Parkinson's: adaptive brain pacemakers, focused ultrasound and AI early detection

A brain stimulator that listens and adjusts itself, and a treatment with no incision at all. Both can calm symptoms; neither slows the disease itself.

What these technologies do, and what they do not

Parkinson's symptoms such as tremor, stiffness and slowness come from the loss of dopamine-producing brain cells. Medicines like levodopa replace dopamine, but their effect rises and falls through the day. The technologies below aim to smooth out symptoms. None of them slows the loss of brain cells or stops the disease from progressing.

Adaptive deep brain stimulation

Standard deep brain stimulation (DBS) uses thin wires in the brain connected to a pacemaker-like device in the chest. It delivers the same electrical stimulation around the clock. When medication levels peak, that fixed stimulation can add to side effects such as involuntary movements or slurred speech; when medication wears off, it can be too weak.

Adaptive DBS listens as well as stimulates. Through the same wires it records brain activity in the beta frequency range, which rises when dopamine is low. When beta activity climbs, stimulation increases; when it falls after a medication dose, stimulation is turned down.

• Approval: Medtronic's BrainSense adaptive DBS received a CE mark in January 2025 and FDA approval in February 2025, for use with its Percept PC and RC devices.
• Evidence: the ADAPT-PD trial, published in JAMA Neurology in 2025, found adaptive stimulation kept good "on" time comparable to well-programmed standard DBS. One mode used about 15% less battery energy. Around 98% of those who completed set-up chose to stay on adaptive mode.
• Limits: the study showed equivalence rather than clear superiority, was partly unblinded and was sponsored by Medtronic. Programming is demanding, and not every patient could be kept on adaptive mode.

Focused ultrasound: treatment without an incision

The patient lies awake in an MRI scanner wearing a helmet with about a thousand ultrasound emitters. Each beam is harmless on its own; where they meet deep in the brain, they heat a spot the size of a pea. The team first warms the target slightly to check that tremor eases and no side effects appear, then heats it to create a small permanent lesion.

Targets include the thalamus for tremor and the globus pallidus or nearby pathways for stiffness, slowness and medication-induced movements.

• US approvals: one-sided treatment of tremor-dominant Parkinson's in 2018, one-sided pallidotomy in 2021, and staged treatment of both sides in July 2025.
• Evidence: a multicentre study of staged two-sided treatment, published in The Lancet Neurology in 2026, reported meaningful improvement in motor symptoms and a large reduction in motor fluctuations, sustained at 12 months.
• UK position: in December 2024 NICE concluded that evidence on long-term safety and effectiveness was not yet sufficient and recommended use only in research.

Spotting Parkinson's earlier with wearables and AI

By the time movement symptoms appear, a large share of the dopamine cells are already lost. Researchers are looking for earlier signals.

• Smartwatch movement data: a study of more than 100,000 UK Biobank participants, published in Nature Medicine in 2023, found that a week of wrist accelerometer data could identify people diagnosed with Parkinson's up to seven years later, more accurately than genetics or blood tests in that study.
• Voice analysis: Parkinson's can subtly change the voice early on. Machine learning tools that analyse recordings have shown promise in small studies.

The catch is that Parkinson's affects only about 1% to 2% of older adults. Even an accurate test applied to everyone would produce many false alarms. That is why the devices currently cleared, such as StrivePD on the Apple Watch and the Personal KinetiGraph, are approved only to monitor symptoms in people already diagnosed.

Access and cost

• United States: DBS is covered by Medicare and private insurers for medically refractory Parkinson's. Medicare covers one-sided focused ultrasound for tremor; coverage for other uses varies by insurer.
• United Kingdom: DBS is available through NHS specialist centres. Focused ultrasound for Parkinson's is limited to research or private payment.
• European Union: both are offered in specialist university centres, with reimbursement varying by country.
• Turkey: DBS is covered by SGK at contracted hospitals for eligible patients. Focused ultrasound is offered at some private centres in Istanbul; we found no official data on SGK coverage.

What is being tested next

• Two-signal adaptive DBS: trials such as one at UCSF combine deep brain signals with recordings from the brain's surface to control both stiffness and involuntary movements.
• Ultrasound to open the blood-brain barrier: low-energy ultrasound with tiny bubbles can briefly let drugs into specific brain areas. Early trials are testing delivery of enzyme therapy in Parkinson's.
• Disease-modifying drugs: a phase 3 trial of prasinezumab, an antibody targeting the protein alpha-synuclein, began in 2025.

What to ask your neurologist

• Am I a candidate for DBS, focused ultrasound or neither, and why?
• If I already have a Percept device, can adaptive mode be switched on?
• What are the realistic risks to speech and balance for me?
• Which centre near me has the most experience with this procedure?

Questions patients ask