AQMFB-DWT: A Preprocessing Technique for Removing Blink Artifacts Before Extracting Pain-evoked Potential EEG

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Pain assessment often relies on electroencephalogram (EEG) signals.
• Extracting pain-evoked potentials from EEG is difficult due to blink artifacts and background noise.
• Current blink artifact removal methods have limitations, including reliance on reference signals and user input.

Purpose of the Study:

• To develop a novel framework for removing blink artifacts from pain-evoked potential EEG signals.
• To improve the accuracy and practicality of EEG-based pain assessment.

Main Methods:

• Proposed a framework using adaptive quadrature mirror filter banks (AQMFB) combined with discrete wavelet transform (DWT).
• Developed an adaptive wavelet construction method tailored to individual EEG characteristics, unlike traditional fixed DWT approaches.
• Evaluated the method's performance against four leading artifact removal techniques.

Main Results:

• The proposed AQMFB-DWT method effectively removed blink artifacts from pain EEG.
• Demonstrated superior performance compared to four existing methods in artifact removal.
• Achieved minimal distortion of crucial pain information while maintaining acceptable processing times.

Conclusions:

• The AQMFB-DWT framework offers a significant advancement in blink artifact removal for pain EEG.
• This technique serves as a valuable preprocessing step for more accurate pain-evoked potential extraction.
• The adaptive nature of the method enhances its applicability in clinical pain assessments.