Automatic method for thalamus parcellation using multi-modal feature classification

Joshua V. Stough, Jeffrey Glaister, Chuyang Ye, Sarah H. Ying, Jerry Ladd Prince, Aaron Carass

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Segmentation and parcellation of the thalamus is an important step in providing volumetric assessment of the impact of disease on brain structures. Conventionally, segmentation is carried out on T1-weighted magnetic resonance (MR) images and nuclear parcellation using diffusion weighted MR images. We present the first fully automatic method that incorporates both tissue contrasts and several derived features to first segment and then parcellate the thalamus. We incorporate fractional anisotrophy, fiber orientation from the 5D Knutsson representation of the principal eigenvectors, and connectivity between the thalamus and the cortical lobes, as features. Combining these multiple information sources allows us to identify discriminating dimensions and thus parcellate the thalamic nuclei. A hierarchical random forest framework with a multidimensional feature per voxel, first distinguishes thalamus from background, and then separates each group of thalamic nuclei. Using a leave one out cross-validation on 12 subjects we have a mean Dice score of 0.805 and 0.799 for the left and right thalami, respectively. We also report overlap for the thalamic nuclear groups.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages169-176
Number of pages8
Volume8675 LNCS
EditionPART 3
ISBN (Print)9783319104423
DOIs
StatePublished - 2014
Event17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014 - Boston, MA, United States
Duration: Sep 14 2014Sep 18 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 3
Volume8675 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014
CountryUnited States
CityBoston, MA
Period9/14/149/18/14

Fingerprint

Magnetic Resonance Image
Magnetic resonance
Nucleus
Segmentation
Dice
Fiber Orientation
Random Forest
Voxel
Fiber reinforced materials
Cross-validation
Eigenvalues and eigenfunctions
Eigenvector
Overlap
Brain
Connectivity
Fractional
Tissue
Framework
Background

Keywords

  • Brain imaging
  • diffusion MRI
  • machine learning
  • magnetic resonance imaging
  • segmentation
  • thalamus parcellation

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Stough, J. V., Glaister, J., Ye, C., Ying, S. H., Prince, J. L., & Carass, A. (2014). Automatic method for thalamus parcellation using multi-modal feature classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 3 ed., Vol. 8675 LNCS, pp. 169-176). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8675 LNCS, No. PART 3). Springer Verlag. https://doi.org/10.1007/978-3-319-10443-0_22

Automatic method for thalamus parcellation using multi-modal feature classification. / Stough, Joshua V.; Glaister, Jeffrey; Ye, Chuyang; Ying, Sarah H.; Prince, Jerry Ladd; Carass, Aaron.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8675 LNCS PART 3. ed. Springer Verlag, 2014. p. 169-176 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8675 LNCS, No. PART 3).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Stough, JV, Glaister, J, Ye, C, Ying, SH, Prince, JL & Carass, A 2014, Automatic method for thalamus parcellation using multi-modal feature classification. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 3 edn, vol. 8675 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 3, vol. 8675 LNCS, Springer Verlag, pp. 169-176, 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014, Boston, MA, United States, 9/14/14. https://doi.org/10.1007/978-3-319-10443-0_22
Stough JV, Glaister J, Ye C, Ying SH, Prince JL, Carass A. Automatic method for thalamus parcellation using multi-modal feature classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 3 ed. Vol. 8675 LNCS. Springer Verlag. 2014. p. 169-176. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3). https://doi.org/10.1007/978-3-319-10443-0_22
Stough, Joshua V. ; Glaister, Jeffrey ; Ye, Chuyang ; Ying, Sarah H. ; Prince, Jerry Ladd ; Carass, Aaron. / Automatic method for thalamus parcellation using multi-modal feature classification. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8675 LNCS PART 3. ed. Springer Verlag, 2014. pp. 169-176 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3).
@inproceedings{e4fae8464a574fd1a6e91960169bce13,
title = "Automatic method for thalamus parcellation using multi-modal feature classification",
abstract = "Segmentation and parcellation of the thalamus is an important step in providing volumetric assessment of the impact of disease on brain structures. Conventionally, segmentation is carried out on T1-weighted magnetic resonance (MR) images and nuclear parcellation using diffusion weighted MR images. We present the first fully automatic method that incorporates both tissue contrasts and several derived features to first segment and then parcellate the thalamus. We incorporate fractional anisotrophy, fiber orientation from the 5D Knutsson representation of the principal eigenvectors, and connectivity between the thalamus and the cortical lobes, as features. Combining these multiple information sources allows us to identify discriminating dimensions and thus parcellate the thalamic nuclei. A hierarchical random forest framework with a multidimensional feature per voxel, first distinguishes thalamus from background, and then separates each group of thalamic nuclei. Using a leave one out cross-validation on 12 subjects we have a mean Dice score of 0.805 and 0.799 for the left and right thalami, respectively. We also report overlap for the thalamic nuclear groups.",
keywords = "Brain imaging, diffusion MRI, machine learning, magnetic resonance imaging, segmentation, thalamus parcellation",
author = "Stough, {Joshua V.} and Jeffrey Glaister and Chuyang Ye and Ying, {Sarah H.} and Prince, {Jerry Ladd} and Aaron Carass",
year = "2014",
doi = "10.1007/978-3-319-10443-0_22",
language = "English (US)",
isbn = "9783319104423",
volume = "8675 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
number = "PART 3",
pages = "169--176",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 3",

}

TY - GEN

T1 - Automatic method for thalamus parcellation using multi-modal feature classification

AU - Stough, Joshua V.

AU - Glaister, Jeffrey

AU - Ye, Chuyang

AU - Ying, Sarah H.

AU - Prince, Jerry Ladd

AU - Carass, Aaron

PY - 2014

Y1 - 2014

N2 - Segmentation and parcellation of the thalamus is an important step in providing volumetric assessment of the impact of disease on brain structures. Conventionally, segmentation is carried out on T1-weighted magnetic resonance (MR) images and nuclear parcellation using diffusion weighted MR images. We present the first fully automatic method that incorporates both tissue contrasts and several derived features to first segment and then parcellate the thalamus. We incorporate fractional anisotrophy, fiber orientation from the 5D Knutsson representation of the principal eigenvectors, and connectivity between the thalamus and the cortical lobes, as features. Combining these multiple information sources allows us to identify discriminating dimensions and thus parcellate the thalamic nuclei. A hierarchical random forest framework with a multidimensional feature per voxel, first distinguishes thalamus from background, and then separates each group of thalamic nuclei. Using a leave one out cross-validation on 12 subjects we have a mean Dice score of 0.805 and 0.799 for the left and right thalami, respectively. We also report overlap for the thalamic nuclear groups.

AB - Segmentation and parcellation of the thalamus is an important step in providing volumetric assessment of the impact of disease on brain structures. Conventionally, segmentation is carried out on T1-weighted magnetic resonance (MR) images and nuclear parcellation using diffusion weighted MR images. We present the first fully automatic method that incorporates both tissue contrasts and several derived features to first segment and then parcellate the thalamus. We incorporate fractional anisotrophy, fiber orientation from the 5D Knutsson representation of the principal eigenvectors, and connectivity between the thalamus and the cortical lobes, as features. Combining these multiple information sources allows us to identify discriminating dimensions and thus parcellate the thalamic nuclei. A hierarchical random forest framework with a multidimensional feature per voxel, first distinguishes thalamus from background, and then separates each group of thalamic nuclei. Using a leave one out cross-validation on 12 subjects we have a mean Dice score of 0.805 and 0.799 for the left and right thalami, respectively. We also report overlap for the thalamic nuclear groups.

KW - Brain imaging

KW - diffusion MRI

KW - machine learning

KW - magnetic resonance imaging

KW - segmentation

KW - thalamus parcellation

UR - http://www.scopus.com/inward/record.url?scp=84906972347&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906972347&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-10443-0_22

DO - 10.1007/978-3-319-10443-0_22

M3 - Conference contribution

SN - 9783319104423

VL - 8675 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 169

EP - 176

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -