Early adolescence is a transition period during which many mental health disorders emerge. The interplay between different internalizing and externalizing mental health problems in adolescence is poorly understood at the within-person level. Executive functioning (EF) in early adolescence has been shown to constitute a transdiagnostic risk factor, but the specificity of the associations between different domains of EF and mental health problems remains unclear. Network dynamics (ie, temporal effects) of different internalizing and externalizing symptoms were investigated leveraging data from the Tracking Adolescents’ Individual Lives Survey (TRAILS), a large longitudinal panel study of adolescents (>1,641 participants) assessed at ages 11, 13, and 15. Two novel methodological panel network approaches were used: cross-lagged panel network models and graphical vector autoregressive models. Hierarchical regression models were used to investigate prospective associations between different measures of EF and broadband transdiagnostic dimensions. Depressive problems predicted a range of other internalizing symptoms (ie, panic, somatic problems, separation anxiety, general anxiety, social phobia) over time, particularly during early adolescence. Important feedback loops with reciprocal associations between different anxiety symptoms were identified. Different facets of EF assessed at age 11, particularly sustained attention, showed weak but significant prospective associations with internalizing and externalizing symptoms at age 13. The present findings emphasize the importance of targeting depressive problems in early adolescence to prevent a spiral of different internalizing symptoms from arising later on.
Background and aims: Models of alcohol use risk suggest that drinking motives represent the most proximal risk factors on which more distal factors converge. However, little is known about how distinct risk factors influence each other and alcohol use on different temporal scales (within a given moment versus over time). We aimed to estimate the dynamic associations of distal (personality and life stressors) and proximal (drinking motives) risk factors, and their relationship to alcohol use in adolescence and early adulthood using a novel graphical vector autoregressive (GVAR) panel network approach.Design, setting and cases: We estimated panel networks on data from the IMAGEN study, a longitudinal European cohort study following adolescents across three waves (aged 16, 19 and 22 years). Our sample consisted of 1829 adolescents (51% females) who reported alcohol use on at least one assessment wave.Measurements: Risk factors included personality traits (NEO‐FFI: neuroticism, extraversion, openness, agreeableness and conscientiousness; SURPS: impulsivity and sensation‐seeking), stressful life events (LEQ: sum scores of stressful life events), and drinking motives [drinking motives questionnaire (DMQ): social, enhancement, conformity, coping anxiety and coping depression]. We assessed alcohol use [alcohol use disorders identification test (AUDIT): quantity and frequency] and alcohol‐related problems (AUDIT: related problems).Findings: Within a given moment, social [partial correlation (pcor) = 0.17] and enhancement motives (pcor = 0.15) co‐occurred most strongly with drinking quantity and frequency, while coping depression motives (pcor = 0.13), openness (pcor = 0.05) and impulsivity (pcor = 0.09) were related to alcohol‐related problems. The temporal network showed no predictive associations between distal risk factors and drinking motives. Social motives (beta = 0.21), previous alcohol use (beta = 0.11) and openness (beta = 0.10) predicted alcohol‐related problems over time (all P < 0.01).Conclusions: Heavy and frequent alcohol use, along with social drinking motives, appear to be key targets for preventing the development of alcohol‐related problems throughout late adolescence. We found no evidence for personality traits and life stressors predisposing towards distinct drinking motives over time.