GMAC: a Matlab toolbox for spectral Granger causality analysis of fMRI data

Area of Science:

• Neuroscience
• Computational Neuroscience
• Medical Imaging Analysis

Background:

• Investigating causal interactions in brain networks using functional magnetic resonance imaging (fMRI) is complex.
• Granger causality analysis (GCA) is a key method for understanding neural activity patterns.
• Existing tools may lack comprehensive features or user-friendliness for advanced GCA.

Purpose of the Study:

• To introduce GMAC (Granger multivariate autoregressive connectivity), an open-source software toolbox.
• To implement multivariate spectral GCA for fMRI data analysis.
• To provide a user-friendly platform for brain network connectivity research.

Main Methods:

• Developed GMAC toolbox in Matlab with a graphical interface.
• Integrated fMRI data handling, preprocessing, and network node definition.
• Implemented multivariate autoregressive modeling and spectral GCA index estimation.
• Included statistical significance assessment using surrogate data.
• Enabled network analysis and visualization of connectivity results.

Main Results:

• GMAC offers a comprehensive suite of tools for multivariate spectral GCA.
• The toolbox facilitates fMRI data import/export, preprocessing, and modeling.
• It provides methods for estimating Granger causality indexes and assessing statistical significance.
• Network analysis and visualization capabilities are integrated for connectivity results.

Conclusions:

• GMAC provides an accessible and powerful open-source solution for fMRI-based brain network connectivity analysis.
• The toolbox supports advanced multivariate spectral GCA, benefiting both technical and clinical users.
• GMAC enhances the investigation of causal interactions within neural networks.