Causal influence: advances in neurosignal analysis

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biophysics

Background:

• Analysis of multichannel neurosignals like electroencephalography (EEG) and magnetoencephalography (MEG) is crucial for brain research and clinical applications.
• Multivariate AutoRegressive (MAR) modeling offers high-resolution characterization of functional relations in neuronal data.
• Assessing causal relations in large-scale brain network interactions is a key challenge.

Purpose of the Study:

• To provide a comprehensive review of advances in causal influence measures for neurosignal analysis using MAR spectral analysis.
• To outline the mathematical foundations and estimation procedures for MAR models and causal measures.
• To discuss practical applications and technical considerations for analyzing neurobiological data, including neural spike trains.

Main Methods:

• Review of mathematical formulations for Multivariate AutoRegressive (MAR) models.
• Detailed explanation of causal influence measures within the MAR spectral analysis framework.
• Discussion of adaptation of MAR models for analyzing neural spike train data.

Main Results:

• MAR spectral analysis provides a robust framework for characterizing causal relations in brain networks.
• Advances enable high spatial, temporal, and frequency resolution in analyzing functional connectivity.
• The methodology has been successfully adapted for diverse neurobiological data, including spike trains.

Conclusions:

• Causal influence measures within MAR spectral analysis are powerful tools for understanding brain circuit dynamics.
• The review highlights practical applications in basic neuroscience, clinical diagnosis, and functional neuroimaging.
• Future research directions include further development and application of these advanced analytical techniques.