Granger causality revisited

This technical paper offers a critical re-evaluation of (spectral) Granger causality measures in the analysis of biological timeseries. Using realistic (neural mass) models of coupled neuronal dynamics, we evaluate the robustness of parametric and nonparametric Granger causality. Starting from a bro...

Full description

Bibliographic Details
Main Authors:	Friston, Karl J., Bastos, André M., Oswal, Ashwini, van Wijk, Bernadette, Richter, Craig, Litvak, Vladimir
Format:	Online
Language:	English
Published:	Academic Press 2014
Online Access:	https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4176655/

id	pubmed-4176655
recordtype	oai_dc
spelling	pubmed-41766552014-11-01 Granger causality revisited Friston, Karl J. Bastos, André M. Oswal, Ashwini van Wijk, Bernadette Richter, Craig Litvak, Vladimir Technical Note This technical paper offers a critical re-evaluation of (spectral) Granger causality measures in the analysis of biological timeseries. Using realistic (neural mass) models of coupled neuronal dynamics, we evaluate the robustness of parametric and nonparametric Granger causality. Starting from a broad class of generative (state-space) models of neuronal dynamics, we show how their Volterra kernels prescribe the second-order statistics of their response to random fluctuations; characterised in terms of cross-spectral density, cross-covariance, autoregressive coefficients and directed transfer functions. These quantities in turn specify Granger causality — providing a direct (analytic) link between the parameters of a generative model and the expected Granger causality. We use this link to show that Granger causality measures based upon autoregressive models can become unreliable when the underlying dynamics is dominated by slow (unstable) modes — as quantified by the principal Lyapunov exponent. However, nonparametric measures based on causal spectral factors are robust to dynamical instability. We then demonstrate how both parametric and nonparametric spectral causality measures can become unreliable in the presence of measurement noise. Finally, we show that this problem can be finessed by deriving spectral causality measures from Volterra kernels, estimated using dynamic causal modelling. Academic Press 2014-11-01 /pmc/articles/PMC4176655/ /pubmed/25003817 http://dx.doi.org/10.1016/j.neuroimage.2014.06.062 Text en © 2014 The Authors
repository_type	Open Access Journal
institution_category	Foreign Institution
institution	US National Center for Biotechnology Information
building	NCBI PubMed
collection	Online Access
language	English
format	Online
author	Friston, Karl J. Bastos, André M. Oswal, Ashwini van Wijk, Bernadette Richter, Craig Litvak, Vladimir
spellingShingle	Friston, Karl J. Bastos, André M. Oswal, Ashwini van Wijk, Bernadette Richter, Craig Litvak, Vladimir Granger causality revisited
author_facet	Friston, Karl J. Bastos, André M. Oswal, Ashwini van Wijk, Bernadette Richter, Craig Litvak, Vladimir
author_sort	Friston, Karl J.
title	Granger causality revisited
title_short	Granger causality revisited
title_full	Granger causality revisited
title_fullStr	Granger causality revisited
title_full_unstemmed	Granger causality revisited
title_sort	granger causality revisited
description	This technical paper offers a critical re-evaluation of (spectral) Granger causality measures in the analysis of biological timeseries. Using realistic (neural mass) models of coupled neuronal dynamics, we evaluate the robustness of parametric and nonparametric Granger causality. Starting from a broad class of generative (state-space) models of neuronal dynamics, we show how their Volterra kernels prescribe the second-order statistics of their response to random fluctuations; characterised in terms of cross-spectral density, cross-covariance, autoregressive coefficients and directed transfer functions. These quantities in turn specify Granger causality — providing a direct (analytic) link between the parameters of a generative model and the expected Granger causality. We use this link to show that Granger causality measures based upon autoregressive models can become unreliable when the underlying dynamics is dominated by slow (unstable) modes — as quantified by the principal Lyapunov exponent. However, nonparametric measures based on causal spectral factors are robust to dynamical instability. We then demonstrate how both parametric and nonparametric spectral causality measures can become unreliable in the presence of measurement noise. Finally, we show that this problem can be finessed by deriving spectral causality measures from Volterra kernels, estimated using dynamic causal modelling.
publisher	Academic Press
publishDate	2014
url	https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4176655/
_version_	1613137850158546944

Granger causality revisited

Similar Items