We are living in a historic period where artificial intelligence (AI) machine learning provides the foundation to rapidly transform human health and wellness. A new peer-reviewed study by researchers at the University of Southern California (USC) and the University of California, Los Angeles (UCLA) showcases a wireless device that uses AI to decode pain levels from brain activity in order to personalize spinal cord stimulation to treat chronic pain.
“Here we report an integrated flexible ultrasound-induced wireless implantable stimulator combined with a pain detection and management system for personalized chronic pain management,” reported corresponding co-authors Laiming Jiang, Jun Chen, and Qifa Zhou in collaboration with twenty other co-authors.
Chronic pain, a global problem
Chronic pain is pain that lasts more than three months. According to estimates from the International Association on the Study of Pain (IASP), one out of every five people worldwide suffers from chronic pain, and 50 million Americans live with chronic pain daily.
Clinically significant anxiety and depression occur frequently among chronic pain sufferers. Roughly 40 percent of adults with chronic pain also have anxiety and depression according to an analysis, led by researchers at Johns Hopkins University School of Medicine, of 376 published studies consisting of over 347,000 chronic pain sufferers from 50 countries. Those most at risk for co-occurrence are fibromyalgia patients, younger adults, and women, according to the study.
The need for non-drug pain treatment
“Chronic pain management typically involves opioids, which are associated with severe side effects such as addiction,” wrote the researchers. Opioids, a class of drugs that are addictive, can be either synthetic (fentanyl, loperamide, methadone, tramadol), semi-synthetic (oxycodone, oxymorphone, hydrocodone, heroin), or natural (opium, codeine, thebaine, morphine). Opioid addiction affects roughly 60 million people and causes over 100,000 overdose deaths (mostly from fentanyl) annually worldwide, according to the July 2023 issue of The Lancet Regional Health—Americas. The American Psychiatric Association estimates that in 2022, 74,000 out of the nearly 110,000 drug overdoses among Americans were due to the synthetic opioid fentanyl.
As an alternative to addictive painkillers and medication, there is an urgent call for drug-free chronic pain management that is customized to the individual patient’s needs.
Personalized Pain Treatment
The research team created a three-part pain management solution consisting of a flexible ultrasound-induced implantable stimulator that acts on the spinal cord to disrupt pain signals to the brain, a wearable ultrasound transmitter, and an AI deep learning algorithm that decodes brain activity in order to classify pain levels in three categories (slight, moderate, extreme). The electroencephalogram (EEG) brain activity of laboratory rats outfitted with the devices was monitored in order for the AI to classify pain levels. The type of AI artificial neural network used was based on ResNet-18, a convolutional neural network (CNN) with 18 layers that is a deep learning model often used to classify images. According to the USC Viterbi School of Engineering report, the deep learning model had a high accuracy rate of 94.8 percent in classifying pain levels of lab rodents.
“We classify pain stimuli from brain recordings by developing a machine learning model and program the acoustic energy from the ultrasound transmitter and therefore the intensity of electrical stimulation,” the researchers shared. After the AI classifies the pain level, the acoustic energy of the ultrasound transmitter is adjusted accordingly to control the power level of the electrical stimulation for pain management, allowing it to be more tailored and specific to the pain need.
“We show that the implant can generate targeted, self-adaptive, and quantitative electrical stimulation to the spinal cord, according to the classified pain levels for chronic pain management in free-moving animal models,” concluded the scientists.
By harnessing the power of AI to predict pain levels from complex image recordings of brain activity, a pioneering team of researchers are making significant headway in providing personalized chronic pain management with electrical stimulation without medication.
Copyright © 2025 Cami Rosso All rights reserved.
