Sackler Centre for Consciousness Science

The Multivariate Granger Causality (MVGC) toolbox

The Sackler Centre pioneers methods for analysing directed functional connectivity in neural systems. Our Multivariate Granger Causality (MVGC) toolbox, created by Lionel Barnett and Anil Seth was released in 2009, has long been a standard resource in neuroscience. In April 2014 we released a major new update, completely rewritten from the ground up.

The Multivariate Granger Causality (MVGC) toolbox is based on Granger causality, a concept originating from econometrics which sees 'causality' as involving both predictability and precedence. Put simply, according to Granger causality a variable A 'causes' a variable B if the past of A helps predict the future of B better than can be done by knowing only the past of B.

The MVGC Matlab® Toolbox is open-source software designed to facilitate Granger-causal analysis with multivariate and (optionally) multi-trial time series data. It supersedes the popular Granger Causal Connectivity Analysis (GCCA) Toolbox, and to a large extent subsumes, enhances and extends GCCA functionality.

The toolbox uses a novel, accurate and efficient approach to numerical computation and statistical inference of Granger causality, conditional and unconditional, in both time and frequency domains, as described in the accompanying reference document [1]. It is not "black box" software; there is no GUI, but rather a set of functions designed to be used in your own Matlab® programs. Annotated demonstration scripts are available which may be used as templates to assist in this task.

The software is developed and maintained by Lionel Barnett and Anil K. Seth.

[1] L. Barnett and A. K. Seth, The MVGC Multivariate Granger Causality Toolbox: A New Approach to Granger-causal Inference, J. Neurosci. Methods 223, 2014. 

For more information and to download the new software, which is written for the MATLAB environment, see here.

MVGC [ZIP 387.17KB]