On the Evaluation of Information Flow in Multivariate Systems by the Directed Transfer Function

Authors:
Michael Eichler
Affiliations:
Institut für Angewandte Mathematik, Universität Heidelberg, Im Neuenheimer Feld 294, 69120, Heidelberg, Germany
Venue:
Biological Cybernetics
Year:
2006

Citing 0
Cited 5

Causal pattern recovery from neural spike train data using the Snap Shot Score

Journal of Computational Neuroscience
On directed information theory and Granger causality graphs

Journal of Computational Neuroscience
A comparison of multivariate causality based measures of effective connectivity

Computers in Biology and Medicine
Statistical pitfalls in the comparison of multivariate causality measures for effective causality

Computers in Biology and Medicine
Quantification of interaction in multiloop control systems using directed spectral decomposition

Automatica (Journal of IFAC)

Quantified Score

Hi-index	0.00

Visualization

Abstract

The directed transfer function (DTF) has been proposed as a measure of information flow between the components of multivariate time series. In this paper, we discuss the interpretation of the DTF and compare it with other measures for directed relationships. In particular, we show that the DTF does not indicate multivariate or bivariate Granger causality, but that it is closely related to the concept of impulse response function and can be viewed as a spectral measure for the total causal influence from one component to another. Furthermore, we investigate the statistical properties of the DTF and establish a simple significance level for testing for the null hypothesis of no information flow.