Symmetric diffeomorphic registration of fibre orientation distributions

Raffelt, David; Tournier, Jacques‐Donald; Fripp, Jürgen; Crozier, Stuart; Connelly, Alan; Salvado, Olivier

doi:10.1016/j.neuroimage.2011.02.014

Cited by 246 publications

(238 citation statements)

References 59 publications

Supporting

Mentioning

237

Contrasting

Order By: Relevance

“…For example, the PSF reorientation method presented here was recently used in a sister publication which demonstrated the benefit of higher order information for FOD image registration (23). PSF reorientation was also recently used to permit group average FOD fiber tractography in a novel approach to white matter tensor-based morphometry (34).…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Reorientation of fiber orientation distributions using apodized point spread functions

Raffelt¹,

Tournier²,

Crozier³

et al. 2011

Magnetic Resonance in Med

Self Cite

109

105

View full text Add to dashboard Cite

Using high angular resolution diffusion-weighted images, spherical deconvolution enables multiple white matter fiber populations to be resolved within a single voxel by computing the fiber orientation distribution (FOD). Higher order information provided by FODs could improve several methods for investigating population differences in white matter, including image registration, voxel-based analysis, atlas-based segmentation and labeling, and group average fiber tractography. All of these methods require spatial normalization of FODs. In this article, a novel method to reorient the FOD is presented, which is an important step required for FOD spatial normalization. The proposed method was assessed using both qualitative and quantitative experiments, with numerical simulations and in vivo human data. Results demonstrate that the proposed method improves FOD reorientation accuracy, removes undesired artefacts, and decreases computation time compared to a previous approach. The utility of the proposed method is illustrated by nonlinear FOD spatial normalization of 10 human subjects. Key words: fiber orientation distribution; reorientation; normalization; diffusion MRI Diffusion-weighted imaging provides a unique method to investigate the architecture of white matter in vivo. To investigate group differences in the diffusion properties of white matter, correspondence between images is required to ensure that the same anatomical structures are being compared. The most common method for obtaining correspondence involves transforming data from each individual subject into a common co-ordinate or template space (a process called spatial normalization). For the remainder of this text, we refer to spatial normalization as simply normalization.The diffusion of water within each imaging voxel is commonly modeled using a diffusion tensor (1). In recent years, normalization of diffusion tensor measures such as fractional anisotropy (FA) (2) has been widely used to investigate voxelwise differences between diseased and healthy populations (see Assaf and Pasternak (3) for a recent review). However, the popular rank-2 diffusion tensor is only capable of modeling a single fiber population within each voxel. This is a serious limitation given recent work suggesting that 90% of white matter voxels contain multiple fiber populations (4). Consequently, in the majority of white matter, population differences in diffusion tensor measures can be influenced not only by differences in so-called white matter "integrity", but also by partial volume effects within crossing fiber regions (5,6).Using high angular resolution diffusion-weighted imaging, higher order models enable crossing fibers to be resolved and, therefore, provide more information than the diffusion tensor at a subvoxel level. Spherical deconvolution is one method that can resolve crossing fibers by estimating the so-called fiber orientation distribution (FOD), a continuous distribution representing the partial volume of the underlying fibers as a function of orientation (7-12...

show abstract

Section: Discussionmentioning

confidence: 99%

“…One application of FOD reorientation is FOD image registration (22,23). Reorientation of FODs should be performed at each iteration of the registration process and, therefore, computational cost is of importance.…”

Section: Experiments 4: Comparison Of Fod Reorientation Computational mentioning

confidence: 99%

Reorientation of fiber orientation distributions using apodized point spread functions

Raffelt¹,

Tournier²,

Crozier³

et al. 2011

Magnetic Resonance in Med

Self Cite

109

105

View full text Add to dashboard Cite

show abstract

“…In the last decades, researchers have spent great efforts on developing registration algorithms to align diffusion tensors derived from DTI and orientation distribution functions (ODFs) derived from HARDI [e.g., [1,14,8]]. However, registration algorithms directly based on DWIs are few.…”

Section: Introductionmentioning

confidence: 99%

Diffeomorphic Metric Mapping of Hybrid Diffusion Imaging Based on BFOR Signal Basis

Hosseinbor

Chung

et al. 2013

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. In this paper, we propose a large deformation diffeomorphic metric mapping algorithm to align multiple b-value diffusion weighted imaging (mDWI) data, specifically acquired via hybrid diffusion imaging (HYDI), denoted as LDDMM-HYDI. We adopt the work given in Hosseinbor et al. (2012) and represent the q-space diffusion signal with the Bessel Fourier orientation reconstruction (BFOR) signal basis. The BFOR framework provides the representation of mDWI in the q-space and thus reduces memory requirement. In addition, since the BFOR signal basis is orthonormal, the L 2 norm that quantifies the differences in q-space signals of any two mDWI datasets can be easily computed as the sum of the squared differences in the BFOR expansion coefficients. In this work, we show that the reorientation of the q-space signal due to spatial transformation can be easily defined on the BFOR signal basis. We incorporate the BFOR signal basis into the LDDMM framework and derive the gradient descent algorithm for LDDMM-HYDI with explicit orientation optimization. Using real HYDI datasets, we show that it is important to consider the variation of mDWI reorientation due to a small change in diffeomorphic transformation in the LDDMM-HYDI optimization.

show abstract

“…Tools for normalizing human brain HARDI datasets are essential to investigate brain development and the onset and progression of neurological disorders by allowing one to quantify intra-and inter-subject differences in longitudinal and population studies. Recent works addressing the problem of normalization include but not limited to [3,4,5,6,7]. The method in [3] uses the spherical harmonic (SH) representation of ODFs for extracting spatial features that are hierarchically updated by increasing the harmonic degree.…”

Section: Introductionmentioning

confidence: 99%

“…[4] performs registration of ODF images by using the SH band energies as rotationally invariant features in a multi-channel diffeomorphic demons algorithm. [5,6] use the SH coefficients of the ODFs as features to find the diffeomorphism between the source and target images. [7] incorporates the Riemannian metric of ODFs for quantifying the similarity between the two images into a variational problem defined under the large deformation diffeomorphic metric mapping (LDDMM) framework.…”

Section: Introductionmentioning

confidence: 99%

An algebraic solution to rotation recovery in HARDI from correspondences of orientation distribution functions

Çetingül

Afsari

Vidal

2012

2012 9th IEEE International Symposium on Biomedical Imaging (ISBI)

View full text Add to dashboard Cite

We consider the problem of aligning high angular resolution diffusion images characterized by orientation distribution functions (ODFs). We cast this problem as an optimization problem where we seek the rotation that aligns the source and target ODFs. This rotation induces a linear transformation of the spherical harmonic coefficients of the ODFs, which can be parametrized by the rotation Euler angles. We propose an algebraic approach to estimate this transformation from a number of ODF correspondences. We evaluate the proposed method on synthetic ODFs as well as on a diffusion MR phantom dataset.

show abstract

Symmetric diffeomorphic registration of fibre orientation distributions

Cited by 246 publications

References 59 publications

Reorientation of fiber orientation distributions using apodized point spread functions

Reorientation of fiber orientation distributions using apodized point spread functions

Diffeomorphic Metric Mapping of Hybrid Diffusion Imaging Based on BFOR Signal Basis

An algebraic solution to rotation recovery in HARDI from correspondences of orientation distribution functions

Contact Info

Product

Resources

About