Constrained sparse functional connectivity networks for MCI classification

Chong Yaw Wee, Pew Thian Yap, Daoqiang Zhang, Lihong Wang, Dinggang Shen

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

Abstract

Mild cognitive impairment (MCI) is difficult to diagnose due to its subtlety. Recent emergence of advanced network analysis techniques utilizing resting-state functional Magnetic Resonance Imaging (rs-fMRI) has made the understanding of neurological disorders more comprehensively at a whole-brain connectivity level. However, inferring effective brain connectivity from fMRI data is a challenging task, particularly when the ultimate goal is to obtain good control-patient classification performance. Incorporating sparsity into connectivity modeling can potentially produce results that are biologically more meaningful since most biologically networks are formed by a relatively few number of connections. However, this constraint, when applied at an individual level, will degrade classification performance due to inter-subject variability. To address this problem, we consider a constrained sparse linear regression model associated with the least absolute shrinkage and selection operator (LASSO). Specifically, we introduced sparsity into brain connectivity via l1-norm penalization, and ensured consistent non-zero connections across subjects via l2-norm penalization. Our results demonstrate that the constrained sparse network gives better classification performance than the conventional correlation-based network, indicating its greater sensitivity to early stage brain pathologies.

Original languageEnglish
Title of host publicationMedical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings
PublisherSpringer Verlag
Pages212-219
Number of pages8
Volume7511 LNCS
ISBN (Print)9783642334177
Publication statusPublished - 2012
Externally publishedYes
Event2nd MICCAI Workshop on Medical Computer Vision, MICCAI-MCV 2012, Held in Conjunction with the 15th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2012 - Nice, France
Duration: 2012 Oct 52012 Oct 5

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7511 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other2nd MICCAI Workshop on Medical Computer Vision, MICCAI-MCV 2012, Held in Conjunction with the 15th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2012
CountryFrance
CityNice
Period12/10/512/10/5

Fingerprint

Network Connectivity
Brain
Connectivity
Functional Magnetic Resonance Imaging
Penalization
Sparsity
L1-norm
Network Analysis
Pathology
Shrinkage
Electric network analysis
Linear Regression Model
Linear regression
Disorder
Norm
Operator
Modeling
Demonstrate
Magnetic Resonance Imaging

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Wee, C. Y., Yap, P. T., Zhang, D., Wang, L., & Shen, D. (2012). Constrained sparse functional connectivity networks for MCI classification. In Medical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings (Vol. 7511 LNCS, pp. 212-219). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7511 LNCS). Springer Verlag.

Constrained sparse functional connectivity networks for MCI classification. / Wee, Chong Yaw; Yap, Pew Thian; Zhang, Daoqiang; Wang, Lihong; Shen, Dinggang.

Medical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings. Vol. 7511 LNCS Springer Verlag, 2012. p. 212-219 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7511 LNCS).

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

Wee, CY, Yap, PT, Zhang, D, Wang, L & Shen, D 2012, Constrained sparse functional connectivity networks for MCI classification. in Medical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings. vol. 7511 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7511 LNCS, Springer Verlag, pp. 212-219, 2nd MICCAI Workshop on Medical Computer Vision, MICCAI-MCV 2012, Held in Conjunction with the 15th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2012, Nice, France, 12/10/5.
Wee CY, Yap PT, Zhang D, Wang L, Shen D. Constrained sparse functional connectivity networks for MCI classification. In Medical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings. Vol. 7511 LNCS. Springer Verlag. 2012. p. 212-219. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Wee, Chong Yaw ; Yap, Pew Thian ; Zhang, Daoqiang ; Wang, Lihong ; Shen, Dinggang. / Constrained sparse functional connectivity networks for MCI classification. Medical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings. Vol. 7511 LNCS Springer Verlag, 2012. pp. 212-219 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{01e67df46a814458855eb903b1c152c7,
title = "Constrained sparse functional connectivity networks for MCI classification",
abstract = "Mild cognitive impairment (MCI) is difficult to diagnose due to its subtlety. Recent emergence of advanced network analysis techniques utilizing resting-state functional Magnetic Resonance Imaging (rs-fMRI) has made the understanding of neurological disorders more comprehensively at a whole-brain connectivity level. However, inferring effective brain connectivity from fMRI data is a challenging task, particularly when the ultimate goal is to obtain good control-patient classification performance. Incorporating sparsity into connectivity modeling can potentially produce results that are biologically more meaningful since most biologically networks are formed by a relatively few number of connections. However, this constraint, when applied at an individual level, will degrade classification performance due to inter-subject variability. To address this problem, we consider a constrained sparse linear regression model associated with the least absolute shrinkage and selection operator (LASSO). Specifically, we introduced sparsity into brain connectivity via l1-norm penalization, and ensured consistent non-zero connections across subjects via l2-norm penalization. Our results demonstrate that the constrained sparse network gives better classification performance than the conventional correlation-based network, indicating its greater sensitivity to early stage brain pathologies.",
author = "Wee, {Chong Yaw} and Yap, {Pew Thian} and Daoqiang Zhang and Lihong Wang and Dinggang Shen",
year = "2012",
language = "English",
isbn = "9783642334177",
volume = "7511 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "212--219",
booktitle = "Medical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings",

}

TY - GEN

T1 - Constrained sparse functional connectivity networks for MCI classification

AU - Wee, Chong Yaw

AU - Yap, Pew Thian

AU - Zhang, Daoqiang

AU - Wang, Lihong

AU - Shen, Dinggang

PY - 2012

Y1 - 2012

N2 - Mild cognitive impairment (MCI) is difficult to diagnose due to its subtlety. Recent emergence of advanced network analysis techniques utilizing resting-state functional Magnetic Resonance Imaging (rs-fMRI) has made the understanding of neurological disorders more comprehensively at a whole-brain connectivity level. However, inferring effective brain connectivity from fMRI data is a challenging task, particularly when the ultimate goal is to obtain good control-patient classification performance. Incorporating sparsity into connectivity modeling can potentially produce results that are biologically more meaningful since most biologically networks are formed by a relatively few number of connections. However, this constraint, when applied at an individual level, will degrade classification performance due to inter-subject variability. To address this problem, we consider a constrained sparse linear regression model associated with the least absolute shrinkage and selection operator (LASSO). Specifically, we introduced sparsity into brain connectivity via l1-norm penalization, and ensured consistent non-zero connections across subjects via l2-norm penalization. Our results demonstrate that the constrained sparse network gives better classification performance than the conventional correlation-based network, indicating its greater sensitivity to early stage brain pathologies.

AB - Mild cognitive impairment (MCI) is difficult to diagnose due to its subtlety. Recent emergence of advanced network analysis techniques utilizing resting-state functional Magnetic Resonance Imaging (rs-fMRI) has made the understanding of neurological disorders more comprehensively at a whole-brain connectivity level. However, inferring effective brain connectivity from fMRI data is a challenging task, particularly when the ultimate goal is to obtain good control-patient classification performance. Incorporating sparsity into connectivity modeling can potentially produce results that are biologically more meaningful since most biologically networks are formed by a relatively few number of connections. However, this constraint, when applied at an individual level, will degrade classification performance due to inter-subject variability. To address this problem, we consider a constrained sparse linear regression model associated with the least absolute shrinkage and selection operator (LASSO). Specifically, we introduced sparsity into brain connectivity via l1-norm penalization, and ensured consistent non-zero connections across subjects via l2-norm penalization. Our results demonstrate that the constrained sparse network gives better classification performance than the conventional correlation-based network, indicating its greater sensitivity to early stage brain pathologies.

UR - http://www.scopus.com/inward/record.url?scp=84988892523&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84988892523&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9783642334177

VL - 7511 LNCS

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

SP - 212

EP - 219

BT - Medical Image Computing and Computer-Assisted Intervention - MICCAI2012 - 15th International Conference, Proceedings

PB - Springer Verlag

ER -