Related Experiment Video
Updated: Jun 20, 2026
11:15
Applications of EEG Neuroimaging Data: Event-related Potentials, Spectral Power, and Multiscale Entropy
Published on: June 27, 2013
Exploring transient transfer entropy based on a group-wise ICA decomposition of EEG data
Vasily A Vakorin1, Natasa Kovacevic, Anthony R McIntosh
1Rotman Research Institute of Baycrest, Canada. vasenka@gmail.com
Neuroimage
|August 25, 2009
Summary
This study introduces a new pipeline to analyze brain signal interactions using independent component analysis (ICA) and transfer entropy. It reveals task-specific functional integration patterns in electroencephalogram (EEG) data.
Area of Science:
- Neuroscience
- Signal Processing
- Computational Biology
Background:
- Estimating electroencephalogram (EEG) coherence directly from scalp electrodes can yield spurious results due to volume conductance.
- Independent Component Analysis (ICA) offers an alternative representation by separating mixed signals into independent components.
- Understanding interdependencies between these components is crucial for accurate brain activity analysis.
Purpose of the Study:
- To develop and validate a data-driven pipeline for analyzing asymmetries in mutual interdependencies between distinct EEG signal components.
- To identify functional roles of ICA components and quantify functional integration using information-theoretic approaches.
- To assess task-specific changes in functional integration within EEG data.
Main Methods:
- A group-based independent component analysis (ICA) was applied across subjects and conditions simultaneously.
- Partial Least Squares (PLS) analysis was used to specify functional roles of ICA components based on task effects.
- Transfer entropy analysis was employed to estimate functional integration from reconstructed phase dynamics.
- A secondary PLS analysis investigated task-specific changes in transfer entropy between functionally specified components.
Main Results:
- The pipeline successfully identified distinct components within EEG signals.
- Functional roles of ICA components were specified, and task effects were analyzed.
- Asymmetries in mutual interdependencies were quantified using transfer entropy.
- Robust, task-specific changes in functional integration were detected between components.
Conclusions:
- The proposed data-driven pipeline effectively addresses limitations of traditional EEG coherence analysis.
- The integration of ICA, PLS, and transfer entropy provides a powerful framework for studying brain functional integration.
- This approach enables the discovery of complex, task-dependent neural communication patterns.
