Applying tensor-based morphometry to parametric surfaces can improve MRI-based disease diagnosis

Yalin Wang, Lei Yuan, Jie Shi, Alexander Greve, Jieping Ye, Arthur W. Toga, Allan L. Reiss, Paul M. Thompson

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

Many methods have been proposed for computer-assisted diagnostic classification. Full tensor information and machine learning with 3D maps derived from brain images may help detect subtle differences or classify subjects into different groups. Here we develop a new approach to apply tensor-based morphometry to parametric surface models for diagnostic classification. We use this approach to identify cortical surface features for use in diagnostic classifiers. First, with holomorphic 1-forms, we compute an efficient and accurate conformal mapping from a multiply connected mesh to the so-called slit domain. Next, the surface parameterization approach provides a natural way to register anatomical surfaces across subjects using a constrained harmonic map. To analyze anatomical differences, we then analyze the full Riemannian surface metric tensors, which retain multivariate information on local surface geometry. As the number of voxels in a 3D image is large, sparse learning is a promising method to select a subset of imaging features and to improve classification accuracy. Focusing on vertices with greatest effect sizes, we train a diagnostic classifier using the surface features selected by an L1-norm based sparse learning method. Stability selection is applied to validate the selected feature sets. We tested the algorithm on MRI-derived cortical surfaces from 42 subjects with genetically confirmed Williams syndrome and 40 age-matched controls, multivariate statistics on the local tensors gave greater effect sizes for detecting group differences relative to other TBM-based statistics including analysis of the Jacobian determinant and the largest eigenvalue of the surface metric. Our method also gave reasonable classification results relative to the Jacobian determinant, the pair of eigenvalues of the Jacobian matrix and volume features. This analysis pipeline may boost the power of morphometry studies, and may assist with image-based classification.

Original languageEnglish (US)
Pages (from-to)209-230
Number of pages22
JournalNeuroImage
Volume74
DOIs
StatePublished - Jul 1 2013

Fingerprint

Learning
Williams Syndrome
Brain
Power (Psychology)
Machine Learning

Keywords

  • Sparse learning
  • Stability selection
  • Surface conformal parameterization
  • Surface registration
  • Tensor-based morphometry

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

Applying tensor-based morphometry to parametric surfaces can improve MRI-based disease diagnosis. / Wang, Yalin; Yuan, Lei; Shi, Jie; Greve, Alexander; Ye, Jieping; Toga, Arthur W.; Reiss, Allan L.; Thompson, Paul M.

In: NeuroImage, Vol. 74, 01.07.2013, p. 209-230.

Research output: Contribution to journalArticle

Wang, Yalin ; Yuan, Lei ; Shi, Jie ; Greve, Alexander ; Ye, Jieping ; Toga, Arthur W. ; Reiss, Allan L. ; Thompson, Paul M. / Applying tensor-based morphometry to parametric surfaces can improve MRI-based disease diagnosis. In: NeuroImage. 2013 ; Vol. 74. pp. 209-230.
@article{f7d6655c68f844c9924f9a34ff1b470c,
title = "Applying tensor-based morphometry to parametric surfaces can improve MRI-based disease diagnosis",
abstract = "Many methods have been proposed for computer-assisted diagnostic classification. Full tensor information and machine learning with 3D maps derived from brain images may help detect subtle differences or classify subjects into different groups. Here we develop a new approach to apply tensor-based morphometry to parametric surface models for diagnostic classification. We use this approach to identify cortical surface features for use in diagnostic classifiers. First, with holomorphic 1-forms, we compute an efficient and accurate conformal mapping from a multiply connected mesh to the so-called slit domain. Next, the surface parameterization approach provides a natural way to register anatomical surfaces across subjects using a constrained harmonic map. To analyze anatomical differences, we then analyze the full Riemannian surface metric tensors, which retain multivariate information on local surface geometry. As the number of voxels in a 3D image is large, sparse learning is a promising method to select a subset of imaging features and to improve classification accuracy. Focusing on vertices with greatest effect sizes, we train a diagnostic classifier using the surface features selected by an L1-norm based sparse learning method. Stability selection is applied to validate the selected feature sets. We tested the algorithm on MRI-derived cortical surfaces from 42 subjects with genetically confirmed Williams syndrome and 40 age-matched controls, multivariate statistics on the local tensors gave greater effect sizes for detecting group differences relative to other TBM-based statistics including analysis of the Jacobian determinant and the largest eigenvalue of the surface metric. Our method also gave reasonable classification results relative to the Jacobian determinant, the pair of eigenvalues of the Jacobian matrix and volume features. This analysis pipeline may boost the power of morphometry studies, and may assist with image-based classification.",
keywords = "Sparse learning, Stability selection, Surface conformal parameterization, Surface registration, Tensor-based morphometry",
author = "Yalin Wang and Lei Yuan and Jie Shi and Alexander Greve and Jieping Ye and Toga, {Arthur W.} and Reiss, {Allan L.} and Thompson, {Paul M.}",
year = "2013",
month = "7",
day = "1",
doi = "10.1016/j.neuroimage.2013.02.011",
language = "English (US)",
volume = "74",
pages = "209--230",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Applying tensor-based morphometry to parametric surfaces can improve MRI-based disease diagnosis

AU - Wang, Yalin

AU - Yuan, Lei

AU - Shi, Jie

AU - Greve, Alexander

AU - Ye, Jieping

AU - Toga, Arthur W.

AU - Reiss, Allan L.

AU - Thompson, Paul M.

PY - 2013/7/1

Y1 - 2013/7/1

N2 - Many methods have been proposed for computer-assisted diagnostic classification. Full tensor information and machine learning with 3D maps derived from brain images may help detect subtle differences or classify subjects into different groups. Here we develop a new approach to apply tensor-based morphometry to parametric surface models for diagnostic classification. We use this approach to identify cortical surface features for use in diagnostic classifiers. First, with holomorphic 1-forms, we compute an efficient and accurate conformal mapping from a multiply connected mesh to the so-called slit domain. Next, the surface parameterization approach provides a natural way to register anatomical surfaces across subjects using a constrained harmonic map. To analyze anatomical differences, we then analyze the full Riemannian surface metric tensors, which retain multivariate information on local surface geometry. As the number of voxels in a 3D image is large, sparse learning is a promising method to select a subset of imaging features and to improve classification accuracy. Focusing on vertices with greatest effect sizes, we train a diagnostic classifier using the surface features selected by an L1-norm based sparse learning method. Stability selection is applied to validate the selected feature sets. We tested the algorithm on MRI-derived cortical surfaces from 42 subjects with genetically confirmed Williams syndrome and 40 age-matched controls, multivariate statistics on the local tensors gave greater effect sizes for detecting group differences relative to other TBM-based statistics including analysis of the Jacobian determinant and the largest eigenvalue of the surface metric. Our method also gave reasonable classification results relative to the Jacobian determinant, the pair of eigenvalues of the Jacobian matrix and volume features. This analysis pipeline may boost the power of morphometry studies, and may assist with image-based classification.

AB - Many methods have been proposed for computer-assisted diagnostic classification. Full tensor information and machine learning with 3D maps derived from brain images may help detect subtle differences or classify subjects into different groups. Here we develop a new approach to apply tensor-based morphometry to parametric surface models for diagnostic classification. We use this approach to identify cortical surface features for use in diagnostic classifiers. First, with holomorphic 1-forms, we compute an efficient and accurate conformal mapping from a multiply connected mesh to the so-called slit domain. Next, the surface parameterization approach provides a natural way to register anatomical surfaces across subjects using a constrained harmonic map. To analyze anatomical differences, we then analyze the full Riemannian surface metric tensors, which retain multivariate information on local surface geometry. As the number of voxels in a 3D image is large, sparse learning is a promising method to select a subset of imaging features and to improve classification accuracy. Focusing on vertices with greatest effect sizes, we train a diagnostic classifier using the surface features selected by an L1-norm based sparse learning method. Stability selection is applied to validate the selected feature sets. We tested the algorithm on MRI-derived cortical surfaces from 42 subjects with genetically confirmed Williams syndrome and 40 age-matched controls, multivariate statistics on the local tensors gave greater effect sizes for detecting group differences relative to other TBM-based statistics including analysis of the Jacobian determinant and the largest eigenvalue of the surface metric. Our method also gave reasonable classification results relative to the Jacobian determinant, the pair of eigenvalues of the Jacobian matrix and volume features. This analysis pipeline may boost the power of morphometry studies, and may assist with image-based classification.

KW - Sparse learning

KW - Stability selection

KW - Surface conformal parameterization

KW - Surface registration

KW - Tensor-based morphometry

UR - http://www.scopus.com/inward/record.url?scp=84874985641&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84874985641&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2013.02.011

DO - 10.1016/j.neuroimage.2013.02.011

M3 - Article

VL - 74

SP - 209

EP - 230

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -