AI-Guided Wrist Device Shows Promise for Treating Essential Tremor

An artificial intelligence–driven wearable device may offer a meaningful, noninvasive treatment option for people with essential tremor, according to a Houston Methodist Hospital-led clinical trial about the common movement disorder that can severely interfere with daily life.

The randomized trial, published in JAMA Neurology in October, found that a transcutaneous peripheral nerve stimulation (TPNS) wrist device designed to continuously modulate tremor-related brain circuits through stimulation of the median, ulnar and radial nerves significantly improved symptoms in patients who wore it.

"This fills an important therapeutic gap," said Dr. William Ondo, a Houston Methodist neurologist and the study's first author. "It's not as powerful as brain surgery, but it's far less invasive, making it a more appealing option for many patients."

Dr. Ondo noted that essential tremor affects an estimated 1% to 2% of people worldwide and worsens over time, impacting activities central to quality of life such as eating, drinking, dressing and using a phone.

Medications can help some patients, but up to half experience inadequate relief or intolerable side effects, leaving a large gap between drug therapy and invasive surgical options such as deep brain stimulation or focused ultrasound.

In the multicenter, double-blinded trial, 125 adults with essential tremor were randomized in a 2-to-1 ratio to wear either the active TPNS device or a sham device for 90 days.

Participants wore the wristband throughout most of their waking hours. The primary outcome was change in a modified Activities of Daily Living (mADL) score derived from the Essential Tremor Rating Assessment Scale, a measure favored by the U.S. Food and Drug Administration that reflects how tremors affect everyday function.

At 90 days, patients using the active TPNS device experienced a 6.9-point improvement in mADL scores, compared with a 2.7-point improvement in the sham group — a statistically and clinically meaningful difference, according to the investigators.

A 20% improvement in disability scores was reported by more than 60% of patients in the active-treatment group versus 25% in the control group. Improvements began to diverge after two weeks and continued to increase over time in the active group, while scores in the sham group plateaued and later worsened.

"What's particularly interesting is that the benefit didn't peak early and flatten out, as we often see with devices," Dr. Ondo said. "The improvement continued to grow over the 90 days, suggesting that the AI-driven adjustments may be particularly helpful at optimizing stimulation over time."

The device uses onboard sensors to continuously measure wrist motion and tremor characteristics, feeding that data to a cloud-based artificial intelligence algorithm. The system automatically adjusts stimulation intensity, pulse frequency and the balance of stimulation across the three nerves in real time.

Although tremor manifests in the hands, the abnormal rhythmic signals originate in the brain. Peripheral nerve stimulation is thought to send afferent signals back to the central tremor network, dampening pathological oscillations rather than mechanically counteracting hand movement.

Secondary outcomes also favored the active device, including patient- and clinician-reported global impressions of improvement and quality-of-life measures focused on hand function. Performance-based tremor scores improved modestly but did not reach statistical significance, a finding Dr. Ondo suggests highlights the importance of patient-centered measures.

"A one-point change on an activities-of-daily-living item can mean going from 'can't do it at all' to 'can do it independently,' which is a big improvement for patients," Dr. Ondo emphasized.

The device was generally well tolerated. The most common adverse event was mild skin irritation at the electrode site, and no serious device-related complications were reported. The authors note that longer-term studies are ongoing to determine whether benefits persist or increase with extended use.

For clinicians, the findings point to a potential new option for patients whose tremors are inadequately controlled with medication but who are not yet ready or eligible for surgery.