Effects of Dinner Timing on Sleep Stage Distribution and EEG Power Spectrum in Healthy Volunteers

Figures & data

Table 1 Percentages of Time Spent in Each Sleep Stage in Each Quarter of the Night

Stage	Routine Dinner	Late Dinner	P value
1st quarter of the night (0–120 minutes)
Wake (%)	17.4 [9.3–36.8]	13.9 [7.9–29.6]	0.60
NREM 1 (%)	6.6 [2.9–10.0]	5.4 [4.1–7.6]	0.24
NREM 2 (%)	32.4 [25–40.6]	31.1 [25.6–40.5]	0.67
NREM 3 (%)	28.8 [17.8–37.4]	34.9 [25.5–45.5]	0.18
REM (%)	4.6 [0–10.4]	3.3 [0.3–9.6]	0.63
2nd quarter of the night (120–240 minutes)
Wake (%)	12.9 [4.0–17.0]	4.8 [2.0–8.3]	0.24
NREM 1 (%)	5.8 [3.4–7.6]	4.2 [2.1–6.4]	0.18
NREM 2 (%)	36.5 [28.9–44.7]	49.0 [41.4–52.0]	0.0056
NREM 3 (%)	28.5 [19.9–40.4]	26.0 [14.8–36.3]	0.39
REM (%)	15.2 [12.1–19.0]	15.6 [12.1–19.4]	0.75
3rd quarter of the night (240–360 minutes)
Wake (%)	3.3 [1.6–6.4]	4.6 [2.0–6.8]	0.84
NREM 1 (%)	4.4 [2.4–6.7]	5.2 [3.2–7.3]	0.36
NREM 2 (%)	52.9 [36.3–60]	45.4 [40.8–54.4]	0.55
NREM 3 (%)	14.6 [5.8–26.9]	10.6 [1.6–15.5]	0.41
REM (%)	21.9 [14.1–30.2]	28.1 [19.5–34.3]	0.035
4th quarter of the night (360–480 minutes)
Wake (%)	7.1 [4.9–10.5]	5.2 [3.3–16.0]	0.81
NREM 1 (%)	8.1 [4.9–10.0]	7.6 [6.9–10.1]	0.70
NREM 2 (%)	46.0 [38.0–55.9]	42.0 [37.8–53.2]	0.31
NREM 3 (%)	1.05 [0.3–8.5]	4.6 [0–10.8]	0.51
REM (%)	28.1 [22.1–34.0]	25.4 [19.6–31.2]	0.41

Notes: Values are shown as median [interquartile range]. % for wake, NREM 1, 2, 3 and REM calculated as % of total recording time. Wilcoxon signed-rank test was performed for the comparison of each variable between two visits.

Abbreviations: REM, rapid eye movement sleep; NREM, non-rapid eye movement sleep.

Figure 1 Distributions of sleep stages throughout the night. Stacked area plot of the proportions of sleep stages among all participants as a function of time of the night (from time at lights out at 23:00 until next morning at 7:00) calculated at 10-minute intervals.

Table 2 Mixed-Effect Spline Regression Models of EEG Spectral Power Distributions

Outcome – Proportion of Alpha Band (%)
				Model adjusted for sex
Fixed effects	β coefficient	95% CI	P value	Fixed effects	β coefficient	95% CI	P value
Late dinner	−1.42	−1.74, −1.10	<0.0001	Late dinner	−1.39	−1.71, −1.07	<0.0001
Time from lights out (min)	−0.24	−0.25, −0.23	<0.0001	Time from lights out (min)	−0.24	−0.25, −0.23	<0.0001
Late dinner × time from lights out	0.0058	0.0047, 0.007	<0.0001	Late dinner × time from lights out	0.0058	0.0046, 0.007	<0.0001
				Male sex	−1.20	−1.52, −0.87	<0.001
				Late dinner × male sex	0.0051	−0.0065, 0.0017	0.39
Outcome – Proportion of Beta Band (%)
				Model adjusted for sex
Fixed effects	β coefficient	95% CI	P value	Fixed effects	β coefficient	95% CI	P value
Late dinner	−1.32	−1.54, −1.11	<0.0001	Late dinner	−1.33	−1.54, −1.11	<0.0001
Time from lights out (min)	−0.088	−0.095, −0.082	<0.0001	Time from lights out (min)	−0.088	−0.094, −0.081	<0.0001
Late dinner × time from lights out	0.0023	0.0016, 0.0031	<0.0001	Late dinner × time from lights out	0.0024	0.0016, 0.0031	<0.0001
				Male sex	−0.099	−0.31, 0.11	0.36
				Late dinner × male sex	−0.0012	−0.002, −0.0047	0.001
Outcome – Proportion of Delta Band (%)
				Model adjusted for sex
Fixed effects	β coefficient	95% CI	P value	Fixed effects	β coefficient	95% CI	P value
Late dinner	2.54	2.07, 3.01	<0.0001	Late dinner	2.56	2.09, 3.03	<0.0001
Time from lights out (min)	0.28	0.27, 0.30	<0.0001	Time from lights out (min)	0.28	0.26, 0.29	<0.0001
Late dinner × time from lights out	−0.0084	−0.01, −0.0067	<0.0001	Late dinner × time from lights out	−0.0085	−0.01, −0.0068	<0.0001
				Male sex	−0.58	−1.06, −0.11	0.015
				Late dinner × male sex	0.0015	−0.0015, 0.0032	0.074
Outcome – Proportion of Theta Band (%)
				Model adjusted for sex
Fixed effects	β coefficient	95% CI	P value	Fixed effects	β coefficient	95% CI	P value
Late dinner	−0.68	−0.94, −0.41	<0.0001	Late dinner	−0.65	−0.92, −0.39	<0.0001
Time from lights out (min)	−0.07	−0.08, −0.06	<0.0001	Time from lights out (min)	−0.07	−0.08, −0.06	<0.001
Late dinner × time from lights out	0.0011	0.00015, 0.0021	0.023	Late dinner × time from lights out	0.0012	0.0014, 0.0021	0.014
				Male sex	−1.25	−1.51, −0.98	<0.001
				Late dinner × male sex	−0.0014	−0.0024, −0.0047	0.0032

Figure 2 Distributions of EEG spectral power bands throughout the night. Median percentages (solid red and blue lines) and interquartile ranges (shaded red and blue areas) for each spectral power band over the course of the night (from time at lights out at 23:00 until next morning at 7:00) are shown. In the beginning of the night, late dinner increased delta power and reduced alpha and beta power. These differences were attenuated in the latter part of the night.

Figure 3 Spline regression models of the distributions of EEG spectral power bands as a function of dinner timing and time of night. Fitted mixed-effect spline regression model of mean percentages of spectral in each power band vs time from lights out after routine (solid line) and late (dashed line) dinners. Knots (circle and triangles after routine and late dinners, respectively), were placed at every 60 minutes. In the beginning of the night, late dinner was associated with increased delta power (Panel A) and reduced theta (Panel B), (Panel C) alpha, and (Panel D) beta power. In the latter part of the night, these differences were reversed for alpha and delta powers and were attenuated for beta and theta powers.

Figure 4 Arousal frequencies throughout the night. Arousal frequencies (number of arousal periods per hour of sleep) were tallied for each quarter of the night (each 120 minutes) for both routine dinner (solid line) and late dinner (dashed line) conditions. Overall, the arousal frequencies decreased over the course of the night. There were no significant differences in arousal frequencies between the 2 dinner conditions in any quarter of the night (p > 0.05 using Wilcoxon signed-rank tests).