Exercise, Sleep and Mood: How Physical Activity May Shield Teens from Problematic Internet Use

Pathological internet use among adolescents has become one of the defining public health puzzles of the digital age, and a new study published in BMC Psychology offers one of the most detailed maps yet of how it may connect to something as simple as moving the body. Drawing on data from 1,245 Chinese adolescents aged 12 to 18, researchers led by Xiaoming Zhou and colleagues constructed a statistical model in which physical activity, sleep problems, and negative emotions form a chain of interlinked pathways leading toward or away from compulsive internet behavior. The results, while cross-sectional and therefore not proof of causation, suggest that the protective association of exercise with problematic internet use may operate substantially through better sleep and fewer depressive, anxious, and stressed emotional states.

The theoretical backbone of the study is Compensatory Internet Use Theory, the idea that people retreat into the online world to escape or regulate unmet psychological needs. Under this framework, an adolescent who sleeps poorly and feels persistently negative emotions may turn to the internet for relief, and physical activity, by improving sleep and mood, could cut off that compensatory drive at its source. Rather than treating exercise as a simple antidote, the researchers modeled it as the first link in a chain: more activity predicts fewer sleep problems, fewer sleep problems predict fewer negative emotions, and fewer negative emotions predict lower levels of pathological internet use.

The data came from a stratified cluster sample of students at four public secondary schools in a single major metropolitan area in China. Each participant completed a battery of standardized instruments with published Chinese-language validation studies: the International Physical Activity Questionnaire Short Form to quantify weekly activity in metabolic equivalent of task units, the Pittsburgh Sleep Quality Index to capture sleep disturbance, the Depression Anxiety Stress Scales-21 to measure negative emotional states, and the Compulsive Internet Use Scale to assess pathological internet use. The authors are candid that some of these validation populations differ from the study sample; the Chinese IPAQ-SF and DASS-21 were validated primarily in college students, and the PSQI in a mixed adult and clinical sample, while the CIUS was validated specifically in Chinese and Taiwanese high-school adolescents.

Before testing any structural relationships, the team established that their measurement instruments worked equivalently across boys and girls, a step that many published studies skip. Using confirmatory factor analysis, they demonstrated measurement invariance across gender groups, with changes in the comparative fit index of minus 0.002 and in the root mean square error of approximation of minus 0.001, both well within conventional thresholds. Only after this foundation was secured did they estimate the structural model, using structural equation modeling with a bootstrapping procedure based on 5,000 resamples to compute confidence intervals for the indirect pathways. Bootstrapping is widely regarded as the preferred method for testing mediation because indirect effects are rarely normally distributed, and large resample counts reduce the instability of the estimates.

The headline finding is straightforward: physical activity was negatively associated with pathological internet use. More revealing, however, is the architecture behind that association. Sleep problems and negative emotions each carried a statistically significant indirect association between activity and internet misuse, and, crucially, the two mediators also formed a significant sequential chain, in which activity was linked to fewer sleep problems, which in turn were linked to fewer negative emotions, which in turn were linked to lower pathological internet use. This chain mediation design matters because it implies that sleep is not merely another symptom of poor mood; it may sit upstream of emotional regulation, so that a night of fragmented sleep depletes the emotional resources an adolescent needs to resist compulsive scrolling and gaming.

The multi-group analysis then uncovered a striking gender asymmetry. Two specific structural paths differed significantly between boys and girls. The association between physical activity and sleep problems was substantially stronger in females, with a standardized coefficient of minus 0.38 compared with minus 0.22 in males, a difference confirmed by a critical-ratio test of differences yielding z equal to 2.94 and p less than 0.01. In other words, the same increase in physical activity appears to be linked to a larger reduction in sleep problems for girls than for boys. Conversely, the association between negative emotions and pathological internet use was significantly stronger in males, with a standardized coefficient of 0.51 versus 0.34 in females, z equal to 3.12 and p less than 0.01, suggesting that when boys experience negative emotions, those emotions are more tightly coupled to compulsive internet behavior than they are for girls.

These gender differences carry practical weight. If exercise exerts a stronger association with sleep quality in girls, then physical activity programs might be a particularly efficient lever for improving sleep in female students, which could then cascade through the chain toward reduced internet misuse. If negative emotions are more strongly tied to problematic internet use in boys, then interventions for male adolescents might benefit from pairing exercise with explicit emotional regulation training, teaching skills for identifying and managing distress before it migrates online. The authors suggest that comprehensive strategies integrating physical exercise, sleep hygiene, and gender-sensitive emotional regulation may inform the design of future intervention research, a formulation that is deliberately cautious about what the data can and cannot support.

That caution is essential, because the study is cross-sectional. All variables were measured at a single point in time, which means the arrows in the model reflect theoretical assumptions as much as empirical demonstration. It is equally plausible that adolescents who are already deeply engaged in pathological internet use withdraw from physical activity, sleep badly, and feel worse, reversing the direction of the chain. The authors themselves state that longitudinal and experimental studies are needed to establish the temporal and causal structure of these associations. They also restrict the scope of their claims with unusual precision: because the sample came from four public secondary schools in one metropolitan area, the findings are most directly applicable to those four schools and should not be assumed to generalize to other schools in the same city, other metropolitan areas, or Chinese adolescents in general without further replication.

Methodologically, the study shows the kind of rigor that strengthens confidence even within those limits. The researchers checked for common method bias using a common latent factor, examined intraclass correlations to account for the clustered sampling design, and reported composite reliability and average variance extracted figures supporting the quality of their measures. Model fit was evaluated with the standard family of indices, including the comparative fit index, the Tucker-Lewis index, the root mean square error of approximation, and the standardized root mean square residual. The ethics protocol was reviewed and approved by the Professor Committee of the School of Journalism and Communication at Beijing Sport University, and written informed consent was obtained from all participating adolescents and their parents or legal guardians, in accordance with the Declaration of Helsinki. The research received no external funding, and the authors declare no competing interests.

For parents, educators, and clinicians watching teenagers disappear into screens, the study offers a reframing that is more useful than simple scolding. Pathological internet use, in this account, is not an isolated vice to be banned but the visible end of a psychophysiological chain that begins with how much a young person moves, how well they sleep, and how they feel when they wake up. Exercise emerges not as a panacea but as a plausible upstream entry point, one whose downstream benefits appear to travel along different routes in boys and girls. Until longitudinal and experimental work confirms the causal structure, the safest reading is that physical activity, sleep, and emotional regulation are interrelated correlates of adolescent internet misuse, and that treating them together, with attention to gender, may be more productive than targeting screen time alone.

Subject of Research: Chain mediation of physical activity, sleep problems, and negative emotions in adolescent pathological internet use

Article Title: Association between physical activity and adolescent pathological internet use: a chain mediation model of sleep problems and negative emotions, with multi-group SEM analysis by gender

Article References: Association between physical activity and adolescent pathological internet use: a chain mediation model of sleep problems and negative emotions, with multi-group SEM analysis by gender. (n.d.). https://doi.org/10.1186/s40359-026-05370-z

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05370-z

Keywords: physical activity, pathological internet use, adolescents, sleep problems, negative emotions, chain mediation, structural equation modeling, Compensatory Internet Use Theory, gender differences, BMC Psychology, cross-sectional study, mental health