Nonlinear dimensionality reduction combining MR imaging with non-imaging information

Robin Wolz (Corresponding Author), Paul Aljabar, Joseph V. Hajnal, Jyrki Lötjönen, Daniel Rueckert, The Alzheimer’s Disease Neuroimaging Initiative

Research output: Contribution to journalArticleScientificpeer-review

33 Citations (Scopus)

Abstract

We propose a framework for the extraction of biomarkers from low-dimensional manifolds representing inter-subject brain variation. Manifold coordinates of each image capture information about structural shape and appearance and, when a phenotype exists, about the subject’s clinical state. Our framework incorporates subject meta-information into the manifold learning step. Apart from gender and age, information such as genotype or a derived biomarker is often available in clinical studies and can inform the classification of a query subject. Such information, whether discrete or continuous, is used as an additional input to manifold learning, extending the Laplacian Eigenmap objective function and enriching a similarity measure derived from pairwise image similarities. The biomarkers identified with the proposed method are
data-driven in contrast to a priori defined biomarkers derived from, e.g., manual or automated segmentations. They form a unified representation of both the imaging and non-imaging measurements, providing a natural use for data analysis and visualization. We test the method to classify subjects with Alzheimer’s Disease (AD), mild cognitive impairment (MCI) and healthy controls enrolled in the ADNI study. Non-imaging metadata used are ApoE genotype, a risk factor associated with AD, and the CSF-concentration of Αβ1-42, an established biomarker for AD. In addition, we use hippocampal volume as a derived imaging-biomarker to enrich the learned manifold. Our classification results compare favorably to what has been reported in a recent meta-analysis using established neuroimaging methods on the same database.
Original languageEnglish
Pages (from-to)819-830
JournalMedical Image Analysis
Volume16
Issue number4
DOIs
Publication statusPublished - 2012
MoE publication typeA1 Journal article-refereed

Fingerprint

Biomarkers
Imaging techniques
Alzheimer Disease
Genotype
Learning
Neuroimaging
Macrophage Colony-Stimulating Factor
Data visualization
Apolipoproteins E
Metadata
Meta-Analysis
Brain
Databases
Phenotype

Keywords

  • Manifold learning
  • Laplacian Eigenmaps
  • Classification
  • Metadata
  • Alzheimer's disease

Cite this

Wolz, R., Aljabar, P., Hajnal, J. V., Lötjönen, J., Rueckert, D., & The Alzheimer’s Disease Neuroimaging Initiative (2012). Nonlinear dimensionality reduction combining MR imaging with non-imaging information. Medical Image Analysis, 16(4), 819-830. https://doi.org/10.1016/j.media.2011.12.003
Wolz, Robin ; Aljabar, Paul ; Hajnal, Joseph V. ; Lötjönen, Jyrki ; Rueckert, Daniel ; The Alzheimer’s Disease Neuroimaging Initiative. / Nonlinear dimensionality reduction combining MR imaging with non-imaging information. In: Medical Image Analysis. 2012 ; Vol. 16, No. 4. pp. 819-830.
@article{a5e1fc2b6f5a411387907781edea2605,
title = "Nonlinear dimensionality reduction combining MR imaging with non-imaging information",
abstract = "We propose a framework for the extraction of biomarkers from low-dimensional manifolds representing inter-subject brain variation. Manifold coordinates of each image capture information about structural shape and appearance and, when a phenotype exists, about the subject’s clinical state. Our framework incorporates subject meta-information into the manifold learning step. Apart from gender and age, information such as genotype or a derived biomarker is often available in clinical studies and can inform the classification of a query subject. Such information, whether discrete or continuous, is used as an additional input to manifold learning, extending the Laplacian Eigenmap objective function and enriching a similarity measure derived from pairwise image similarities. The biomarkers identified with the proposed method are data-driven in contrast to a priori defined biomarkers derived from, e.g., manual or automated segmentations. They form a unified representation of both the imaging and non-imaging measurements, providing a natural use for data analysis and visualization. We test the method to classify subjects with Alzheimer’s Disease (AD), mild cognitive impairment (MCI) and healthy controls enrolled in the ADNI study. Non-imaging metadata used are ApoE genotype, a risk factor associated with AD, and the CSF-concentration of Αβ1-42, an established biomarker for AD. In addition, we use hippocampal volume as a derived imaging-biomarker to enrich the learned manifold. Our classification results compare favorably to what has been reported in a recent meta-analysis using established neuroimaging methods on the same database.",
keywords = "Manifold learning, Laplacian Eigenmaps, Classification, Metadata, Alzheimer's disease",
author = "Robin Wolz and Paul Aljabar and Hajnal, {Joseph V.} and Jyrki L{\"o}tj{\"o}nen and Daniel Rueckert and {The Alzheimer’s Disease Neuroimaging Initiative}",
year = "2012",
doi = "10.1016/j.media.2011.12.003",
language = "English",
volume = "16",
pages = "819--830",
journal = "Medical Image Analysis",
issn = "1361-8415",
publisher = "Elsevier",
number = "4",

}

Wolz, R, Aljabar, P, Hajnal, JV, Lötjönen, J, Rueckert, D & The Alzheimer’s Disease Neuroimaging Initiative 2012, 'Nonlinear dimensionality reduction combining MR imaging with non-imaging information', Medical Image Analysis, vol. 16, no. 4, pp. 819-830. https://doi.org/10.1016/j.media.2011.12.003

Nonlinear dimensionality reduction combining MR imaging with non-imaging information. / Wolz, Robin (Corresponding Author); Aljabar, Paul; Hajnal, Joseph V.; Lötjönen, Jyrki; Rueckert, Daniel; The Alzheimer’s Disease Neuroimaging Initiative.

In: Medical Image Analysis, Vol. 16, No. 4, 2012, p. 819-830.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Nonlinear dimensionality reduction combining MR imaging with non-imaging information

AU - Wolz, Robin

AU - Aljabar, Paul

AU - Hajnal, Joseph V.

AU - Lötjönen, Jyrki

AU - Rueckert, Daniel

AU - The Alzheimer’s Disease Neuroimaging Initiative

PY - 2012

Y1 - 2012

N2 - We propose a framework for the extraction of biomarkers from low-dimensional manifolds representing inter-subject brain variation. Manifold coordinates of each image capture information about structural shape and appearance and, when a phenotype exists, about the subject’s clinical state. Our framework incorporates subject meta-information into the manifold learning step. Apart from gender and age, information such as genotype or a derived biomarker is often available in clinical studies and can inform the classification of a query subject. Such information, whether discrete or continuous, is used as an additional input to manifold learning, extending the Laplacian Eigenmap objective function and enriching a similarity measure derived from pairwise image similarities. The biomarkers identified with the proposed method are data-driven in contrast to a priori defined biomarkers derived from, e.g., manual or automated segmentations. They form a unified representation of both the imaging and non-imaging measurements, providing a natural use for data analysis and visualization. We test the method to classify subjects with Alzheimer’s Disease (AD), mild cognitive impairment (MCI) and healthy controls enrolled in the ADNI study. Non-imaging metadata used are ApoE genotype, a risk factor associated with AD, and the CSF-concentration of Αβ1-42, an established biomarker for AD. In addition, we use hippocampal volume as a derived imaging-biomarker to enrich the learned manifold. Our classification results compare favorably to what has been reported in a recent meta-analysis using established neuroimaging methods on the same database.

AB - We propose a framework for the extraction of biomarkers from low-dimensional manifolds representing inter-subject brain variation. Manifold coordinates of each image capture information about structural shape and appearance and, when a phenotype exists, about the subject’s clinical state. Our framework incorporates subject meta-information into the manifold learning step. Apart from gender and age, information such as genotype or a derived biomarker is often available in clinical studies and can inform the classification of a query subject. Such information, whether discrete or continuous, is used as an additional input to manifold learning, extending the Laplacian Eigenmap objective function and enriching a similarity measure derived from pairwise image similarities. The biomarkers identified with the proposed method are data-driven in contrast to a priori defined biomarkers derived from, e.g., manual or automated segmentations. They form a unified representation of both the imaging and non-imaging measurements, providing a natural use for data analysis and visualization. We test the method to classify subjects with Alzheimer’s Disease (AD), mild cognitive impairment (MCI) and healthy controls enrolled in the ADNI study. Non-imaging metadata used are ApoE genotype, a risk factor associated with AD, and the CSF-concentration of Αβ1-42, an established biomarker for AD. In addition, we use hippocampal volume as a derived imaging-biomarker to enrich the learned manifold. Our classification results compare favorably to what has been reported in a recent meta-analysis using established neuroimaging methods on the same database.

KW - Manifold learning

KW - Laplacian Eigenmaps

KW - Classification

KW - Metadata

KW - Alzheimer's disease

U2 - 10.1016/j.media.2011.12.003

DO - 10.1016/j.media.2011.12.003

M3 - Article

VL - 16

SP - 819

EP - 830

JO - Medical Image Analysis

JF - Medical Image Analysis

SN - 1361-8415

IS - 4

ER -

Wolz R, Aljabar P, Hajnal JV, Lötjönen J, Rueckert D, The Alzheimer’s Disease Neuroimaging Initiative. Nonlinear dimensionality reduction combining MR imaging with non-imaging information. Medical Image Analysis. 2012;16(4):819-830. https://doi.org/10.1016/j.media.2011.12.003