Coherence, phase differences, phase shift, and phase lock in EEG/ERP analyses

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electroencephalogram (EEG) coherence is a critical measure for understanding brain dynamics.
• Spontaneous EEG signals exhibit phase locking and phase shifts, which are essential for neural communication.
• Traditional EEG analysis methods can be susceptible to artifacts that distort these physiological signals.

Purpose of the Study:

• To evaluate the impact of common EEG reference techniques on coherence and phase difference calculations.
• To identify artifact-resistant methods for analyzing EEG phase dynamics.
• To explore the utility of cross-frequency synchrony in surface EEG and LORETA for brain function insights.

Main Methods:

• Comparison of coherence and phase difference calculations using average reference, Laplacian transforms, and independent component analysis (ICA).
• Assessment of time domain measures for phase shift and phase lock.
• Analysis of cross-frequency synchrony in surface EEG and using Low Resolution Electromagnetic Tomography (LORETA).

Main Results:

• Average reference, Laplacian transforms, and ICA can distort physiologically generated phase differences, invalidating coherence and phase reset computations.
• Time domain measures of phase shift and phase lock are less prone to artifact.
• Time domain measures are independent of volume conduction effects.

Conclusions:

• Standard EEG reference methods can introduce artifacts that compromise the validity of coherence and phase difference analyses.
• Time domain measures provide a more robust approach to studying EEG phase dynamics, unaffected by volume conduction.
• Cross-frequency synchrony analysis in EEG and LORETA offers valuable insights into the brain's dynamic functional organization.