Identifying frequency-domain features for an EEG-based pain measurement system

An objective pain measurement device would provide a revolutionary tool for physicians to treat patients. Electroencephalography (EEG) is a widely-used, inexpensive, simply-administered method for obtaining robust information on brain activity. This study shows how frequency domain data from EEG could provide features toward a pattern recognition system to detect and possibly quantify pain in humans. In addition artifacts from pain trials not exclusively linked to pain state (i.e., muscle, arousal, anxiety, etc.) will be identified. EEG data was recorded from human subjects during pain and control trials. Scalp and facial muscle artifact primarily affected only the Gamma band (26-98 Hz), while state of anxiety caused significant changes only in the frontal electrodes of the Theta band (5-6 Hz). State of arousal appeared to be the primary factor in only the Beta2 band (19-25 Hz). Decreased visual activity during pain is likely indicated by the observed increased Alpha (8-13 Hz) (associated with increased inhibition) over the visual cortex. Decreased Theta and Alpha during pain in right temporal-parietal region is consistent with a pain-related activation of brain areas in the parietal operculum and insula known to be involved in pain processes. Given these results, evidence that Alpha indexes cortical gating and the physiological significance of the observed gating, Alpha over the temporal-parietal region is a promising feature towards a pain measurement system.