This work explores if combining information from multiple Magnetic Resonance Imaging (MRI) modalities provides improved interpretation of brain biological architecture as each MR modality can reveal different characteristics of underlying anatomical structures. Structural MRI provides a means for high-resolution quantitative study of brain morphometry. Diffusion-weighted MR imaging (DWI) allows for low-resolution modeling of diffusivity properties of water molecules. Structural and diffusion-weighted MRI modalities are commonly used for monitoring the biological architecture of the brain in normal development or neurodegenerative disease processes. Structural MRI provides an overall map of brain tissue organization that is useful for identifying distinct anatomical boundaries that define gross organization of the brain. DWI models provide a reflection of the micro-structure of white matter (WM), thereby providing insightful information for measuring localized tissue properties or for generating maps of brain connectivity. Multispectral information from different structural MR modalities can lead to better delineation of anatomical boundaries, but careful considerations should be taken to deal with increased partial volume effects (PVE) when input modalities are provided in different spatial resolutions. Interpretation of diffusion-weighted MRI is strongly limited by its relatively low spatial resolution. PVE's are an inherent consequence of the limited spatial resolution in low-resolution images like DWI. This work develops novel methods to enhance tissue classification by addressing challenges of partial volume effects encountered from multi-modal data that are provided in different spatial resolutions. Additionally, this project addresses PVE in low-resolution DWI scans by introducing a novel super-resolution reconstruction approach that uses prior information from multi-modal structural MR images provided in higher spatial resolution. The major contributions of this work include: 1) Enhancing multi-modal tissue classification by addressing increased PVE when multispectral information come from different spatial resolutions. A novel method was introduced to find pure spatial samples that are not affected by partial volume composition. Once detecting pure samples, we can safely integrate multi-modal information in training/initialization of the classifier for an enhanced segmentation quality. Our method operates in physical spatial domain and is not limited by the constraints of voxel lattice spaces of different input modalities. 2) Enhancing the spatial resolution of DWI scans by introducing a novel method for super-resolution reconstruction of diffusion-weighted imaging data using high biological-resolution information provided by structural MRI data such that the voxel values at tissue boundaries of the reconstructed DWI image will be in agreement with the actual anatomical definitions of morphological data. We used 2D phantom data and 3D simulated multi-modal MR scans for quantitative evaluation of introduced tissue classification approach. The phantom study result demonstrates that the segmentation error rate is reduced when training samples were selected only from the pure samples. Quantitative results using Dice index from 3D simulated MR scans proves that the multi-modal segmentation quality with low-resolution second modality can approach the accuracy of high-resolution multi-modal segmentation when pure samples are incorporated in the training of classifier. We used high-resolution DWI from Human Connectome Project (HCP) as a gold standard for super-resolution reconstruction evaluation to measure the effectiveness of our method to recover high-resolution extrapolations from low-resolution DWI data using three evaluation approaches consisting of brain tractography, rotationally invariant scalars and tensor properties. Our validation demonstrates a significant improvement in the performance of developed approach in providing accurate assessment of brain connectivity and recovering the high-resolution rotationally invariant scalars (RIS) and tensor property measurements when our approach was compared with two common methods in the literature. The novel methods of this work provide important improvements in tools that assist with improving interpretation of brain biological architecture. We demonstrate an increased sensitivity for volumetric and diffusion measures commonly used in clinical trials to advance our understanding of both normal development and disease induced degeneration. The improved sensitivity may lead to a substantial decrease in the necessary sample size required to demonstrate statistical significance and thereby may reduce the cost of future studies or may allow more clinical and observational trials to be performed in parallel.

L'Imagerie par Résonance Magnétique pondérée en Diffusion (IRM-D) permet l'accès à l'information structurelle des tissus au travers de la lecture du mouvement brownien des molécules d'eau. Ses applications sont nombreuses en imagerie cérébrale, tant en milieu clinique qu'en recherche. Néanmoins le mouvement physiologique créé une perte de signal supplémentaire au cours de l'encodage de la diffusion. Cette perte de signal liée au mouvement limite les applications de l'IRM-D quant à l'imagerie cardiaque. L'utilisation de faibles valeurs de pondération (b) réduit cette sensibilité mais permet seulement l'imagerie du mouvement incohérent intra-voxel (IVIM) qui contient la circulation sanguine et la diffusion des molécules d'eau. L'imagerie IVIM possède pourtant de nombreuses applications en IRM de l'abdomen, depuis la caractérisation tissulaire à la quantification de la perfusion, mais reste inexplorée pour l'imagerie du coeur. Mon travail de thèse correspond à l'évaluation des conditions d'application de l'IRM-D à faibles valeurs de b pour le coeur humain, afin de proposer des contributions méthodologiques et d'appliquer les techniques développées expérimentalement. Nous avons identifié le mouvement cardiaque comme une des sources majeures de perte de signal. Bien que le mouvement global puisse être corrigé par un recalage non-rigide, la perte de signal induite par le mouvement perdure et empêche l'analyse précise par IRM-D du myocarde. L'étude de cette perte de signal chez un volontaire a fourni une fenêtre temporelle durable où le mouvement cardiaque est au minimum en diastole. Au sein de cette fenêtre optimale, la fluctuation de l'intensité atteste d'un mouvement variable résiduel. Une solution de répéter les acquisitions avec un déclenchement décalé dans le temps permet la capture des minimas du mouvement, c.-à-d. des maximas d'intensité en IRM-D. La projection du maximum d'intensité dans le temps (TMIP) permet ensuite de récupérer des images pondérées en diffusion avec un minimum de perte de signal lié au mouvement. Nous avons développé et évalué différentes séquences d'acquisition combinées avec TMIP : la séquence d'imagerie écho-planaire classique par écho de spin (SE-EPI) peut être adaptée mais souffre du repliement d'image ; une séquence Carr-Purcell-Meiboom-Gill combinée avec une préparation d'encodage de diffusion est plus robuste aux distorsions spatiales mais des artefacts de bandes noires empêchent son applicabilité ; finalement une séquence double-SE-EPI compensant les courants de Foucault et pleinement optimisée produit des images IRM-D moins artefactées. Avec cette séquence, l'IRM-D-TMIP permet la réduction significative de la perte de signal liée au mouvement pour l'imagerie cardiaque pondérée en diffusion. L'inconvénient avec TMIP vient de l'amplification du bruit positif d'intensité. Afin de compenser cette sensibilité du TMIP, nous séparons le bruit d'intensité des fluctuations lentes liées au mouvement grâce à une nouvelle approche basée sur l'analyse en composantes principales (PCA). La décomposition préserve les détails anatomiques tout en augmentant les rapports signal et contraste-à-bruit (SNR, CNR). Avec l'IRM-D-PCATMIP, nous augmentons à la fois l'intensité finale et la qualité d'image (SNR) en théorie et expérimentalement. Les bénéfices ont été quantifiés en simulation avant d'être validés sur des volontaires. De plus la technique a montré des résultats reproductibles sur des patients post-infarctus aigue du myocarde, avec un contraste cohérent avec la position et l'étendue de la zone pathologique. Contrairement à l'imagerie cérébrale, l'imagerie IRM-D par faibles valeurs de pondération in vivo doit être différentiée des analyses IRM-D ex-vivo. Ainsi l'IRM-D-PCATMIP offre une technique sans injection pour l'exploration du myocarde par imagerie IVIM. Les premiers résultats sont encourageants pour envisager l'application sur un modèle expérimental d'une maladie cardiovasculaire [etc...] / Diffusion weighted magnetic resonance imaging (DW-MRI, or DWI) enables the access to the structural information of body tissues through the reading of water molecules Brownian motion. Its applications are many in brain imaging, from clinical practice to research. However physiological motion induces an additional signal-loss when diffusion encoding is applied. This motion-induced signal-loss limits greatly its applications in cardiac imaging. Using low diffusion-weighting values (b) DWI reduces this sensitivity but permits only the imaging of intravoxel incoherent motion (IVIM), which combines both water diffusion and perfusion. IVIM imaging has many applications in body MRI, from tissue characterization to perfusion quantification but remains unexplored for the imaging of the heart. The purpose of this work was to evaluate the context of low b-values DWI imaging of the heart, propose methodological contributions and then apply the developed techniques experimentally. We identified cardiac motion as one of the major sources of motion-induced signal loss. Although bulk motion can be corrected with a non-rigid registration algorithm, additional signal-loss remains uncorrected for and prevents accurate DWI of the myocardium. The study of diffusion-weighted signal-loss induced by cardiac motion in a volunteer provided a time-window when motion is at minimum in diastole. Within this optimal time-window, fluctuation of intensity attests of variable remaining physiological motion. A solution to repeat acquisition with shifted trigger-times ease the capture of motion amplitude minima, i.e. DWI-intensity maxima. Temporal maximum intensity projection (TMIP) finally retrieves diffusion weighted images of minimal motion-induced signal-loss. We evaluated various attempts of sequence development with TMIP: usual spin-echo echo-planar imaging (se-EPI) sequence can be improved but suffers aliasing issues; a balanced steady-state free-precession (b-SSFP) combined with a diffusion preparation is more robust to spatial distortions but typical banding artifacts prevent its applicability; finally a state-of-the-art double-spin-echo EPI sequence produces less artifacted DWI results. With this sequence, TMIP-DWI proves to significantly reduce motion-induced signal-loss in the imaging of the myocardium. The drawback with TMIP comes from noise spikes that can easily be highlighted. To compensate for TMIP noise sensitivity, we separated noise spikes from smooth fluctuation of intensity using a novel approach based on localized principal component analysis (PCA). The decomposition was made so as to preserve anatomical features while increasing signal and contrast to noise ratios (SNR, CNR). With PCATMIP-DWI, both signal-intensity and SNR are increased theoretically and experimentally. Benefits were quantified in a simulation before being validated in volunteers. Additionally the technique showed reproducible results in a sample of acute myocardial infarction (AMI) patients, with a contrast matching the extent and location of the injured area. Contrarily to brain imaging, in vivo low b-values DWI should be differentiated from ex vivo DWI pure diffusion measurements. Thus PCATMIP-DWI might provide an injection-free technique for exploring cardiac IVIM imaging. Early results encourage the exploration of PCATMIP-DWI in an experimental model of cardiac diseases. Moreover the access to higher b values would permit the study of the full IVIM model for the human heart that retrieves and separates both perfusion and diffusion information