Supplemental Information
Entrainment of the human circadian clock to the natural light-dark cycle
Kenneth P. Wright Jr, Andrew W. McHill, Brian R. Birks, Brandon R. Griffin, Thomas Rusterholz, Evan D. Chinoy
1. Inventory of Supplemental Information ( Figures S1, S2 and S3, Table S1, Supplemental Experimental Procedures and References).
Fig S1 is related to main Fig 3
Fig S2 is related to main Fig 3
Fig S3 is related to main Figs 2 and 3
Table S1 is related to main Figs 2 and 3
2. Supplemental Figures and Legends, Tables, Experimental Procedures, and References
Supplemental Figure S1. Histograms of the phase angle of entrainment (Ψ) between the DLMO25% and sleep start time. Prior findings indicate that the Ψ in humans is determined by the intrinsic period of the circadian clock and prior light exposure[S10,S16,S17,S18]. The average Ψ of -2.01±0.63 h (±SD) observed following exposure to a week of typical electrical lighting (Panel A) is consistent with prior findings[S10, S17,S18]. The finding that the Ψ is significantly wider (p=0.019, two tailed) following a week of exposure to the natural light-dark cycle camping (Panel B) -2.82±0.85 h provides evidence that electrical lighting in the constructed environment has altered the biological timing of sleep and wakefulness. Bars represent the number of subjects and their respective Ψ.
Supplemental Figure S2. Association between the timing of the start, middle, and end of the sleep episode with circadian timing. As noted in Table S1, the timing of the DLMO25% was significantly correlated with the average midsleep (Panel A) and sleep end (Panel C) times derived from average weekly actigraphy recordings after exposure to electrical lighting in the constructed environment such that those with a later DLMO25% showed later midsleep and/or sleep end times; consistent with prior research [S6,S7,S8,S9,S10, S11, S12,S13]. After exposure to only natural light there was a non-significant trend (Table S1) for an association between the DLMO25% and midsleep (Panel B) and a significant relationship (Table S1) with sleep end (Panel D). As discussed, the reduction in the strength of the association between circadian and sleep timing after exposure to natural light is likely due to a reduction in the range of scores (Legend Table S1). Symbols represent individual subjects and numbers represent clock time. The solid line represents a linear fit of the data.
Supplemental Figure S3. Physical activity. Average weekly activity counts (arbitrary units) are plotted in 10 minute bins during the week of exposure to electrical lighting in the constructed environment and the week of exposure to the natural light-dark cycle camping. As seen, activity levels are higher during the daytime when camping whereas activity levels are higher after midnight during the week in the electrical lighting-constructed environment; consistent with the light exposure levels in Fig. 2 of the main text. Participants’ average activity levels were 70% higher during wakefulness [557.2±228.5 counts (±SD)] and 14% higher during sleep (15.5±4.8 counts) episodes the week of camping when compared to activity levels during wakefulness (326.3±135.6 counts) and sleep (13.6±5.4 counts) episodes the week of electrical lighting, respectively (p < 0.05; two tailed for condition comparisons). Data are double plotted so that activity counts across midnight (24h local clock time) can be more easily observed.
Supplemental Table S1. Association between the timing of the start, middle, and end of the sleep episode with circadian timing. Sleep is well established to be strongly modulated by circadian phase [S1,S2,S3,S4,S5] and prior findings indicate an association between melatonin circadian phase and the timing of the beginning, midpoint and/or end of the sleep episode [S6,S7,S8,S9,S10, S11, S12,S13], with later circadian phases of evening versus morning types [S6,S9]. The strong association we observed between circadian and sleep timing observed following exposure to a week of typical electrical lighting in the constructed environment is consistent with prior findings. The finding of weaker associations between circadian and sleep timing, most of which were no longer significant, following a week of exposure to the natural light-dark cycle camping is likely due to the reduction in individual differences, or statistically a reduction in the range of scores for both circadian and sleep timing (Supplemental Fig S2). The reduction in the range of scores is likely due to the influence of the strong environmental time cue of sunlight on circadian timing and to changes in sleep timing related to social behaviors, removal of electrical lighting at night, as well as increased sunlight and physical activity during the day. The relative contribution of these factors cannot be determined by the current study and requires follow-up in future studies. *Midsleep is calculated as the midpoint between actigraphy derived sleep start and end times and is not to be confused with midsleep timing on free days (MSFc)[S14,S15], which is thought to be an indicator of chronotype. Data are rounded.
Circadian and sleep
measures correlated
Exposure to electrical plus natural light in the work-home-social constructed environment
Pearson r (p value; two-tailed)
Exposure to only natural light while camping
Pearson r (p value; two-tailed)
DLMO25%
Sleep Start 0.93 (< 0.001) 0.49 (0.21)
Midsleep*
0.96 (< 0.001) 0.70 (0.053)
Sleep End
0.93 (< 0.001) 0.80 (0.017)
MP25%
Sleep Start
0.88 (0.004) 0.28 (0.50)
Midsleep
0.93 (< 0.001) 0.30 (0.48)
Sleep End
0.92 (0.001) 0.28 (0.50)
DLMOff25%
Sleep Start
0.77 (0.027) -0.13 (0.75)
Midsleep
0.82 (0.013) -0.29 (0.49)
Sleep End
0.83 (0.01) -0.39 (0.34)
