A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces

Niharika Shimona D’Souza, Mary Beth Nebel, Nicholas Wymbs, Stewart H Mostofsky, Archana Venkataraman

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

Abstract

The problem of linking functional connectomics to behavior is extremely challenging due to the complex interactions between the two distinct, but related, data domains. We propose a coupled manifold optimization framework which projects fMRI data onto a low dimensional matrix manifold common to the cohort. The patient specific loadings simultaneously map onto a behavioral measure of interest via a second, non-linear, manifold. By leveraging the kernel trick, we can optimize over a potentially infinite dimensional space without explicitly computing the embeddings. As opposed to conventional manifold learning, which assumes a fixed input representation, our framework directly optimizes for embedding directions that predict behavior. Our optimization algorithm combines proximal gradient descent with the trust region method, which has good convergence guarantees. We validate our framework on resting state fMRI from fifty-eight patients with Autism Spectrum Disorder using three distinct measures of clinical severity. Our method outperforms traditional representation learning techniques in a cross validated setting, thus demonstrating the predictive power of our coupled objective.

Original languageEnglish (US)
Title of host publicationInformation Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings
EditorsSiqi Bao, Albert C.S. Chung, James C. Gee, Paul A. Yushkevich
PublisherSpringer Verlag
Pages605-616
Number of pages12
ISBN (Print)9783030203504
DOIs
StatePublished - Jan 1 2019
Event26th International Conference on Information Processing in Medical Imaging, IPMI 2019 - Hong Kong, China
Duration: Jun 2 2019Jun 7 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11492 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference26th International Conference on Information Processing in Medical Imaging, IPMI 2019
CountryChina
CityHong Kong
Period6/2/196/7/19

Fingerprint

Functional Magnetic Resonance Imaging
Optimization
Optimise
Distinct
Manifold Learning
Trust Region Method
Infinite-dimensional Spaces
Gradient Descent
Linking
Disorder
Optimization Algorithm
Model
kernel
Predict
Computing
Interaction
Framework
Magnetic Resonance Imaging
Learning

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

D’Souza, N. S., Nebel, M. B., Wymbs, N., Mostofsky, S. H., & Venkataraman, A. (2019). A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces. In S. Bao, A. C. S. Chung, J. C. Gee, & P. A. Yushkevich (Eds.), Information Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings (pp. 605-616). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11492 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-20351-1_47

A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces. / D’Souza, Niharika Shimona; Nebel, Mary Beth; Wymbs, Nicholas; Mostofsky, Stewart H; Venkataraman, Archana.

Information Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings. ed. / Siqi Bao; Albert C.S. Chung; James C. Gee; Paul A. Yushkevich. Springer Verlag, 2019. p. 605-616 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11492 LNCS).

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

D’Souza, NS, Nebel, MB, Wymbs, N, Mostofsky, SH & Venkataraman, A 2019, A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces. in S Bao, ACS Chung, JC Gee & PA Yushkevich (eds), Information Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11492 LNCS, Springer Verlag, pp. 605-616, 26th International Conference on Information Processing in Medical Imaging, IPMI 2019, Hong Kong, China, 6/2/19. https://doi.org/10.1007/978-3-030-20351-1_47
D’Souza NS, Nebel MB, Wymbs N, Mostofsky SH, Venkataraman A. A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces. In Bao S, Chung ACS, Gee JC, Yushkevich PA, editors, Information Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings. Springer Verlag. 2019. p. 605-616. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-20351-1_47
D’Souza, Niharika Shimona ; Nebel, Mary Beth ; Wymbs, Nicholas ; Mostofsky, Stewart H ; Venkataraman, Archana. / A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces. Information Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings. editor / Siqi Bao ; Albert C.S. Chung ; James C. Gee ; Paul A. Yushkevich. Springer Verlag, 2019. pp. 605-616 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{041ebdc8163c43c9ab70a528f9455959,
title = "A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces",
abstract = "The problem of linking functional connectomics to behavior is extremely challenging due to the complex interactions between the two distinct, but related, data domains. We propose a coupled manifold optimization framework which projects fMRI data onto a low dimensional matrix manifold common to the cohort. The patient specific loadings simultaneously map onto a behavioral measure of interest via a second, non-linear, manifold. By leveraging the kernel trick, we can optimize over a potentially infinite dimensional space without explicitly computing the embeddings. As opposed to conventional manifold learning, which assumes a fixed input representation, our framework directly optimizes for embedding directions that predict behavior. Our optimization algorithm combines proximal gradient descent with the trust region method, which has good convergence guarantees. We validate our framework on resting state fMRI from fifty-eight patients with Autism Spectrum Disorder using three distinct measures of clinical severity. Our method outperforms traditional representation learning techniques in a cross validated setting, thus demonstrating the predictive power of our coupled objective.",
author = "D’Souza, {Niharika Shimona} and Nebel, {Mary Beth} and Nicholas Wymbs and Mostofsky, {Stewart H} and Archana Venkataraman",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-20351-1_47",
language = "English (US)",
isbn = "9783030203504",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "605--616",
editor = "Siqi Bao and Chung, {Albert C.S.} and Gee, {James C.} and Yushkevich, {Paul A.}",
booktitle = "Information Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings",

}

TY - GEN

T1 - A Coupled Manifold Optimization Framework to Jointly Model the Functional Connectomics and Behavioral Data Spaces

AU - D’Souza, Niharika Shimona

AU - Nebel, Mary Beth

AU - Wymbs, Nicholas

AU - Mostofsky, Stewart H

AU - Venkataraman, Archana

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The problem of linking functional connectomics to behavior is extremely challenging due to the complex interactions between the two distinct, but related, data domains. We propose a coupled manifold optimization framework which projects fMRI data onto a low dimensional matrix manifold common to the cohort. The patient specific loadings simultaneously map onto a behavioral measure of interest via a second, non-linear, manifold. By leveraging the kernel trick, we can optimize over a potentially infinite dimensional space without explicitly computing the embeddings. As opposed to conventional manifold learning, which assumes a fixed input representation, our framework directly optimizes for embedding directions that predict behavior. Our optimization algorithm combines proximal gradient descent with the trust region method, which has good convergence guarantees. We validate our framework on resting state fMRI from fifty-eight patients with Autism Spectrum Disorder using three distinct measures of clinical severity. Our method outperforms traditional representation learning techniques in a cross validated setting, thus demonstrating the predictive power of our coupled objective.

AB - The problem of linking functional connectomics to behavior is extremely challenging due to the complex interactions between the two distinct, but related, data domains. We propose a coupled manifold optimization framework which projects fMRI data onto a low dimensional matrix manifold common to the cohort. The patient specific loadings simultaneously map onto a behavioral measure of interest via a second, non-linear, manifold. By leveraging the kernel trick, we can optimize over a potentially infinite dimensional space without explicitly computing the embeddings. As opposed to conventional manifold learning, which assumes a fixed input representation, our framework directly optimizes for embedding directions that predict behavior. Our optimization algorithm combines proximal gradient descent with the trust region method, which has good convergence guarantees. We validate our framework on resting state fMRI from fifty-eight patients with Autism Spectrum Disorder using three distinct measures of clinical severity. Our method outperforms traditional representation learning techniques in a cross validated setting, thus demonstrating the predictive power of our coupled objective.

UR - http://www.scopus.com/inward/record.url?scp=85066114568&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066114568&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-20351-1_47

DO - 10.1007/978-3-030-20351-1_47

M3 - Conference contribution

SN - 9783030203504

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

SP - 605

EP - 616

BT - Information Processing in Medical Imaging - 26th International Conference, IPMI 2019, Proceedings

A2 - Bao, Siqi

A2 - Chung, Albert C.S.

A2 - Gee, James C.

A2 - Yushkevich, Paul A.

PB - Springer Verlag

ER -