Projet de thèse en Imagerie et physique médicale
Sous la direction de Cyril Poupon.
Thèses en préparation à université Paris-Saclay , dans le cadre de École doctorale Electrical, optical, bio-physics and engineering , en partenariat avec Construction de grands instruments pour la neuroimagerie : de l'imagerie en population aux champs magnétiques ultra-hauts (laboratoire) , Faculté des sciences d'Orsay (référent) et de Université Paris-Saclay. Graduate School Sciences de l’ingénierie et des systèmes (2020-….) (graduate school) depuis le 01-10-2018 .
Résumé: Le cortex cérébral constitue le substrat anatomique de l'activité cérébrale, dont l'organisation à l'échelle cellulaire reste encore à découvrir. Bien que son organisation en couches ait été mise en évidence à partir d'études histologiques dès le début du 20ème siècle, on ne dispose aujourd'hui pas d'outil permettant de la caractériser in vivo à l'échelle individuelle. Il est en de même pour l'organisation cellulaire ou cytoarchitecture de ces couches corticales. Le lien entre aires cytoarchitectoniques et aires fonctionnelles reste peu connu, et cette thèse vise donc à développer de nouveaux outils d'imagerie permettant de caractériser finement la structure laminaire du cortex cérébral et l'organisation de ces couches à l'échelle cellulaire, in vivo, chez l'homme. Les méthodes mises en œuvre seront mises à profit pour cartographier finement la structure laminiaire et la cytoarchitecture du cortex d'individus d'une population de sujets sains et validées au travers d'une étude réalisée post-mortem sur un cerveau humain bénéficiant d'un marquage histologique réalisé a posteriori.
Ultra-high field diffusion and quantitative MRI with strong gradients to explore the connectivity, cytoarchitecture and myeloarchitecture of animal and human brains at the mesoscale
Summary: The cerebral cortex is known to be the anatomical substrate supporting brain activity, but its organization at cellular scales still have to be investigated more in details. In spite of the fact that the existence of several cortical layers was established from histological studies back in the beginning of the 20th century, it does not exist any method available to characterize it in vivo at the individual scale yet. Access to the cellular organization (or cytoarchitecture) of these cortical layers suffers from the same limitation. In addtion, the relationship between cytoarchitectonics and functional areas is still poorly understood, and this PhD thesis aims at developing novel in vivo imaging tools to accurately map and characterize the laminar structure of the cerebral cortex as well as its cellular composition, in humans. After validation on a post-mortem human brain sample using both MRI and histological stainins, the methods developed during the thesis will be exploited to map the cortex layers and their microstructure of a small cohort of healthy subjects and to create a novel probabilistic atlas of the human brain cytoarchitecture. Details: Since the beginning of the 20th century, neuroanatomists have studied the human cortex and have shown that it depicts a laminar structure with several layers being characterized by different cellular organizations. This observation allowed Korbinian Brodmann, to define in 1905 43 distinct areas at the cortical surface based on the observation, using optical microscopy, of the differences of these cellular compositions between layers, from a human brain ex vivo sample, giving birth to the famous Brodmann atlas, still widely used by the community. Because it was established from a single human brain sample, it is unfortunately not possible to capture the inter-subject variability of these areas. In addition, it does not exit any consencious in the community that the boundaries of such cytoarchitecture-based cortex areas correspond to the boundaries separating specific functional networks. Having the possibility to further investigate this question using in vivo functional magnetic resonance imaging (fMRI) would certainly help investigating the correlates existing between the localization of functional networks and their underlying cytoarchitecture. Developing new in vivo atlases of the human brain is one of the main challenges of Human Brain Project, a European flagship aiming at creating a multi-modal / multi-scale atlas of the human brain. The Neurospin team is strongly involved in this project, having in charge the acquisition of post-mortem and in vivo MRI databases as well as to contribute to the construction of the atlas through the development of novel tools to extract various brain structures (its structural connectivity, its cortical folding, its cytoarchitecture, ). More specifically, the purpose of this thesis is to develop methods to map in vivo the cortex cytoarchitecture and to study the correlates between functional networks and cytoarchitectonics. To achieve this goal, the PhD candidate will follow a four-fold strategy : - first, the candidate will develop new methods to automatically segregate the laminar structure of the cortical ribbon. It has been recently shown that the sensitivity of both quantitative and diffusion magnetic resonance imaging (qMRI and dMRI) to the cellular environment of tissues allow to infer Gaussian mixture models that help characterizing the local myelo- and cyto- architectures. The same approach can be used in vivo and to this aim, the candidate will actively participate to the acquisition of the HBP/BICKET (Brain Imaging of the Cytoarchitecture using a Key Ermerging Technology) cohort acquired on the clinical Prisma MRI system of Neurospin and providing a massive set of quantitative and diffusion MRI data for 15 healthy human subjects (each subject will follow 12 imaging sessions providing ~3400 qMRI, dMRI and fMRI different contrasts). These massive individual datasets will enable the robust fitting of such Gaussian mixture, and the accurate delineation of individual cortical layers ; - second, the candidate will develop novel clustering approaches to define new individual and inter-subject probabilistic cortical areas using the information provided by myelo- and cyto- architectural features inferred from qMRI and dMRI in each layer and perpendicular to the cortex ; this will allow to define myelo- and cyto-based areas over the pial surface ; - third, the candidate will validate results on a post-mortem brain sample acquired on the preclinical Bruker 11.7T MRI system allowing to reach very high spatial resolution ; histological stainings will be performed for specific cortical regions to assess the segregation obtained from qMRI and dMRI and thus validate the approach - last, using the in vivo MRI datasets available from the HBP/BICKET cohort, the candidate will study the relationship between some specific functional networks (visual network, language network, motor network), and their underlying cytoarchitecture, to investigate whether the boundaries of cytoarchitectonic maps correspond to the frontiers of functional networks. The outcome of the PhD thesis will consist of a new acquisition and processing pipeline allowing to extract at the individual scale the cortical layers and to parcellate the pial surface into cytoarchitectural areas. It will also consist in the construction of a probabilistic atlas of the cytoarchitecture of the cortical ribbon at the scale of the scanned population. This thesis will involve the collaboration between several teams from MR physics to image analysis, clinical and cognitive neurosciences. The candidate should be graduated with a Master or Engineer diploma in medical imaging or medical physics either from a university or from an engineering school, should be familiar with magnetic resonance imaging. Skills in neuroimaging and/or neurosciences would be appreciated. The candidate should also be able to interact within a multidisciplinary environment, be fluent in english and familiar with IT tools (computer programming in C++ and python would be appreciated).