Neurodevelopmental and psychiatric disorders are characterized by complex patterns of comorbidity that unfold dynamically across development. Yet, current diagnostic approaches often reduce this complexity to static categories, potentially obscuring mechanisms linking early vulnerabilities to later psychiatric outcomes. This thesis aimed to develop multivariate and dynamic analytical approaches to characterize these developmental pathways, using 22q11.2 deletion syndrome (22q11DS) as a neurogenetic model of psychiatric vulnerability. Particular emphasis was placed on sleep and metabolic dysregulation as potentially modifiable mechanisms linking genetic risk, behavior, brain development, and psychopathology.

Four complementary studies were conducted. First, a dynamic network framework based on graph signal processing was developed to characterize longitudinal interactions among psychiatric symptoms and predict individual clinical trajectories. This approach identified reproducible pathways of vulnerability across independent genetic and clinical high-risk cohorts. Second, multivariate integration of objective and subjective sleep measures identified patterns of non-restorative sleep in 22q11DS that predicted persistent affective dysregulation and subsequent psychotic symptoms. Third, high-density polysomnography combined with deep behavioral phenotyping linked non-restorative sleep to disrupted slow-wave dynamics and to a transdiagnostic profile encompassing executive dysfunction, emotional dysregulation, motivational difficulties, and daytime somnolence. Finally, longitudinal metabolic, behavioral, and neuroimaging analyses identified a developmental pathway linking childhood impulsivity to persistent elevations in body mass index and, subsequently, increased psychosis vulnerability. The duration of elevated BMI showed a dose-dependent association with cerebellar gray-matter alterations and psychosis-related disorganization.

Together, these findings support a developmental and transdiagnostic conceptualization of psychopathology in which genetic vulnerability does not determine psychiatric outcomes in isolation but interacts dynamically with behavioral, physiological, and neurodevelopmental processes. They further identify sleep disturbance, impulsivity, and metabolic dysregulation as clinically actionable targets that may modify adverse developmental trajectories. More broadly, this work demonstrates how precision phenotyping and dynamic multivariate approaches can embrace the complexity of neurodevelopmental disorders while generating clinically interpretable insights, providing a framework for more personalized and preventive strategies in youth at elevated psychiatric risk.