sit in front of screens in the evening or at night to fall asleep faster, feel sleepier, and have a more stable melatonin phase by spectrally tuning the visual display light without compromising the visual appearance.

Fig. 1. Lighting conditions and experimental protocol of the study.

a Participants were assigned to one of the four light intensity groups (n = 18 per group) that differed in luminance [cd/m²]. b Irradiance of all four light intensity groups [W/(sqm*nm)]. Irradiances of the low melanopic condition (LM) are depicted with orange lines and the high melanopic condition (HM) with blue lines. c Summation of the alpha-opic equivalent daylight illuminance for the three cone types (CIE S 0 26) [lx] in the four intensity groups (S-cones: blue bars; M-cones: green bars; L-cones: red bars). d Melanopic equivalent daylight illuminance (mEDI) in the four intensity groups shows the ~200–300% melanopsin contrast (Contrast = (HM-LM)/LM [%]; Ratio = HM/LM) between the low melanopic (LM: orange bars) and high melanopic (HM: blue bars) condition. e Photo illustrating screen and experimental setup. Participants were sitting at a distance of 60 cm in front of the screen. f The study protocol consisted of one adaptation night and two 17-h blocks either under the LM or HM condition. The order of both conditions was balanced. Both experimental blocks started 7 h prior to habitual bedtime under standard fluorescent light (~67 lx). Before light exposure, there were two episodes of dark adaptation (~0.1 lx) and one dim light episode (~0–7 lx) for adaptation of photoreceptor sensitivity and to collect baseline measurements. Participants were exposed to LM or HM, respectively, for 3.5 h, starting 4 h before habitual bedtime. The light exposure was followed by an 8-h sleep episode, which was polysomnographically recorded (for details, see “Methods” section) and a 1-h dim light episode the next morning. The yellow triangles indicate the timing of the cognitive test batteries. The results of the cognitive tasks will be published elsewhere. Before, during, and in the morning after light exposure, salivary melatonin levels were measured, and subjective levels of sleepiness were rated in half-hourly intervals throughout scheduled wakefulness.

Fig. 2. Sleep latencies to stage N2 in minutes after lights off.

Depicted are back-transformed means for the low melanopic condition (LM: orange points) and the high melanopic condition (blue points) and 95% CIs (black lines) in the four light intensity groups (n = 18 per group, except LM intensity 1 n = 17). Grey points represent the individual values of participants. Black asterisks indicate the statistically significant main effect (p < 0.05) of the fixed factor light condition, computed in a linear mixed model.

Fig. 3. Results for melatonin during light exposure.

a Time course of salivary melatonin concentrations in pg/mL during the low melanopic (LM: orange points and lines) and high melanopic (HM: blue points and lines) light conditions plotted against the hours relative to bedtime [h]. Depicted are means ± 1SEM of the four light intensity groups (Intensity 1: n = 15, intensity 2: n = 14, intensity 3: n = 18, intensity 4: n = 18). In total, data of seven participants were not included in the data analysis due to low melatonin concentrations during the evening (for details see “Methods” section) b Melatonin AUC (area under the curve) in pg/mL/h during light exposure. Depicted are the means ± 1SEM for LM (Mean: orange points; SEM: black lines) and HM (Mean: blue points; SEM: black lines). (Intensity 1: n = 15, intensity 2: n = 14, intensity 3: n = 18, intensity 4: n = 18) c Melatonin onset relative to bedtime (HPB: Hours prior bedtime) during LM (Mean: orange points; SEM: black lines) and HM (Mean: blue points; SEM: black lines). Depicted are means ± 1SEM (Intensity 1: LM n = 15, HM n = 17; intensity 2: LM n = 15, HM n = 15; intensity 3: LM n = 16, HM n = 16; intensity 4: LM n = 16, HM n = 15). Grey points represent the individual values of participants. The black asterisks indicate the statistically significant main effect (p < 0.05) of the fixed factor light condition, computed in a linear mixed model.

Fig. 4. Results of subjective alerting response during light exposure.

a Time course of subjective sleepiness ratings (KSS) during the low melanopic (LM: orange points and lines) and high melanopic (HM: blue points and lines) light conditions plotted against the hours relative to bedtime [h]. Individual sleepiness ratings were corrected with the last sleepiness rating before light exposure, showing the difference to the pre-light exposure level. Depicted are means ± 1SEM (n = 18 per intensity group). b Baseline-corrected subjective sleepiness ratings (KSS) for LM (Mean: orange points; SEM: black lines) and HM (Mean: blue points; SEM: black lines) during light exposure. Values shown are means of eight time assessments ±1SEM. Grey points represent the individual values of participants. The black asterisks indicate the statistically significant main effect (p < 0.05) of the fixed factor light condition, computed in a linear mixed model.

Fig. 5. Dose-response relationships with log10-transformed mEDI [lx] and photopic illuminance [lx].

Averages per light condition over all light intensity groups entered the calculation of the regression models. Dose dependencies were calculated for a Sleep latency [min], depicted are back-transformed means and 95% CIs (n = 18 per group, except LM intensity 1 n = 17) and the regression line (Mean values of all light conditions, n = 8), b Melatonin area under the curve (AUC) [pg/mL/h], depicted are means ± 1SEM groups (Intensity 1: n = 15, intensity 2: n = 14, intensity 3: n = 18, intensity 4: n = 18) and the regression line (Mean values of all light conditions, n = 8), c Melatonin onset in hours prior individual bedtime (HPB), depicted are means ± 1SEM (Intensity 1: LM n = 15, HM n = 17; intensity 2: LM n = 15, HM n = 15; intensity 3: LM n = 16, HM n = 16; intensity 4: LM n = 16, HM n = 15) and the regression line (Mean values of all light conditions, n = 8), d Baseline-corrected subjective sleepiness (KSS), depicted are means ± 1SEM (n = 18 per intensity group) and the regression line (Mean values of all light conditions, n = 8). Orange points show means of the low melanopic and orange lines 95% CIs or SEM. Blue points depict means of the high melanopic and blue lines 95% CIs or SEM. The grey bands represent the 95% confidence interval limits.