Detection of phase synchronization in EEG with Bivariate Empirical Mode Decomposition

Area of Science:

• Neuroscience
• Signal Processing
• Computational Biology

Background:

• Phase synchronization in electroencephalography (EEG) is crucial for understanding cortical interactions.
• Traditional coherence methods require stationary signals, limiting their application to dynamic brain activity.
• Non-stationary EEG signals present challenges for accurately measuring neural synchrony.

Purpose of the Study:

• To develop and validate a novel technique for assessing phase synchrony in non-stationary EEG signals.
• To compare the proposed method's efficacy against conventional techniques during a cognitive task.
• To enhance the analysis of brain network dynamics in real-time or during complex tasks.

Main Methods:

• Application of the Phase Locking Value (PLV) metric.
• Integration of the Hilbert transform for phase extraction.
• Utilizing Bivariate Empirical Mode Decomposition (BEMD) to handle non-stationary signal characteristics.
• Analysis of EEG data recorded during Dynamical Dot Quartet (DDQ) tasks.

Main Results:

• The proposed method successfully measured phase synchrony in non-stationary EEG signals.
• The Phase Locking Value (PLV) combined with Bivariate Empirical Mode Decomposition (BEMD) demonstrated superior performance compared to traditional coherence.
• The technique effectively captured phase synchrony patterns during the Dynamical Dot Quartet (DDQ) tasks.

Conclusions:

• The novel technique offers a more suitable approach for analyzing phase synchrony in non-stationary EEG data, particularly during dynamic cognitive tasks like DDQ.
• This advancement can improve the understanding of brain connectivity and neural interactions in complex scenarios.
• The proposed method enhances the reliability of EEG signal analysis for non-stationary brain activity.