Unmixing Oscillatory Brain Activity by EEG Source Localization and Empirical Mode Decomposition

Area of Science:

• Neuroscience
• Biophysics
• Signal Processing

Background:

• Neuronal activity comprises synchronous and asynchronous oscillations at various frequencies.
• Electroencephalography (EEG) measures electrical brain activity from the scalp.
• Decomposing EEG signals in space and time is crucial for interpreting brain activity.

Purpose of the Study:

• To investigate source-based signal decomposition of EEG signals.
• To analyze neuronal dynamics using empirical mode decomposition (EMD) on spatially unmixed EEG data.
• To explore the relationship between localized electrical activity and motor responses.

Main Methods:

• Source localization was used to spatially unmix EEG signals.
• Empirical mode decomposition (EMD) was applied to analyze neuronal dynamics.
• Simulations and real EEG data recorded during transcranial magnetic stimulation (TMS) were utilized.

Main Results:

• EMD demonstrated higher accuracy when applied to spatially unmixed EEG signals compared to scalp-level signals.
• A correlation was observed between motor-evoked potential (MEP) amplitude and the phase of localized electrical activity preceding TMS.
• The study validates the effectiveness of combining source localization and EMD for EEG analysis.

Conclusions:

• Source localization combined with EMD offers a more accurate method for analyzing EEG signals.
• This integrated approach provides deeper insights into the phase and frequency of electrical oscillations and their cortical origins.
• The findings support the use of this combined technique for understanding brain mechanisms.