MULTIMODAL MAGNETIC RESONANCE IMAGING IN PEDIATRIC STROKE AND CEREBROVASCULAR DISORDERS

Background Pediatric and perinatal stroke represent major causes of long-term neurodevelopmental disability, frequently leading to persistent motor, cognitive, behavioral, and visuospatial impairments. In contrast to adult stroke, ischemic injury in the developing brain occurs during critical phases of cerebral maturation, profoundly influencing structural and functional network organization. Recent advances in multimodal neuroimaging and connectomics have progressively shifted the conceptual framework of pediatric stroke from a focal vascular lesion toward a distributed network disorder characterized by widespread alterations in white matter architecture, functional connectivity, and neurodevelopmental plasticity. Objectives This thesis aimed to investigate the structural, vascular, and connectomic correlates of pediatric cerebrovascular injury through advanced multimodal neuroimaging techniques, with particular focus on developmental brain reorganization, white matter connectivity, cerebral perfusion, and network-related cognitive outcomes. Methods A multimodal neuroradiological approach integrating conventional MR, MR-angiography, diffusion tensor imaging (DTI) and structural connectomics, was applied across different pediatric cerebrovascular conditions, including perinatal arterial ischemic stroke. Structural connectivity analyses were performed using graph theoretical methods to characterize global and regional network topology. Imaging findings were correlated with neurodevelopmental and cognitive measures, including visuospatial and executive functioning. Results The studies included in this thesis demonstrate that pediatric cerebrovascular injury induces widespread alterations extending beyond the primary lesion site and involving distributed structural and functional brain networks. Connectomic analyses revealed reduced global efficiency, decreased mean network strength and density, and increased path length, supporting the presence of impaired large-scale network integration after stroke. Regional analyses identified selective disruption of fronto-parietal, temporo-parietal, insular, and orbitofrontal connectivity, regions critically involved in visuospatial processing, attention, and higher-order cognitive integration. Importantly, contralesional hemispheric networks appeared relatively preserved and occasionally demonstrated increased connectivity, supporting the presence of adaptive developmental neuroplasticity and compensatory reorganization within the intact hemisphere. Structural connectivity alterations were associated with cognitive performance, particularly visuospatial reasoning abilities, suggesting that disruption of integrative network hubs contributes to long-term neurocognitive vulnerability after pediatric stroke. Overall, the findings consistently support a network-based model of pediatric brain injury in which long-term neurological outcome depends not only on lesion size and location but also on the integrity, efficiency, and adaptive reorganization of distributed brain networks. Conclusions This thesis highlights the central role of multimodal neuroimaging and connectomic analysis in understanding pediatric cerebrovascular disease. Advanced quantitative imaging techniques provide valuable insights into mechanisms of developmental neuroplasticity, network disruption, and cognitive vulnerability following early brain injury. The integration of structural, perfusional, and connectomic biomarkers may contribute to improved prognostic stratification, personalized therapeutic planning, and the future development of precision neurorehabilitation strategies in pediatric stroke.