Estimation of white matter fiber parameters from compressed multiresolution diffusion MRI using sparse Bayesian learning

Pramod K Pisharady, Stamatios N. Sotiropoulos, Julio M. Duarte-Carvajalino, Guillermo Sapiro, Christophe Lenglet

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

We present a sparse Bayesian unmixing algorithm BusineX: Bayesian Unmixing for Sparse Inference-based Estimation of Fiber Crossings (X), for estimation of white matter fiber parameters from compressed (under-sampled) diffusion MRI (dMRI) data. BusineX combines compressive sensing with linear unmixing and introduces sparsity to the previously proposed multiresolution data fusion algorithm RubiX, resulting in a method for improved reconstruction, especially from data with lower number of diffusion gradients. We formulate the estimation of fiber parameters as a sparse signal recovery problem and propose a linear unmixing framework with sparse Bayesian learning for the recovery of sparse signals, the fiber orientations and volume fractions. The data is modeled using a parametric spherical deconvolution approach and represented using a dictionary created with the exponential decay components along different possible diffusion directions. Volume fractions of fibers along these directions define the dictionary weights. The proposed sparse inference, which is based on the dictionary representation, considers the sparsity of fiber populations and exploits the spatial redundancy in data representation, thereby facilitating inference from under-sampled q-space. The algorithm improves parameter estimation from dMRI through data-dependent local learning of hyperparameters, at each voxel and for each possible fiber orientation, that moderate the strength of priors governing the parameter variances. Experimental results on synthetic and in-vivo data show improved accuracy with a lower uncertainty in fiber parameter estimates. BusineX resolves a higher number of second and third fiber crossings. For under-sampled data, the algorithm is also shown to produce more reliable estimates.

Original languageEnglish (US)
Pages (from-to)488-503
Number of pages16
JournalNeuroImage
Volume167
DOIs
StatePublished - Feb 15 2018

Fingerprint

Diffusion Magnetic Resonance Imaging
Learning
Uncertainty
Weights and Measures
White Matter
Population
Direction compound

Keywords

  • Compressive sensing
  • Diffusion MRI
  • Fiber orientation
  • Linear unmixing
  • Sparse Bayesian learning
  • Sparse signal recovery

Cite this

Estimation of white matter fiber parameters from compressed multiresolution diffusion MRI using sparse Bayesian learning. / Pisharady, Pramod K; Sotiropoulos, Stamatios N.; Duarte-Carvajalino, Julio M.; Sapiro, Guillermo; Lenglet, Christophe.

In: NeuroImage, Vol. 167, 15.02.2018, p. 488-503.

Research output: Contribution to journalArticle

Pisharady, Pramod K ; Sotiropoulos, Stamatios N. ; Duarte-Carvajalino, Julio M. ; Sapiro, Guillermo ; Lenglet, Christophe. / Estimation of white matter fiber parameters from compressed multiresolution diffusion MRI using sparse Bayesian learning. In: NeuroImage. 2018 ; Vol. 167. pp. 488-503.
@article{fec3ceb187804ddaa25fa9bc3926cc01,
title = "Estimation of white matter fiber parameters from compressed multiresolution diffusion MRI using sparse Bayesian learning",
abstract = "We present a sparse Bayesian unmixing algorithm BusineX: Bayesian Unmixing for Sparse Inference-based Estimation of Fiber Crossings (X), for estimation of white matter fiber parameters from compressed (under-sampled) diffusion MRI (dMRI) data. BusineX combines compressive sensing with linear unmixing and introduces sparsity to the previously proposed multiresolution data fusion algorithm RubiX, resulting in a method for improved reconstruction, especially from data with lower number of diffusion gradients. We formulate the estimation of fiber parameters as a sparse signal recovery problem and propose a linear unmixing framework with sparse Bayesian learning for the recovery of sparse signals, the fiber orientations and volume fractions. The data is modeled using a parametric spherical deconvolution approach and represented using a dictionary created with the exponential decay components along different possible diffusion directions. Volume fractions of fibers along these directions define the dictionary weights. The proposed sparse inference, which is based on the dictionary representation, considers the sparsity of fiber populations and exploits the spatial redundancy in data representation, thereby facilitating inference from under-sampled q-space. The algorithm improves parameter estimation from dMRI through data-dependent local learning of hyperparameters, at each voxel and for each possible fiber orientation, that moderate the strength of priors governing the parameter variances. Experimental results on synthetic and in-vivo data show improved accuracy with a lower uncertainty in fiber parameter estimates. BusineX resolves a higher number of second and third fiber crossings. For under-sampled data, the algorithm is also shown to produce more reliable estimates.",
keywords = "Compressive sensing, Diffusion MRI, Fiber orientation, Linear unmixing, Sparse Bayesian learning, Sparse signal recovery",
author = "Pisharady, {Pramod K} and Sotiropoulos, {Stamatios N.} and Duarte-Carvajalino, {Julio M.} and Guillermo Sapiro and Christophe Lenglet",
year = "2018",
month = "2",
day = "15",
doi = "10.1016/j.neuroimage.2017.06.052",
language = "English (US)",
volume = "167",
pages = "488--503",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Estimation of white matter fiber parameters from compressed multiresolution diffusion MRI using sparse Bayesian learning

AU - Pisharady, Pramod K

AU - Sotiropoulos, Stamatios N.

AU - Duarte-Carvajalino, Julio M.

AU - Sapiro, Guillermo

AU - Lenglet, Christophe

PY - 2018/2/15

Y1 - 2018/2/15

N2 - We present a sparse Bayesian unmixing algorithm BusineX: Bayesian Unmixing for Sparse Inference-based Estimation of Fiber Crossings (X), for estimation of white matter fiber parameters from compressed (under-sampled) diffusion MRI (dMRI) data. BusineX combines compressive sensing with linear unmixing and introduces sparsity to the previously proposed multiresolution data fusion algorithm RubiX, resulting in a method for improved reconstruction, especially from data with lower number of diffusion gradients. We formulate the estimation of fiber parameters as a sparse signal recovery problem and propose a linear unmixing framework with sparse Bayesian learning for the recovery of sparse signals, the fiber orientations and volume fractions. The data is modeled using a parametric spherical deconvolution approach and represented using a dictionary created with the exponential decay components along different possible diffusion directions. Volume fractions of fibers along these directions define the dictionary weights. The proposed sparse inference, which is based on the dictionary representation, considers the sparsity of fiber populations and exploits the spatial redundancy in data representation, thereby facilitating inference from under-sampled q-space. The algorithm improves parameter estimation from dMRI through data-dependent local learning of hyperparameters, at each voxel and for each possible fiber orientation, that moderate the strength of priors governing the parameter variances. Experimental results on synthetic and in-vivo data show improved accuracy with a lower uncertainty in fiber parameter estimates. BusineX resolves a higher number of second and third fiber crossings. For under-sampled data, the algorithm is also shown to produce more reliable estimates.

AB - We present a sparse Bayesian unmixing algorithm BusineX: Bayesian Unmixing for Sparse Inference-based Estimation of Fiber Crossings (X), for estimation of white matter fiber parameters from compressed (under-sampled) diffusion MRI (dMRI) data. BusineX combines compressive sensing with linear unmixing and introduces sparsity to the previously proposed multiresolution data fusion algorithm RubiX, resulting in a method for improved reconstruction, especially from data with lower number of diffusion gradients. We formulate the estimation of fiber parameters as a sparse signal recovery problem and propose a linear unmixing framework with sparse Bayesian learning for the recovery of sparse signals, the fiber orientations and volume fractions. The data is modeled using a parametric spherical deconvolution approach and represented using a dictionary created with the exponential decay components along different possible diffusion directions. Volume fractions of fibers along these directions define the dictionary weights. The proposed sparse inference, which is based on the dictionary representation, considers the sparsity of fiber populations and exploits the spatial redundancy in data representation, thereby facilitating inference from under-sampled q-space. The algorithm improves parameter estimation from dMRI through data-dependent local learning of hyperparameters, at each voxel and for each possible fiber orientation, that moderate the strength of priors governing the parameter variances. Experimental results on synthetic and in-vivo data show improved accuracy with a lower uncertainty in fiber parameter estimates. BusineX resolves a higher number of second and third fiber crossings. For under-sampled data, the algorithm is also shown to produce more reliable estimates.

KW - Compressive sensing

KW - Diffusion MRI

KW - Fiber orientation

KW - Linear unmixing

KW - Sparse Bayesian learning

KW - Sparse signal recovery

UR - http://www.scopus.com/inward/record.url?scp=85024500662&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024500662&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2017.06.052

DO - 10.1016/j.neuroimage.2017.06.052

M3 - Article

VL - 167

SP - 488

EP - 503

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -