Analysis of dynamic brain connectivity through geodesic clustering

A. Yamin, M. Dayan, L. Squarcina, P. Brambilla, V. Murino, V. Diwadkar, D. Sona

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

Abstract

Analysis of dynamic functional connectivity allows for studying the time variant behavior of brain connectivity during specific tasks or at rest. There is, however, a debate around the significance of studies analyzing the dynamic connectivity, as it is usually estimated using short subsequences of the entire time-series. Therefore, a question that naturally arises is whether the dynamic connectivity information is robust enough to compare connectivity matrices. In this paper we investigate the importance of the choice of metric on the space of graphs to answer this question, using a dataset of twins under the assumption that twins connectivity is more similar than in any other pair of unrelated subjects. Specifically, the problem was formulated as a classification task between twin and non-twin pairs. The approach described in the paper relies on geodesic clustering of dynamic connectivity matrices to find a subset of brain states, which were then used to encode the pairwise connectivity similarities between subjects. Experiments were performed to compare the use of Euclidean distance in a vectorial space and a geodesic distance in the Riemannian space of symmetric positive definite matrices. We showed that the geodesic distance provided a better classification of twins subjects, suggesting this use of this distance can robustly compare dynamic connectivity matrices.

Original languageEnglish
Title of host publicationImage Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings
EditorsElisa Ricci, Nicu Sebe, Samuel Rota Bulò, Cees Snoek, Oswald Lanz, Stefano Messelodi
PublisherSpringer Verlag
Pages640-648
Number of pages9
ISBN (Print)9783030306441
DOIs
Publication statusPublished - Jan 1 2019
Event20th International Conference on Image Analysis and Processing, ICIAP 2019 - Trento, Italy
Duration: Sep 9 2019Sep 13 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11752 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference20th International Conference on Image Analysis and Processing, ICIAP 2019
CountryItaly
CityTrento
Period9/9/199/13/19

Fingerprint

Geodesic
Brain
Connectivity
Clustering
Geodesic Distance
Time series
Symmetric Positive Definite Matrix
Euclidean Distance
Subsequence
Pairwise
Entire
Experiments
Metric
Subset
Graph in graph theory
Experiment

Keywords

  • Connectomes
  • Dynamic functional connectivity
  • Geodesic clustering
  • SVM
  • Symmetric positive definite matrices
  • Task-based fMRI

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Yamin, A., Dayan, M., Squarcina, L., Brambilla, P., Murino, V., Diwadkar, V., & Sona, D. (2019). Analysis of dynamic brain connectivity through geodesic clustering. In E. Ricci, N. Sebe, S. Rota Bulò, C. Snoek, O. Lanz, & S. Messelodi (Eds.), Image Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings (pp. 640-648). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11752 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-30645-8_58

Analysis of dynamic brain connectivity through geodesic clustering. / Yamin, A.; Dayan, M.; Squarcina, L.; Brambilla, P.; Murino, V.; Diwadkar, V.; Sona, D.

Image Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings. ed. / Elisa Ricci; Nicu Sebe; Samuel Rota Bulò; Cees Snoek; Oswald Lanz; Stefano Messelodi. Springer Verlag, 2019. p. 640-648 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11752 LNCS).

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

Yamin, A, Dayan, M, Squarcina, L, Brambilla, P, Murino, V, Diwadkar, V & Sona, D 2019, Analysis of dynamic brain connectivity through geodesic clustering. in E Ricci, N Sebe, S Rota Bulò, C Snoek, O Lanz & S Messelodi (eds), Image Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11752 LNCS, Springer Verlag, pp. 640-648, 20th International Conference on Image Analysis and Processing, ICIAP 2019, Trento, Italy, 9/9/19. https://doi.org/10.1007/978-3-030-30645-8_58
Yamin A, Dayan M, Squarcina L, Brambilla P, Murino V, Diwadkar V et al. Analysis of dynamic brain connectivity through geodesic clustering. In Ricci E, Sebe N, Rota Bulò S, Snoek C, Lanz O, Messelodi S, editors, Image Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings. Springer Verlag. 2019. p. 640-648. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-30645-8_58
Yamin, A. ; Dayan, M. ; Squarcina, L. ; Brambilla, P. ; Murino, V. ; Diwadkar, V. ; Sona, D. / Analysis of dynamic brain connectivity through geodesic clustering. Image Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings. editor / Elisa Ricci ; Nicu Sebe ; Samuel Rota Bulò ; Cees Snoek ; Oswald Lanz ; Stefano Messelodi. Springer Verlag, 2019. pp. 640-648 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{0ba025a581904b0a9a0cd709654411d7,
title = "Analysis of dynamic brain connectivity through geodesic clustering",
abstract = "Analysis of dynamic functional connectivity allows for studying the time variant behavior of brain connectivity during specific tasks or at rest. There is, however, a debate around the significance of studies analyzing the dynamic connectivity, as it is usually estimated using short subsequences of the entire time-series. Therefore, a question that naturally arises is whether the dynamic connectivity information is robust enough to compare connectivity matrices. In this paper we investigate the importance of the choice of metric on the space of graphs to answer this question, using a dataset of twins under the assumption that twins connectivity is more similar than in any other pair of unrelated subjects. Specifically, the problem was formulated as a classification task between twin and non-twin pairs. The approach described in the paper relies on geodesic clustering of dynamic connectivity matrices to find a subset of brain states, which were then used to encode the pairwise connectivity similarities between subjects. Experiments were performed to compare the use of Euclidean distance in a vectorial space and a geodesic distance in the Riemannian space of symmetric positive definite matrices. We showed that the geodesic distance provided a better classification of twins subjects, suggesting this use of this distance can robustly compare dynamic connectivity matrices.",
keywords = "Connectomes, Dynamic functional connectivity, Geodesic clustering, SVM, Symmetric positive definite matrices, Task-based fMRI",
author = "A. Yamin and M. Dayan and L. Squarcina and P. Brambilla and V. Murino and V. Diwadkar and D. Sona",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-30645-8_58",
language = "English",
isbn = "9783030306441",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "640--648",
editor = "Elisa Ricci and Nicu Sebe and {Rota Bul{\`o}}, Samuel and Cees Snoek and Oswald Lanz and Stefano Messelodi",
booktitle = "Image Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings",
address = "Germany",

}

TY - GEN

T1 - Analysis of dynamic brain connectivity through geodesic clustering

AU - Yamin, A.

AU - Dayan, M.

AU - Squarcina, L.

AU - Brambilla, P.

AU - Murino, V.

AU - Diwadkar, V.

AU - Sona, D.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Analysis of dynamic functional connectivity allows for studying the time variant behavior of brain connectivity during specific tasks or at rest. There is, however, a debate around the significance of studies analyzing the dynamic connectivity, as it is usually estimated using short subsequences of the entire time-series. Therefore, a question that naturally arises is whether the dynamic connectivity information is robust enough to compare connectivity matrices. In this paper we investigate the importance of the choice of metric on the space of graphs to answer this question, using a dataset of twins under the assumption that twins connectivity is more similar than in any other pair of unrelated subjects. Specifically, the problem was formulated as a classification task between twin and non-twin pairs. The approach described in the paper relies on geodesic clustering of dynamic connectivity matrices to find a subset of brain states, which were then used to encode the pairwise connectivity similarities between subjects. Experiments were performed to compare the use of Euclidean distance in a vectorial space and a geodesic distance in the Riemannian space of symmetric positive definite matrices. We showed that the geodesic distance provided a better classification of twins subjects, suggesting this use of this distance can robustly compare dynamic connectivity matrices.

AB - Analysis of dynamic functional connectivity allows for studying the time variant behavior of brain connectivity during specific tasks or at rest. There is, however, a debate around the significance of studies analyzing the dynamic connectivity, as it is usually estimated using short subsequences of the entire time-series. Therefore, a question that naturally arises is whether the dynamic connectivity information is robust enough to compare connectivity matrices. In this paper we investigate the importance of the choice of metric on the space of graphs to answer this question, using a dataset of twins under the assumption that twins connectivity is more similar than in any other pair of unrelated subjects. Specifically, the problem was formulated as a classification task between twin and non-twin pairs. The approach described in the paper relies on geodesic clustering of dynamic connectivity matrices to find a subset of brain states, which were then used to encode the pairwise connectivity similarities between subjects. Experiments were performed to compare the use of Euclidean distance in a vectorial space and a geodesic distance in the Riemannian space of symmetric positive definite matrices. We showed that the geodesic distance provided a better classification of twins subjects, suggesting this use of this distance can robustly compare dynamic connectivity matrices.

KW - Connectomes

KW - Dynamic functional connectivity

KW - Geodesic clustering

KW - SVM

KW - Symmetric positive definite matrices

KW - Task-based fMRI

UR - http://www.scopus.com/inward/record.url?scp=85072902321&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85072902321&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-30645-8_58

DO - 10.1007/978-3-030-30645-8_58

M3 - Conference contribution

AN - SCOPUS:85072902321

SN - 9783030306441

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

SP - 640

EP - 648

BT - Image Analysis and Processing – ICIAP 2019 - 20th International Conference, Proceedings

A2 - Ricci, Elisa

A2 - Sebe, Nicu

A2 - Rota Bulò, Samuel

A2 - Snoek, Cees

A2 - Lanz, Oswald

A2 - Messelodi, Stefano

PB - Springer Verlag

ER -