Public Thesis defense - ICTEAM

White matter microstructure estimation and visualization in neuropathological brains via diffusion MRI by Nicolas Delinte

The brain, with its intricate network of neural cells, their connections, and associated functions, forms a complex structure that has long fascinated researchers. Recent advances in MRI technology, particularly diffusion MRI, have enabled the in vivo investigation of connections between brain regions and the microstructure of the neural fibers linking those regions. However, these advancements and the methodologies they enabled could benefit from further improvements. On the macroscopic scale, identifying these connections relies on tractography algorithms that require either pre-defined atlases or extensive manual corrections and expertise to delineate the streamlines accurately. On the microscopic scale, advancements in microstructural models have enabled more precise estimations of the direction and properties of each fiber bundle within a voxel. However, the increasing number and complexity of these models have delayed their adoption in medical research. During this thesis, tools were developed to ease the generation and cleaning of tractography outputs and to facilitate the comparison and interpretation of models estimating multiple fiber populations per voxel. These tools included a modification to the tractography algorithm, a method for filtering spurious streamlines, and a framework attributing multi-fixel metrics to specific tracts. These methods were validated using synthetic phantoms, in vivo scans, and population studies involving patients with brain damage. These studies included the effects of abstinence in alcohol-dependent participants and the impact of intense motor training on participants with impaired motor control due to cerebral palsy and stroke. The tools developed are not dependent on specific methodologies or algorithms, and all the code has been made open source to enhance ease of use. Overall, these tools could benefit researchers analyzing specific white matter tracts, especially in non-conventional brains.

Jury members:

Dr.Laurence Dricot (UCLouvain), supervisor

Prof. Benoît Macq (UCLouvain), supervisor

Prof. Philippe Lefèvre (UCLouvain), chairperson

Prof. Philippe de Timary (UCLouvain), secretary

Dr.Gaëtan Rensonnet (UCLouvain)

Prof. Gloria Menegaz (University of Verona, Italy)

Prof. Jean-Philippe Thiran (EPFL, Suisse)

Pay attention :

The public defense of Nicolas Delinte scheduled for Monday September 16 at 05:00 p.m. will also take place in the form of a video conference.

Télécharger l’annonce.