A procedure to reduce the parameters estimate in Granger-causal analysis through temporal smoothing.
Code to infer networks using a modified version of conditional Granger causality as described in A procedure to increase power of Granger-causal analysis through temporal smoothing.
Must have MVGC Multivariate Granger Causality toolbox to run as implemented in:
Barnett, L., Seth, A.K., 2014. The MVGC multivariate Granger causality toolbox: A new approach to Granger-causal inference. J. Neurosci. Methods 223, 50–68. doi:10.1016/j.jneumeth.2013.10.018. .
Three example simulations are provided:
Example 1) main_sim_1N_high_freq.m Example 2) main_sim_1N_low_freq.m Example 3) main_sim_9N.m
Examples 1 and 2 fit models for one signal dominated by high or low frequencies, respectively. Example 3 fits networks to a multivariate system of nine signals.
To execute an example, call the simualtion at the MATLAB command line:
>> main_sim_1N_high_freq