Extracting orthogonal subject- and condition-specific signatures from fMRI data using whole-brain effective connectivity

Vicente Pallarés, Andrea Insabato, Ana Sanjuán, Simone Kühn, Dante Mantini, Gustavo Deco, Matthieu Gilson

Research output: Contribution to journalArticle

Abstract

The study of brain communication based on fMRI data is often limited because such measurements are a mixture of session-to-session variability with subject- and condition-related information. Disentangling these contributions is crucial for real-life applications, in particular when only a few recording sessions are available. The present study aims to define a reliable standard for the extraction of multiple signatures from fMRI data, while verifying that they do not mix information about the different modalities (e.g., subjects and conditions such as tasks performed by them). In particular, condition-specific signatures should not be contaminated by subject-related information, since they aim to generalize over subjects. Practically, signatures correspond to subnetworks of directed interactions between brain regions (typically 100 covering the whole brain) supporting the subject and condition identification for single fMRI sessions. The key for robust prediction is using effective connectivity instead of functional connectivity. Our method demonstrates excellent generalization capabilities for subject identification in two datasets, using only a few sessions per subject as reference. Using another dataset with resting state and movie viewing, we show that the two signatures related to subjects and tasks correspond to distinct subnetworks, which are thus topologically orthogonal. Our results set solid foundations for applications tailored to individual subjects, such as clinical diagnostic.

Original languageEnglish
Pages (from-to)238-254
Number of pages17
JournalNeuroImage
Volume178
DOIs
Publication statusPublished - Sep 1 2018

Fingerprint

Magnetic Resonance Imaging
Brain
Motion Pictures
Datasets

ASJC Scopus subject areas

  • Neurology
  • Cognitive Neuroscience

Cite this

Extracting orthogonal subject- and condition-specific signatures from fMRI data using whole-brain effective connectivity. / Pallarés, Vicente; Insabato, Andrea; Sanjuán, Ana; Kühn, Simone; Mantini, Dante; Deco, Gustavo; Gilson, Matthieu.

In: NeuroImage, Vol. 178, 01.09.2018, p. 238-254.

Research output: Contribution to journalArticle

Pallarés, Vicente ; Insabato, Andrea ; Sanjuán, Ana ; Kühn, Simone ; Mantini, Dante ; Deco, Gustavo ; Gilson, Matthieu. / Extracting orthogonal subject- and condition-specific signatures from fMRI data using whole-brain effective connectivity. In: NeuroImage. 2018 ; Vol. 178. pp. 238-254.
@article{a11feeb618e84cb5ac0cf9b26f9fc8b3,
title = "Extracting orthogonal subject- and condition-specific signatures from fMRI data using whole-brain effective connectivity",
abstract = "The study of brain communication based on fMRI data is often limited because such measurements are a mixture of session-to-session variability with subject- and condition-related information. Disentangling these contributions is crucial for real-life applications, in particular when only a few recording sessions are available. The present study aims to define a reliable standard for the extraction of multiple signatures from fMRI data, while verifying that they do not mix information about the different modalities (e.g., subjects and conditions such as tasks performed by them). In particular, condition-specific signatures should not be contaminated by subject-related information, since they aim to generalize over subjects. Practically, signatures correspond to subnetworks of directed interactions between brain regions (typically 100 covering the whole brain) supporting the subject and condition identification for single fMRI sessions. The key for robust prediction is using effective connectivity instead of functional connectivity. Our method demonstrates excellent generalization capabilities for subject identification in two datasets, using only a few sessions per subject as reference. Using another dataset with resting state and movie viewing, we show that the two signatures related to subjects and tasks correspond to distinct subnetworks, which are thus topologically orthogonal. Our results set solid foundations for applications tailored to individual subjects, such as clinical diagnostic.",
author = "Vicente Pallar{\'e}s and Andrea Insabato and Ana Sanju{\'a}n and Simone K{\"u}hn and Dante Mantini and Gustavo Deco and Matthieu Gilson",
year = "2018",
month = "9",
day = "1",
doi = "10.1016/j.neuroimage.2018.04.070",
language = "English",
volume = "178",
pages = "238--254",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Extracting orthogonal subject- and condition-specific signatures from fMRI data using whole-brain effective connectivity

AU - Pallarés, Vicente

AU - Insabato, Andrea

AU - Sanjuán, Ana

AU - Kühn, Simone

AU - Mantini, Dante

AU - Deco, Gustavo

AU - Gilson, Matthieu

PY - 2018/9/1

Y1 - 2018/9/1

N2 - The study of brain communication based on fMRI data is often limited because such measurements are a mixture of session-to-session variability with subject- and condition-related information. Disentangling these contributions is crucial for real-life applications, in particular when only a few recording sessions are available. The present study aims to define a reliable standard for the extraction of multiple signatures from fMRI data, while verifying that they do not mix information about the different modalities (e.g., subjects and conditions such as tasks performed by them). In particular, condition-specific signatures should not be contaminated by subject-related information, since they aim to generalize over subjects. Practically, signatures correspond to subnetworks of directed interactions between brain regions (typically 100 covering the whole brain) supporting the subject and condition identification for single fMRI sessions. The key for robust prediction is using effective connectivity instead of functional connectivity. Our method demonstrates excellent generalization capabilities for subject identification in two datasets, using only a few sessions per subject as reference. Using another dataset with resting state and movie viewing, we show that the two signatures related to subjects and tasks correspond to distinct subnetworks, which are thus topologically orthogonal. Our results set solid foundations for applications tailored to individual subjects, such as clinical diagnostic.

AB - The study of brain communication based on fMRI data is often limited because such measurements are a mixture of session-to-session variability with subject- and condition-related information. Disentangling these contributions is crucial for real-life applications, in particular when only a few recording sessions are available. The present study aims to define a reliable standard for the extraction of multiple signatures from fMRI data, while verifying that they do not mix information about the different modalities (e.g., subjects and conditions such as tasks performed by them). In particular, condition-specific signatures should not be contaminated by subject-related information, since they aim to generalize over subjects. Practically, signatures correspond to subnetworks of directed interactions between brain regions (typically 100 covering the whole brain) supporting the subject and condition identification for single fMRI sessions. The key for robust prediction is using effective connectivity instead of functional connectivity. Our method demonstrates excellent generalization capabilities for subject identification in two datasets, using only a few sessions per subject as reference. Using another dataset with resting state and movie viewing, we show that the two signatures related to subjects and tasks correspond to distinct subnetworks, which are thus topologically orthogonal. Our results set solid foundations for applications tailored to individual subjects, such as clinical diagnostic.

UR - http://www.scopus.com/inward/record.url?scp=85047449032&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047449032&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2018.04.070

DO - 10.1016/j.neuroimage.2018.04.070

M3 - Article

C2 - 29753842

AN - SCOPUS:85047449032

VL - 178

SP - 238

EP - 254

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -