Sometimes the pain felt like lightning bolts. Or snakes biting. Or needles.

“Just imagine the worst burn you’ve ever had, all over your body, never going away,” said Ed Mowery, 55, describing his life with chronic pain. “I would wake up in the middle of night, screaming at the top of my lungs.”

Beginning with a severe knee injury he got playing soccer at 15, he underwent about 30 major surgeries for various injuries over the decades, including procedures on his knees, spine and ankles. Doctors put in a spinal cord stimulator, which delivers electrical pulses to relieve pain, and prescribed morphine, oxycodone and other medications, 17 a day at one point. Nothing helped.

Unable to walk or sit for more than 10 minutes, Mr. Mowery, of Rio Rancho, N.M., had to stop working at his job selling electronics to engineering companies and stop playing guitar with his death metal band.

Out of options four years ago, Mr. Mowery signed up for a cutting-edge experiment: a clinical trial involving personalized deep brain stimulation to try to ease chronic pain.

The study, published on Wednesday, outlines a new approach for the most devastating cases of chronic pain, and could also provide insights to help drive invention of less invasive therapies, pain experts said.

“It’s highly innovative work, using the experience and technology they have developed and applying it to an underserved area of medicine,” said Dr. Andre Machado, chief of the Neurological Institute at Cleveland Clinic, who was not involved in the study.

Chronic pain, defined as lasting at least three months, afflicts about 20 percent of adults in the United States, an estimated 50 million people, according to the Centers for Disease Control and Prevention. In about a third of cases, the pain substantially limits daily activities, the C.D.C. reported.

Some patients find relief in medication, nerve blocks or spinal cord stimulation, but doctors say these methods don’t work for many patients and have led to opioid addiction for some.

The study Mr. Mowery participated in was small, involving six patients with various types of longstanding chronic pain unrelieved by other treatments. But the trial was intensive and rigorous. Researchers implanted electrodes, identified brain areas and neural signals linked to individual patients’ pain, and then developed personalized algorithms using A.I. that delivered bursts of stimulation when those signals surged.

Patients were followed for an average of 22 months, including a phase in which they randomly received either active stimulation or inactive sham stimulation, without knowing which.

The study found that personalized stimulation received in response to high-pain signals reduced average daily pain by about 60 percent. Sham stimulation, which can have placebo effects, provided almost no benefit.

Patients reported that they could walk more, that their mood was improved and that pain interfered less with their daily activities, sleep and enjoyment of life after personalized stimulation compared to sham. One patient was thrilled to be able to hug his wife without feeling pain. Another, a woman who had been virtually homebound, was able to travel and attend church and a family wedding.

Since the trial, two patients have stopped taking all opioid and pain medications and two others, including Mr. Mowery, have drastically reduced their use of those medications.

Currently, deep brain stimulation is approved by the Food and Drug Administration for only a few conditions, including epilepsy and Parkinson’s disease. But this conventional stimulation is not individualized: electrodes are implanted in one or two brain areas for each condition, and stimulation is usually delivered continuously or on a regular schedule.

An author of the new study, Dr. Prasad Shirvalkar, a neurologist at University of California, San Francisco, said deep brain stimulation had been briefly approved for chronic pain about 40 years ago, but approval was withdrawn a few years later because results were inconsistent. Some patients were not helped, while others found that their relief didn’t last. And solid data was limited because many studies did not compare active stimulation to sham stimulation, he said.

“Pain has been a mystery for a long time,” Dr. Shirvalkar said.

Pain is complex because it involves different interconnected brain areas, not one pain center, and because it has three neurological dimensions, Dr. Machado said: “What you feel, the sensation; how you understand it, which is cognitive; and how you process the emotion of it, the suffering.”

Reducing suffering, if not the pain itself, was one aim of the new study, said Dr. Edward Chang, chairman of neurological surgery at U.C.S.F. and a senior author. Another goal was a better understanding of how the brain varies between patients, said Dr. Chang, who, with colleagues, has helped lead the field of brain-computer interface research, including devising personalized deep brain stimulation for Parkinson’s and depression.

“We’re building these computational models of how those parts of the brain work,” he said. “We look across people who have different conditions and we look at when people have the symptoms and when they don’t.”

Dr. Shirvalkar said his own diagnoses of Tourette syndrome and obsessive-compulsive disorder made him particularly interested in understanding the brain.

The theory behind the trial was that delivering stimulation in response to a patient’s distinctive pain signals (called a closed-loop system) could operate like “a thermostat for pain,” Dr. Shirvalkar said. Besides providing individualized treatment, it might avoid a possible pitfall of continuous stimulation, in which “the brain might adapt and learn to ignore it,” he said.

The study was published on MedRxiv, a site for research that has not yet been peer-reviewed, but the research, funded by the National Institutes of Health, was previously presented at a conference and before a congressional committee.

Three of the trial participants had post-stroke pain; one had chemotherapy-induced peripheral neuropathy; one had a spinal cord injury; and Mr. Mowery had complex regional pain syndrome, which can cause severe pain and other symptoms in the limbs, skin and joints.

First, researchers implanted electrodes to read signals from and stimulate 14 different brain locations. Over ten days, they tested various combinations to decipher pain biomarkers — patterns that correspond to low-pain and high-pain states — and identify optimal stimulation characteristics to respond to each patient’s high-pain condition.

One patient with a rare post-stroke syndrome experienced little pain relief, so he stopped participating. With the others, researchers implanted permanent electrodes and followed the patients, randomizing the stimulation between the active and sham versions. Patients reported about their pain regularly and completed surveys about their mood and daily activities.

Under the personalized model, which also adjusted for sleep, stimulation was delivered between 7 percent and 55 percent of the time, much less than is provided by conventional stimulation that is always on.

The trial was not without risks, primarily involving the implant surgery, researchers reported. After implantation, one participant developed an infection, requiring the device on one side of her brain to be removed. Another experienced speech problems that took several months to resolve. Mr. Mowery experienced a brief seizure during the trial’s exploratory phase. The stimulation itself caused no adverse effects, researchers said. They also said that participants were given extensive psychiatric support throughout the trial.

Dr. Alex Green, a professor of neurosurgery at Oxford, who was not involved in the study, called it “quite novel,” adding, “we’ve never before seen what happens in real time in terms of electrical activities in these different brain areas.”

Outside experts and the authors noted several limitations of the study. Given the small number of participants, with different medical conditions, “it’s hard to understand the consistency of effect,” said Dr. Machado, who led one of the few previous randomized stimulation trials for chronic pain. And, he said, because it did not involve comparisons to stimulation that runs continuously or at random or regular intervals, it’s unclear if those models would also help these patients.

Dr. Shirvalkar said the trial showed that personalized stimulation “is possible and that you can implement it in a way that actually helps people, but it’s not proof that it’s necessary.” Additional studies are planned, he said.

Unless it can be done more simply, the invasiveness and expense might make it inaccessible for many patients, experts said. “I think it does have potential, but we don’t quite know what that potential is,” Dr. Green, a longtime deep brain stimulation researcher, said.

Still, Dr. Chang said, the devices and procedures may eventually become as accessible as cardiac pacemakers, and that by identifying pain biomarkers such research could “unlock many other therapies, both pharmacologic and stimulation-based.”

When Mr. Mowery first learned about the study, the prospect of undergoing brain surgery deterred him. But 18 months later, “the pain got so terrible” that it felt “like someone tying fishing line around my toes and yanking,” he said. “Taking a shower at times was like razor blades hitting the skin.”

Toward the end of the 10-day exploratory phase in 2021, one stimulation algorithm suddenly did the trick.

“The pain’s gone,” Mr. Mowery said from his hospital bed, head swathed in bandages from surgery, a video shows. “Pain’s stopped.”

Dr. Shirvalkar was incredulous. “You’re not kidding me?” he asked.

“I’m not kidding you,” Mr. Mowery replied.

After the permanent implant, researchers followed him until 2024. “It took about a year to figure out how to get his biomarkers to reliably activate his brain stimulation,” Dr. Shirvalkar said.

Now, Mr. Mowery said, he has only “maybe one bad day a week.” While he can set his stimulation to a specific pattern, setting it to respond to his individual pain signals usually works best. “I’ll feel the pain in my foot pop up, and then it just goes away,” he said.

He’s sharply tapered his morphine and other pain medications and aims to wean himself off all of them this year. He said he had shocked his lawyer by asking to discontinue his disability benefits.

He’s been able to perform and finish an album with his band, Tiwanaku, named after a Bolivian archaeological site. He is even writing a song about his deep brain stimulation experience.

“It’s definitely going to be one of the more aggressive songs we do,” he said. “Something to do with the euphoria after being dropped out of pain.”