Inference of Granger causal time-dependent influences in noisy multivariate time series

Linda Sommerlade, Marco Thiel, Bettina Platt, Andrea Plano, Gernot Riedel, Celso Grebogi, Jens Timmer, Bjoern Olaf Schelter

Research output: Contribution to journalArticle

38 Citations (Scopus)

Abstract

Inferring Granger-causal interactions between processes promises deeper insights into mechanisms underlying network phenomena, e.g. in the neurosciences where the level of connectivity in neural networks is of particular interest. Renormalized partial directed coherence has been introduced as a means to investigate Granger causality in such multivariate systems. A major challenge in estimating respective coherences is a reliable parameter estimation of vector autoregressive processes. We discuss two shortcomings typical in relevant applications, i.e. non-stationarity of the processes generating the time series and contamination with observational noise. To overcome both, we present a new approach by combining renormalized partial directed coherence with state space modeling. A numerical efficient way to perform both the estimation as well as the statistical inference will be presented. (C) 2011 Elsevier B.V. All rights reserved.

Original languageEnglish
Pages (from-to)173-185
Number of pages13
JournalJournal of Neuroscience Methods
Volume203
Issue number1
Early online date16 Sep 2011
DOIs
Publication statusPublished - 15 Jan 2012

Fingerprint

Neurosciences
Causality
Noise

Keywords

  • non-stationary causal influences
  • time-resolved partial directed coherence
  • vector autoregressive processes
  • state space models
  • expectation-maximization algorithm
  • partial directed coherence
  • functional connectivity
  • maximum-likelihood
  • information-flow
  • signals
  • sleep
  • EEG
  • algorithm
  • behavior
  • systems

Cite this

Inference of Granger causal time-dependent influences in noisy multivariate time series. / Sommerlade, Linda; Thiel, Marco; Platt, Bettina; Plano, Andrea; Riedel, Gernot; Grebogi, Celso; Timmer, Jens; Schelter, Bjoern Olaf.

In: Journal of Neuroscience Methods, Vol. 203, No. 1, 15.01.2012, p. 173-185.

Research output: Contribution to journalArticle

@article{04f7dbe8ba7d406989d5f961fea8f0fd,
title = "Inference of Granger causal time-dependent influences in noisy multivariate time series",
abstract = "Inferring Granger-causal interactions between processes promises deeper insights into mechanisms underlying network phenomena, e.g. in the neurosciences where the level of connectivity in neural networks is of particular interest. Renormalized partial directed coherence has been introduced as a means to investigate Granger causality in such multivariate systems. A major challenge in estimating respective coherences is a reliable parameter estimation of vector autoregressive processes. We discuss two shortcomings typical in relevant applications, i.e. non-stationarity of the processes generating the time series and contamination with observational noise. To overcome both, we present a new approach by combining renormalized partial directed coherence with state space modeling. A numerical efficient way to perform both the estimation as well as the statistical inference will be presented. (C) 2011 Elsevier B.V. All rights reserved.",
keywords = "non-stationary causal influences, time-resolved partial directed coherence, vector autoregressive processes, state space models, expectation-maximization algorithm, partial directed coherence, functional connectivity, maximum-likelihood, information-flow, signals, sleep, EEG, algorithm, behavior, systems",
author = "Linda Sommerlade and Marco Thiel and Bettina Platt and Andrea Plano and Gernot Riedel and Celso Grebogi and Jens Timmer and Schelter, {Bjoern Olaf}",
year = "2012",
month = "1",
day = "15",
doi = "10.1016/j.jneumeth.2011.08.042",
language = "English",
volume = "203",
pages = "173--185",
journal = "Journal of Neuroscience Methods",
issn = "0165-0270",
publisher = "Elsevier",
number = "1",

}

TY - JOUR

T1 - Inference of Granger causal time-dependent influences in noisy multivariate time series

AU - Sommerlade, Linda

AU - Thiel, Marco

AU - Platt, Bettina

AU - Plano, Andrea

AU - Riedel, Gernot

AU - Grebogi, Celso

AU - Timmer, Jens

AU - Schelter, Bjoern Olaf

PY - 2012/1/15

Y1 - 2012/1/15

N2 - Inferring Granger-causal interactions between processes promises deeper insights into mechanisms underlying network phenomena, e.g. in the neurosciences where the level of connectivity in neural networks is of particular interest. Renormalized partial directed coherence has been introduced as a means to investigate Granger causality in such multivariate systems. A major challenge in estimating respective coherences is a reliable parameter estimation of vector autoregressive processes. We discuss two shortcomings typical in relevant applications, i.e. non-stationarity of the processes generating the time series and contamination with observational noise. To overcome both, we present a new approach by combining renormalized partial directed coherence with state space modeling. A numerical efficient way to perform both the estimation as well as the statistical inference will be presented. (C) 2011 Elsevier B.V. All rights reserved.

AB - Inferring Granger-causal interactions between processes promises deeper insights into mechanisms underlying network phenomena, e.g. in the neurosciences where the level of connectivity in neural networks is of particular interest. Renormalized partial directed coherence has been introduced as a means to investigate Granger causality in such multivariate systems. A major challenge in estimating respective coherences is a reliable parameter estimation of vector autoregressive processes. We discuss two shortcomings typical in relevant applications, i.e. non-stationarity of the processes generating the time series and contamination with observational noise. To overcome both, we present a new approach by combining renormalized partial directed coherence with state space modeling. A numerical efficient way to perform both the estimation as well as the statistical inference will be presented. (C) 2011 Elsevier B.V. All rights reserved.

KW - non-stationary causal influences

KW - time-resolved partial directed coherence

KW - vector autoregressive processes

KW - state space models

KW - expectation-maximization algorithm

KW - partial directed coherence

KW - functional connectivity

KW - maximum-likelihood

KW - information-flow

KW - signals

KW - sleep

KW - EEG

KW - algorithm

KW - behavior

KW - systems

U2 - 10.1016/j.jneumeth.2011.08.042

DO - 10.1016/j.jneumeth.2011.08.042

M3 - Article

VL - 203

SP - 173

EP - 185

JO - Journal of Neuroscience Methods

JF - Journal of Neuroscience Methods

SN - 0165-0270

IS - 1

ER -