Holo-Hilbert spectral-based noise removal method for EEG high-frequency bands

Area of Science:

• Neuroscience
• Biomedical Engineering

Background:

• Simultaneous electroencephalography-functional magnetic resonance imaging (EEG-fMRI) offers insights into brain function dynamics.
• Gamma band EEG frequencies (>30 Hz) are linked to cognition and neurological disorders but are susceptible to MR artifacts.
• Existing methods struggle with MR-induced artifacts in the gamma band, limiting EEG-fMRI research.

Purpose of the Study:

• To develop and validate a novel noise removal method for the gamma band of EEG signals.
• To improve the quality of EEG data acquired during simultaneous EEG-fMRI scans.
• To facilitate the study of gamma band activity in relation to cognitive processes and neurological conditions.

Main Methods:

• The proposed method utilizes a modified Holo-Hilbert Spectral Analysis (HHSA) approach.
• It employs a nested empirical mode decomposition (EMD) strategy applied to amplitude and frequency modulation (AM/FM) components.
• Artifacts are reduced by removing low-power components based on the power-instantaneous frequency spectrum, followed by signal reconstruction.

Main Results:

• Simulations demonstrated significant reduction of MR-induced artifacts in the gamma band EEG.
• The method effectively preserved the original gamma band signal characteristics.
• The denoising technique proved efficient for simultaneous EEG/fMRI applications.

Conclusions:

• The novel HHSA-based method successfully removes artifacts from gamma band EEG signals.
• This technique is critical for advancing simultaneous EEG-fMRI research by ensuring signal integrity.
• The improved EEG data quality will aid in understanding brain function and disorders.