RefineNet: An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis

Naresh Nandakumar; Komal Manzoor; Shruti Agarwal; Haris I. Sair; Archana Venkataraman

doi:10.1007/978-3-031-16431-6_30

RefineNet: An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis

Naresh Nandakumar, Komal Manzoor, Shruti Agarwal, Haris I. Sair, Archana Venkataraman

School of Medicine

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Parcellations used in resting-state fMRI (rs-fMRI) analyses are derived from group-level information, and thus ignore both subject-level functional differences and the downstream task. In this paper, we introduce RefineNet, a Bayesian-inspired deep network architecture that adjusts region boundaries based on individual functional connectivity profiles. RefineNet uses an iterative voxel reassignment procedure that considers neighborhood information while balancing temporal coherence of the refined parcellation. We validate RefineNet on rs-fMRI data from three different datasets, each one geared towards a different predictive task: (1) cognitive fluid intelligence prediction using the HCP dataset (regression), (2) autism versus control diagnosis using the ABIDE II dataset (classification), and (3) language localization using an rs-fMRI brain tumor dataset (segmentation). We demonstrate that RefineNet improves the performance of existing deep networks from the literature on each of these tasks. We also show that RefineNet produces anatomically meaningful subject-level parcellations with higher temporal coherence.

Original language	English (US)
Title of host publication	Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings
Editors	Linwei Wang, Qi Dou, P. Thomas Fletcher, Stefanie Speidel, Shuo Li
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	315-325
Number of pages	11
ISBN (Print)	9783031164309
DOIs	https://doi.org/10.1007/978-3-031-16431-6_30
State	Published - 2022
Event	25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 - Singapore, Singapore Duration: Sep 18 2022 → Sep 22 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13431 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
Country/Territory	Singapore
City	Singapore
Period	9/18/22 → 9/22/22

Keywords

Parcellation refinement
Rs-fMRI
Task optimization

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-16431-6_30

Cite this

Nandakumar, N., Manzoor, K., Agarwal, S., Sair, H. I., & Venkataraman, A. (2022). RefineNet: An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis. In L. Wang, Q. Dou, P. T. Fletcher, S. Speidel, & S. Li (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings (pp. 315-325). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13431 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-16431-6_30

RefineNet: An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis. / Nandakumar, Naresh; Manzoor, Komal; Agarwal, Shruti et al.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings. ed. / Linwei Wang; Qi Dou; P. Thomas Fletcher; Stefanie Speidel; Shuo Li. Springer Science and Business Media Deutschland GmbH, 2022. p. 315-325 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13431 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Nandakumar, N, Manzoor, K, Agarwal, S, Sair, HI & Venkataraman, A 2022, RefineNet: An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis. in L Wang, Q Dou, PT Fletcher, S Speidel & S Li (eds), Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13431 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 315-325, 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022, Singapore, Singapore, 9/18/22. https://doi.org/10.1007/978-3-031-16431-6_30

Nandakumar N, Manzoor K, Agarwal S, Sair HI, Venkataraman A. RefineNet: An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis. In Wang L, Dou Q, Fletcher PT, Speidel S, Li S, editors, Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 315-325. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-16431-6_30

Nandakumar, Naresh ; Manzoor, Komal ; Agarwal, Shruti et al. / RefineNet : An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis. Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings. editor / Linwei Wang ; Qi Dou ; P. Thomas Fletcher ; Stefanie Speidel ; Shuo Li. Springer Science and Business Media Deutschland GmbH, 2022. pp. 315-325 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{87bd10ebd98948d1a1d7edb594fa4abf,

title = "RefineNet: An Automated Framework to Generate Task and Subject-Specific Brain Parcellations for Resting-State fMRI Analysis",

abstract = "Parcellations used in resting-state fMRI (rs-fMRI) analyses are derived from group-level information, and thus ignore both subject-level functional differences and the downstream task. In this paper, we introduce RefineNet, a Bayesian-inspired deep network architecture that adjusts region boundaries based on individual functional connectivity profiles. RefineNet uses an iterative voxel reassignment procedure that considers neighborhood information while balancing temporal coherence of the refined parcellation. We validate RefineNet on rs-fMRI data from three different datasets, each one geared towards a different predictive task: (1) cognitive fluid intelligence prediction using the HCP dataset (regression), (2) autism versus control diagnosis using the ABIDE II dataset (classification), and (3) language localization using an rs-fMRI brain tumor dataset (segmentation). We demonstrate that RefineNet improves the performance of existing deep networks from the literature on each of these tasks. We also show that RefineNet produces anatomically meaningful subject-level parcellations with higher temporal coherence.",

keywords = "Parcellation refinement, Rs-fMRI, Task optimization",

author = "Naresh Nandakumar and Komal Manzoor and Shruti Agarwal and Sair, {Haris I.} and Archana Venkataraman",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 ; Conference date: 18-09-2022 Through 22-09-2022",

year = "2022",

doi = "10.1007/978-3-031-16431-6_30",

language = "English (US)",

isbn = "9783031164309",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "315--325",

editor = "Linwei Wang and Qi Dou and Fletcher, {P. Thomas} and Stefanie Speidel and Shuo Li",

booktitle = "Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings",

address = "Germany",

}

TY - GEN

T1 - RefineNet

T2 - 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022

AU - Nandakumar, Naresh

AU - Manzoor, Komal

AU - Agarwal, Shruti

AU - Sair, Haris I.

AU - Venkataraman, Archana

PY - 2022

Y1 - 2022

N2 - Parcellations used in resting-state fMRI (rs-fMRI) analyses are derived from group-level information, and thus ignore both subject-level functional differences and the downstream task. In this paper, we introduce RefineNet, a Bayesian-inspired deep network architecture that adjusts region boundaries based on individual functional connectivity profiles. RefineNet uses an iterative voxel reassignment procedure that considers neighborhood information while balancing temporal coherence of the refined parcellation. We validate RefineNet on rs-fMRI data from three different datasets, each one geared towards a different predictive task: (1) cognitive fluid intelligence prediction using the HCP dataset (regression), (2) autism versus control diagnosis using the ABIDE II dataset (classification), and (3) language localization using an rs-fMRI brain tumor dataset (segmentation). We demonstrate that RefineNet improves the performance of existing deep networks from the literature on each of these tasks. We also show that RefineNet produces anatomically meaningful subject-level parcellations with higher temporal coherence.

AB - Parcellations used in resting-state fMRI (rs-fMRI) analyses are derived from group-level information, and thus ignore both subject-level functional differences and the downstream task. In this paper, we introduce RefineNet, a Bayesian-inspired deep network architecture that adjusts region boundaries based on individual functional connectivity profiles. RefineNet uses an iterative voxel reassignment procedure that considers neighborhood information while balancing temporal coherence of the refined parcellation. We validate RefineNet on rs-fMRI data from three different datasets, each one geared towards a different predictive task: (1) cognitive fluid intelligence prediction using the HCP dataset (regression), (2) autism versus control diagnosis using the ABIDE II dataset (classification), and (3) language localization using an rs-fMRI brain tumor dataset (segmentation). We demonstrate that RefineNet improves the performance of existing deep networks from the literature on each of these tasks. We also show that RefineNet produces anatomically meaningful subject-level parcellations with higher temporal coherence.

KW - Parcellation refinement

KW - Rs-fMRI

KW - Task optimization

UR - http://www.scopus.com/inward/record.url?scp=85138817884&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85138817884&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-16431-6_30

DO - 10.1007/978-3-031-16431-6_30

M3 - Conference contribution

AN - SCOPUS:85138817884

SN - 9783031164309

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

SP - 315

EP - 325

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings

A2 - Wang, Linwei

A2 - Dou, Qi

A2 - Fletcher, P. Thomas

A2 - Speidel, Stefanie

A2 - Li, Shuo

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 18 September 2022 through 22 September 2022

ER -