Section: New Results

Computational Diffusion MRI

Coarse-Grained Spatiotemporal Acquisition Design for Diffusion MRI

Participants : Patryk Filipiak, Rutger Fick [TheraPanacea, Paris] , Alexandra Petiet [ICM, CENIR, Paris] , Mathieu Santin [ICM, CENIR, Paris] , Anne-Charlotte Philippe [ICM, CENIR, Paris] , Stéphane Lehericy [ICM, CENIR, Paris] , Demian Wassermann [Inria Parietal] , Rachid Deriche.

Acquisition protocols that allow to capture time-dependent changes in diffusion signal require long imaging time. We address this issue through an optimized subsampling scheme that maximizes accuracy of the spatiotemporal diffusion signal representation, qτ-dMRI, for given time constraints. Our proposed coarse-grained variant of the problem reduces the space of feasible acquisition parameters compared to the fine-grained approach causing no significant deterioration of a reconstruction accuracy in most of the studied cases.

This work has been published in [25].

A Computational Framework For Generating Rotation Invariant Features And Its Application In Diffusion MRI

Participants : Mauro Zucchelli, Samuel Deslauriers-Gauthier, Rachid Deriche.

In this work, we present a novel computational framework for analytically generating a complete set of algebraically independent Rotation Invariant Features (RIF) given the Laplace-series expansion of a spherical function. Our computational framework provides a closed-form solution for these new invariants, which are the natural expansion of the well known spherical mean, power-spectrum and bispectrum invariants. We highlight the maximal number of algebraically independent invariants which can be obtained from a truncated Spherical Harmonic (SH) representation of a spherical function and show that most of these new invariants can be linked to statistical and geometrical measures of spherical functions, such as the mean, the variance and the volume of the spherical signal. Moreover, we demonstrate their application to dMRI signal modeling including the Apparent Diffusion Coefficient (ADC), the diffusion signal and the fiber Orientation Distribution Function (fODF). In addition, using both synthetic and real data, we test the ability of our invariants to estimate brain tissue microstructure in healthy subjects and show that our framework provides more flexibility and open up new opportunities for innovative development in the domain of microstructure recovery from diffusion MRI.

This work has been published in [20].

A Novel Characterization of Traumatic Brain Injury in White Matter with Diffusion MRI Spherical-Harmonics Rotation Invariants

Participants : Mauro Zucchelli, Samuel Deslauriers-Gauthier, Drew Parker [Penn Applied Connectomics and Imaging Group, Philadelphia] , Junghoon John Kim [Department of Molecular, Cellular & Biomedical Sciences, New York] , Ragini Verma [Penn Applied Connectomics and Imaging Group, Philadelphia] , Rachid Deriche.

The current DTI-based markers of traumatic brain injury are able to capture affected WM in the brain, but missthe areas of crossing fibers and complex WM due to the simplicity of the model. In this work, we use a novelset of spherical-harmonics rotation invariant indices, recently proposed in the literature. We demonstrate thatthese 12 invariants capture all the information provided by DTI. But in addition, they capture differences incomplex WM, beyond DTI measures. This combined with the clinical feasibility of the method, paves the wayfor them to be used as better markers of brain injury.

This work has been published in [31].

The Dmipy Toolbox: Diffusion MRI Multi-Compartment Modeling and Microstructure Recovery Made Easy

Participants : Rutger Fick [TheraPanacea, Paris] , Demian Wassermann [Inria Parietal] , Rachid Deriche.

Non-invasive estimation of brain microstructure features using diffusion MRI (dMRI)—known as Microstructure Imaging—has become an increasingly diverse and complicated field over the last decades. Multi-compartment (MC)-models, representing the measured diffusion signal as a linear combination of signal models of distinct tissue types, have been developed in many forms to estimate these features. However, a generalized implementation of MC-modeling as a whole, providing deeper insights in its capabilities, remains missing. To address this fact, we present Diffusion Microstructure Imaging in Python (Dmipy), an open-source toolbox implementing PGSE-based MC-modeling in its most general form. Dmipy allows on-the-fly implementation, signal modeling, and optimization of any user-defined MC-model, for any PGSE acquisition scheme. Dmipy follows a “building block”-based philosophy to Microstructure Imaging, meaning MC-models are modularly constructed to include any number and type of tissue models, allowing simultaneous representation of a tissue's diffusivity, orientation, volume fractions, axon orientation dispersion, and axon diameter distribution. In particular, Dmipy is geared toward facilitating reproducible, reliable MC-modeling pipelines, often allowing the whole process from model construction to parameter map recovery in fewer than 10 lines of code. To demonstrate Dmipy's ease of use and potential, we implement a wide range of well-known MC-models, including IVIM, AxCaliber, NODDI(x), Bingham-NODDI, the spherical mean-based SMT and MC-MDI, and spherical convolution-based single- and multi-tissue CSD. By allowing parameter cascading between MC-models, Dmipy also facilitates implementation of advanced approaches like CSD with voxel-varying kernels and single-shell 3-tissue CSD. By providing a well-tested, user-friendly toolbox that simplifies the interaction with the otherwise complicated field of dMRI-based Microstructure Imaging, Dmipy contributes to more reproducible, high-quality research.

This work has been published in [12].

Effects of tractography filtering on the topology and interpretability of connectomes.

Participants : Matteo Frigo, Samuel Deslauriers-Gauthier, Drew Parker [Penn Applied Connectomics and Imaging Group, Philadelphia] , Abdol Aziz Ould Ismail [Penn Applied Connectomics and Imaging Group, Philadelphia] , Junghoon John Kim [Department of Molecular, Cellular & Biomedical Sciences, New York] , Ragini Verma [Penn Applied Connectomics and Imaging Group, Philadelphia] , Rachid Deriche.

The analysis of connectomes and their associated network metrics forms an important part of clinical studies. These connectomes are based on tractography algorithms to estimate the structural connectivity between brain regions. However, tractography algorithms, are prone to false positive connections and this affects the quality of the connectomes. Several tractography filtering techniques (TFTs) have been proposed to alleviate this issue in studies, but their effect on connectomic analyses of pathology has not been investigated. The aim of this work is to investigate how TFTs affect network metrics and their interpretation in the context of clinical studies.

This work has been published in [29]

Spherical convolutional neural network for fiber orientation distribution function and micro-structure parameter estimation from dMRI

Participants : Sara Sedlar, Samuel Deslauriers-Gauthier, Théodore Papadopoulo, Rachid Deriche.

Convolutional neural networks (CNNs) are proven to be a powerful tool for many computer vision problems where the data is acquired on a regular grid in Euclidean space. As the dMRI signals used in our experiments are acquired on spheres, we have investigated spherical CNN model (S2-CNN). In regular CNNs, during convolution, kernels are translated over the input feature maps with equidistant steps. In S2-CNN, both kernels and feature maps are represented in the 3D rotation group - SO(3) manifold. A rotation in SO(3) is analogous to a translation in Euclidean space. However, there is no regular equidistant grid in SO(3). As a consequence, the convolution is performed in the rotational harmonics (Fourier) domain. In this work, we investigate how the S2-CNN can be adapted to properties of dMRI data, such as antipodal symmetry, the presence of Rician noise, multiple sampling shells, etc.

This work currently in progress.

Adaptive phase correction of diffusion-weighted images

Participants : Marco Pizzolato [Signal Processing Lab (LTS5), EPFL, Lausanne] , Guillaume Gilbertb [MR Clinical Science, Philips Healthcare Canada, Markham, ON] , Jean-Philippe Thiran [Signal Processing Lab (LTS5), EPFL, Lausanne] , Maxime Descoteaux [Université de Sherbrooke, Sherbrooke] , Rachid Deriche.

Phase correction (PC) is a preprocessing technique that exploits the phase of images acquired in Magnetic Resonance Imaging (MRI) to obtain real-valued images containing tissue contrast with additive Gaussian noise, as opposed to magnitude images which follow a non-Gaussian distribution, e.g. Rician. PC finds its natural application to diffusion-weighted images (DWIs) due to their inherent low signal-to-noise ratio and consequent non-Gaussianity that induces a signal overestimation bias that propagates to the calculated diffusion indices. PC effectiveness depends upon the quality of the phase estimation, which is often performed via a regularization procedure. We show that a suboptimal regularization can produce alterations of the true image contrast in the real-valued phase-corrected images. We propose adaptive phase correction (APC), a method where the phase is estimated by using MRI noise information to perform a complex-valued image regularization that accounts for the local variance of the noise. We show, on synthetic and acquired data, that APC leads to phase-corrected real-valued DWIs that present a reduced number of alterations and a reduced bias. The substantial absence of parameters for which human input is required favors a straightforward integration of APC in MRI processing pipelines.

This work has been published in [17].

Towards validation of diffusion MRI tractography: bridging the resolution gap with 3D Polarized Light Imaging

Participants : Abib Olushola Yessouffou Alimi, Samuel Deslauriers-Gauthier, Rachid Deriche.

Three-dimensional Polarized Light Imaging (3D-PLI) is an optical approach presented as a good candidate for validation of diffusion Magnetic Resonance Imaging (dMRI) results such as orientation estimates (fiber Orientation Distribution Functions) and tractography. We developed an anlytical approach to reconstruct fiber ODFs from 3D-PLI datasets. From these fODFs, here we compute brain fiber tracts via dMRI-based probabilistic tractography algorithm. Reconstructed fODFs at different scales proves the ability to bridge the resolution gap between 3D-PLI and dMRI, demonstrating, therefore, a great promise to validate diffusion MRI tractography thanks to multi-scale fiber tracking based on 3D-PLI.

This work has been published in [21].

Analytical Fiber ODF Reconstruction in 3D Polarized Light Imaging: Performance Assessment

Participants : Abib Olushola Yessouffou Alimi, Samuel Deslauriers-Gauthier, Felix Matuschke [INM-1 - Institute of Neuroscience and Medicine, Jülich] , Daniel Schmitz [INM-1 - Institute of Neuroscience and Medicine, Jülich] , Markus Axer [INM-1 - Institute of Neuroscience and Medicine, Jülich] , Rachid Deriche.

Three dimensional Polarized Light Imaging (3D-PLI) allows to map the spatial fiber structure of postmortem tissue at a sub-millimeter resolution, thanks to its birefringence property. Different methods have been recently proposed to reconstruct the fiber orientation distribution function (fODF) from high-resolution vector data provided by 3D-PLI. Here, we focus on the analytical fODF computation approach, which uses the spherical harmonics to represent the fODF and analytically computes the spherical harmonics coefficients via the spherical Fourier transform. This work deals with the assessment of the performance of this approach on rich synthetic data which simulates the geometry of the neuronal fibers and on human brain dataset. A computational complexity and robustness to noise analysis demonstrate the interest and great potential of the approach.

This work has been published in [22].