Spatio-Temporal Dynamics of Intrinsic Networks in Functional Magnetic Imaging Data Using Recurrent Neural Networks

R. Devon Hjelm, Eswar Damaraju, Kyunghyun Cho, Helmut Laufs, Sergey M. Plis, Vince D. Calhoun

Research output: Contribution to journalArticle

Abstract

We introduce a novel recurrent neural network (RNN) approach to account for temporal dynamics and dependencies in brain networks observed via functional magnetic resonance imaging (fMRI). Our approach directly parameterizes temporal dynamics through recurrent connections, which can be used to formulate blind source separation with a conditional (rather than marginal) independence assumption, which we call RNN-ICA. This formulation enables us to visualize the temporal dynamics of both first order (activity) and second order (directed connectivity) information in brain networks that are widely studied in a static sense, but not well-characterized dynamically. RNN-ICA predicts dynamics directly from the recurrent states of the RNN in both task and resting state fMRI. Our results show both task-related and group-differentiating directed connectivity.

Original languageEnglish (US)
Article number00600
JournalFrontiers in Neuroscience
Volume12
Issue numberSEP
DOIs
StatePublished - Sep 20 2018

Fingerprint

Magnetic Resonance Imaging
Brain

Keywords

  • Aod
  • Deep learning
  • Fmri
  • Ica
  • Neuroimaging methods
  • Resting-state fmri
  • Rnn

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

Spatio-Temporal Dynamics of Intrinsic Networks in Functional Magnetic Imaging Data Using Recurrent Neural Networks. / Hjelm, R. Devon; Damaraju, Eswar; Cho, Kyunghyun; Laufs, Helmut; Plis, Sergey M.; Calhoun, Vince D.

In: Frontiers in Neuroscience, Vol. 12, No. SEP, 00600, 20.09.2018.

Research output: Contribution to journalArticle

Hjelm, R. Devon ; Damaraju, Eswar ; Cho, Kyunghyun ; Laufs, Helmut ; Plis, Sergey M. ; Calhoun, Vince D. / Spatio-Temporal Dynamics of Intrinsic Networks in Functional Magnetic Imaging Data Using Recurrent Neural Networks. In: Frontiers in Neuroscience. 2018 ; Vol. 12, No. SEP.
@article{73bb0723f3284800a03576d769a01b3a,
title = "Spatio-Temporal Dynamics of Intrinsic Networks in Functional Magnetic Imaging Data Using Recurrent Neural Networks",
abstract = "We introduce a novel recurrent neural network (RNN) approach to account for temporal dynamics and dependencies in brain networks observed via functional magnetic resonance imaging (fMRI). Our approach directly parameterizes temporal dynamics through recurrent connections, which can be used to formulate blind source separation with a conditional (rather than marginal) independence assumption, which we call RNN-ICA. This formulation enables us to visualize the temporal dynamics of both first order (activity) and second order (directed connectivity) information in brain networks that are widely studied in a static sense, but not well-characterized dynamically. RNN-ICA predicts dynamics directly from the recurrent states of the RNN in both task and resting state fMRI. Our results show both task-related and group-differentiating directed connectivity.",
keywords = "Aod, Deep learning, Fmri, Ica, Neuroimaging methods, Resting-state fmri, Rnn",
author = "Hjelm, {R. Devon} and Eswar Damaraju and Kyunghyun Cho and Helmut Laufs and Plis, {Sergey M.} and Calhoun, {Vince D.}",
year = "2018",
month = "9",
day = "20",
doi = "10.3389/fnins.2018.00600",
language = "English (US)",
volume = "12",
journal = "Frontiers in Neuroscience",
issn = "1662-4548",
publisher = "Frontiers Research Foundation",
number = "SEP",

}

TY - JOUR

T1 - Spatio-Temporal Dynamics of Intrinsic Networks in Functional Magnetic Imaging Data Using Recurrent Neural Networks

AU - Hjelm, R. Devon

AU - Damaraju, Eswar

AU - Cho, Kyunghyun

AU - Laufs, Helmut

AU - Plis, Sergey M.

AU - Calhoun, Vince D.

PY - 2018/9/20

Y1 - 2018/9/20

N2 - We introduce a novel recurrent neural network (RNN) approach to account for temporal dynamics and dependencies in brain networks observed via functional magnetic resonance imaging (fMRI). Our approach directly parameterizes temporal dynamics through recurrent connections, which can be used to formulate blind source separation with a conditional (rather than marginal) independence assumption, which we call RNN-ICA. This formulation enables us to visualize the temporal dynamics of both first order (activity) and second order (directed connectivity) information in brain networks that are widely studied in a static sense, but not well-characterized dynamically. RNN-ICA predicts dynamics directly from the recurrent states of the RNN in both task and resting state fMRI. Our results show both task-related and group-differentiating directed connectivity.

AB - We introduce a novel recurrent neural network (RNN) approach to account for temporal dynamics and dependencies in brain networks observed via functional magnetic resonance imaging (fMRI). Our approach directly parameterizes temporal dynamics through recurrent connections, which can be used to formulate blind source separation with a conditional (rather than marginal) independence assumption, which we call RNN-ICA. This formulation enables us to visualize the temporal dynamics of both first order (activity) and second order (directed connectivity) information in brain networks that are widely studied in a static sense, but not well-characterized dynamically. RNN-ICA predicts dynamics directly from the recurrent states of the RNN in both task and resting state fMRI. Our results show both task-related and group-differentiating directed connectivity.

KW - Aod

KW - Deep learning

KW - Fmri

KW - Ica

KW - Neuroimaging methods

KW - Resting-state fmri

KW - Rnn

UR - http://www.scopus.com/inward/record.url?scp=85055199148&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055199148&partnerID=8YFLogxK

U2 - 10.3389/fnins.2018.00600

DO - 10.3389/fnins.2018.00600

M3 - Article

VL - 12

JO - Frontiers in Neuroscience

JF - Frontiers in Neuroscience

SN - 1662-4548

IS - SEP

M1 - 00600

ER -