Sparse inverse covariance analysis of human brain for Alzheimer's disease study

Rinkal Patel, Jun Liu, Kewei Chen, Eric Reiman, Gene Alexander, Jieping Ye

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

Abstract

Analysis of functional neuroimaging data in the studies of human brain has become very critical in understanding neuro-degenerative diseases such as Alzheimer's disease (AD). The most common approach in AD neuroimaging studies has been of univariate nature, where individual brain regions/voxels are analyzed separately. In many cases these techniques prove to be effective. However, they could not shed light on inter brain region connectivity associated with the brain function or disease of interest. Indeed, human brain is a very complex organ anatomically and the functional interactions between its regions are even more. As a result, there is a need to understand this interdependency or inter-connection of brain regions. There are several existing techniques to address this issue. They include principal component analysis (PCA), PCA based scaled subprofile modeling (SSM), Bayesian network approach and independent component analysis (ICA). In this study, we propose a machine learning technique called "Sparse Inverse Covariance Analysis" to learn the brain region interactivity, with minimal computational cost and appropriate degree of sparsity. Under Gaussian assumption, each element of the inverse covariance matrix represents conditional dependence between the constituent pair of variables, given all other variables. By introducing sparsity constraint, unnecessary/noisy functional dependencies are eliminated by setting the constituent element to zero, resulting into conditional independence between the variable pairs. Using functional FDG-PET data acquired from 49 AD and 67 normal subjects from the Alzheimer's disease neuroimaging initiative (ADNI) project, we evaluate this technique in terms of distinct brain region connectivity pattern in patients with AD compared to that in normal control subjects. It was found that the patients with AD had disconnections that are not present in the normal controls. This different connectivity pattern is potentially usable for clinical diagnosis and for establishing sensitive markers for the disease progression and treatment evaluation.

Original languageEnglish (US)
Title of host publication2009 ICME International Conference on Complex Medical Engineering, CME 2009
DOIs
StatePublished - 2009
Event2009 ICME International Conference on Complex Medical Engineering, CME 2009 - Tempe, AZ, United States
Duration: Apr 9 2009Apr 11 2009

Other

Other2009 ICME International Conference on Complex Medical Engineering, CME 2009
CountryUnited States
CityTempe, AZ
Period4/9/094/11/09

Fingerprint

Brain Diseases
Brain
Alzheimer Disease
Neuroimaging
Principal Component Analysis
Principal component analysis
Functional neuroimaging
Neurodegenerative diseases
Functional Neuroimaging
Bayes Theorem
Independent component analysis
Bayesian networks
Covariance matrix
Disease Progression
Learning systems
Costs and Cost Analysis

ASJC Scopus subject areas

  • Biomedical Engineering
  • Health Informatics

Cite this

Patel, R., Liu, J., Chen, K., Reiman, E., Alexander, G., & Ye, J. (2009). Sparse inverse covariance analysis of human brain for Alzheimer's disease study. In 2009 ICME International Conference on Complex Medical Engineering, CME 2009 [4906604] https://doi.org/10.1109/ICCME.2009.4906604

Sparse inverse covariance analysis of human brain for Alzheimer's disease study. / Patel, Rinkal; Liu, Jun; Chen, Kewei; Reiman, Eric; Alexander, Gene; Ye, Jieping.

2009 ICME International Conference on Complex Medical Engineering, CME 2009. 2009. 4906604.

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

Patel, R, Liu, J, Chen, K, Reiman, E, Alexander, G & Ye, J 2009, Sparse inverse covariance analysis of human brain for Alzheimer's disease study. in 2009 ICME International Conference on Complex Medical Engineering, CME 2009., 4906604, 2009 ICME International Conference on Complex Medical Engineering, CME 2009, Tempe, AZ, United States, 4/9/09. https://doi.org/10.1109/ICCME.2009.4906604
Patel R, Liu J, Chen K, Reiman E, Alexander G, Ye J. Sparse inverse covariance analysis of human brain for Alzheimer's disease study. In 2009 ICME International Conference on Complex Medical Engineering, CME 2009. 2009. 4906604 https://doi.org/10.1109/ICCME.2009.4906604
Patel, Rinkal ; Liu, Jun ; Chen, Kewei ; Reiman, Eric ; Alexander, Gene ; Ye, Jieping. / Sparse inverse covariance analysis of human brain for Alzheimer's disease study. 2009 ICME International Conference on Complex Medical Engineering, CME 2009. 2009.
@inproceedings{76747d77451947b884f0ff1f027b40ad,
title = "Sparse inverse covariance analysis of human brain for Alzheimer's disease study",
abstract = "Analysis of functional neuroimaging data in the studies of human brain has become very critical in understanding neuro-degenerative diseases such as Alzheimer's disease (AD). The most common approach in AD neuroimaging studies has been of univariate nature, where individual brain regions/voxels are analyzed separately. In many cases these techniques prove to be effective. However, they could not shed light on inter brain region connectivity associated with the brain function or disease of interest. Indeed, human brain is a very complex organ anatomically and the functional interactions between its regions are even more. As a result, there is a need to understand this interdependency or inter-connection of brain regions. There are several existing techniques to address this issue. They include principal component analysis (PCA), PCA based scaled subprofile modeling (SSM), Bayesian network approach and independent component analysis (ICA). In this study, we propose a machine learning technique called {"}Sparse Inverse Covariance Analysis{"} to learn the brain region interactivity, with minimal computational cost and appropriate degree of sparsity. Under Gaussian assumption, each element of the inverse covariance matrix represents conditional dependence between the constituent pair of variables, given all other variables. By introducing sparsity constraint, unnecessary/noisy functional dependencies are eliminated by setting the constituent element to zero, resulting into conditional independence between the variable pairs. Using functional FDG-PET data acquired from 49 AD and 67 normal subjects from the Alzheimer's disease neuroimaging initiative (ADNI) project, we evaluate this technique in terms of distinct brain region connectivity pattern in patients with AD compared to that in normal control subjects. It was found that the patients with AD had disconnections that are not present in the normal controls. This different connectivity pattern is potentially usable for clinical diagnosis and for establishing sensitive markers for the disease progression and treatment evaluation.",
author = "Rinkal Patel and Jun Liu and Kewei Chen and Eric Reiman and Gene Alexander and Jieping Ye",
year = "2009",
doi = "10.1109/ICCME.2009.4906604",
language = "English (US)",
isbn = "9781424433162",
booktitle = "2009 ICME International Conference on Complex Medical Engineering, CME 2009",

}

TY - GEN

T1 - Sparse inverse covariance analysis of human brain for Alzheimer's disease study

AU - Patel, Rinkal

AU - Liu, Jun

AU - Chen, Kewei

AU - Reiman, Eric

AU - Alexander, Gene

AU - Ye, Jieping

PY - 2009

Y1 - 2009

N2 - Analysis of functional neuroimaging data in the studies of human brain has become very critical in understanding neuro-degenerative diseases such as Alzheimer's disease (AD). The most common approach in AD neuroimaging studies has been of univariate nature, where individual brain regions/voxels are analyzed separately. In many cases these techniques prove to be effective. However, they could not shed light on inter brain region connectivity associated with the brain function or disease of interest. Indeed, human brain is a very complex organ anatomically and the functional interactions between its regions are even more. As a result, there is a need to understand this interdependency or inter-connection of brain regions. There are several existing techniques to address this issue. They include principal component analysis (PCA), PCA based scaled subprofile modeling (SSM), Bayesian network approach and independent component analysis (ICA). In this study, we propose a machine learning technique called "Sparse Inverse Covariance Analysis" to learn the brain region interactivity, with minimal computational cost and appropriate degree of sparsity. Under Gaussian assumption, each element of the inverse covariance matrix represents conditional dependence between the constituent pair of variables, given all other variables. By introducing sparsity constraint, unnecessary/noisy functional dependencies are eliminated by setting the constituent element to zero, resulting into conditional independence between the variable pairs. Using functional FDG-PET data acquired from 49 AD and 67 normal subjects from the Alzheimer's disease neuroimaging initiative (ADNI) project, we evaluate this technique in terms of distinct brain region connectivity pattern in patients with AD compared to that in normal control subjects. It was found that the patients with AD had disconnections that are not present in the normal controls. This different connectivity pattern is potentially usable for clinical diagnosis and for establishing sensitive markers for the disease progression and treatment evaluation.

AB - Analysis of functional neuroimaging data in the studies of human brain has become very critical in understanding neuro-degenerative diseases such as Alzheimer's disease (AD). The most common approach in AD neuroimaging studies has been of univariate nature, where individual brain regions/voxels are analyzed separately. In many cases these techniques prove to be effective. However, they could not shed light on inter brain region connectivity associated with the brain function or disease of interest. Indeed, human brain is a very complex organ anatomically and the functional interactions between its regions are even more. As a result, there is a need to understand this interdependency or inter-connection of brain regions. There are several existing techniques to address this issue. They include principal component analysis (PCA), PCA based scaled subprofile modeling (SSM), Bayesian network approach and independent component analysis (ICA). In this study, we propose a machine learning technique called "Sparse Inverse Covariance Analysis" to learn the brain region interactivity, with minimal computational cost and appropriate degree of sparsity. Under Gaussian assumption, each element of the inverse covariance matrix represents conditional dependence between the constituent pair of variables, given all other variables. By introducing sparsity constraint, unnecessary/noisy functional dependencies are eliminated by setting the constituent element to zero, resulting into conditional independence between the variable pairs. Using functional FDG-PET data acquired from 49 AD and 67 normal subjects from the Alzheimer's disease neuroimaging initiative (ADNI) project, we evaluate this technique in terms of distinct brain region connectivity pattern in patients with AD compared to that in normal control subjects. It was found that the patients with AD had disconnections that are not present in the normal controls. This different connectivity pattern is potentially usable for clinical diagnosis and for establishing sensitive markers for the disease progression and treatment evaluation.

UR - http://www.scopus.com/inward/record.url?scp=67650648738&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67650648738&partnerID=8YFLogxK

U2 - 10.1109/ICCME.2009.4906604

DO - 10.1109/ICCME.2009.4906604

M3 - Conference contribution

SN - 9781424433162

BT - 2009 ICME International Conference on Complex Medical Engineering, CME 2009

ER -