Towards precision EEG connectomics: Evaluating the benefits of dense sampling

Area of Science:

• Neuroscience
• Brain Imaging
• Computational Neuroscience

Background:

• EEG connectomics is crucial for understanding neurodevelopment, brain disorders, and behavior.
• Existing functional connectivity (FC) measures lack direct comparison regarding reliability and data quantity influence.
• Reliability is key for robust EEG connectomics studies.

Purpose of the Study:

• To compare the reliability of 9 popular EEG functional connectivity measures.
• To assess the impact of data quantity on measure reliability and individualization.
• To identify optimal EEG connectomics measures for study design.

Main Methods:

• Collected 80 minutes of EEG data from 25 parent-child pairs over 4 sessions.
• Calculated connectomes using 9 phase- and envelope-based FC measures and 1 effective connectivity measure.
• Employed connectome individualization (fingerprinting) as a multivariate reliability metric.
• Used simulations to assess volume conduction influence on identifiability.

Main Results:

• Coherence (COH) and Phase Locking Value (PLV) were vulnerable to volume conduction.
• Imaginary Coherence (IMCOH), Weighted Phase-Lag Index (WPLI), Phase-Lag Index (PLI), and Phase Slope Index (PSI) showed minimal vulnerability.
• IMCOH exhibited the best performance for individualization and group effect reliability.
• Reliability improved with 25-30 minutes of data for IMCOH and other corrected measures.

Conclusions:

• IMCOH is a highly reliable EEG FC measure, especially when volume conduction is considered.
• Sufficient data (25-30 minutes) enhances the reliability of corrected EEG FC measures.
• Findings guide the selection of EEG FC measures and data collection duration for robust connectomics research.