Equivariant Spherical Deconvolution: Learning Sparse Orientation Distribution Functions from Spherical Data

We present a rotation-equivariant unsupervised learning framework for the sparse deconvolution of non-negative scalar fields defined on the unit sphere. Spherical signals with multiple peaks naturally arise in Diffusion MRI (dMRI), where each voxel consists of one or more signal sources corresponding to anisotropic tissue structure such as white matter. Due to spatial and spectral partial voluming, clinically-feasible dMRI struggles to resolve crossing-fiber white matter configurations, leading to extensive development in spherical deconvolution methodology to recover underlying fiber directions. However, these methods are typically linear and struggle with small crossing-angles and partial volume fraction estimation. In this work, we improve on current methodologies by nonlinearly estimating fiber structures via unsupervised spherical convolutional networks with guaranteed equivariance to spherical rotation. Experimentally, we first validate our proposition via extensive single and multi-shell synthetic benchmarks demonstrating competitive performance against common baselines. We then show improved downstream performance on fiber tractography measures on the Tractometer benchmark dataset. Finally, we show downstream improvements in terms of tractography and partial volume estimation on a multi-shell dataset of human subjects.

[1]  Colin B. Hansen,et al.  Deep learning reveals untapped information for local white-matter fiber reconstruction in diffusion-weighted MRI. , 2019, Magnetic resonance imaging.

[2]  Jan Sijbers,et al.  Multi-tissue constrained spherical deconvolution for improved analysis of multi-shell diffusion MRI data , 2014, NeuroImage.

[3]  Gabriel Girard,et al.  Sparse wars: A survey and comparative study of spherical deconvolution algorithms for diffusion MRI , 2019, NeuroImage.

[4]  Max Welling,et al.  Group Equivariant Convolutional Networks , 2016, ICML.

[5]  Alan Connelly,et al.  MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation , 2019, NeuroImage.

[6]  Simon K. Warfield,et al.  A machine learning-based method for estimating the number and orientations of major fascicles in diffusion-weighted magnetic resonance imaging , 2020, Medical Image Anal..

[7]  Thomas Schultz,et al.  Better Fiber ODFs from Suboptimal Data with Autoencoder Based Regularization , 2018, MICCAI.

[8]  Nathanael Perraudin,et al.  DeepSphere: Efficient spherical Convolutional Neural Network with HEALPix sampling for cosmological applications , 2018, Astron. Comput..

[9]  Alan Connelly,et al.  Robust determination of the fibre orientation distribution in diffusion MRI: Non-negativity constrained super-resolved spherical deconvolution , 2007, NeuroImage.

[10]  Peter F. Neher,et al.  The challenge of mapping the human connectome based on diffusion tractography , 2017, Nature Communications.

[11]  R. Deriche,et al.  Diffusion MRI Fiber Orientation Distribution Function Estimation Using Voxel-Wise Spherical U-Net , 2021, Computational Diffusion MRI.

[12]  Hao Yan,et al.  Estimating fiber orientation distribution from diffusion MRI with spherical needlets , 2016, Medical Image Anal..

[14]  Max Welling,et al.  Spherical CNNs , 2018, ICLR.

[15]  Rachid Deriche,et al.  Quantitative Comparison of Reconstruction Methods for Intra-Voxel Fiber Recovery From Diffusion MRI , 2014, IEEE Transactions on Medical Imaging.

[16]  Stephen M. Smith,et al.  Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm , 2001, IEEE Transactions on Medical Imaging.

[17]  Krzysztof M. Gorski,et al.  The HEALPix Primer , 1999, astro-ph/9905275.

[18]  Sebastien Ourselin,et al.  Using convolution neural networks to learn enhanced fiber orientation distribution models from commercially available diffusion magnetic resonance imaging , 2020, ArXiv.

[19]  Chun-Hung Yeh,et al.  MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation , 2019, NeuroImage.

[20]  V. Wedeen,et al.  Mapping fiber orientation spectra in cerebral white matter with Fourier-transform diffusion MRI , 2000 .

[21]  Peipeng Liang,et al.  Multicenter dataset of multi-shell diffusion MRI in healthy traveling adults with identical settings , 2020, Scientific Data.

[22]  Zhiwei Li,et al.  Fast learning of fiber orientation distribution function for MR tractography using convolutional neural network. , 2019, Medical physics.