In a sample of high school students, (1) to characterize within-person changes in sleep and circadian characteristics from school nights to weekend nights, (2) to examine whether later circadian phase relates to weekday-weekend changes in sleep/circadian characteristics, and (3) to examine correlations between biological and proxy measures of circadian phase. Sample included 95 high school students reporting at least one drink of alcohol in their lifetime. Participants completed baseline self-report measures, wrist actigraphy for 8 days, and two overnight laboratory visits (Thursday and Sunday) for salivary melatonin sample collection. Circadian phase was calculated as the dim light melatonin onset (DLMO; 4 pg/mL threshold). Proxy circadian phase measures included the Composite Scale of Morningness (CSM), Munich Chronotype Questionnaire (MCTQ), and actigraphy-based midsleep. Other than nap duration, all examined actigraphy-based sleep characteristics, DLMO, and DLMO-sleep phase angles showed weekday-weekend differences (adjusted p-values <0.05). Later mean DLMO was associated with larger weekday-weekend changes in total sleep time (b=0.39, padjusted=0.010). CSM and actigraphy-based midsleep showed small-to-moderate (rho=~0.3) and moderate (rho=~0.5) correlations with DLMO, respectively, but chronotype based on the MCTQ was not correlated with DLMO. In the largest published sample to date, circadian phase substantially shifted from the school week to weekend, underscoring the "social jetlag" imposed by early school start times. Similarly, teens with the latest circadian phase exhibited the greatest weekend catch-up sleep. Finally, perhaps due to the instability of circadian phase in this context, self-reported proxies for circadian timing were poor approximations of biological circadian phase.
Read full abstract