A comparative study of different references for EEG default mode network: the use of the infinity reference

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electroencephalograph (EEG) reference selection critically impacts brain activity studies.
• The default mode network (DMN) is a key brain network studied using EEG.
• Investigating the impact of different EEG references on DMN analysis is essential.

Purpose of the Study:

• To evaluate the Reference Electrode Standardization Technique (REST) for obtaining an 'infinity' reference in EEG.
• To assess the influence of REST, average reference (AR), linked mastoids (LM), and left mastoid (L) references on EEG spectral properties and DMN.
• To determine the most accurate reference for studying functional connectivity and DMN.

Main Methods:

• Simulated 129-channel EEG data using 100 source configurations with coherent waveforms.
• Recorded eyes-closed EEG from 15 subjects.
• Re-referenced simulated and real EEG data to REST, AR, LM, and L.
• Analyzed spectral properties and constructed DMN based on coherence for each reference.

Main Results:

• REST accurately recovered simulated EEG network configurations.
• Significant differences in power spectra, coherence, and DMN were observed across references for real EEG data.
• AR strengthened long-distance anterior-posterior connectivity, while LM and L degraded posterior connectivity compared to REST.
• Reference effects were frequency-dependent.

Conclusions:

• Non-neutral EEG references significantly alter spectral properties, coherence, and network analysis, including DMN.
• REST is a valid and promising reference technique for objective EEG comparisons.
• While AR is closer to REST in spectral and coherence measures, it substantially alters the DMN.