Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks

Roberto C. Sotero; Jose M. Sanchez-Bornot

doi:10.1007/978-3-031-48232-8_21

Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks

Roberto C. Sotero, Jose M. Sanchez-Bornot

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

2 Downloads (Pure)

Abstract

Community detection in complex signed networks is a significant challenge, traditionally addressed using the Louvain method directly applied to the correlation matrix. This study introduces a two-tier approach that integrates a Hebbian learning rule within an adaptive signed random walk (ASRW) framework, then applies the Louvain method to the final weight matrix. This approach refines the network analysis process, providing a new tool for exploring community structure. Tested extensively on synthetic signed networks with defined community structures, our methodology consistently outperformed the traditional Louvain approach, particularly when communities were less clearly demarcated. Further application to resting-state functional MRI data from the ABIDE Preprocessed Initiative highlighted functional connectivity differences between neurotypical individuals and those diagnosed with Autism Spectrum Disorder (ASD). Our approach found key areas of significant difference, including several cerebellum regions, consistent with existing ASD literature. Our findings underscore the potential of the proposed technique to advance community detection in correlation-based networks.

Original language	English
Title of host publication	Lecture Notes in Computer Science
Subtitle of host publication	Intelligent Data Engineering and Automated Learning – IDEAL 2023
Editors	Paulo Quaresma, Teresa Gonçalves, David Camacho, Hujun Yin, Vicente Julian, Antonio J. Tallón-Ballesteros
Publisher	Springer Nature
Pages	222–232
Number of pages	11
Volume	14404
ISBN (Electronic)	978-3-031-48232-8
ISBN (Print)	9783031482311
DOIs	https://doi.org/10.1007/978-3-031-48232-8_21
Publication status	Published online - 15 Nov 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14404 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Bibliographical note

Funding Information:
This work was supported by Grant 222300868 from the Alberta Innovates LevMax program, and by RGPIN-2022-03042 from Natural Sciences and Engineering Research Council of Canada.

Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

Keywords

Community Detection
Hebbian Learning
Random Walks
Brain Networks
Autism Spectrum Disorder

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1007/978-3-031-48232-8_21

Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain NetworksAuthor Accepted Manuscript, 2.39 MB

Cite this

Sotero, R. C., & Sanchez-Bornot, J. M. (2023). Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks. In P. Quaresma, T. Gonçalves, D. Camacho, H. Yin, V. Julian, & A. J. Tallón-Ballesteros (Eds.), Lecture Notes in Computer Science: Intelligent Data Engineering and Automated Learning – IDEAL 2023 (Vol. 14404, pp. 222–232). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14404 LNCS). Springer Nature. Advance online publication. https://doi.org/10.1007/978-3-031-48232-8_21

Sotero, Roberto C. ; Sanchez-Bornot, Jose M. / Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks. Lecture Notes in Computer Science: Intelligent Data Engineering and Automated Learning – IDEAL 2023. editor / Paulo Quaresma ; Teresa Gonçalves ; David Camacho ; Hujun Yin ; Vicente Julian ; Antonio J. Tallón-Ballesteros. Vol. 14404 Springer Nature, 2023. pp. 222–232 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inbook{9daeb118b1c64b20ac70d13ceb93da0f,

title = "Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks",

abstract = "Community detection in complex signed networks is a significant challenge, traditionally addressed using the Louvain method directly applied to the correlation matrix. This study introduces a two-tier approach that integrates a Hebbian learning rule within an adaptive signed random walk (ASRW) framework, then applies the Louvain method to the final weight matrix. This approach refines the network analysis process, providing a new tool for exploring community structure. Tested extensively on synthetic signed networks with defined community structures, our methodology consistently outperformed the traditional Louvain approach, particularly when communities were less clearly demarcated. Further application to resting-state functional MRI data from the ABIDE Preprocessed Initiative highlighted functional connectivity differences between neurotypical individuals and those diagnosed with Autism Spectrum Disorder (ASD). Our approach found key areas of significant difference, including several cerebellum regions, consistent with existing ASD literature. Our findings underscore the potential of the proposed technique to advance community detection in correlation-based networks.",

keywords = "Community Detection, Hebbian Learning, Random Walks, Brain Networks, Autism Spectrum Disorder",

author = "Sotero, {Roberto C.} and Sanchez-Bornot, {Jose M.}",

note = "Funding Information: This work was supported by Grant 222300868 from the Alberta Innovates LevMax program, and by RGPIN-2022-03042 from Natural Sciences and Engineering Research Council of Canada. Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.",

year = "2023",

month = nov,

day = "15",

doi = "10.1007/978-3-031-48232-8_21",

language = "English",

isbn = "9783031482311",

volume = "14404",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Nature",

pages = "222–232",

editor = "Paulo Quaresma and Teresa Gon{\c c}alves and David Camacho and Hujun Yin and Vicente Julian and Tall{\'o}n-Ballesteros, {Antonio J.}",

booktitle = "Lecture Notes in Computer Science",

address = "Switzerland",

}

Sotero, RC & Sanchez-Bornot, JM 2023, Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks. in P Quaresma, T Gonçalves, D Camacho, H Yin, V Julian & AJ Tallón-Ballesteros (eds), Lecture Notes in Computer Science: Intelligent Data Engineering and Automated Learning – IDEAL 2023. vol. 14404, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14404 LNCS, Springer Nature, pp. 222–232. https://doi.org/10.1007/978-3-031-48232-8_21

Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks. / Sotero, Roberto C.; Sanchez-Bornot, Jose M.
Lecture Notes in Computer Science: Intelligent Data Engineering and Automated Learning – IDEAL 2023. ed. / Paulo Quaresma; Teresa Gonçalves; David Camacho; Hujun Yin; Vicente Julian; Antonio J. Tallón-Ballesteros. Vol. 14404 Springer Nature, 2023. p. 222–232 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14404 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

TY - CHAP

T1 - Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks

AU - Sotero, Roberto C.

AU - Sanchez-Bornot, Jose M.

N1 - Funding Information: This work was supported by Grant 222300868 from the Alberta Innovates LevMax program, and by RGPIN-2022-03042 from Natural Sciences and Engineering Research Council of Canada. Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

PY - 2023/11/15

Y1 - 2023/11/15

N2 - Community detection in complex signed networks is a significant challenge, traditionally addressed using the Louvain method directly applied to the correlation matrix. This study introduces a two-tier approach that integrates a Hebbian learning rule within an adaptive signed random walk (ASRW) framework, then applies the Louvain method to the final weight matrix. This approach refines the network analysis process, providing a new tool for exploring community structure. Tested extensively on synthetic signed networks with defined community structures, our methodology consistently outperformed the traditional Louvain approach, particularly when communities were less clearly demarcated. Further application to resting-state functional MRI data from the ABIDE Preprocessed Initiative highlighted functional connectivity differences between neurotypical individuals and those diagnosed with Autism Spectrum Disorder (ASD). Our approach found key areas of significant difference, including several cerebellum regions, consistent with existing ASD literature. Our findings underscore the potential of the proposed technique to advance community detection in correlation-based networks.

AB - Community detection in complex signed networks is a significant challenge, traditionally addressed using the Louvain method directly applied to the correlation matrix. This study introduces a two-tier approach that integrates a Hebbian learning rule within an adaptive signed random walk (ASRW) framework, then applies the Louvain method to the final weight matrix. This approach refines the network analysis process, providing a new tool for exploring community structure. Tested extensively on synthetic signed networks with defined community structures, our methodology consistently outperformed the traditional Louvain approach, particularly when communities were less clearly demarcated. Further application to resting-state functional MRI data from the ABIDE Preprocessed Initiative highlighted functional connectivity differences between neurotypical individuals and those diagnosed with Autism Spectrum Disorder (ASD). Our approach found key areas of significant difference, including several cerebellum regions, consistent with existing ASD literature. Our findings underscore the potential of the proposed technique to advance community detection in correlation-based networks.

KW - Community Detection

KW - Hebbian Learning

KW - Random Walks

KW - Brain Networks

KW - Autism Spectrum Disorder

UR - https://doi.org/10.1007/978-3-031-48232-8_21

UR - http://www.scopus.com/inward/record.url?scp=85177830317&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-48232-8_21

DO - 10.1007/978-3-031-48232-8_21

M3 - Chapter

SN - 9783031482311

VL - 14404

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

SP - 222

EP - 232

BT - Lecture Notes in Computer Science

A2 - Quaresma, Paulo

A2 - Gonçalves, Teresa

A2 - Camacho, David

A2 - Yin, Hujun

A2 - Julian, Vicente

A2 - Tallón-Ballesteros, Antonio J.

PB - Springer Nature

ER -

Sotero RC, Sanchez-Bornot JM. Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks. In Quaresma P, Gonçalves T, Camacho D, Yin H, Julian V, Tallón-Ballesteros AJ, editors, Lecture Notes in Computer Science: Intelligent Data Engineering and Automated Learning – IDEAL 2023. Vol. 14404. Springer Nature. 2023. p. 222–232. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Epub 2023 Nov 15. doi: 10.1007/978-3-031-48232-8_21

Hebbian Learning-Guided Random Walks for Enhanced Community Detection in Correlation-Based Brain Networks

Abstract

Publication series

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this