Caffeine Consuming Children and Adolescents Show Altered Sleep Behavior and Deep Sleep

Abstract

Caffeine is the most commonly ingested psychoactive drug worldwide with increasing consumption rates among young individuals. While caffeine leads to decreased sleep quality in adults, studies investigating how caffeine consumption affects children's and adolescents' sleep remain scarce. We explored the effects of regular caffeine consumption on sleep behavior and the sleep electroencephalogram (EEG) in children and adolescents (10-16 years). While later habitual bedtimes (Caffeine 23:14 ± 11.4, Controls 22:17 ± 15.4) and less time in bed were found in caffeine consumers compared to the control group (Caffeine 08:10 ± 13.3, Controls 09:03 ± 16.1), morning tiredness was unaffected. Furthermore, caffeine consumers exhibited reduced sleep EEG slow-wave activity (SWA, 1-4.5 Hz) at the beginning of the night compared to controls (20% ± 9% average reduction across all electrodes and subjects). Comparable reductions were found for alpha activity (8.25-9.75 Hz). These effects, however, disappeared in the morning hours. Our findings suggest that caffeine consumption in adolescents may lead to later bedtimes and reduced SWA, a well-established marker of sleep depth. Because deep sleep is involved in recovery processes during sleep, further research is needed to understand whether a caffeine-induced loss of sleep depth interacts with neuronal network refinement processes that occur during the sensitive period of adolescent development.

Keywords: adolescents; caffeine; children; development; sleep EEG topography; slow-wave activity.

Figure 1

Power spectrum of the sleep electroencephalogram (EEG) in caffeine consumers (red, n = 16) and controls (blue, n = 16). Group-wise comparisons were performed with the first two hours of NREM sleep EEG data using the average across all 109 electrode channels for each subject within each frequency bin within 0.25 and 25 Hz. Commonly used EEG frequencies are separated by vertical lines: SWA = 1–4.5 Hz; theta = 4.75–7.75Hz; alpha = 8–9.75 Hz; beta = 20–25 Hz. Frequency bins (0.25-Hz) within which power differed significantly (p < 0.05; two-tailed, unpaired Student’s t-test) and the level of their significance are shown in the lower panel, only significant p values (after FDR correction) are plotted.

Figure 2

Slow-wave activity (SWA) across the night in caffeine consumers (red, n = 16) and controls (blue, n = 16). SWA (1–4.5 Hz) was averaged across all electrodes (M ± SEM) for the first and last two hours (latest possible artifact-free, common two hours among all subjects) of NREM sleep stages N2–3. The asterisk indicates a significant difference between caffeine consumers and controls in the first two hours of NREM sleep (p < 0.05, two-tailed, unpaired Student’s t-test).

Figure 3

Regional sleep slow-wave activity (SWA, 1–4.5 Hz) and alpha activity (8.25–9.75) in caffeine consumers (n = 16) and controls (n = 16). Upper panels show the topographical distribution of SWA averaged for the two groups in NREM stages 2 and 3 during the first two hours of NREM sleep, group differences and the level of significance (t < −2.21, two-tailed, unpaired t-test). Lower panels show the topographical distribution of alpha activity averaged for the two groups in NREM stages 2 and 3 during the first two hours of NREM sleep, group differences and the level of significance (t < −2.25, two-tailed, unpaired t-test). The significance level is indicated with a red arrow. Values are color coded (maxima in red, minima in blue, respectively maxima in white, minima in black for t-values) and plotted on the planar projection of a hemispheric scalp model. Maps for caffeine consumers and controls are presented using the same color scale, and statistical maps are proportionally scaled to optimize color contrast. Values between electrodes were interpolated in each plot.