The brain is probably the most complex system of the human body, composed of numerous neural units interconnected at dierent scales. This highly structured architecture provides the ability to communicate, synthesize information and perform the analytical tasks of human beings. Its development starts during the transition between the embryonic and fetal periods, from a simple tubular to a highly complex folded structure. It is globally organized as early as birth. This developing process is highly vulnerable to antenatal adverse conditions. Indeed, extreme prematurity and intra uterine growth restriction are major risk factors for long-term morbidities, including developmental ailments such as cerebral palsy, mental retardation and a wide spectrum of learning disabilities and behavior disorders. In this context, the characterization of the brainâ s normative wiring pattern is crucial for our understanding of its architecture and workings, as the origin of many neurological and neurobehavioral disorders is found in early structural brain development. Diusion magnetic resonance imaging (dMRI) allows the in vivo assessment of biological tissues at the microstructural level. It has emerged as a powerful tool to study brain connectivity and analyse the underlying substrate of the human brain, comprising its structurally integrated and functionally specialized architecture. dMRI has been widely used in adult studies. Nevertheless, due to technical constraints, this mapping at earlier stages of development has not yet been accomplished. Yet, this time period is of extreme importance to comprehend the structural and functional integrity of the brain. This thesis is motivated by this shortfall, and intends to fill the gap between the clinical and neuroscience demands and the methodological developments needed to fulfill them. In our work, we comprehensibly study the brain structural connectivity of children born extremely prematurely and/or with additional prenatal restriction at school-age. We provide evidence that brain systems that mature early in development are the most vulnerable to antenatal insults. Interestingly, the alterations highlighted in these systems correlate with the neurobehavioral and cognitive impairments seen in these children at school-age. The overall brain organization appear also altered after preterm birth and prenatal restriction. Indeed, these children show dierent brain network modular topology, with a reduction in the overall network capacity. What remains unclear is whether the alterations seen at school age are already present at birth and, if yes, to what extent. In this thesis we set the technical basis to enable the connectome analysis as early as at birth. This task is challenging when dealing with neonatal data. Indeed, most of the assumptions used in adult data processing methods do not hold, due to the inverted image contrast and other MRI artefacts such as motion, partial volume and intensity inhomogeneities. Here, we propose a novel technique for surface reconstruction, and provide a fully-automatic procedure to delineate the newborn cortical surface, opening the way to establish the newborn connectome.