Tehran University of Medical Sciences Iranian Journal of Psychiatry 1735-4587 0 0 2026 09 13 1 12 Sevara Mukhammadieva Department of Faculty and Hospital Therapy No.1, Rheumatology, Occupational Pathology of Tashkent State Medical University, Tashkent, Uzbekistan. Kadirova Munira Fergana Medical Institute of Public Health, Fergana, Uzbekistan. Jumaniyozova Muhayyo Uzbek State World Languages University, Tashkent, Uzbekistan. Turakulova Iroda Department of Psychology, Samarkand State Pedagogical Institute, Samarkand, Uzbekistan. Oripova Mokhidil Department of Philosophy, Fergana State University, Fergana, 150100, Uzbekistan. Rasulov Ilkhom Department of Russian Language and Literature, Kokand State University, Kokand, Uzbekistan. Tulenova Gulmira Department of Humanities (Liberal Arts), Tashkent University of Information Technologies, Tashkent, Uzbekistan. Khatamova Sarvinoz Department of Forensic Medicine, Bukhara State Medical Institute, Bukhara, Uzbekistan. Erkin Ismoilov Department of Social Sciences and Humanities, Termez State University of Engineering and Agrotechnologies, Termez, Uzbekistan. 2026 07 09 2026 08 18 Objective: This study aims to transcend the limitations of aggregate screen time metrics by investigating the granular affective dynamics and emotional reactivity associated with specific social media behaviors in adolescents, while also incorporating the constructs of sleep quality and fear of missing out (FoMO). Method: A cohort of 346 adolescents participated in a 21-day ecological momentary assessment (EMA) protocol, receiving five prompts daily. Participants reported real-time social media activities (passive scrolling versus active interaction), concurrent affect, and pre-sleep social media use. Baseline and follow-up depressive symptoms were assessed using the patient health questionnaire-9 (PHQ-9). Additional measures included the FoMO scale and the Pittsburgh sleep quality index (PSQI). We utilized dynamic structural equation modeling (DSEM) to analyze within-person temporal associations and machine learning (random forest) to identify behavioral predictors of depression. Results: For 328 adolescents who completed the EMA protocol, DSEM revealed that passive scrolling predicted subsequent increases in negative affect (β = 0.31, P < 0.001), whereas active interaction predicted increased positive affect (β = 0.22, P = 0.003). High FoMO significantly moderated the relationship between passive use and negative affect (β = 0.15, P = 0.01). Furthermore, two distinct mediation pathways were identified. First, nighttime social media use mediated the association between daily passive scrolling and poorer next-day sleep quality (indirect effect = -0.10, 95% CI [-0.15, -0.05]). Second, aggregated poor sleep quality over the study period mediated the relationship between average passive scrolling and increased depressive symptoms at follow-up (indirect effect = 0.12, 95% CI [0.05, 0.19]). The machine learning model achieved high predictive accuracy for PHQ-9 scores (R² = 0.76, RMSE = 2.31), identifying high-frequency passive use, elevated FoMO, and delayed sleep onset as primary risk factors. Conclusion: The findings demonstrate that the emotional impact of social media is contingent upon usage patterns and individual traits like FoMO. Passive consumption was prospectively associated with negative emotional spirals and disrupted sleep, highlighting the need for interventions targeting specific digital behaviors and evening routines rather than total screen time. https://ijps.tums.ac.ir/index.php/ijps/article/view/4710 https://ijps.tums.ac.ir/index.php/ijps/article/download/4710/1356