A Novel Pain-Evoked EEG Extraction Method Guided by Pain Brain Sources

Pain-evoked potentials are widely recognized as a crucial electrophysiological indicator for assessing pain. However, isolating pain-evoked potentials from raw electroencephalogram (EEG) signals is often hindered by interference from background brain activity and external noise. Traditional methods, such as those based on wavelets, typically average data from all trials to generate templates. This approach, however, can mask subtle yet significant variations in EEG patterns associated with different pain intensities. To address this limitation, our study proposes a novel framework that utilizes pain-related brain source signals. By selectively extracting relevant EEG data from a historical database, the framework constructs more accurate templates, significantly improving the extraction of pain-related neural activity. Experiments show that our method surpasses four alternative techniques in accurately extracting pain-evoked potentials. This work not only tackles the challenge of weak pain representation in raw EEG data but also advances the development of reliable, objective biomarkers for pain assessment, representing a significant step forward in the field of pain diagnostics.