Multilevel Laser-Induced Pain Measurement with Wasserstein Generative Adversarial Network - Gradient Penalty Model

Area of Science:

• Neuroscience
• Biomedical Engineering
• Pain Research

Background:

• Pain disorders affect billions globally, posing significant measurement challenges.
• Current pain assessment relies on subjective scores and limited objective biomarker-based measures like electroencephalography (EEG).
• Existing methods often fail to capture nuanced, multi-level pain experiences, necessitating advanced diagnostic tools.

Purpose of the Study:

• To develop and validate novel electroencephalography (EEG)-based methods for objective, multi-level pain assessment.
• To enhance EEG data for improved pain recognition accuracy using advanced augmentation techniques.
• To investigate the neural correlates of pain perception, specifically focusing on the parietal lobe's sensory function.

Main Methods:

• Utilized a high-power laser stimulation paradigm for pain induction.
• Extracted EEG features using a modified S-transform, preserving time-frequency information.
• Optimized 20-40 Hz frequency band features and employed Wasserstein generative adversarial network with gradient penalty for data augmentation.
• Classified five pain levels using advanced algorithms and compared performance against existing methods.

Main Results:

• The proposed EEG feature extraction and augmentation method demonstrated significant advantages for classifying a five-level pain dataset.
• Effective classification performance was observed using features from the parietal brain region, indicating its role in pain sensory processing.
• The developed techniques offer a more objective and detailed assessment of pain intensity compared to traditional methods.

Conclusions:

• This research presents a novel approach for quantitative pain measurement using enhanced EEG analysis.
• The findings highlight the potential of EEG-based biomarkers for objective pain assessment in clinical settings.
• The study provides new insights into the neural mechanisms of pain and offers advanced tools for pain recognition research.