L’étude de la complexité de l’anatomie du cerveau humain nécessite la caractérisation des paramètres multimodaux et multi-échelle obtenus par des techniques de neuroimagerie récentes. Pour ce travail de thèse nous avons tiré profit d’un logiciel automatique actuel d’analyse surfacique d’images cérébrales afin d’extraire les phénotypes structuraux du cortex cérébral humain, c’est-à-dire l’épaisseur corticale, l’aire de la surface, la profondeur sulcale, la courbure et le contenu en myéline intracorticale. L’objectif principal de ce travail a été de caractériser des variables structurales multimodales sur une large base de données de plus de 450 adultes sains âgés de 18 à 57 ans (base de données BIL&GIN) dans le but de décrire la variabilité interindividuelle de l’organisation structurale du cerveau et notamment la recherche de marqueurs de la maturation cérébrale et de la latéralisation. Nous avons tout d’abord pris l’exemple du gyrus de Heschl, support anatomique du cortex auditif primaire, qui possède une grande variabilité en lien avec l’existence de différents profils de duplication du gyrus couplée à de fortes différences interhémisphériques. Nous avons montré qu’une duplication partielle ou complète du gyrus de Heschl était associée à des modifications locorégionales d’épaisseur corticale, d’aire de la surface et de myéline localisée postérieurement à ce gyrus et dans le planum temporale, ces deux régions étant impliquées dans le traitement du langage. Dans une deuxième étude, nous avons recherché les modifications structurales du cortex associées à la maturation tardive (entre 18 et 30 ans) et à l’atrophie corticale liée au vieillissement. Nous avons montré que l’établissement d’un index de maturation basé sur l’intégration de l’épaisseur corticale et de la myéline intracorticale améliorait la discrimination entre les 2 profils de modifications de la substance grise pendant ces deux périodes de la vie. Finalement, nous avons caractérisé les asymétries corticales en utilisant un recalage surfacique des hémisphères qui s’affranchit des différences de morphologie sulcale et de position entre les deux hémisphères. Nous avons mis en évidence des régions pour lesquelles les asymétries d’épaisseur et de surface étaient concordantes (asymétrie gauche ou droite pour les deux variables anatomiques) et des régions pour lesquelles les asymétries étaient opposées (gauche pour l’une des variables et droite pour l’autre). Environ 20% des régions qui montraient une asymétrie d’épaisseur et d’aire présentaient des corrélations négatives entre ces variables. Il est frappant de constater que les deux régions ayant les asymétries les plus fortes, le planum temporale et le sillon temporal supérieur, ont des corrélations positives entre leurs asymétries d’épaisseur et d’aire. Le planum temporale possède une asymétrie gauche à la fois pour l’épaisseur et l’aire alors que le sillon temporal supérieur a une asymétrie droite pour les deux variables. Cette étude démontre qu’il existe des corrélations entre les asymétries d’épaisseur et d’aire qui sont caractéristiques de l’organisation du cortex. Ces régions sont des sites clé pour lesquels il reste maintenant à étudier la pertinence en tant que marqueurs de la latéralisation cérébrale et leurs corrélats fonctionnels. / Studying the complexity of the human brain anatomy requires the characterization of multimodal and multiscale features obtained by recent in vivo neuroimaging techniques. In the present thesis, we benefited from up to date automated surface-based brain image analysis software to extract structural phenotypes of the human cerebral cortex, namely the cortical thickness, the surface area, the sulcal depth, the curvature and the intracortical myelin content. The principal aim of this work was to characterize multimodal structural variables on a large database of 450 healthy adults aged from 18 to 57 years (the BIL&GIN database) in order to describe the inter-individual variability of brain structural organization and notably the research of candidate markers for brain maturation and lateralization. We first took the example of the Heschl’s gyrus hosting the primary auditory cortex and having high variability due to the presence of different pattern of gyrus duplication coupled with strong interhemispheric differences. We showed that the partial or complete duplication of the Heschl’s gyrus was associated to loco-regional modifications in terms of cortical thickness, surface area and myelin located posteriorly to this gyrus and in the planum temporale, this two regions being implied in language processing. In a second study, we investigated the cortical structural modifications associated to late maturation (between 18 and 30 years) and cortical atrophy linked to aging. We revealed that the computation of a maturation index based on an integration of cortical thickness and intracortical myelin improved the discrimination of two different patterns of grey matter changes during these different stages of life. Finally, we characterized cortical asymmetries using a specific hemisphere surface matching which removed differences in sulcal morphology and position between both hemispheres. We highlighted regions where thickness and surface area asymmetries were concordant (leftward or rightward asymmetry for both anatomical variables) and regions of opposite asymmetries (leftward for one and rightward for the other). About 20% of regions that showed cortical thickness and surface area asymmetries presented negative correlation between these variables. It is striking that the two regions with the strongest anatomical asymmetries; the planum temporale and the superior temporal sulcus had rather positive asymmetry correlations. The planum temporale presented a leftward asymmetry of both cortical thickness and area while the superior temporal sulcus showed a right asymmetry of the two variables. This study demonstrated that there were correlations between thickness and surface area asymmetries, characteristics of the cortex organization. These areas are key sites for which it now remains to study the anatomical relevance as markers of brain lateralization and its functional correlates.

Multimodality MR imaging of pathological changes in schizophrenia Aim: To prove structural changes of the neocortex and white matter of the brain indicating connectivity disorder in early phases of schizophrenia. Material and methods: A prospective monocentric study comparing a cohort of patients after the first episode of schizophrenia (on average 15.6 days after the initial hospitalization) with a control group of healthy persons. Probands were examined using a complex MRI protocol. Twenty-six patients and twenty-four healthy persons were examined in total. Three dimensional T1 and T2 data and DWI data were analyzed using TBSS FA, FBA a surface- based morphometry. Results: Large areas of dispersively decreased FA were found in patients compared to control group using TBSS. Several fixels of decreased FD metric were found using FBA in the anterior commissure of patients and one sporadic fixel of decreased FDC metric was found in frontal white matter of the brain. No statistically significant areas of cortical surface area and cortical thickness differences were found using SBM. Conclusions: Large areas of decreased microstructural integrity of the white matter of the brain were found. However, it was not possible to specify the nature of its corruption using FBA. Our findings indicate the crucial role of...

This thesis is divided into three main parts. In the first, we discuss that, although permutation tests can provide exact control of false positives under the reasonable assumption of exchangeability, there are common examples in which global exchangeability does not hold, such as in experiments with repeated measurements or tests in which subjects are related to each other. To allow permutation inference in such cases, we propose an extension of the well known concept of exchangeability blocks, allowing these to be nested in a hierarchical, multi-level definition. This definition allows permutations that retain the original joint distribution unaltered, thus preserving exchangeability. The null hypothesis is tested using only a subset of all otherwise possible permutations. We do not need to explicitly model the degree of dependence between observations; rather the use of such permutation scheme leaves any dependence intact. The strategy is compatible with heteroscedasticity and can be used with permutations, sign flippings, or both combined. In the second part, we exploit properties of test statistics to obtain accelerations irrespective of generic software or hardware improvements. We compare six different approaches using synthetic and real data, assessing the methods in terms of their error rates, power, agreement with a reference result, and the risk of taking a different decision regarding the rejection of the null hypotheses (known as the resampling risk). In the third part, we investigate and compare the different methods for assessment of cortical volume and area from magnetic resonance images using surface-based methods. Using data from young adults born with very low birth weight and coetaneous controls, we show that instead of volume, the permutation-based non-parametric combination (NPC) of thickness and area is a more sensitive option for studying joint effects on these two quantities, giving equal weight to variation in both, and allowing a better characterisation of biological processes that can affect brain morphology.