Deep Collaborative Learning With Application to the Study of Multimodal Brain Development

Wenxing Hu, Biao Cai, Aiying Zhang, Vince D. Calhoun, Yu Ping Wang

Research output: Contribution to journalArticle

Abstract

OBJECTIVE: Multi-modal functional magnetic resonance imaging has been widely used for brain research. Conventional data-fusion methods cannot capture complex relationship (e.g., nonlinear predictive relationship) between multiple data. This paper aims to develop a neural network framework to extract phenotype related cross-data relationships and use it to study the brain development. METHODS: We propose a novel method, deep collaborative learning (DCL), to address the limitation of existing methods. DCL first uses a deep network to represent original data and then seeks their correlations, while also linking the data representation with phenotypical information. RESULTS: We studied the difference of functional connectivity (FCs) between different age groups and also use FCs as a fingerprint to predict cognitive abilities. Our experiments demonstrated higher accuracy of using DCL over other conventional models when classifying populations of different ages and cognitive scores. Moreover, DCL revealed that brain connections became stronger at adolescence stage. Furthermore, DCL detected strong correlations between default mode network and other networks which were overlooked by linear canonical correlation analysis, demonstrating DCL's ability of detecting nonlinear correlations. CONCLUSION: The results verified the superiority of DCL over conventional data-fusion methods. In addition, the stronger brain connection demonstrated the importance of adolescence stage for brain development. SIGNIFICANCE: DCL can better combine complex correlations between multiple data sets in addition to their fitting to phenotypes, with the potential to overcome the limitations of several current data-fusion models.

Original languageEnglish (US)
Pages (from-to)3346-3359
Number of pages14
JournalIEEE transactions on bio-medical engineering
Volume66
Issue number12
DOIs
StatePublished - Dec 1 2019

Fingerprint

Brain
Data fusion
Neural networks
Experiments

ASJC Scopus subject areas

  • Biomedical Engineering

Cite this

Deep Collaborative Learning With Application to the Study of Multimodal Brain Development. / Hu, Wenxing; Cai, Biao; Zhang, Aiying; Calhoun, Vince D.; Wang, Yu Ping.

In: IEEE transactions on bio-medical engineering, Vol. 66, No. 12, 01.12.2019, p. 3346-3359.

Research output: Contribution to journalArticle

Hu, Wenxing ; Cai, Biao ; Zhang, Aiying ; Calhoun, Vince D. ; Wang, Yu Ping. / Deep Collaborative Learning With Application to the Study of Multimodal Brain Development. In: IEEE transactions on bio-medical engineering. 2019 ; Vol. 66, No. 12. pp. 3346-3359.
@article{440daac20566478084dbf69f6b28f48a,
title = "Deep Collaborative Learning With Application to the Study of Multimodal Brain Development",
abstract = "OBJECTIVE: Multi-modal functional magnetic resonance imaging has been widely used for brain research. Conventional data-fusion methods cannot capture complex relationship (e.g., nonlinear predictive relationship) between multiple data. This paper aims to develop a neural network framework to extract phenotype related cross-data relationships and use it to study the brain development. METHODS: We propose a novel method, deep collaborative learning (DCL), to address the limitation of existing methods. DCL first uses a deep network to represent original data and then seeks their correlations, while also linking the data representation with phenotypical information. RESULTS: We studied the difference of functional connectivity (FCs) between different age groups and also use FCs as a fingerprint to predict cognitive abilities. Our experiments demonstrated higher accuracy of using DCL over other conventional models when classifying populations of different ages and cognitive scores. Moreover, DCL revealed that brain connections became stronger at adolescence stage. Furthermore, DCL detected strong correlations between default mode network and other networks which were overlooked by linear canonical correlation analysis, demonstrating DCL's ability of detecting nonlinear correlations. CONCLUSION: The results verified the superiority of DCL over conventional data-fusion methods. In addition, the stronger brain connection demonstrated the importance of adolescence stage for brain development. SIGNIFICANCE: DCL can better combine complex correlations between multiple data sets in addition to their fitting to phenotypes, with the potential to overcome the limitations of several current data-fusion models.",
author = "Wenxing Hu and Biao Cai and Aiying Zhang and Calhoun, {Vince D.} and Wang, {Yu Ping}",
year = "2019",
month = "12",
day = "1",
doi = "10.1109/TBME.2019.2904301",
language = "English (US)",
volume = "66",
pages = "3346--3359",
journal = "IEEE Transactions on Biomedical Engineering",
issn = "0018-9294",
publisher = "IEEE Computer Society",
number = "12",

}

TY - JOUR

T1 - Deep Collaborative Learning With Application to the Study of Multimodal Brain Development

AU - Hu, Wenxing

AU - Cai, Biao

AU - Zhang, Aiying

AU - Calhoun, Vince D.

AU - Wang, Yu Ping

PY - 2019/12/1

Y1 - 2019/12/1

N2 - OBJECTIVE: Multi-modal functional magnetic resonance imaging has been widely used for brain research. Conventional data-fusion methods cannot capture complex relationship (e.g., nonlinear predictive relationship) between multiple data. This paper aims to develop a neural network framework to extract phenotype related cross-data relationships and use it to study the brain development. METHODS: We propose a novel method, deep collaborative learning (DCL), to address the limitation of existing methods. DCL first uses a deep network to represent original data and then seeks their correlations, while also linking the data representation with phenotypical information. RESULTS: We studied the difference of functional connectivity (FCs) between different age groups and also use FCs as a fingerprint to predict cognitive abilities. Our experiments demonstrated higher accuracy of using DCL over other conventional models when classifying populations of different ages and cognitive scores. Moreover, DCL revealed that brain connections became stronger at adolescence stage. Furthermore, DCL detected strong correlations between default mode network and other networks which were overlooked by linear canonical correlation analysis, demonstrating DCL's ability of detecting nonlinear correlations. CONCLUSION: The results verified the superiority of DCL over conventional data-fusion methods. In addition, the stronger brain connection demonstrated the importance of adolescence stage for brain development. SIGNIFICANCE: DCL can better combine complex correlations between multiple data sets in addition to their fitting to phenotypes, with the potential to overcome the limitations of several current data-fusion models.

AB - OBJECTIVE: Multi-modal functional magnetic resonance imaging has been widely used for brain research. Conventional data-fusion methods cannot capture complex relationship (e.g., nonlinear predictive relationship) between multiple data. This paper aims to develop a neural network framework to extract phenotype related cross-data relationships and use it to study the brain development. METHODS: We propose a novel method, deep collaborative learning (DCL), to address the limitation of existing methods. DCL first uses a deep network to represent original data and then seeks their correlations, while also linking the data representation with phenotypical information. RESULTS: We studied the difference of functional connectivity (FCs) between different age groups and also use FCs as a fingerprint to predict cognitive abilities. Our experiments demonstrated higher accuracy of using DCL over other conventional models when classifying populations of different ages and cognitive scores. Moreover, DCL revealed that brain connections became stronger at adolescence stage. Furthermore, DCL detected strong correlations between default mode network and other networks which were overlooked by linear canonical correlation analysis, demonstrating DCL's ability of detecting nonlinear correlations. CONCLUSION: The results verified the superiority of DCL over conventional data-fusion methods. In addition, the stronger brain connection demonstrated the importance of adolescence stage for brain development. SIGNIFICANCE: DCL can better combine complex correlations between multiple data sets in addition to their fitting to phenotypes, with the potential to overcome the limitations of several current data-fusion models.

UR - http://www.scopus.com/inward/record.url?scp=85075805428&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85075805428&partnerID=8YFLogxK

U2 - 10.1109/TBME.2019.2904301

DO - 10.1109/TBME.2019.2904301

M3 - Article

C2 - 30872216

AN - SCOPUS:85075805428

VL - 66

SP - 3346

EP - 3359

JO - IEEE Transactions on Biomedical Engineering

JF - IEEE Transactions on Biomedical Engineering

SN - 0018-9294

IS - 12

ER -