Validity of Time Reversal for Testing Granger Causality

Inferring causal interactions from observed data is a challenging problem, especially in the presence of measurement noise. To alleviate the problem of spurious causality, Haufe (2013) proposed to contrast measures of information flow obtained on the original data against the same measures obtained on time-reversed data. They show that this procedure, time-reversed Granger causality (TRGC), robustly rejects causal interpretations on mixtures of independent signals. While promising results have been achieved in simulations, it was so far unknown whether time reversal leads to valid measures of information flow in the presence of true interaction. Here, we prove that, for linear finite-order autoregressive processes with unidirectional information flow between two variables, the application of time reversal for testing Granger causality indeed leads to correct estimates of information flow and its directionality. Using simulations, we further show that TRGC is able to infer correct directionality with similar statistical power as the net Granger causality between two variables, while being much more robust to the presence of measurement noise.

[1]  M. G. Kendall,et al.  A Study in the Analysis of Stationary Time-Series. , 1955 .

[2]  J. Kiefer,et al.  An Introduction to Stochastic Processes. , 1956 .

[3]  P. Whittle.  On the fitting of multivariate autoregressions, and the approximate canonical factorization of a spectral density matrix , 1963 .

[4]  J. Andel.  Symmetric and Reversed Multiple Stationary Autoregressive Series , 1972 .

[5]  G. C. Tiao,et al.  Effect of temporal aggregation on the dynamic relationship of two time series variables , 1976 .

[6]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[7]  P. Newbold.  Feedback Induced by Measurement Errors , 1978 .

[8]  Diane Valérie Ouellette.  Schur complements and statistics , 1981 .

[9]  J. Geweke,et al.  Measurement of Linear Dependence and Feedback between Multiple Time Series , 1982 .

[10]  Jonathan D. Cryer,et al.  Time Series Analysis , 1986 .

[11]  Evgenia K. Akritas,et al.  Various proofs of Sylvester's (determinant) identity , 1996 .

[12]  D. Stern,et al.  Evidence for human influence on climate from hemispheric temperature relations , 1997, Nature.

[13]  C. Granger.  Investigating causal relations by econometric models and cross-spectral methods , 1969 .

[14]  Luiz A. Baccalá,et al.  Partial directed coherence: a new concept in neural structure determination , 2001, Biological Cybernetics.

[15]  Umberto Triacca,et al.  On the use of Granger causality to investigate the human influence on climate , 2001 .

[16]  Michael Breakspear,et al.  Spatiotemporal wavelet resampling for functional neuroimaging data , 2004, Human brain mapping.

[17]  Marcus J. Chambers,et al.  Granger Causality and the Sampling of Economic Processes , 2006 .

[18]  Katarzyna J. Blinowska,et al.  A new method of the description of the information flow in the brain structures , 1991, Biological Cybernetics.

[19]  M. Hallett,et al.  Identifying true brain interaction from EEG data using the imaginary part of coherency , 2004, Clinical Neurophysiology.

[20]  Karin Schwab,et al.  Comparison of linear signal processing techniques to infer directed interactions in multivariate neural systems , 2005, Signal Process..

[21]  Kaare Brandt Petersen,et al.  The Matrix Cookbook , 2006 .

[22]  Andreas Ziehe,et al.  Identifying interactions in mixed and noisy complex systems. , 2006, Physical review. E, Statistical, nonlinear, and soft matter physics.

[23]  G. Rangarajan,et al.  Mitigating the effects of measurement noise on Granger causality. , 2007, Physical review. E, Statistical, nonlinear, and soft matter physics.

[24]  Helmut Ltkepohl,et al.  New Introduction to Multiple Time Series Analysis , 2007 .

[25]  Karen O. Egiazarian,et al.  Measuring directional coupling between EEG sources , 2008, NeuroImage.

[26]  Volkmar Glauche,et al.  On the Detection of Direct Directed Information Flow in fMRI , 2008, IEEE Journal of Selected Topics in Signal Processing.

[27]  Daniele Marinazzo,et al.  Kernel method for nonlinear granger causality. , 2007, Physical review letters.

[28]  Moritz Grosse-Wentrup,et al.  Understanding Brain Connectivity Patterns during Motor Imagery for Brain-Computer Interfacing , 2008, NIPS.

[29]  K. Müller,et al.  Robustly estimating the flow direction of information in complex physical systems. , 2007, Physical review letters.

[30]  J. Schoffelen,et al.  Source connectivity analysis with MEG and EEG , 2009, Human brain mapping.

[31]  Motoaki Kawanabe,et al.  Modeling Sparse Connectivity Between Underlying Brain Sources for EEG/MEG , 2009, IEEE Transactions on Biomedical Engineering.

[32]  Andreas Ziehe,et al.  Comparison of Granger Causality and Phase Slope Index , 2008, NIPS Causality: Objectives and Assessment.

[33]  Aapo Hyvärinen,et al.  Estimation of a Structural Vector Autoregression Model Using Non-Gaussianity , 2010, J. Mach. Learn. Res..

[34]  Gordon Pipa,et al.  Transfer entropy—a model-free measure of effective connectivity for the neurosciences , 2010, Journal of Computational Neuroscience.

[35]  Mark W. Woolrich,et al.  Network modelling methods for FMRI , 2011, NeuroImage.

[36]  Steven L. Bressler,et al.  Wiener–Granger Causality: A well established methodology , 2011, NeuroImage.

[37]  Robert D. Nowak,et al.  Causal Network Inference Via Group Sparse Regularization , 2011, IEEE Transactions on Signal Processing.

[38]  Patrik O. Hoyer,et al.  Causal Search in Structural Vector Autoregressive Models , 2009, NIPS Mini-Symposium on Causality in Time Series.

[39]  C. Grebogi,et al.  Inference of Granger causal time-dependent influences in noisy multivariate time series , 2012, Journal of Neuroscience Methods.

[40]  Stefan Haufe,et al.  Alleviating the Influence of Weak Data Asymmetries on Granger-Causal Analyses , 2012, LVA/ICA.

[41]  Martin J. McKeown,et al.  A Generalized Multivariate Autoregressive (GmAR)-Based Approach for EEG Source Connectivity Analysis , 2012, IEEE Transactions on Signal Processing.

[42]  Stefan Haufe,et al.  A critical assessment of connectivity measures for EEG data: A simulation study , 2013, NeuroImage.

[43]  Bernhard Schölkopf,et al.  Causal Inference on Time Series using Restricted Structural Equation Models , 2013, NIPS.

[44]  Anil K. Seth,et al.  Granger causality analysis of fMRI BOLD signals is invariant to hemodynamic convolution but not downsampling , 2013, NeuroImage.

[45]  Douglas Zhou,et al.  Reliability of the Granger causality inference , 2014 .

[46]  G. Rangarajan,et al.  Effect of measurement noise on Granger causality. , 2014, Physical review. E, Statistical, nonlinear, and soft matter physics.

[47]  Conrado A. Bosman,et al.  How to detect the Granger-causal flow direction in the presence of additive noise? , 2015, NeuroImage.

[48]  L. Breuer.  Introduction to Stochastic Processes , 2022, Statistical Methods for Climate Scientists.