Circadian clocks shift later (delay) with the progression of puberty; this shift contributes to late sleep onsets in older adolescents. Early school start times, however, force teenagers to awaken earlier than their spontaneous wake time and the opportunity for sleep shortens. Indeed, a majority of adolescents in the U.S. are chronically sleep deprived, getting 1-2 hours less sleep than recommended. Also, many adolescents wake for school at the "wrong" circadian time. Chronic circadian misalignment and sleep restriction are at their peak during late adolescence, and are associated with morning daytime sleepiness, poor academic performance, conduct problems, depressed mood, suicidal ideation, substance use, and obesity. Morning bright light exposure from light boxes can shift rhythms earlier (phase advance) to facilitate earlier sleep onset, and reduce circadian misalignment and the associated risks. The investigators constructed the first phase response curve (PRC) to light for adolescents and determined that the optimal time for bright light exposure to produce phase advances was about 1 hour before habitual wake and light should be avoided around the time of habitual bedtime because it causes rhythms to shift later (delay). Studies of adults, however, indicate that restricted sleep and exposure to evening light due to late bedtimes - two classic features of older adolescent sleep - make morning bright light less effective in producing advances. Pilot data collected from adolescents mimic this finding, but also suggest that staying awake late in normal household lighting and the subsequent sleep restriction before and during a 3-day morning bright light regimen, can shift the system in the wrong direction (phase delay). The overarching goal of this study is to examine the dose of sleep restriction and evening household light that prevents the desired phase advance to morning bright light in adolescents aged 14-17 years. Study 1 aims to construct a sleep restriction with normal household evening light dose-response curve to determine the point at which morning bright light begins to lose its effectiveness. Following a baseline week with 10-hour sleep opportunities, participants will keep the same wake time but be randomly assigned to one of 4 bedtimes which will be the same or later than baseline to produce 4 levels of sleep restriction with evening light (0, 1.5, 3, or 4.5 hours). After 2 nights, the sleep schedule will gradually shift earlier over 3 nights, and participants will receive bright light each morning. The investigators hypothesize that the circadian system will advance with sufficient sleep, but with increasing sleep restriction/evening light, circadian rhythms will not shift or will delay despite the phase advancing morning bright light. Study 2 will test whether reducing evening light exposure by wearing sunglasses before bedtime during sleep restriction can facilitate phase advances. Study 2 will test the same 4 "doses" of sleep restriction. The main outcome measures to build the dose-response curve will be phase shifts of the central circadian clock marked by the dim light melatonin onset (DLMO) and total sleep time measured from actigraphy in the laboratory. Secondary outcomes include cognitive performance, sleepiness, and mood.
