Muscle artifacts in multichannel EEG: characteristics and reduction.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Unintentional muscle activities are common artifacts in clinical electroencephalography (EEG).
• These artifacts can obscure genuine drug or biological effects in EEG recordings.
• Existing methods for artifact reduction may not be efficient or effective for large clinical datasets.

Purpose of the Study:

• To characterize the spectral, temporal, and spatial properties of unintentional muscle artifacts in clinical EEG.
• To develop and validate a high-throughput method for reducing muscle artifacts in EEG.
• To improve the clarity of EEG signals for detecting drug or biological effects.

Main Methods:

• Independent Component Analysis (ICA) was used to extract pure muscle signals from EEG datasets.
• A novel high-throughput artifact reduction method, ICA-SR, was developed, utilizing a new feature called Spectral Ratio (SR).
• The method was evaluated on two clinical EEG datasets and compared against an existing artifact reduction technique.

Main Results:

• The spectral, temporal, and spatial characteristics of muscle artifacts were elucidated.
• The ICA-SR method demonstrated superior artifact reduction compared to an existing method, with less distortion introduced to the EEG signal.
• The effectiveness of ICA-SR was validated in real-world clinical EEG recordings, including a CO(2)-inhalation study.

Conclusions:

• The study confirmed that unintentional muscle activity characteristics are consistent with controlled muscle activities.
• Spatial characteristics of artifacts can be influenced by EEG equipment.
• The developed ICA-SR method offers an effective and efficient solution for processing clinical EEG, enhancing its utility in research and diagnostics.