Integrative analysis of multi-dimensional imaging genomics data for alzheimer's disease prediction

The Alzheimer's Disease Neuroimaging Initiative

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

In this paper, we explore the effects of integrating multi-dimensional imaging genomics data for Alzheimer's disease (AD) prediction using machine learning approaches. Precisely, we compare our three recent proposed feature selection methods (i.e. multiple kernel learning (MKL), high order graph matching based feature selection (HGM-FS), sparse multimodal learning (SMML)) using four widely-used modalities (i.e. magnetic resonance imaging (MRI), positron emission tomography (PET), cerebrospinal fluid (CSF), and genetic modality single-nucleotide polymor phism (SNP)). This study demonstrates the performance of each method using these modalities individually or integratively, and may be valuable to clinical tests in practice. Our experimental results suggest that for AD prediction, in general, (1) in terms of accuracy, PET is the best modality; (2) Even though the discriminant power of genetic SNP features is weak, adding this modality to other modalities does help improve the classification accuracy; (3) HGM-FS works best among the three feature selection methods; (4) Some of the selected features are shared by all the feature selection methods, which may have high correlation with the disease. Using all the modalities on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, the best accuracies, described as (mean±standard deviation)%, among the three methods are (76.2±11.3)% for AD vs. MCI, (94.8±7.3)% for AD vs. HC, (76.5±11.1)% for MCI vs. HC, and (71.0±8.4)% for AD vs. MCI vs. HC, respectively.

Original languageEnglish
Article number260
JournalFrontiers in Aging Neuroscience
Volume6
Issue numberSEP
DOIs
Publication statusPublished - 2014 Jan 1
Externally publishedYes

Fingerprint

Genomics
Alzheimer Disease
Positron-Emission Tomography
Nucleotides
Learning
Neuroimaging
Cerebrospinal Fluid
Magnetic Resonance Imaging

ASJC Scopus subject areas

  • Ageing
  • Cognitive Neuroscience

Cite this

Integrative analysis of multi-dimensional imaging genomics data for alzheimer's disease prediction. / The Alzheimer's Disease Neuroimaging Initiative.

In: Frontiers in Aging Neuroscience, Vol. 6, No. SEP, 260, 01.01.2014.

Research output: Contribution to journalArticle

@article{7a3574e1f84145689933724760922b9d,
title = "Integrative analysis of multi-dimensional imaging genomics data for alzheimer's disease prediction",
abstract = "In this paper, we explore the effects of integrating multi-dimensional imaging genomics data for Alzheimer's disease (AD) prediction using machine learning approaches. Precisely, we compare our three recent proposed feature selection methods (i.e. multiple kernel learning (MKL), high order graph matching based feature selection (HGM-FS), sparse multimodal learning (SMML)) using four widely-used modalities (i.e. magnetic resonance imaging (MRI), positron emission tomography (PET), cerebrospinal fluid (CSF), and genetic modality single-nucleotide polymor phism (SNP)). This study demonstrates the performance of each method using these modalities individually or integratively, and may be valuable to clinical tests in practice. Our experimental results suggest that for AD prediction, in general, (1) in terms of accuracy, PET is the best modality; (2) Even though the discriminant power of genetic SNP features is weak, adding this modality to other modalities does help improve the classification accuracy; (3) HGM-FS works best among the three feature selection methods; (4) Some of the selected features are shared by all the feature selection methods, which may have high correlation with the disease. Using all the modalities on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, the best accuracies, described as (mean±standard deviation){\%}, among the three methods are (76.2±11.3){\%} for AD vs. MCI, (94.8±7.3){\%} for AD vs. HC, (76.5±11.1){\%} for MCI vs. HC, and (71.0±8.4){\%} for AD vs. MCI vs. HC, respectively.",
keywords = "Alzheimer's disease prediction, Binary and multiclass classification, Feature selection, Imaging genomics data, Modality integration",
author = "{The Alzheimer's Disease Neuroimaging Initiative} and Ziming Zhang and Heng Huang and Dinggang Shen",
year = "2014",
month = "1",
day = "1",
doi = "10.3389/fnagi.2014.00260",
language = "English",
volume = "6",
journal = "Frontiers in Aging Neuroscience",
issn = "1663-4365",
publisher = "Frontiers Research Foundation",
number = "SEP",

}

TY - JOUR

T1 - Integrative analysis of multi-dimensional imaging genomics data for alzheimer's disease prediction

AU - The Alzheimer's Disease Neuroimaging Initiative

AU - Zhang, Ziming

AU - Huang, Heng

AU - Shen, Dinggang

PY - 2014/1/1

Y1 - 2014/1/1

N2 - In this paper, we explore the effects of integrating multi-dimensional imaging genomics data for Alzheimer's disease (AD) prediction using machine learning approaches. Precisely, we compare our three recent proposed feature selection methods (i.e. multiple kernel learning (MKL), high order graph matching based feature selection (HGM-FS), sparse multimodal learning (SMML)) using four widely-used modalities (i.e. magnetic resonance imaging (MRI), positron emission tomography (PET), cerebrospinal fluid (CSF), and genetic modality single-nucleotide polymor phism (SNP)). This study demonstrates the performance of each method using these modalities individually or integratively, and may be valuable to clinical tests in practice. Our experimental results suggest that for AD prediction, in general, (1) in terms of accuracy, PET is the best modality; (2) Even though the discriminant power of genetic SNP features is weak, adding this modality to other modalities does help improve the classification accuracy; (3) HGM-FS works best among the three feature selection methods; (4) Some of the selected features are shared by all the feature selection methods, which may have high correlation with the disease. Using all the modalities on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, the best accuracies, described as (mean±standard deviation)%, among the three methods are (76.2±11.3)% for AD vs. MCI, (94.8±7.3)% for AD vs. HC, (76.5±11.1)% for MCI vs. HC, and (71.0±8.4)% for AD vs. MCI vs. HC, respectively.

AB - In this paper, we explore the effects of integrating multi-dimensional imaging genomics data for Alzheimer's disease (AD) prediction using machine learning approaches. Precisely, we compare our three recent proposed feature selection methods (i.e. multiple kernel learning (MKL), high order graph matching based feature selection (HGM-FS), sparse multimodal learning (SMML)) using four widely-used modalities (i.e. magnetic resonance imaging (MRI), positron emission tomography (PET), cerebrospinal fluid (CSF), and genetic modality single-nucleotide polymor phism (SNP)). This study demonstrates the performance of each method using these modalities individually or integratively, and may be valuable to clinical tests in practice. Our experimental results suggest that for AD prediction, in general, (1) in terms of accuracy, PET is the best modality; (2) Even though the discriminant power of genetic SNP features is weak, adding this modality to other modalities does help improve the classification accuracy; (3) HGM-FS works best among the three feature selection methods; (4) Some of the selected features are shared by all the feature selection methods, which may have high correlation with the disease. Using all the modalities on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, the best accuracies, described as (mean±standard deviation)%, among the three methods are (76.2±11.3)% for AD vs. MCI, (94.8±7.3)% for AD vs. HC, (76.5±11.1)% for MCI vs. HC, and (71.0±8.4)% for AD vs. MCI vs. HC, respectively.

KW - Alzheimer's disease prediction

KW - Binary and multiclass classification

KW - Feature selection

KW - Imaging genomics data

KW - Modality integration

UR - http://www.scopus.com/inward/record.url?scp=84907155856&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907155856&partnerID=8YFLogxK

U2 - 10.3389/fnagi.2014.00260

DO - 10.3389/fnagi.2014.00260

M3 - Article

AN - SCOPUS:84907155856

VL - 6

JO - Frontiers in Aging Neuroscience

JF - Frontiers in Aging Neuroscience

SN - 1663-4365

IS - SEP

M1 - 260

ER -