In graduation research by students at Kanazawa Institute of Technology's Department of Information Engineering, a deep learning model that uses brain waves to estimate the location of pain achieved an accuracy rate of 90.8±5.9%.
Personal and subjective pain is difficult to objectively evaluate by a third party, but in recent years, technology that uses brain waves to estimate pain intensity (how much it hurts) has been attracting attention. On the other hand, research on estimating the location of pain using brain waves is still insufficient.
Therefore, in this study, we captured the brain wave responses when painful stimuli were applied to the palms of subjects and examined whether it was possible to estimate the pain site from the brain waves. Electrodes were placed based on the somatosensory cortex of the brain, and data were analyzed from a total of 40 trials in which mild painful stimulation was applied to the palms of the right and left hands.
As a result, all three subjects showed particularly increased responses at the CP3 electrode in the left hemisphere when the right hand was stimulated, and at the CP4 electrode in the right hemisphere when the left hand was stimulated.
Furthermore, when they created a deep learning model based on this brainwave data, they achieved a high accuracy rate of 90.8±5.9% for each subject's data. This revealed that EEG patterns are strongly related to the localization of pain (where the pain is located), and a general-purpose learning model that estimates the location of pain from EEG, regardless of the characteristics of the subject, was developed. The effectiveness was suggested.
This time, the analysis results were only for a limited area of the palm of the hand, but in the future, we will collect and analyze a wide range of brain wave responses to pain stimulation in different body parts, and develop technology for objectively estimating pain locations using brain waves. Further progress is expected. If clinical application of this technology is realized, it is thought that it will be possible to provide prompt and appropriate treatment to hospitalized patients and infants who are unable to complain of pain on their own.
