Blue-Enriched White Light Improves Performance but Not Subjective Alertness and Circadian Adaptation During Three Consecutive Simulated Night Shifts

Abstract

Use of blue-enriched light has received increasing interest regarding its activating and performance sustaining effects. However, studies assessing effects of such light during night work are few, and novel strategies for lighting using light emitting diode (LED) technology need to be researched. In a counterbalanced crossover design, we investigated the effects of a standard polychromatic blue-enriched white light (7000 K; ∼200 lx) compared to a warm white light (2500 K), of similar photon density (∼1.6 × 1014 photons/cm2/s), during three consecutive simulated night shifts. A total of 30 healthy participants [10 males, mean age 23.3 (SD = 2.9) years] were included in the study. Dependent variables comprised subjective alertness using the Karolinska Sleepiness Scale, a psychomotor vigilance task (PVT) and a digit symbol substitution test (DSST), all administered at five time points throughout each night shift. We also assessed dim-light melatonin onset (DLMO) before and after the night shifts, as well as participants' opinion of the light conditions. Subjective alertness and performance on the PVT and DSST deteriorated during the night shifts, but 7000 K light was more beneficial for performance, mainly in terms of fewer errors on the PVT, at the end of the first- and second- night shift, compared to 2500 K light. Blue-enriched light only had a minor impact on PVT response times (RTs), as only the fastest 10% of the RTs were significantly improved in 7000 K compared to 2500 K light. In both 7000 and 2500 K light, the DLMO was delayed in those participants with valid assessment of this parameter [n = 20 (69.0%) in 7000 K light, n = 22 (78.6%) in 2500 K light], with a mean of 2:34 (SE = 0:14) and 2:12 (SE = 0:14) hours, respectively, which was not significantly different between the light conditions. Both light conditions were positively rated, although participants found 7000 K to be more suitable for work yet evaluated 2500 K light as more pleasant. The data indicate minor, but beneficial, effects of 7000 K light compared to 2500 K light on performance during night work. Circadian adaptation did not differ significantly between light conditions, though caution should be taken when interpreting these findings due to missing data. Field studies are needed to investigate similar light interventions in real-life settings, to develop recommendations regarding illumination for night workers. Clinical Trial Registration: www.ClinicalTrials.gov, identifier NCT03203538.

Keywords: Fatigue; alertness; countermeasures; light; light emitting diode; night work; performance.

Copyright © 2020 Sunde, Pedersen, Mrdalj, Thun, Grønli, Harris, Bjorvatn, Waage, Skene and Pallesen.

Figures

FIGURE 1

Double-raster plot of the simulated night work protocol. Clock hour is indicated on the x-axis and study day on the y-axis. The night work protocol included two study periods with three simulated night shifts (from 23:00 to 06:45 h) performed in a laboratory with different lighting conditions. (A) 2500 K light. (B) 7000 K light. The study periods were separated by 4 weeks and the order of conditions was counterbalanced. White bars indicate enrollment and practice session (before the first study period only) in the laboratory. Black bars indicate assumed baseline sleep at home. Colored bars indicate night shifts in the laboratory. Gray hatched bars indicate assumed daytime sleep at home. Black dots indicate primary test bouts including the Karolinska Sleepiness Scale (KSS), a Psychomotor Vigilance Task (PVT), and a Digit Symbol Substitution Test (DSST). White diamonds indicate salivary dim-light melatonin sampling at home.

FIGURE 2

Spectral distribution of the 7000 K light (solid line) and the 2500 K light (dotted line).

FIGURE 3

Estimated marginal means and standard error plotted as a function of light condition (2500 K vs. 7000 K light), night shift, and time of testing. (A) Rating on the Karolinska Sleepiness Scale (KSS). (B) Mean reciprocal response time (1/RT) on the Psychomotor Vigilance Task (PVT). (C) Number of lapses (RTs ≥ 500 ms) on the PVT. (D) Number of false starts (response without stimulus) on the PVT. (E) Number of correct responses on the Digit Symbol Substitution Test (DSST). *p < 0.05 between light conditions (only for variables with significant light × night × time interactions).

FIGURE 4

Estimated marginal means and standard error plotted as a function of light condition (2500 K vs. 7000 K light) and time of testing (all night shifts included). (A) Response times (RT) for the 10% fastest RTs on the Psychomotor Vigilance Task (PVT). (B) Mean resiprocal RTs (1/RT) for the 10% slowest RTs on the PVT. *p < 0.05 between light conditions.

FIGURE 5

Phase markers for individual participants before (baseline) and after (final) three consecutive night shifts. (A) Night shifts in 2500 K light. (B) Night shifts in 7000 K light. Open circles indicate salivary dim-light melatonin onset (DLMO) for each participant. Filled diamond squares indicate estimated temperature minimum (DLMO + 7 h) for each participant. Lines are drawn between the baseline and final markers for each participant with complete baseline and final markers. The vertical dotted lines and the horizontal bars indicates the start and end times of the night shifts and light exposure.

FIGURE 6

Estimated marginal means and standard error plotted as a function of light condition (2500 K vs. 7000 K light) and daytime sleep period (after night shift 1–3). Sleep variables were derived from actigraphy. Estimates are provided as clock time (hh:mm) for (A,C) and duration (h:mm) for (B,D). No statistically significant differences between light conditions were found.