White Matter Supervoxel Segmentation by Axial DP-Means Clustering

Cabeen, Ryan P.; Laidlaw, David H.

doi:10.1007/978-3-319-14104-6_10

Cited by 4 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The second region-based analysis (denoted SUPER) used automatically defined “supervoxel” ROIs that were computed for each template using a clustering algorithm [53]. The clustering algorithm includes parameters to control the relative contribution the voxel positions (α), fiber orientations (β), and number of clusters (λ) make to the overall optimization.…”

Section: Methodsmentioning

confidence: 99%

A Comparative evaluation of voxel-based spatial mapping in diffusion tensor imaging

2017

Self Cite

View full text Add to dashboard Cite

This paper presents a comparative evaluation of methods for automated voxel-based spatial mapping in diffusion tensor imaging studies. Such methods are an essential step in computational pipelines and provide anatomically comparable measurements across a population in atlas-based studies. To better understand their strengths and weaknesses, we tested a total of eight methods for voxel-based spatial mapping in two types of diffusion tensor templates. The methods were evaluated with respect to scan-rescan reliability and an application to normal aging. The methods included voxel-based analysis with and without smoothing, two types of region-based analysis, and combinations thereof with skeletonization. The templates included a study-specific template created with DTI-TK and the IIT template serving as a standard template. To control for other factors in the pipeline, the experiments used a common dataset, acquired at 1.5T with a single shell high angular resolution diffusion MR imaging protocol, and tensor-based spatial normalization with DTI-TK. Scan-rescan reliability was assessed using the coefficient of variation (CV) and intraclass correlation (ICC) in eight subjects with three scans each. Sensitivity to normal aging was assessed in a population of 80 subjects aged 25 to 65 years old, and methods were compared with respect to the anatomical agreement of significant findings and the R2 of the associated models of fractional anisotropy. The results show that reliability depended greatly on the method used for spatial mapping. The largest differences in reliability were found when adding smoothing and comparing voxel-based and region-based analyses. Skeletonization and template type were found to have either a small or negligible effect on reliability. The aging results showed agreement among the methods in nine brain areas, with some methods showing more sensitivity than others. Skeletonization and smoothing were not major factors affecting sensitivity to aging, but the standard template showed higher R2 in several conditions. A structural comparison of the templates showed that large deformations between them may be related to observed differences in patterns of significant voxels. Most areas showed significantly higher R2 with voxel-based analysis, particularly when clusters were smaller than the available regions-of-interest. Looking forward, these results can potentially help to interpret results from existing white matter imaging studies, as well as provide a resource to help in planning future studies to maximize reliability and sensitivity with regard to the scientific goals at hand.

show abstract

Section: Methodsmentioning

confidence: 99%

A Comparative evaluation of voxel-based spatial mapping in diffusion tensor imaging

2017

Self Cite

View full text Add to dashboard Cite

show abstract

“…10. This is equivalent to solving the weighted axial DP-means clustering problem [27], which is similar to the k-means algorithm with two extensions. First, clustering is performed with axial variables [28], which are equivalent to fiber orientations.…”

Section: Methodsmentioning

confidence: 99%

“…12. This is theoretically well-grounded because the Watson distribution is an exponential family, and d f is the associated Bregman divergence [27]. The work of Jiang et al also demonstrates that the general form of Algo.…”

mentioning

confidence: 99%

“…This can be solved in closed form by finding the principal eigenvector of μ [66]. This formulation also has a statistical interpretation, as

d_{f}^{2}

is equivalent to the Bregman divergence between Watson distributions of equal dispersion [68] [27], and the weighted average is equivalent to the maximum likelihood estimate of the direction of a Watson distribution [69]. These results support prior work suggesting a relationship between extrinsic means and maximum likelihood estimates for data in ℝℙ n [70].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Kernel regression estimation of fiber orientation mixtures in diffusion MRI

2016

Self Cite

View full text Add to dashboard Cite

We present and evaluate a method for kernel regression estimation of fiber orientations and associated volume fractions for diffusion MR tractography and population-based atlas construction in clinical imaging studies of brain white matter. This is a model-based image processing technique in which representative fiber models are estimated from collections of component fiber models in model-valued image data. This extends prior work in nonparametric image processing and multi-compartment processing to provide computational tools for image interpolation, smoothing, and fusion with fiber orientation mixtures. In contrast to related work on multi-compartment processing, this approach is based on directional measures of divergence and includes data-adaptive extensions for model selection and bilateral filtering. This is useful for reconstructing complex anatomical features in clinical datasets analyzed with the ball-and-sticks model, and our framework’s data-adaptive extensions are potentially useful for general multi-compartment image processing. We experimentally evaluate our approach with both synthetic data from computational phantoms and in vivo clinical data from human subjects. With synthetic data experiments, we evaluate performance based on errors in fiber orientation, volume fraction, compartment count, and tractography-based connectivity. With in vivo data experiments, we first show improved scan-rescan reproducibility and reliability of quantitative fiber bundle metrics, including mean length, volume, streamline count, and mean volume fraction. We then demonstrate the creation of a multi-fiber tractography atlas from a population of 80 human subjects. In comparison to single tensor atlasing, our multi-fiber atlas shows more complete features of known fiber bundles and includes reconstructions of the lateral projections of the corpus callosum and complex fronto-parietal connections of the superior longitudinal fasciculus I, II, and III.

show abstract

Bilateral Filtering of Multiple Fiber Orientations in Diffusion MRI

Cabeen

Laidlaw

2014

Computational Diffusion MRI

View full text Add to dashboard Cite

White Matter Supervoxel Segmentation by Axial DP-Means Clustering

Cited by 4 publications

References 16 publications

A Comparative evaluation of voxel-based spatial mapping in diffusion tensor imaging

A Comparative evaluation of voxel-based spatial mapping in diffusion tensor imaging

Kernel regression estimation of fiber orientation mixtures in diffusion MRI

Bilateral Filtering of Multiple Fiber Orientations in Diffusion MRI

Contact Info

Product

Resources

About