Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder Identification
Functional brain network (FBN) has been becoming an increasingly important way to model the statistical dependence among neural time courses of brain, and provides effective imaging biomarkers for diagnosis of some neurological or psychological disorders. Currently, Pearson's Correlation (PC) i...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Frontiers Media S.A.
2017-08-01
|
Series: | Frontiers in Neuroinformatics |
Subjects: | |
Online Access: | http://journal.frontiersin.org/article/10.3389/fninf.2017.00055/full |
id |
doaj-6107d03af2964313974871e7dbddb815 |
---|---|
record_format |
Article |
spelling |
doaj-6107d03af2964313974871e7dbddb8152020-11-25T00:34:31ZengFrontiers Media S.A.Frontiers in Neuroinformatics1662-51962017-08-011110.3389/fninf.2017.00055251582Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder IdentificationWeikai Li0Weikai Li1Zhengxia Wang2Limei Zhang3Lishan Qiao4Dinggang Shen5Dinggang Shen6College of Information Science and Engineering, Chongqing Jiaotong UniversityChongqing, ChinaSchool of Mathematics, Liaocheng UniversityLiaocheng, ChinaCollege of Information Science and Engineering, Chongqing Jiaotong UniversityChongqing, ChinaSchool of Mathematics, Liaocheng UniversityLiaocheng, ChinaSchool of Mathematics, Liaocheng UniversityLiaocheng, ChinaDepartment of Radiology and BRIC, University of North Carolina at Chapel HillChapel Hill, NC, United StatesDepartment of Brain and Cognitive Engineering, Korea UniversitySeoul, South KoreaFunctional brain network (FBN) has been becoming an increasingly important way to model the statistical dependence among neural time courses of brain, and provides effective imaging biomarkers for diagnosis of some neurological or psychological disorders. Currently, Pearson's Correlation (PC) is the simplest and most widely-used method in constructing FBNs. Despite its advantages in statistical meaning and calculated performance, the PC tends to result in a FBN with dense connections. Therefore, in practice, the PC-based FBN needs to be sparsified by removing weak (potential noisy) connections. However, such a scheme depends on a hard-threshold without enough flexibility. Different from this traditional strategy, in this paper, we propose a new approach for estimating FBNs by remodeling PC as an optimization problem, which provides a way to incorporate biological/physical priors into the FBNs. In particular, we introduce an L1-norm regularizer into the optimization model for obtaining a sparse solution. Compared with the hard-threshold scheme, the proposed framework gives an elegant mathematical formulation for sparsifying PC-based networks. More importantly, it provides a platform to encode other biological/physical priors into the PC-based FBNs. To further illustrate the flexibility of the proposed method, we extend the model to a weighted counterpart for learning both sparse and scale-free networks, and then conduct experiments to identify autism spectrum disorders (ASD) from normal controls (NC) based on the constructed FBNs. Consequently, we achieved an 81.52% classification accuracy which outperforms the baseline and state-of-the-art methods.http://journal.frontiersin.org/article/10.3389/fninf.2017.00055/fullfunctional brain networkfunctional magnetic resonance imagingPearson's correlationsparse representationscale-freeautism spectrum disorder |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Weikai Li Weikai Li Zhengxia Wang Limei Zhang Lishan Qiao Dinggang Shen Dinggang Shen |
spellingShingle |
Weikai Li Weikai Li Zhengxia Wang Limei Zhang Lishan Qiao Dinggang Shen Dinggang Shen Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder Identification Frontiers in Neuroinformatics functional brain network functional magnetic resonance imaging Pearson's correlation sparse representation scale-free autism spectrum disorder |
author_facet |
Weikai Li Weikai Li Zhengxia Wang Limei Zhang Lishan Qiao Dinggang Shen Dinggang Shen |
author_sort |
Weikai Li |
title |
Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder Identification |
title_short |
Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder Identification |
title_full |
Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder Identification |
title_fullStr |
Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder Identification |
title_full_unstemmed |
Remodeling Pearson's Correlation for Functional Brain Network Estimation and Autism Spectrum Disorder Identification |
title_sort |
remodeling pearson's correlation for functional brain network estimation and autism spectrum disorder identification |
publisher |
Frontiers Media S.A. |
series |
Frontiers in Neuroinformatics |
issn |
1662-5196 |
publishDate |
2017-08-01 |
description |
Functional brain network (FBN) has been becoming an increasingly important way to model the statistical dependence among neural time courses of brain, and provides effective imaging biomarkers for diagnosis of some neurological or psychological disorders. Currently, Pearson's Correlation (PC) is the simplest and most widely-used method in constructing FBNs. Despite its advantages in statistical meaning and calculated performance, the PC tends to result in a FBN with dense connections. Therefore, in practice, the PC-based FBN needs to be sparsified by removing weak (potential noisy) connections. However, such a scheme depends on a hard-threshold without enough flexibility. Different from this traditional strategy, in this paper, we propose a new approach for estimating FBNs by remodeling PC as an optimization problem, which provides a way to incorporate biological/physical priors into the FBNs. In particular, we introduce an L1-norm regularizer into the optimization model for obtaining a sparse solution. Compared with the hard-threshold scheme, the proposed framework gives an elegant mathematical formulation for sparsifying PC-based networks. More importantly, it provides a platform to encode other biological/physical priors into the PC-based FBNs. To further illustrate the flexibility of the proposed method, we extend the model to a weighted counterpart for learning both sparse and scale-free networks, and then conduct experiments to identify autism spectrum disorders (ASD) from normal controls (NC) based on the constructed FBNs. Consequently, we achieved an 81.52% classification accuracy which outperforms the baseline and state-of-the-art methods. |
topic |
functional brain network functional magnetic resonance imaging Pearson's correlation sparse representation scale-free autism spectrum disorder |
url |
http://journal.frontiersin.org/article/10.3389/fninf.2017.00055/full |
work_keys_str_mv |
AT weikaili remodelingpearsonscorrelationforfunctionalbrainnetworkestimationandautismspectrumdisorderidentification AT weikaili remodelingpearsonscorrelationforfunctionalbrainnetworkestimationandautismspectrumdisorderidentification AT zhengxiawang remodelingpearsonscorrelationforfunctionalbrainnetworkestimationandautismspectrumdisorderidentification AT limeizhang remodelingpearsonscorrelationforfunctionalbrainnetworkestimationandautismspectrumdisorderidentification AT lishanqiao remodelingpearsonscorrelationforfunctionalbrainnetworkestimationandautismspectrumdisorderidentification AT dinggangshen remodelingpearsonscorrelationforfunctionalbrainnetworkestimationandautismspectrumdisorderidentification AT dinggangshen remodelingpearsonscorrelationforfunctionalbrainnetworkestimationandautismspectrumdisorderidentification |
_version_ |
1725312989279551488 |