Volume conduction effects in brain network inference from electroencephalographic recordings using phase lag index.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biomedical Engineering

Background:

• Electroencephalography (EEG) is crucial for brain network inference.
• Volume conduction in EEG can distort connectivity estimates.
• Phase Lag Index (PLI) is a common synchrony measure used in neuroscience.

Purpose of the Study:

• To assess the performance of the Phase Lag Index (PLI) in brain network inference using EEG data.
• To investigate the impact of volume conduction on PLI-based network estimation.
• To compare PLI with other synchrony measures like coherence (R) and phase coherence (PC) under simulated volume conduction.

Main Methods:

• Simulated EEG data using a four-sphere head model to mimic volume conduction.
• Estimated brain networks using PLI, R, and PC under conditions with and without volume conduction.
• Analyzed network properties, including clustering coefficient and small-worldness, under the null hypothesis of independent sources.

Main Results:

• Coherence (R) and Phase Coherence (PC) were highly influenced by volume conduction, inferring clustered networks.
• Phase Lag Index (PLI) demonstrated partial invariance to volume conduction.
• PLI-derived networks exhibited small-world properties, but high clustering was attributed to volume conduction artifacts when compared to PLI-NVC (no volume conduction).

Conclusions:

• Volume conduction can influence Phase Lag Index (PLI) results in EEG brain network inference.
• Ignoring the effects of volume conduction may lead to biased network estimations.
• Further research is needed to refine PLI or develop alternative methods to mitigate volume conduction artifacts in EEG connectivity analysis.