Section: New Results

Image processing on Diffusion Weighted Magnetic Resonance Imaging

Non-Local Robust Detection of DTI White Matter Differences with Small Databases

Participants : Olivier Commowick, Aymeric Stamm.

Diffusion imaging, through the study of water diffusion, al- lows for the characterization of brain white matter, both at the population and individual level. In recent years, it has been employed to detect brain abnormalities in patients suffering from a disease, e.g. from multiple sclerosis (MS). State-of-the-art methods usually utilize a database of matched (age, sex, ...) controls, registered onto a template, to test for differences in the patient white matter. Such approaches however suffer from two main drawbacks. First, registration algorithms are prone to local errors, thereby degrading the comparison results. Second, the database needs to be large enough to obtain reliable results. However, in medical imaging, such large databases are hardly available. In this paper, we propose a new method that addresses these two issues. It relies on the search for samples in a local neighborhood of each pixel to increase the size of the database. Then, we propose a new test based on these samples to perform a voxelwise comparison of a patient image with respect to a population of controls. We demonstrate on simulated and real MS patient data how such a framework allows for an improved detection power and a better robustness and reproducibility, even with a small database [26] .

Registration and Analysis of White Matter Group Differences with a Multi-Fiber Model

Participant : Olivier Commowick.

Diffusion magnetic resonance imaging has been used extensively to probe the white matter in vivo. Typically, the raw diffusion images are used to reconstruct a diffusion tensor image (DTI). The incapacity of DTI to represent crossing fibers leaded to the development of more sophisticated diffusion models. Among them, multi-fiber models represent each fiber bundle independently, allowing the direct extraction of diffusion features for population analysis. However, no method exists to properly register multi-fiber models, seriously limiting their use in group comparisons. This paper presents a registration and atlas construction method for multi-fiber models. The validity of the registration is demonstrated on a dataset of 45 subjects, including both healthy and unhealthy subjects. Morphometry analysis and tract-based statistics are then carried out, proving that multi-fiber models registration is better at detecting white matter local differences than single tensor registration.

This work was done in collaboration with Maxime Taquet, Benoit Scherrer, Jurriaan Peters, Mustafa Sahin, Benoît Macq and Simon K. Warfield [44] .

Automated delineation of white matter fiber tracts with a multiple region-of-interest approach

Participant : Olivier Commowick.

White matter fiber bundles of the brain can be delineated by tractography utilizing multiple regions-of-interest (MROI) defined by anatomical landmarks. These MROI can be used to specify regions in which to seed, select, or reject tractography fibers. Manual identification of anatomical MROI enables the delineation of white matter fiber bundles, but requires considerable training to develop expertise, considerable time to carry out and suffers from unwanted inter- and intra-rater variability. In a study of 20 healthy volunteers, we compared three methodologies for automated delineation of the white matter fiber bundles. Using these methodologies, fiber bundle MROI for each volunteer were automatically generated. We assessed three strategies for inferring the automatic MROI utilizing nonrigid alignment of reference images and projection of template MROI. We assessed the bundle delineation error associated with alignment utilizing T1-weighted MRI, fractional anisotropy images, and full tensor images. We confirmed the smallest delineation error was achieved using the full tensor images. We then assessed three projection strategies for automatic determination of MROI in each volunteer. Quantitative comparisons were made using the root-mean-squared error observed between streamline density images constructed from fiber bundles identified automatically and by manually drawn MROI in the same subjects. We demonstrate that a multiple template consensus label fusion algorithm generated fiber bundles most consistent with the manual reference standard.

This work was done in collaboration with Ralph Suarez, Sanjay Prabhu and Simon K. Warfield [23] .

Corticospinal tractography with morphological, functional and diffusion tensor MRI: a comparative study of four deterministic algorithms used in clinical routine

Participants : Sylvain Prima, Camille Maumet, Jean-Christophe Ferré.

In collaboration with Romuald Seizeur, Nicolas Wiest-Daesslé and Xavier Morandi, we aimed to compare four deterministic tractography algorithms used in clinical routine for the study of the corticospinal tract (the bundle mediating voluntary movement) in 15 right-handed volunteers. We found no difference between right and left sides of the brain for any of the algorithms [22] .

A new multi-directional fiber model for low angular resolution diffusion imaging

Participants : Aymeric Stamm, Christian Barillot.

Diffusion MRI is a tool of choice for the analysis of the brain white matter fiber pathways. When translated to clinics, the short acquisition time leads to low angular resolution diffusion (LARD) images. Fiber pathways are then inferred assuming Gaussian diffusion (a.k.a. DTI) that provides one fiber orientation per voxel. In the past decade, recent researches highlight more intricate intra-voxel fiber configurations using higher angular resolution diffusion images. In collaboration with Patrick Perez (Technocolor), we have proposed a non-Gaussian diffusion model of the white matter fibers able to recover from crossing of fibers even from low angular resolution.This model enables crossing fibers to be theoretically estimated from only 8 diffusion MR images. In particular, this model allows for the retrospective study of DW data sets acquired in the past.  [42]  [43] .