A Korean analysis team has developed AI that objectively analyzes pain based mostly on mind waves (EEG). Provided by Getty Images Bank.
Technology that objectively quantifies subjective pain, which varies from individual to individual, has been developed. By analyzing mind waves alone, it may assess pain depth even when the affected person doesn’t categorical it immediately, suggesting a method to objectively consider pain in sufferers with impaired consciousness or issue speaking.
Daegu Gyeongbuk Institute of Science and Technology (DGIST) introduced on the twenty sixth {that a} team led by Principal Researcher Jinwoong Ahn on the Industrial AX Innovation Headquarters, along with a team led by Professor Sungchan Jeon at Gwangju Institute of Science and Technology (GIST), developed a expertise that uses synthetic intelligence (AI) to analyze electroencephalography (EEG) indicators generated by thermal stimulation and classify pain depth. The outcomes had been revealed within the May challenge of the worldwide journal ‘IEEE Transactions on Neural Systems and Rehabilitation Engineering’.
In medical observe, physicians primarily depend on subjective pain assessments by which sufferers report their pain utilizing numbers or phrases. However, even with the identical stimulus, individuals describe pain in a different way, which leads to giant variations in evaluation. The tougher communication is—for instance in pediatric sufferers, older adults, or critically unwell sufferers—the more durable it turns into to consider pain precisely.
The analysis team developed an AI expertise that analyzes EEG information recorded throughout numerous levels of thermal stimulation. EEG is a check that measures and data electrical indicators generated by neuronal exercise utilizing electrodes hooked up to the scalp.
Instead of utilizing sufferers’ self-reported subjective pain scores to prepare the AI, the team employed an inner AI validation system to enhance the objectivity of pain analysis. Two AI fashions every predicted the pain stage after which in contrast their outputs, and the coaching course of was designed in order that the system selectively discovered solely from information judged to be extremely dependable.
When the team validated the AI’s efficiency utilizing EEG information from 41 individuals, they discovered that it outperformed current EEG-based pain classification fashions. The mannequin additionally maintained secure predictive efficiency for forms of stimuli it had not beforehand encountered throughout coaching.
They additional confirmed that delta wave exercise within the left and proper anterior temporal lobes is carefully related to pain depth. This is thought to be offering neurophysiological proof for creating future pain indices based mostly on mind indicators.
Principal Researcher Ahn mentioned, “This research directly tackles the bias from subjective self-reporting, which has been a major limitation of pain analysis based on brain waves,” including, “We plan to integrate various biosignals and develop it into a general-purpose AI pain platform that can be used in real clinical settings.”
doi.org/10.1109/TNSRE.2026.3692232
From left: Principal Researcher Jinwoong Ahn, Postdoctoral Researcher Euijin Jung, and GIST Professor Sungchan Jeon. Provided by DGIST.
Copyright ⓒ DongA Science. All rights reserved.