Globally, preterm birth (defined as birth at < 37 weeks of gestation) affects around 11% of deliveries and it is closely associated with cerebral palsy, cognitive impairments and neuropsychiatric diseases in later life. Magnetic Resonance Imaging (MRI) has utility for measuring different properties of the brain during the lifespan. Specially, diffusion MRI has been used in the neonatal period to quantify the effect of preterm birth on white matter structure, which enables inference about brain development and injury. By combining information from both structural and diffusion MRI, is it possible to calculate structural connectivity of the brain. This involves calculating a model of the brain as a network to extract features of interest. The process starts by defining a series of nodes (anatomical regions) and edges (connections between two anatomical regions). Once the network is created, different types of analysis can be performed to find features of interest, thereby allowing group wise comparisons. The main frameworks/tools designed to construct the brain connectome have been developed and tested in the adult human brain. There are several differences between the adult and the neonatal brain: marked variation in head size and shape, maturational processes leading to changes in signal intensity profiles, relatively lower spatial resolution, and lower contrast between tissue classes in the T1 weighted image. All of these issues make the standard processes to construct the brain connectome very challenging to apply in the neonatal population. Several groups have studied the neonatal structural connectivity proposing several alternatives to overcome these limitations. The aim of this thesis was to optimise the different steps involved in connectome analysis for neonatal data. First, to provide accurate parcellation of the cortex a new atlas was created based on a control population of term infants; this was achieved by propagating the atlas from an adult atlas through intermediate childhood spatio-temporal atlases using image registration. After this the advanced anatomically-constrained tractography framework was adapted for the neonatal population, refined using software tools for skull-stripping, tissue segmentation and parcellation specially designed and tested for the neonatal brain. Finally, the method was used to test the effect of early nutrition, specifically breast milk exposure, on structural connectivity in preterm infants. We found that infants with higher exposure to breastmilk in the weeks after preterm birth had improved structural connectivity of developing networks and greater fractional anisotropy in major white matter fasciculi. These data also show that the benefits are dose dependent with higher exposure correlating with increased white matter connectivity. In conclusion, structural connectivity is a robust method to investigate the developing human brain. We propose an optimised framework for the neonatal brain, designed for our data and using tools developed for the neonatal brain, and apply it to test the effect of breastmilk exposure on preterm infants.

La parcellisation du cerveau humain en aires fonctionnelles est un problème complexe mais majeur pour la compréhension du fonctionnement du cerveau et pourrait avoir des applications médicales importantes en neurochirurgie par exemple pour mieux identifier les zones fonctionnelles à sauvegarder. Cet objectif va de pair avec l’idée de construire le connectome cérébral humain, qui n’est autre que le réseau de ses connexions.Pour définir un tel réseau, il faut en effet définir les éléments de ce réseau de connexions : c’est-à-dire avoir un découpage du cerveau en régions. Il existe de multiples manières et critères pour identifier ces régions et à ce jour il n’y a pas de parcellisation universelle du cortex. Dans cette thèse nous étudierons la possibilité d’effectuer cette parcellisation en fonction des données de connectivité anatomique, issues de l’imagerie par résonance magnétique de diffusion, qui est une technique d’acquisition permettant de reconstruire les faisceaux de neurones cérébraux de manière non invasive. Nous nous placerons dans un cadre surfacique en étudiant seulement la surface corticale et les connexions anatomiques sous-jacentes. Dans ce contexte nous présenterons un ensemble de nouveaux outils pour construire, visualiser et simuler le connectome cérébral humain, dans un cadre surfacique et à partir des données de connectivité anatomique reconstruites par IRM, et ceci pour un groupe de sujets. A partir de ces outils nous présenterons des méthodes de réduction de dimension des données de connectivité, que nous appliquerons pour parcelliser le cortex entier de quelques sujets. Nous proposons aussi une nouvelle manière de décomposer les données de connectivité au niveau d’un groupe de sujets en tenant compte de la variabilité inter-individuelle. Cette méthode sera testée et comparée à d’autres méthodes sur des données simulées et des données réelles. Les enjeux de ce travail sont multiples, tant au niveau méthodologique (comparaison de différents algorithmes de tractographie par exemple) que clinique (étude du lien entre altérations des connexions et pathologie). / In-vivo parcellation of the human cortex into functional brain areas is a major goal to better understand how the brain works and could have a lot of medical applications and give useful information to guide neurosurgery for example. This objective is related to the buildong of the human brain connectome, which is the networks of brain connections.Indeed, it is necessary to define the basic element of this connectome, and for doing this to have a subdivision of the cortex into brain regions. Actually, there is no such gold standard parcellation : there are a lot of techniques and methods to achieve this goal. During this PhD., anatomical connectivité based on diffusion-weighted imaging hase been used to address this problem, with a surfacic approach. In this context, we will present a set of new tools to create, visualize and simulate the human brain connectome for a group of subjects. We will introduce dimension reduction methods to compile the cortical connectivity profiles taking into account the interindividual variability. These methods will be apply to parcellate the cortex, for one subject or for a group of subjects simultaneously.There are many applications of this work, in methodology, to compare tractography algorithms for example or in clinical, to look at the relations between connections damages and pathology.

The current thesis addresses some of the unresolved predictions of recent models of the semantic brain system, such as the hub-and-spokes model. In particular, we tackle different aspects of the hypothesis that a widespread network of interacting heteromodal (hub(s)) and unimodal (spokes) cortices underlie semantic cognition. For this purpose, we use connectivity analyses, measures of graph theory and permutation-based statistics with source reconstructed Electro-/MagnetoEncephaloGraphy (EEG/MEG) data in order to track dynamic modulations of activity and connectivity within the semantic networks while a concept unfolds in the brain. Moreover, in order to obtain more accurate connectivity estimates of the semantic networks, we propose novel methods for some of the challenges associated with EEG/MEG connectivity analysis in source space. We utilised data-driven analyses of EEG/MEG recordings of visual word recognition paradigms and found that: 1) Bilateral Anterior Temporal Lobes (ATLs) acted as potential processor hubs for higher-level abstract representation of concepts. This was reflected in modulations of activity by multiple contrasts of semantic variables; 2) ATL and Angular Gyrus (AG) acted as potential integrator hubs for integration of information produced in distributed semantic areas. This was observed using Dynamic Causal Modelling of connectivity among the main left-hemispheric candidate hubs and modulations of functional connectivity of ATL and AG to semantic spokes by word concreteness. Furthermore, examining whole-brain connectomes using measures of graph theory revealed modules in the right ATL and parietal cortex as global hubs; 3) Brain oscillations associated with perception and action in low-level cortices, in particular Alpha and Gamma rhythms, were modulated in response to words with those sensory-motor attributes in the corresponding spokes, shedding light on the mechanism of semantic representations in spokes; 4) Three types of hub-hub, hub-spoke and spoke-spoke connectivity were found to underlie dynamic semantic graphs. Importantly, these results were obtained using novel approaches proposed to address two challenges associated with EEG/MEG connectivity. Firstly, in order to find the most suitable of several connectivity metrics, we utilised principal component analysis (PCA) to find commonalities and differences of those methods when applied to a dataset and identified the most suitable metric based on the maximum explained variance. Secondly, reconstruction of EEG/MEG connectomes using anatomical or fMRI-based parcellations can be significantly contaminated by spurious leakage-induced connections in source space. We, therefore, utilised cross-talk functions in order to optimise the number, size and locations of cortical parcels, obtaining EEG/MEG-adaptive parcellations. In summary, this thesis proposes approaches for optimising EEG/MEG connectivity analyses and applies them to provide the first empirical evidence regarding some of the core predictions of the hub-and-spokes model. The key findings support the general framework of the hub(s)-and-spokes, but also suggest modifications to the model, particularly regarding the definition of semantic hub(s).