Introduction

Sleep regulation is well known to be governed by two primary processes: homeostatic drive (process S) and circadian rhythm (process C)1. Building on this framework, sleep inertia (SI) is proposed as a third pivotal process in sleep regulation2, characterized by a temporary state of reduced cognitive function and lower arousal following awakening3. Sleep inertia (SI), influenced by factors such as sleep duration, sleep stage prior to awakening, and the time of day, typically diminishes within 30 min. However, substantial sleep deprivation can intensify its effects3. Previous research in which subjects were awakened hourly for addition tests on consecutive nights found that performance during the SI period progressively worsens with increased sleep deprivation4.

Circadian rhythm is another factor potentially affecting SI, but previous studies provided inconsistent evidence regarding its influence. Naitoh et al. (1993) explored the relationship between circadian timing and SI as assessed through a logical reasoning task, concluding that SI does not vary significantly based on the time of awakening5. In contrast, Dinges et al. (1985) found that cognitive decline due to SI is more severe after naps taken during the circadian trough (nighttime) than naps taken during the circadian peak (afternoon)6.

An additional circadian factor affecting SI is chronotype, which refers to individual differences in sleep and wake patterns. Ritchie et al. (2017) found that individuals showing greater eveningness take more time to overcome SI compared with those showing earlier chronotypes7. A misalignment between the time of day and an individual’s chronotype exacerbates the effects of sleepiness and sleep inertia, leading to greater impairments in driving performance, particularly among morning-type individuals8. Moreover, chronotype influences circadian rhythms and sleep parameters, with evening-type individuals showing delayed activity peaks and poorer sleep quality during weekdays compared with morning and intermediate types, but recovering on weekends due to fewer social and academic constraints9. This pattern, known as “social jetlag” (SJL), reflects a misalignment between social and biological clocks due to significant differences in their sleep schedules between workdays and free days: evening types often compensate for sleep debt accrued during workdays by oversleeping on free days10. The relationship between chronotype and SJL is well-established, with SJL being more pronounced in individuals with greater eveningness11. Therefore, it is important to differentiate factors such as sleep deficiency and chronotype when considering SJL’s impact on SI.

Roenneberg et al. mentioned that evening types require longer durations to feel fully awake, with reduced sleep leading to longer SI on workdays. However, SI lasts approximately one hour, regardless of chronotype or sleep duration on free days12. Nevertheless, extreme evening types experience the greatest SJL and its negative impacts on cognitive functioning, such as SI13. Furthermore, research shows that SJL is associated with lower academic performance and reduced cognitive abilities, with girls being disproportionately affected14. Among female adolescents, SJL has been positively linked to depressive symptoms, a pattern not observed in males15. Longitudinal studies have also highlighted that persistent or increasing SJL during adolescence is associated with abnormal BMI in young adulthood, with notable sex-specific differences16. In a study conducted in Northern Russia, 65% of young adults reported experiencing SJL, which was associated with disruptions in circadian rhythms, such as reduced wrist temperature amplitude and elevated cortisol awakening responses. Importantly, these negative effects were more pronounced in females, affecting their sleep quality, duration, and overall circadian rhythm stability17. These findings emphasize the importance of addressing SJL, particularly in adolescents and young adults, to mitigate its adverse health and functional impacts. Hilditch et al. found that females reported greater increases in sleepiness than males after nighttime awakenings, despite similar cognitive performance18. Based on this evidence, we hypothesized that chronotype influences SI through the mediating role of SJL. Specifically, evening types may be more likely to experience greater SJL due to the misalignment between their biological rhythms and socially imposed schedules, leading to heightened SI when waking at suboptimal times (e.g., early mornings on workdays). Recognizing the potential influence of sex on SJL and SI, this study also accounts for sex as a covariate to better isolate the effects of chronotype and SJL on SI. However, there is lack of research exploring this connection to date. Thus, this study aims to explore the relationship between chronotype and SI, investigating the potential mediating role of SJL in this relationship.

Methods

A total of 319 participants were initially recruited; however, 88 participants were excluded for not meeting the inclusion criteria. The final sample consisted of 231 participants (174 women, 56 men, 1 no answer; mean age = 20.44 ± 1.13 years). The study utilized convenience sampling to recruit participants through advertisements on social media platforms and campus noticeboards. Interested individuals were directed to a study website for further details and eligibility screening. University students aged 18 to 22 years who met the following inclusion criteria were invited to participate: (1) actively enrolled in courses; (2) no history of major physical or mental illnesses or ongoing medical treatment; (3) no use of illegal substances or medications that could affect sleep and no diagnosis of sleep-related disorders; (4) a daily caffeine intake of no more than 300 mg; and (5) no habitual smoking or alcohol consumption. Data were collected using a self-administered questionnaire via Google Forms, distributed through online community platforms such as Facebook, from April 1, 2022, to February 28, 2023. To ensure eligibility, participants reviewed and confirmed the five inclusion criteria during the informed consent process by selecting “yes” or “no” for each statement. Those who did not meet any criteria were prompted to exit the survey. After submission, the research team manually verified responses to ensure compliance with the inclusion criteria and the accuracy of demographic information. The required sample size was calculated according to the method outlined in Table 3 of Fritz and MacKinnon (2007) 19. Given a moderate effect size and the possibility of partial mediation, we targeted a minimum of 224 valid responses to ensure sufficient power. By the project deadline, 231 valid responses were collected, meeting the recommended threshold. Ethical approval for the study was obtained from the Institutional Review Board (IRB) of the research ethics committee of National Taiwan University, Taipei, Taiwan (NTU-REC No.: 202204ES008).

Table 1 Sleep characteristics of participants on weekdays and weekends.
Full size table

This study included the following questionnaires:

  1. (i)

    Demographic Information.

    The questionnaire included items to collect participants’ basic demographic information. These items covered age, gender, current academic year (e.g., freshman, sophomore), living situation (alone or with others), history of major physical illnesses, and experience consulting professionals for psychological or mental health concerns. This information was also used to verify whether participants met the inclusion criteria for the study.

  2. (ii)

    Chinese version of Morningness-Eveningness Questionnaire (MEQ).

    This 19-item questionnaire aims to measure one’s chronotype. The scale includes both Likert-type and time-scale questions. The Likert-type items offer five response options, with lower values indicating stronger eveningness, and are scored from 1 to 5. For example, one item asks, “If you need to wake up at a specific time in the morning, how dependent are you on an alarm clock?” Responses range from “Very dependent” to “Not at all dependent.” The time-scale questions cover a 7-hour range, divided into 15-minute intervals. According to the total score, people are divided into 5 types, including: definitely morning type, moderately morning type, intermediate type, moderately evening type, definitely evening type. Higher scores are associated with earlier wake times and a stronger inclination towards morningness. The internal consistency of this Chinese online version of the questionnaire is 0.75 and the test-retest reliability is 0.5220,21.

  3. (iii)

    Munich ChronoType Questionnaire (MCTQ).

    This 15-item questionnaire includes two sets of questions on living habits for workdays and free days: one item for acquiring participants’ weekly schedules and another with 7 questions each for workdays and free days. In this study, we used participants’ MCTQ self-reports to calculate weekly average sleep duration, weekly sleep deprivation, mid-sleep point on school days (MSW) and free days (MSF), and social jetlag (SJL) for further analysis12. SJL was calculated following Wittmann et al. (2006), using the difference between the MSW (representing the social clock influenced by external demands, such as work or school schedules) and the MSF (representing the biological clock, as it reflects the individual’s natural sleep pattern without external pressures)10. The version of the MCTQ used in this study was a traditional Chinese version provided directly by the original author. This version had been previously translated and made available with the original author’s authorization, ensuring linguistic and conceptual fidelity to the original instrument.

  4. (iv)

    Chinese version of Sleep Inertia Questionnaire (SIQ).

    This 22-item questionnaire includes four subscales that reflect aspects of SI: (1) Physiological, (2) Responses to Sleep Inertia, (3) Cognitive and (4) Emotional. The participants answered on five-point Likert scale ranging from 1 (not at all) to 5 (always). Greater total scores indicate that individuals experience more severe SI. The original questionnaire has high construct validity and good convergent validity with a reliability of 0.95 for the entire scale22. The translation process involved multiple steps to ensure the quality and fidelity of the Chinese version of the questionnaire. With authorization from the original author, the questionnaire was first translated into Chinese and reviewed by two domain experts for linguistic and cultural appropriateness. A back-translation was then performed to identify and resolve any inconsistencies through iterative refinement. Additionally, reliability and validity testing were conducted to confirm the suitability of the Chinese version for the current research. The Cronbach’s alpha for the Chinese version was 0.90. The test-retest reliability was calculated as 0.793, based on data from a separate study (unpublished data).

Data analysis

All statistical analyses were performed using IBM SPSS Statistics for Windows software and mediation model analyses were performed with PROCESS procedure for SPSS Version 4.023. Beyond the Pearson correlation analysis, we also conducted the simple mediation analysis between Chronotype, SJL and overall SIQ (Model A); and we further decomposed the SIQ into SIQ subscale factors in the mediation analysis (Model B) to test the hypothesis that different sleep inertia factors are mediated by SJL. To account for potential confounding effects, sex was included as a control variable in all mediation models.

Results

Among the 231 participants, 5.2% (n = 12) were definitely evening type, 41.6% (n = 96) were moderately evening type, 48.1% (n = 111) were intermediate type, and 5.2% (n = 12) were moderately morning type. The average SJL was 1.54 ± 1.64 h for the participants. The SIQ scores averaged 60.74 ± 14.66. Table 1 exhibits the overall sleep characteristics on weekdays and weekends. 

Table 2 provides a comprehensive overview of correlations among key variables, shedding light on the relationships between sleep inertia, chronotype, and SJL. Starting with the SIQ scores, we observed a significant negative correlation between SIQ scores and MEQ scores (r = − .243, p < .01). This suggests that individuals showing more eveningness reported higher levels of sleep inertia, indicating greater difficulty in transitioning from sleep to wakefulness. When examining the SIQ factors, the Physiological and Cognitive factors exhibited a significant negative correlation with chronotype (r = − .131, p < .05; r = − .170, p < .01). Moreover, the “Responses to Sleep Inertia” factor also displayed a significant negative correlation with chronotype (r = − .496, p < .01), implying that individuals showing more eveningness experience more behavioral and cognitive challenges transitioning from sleep to wakefulness, such as a stronger reliance on external cues like alarms and frequent use of the snooze button. However, no significant association was found with the Emotional factor, as indicated by the low correlation coefficients.

Table2 Descriptive statistics and correlations for sleep variables.
Full size table

Regarding the association between SJL and SIQ factors, we found no significant correlation was evident among the Physiological, Cognitive, and Emotional factors with SJL (r = .032, p > .05; r = .004, p > .05; r = .018, p > .05, respectively). However, a positive significant association was found between the factor “Responses to Sleep Inertia” and SJL (r = .231, p < .01), suggesting that individuals with misaligned social and biological clocks report more difficulty transitioning from sleep to wakefulness.

Figure 1 (Model A) shows the result of simple mediation analysis conducted using the Hayes’ PROCESS macro in SPSS (Model A). The total effect of Chronotype on SIQ was significant (β = -0.38, p < .001), indicating a negative relationship between Chronotype and total SIQ total score. When controlling for both SJL and sex, the direct effect of Chronotype on SIQ total score remained significant (β = -0.37, p < .001). This suggests that Chronotype has a direct negative impact on SIQ, independent of SJL and sex. The indirect effect of Chronotype on SIQ through SJL was not statistically significant (β = -0.01, SE = 0.04, Boot LLCI = -0.0443, Boot ULCI = 0.1094). This implies that the mediation effect of SJL in the relationship between Chronotype and SIQ is not supported. The effect size of Chronotype on SIQ is evaluated as follows: The total effect demonstrates a significant and substantial impact (Cohen’s f2 = 0.0718), even after accounting for the effects of SJL and sex. In contrast, the direct effect reveals a minimal effect size (Cohen’s f2 = 0.00075). These findings emphasize the robustness of the relationship between Chronotype and SIQ as a total effect, suggesting that the influence of Chronotype on SIQ persists after controlling SJL and sex. However, the scarce effect size for the direct effect highlights the complexity of this relationship, potentially mediated or moderated by other factors.

Table 3 Mediation analysis: between the chronotypes and responses to sleep inertia scale (Model b).
Full size table
Fig. 1
figure 1

The mediation model of chronotype, social jetlag and response to sleep inertia. *p < .05. **p < .01.

Full size image

Figure 1 (Model B) illustrates the mediation analysis results for the only significant SIQ factor, “Responses to Sleep Inertia”. The total effect of Chronotype on Responses to Sleep Inertia was significant (β = -0.22, p < .001), indicating a negative association between Chronotype and the Responses to Sleep Inertia score (see Table 3). The direct effect, controlling for SJL and sex, also remained significant (β = -0.20, p < .001). This suggests that Chronotype has a direct negative impact on the Responses to Sleep Inertia, independent of its effect on SJL and sex. The indirect effect of Chronotype on Responses to Sleep Inertia through SJL was statistically significant (β = -0.01, Boot SE = 0.01, Boot LLCI = -0.0407, Boot ULCI = -0.0044), indicating that SJL partially mediates the relationship between Chronotype and Responses to Sleep Inertia. The effect size of Chronotype on the Responses to Sleep Inertia score is substantial (Cohen’s f2 = 0.414) for the total effect, indicating a strong and meaningful relationship. Even after controlling for SJL and sex, the direct effect remains significant, albeit with a smaller effect size (Cohen’s f2 = 0.024). These findings underscore the robustness of the relationship between Chronotype and the Responses to Sleep Inertia, suggesting that this relationship is both direct and partially mediated through SJL.

The findings suggest a significant total effect of Chronotype on Responses to Sleep Inertia, with a negative association. Additionally, while no mediation effect was observed for other SI factors, the small partial mediation effect on this relationship supports the hypothesis that SJL plays a role in mediating relationships between eveningness and behavioral difficulty transitioning from sleep to wakefulness.

Discussion

This study investigated the connection between chronotype and sleep inertia, with a particular focus on the potential mediating role of SJL. Our findings show that eveningness heightens sleep inertia, consistent with Ritchie et al.7. Additionally, consistent with Roenneberg et al. (2003), eveningness exhibited stronger SJL11. However, our results showed no robust relationship between SJL and SI, with only a minimal correlation observed (0.081) observed for all SI factors except for behavioral responses to SI, which showed a small mediation effect size. In brief, SJL does not appear to play a strong mediating role in the relationship between chronotype and SI.

Nevertheless, it is noteworthy that evening types reported greater responses to SI in our study. This SIQ factor is primarily related to behavioral aspects of sleep inertia and sleepiness, containing the following items: “Need an alarm to wake up,” “Wish you could sleep more,” “Notice that you feel sleepy,” “Hit the snooze button on the alarm,” and “Have problems getting out of bed”22. Previous studies reported the relationship between chronotype and sleepiness24, especially in adolescents25. Roeser et al. observed that while eveningness negatively impacts school performances through SJL, its effect on academic achievement is indirect, being influenced by daytime sleepiness and learning motivation25. Based on our findings that SJL mediates the relationship between chronotype and responses to sleep inertia, eveningness partially predicts an individual’s responses to sleep inertia that result from SJL. Individuals with late chronotype tend to experience increased sleep inertia and higher levels of daytime sleepiness, which may stem from insufficient sleep during workdays and the accumulation of sleep debt that is compensated for on free days.

Previous research has highlighted notable gender differences in the experience and consequences of SJL and SI. For example, girls appear more susceptible to the negative effects of SJL on cognitive abilities and GPA14. These differences may stem from both biological and social factors. Biologically, girls tend to enter puberty earlier than boys, and this developmental stage is associated with a shift toward later sleep timing26,27. Socially, girls often face expectations to wake earlier on weekdays for activities like grooming or household responsibilities28,29. In contrast, boys may maintain earlier schedules on weekends due to participation in sports or other structured activities, which could mitigate their exposure to SJL30. Despite these findings, when we examined sex as a potential moderator or mediator in the relationships between chronotype, SJL, and responses to SI in our study, the effects were not statistically significant.

The non-significant mediation findings in our study suggest that while SJL is conceptually linked to both chronotype and SI, its role as a mediator requires further investigation. One possible explanation is the relatively homogenous sample of young adults, whose sleep patterns and schedules may differ from those in broader populations, reducing variability in SJL. Additionally, the range of SJL values in this study was relatively narrow, predominantly falling between 0 and 3, which may have further limited its variability and made it more challenging to detect significant differences. Unmeasured factors, such as the sleep stage prior to awakening3, the amount and depth of slow-wave sleep31, or behavioral/environmental factors, may play more significant roles in the relationship between chronotype and SI than SJL alone.

This study has three limitations: First, the absence of a significant robust mediation effect in our findings may be influenced by the specific characteristics of our sample, which consisted primarily of students in university. This sample homogeneity limited the generalizability of the findings to other groups. Consequently, a future cross-sectional design is warranted to probe more nuanced relationships among chronotype, SJL, and SI. Second, this relationship is subject to various other confounding factors. For example, sleep deprivation may potentially influence both SI and SJL. However, in our study, participants appeared to get sufficient sleep (over 7 h) on both work and free days (see Table 1). Similarly, bedtime and wakeup time may influence the relationship between chronotype and both SI and SJL. Third, the method of SJL calculation could affect the findings. For example, another index, social jetlag sleep-corrected (SJLsc), proposed by Jankowski (2017)32, is calculated by the absolute difference between sleep onset on free days and the sleep onset on workdays12. Updating both models with SJLsc, we found that SJLsc also did not mediate the relationship between chronotype and SIQ (SJLsc → SIQ in Model A: β = -0.17, p = .79; in Model B: β = 0.07, p = .68). However, the adequacy of adopting SJLsc (simple calculation by sleep-onset disparity) is questionable in the current study, because the index may not comprehensively represent the cumulative consequences of circadian misalignments and sleep factors in social jetlag, which aligned well with Wittmann’s viewpoint10. Our calculation of SJL adheres to the widely validated method proposed by Wittmann et al. (2006)10, which focuses on quantifying the temporal discrepancy between the social clock—shaped by social obligations and demands—and the biological clock. While this approach captures the core misalignment central to SJL, we acknowledge that other functional aspects of social obligations, such as the nature and variability of these demands, may provide additional insights into the broader context of SJL. Future studies could explore these dimensions to enhance the understanding of SJL and its relationship with sleep inertia and chronotype. Ultimately, the current study sheds light on the relationships between chronotype, SJL, and SI; additionally, further studies are warranted to investigate the neurophysiological mechanisms of the sleep inertia process.