Pain Classification using Evoked EEG Induced by Thermal Grill Illusion - Deep Neural Network Approach.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Pain Research

Background:

• Quantifying pain perception is a growing area in biomedical engineering.
• Electroencephalogram (EEG) signals are explored as biomarkers for pain, with recent studies focusing on thermal stimuli.
• Previous EEG studies on pain often used single stimulus conditions and long stimulation durations.

Purpose of the Study:

• To investigate the classification of EEG signals during distinct thermal pain stimuli: Cool (20 °C), Warm (40 °C), and Thermal Grill Illusion (TGI).
• To assess the consistency between objective EEG-based classification and subjective pain ratings.
• To identify brain regions associated with TGI perception using EEG features.

Main Methods:

• Collected EEG data from 43 subjects exposed to Cool, Warm, and TGI stimuli.
• Utilized a deep convolutional neural network (EEGNet) for classification of EEG signals.
• Performed individual binary classifications for TGI vs. Cool, TGI vs. Warm, and Warm vs. Cool conditions.

Main Results:

• Achieved classification accuracies of 0.74±0.01 for TGI-Cool, 0.71±0.01 for TGI-Warm, and 0.74±0.01 for Warm-Cool.
• Higher TGI-Cool classification accuracy was observed in subjects who perceived TGI as hotter than Warm.
• Discriminative EEG features for TGI were predominantly located in the parietal area for subjects experiencing the illusion.

Conclusions:

• Objective EEG features can effectively differentiate between thermal pain stimuli and align with subjective pain experiences.
• The findings suggest that the somatosensory cortex may be activated during the perception of TGI as hot pain.
• EEG analysis offers a promising avenue for objective pain quantification and understanding complex pain phenomena.