Green light for deep brain stimulator incorporating neurofeedback

A deep brain stimulation device capable of fine-tuning output on the basis of patient neuronal activity has been approved for the first time for use in Parkinson’s and a range of other neurological disorders.

The field of deep brain stimulation took a big step forward in June when the US Food and Drug Administration (FDA) approved a system with brain-wave sensing capabilities. The device, called Percept and developed by Medtronic, is the first commercially available DBS system that can record brain activity while simultaneously stimulating, which had not been clinically available previously. The FDA approved the therapy for Parkinson’s disease, essential tremor, dystonia, epilepsy and obsessive-compulsive disorder.

This is the first marketed deep neurostimulator that can also capture brain signals. Credit: Medtronic

Medtronic’s device not only offers the first commercial brain sensing DBS device for movement and psychiatric disorders, but also pushes the field of neuromodulation toward its ultimate goal: development of a closed-loop system in which a patient receives electrical stimulation adapted to their unique brain activity. “This is an important step in that direction,” says Mark George, director of the Brain Stimulation Lab at the Medical University of South Carolina. “If you can tailor the stimulation to some signal from the person’s brain, and do it in an intelligent way, you may be able to actually change the physiology of the brain.” That strategy could potentially induce plasticity and help treat the underlying illness, he says.

DBS is a therapy in which surgically implanted electrodes send electrical impulses into deep brain structures to thwart unwanted neurological activity. The impulses are generated by a pacemaker-like device typically implanted in the chest and connected via a lead to electrodes implanted in the brain. Since the first US regulatory approval in 1997, hundreds of thousands of DBS devices have been implanted in people with Parkinson’s disease, dystonia, essential tremor, epilepsy, obsessive-compulsive disorder and other diseases. Abbott Laboratories, Aleva Neurotherapeutics (Lausanne, Switzerland), Boston Scientific, Medtronic, PINS Medical (Beijing), and SceneRay (Suzhou, China) all offer commercial DBS systems.

Parkinson’s disease, DBS is typically prescribed after a patient has taken medication for several years. The medication may be working, but tends to wear off between doses. “This is the group of patients for whom the DBS really helps. It improves their off periods when the drugs aren’t working,” says Zelma Kiss, a neurosurgeon and professor of clinical neuroscience at the University of Calgary. “Instead of having these huge peaks and troughs, like going from walking normally to being frozen on the couch, their day is smoother,” she says. Patients with tremor are also good candidates for DBS, as drugs may not control tremor, she says.

DBS can certainly improve quality of life. For example, in a trial of a previous iteration of Medtronic’s commercial DBS device, Parkinson’s patients’ motor scores improved by 53%, hours of good mobility (‘on’ periods) increased by 20%, and medication needs decreased by 39% in the group that received DBS, compared with a control group receiving other medical therapy. The study of 251 people was completed in Europe and published in the New England Journal of Medicine in 2013.

Medtronic’s new dual function device, dubbed Percept, is more than an incremental improvement over the DBS systems clinicians have been using all these years. So far, commercial DBS devices have been unidirectional and can’t tune in to brain activity, and they stimulate regardless of the individual’s state — like shooting blind. So clinicians have to adjust stimulation settings in Parkinson’s patients on the basis of physiological signs, such as tremors or difficulty initiating movement. “That’s the technology that got the first FDA approvals, but everybody realized it was just dumb,” says George. “It was just an elegant crutch. In no way was it changing the physiology of the brain.”

The first approved brain stimulation system that can both sense and stimulate was developed by NeuroPace to treat focal epilepsy. NeuroPace’s device is implanted in the precise brain region where the seizure originates. Called the RNS System, the device is trained to identify the unique neural activity that occurs just before an epileptic seizure. It then sends out a preemptive electrical pulse to stop the seizure from happening. The device is always in sensing mode, except during brief bursts when it delivers stimulation pulses in response to abnormal brain activity that indicate seizures — something that may happen a few minutes each day. A nine-year prospective study with 256 patients completed in 2018 found that three out of four patients responded to the therapy, experiencing at least a 50% seizure reduction, and about a third of patients experienced a 90% reduction.

Medtronic’s new device Percept moves brain sensing beyond epilepsy to a broader range of indications, such as movement and psychiatric disorders. The device can stimulate continuously (24 hours a day) while simultaneously sensing brain signals, or local field potentials. Key to Percept’s advance was in filtering out the noise from the artificial stimulus pulse while tuning in to the relatively subtle signals of the brain. “Hearing that whisper next to that jet engine is really challenging,” says Steven Goetz, a distinguished engineer at Medtronic. His team achieved this using high-quality amplification and frequency-specific filters that enable the device to isolate and subtract the artificial signal from the intrinsic signals of the brain.

The technology for Percept emerged from the company’s research-only devices, the Activa PC+S and RC+S, which can both sense and stimulate and have been under investigation in humans since 2013. “We learned a lot [from the research devices] about how to tease out the signals,” Goetz says. Percept filters noise better than the research devices, simplifying the user experience, he says. Clinicians overseeing DBS treatment get a tablet programmer that tracks the patient’s brain signals both in and out of the clinic.

In simplifying, however, the company had to reduce Percept’s ability to analyze brain activity to a smaller frequency range. This limits what the device can sense, compared with the research version, but it nonetheless includes two important types of frequencies: beta and gamma oscillations. Abnormal beta oscillations are the most well-characterized electrophysiological biomarker of Parkinson’s disease. These frequencies of about 13 to 30 hertz observed in the brain’s subthalamic nucleus are a hallmark of the slow and rigid movements of Parkinson’s disease. Abnormal gamma frequencies in the 60 to 80 hertz range accompany excessive movement or dyskinesia, which is a side effect of high doses of Parkinson’s disease medication or of a combination of medication and stimulation.

But in its commercial iteration, the device cannot sense other suspect activity, such high-frequency oscillations above 100 hertz, or more complex activity such as phase–amplitude coupling — a synchrony of signals coming from both deep structures in the brain and the motor cortex (although these are less well characterized in the literature than beta oscillations). “There’s a limited amount of information that can be gleaned from Percept,” says Kiss at Calgary. “It’s simplified to make it palatable and comprehensible by a clinician who isn’t trained in neural engineering,” she says.

Still, insights can be extracted from Percept. Medtronic plans to aggregate patient data and make them available to researchers, using patient consent protocols and data security measures. This could allow researchers to track beta oscillations as the disease progresses or check how patients respond to medication, for example. And rather than being limited to capturing data during surgery or clinic visits, Percept can record data while patients are living their normal lives. “Never before have we had wires in the brains of so many people walking around,” says George. “So you wonder: is there a plan for the company to pool all this information so that scientists with good ideas can data mine?”

Goetz notes that Percept’s firmware and software can be updated as collective knowledge of biomarkers improves. This could enable the sensing of more complex signals without having to implant new devices in patients, he says.

Some of the data and knowledge for upgrading the system may come from Medtronic’s upcoming ADAPT trial, slated to begin enrolling in early 2021. The purpose of the study is to use DBS to better understand the rhythms of the brain associated with Parkinson’s disease and to develop an automated system for delivering more finely tuned stimulation based on those rhythms. Known as a closed-loop, or adaptive, system, such a technology would rely on algorithms and machine learning to make judgments about when and how to stimulate — something researchers and engineers have been working toward for years.

Adaptive capability exists, but will require a lot more investigation before it can be used more broadly or commercially, says Philip Starr, a professor of neurological surgery at the University of California San Francisco. “Adaptive stimulation is a big leap in sophistication, but it will come eventually,” he says. Broad use of Percept could help with that. In addition to the FDA approval in June, Percept received Europe’s CE mark in January and Japan’s regulatory approval in May. “The adaptive closed-loop stimulation system remains a prominent challenge,” says Hayriye Cagnan, an associate professor in the brain network dynamics unit at the University of Oxford. The fact that Percept’s software can be updated with adaptive capability is key, she says. No doubt the wealth of information that is about to emerge from people with Parkinson’s disease will shift researchers’ perspectives on this electrophysiological disorder.

Author information



Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Waltz, E. Green light for deep brain stimulator incorporating neurofeedback. Nat Biotechnol 38, 1014–1015 (2020).

Download citation


Quick links

Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing