This school-based cohort study was conducted by the Ningbo Center for Disease Control and Prevention (CDC) in Ningbo, China. The baseline survey was administered beginning in October 2016 and included seventh-grade students from 13 middle schools across 10 districts via cluster random sampling. Follow-up assessments were conducted in October 2017 and October 2018. The study enrolled healthy adolescents and excluded those with major illnesses or incomplete data. The study process is illustrated in Figure 1.

A standardized, self-report questionnaire adapted from the U.S. Youth Risk Behavior Survey (YRBS) (29–31) was used to collect data on demographics, sleep duration and quality, physical activity, screen time, smoking, alcohol use, dieting behavior, and mental health. Trained investigators distributed and collected the questionnaires onsite, ensuring completeness.

Descriptive statistics were computed for all the variables. Continuous variables were assessed for normality via the Kolmogorov–Smirnov test. Normally distributed variables are reported as the means ± standard deviations (SDs), whereas nonnormally distributed variables are expressed as medians and interquartile ranges (IQRs). Categorical data are presented as frequencies and percentages.

Between-group comparisons were performed via independent t tests or analysis of variance (ANOVA) for continuous variables and chi-square tests for categorical variables. Pearson correlation analysis or Spearman correlation analysis was used to explore the associations between sleep disorders and cyberbullying victimization, lifestyle factors, and psychological factors, and heatmaps were drawn.

Multivariate logistic regression models were applied to examine the association between sleep disorders and cyberbullying victimization, with the results reported as adjusted odds ratios (ORs) and 95% confidence intervals (CIs). To further explore the dose–response relationship between sleep disorders and cyberbullying victimization, we employed a restricted cubic spline (RCS) to further analyze their nonlinear relationship and plotted the RCS curve.

To assess the robustness of our results, we performed several additional sensitivity analyses. We classified sleep disorders into level 0 (sleep disorder score=0), level 1 (score=1–9), level 2 (score=10–18), and level 3 (score=19–27) groups on the basis of the sleep disorder score according to the PSQI instruction manual (32, 41) and calculated unadjusted and adjusted ORs and 95% confidence intervals (95% CIs) via univariate and multivariate logistic regression analyses.

Mediation analysis was conducted to explore the indirect effects of loneliness, sadness, and nervousness on the relationship between sleep disorders and cyberbullying victimization. All analyses were conducted via R software (version 4.5.1, R Foundation for Statistical Computing, Vienna, Austria). A two-tailed P value of <0.05 was considered statistically significant.

Among the 2164 students initially enrolled, 41 were excluded because of missing data, and 213 were lost to follow-up, resulting in a final sample of 1910 adolescents (Table 1). The mean age of the participants was 12.2 ± 0.47 years, and the mean sleep disorder score was 3.32 ± 3.68 (range: 0–27.0). Significant differences were observed among the sleep disorder groups in terms of sex, parental marriage status, TV screen time, internet screen time, alcohol use, fruit intake, and meat intake (all P < 0.05). Furthermore, adolescents with higher sleep disorder levels reported higher prevalences of loneliness, sadness, and nervousness (all P < 0.05).

Over the two-year follow-up period, 196 adolescents (10.3%) engaged in cyberbullying victimization. The incidence of cyberbullying victimization was significantly associated with sex, parental marital status, muscle strengthening activity, alcohol use, meat intake, and adverse mental health outcomes, including loneliness, sadness, and nervousness (all P < 0.05, Table 2). Compared with girls (6.5%), boys were more likely to be involved in cyberbullying victimization (13.8%). Additionally, adolescents from households with divorced, widowed, or separated parents demonstrated an increased risk of cyberbullying victimization (19.0% vs. 9.5%). Compared with those not involved in cyberbullying victimization, adolescents involved in cyberbullying victimization presented significantly greater levels of loneliness, sadness, and nervousness (all P < 0.05).

Figure 2 shows the correlations of cyberbullying victimization with sleep disorders, lifestyle factors and psychological factors in school-aged adolescents. The heatmap correlation matrices revealed that cyberbullying victimization was significantly associated with sleep disorders, sex, parental marital status, TV screen time, internet screen time, alcohol use, fruit intake, meat intake, loneliness, sadness, and nervousness (all P < 0.05, Table 2, Figure 2).

Multivariate logistic regression demonstrated that adolescents with sleep disorders at baseline had a significantly greater risk of cyberbullying victimization during follow-up (OR: 1.11, CI: 1.07–1.14; Table 3). After adjusting for age, sex, parental marital status, education level, physical activity, screen time, cigarette and alcohol use, and dietary habits, sleep disorders remained an independent predictor of cyberbullying victimization (OR: 1.10, CI: 1.06–1.14). Sensitivity analyses further validated the robustness of the results, with the risk ratio of cyberbullying victimization increasing approximately with increasing incidence of sleep disorders (Table 3). The RCS curve revealed that the risk ratio of cyberbullying victimization increased approximately linearly with increasing incidence of sleep disorders (P for overall<0.001, P for nonlinear=0.915; Figure 3).

Gender-stratified analyses revealed that sleep disorders increased the risk of cyberbullying victimization among both boys (OR: 1.09, 95% CI: 1.05–1.14) and girls (OR: 1.15, 95% CI: 1.08–1.21). After adjusting for confounders, the association remained significant for both genders, although girls with greater sleep disorders presented a slightly greater risk of cyberbullying victimization (adjusted OR: 1.14, 95% CI: 1.07–1.22) than boys did (adjusted OR: 1.08, 95% CI: 1.03–1.14).

Mediation analysis was conducted to explore the roles of loneliness, sadness, and nervousness in the relationship between sleep disorders and cyberbullying victimization (Table 4, Figure 4). Sleep disorders were positively correlated with cyberbullying victimization (β = 0.008, P = 0.001) and significantly predicted higher levels of loneliness, sadness, and nervousness (all P < 0.05). Further analysis revealed that loneliness and nervousness were positively associated with cyberbullying victimization (β = 0.043 and 0.042, both P < 0.05).

Loneliness partially mediated the relationship between sleep disorders and cyberbullying victimization, accounting for 25.00% of the total effect (indirect effect β = 0.003, P < 0.001; Table 4, Figure 4). Similarly, nervousness exhibited a partially mediated effect, contributing to 8.33% of the association (indirect effect β = 0.001, P = 0.038; Table 4, Figure 4). In contrast, sadness exhibited no significant mediating effects (P = 0.369), suggesting that the link between sleep disorders and cyberbullying victimization is driven primarily by increased loneliness and nervousness.

This two-year longitudinal study advances our understanding of the complex relationship between sleep disorders and cyberbullying victimization in school-aged adolescents. After adjusting for lifestyle and sociodemographic confounders, sleep disorders independently predicted cyberbullying victimization (adjusted OR: 1.10, 95% CI: 1.06–1.14), with a stronger association observed in girls (adjusted OR: 1.14 vs. 1.08 for boys). Notably, loneliness and nervousness mediated 25.00% and 8.33% of this association, respectively, whereas sadness had no significant mediating effect. These findings highlight the critical interplay between sleep disturbances, emotional states, and cyberbullying behaviors, emphasizing the need for integrated interventions targeting sleep health and psychological well-being to mitigate the risk of cyberbullying victimization.

Our results align with theoretical frameworks suggesting that sleep disorders impair emotional regulation and executive functioning, thereby increasing vulnerability to aggression and peer conflict (42–44). The observed dose–response relationship (P for overall < 0.001, P for nonlinear = 0.915) supports prior evidence that even subclinical sleep disturbances increase the risk of cyberbullying (11, 45, 46). For example, experimental studies have demonstrated that sleep deprivation heightens amygdala reactivity to negative stimuli while reducing prefrontal cortex activity, fostering impulsivity and hostile social perceptions (47–49). Another cross-sectional survey by Chervin et al. (2003) demonstrated that adolescents with sleep-disordered breathing (SDB) exhibit 2–3 times more bullying and other specific aggressive behaviors (50–52), which is consistent with our observed OR of 1.10. The lower risk magnitude in our cohort may stem from methodological differences: while prior studies focused on sleep disturbances (assessed via the Adolescent Sleep-Wake Scale) (53), we measured broader sleep deficits (e.g., insomnia, sleep apnea and narcolepsy) via the PSQI, a validated tool for adolescent populations (Cronbach’s α = 0.717 in our sample) (54). These neurobiological changes may explain why adolescents with sleep disorders are more prone to misinterpret social cues, escalating conflicts into cyberbullying perpetration or victimization (9, 10).

The gender-specific risk pattern—higher vulnerability in girls—contrasts with cross-sectional studies reporting stronger sleep-aggression associations in boys (55, 56) but aligns with longitudinal data suggesting that girls’ emotional reactivity to sleep loss is amplified during adolescence (57). This disparity may reflect sex differences in stress hormone responses (e.g., cortisol dysregulation) (58) or sociocultural factors, such as girls’ greater reliance on peer relationships for emotional support (59, 60). Further research should disentangle biological (e.g., pubertal hormones) and environmental (e.g., gender norms) contributors to these differences.

The relationship between sleep disorders and cyberbullying may be moderated by lifestyle factors, with different lifestyle behavior variables presenting varying risk levels among adolescents. Interestingly, alcohol use is a lifestyle factor associated with both increased sleep disturbance and increased cyberbullying risk: adolescents who consume alcohol are more likely to experience sleep disorders and cyberbullying victimization, reinforcing links between alcohol use, sleep disruption, and aggression (61). Surprisingly, reduced meat intake has been associated with increased sleep problems in recent dietary studies, yet in some contexts, it is correlated with lower engagement in aggressive behaviors, including cyberbullying—findings that suggest that dietary components may have complex, bidirectional associations with both sleep and peer conflict risk (62, 63). Moreover, while greater internet use time is consistently related to worsened sleep patterns (e.g., delayed sleep onset, lower sleep quality) (64), it does not always correlate with greater cyberbullying victimization in these same cohorts. A possible explanation is that these outcomes pertain to victimization rather than perpetration: high screen or internet time may increase exposure to online environments, but not all users become targets; moreover, lifestyle and dietary moderators may buffer or amplify risk differently for victims vs. perpetrators.

The partial mediation by loneliness (25.00%) and nervousness (8.33%) highlights the importance of social-emotional pathways in the link between sleep disorders and cyberbullying. Sleep disturbances may intensify feelings of isolation by limiting opportunities for positive peer interactions, such as through daytime fatigue that hinders social engagement (44, 65). In contrast, nervousness may heighten negative appraisals of social situations, thereby increasing hostility (66, 67). These mechanisms align with the social information processing model, wherein sleep loss biases cognitive processing toward threat detection and retaliatory behaviors (68, 69).

The absence of mediation by sadness indicates that depressive symptoms may not directly influence cyberbullying victimization, contrasting with findings from cross-sectional studies (70, 71). This discrepancy could stem from differences in measurement (single-item vs. validated scales) or from the longitudinal design capturing the temporal precedence of mediators. Future studies should employ multidimensional mental health assessments to clarify these pathways. This distinction has practical implications: antibullying programs should prioritize social integration over worry reduction alone.

Our results advocate for multitiered interventions:

The strengths of this study include its longitudinal design, adjustment for key confounders (e.g., substance use, screen time), and robust mediation analysis. However, several limitations must be acknowledged (1): Self-report bias and single-item assessment: Subjective measures and single-item assessments of sleep and cyberbullying may underreport sensitive behaviors. Objective tools (e.g., actigraphy, peer nominations) and repeat-item assessment can strengthen future work (74, 75) (2). Cultural specificity: The sample was drawn from a single Chinese region; replication in diverse populations is needed to assess generalizability (3). Temporal granularity: The 2-year follow-up may miss short-term fluctuations in mediators. Ecological momentary assessment can capture dynamic sleep–emotion–behavior interactions (40, 76).

Sleep disorders represent a modifiable risk factor for cyberbullying victimization in adolescents, with loneliness and nervousness serving as critical mediators. By integrating sleep promotion with psychosocial interventions targeting emotional well-being, policymakers and educators can disrupt the pathway from poor sleep to aggressive behaviors. Future research should test the efficacy of such strategies in randomized trials and explore the bidirectional relationships between sleep, emotions, and cyberbullying across developmental stages.