Very Short Sleep and Weekend Sleep Drift Linked to Teen Depression and Suicide Risk

When teenagers trade weeknight sleep for weekend lie-ins, the cost may be measured in more than grogginess. A new analysis of nearly 50,000 South Korean adolescents suggests that two sleep characteristics in particular—getting very little sleep overall and shifting one’s sleep window dramatically between weekdays and weekends—stand out as the most consistent correlates of depressive mood and suicidal outcomes. The study, published in BMC Psychiatry, draws on the 2025 Korea Youth Risk Behavior Survey, a nationally representative government survey, and applies a level of statistical scrutiny that goes well beyond the simple sleep-questionnaire comparisons common in this field.

The research team, led by Heerim Son and corresponding author Jae-Won Choi of Gyeongsang National University Hospital, set out to disentangle three intertwined descriptors of adolescent sleep: weekly average sleep duration, corrected mid-sleep on free days (known in circadian research as MSFsc), and the absolute difference between weekday and weekend sleep midpoints, a measure often described as a proxy for social jetlag. Because these three quantities are mathematically derived from the same underlying bedtime and wake-time reports, the authors argue that treating them as independent predictors—as many previous studies have done—can be misleading. A teenager who sleeps late on weekends necessarily has both a later free-day mid-sleep and a larger weekday-weekend difference, so any analysis that adjusts for one variable while interpreting the coefficient of another as a standalone effect risks obscuring the true structure of the data.

To address this, the researchers used survey-weighted logistic regression fitted with restricted cubic splines, a technique that allows the relationship between a continuous predictor and an outcome to bend and curve rather than being forced into a straight line. This matters because sleep-health effects are rarely linear: the difference in risk between five and six hours of sleep is unlikely to mirror the difference between eight and nine hours. The outcomes examined were four self-reported measures from the survey: depressive mood, suicidal ideation, making a suicide plan, and attempting suicide. Descriptive analyses included 48,426 adolescents with valid sleep-derived variables, and the fully adjusted models included 48,416 complete cases—a scale that lends considerable statistical power to the estimates.

The headline finding is that weekly average sleep duration and the absolute weekday-weekend sleep midpoint difference were associated with all four outcomes. For sleep duration, the association was nonlinear across the board, with odds of depression and suicidality rising steeply at shorter durations. In other words, the harm of cutting sleep appears concentrated at the extreme short end of the distribution, a pattern consistent with the idea that there is a threshold below which the adolescent brain—still undergoing substantial neurodevelopment—struggles to compensate. The midpoint difference, by contrast, was nonlinear only for depressive mood; for the suicidal outcomes, its association was generally monotonic, meaning that the larger the weekday-weekend drift, the higher the odds, without a clear plateau.

The third descriptor, MSFsc, told a more nuanced story. Corrected mid-sleep on free days is widely used in chronobiology as a marker of an individual’s intrinsic circadian phase, because it captures when a person’s body clock prefers to sleep once the pressure of weekday obligations is removed. In this study, MSFsc showed overall and nonlinear associations with depressive mood, suicidal ideation, and suicide plan, but the association with suicide attempt did not reach conventional significance in the primary analysis, with an overall p-value of 0.054 and a nonlinearity p-value of 0.993. Moreover, this result proved sensitive to the exclusion of extreme MSFsc values, suggesting that a small number of adolescents with unusually late or unusual free-day sleep patterns may have been driving part of the signal. The authors are appropriately cautious here: a later chronotype has often been linked to poorer adolescent mental health in prior literature, but this analysis indicates that the association may be less robust for the most severe outcome, suicide attempt, than for mood and ideation.

One of the study’s most instructive findings concerns what happens when sleep-timing variables are categorized and adjusted in sequence. When the researchers grouped adolescents into categorical sleep-timing classes and then progressively adjusted the models for sleep duration and the midpoint difference, the estimates changed substantially. This is a technical point with real practical consequences: it means that much of the apparent risk attributed to being a ‘late sleeper’ in simpler analyses may actually reflect the shorter total sleep or greater weekend drift that tend to travel with late timing. The three descriptors are not independent dials that can each be turned separately; they are overlapping summaries of the same behavior, and any one of them may be carrying the statistical weight of the others.

The team also ran exploratory analyses combining sleep-timing groups with sleep-duration categories, producing 72 possible profiles. The pattern was striking: sleeping less than five hours per night was associated with higher odds of the outcomes across every sleep-timing group, suggesting that extreme sleep restriction dominates regardless of when a teenager’s sleep is anchored. Yet only 4 of the 72 direct contrasts between profiles remained statistically significant after correction for the false discovery rate—a statistical safeguard against piling up spurious positives when many comparisons are made. The lesson is that while very short sleep looks like a robust red flag, finer-grained claims about specific combinations of timing and duration require more data before they can be treated as established.

Methodologically, the study has notable strengths. The sample is large and nationally representative, the survey weights allow estimates to reflect the Korean adolescent population, and the spline-based modeling respects the nonlinear character of sleep-health relationships. The use of secondary, de-identified government survey data was reviewed and deemed exempt by the Institutional Review Board of Gyeongsang National University Hospital, and the analysis was conducted in accordance with the Declaration of Helsinki. The work was supported by a New Faculty Research Support Grant from Gyeongsang National University in 2025, and the authors declare no competing interests.

At the same time, the authors are explicit about the limits of what can be concluded. The design is cross-sectional, capturing a single moment in time, so it cannot establish whether poor sleep causes depressive mood and suicidality, whether depression disrupts sleep, or whether both arise from shared causes such as academic stress, screen use, or underlying psychiatric vulnerability. The sleep measures are self-reported rather than objectively measured with actigraphy or polysomnography, and the survey captures only a snapshot of sleep habits rather than their trajectory over months or years. Perhaps most importantly, the authors emphasize that because the three sleep descriptors share timing information, the mutually adjusted estimates should be read as conditional associations, not as independent causal effects. This is a refreshingly honest framing in a literature where over-adjustment and causal overreach are common.

For clinicians, educators, and policymakers, the practical signal is nonetheless clear. Very short sleep duration—particularly under five hours—and large weekday-weekend shifts in sleep midpoint are the markers most consistently tied to depressive mood and suicidal outcomes in this population. In a country where intense academic schedules push many adolescents toward late bedtimes and early wake times on school days, followed by dramatic weekend catch-up sleep, these findings suggest that stabilizing sleep schedules and protecting minimum sleep duration could be meaningful targets for school-based mental health strategies. The authors call for longitudinal studies using more detailed circadian and sleep measures to determine whether intervening on sleep timing and duration can actually reduce risk. Until then, the study stands as a reminder that the question is no longer simply whether teenagers sleep enough, but how the shape of their sleep across the week—its length, its anchor, and its weekly wobble—relates to the state of their minds.

Subject of Research: Associations of sleep duration and sleep-timing patterns with depressive mood and suicidal outcomes in Korean adolescents

Article Title: Associations of sleep duration and sleep-timing patterns with depressive mood and suicidal outcomes in Korean adolescents

Article References: Associations of sleep duration and sleep-timing patterns with depressive mood and suicidal outcomes in Korean adolescents. (n.d.). https://doi.org/10.1186/s12888-026-08706-6

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08706-6

Keywords: adolescent health, sleep duration, sleep timing, circadian rhythm, depression, suicidal ideation, social jetlag, Korea Youth Risk Behavior Survey, MSFsc, epidemiology, mental health, Associations