Reymbaut, A. | bioRxiv Science

bioRxiv2021

Enabling constrained spherical deconvolution and diffusional variance decomposition with tensor-valued diffusion MRI

Diffusion tensor imaging (DTI) is widely used to extract valuable tissue measurements and white matter (WM) fiber orientations, even though its lack of specificity is now well-known, especially for WM fiber crossings. Models such as constrained spherical deconvolution (CSD) take advantage of high angular resolution diffusion imaging (HARDI) data to compute fiber orientation distribution functions (fODF) and tackle the orientational part of the DTI limitations. Furthermore, the recent introduction of tensor-valued diffusion MRI allows for diffusional variance decomposition (DIVIDE), opening the door to the computation of measures more specific to microstructure than DTI measures, such as microscopic fractional anisotropy (FA). However, tensor-valued diffusion MRI data is not mathematically compatible with latest versions of CSD and the impacts of such atypical data on fODF reconstruction with CSD are yet to be studied. In this work, we lay down the mathematical and computational foundations of a tensor-valued CSD and use simulated data to explore the effects of various combinations of diffusion encodings on the angular resolution of extracted fOFDs. We also compare the combinations with regards to their performance at producing accurate and precise FA with DIVIDE, and present an optimised protocol for both methods. We show that our proposed protocol enables the reconstruction of both fODFs and FA on in vivo data.

neuroscience↗

bioRxiv2020

Visualizing orientation-specific relaxation-diffusion features mapped onto orientation distribution functions estimated via nonparametric Monte Carlo MRI signal inversion

Diffusion MRI techniques are widely used to study in vivo changes in the human brain connectome. However, to resolve and characterise white matter fibres in heterogeneous diffusion MRI voxels remains a challenging problem typically approached with signal models that rely on prior information and restrictive constraints. We have recently introduced a 5D relaxation-diffusion correlation framework wherein multidimensional diffusion encoding strategies are used to acquire data at multiple echo-times in order to increase the amount of information encoded into the signal and ease the constraints needed for signal inversion. Nonparametric Monte Carlo inversion of the resulting datasets yields 5D relaxation-diffusion distributions where contributions from different sub-voxel tissue environments are separated with minimal assumptions on their microscopic properties. Here, we build on the 5D correlation approach to derive fibre-specific metrics that can be mapped throughout the imaged brain volume. Distribution components ascribed to fibrous tissues are resolved, and subsequently mapped to a dense mesh of overlapping orientation bins in order to define a smooth orientation distribution function (ODF). Moreover, relaxation and diffusion measures are correlated to each independent ODF coordinate, thereby allowing the estimation of orientation-specific relaxation rates and diffusivities. The proposed method is tested on a healthy volunteer, where the estimated ODFs were observed to capture major WM tracts, resolve fibre crossings, and, more importantly, inform on the relaxation and diffusion features along distinct fibre bundles. If combined with fibre-tracking algorithms, the methodology presented in this work may be useful for investigating the microstructural properties along individual white matter pathways.

neuroscience↗