Code and simulation data for: A Bayesian method for inference of effective connectivity in brain networks for detecting the Mozart effect

Published: 25 May 2020| Version 1 | DOI: 10.17632/n9yh7t4bxv.1
Rik van Esch


Developed MATLAB code and simulation data for the Mozart effect study. The code requires the Bayesian topology identification code and the MVGC toolbox to run (see references).


Steps to reproduce

To choose different simulation data for analysis, change the .mat file that is loaded in topo_options.m. Make sure both the MVGC toolbox and Bayesian topology identification are installed. To run an analysis of simulation data, run granger_bayes_performance.m. It will produce the ROC curves of the simulation data. For the calculation of marginal hypothesis using your own ICA time series data, some steps need to be taken, which are detailed in testhyp.m.


Engineering, Graph Theory, Brain Dynamics