How Smart Is It to Go to Bed with the Phone? The Impact of Short-Wavelength Light and Affective States on Sleep and Circadian Rhythms

Abstract

Previously, we presented our preliminary results (N = 14) investigating the effects of short-wavelength light from a smartphone during the evening on sleep and circadian rhythms (Höhn et al., 2021). Here, we now demonstrate our full sample (N = 33 men), where polysomnography and body temperature were recorded during three experimental nights and subjects read for 90 min on a smartphone with or without a filter or from a book. Cortisol, melatonin and affectivity were assessed before and after sleep. These results confirm our earlier findings, indicating reduced slow-wave-sleep and -activity in the first night quarter after reading on the smartphone without a filter. The same was true for the cortisol-awakening-response. Although subjective sleepiness was not affected, the evening melatonin increase was attenuated in both smartphone conditions. Accordingly, the distal-proximal skin temperature gradient increased less after short-wavelength light exposure than after reading a book. Interestingly, we could unravel within this full dataset that higher positive affectivity in the evening predicted better subjective but not objective sleep quality. Our results show disruptive consequences of short-wavelength light for sleep and circadian rhythmicity with a partially attenuating effect of blue-light filters. Furthermore, affective states influence subjective sleep quality and should be considered, whenever investigating sleep and circadian rhythms.

Keywords: affective states; blue-light filter; cortisol; distal-proximal skin temperature gradient; light exposure; melatonin; short-wavelength light; sleepiness; slow wave activity; slow wave sleep.

Figure 1

Trajectory of subjective sleepiness (mean and 95% confidence intervals). Subjects were significantly more tired at awakening and by trend more tired 30 min post-awakening after reading a book compared to reading on a smartphone with a filter. Yellow background = light exposure (reading session); gray background = lights turned off (sleep). *: p ≤ 0.05; °: p < 0.10; †: p_adj. > 0.10.

Figure 2

Cortisol awakening response (mean and 95% confidence intervals). The AUCi was calculated for three time points in the morning, i.e., at awakening as well as 30 min and 60 min after awakening. AUCi was significantly smaller in the “no filter” compared to both other light conditions. The dashed lines connect the measurements from each subject. *: p ≤ 0.05; †: p_adj. > 0.10.

Figure 3

Trajectory of baseline corrected salivary melatonin concentration (mean and 95% confidence intervals). Melatonin increased, by trend, less in the “filter” compared to the “book” condition after 30 min and significantly less after 60 min of light exposure and at bedtime. Melatonin increased significantly less in the “no filter” compared to the “book” condition starting after 60 min of light exposure and persisted until bedtime. Yellow background = light exposure (reading session); gray background = lights turned off (sleep). *: p ≤ 0.05; °: p < 0.10; †: p_adj. > 0.10.

Figure 4

Trajectory of the distal-proximal gradient (mean and 95% confidence intervals). DPG was significantly lower in the “no filter” condition compared to the “book” condition from 3:00 to 4:30 and by trend lower in the “filter” condition compared to the “book” condition at 3:00. Yellow background = light exposure (reading session); gray background = lights turned off (sleep). *: p ≤ 0.05; °: p < 0.10; †: p_adj. > 0.10.

Figure 5

Time in SWS (mean and 95% confidence intervals). The amount of SWS in the first night quarter was significantly reduced in the “no filter” condition as compared to the “filter” and “book” condition. In the second night quarter, the amount of SWS was significantly reduced in the “filter” condition compared to the “no filter” and by trend compared to the “book” condition. *: p ≤ 0.05; °: p < 0.10; †: p_adj. > 0.10.

Figure 6

SWA in the first night quarter (mean and 95% confidence intervals). SWA was significantly reduced in the “no filter” compared to the “book” condition at frontal, central, parietal and occipital electrodes. Additionally, SWA was significantly reduced in the “no filter” compared to the “filter” condition at frontal, central and parietal derivations. *: p ≤ 0.05; †: p_adj. > 0.10.

Figure 7

Study design [64]. (A) Overview of the general procedure, which covered a period of 13 days, including an entrance examination, one adaptation night and three experimental nights. (B) Detailed illustration of the study protocol in the experimental night. During the whole study period, participants’ sleep-wake rhythm was monitored by means of wrist actigraphy and sleep diary.