State-Space Analysis of Granger-Geweke Causality Measures with Application to fMRI

The recent interest in the dynamics of networks and the advent, across a range of applications, of measuring modalities that operate on different temporal scales have put the spotlight on some significant gaps in the theory of multivariate time series. Fundamental to the description of network dynamics is the direction of interaction between nodes, accompanied by a measure of the strength of such interactions. Granger causality and its associated frequency domain strength measures (GEMs) (due to Geweke) provide a framework for the formulation and analysis of these issues. In pursuing this setup, three significant unresolved issues emerge. First, computing GEMs involves computing submodels of vector time series models, for which reliable methods do not exist. Second, the impact of filtering on GEMs has never been definitively established. Third, the impact of downsampling on GEMs has never been established. In this work, using state-space methods, we resolve all these issues and illustrate the results with some simulations. Our analysis is motivated by some problems in (fMRI) brain imaging, to which we apply it, but it is of general applicability.

[1]  Hualou Liang,et al.  Short-window spectral analysis of cortical event-related potentials by adaptive multivariate autoregressive modeling: data preprocessing, model validation, and variability assessment , 2000, Biological Cybernetics.

[2]  Karl J. Friston,et al.  Effective connectivity: Influence, causality and biophysical modeling , 2011, NeuroImage.

[3]  Vince D. Calhoun,et al.  Dynamic modeling of neuronal responses in fMRI using cubature Kalman filtering , 2011, NeuroImage.

[4]  Peter E. Caines,et al.  Weak and strong feedback free processes , 1976 .

[5]  K. D. C. Stoodley.  Time Series and Systems Analysis with Applications , 1985 .

[6]  James D. Hamilton.  Time Series Analysis , 1994 .

[7]  Alexandre Andrade,et al.  Lag-based effective connectivity applied to fMRI: A simulation study highlighting dependence on experimental parameters and formulation , 2014, NeuroImage.

[8]  Mohamed-Jalal Fadili,et al.  Activelets: Wavelets for sparse representation of hemodynamic responses , 2011, Signal Process..

[9]  C. Granger.  Investigating Causal Relations by Econometric Models and Cross-Spectral Methods , 1969 .

[10]  Michael Green,et al.  All-pass matrices, the positive real lemma, and unit canonical correlations between future and past , 1988 .

[11]  S. Petersen,et al.  Characterizing the Hemodynamic Response: Effects of Presentation Rate, Sampling Procedure, and the Possibility of Ordering Brain Activity Based on Relative Timing , 2000, NeuroImage.

[12]  C. Sims.  Money, Income, and Causality , 1972 .

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

[14]  E. J. Hannan,et al.  Unit canonical correlations between future and past , 1988 .

[15]  Jean-Marie Dufour,et al.  Short and long run causality measures: Theory and inference , 2008 .

[16]  B. Moor,et al.  Subspace identification for linear systems , 1996 .

[17]  Zafer Dogan,et al.  Detecting spontaneous brain activity in functional magnetic resonance imaging using finite rate of innovation , 2014, 2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI).

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

[19]  Muhong Wang,et al.  Global warming and global dioxide emission : An empirical study , 1996 .

[20]  Karthik Ramakrishnan Sreenivasan,et al.  Nonparametric Hemodynamic Deconvolution of fMRI Using Homomorphic Filtering , 2015, IEEE Transactions on Medical Imaging.

[21]  C. Sims.  DISCRETE APPROXIMATIONS TO CONTINUOUS TIME DISTRIBUTED LAGS IN ECONOMETRICS , 1971 .

[22]  N. L. Johnson,et al.  Multivariate Analysis , 1958, Nature.

[23]  S. Bressler,et al.  Frequency decomposition of conditional Granger causality and application to multivariate neural field potential data , 2006, Journal of Neuroscience Methods.

[24]  C. W. J. Granger,et al.  Economic Processes Involving Feedback , 1963, Inf. Control..

[25]  Guorong Wu,et al.  A blind deconvolution approach to recover effective connectivity brain networks from resting state fMRI data , 2012, Medical Image Anal..

[26]  Tohru Ozaki,et al.  Evaluating frequency-wise directed connectivity of BOLD signals applying relative power contribution with the linear multivariate time-series models , 2005, NeuroImage.

[27]  Victor Solo,et al.  Topics in advanced time series analysis , 1986 .

[28]  Xiaoping Hu,et al.  Effect of hemodynamic variability on Granger causality analysis of fMRI , 2010, NeuroImage.

[29]  D Lehmann,et al.  Isolated effective coherence (iCoh): causal information flow excluding indirect paths , 2014 .

[30]  W. Larimore.  System Identification, Reduced-Order Filtering and Modeling via Canonical Variate Analysis , 1983, 1983 American Control Conference.

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

[32]  Umberto Triacca,et al.  Is Granger causality analysis appropriate to investigate the relationship between atmospheric concentration of carbon dioxide and global surface air temperature? , 2005 .

[33]  Rainer Goebel,et al.  Mapping directed influence over the brain using Granger causality and fMRI , 2005, NeuroImage.

[34]  Mark D'Esposito,et al.  The continuing challenge of understanding and modeling hemodynamic variation in fMRI , 2012, NeuroImage.

[35]  Larry D. Haugh,et al.  Causality in temporal systems: Characterization and a survey , 1977 .

[36]  Fredrik Edin.  Scaling errors in measures of brain activity cause erroneous estimates of effective connectivity , 2010, NeuroImage.

[37]  Essa Yacoub,et al.  The rapid development of high speed, resolution and precision in fMRI , 2012, NeuroImage.

[38]  Manfred Deistler,et al.  Consistency and Relative Efficiency of Subspace Methods , 1994 .

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

[40]  J. Magnus,et al.  Matrix Differential Calculus with Applications in Statistics and Econometrics , 1991 .

[41]  Victor Solo,et al.  Identifying fMRI Model Violations With Lagrange Multiplier Tests , 2012, IEEE Transactions on Medical Imaging.

[42]  Graham C. Goodwin,et al.  Digital control and estimation : a unified approach , 1990 .

[43]  L. Telser,et al.  Discrete Samples and Moving Sums in Stationary Stochastic Processes , 1967 .

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

[45]  Dietmar Bauer,et al.  ESTIMATING LINEAR DYNAMICAL SYSTEMS USING SUBSPACE METHODS , 2005, Econometric Theory.

[46]  John Geweke,et al.  TEMPORAL AGGREGATION IN MULTIPLE-REGRESSION MODEL , 1978 .

[47]  C. Segebarth,et al.  Identifying Neural Drivers with Functional MRI: An Electrophysiological Validation , 2008, PLoS biology.

[48]  G. Glover.  Deconvolution of Impulse Response in Event-Related BOLD fMRI1 , 1999, NeuroImage.

[49]  Rainer Goebel,et al.  The identification of interacting networks in the brain using fMRI: Model selection, causality and deconvolution , 2011, NeuroImage.

[50]  Jorma Rissanen,et al.  Measures of mutual and causal dependence between two time series , 1987, IEEE Trans. Inf. Theory.

[51]  E. Hannan,et al.  The statistical theory of linear systems , 1989 .

[52]  Peter E. Caines,et al.  Feedback between stationary stochastic processes , 1975 .

[53]  Martin Eichenbaum,et al.  Temporal Aggre-gation and Structural Inference in Macroeconomics , 1987 .

[54]  J. Geweke,et al.  Measures of Conditional Linear Dependence and Feedback between Time Series , 1984 .

[55]  Pedro A. Valdes-Sosa,et al.  Spatio-temporal autoregressive models defined over brain manifolds , 2007, Neuroinformatics.

[56]  Jean-Marie Dufour,et al.  Short-Run and Long-Rub Causality in Time Series: Theory. , 1998 .

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

[58]  Rainer Goebel,et al.  Investigating directed cortical interactions in time-resolved fMRI data using vector autoregressive modeling and Granger causality mapping. , 2003, Magnetic resonance imaging.

[59]  Tobias Damm,et al.  Zero Crossings, Overshoot and Initial Undershoot in the Step and Impulse Responses of Linear Systems , 2014, IEEE Transactions on Automatic Control.

[60]  Luiz A. Baccalá,et al.  Partial Directed Coherence Asymptotics for VAR Processes of Infinite Order , 2008 .

[61]  Albert MARCET.  10 Temporal Aggregation of Economic Time Series , 2006 .

[62]  Peter König,et al.  On the directionality of cortical interactions studied by structural analysis of electrophysiological recordings , 1999, Biological Cybernetics.