Her relapse into depression felt like defeat—but it offered vital clues to achieving lasting psychiatric relief.
The 67-year-old woman from Alabama had already endured four major depressive episodes in her decades-long battle with mental illness. After exhausting numerous medications and other therapies, in 2015 she turned to an experimental last resort: deep brain stimulation, or DBS.
Neurosurgeons implanted electrodes a few inches below her skull, targeting a small bundle of neural fibers in a brain region behind the forehead that acts as a crucial hub for mood regulation. Thin wires connected the electrodes to a pulse generator discreetly inserted in her upper chest. Once activated, the device delivered a steady stream of high-frequency electricity, gently buzzing the targeted circuits to disrupt maladaptive patterns and, like a pacemaker for the brain, restore a healthier balance of neural activity.
At first, the treatment seemed to be working. The woman’s despair lifted, and she edged closer to remission. Watching football for hours with her husband on Sundays started to feel tedious—in a good way. Her desire to get off the couch and pursue other activities had returned.
An X-ray image shows two pairs of DBS electrodes implanted for depression treatment. Department of Neurosurgery, Baylor College of Medicine
But four months on, the darkness crept back in. The woman’s sudden downturn blindsided the medical team that had been closely monitoring her recovery. The doctors had to make three adjustments to the implant’s stimulation parameters, slowly increasing the voltage, before her condition finally stabilized—an agonizing couple of months.
When the clinicians reviewed the data later, they realized that the electrodes embedded in the woman’s brain had detected trouble brewing before she did. Subtle shifts in the electrical patterns coursing through her neural fibers had flagged the impending relapse weeks before her outward symptoms reappeared. If clinicians had acted on those signals, they might have adjusted the stimulation settings in time to prevent her relapse.
It’s a thought that weighs on Patricio Riva Posse, the psychiatrist at Emory University School of Medicine, in Atlanta, who treated the woman. Looking back now, he says, had he known that the brain’s circuits were off-kilter, “I would have taken action earlier.”
Fortunately, Riva Posse no longer has to dwell on what could have been. Together with colleagues at the Icahn School of Medicine at Mount Sinai, in New York City, and Georgia Tech, in Atlanta, he is now leveraging advances in DBS hardware and artificial intelligence (AI) to design more precise treatments for depression. The team’s goal is to base treatment on objective neural data rather than the subjective measures—patient accounts, clinical hunches, questionnaires, mood scales—that dominate psychiatry today.
The pioneering neurologist Helen S. Mayberg co-led the team with Riva Posse and Christopher Rozell of Georgia Tech. Ultimately, they hope to enable preemptive interventions rather than regretful, after-the-fact adjustments.
It’s a new frontier for psychiatry. The field has long been one of the few medical disciplines without objective measures to guide treatment decisions. But with the advent of real-time brain monitoring with AI-driven analytics, that could finally change. “It’s a whole different mindset now,” says Martijn Figee, a Mount Sinai psychiatrist involved in the research. “My intuition, unfortunately, is not 100 percent [accurate],” he acknowledges. “So ultimately, I would always trust the brain more.”
Researchers are developing “an automatic alarm system”—an AI-driven tool designed to continuously monitor device output and flag warning signs of relapse.
Other research groups are pursuing similar goals, aiming to move beyond the one-size-fits-all approach that has long defined DBS treatment for mental health and replace it with precise stimulation tailored to individual needs. While standardized protocols benefit around 60 percent of people with treatment-resistant depression, they still leave a substantial minority without meaningful relief.
No DBS platform is yet approved for treating depression, although some first-generation devices are getting close. Those are rooted in decades-old technology, however, while the Mount Sinai team and others are breaking new ground. They are investigating analytical frameworks that harness brain data to predict relapses, optimize stimulation parameters, or dynamically adjust device output in a responsive, closed-loop manner.
“The field is just at a super exciting place,” says Benjamin Davidson, a neurosurgeon at the Sunnybrook Health Sciences Centre, in Toronto. “Things are starting to move at a kind of dizzying pace.”
The Origins of DBS for Depression
That momentum is a relatively recent phenomenon in a field that, for the past two decades, has progressed through baby steps. Beset by commercial and clinical setbacks, little has changed over the years aside from the adoption of newer surgical techniques. The biggest advance was an imaging-guided surgical approach called tractography that allows for greater precision in electrode placement, informed by connectivity patterns between bundles of brain fibers rather than anatomical landmarks alone.
“The story is one of iteration to optimize and refine the targeting using new neuroscience tools,” says Mayberg, who launched the world’s first DBS trial for treatment-resistant depression in the early 2000s at the University of Toronto. “The procedure, as envisioned and published in 2005, is, in essence, what we continue to do today,” she says.
DBS is primarily used to manage movement disorders such as essential tremor and Parkinson’s disease. For those ailments, it’s an established and approved therapy that can drastically reduce symptoms such as shaking and muscle rigidity.
But Mayberg was inspired by the discovery of a brain region called the subgenual cingulate (SGC), which plays a key role in acute sadness and the effects of antidepressant treatments. She theorized that stimulating this area might alleviate severe, treatment-resistant depression. Her patients were people who had typically tried several types of antidepressant medications and more drastic measures, like electroconvulsive therapy, without finding any relief.
While the treatment didn’t work for everyone, many did feel better. Six months after surgery, 12 of the 20-person cohort experienced a profound lifting of their depressive symptoms, with 7 going into full remission. The effect was lasting, with many of those individuals continuing to report benefits to this day, according to Andres Lozano, the University of Toronto neurosurgeon who performed the operations.
Mayberg’s hypothesis, it would seem, had proved correct.
Learning from DBS Failures
Yet, for all its early potential, DBS never gained traction as a mainstream psychiatric treatment. It is occasionally used today for people with debilitating obsessive-compulsive disorder, but the technique remains unapproved for depression and is largely confined to research trials—some of which have ended in dispiriting, high-profile failure.
One of the most notable setbacks occurred in 2013. The device company St. Jude Medical set out to replicate the findings of Mayberg’s study in a randomized trial, with plans to enlist 200 participants. But the study was halted prematurely after only 90 patients had been enrolled. An interim analysis had found the therapy was no more effective than sham stimulation.
It was a crushing blow to the field. Mayberg and others struggled to continue their research, as funding agencies and the scientific community at large grew increasingly skeptical about the viability of DBS for depression.
With the benefit of hindsight, however, many researchers now believe that the St. Jude failure owed more to the study’s design flaws than to any inherent shortcomings of DBS itself. A longer-term follow-up of participants indicated that the treatment’s antidepressant effect steadily strengthened. The trial may simply have measured responses on the wrong timeline. Plus, the neurosurgical placement of the DBS electrodes relied on an outdated understanding of brain connectivity, leading to suboptimal positioning. This may have delayed the therapeutic response past the initial 6- to 12-month assessment window.
These missteps likely undermined the study’s results, the trial investigators later concluded. But with the right trial design, most experts anticipate that future studies will succeed. “That could make a huge difference,” says Darin Dougherty, a psychiatrist at Massachusetts General Hospital, in Boston. “Hopefully those lessons learned will be enough to get it over the top.”
A patient identified as Sarah participates in a trial at UC San Francisco of the first fully closed-loop DBS system for depression. Maurice Ramirez
The biomedical company Abbott (which acquired St. Jude in 2017) is now conducting a do-over study at 22 sites across the United States; Dougherty, Figee, Riva Posse, and other leaders in the field are involved in the effort. The 100-person trial, launched in September 2024, could finally lead to regulatory approval and wider-scale adoption of DBS as a treatment strategy for depression.
But Abbott’s study takes a “set-it-and-forget-it” approach, in which stimulation parameters are programmed during initial visits and remain largely unchanged over time. The settings are generally standardized across patients, with a common pulse width and frequency fixed at around 90 microseconds and 130 hertz, respectively. Only the amplitude of stimulation, measured in volts, is typically adjusted to accommodate individual tolerances or symptom severity.
While this treatment approach is simple and scalable, it lacks the adaptability to respond to the dynamic nature of depression and its varying symptoms from one individual to the next. This limitation stems in part from a technological shortcoming of the Abbott platform: It can deliver precisely tuned electricity, but it lacks the ability to sense and record neural activity. Without this feedback mechanism, the device cannot detect shifts in brain states that might signal a relapse or a need for parameter adjustments, leaving clinicians reliant on patients’ reports.
In contrast, newer DBS devices for epilepsy and movement disorders can both stimulate and record signals. Medtronic’s Percept system and NeuroPace’s Responsive Neurostimulator, for example, offer real-time feedback capabilities, which could allow for more adaptive therapies. Researchers want to bring that flexibility to DBS for depression.
How Responsive DBS for Depression Works
Consider again the example of Riva Posse’s 67-year-old patient. As described in Nature two years ago, this woman received a research-grade version of the Percept platform that detected signs of neural instability five weeks before her clinical symptoms reappeared.
“Before the patient knew anything was wrong—before there was even a hint of behavior that could seem symptomatic of a relapse—the brain signal was headed in the wrong direction,” says Rozell, the neuroengineer at Georgia Tech who developed the AI model used to interpret the woman’s brain activity patterns.
Rozell’s model combined a neural network classification scheme (for analyzing brain signals) with a generative causal explainer (for identifying key activity patterns). His work uncovered a distinct biomarker that reliably differentiated between states of depression relapse and recovery. Intriguingly, the biomarker also reflected changes in sleep quality, a telling early indicator since poor sleep patterns often precede the return of depression symptoms.
Depression can take many forms: Some people experience it as emotional despondency, while others struggle with obsessive thoughts or a loss of pleasure.
But the insights provided by Rozell’s model came too late to help the patient in the moment—they were validated only after her relapse had occurred. To address this limitation, the researchers are now refining the approach for real-time use, aiming to develop what Mayberg calls “an automatic alarm system”—an AI-driven tool designed to continuously monitor device output and flag warning signs of relapse.
Such a system could prompt clinicians to intervene before these brain signals escalate into a full-blown depressive episode. Simultaneously, it could filter out false alerts from patients, providing reassurance to users who might otherwise interpret normal stress or anxiety as signs of an impending relapse. Informed by this neurofeedback, psychiatrists might then choose to fine-tune stimulation settings. Or they might proactively recommend additional support, such as psychotherapy or medication adjustments.
Closing the Loop for DBS
Going one step further, researchers from the University of California, San Francisco, are exploring a fully closed-loop DBS system for depression that removes some of the need for human decision-making. Their approach empowers the device itself to automatically adjust stimulation parameters in real time based on brain activity.
Reporting on their first patient—a woman in her 30s named Sarah, who withheld her last name for privacy—the UC San Francisco team documented transformative improvements in her mood, emotional balance, everyday functioning, and overall outlook on life, all in the first week after the implant was switched on.
Sarah reports that the closed-loop DBS system restored pleasure and purpose to her life. John Lok
“My life took an immediate upward turn,” Sarah said at a 2021 press conference announcing the study’s early findings. “Hobbies I used to distract myself from suicidal thoughts suddenly became pleasurable again. I was able to make small decisions about what to eat without becoming stuck in a morass of indecision for hours,” she said, adding, “the device has kept my depression at bay, allowing me to return to my best self and rebuild a life worth living.”
According to Andrew Krystal, the UC San Francisco psychiatrist leading the effort, similar benefits have since been seen in at least two other recipients of the closed-loop DBS device.
In each case, patients first undergo an intensive 10-day exploration of their typical neural activity, with 10 electrodes—targeting five locations on each side of the brain—temporarily implanted. During this period, researchers administer a battery of tests to identify the most effective sites for both stimulation and sensing. Once the optimal locations are determined, a second surgery is performed to implant the permanent DBS system, now simplified to just two electrodes: one dedicated to delivering stimulation and the other to recording neural activity.
When the recording electrode detects brain activity associated with depression—an event that can happen hundreds of times per day—it prompts the other electrode to deliver a brief burst of electricity lasting a few seconds. This approach stands out not only because it operates automatically in response to real-time brain activity, but also because it employs intermittent, on-demand stimulation rather than the continuous stimulation more commonly employed in DBS for psychiatric conditions.
This adaptive and dynamic feedback strategy may be especially well suited to addressing the day-to-day fluctuations in mood and emotional strain that can make depression so hard to live with, notes Katherine Scangos, a psychiatrist who participated in the study. Patients have told her that receiving stimulation at key moments—like during a stressful interaction at the checkout line of a grocery store—helped prevent them from spiraling into distress. “They could really tell that they were getting the stimulation when they needed it most,” says Scangos, who joined the staff of the University of Pennsylvania last year.
Identifying the right sites and parameters is an intricate and labor-intensive process, and it’s not always immediately clear which settings will work best, according to UC San Francisco neurosurgeon Kristin Sellers. All the data they collect creates a “curse of bounty,” she says. Yet, in her view, the outcomes demonstrate the effectiveness of taking this personalized approach. “No one has an identical implant,” she says.
New Ideas on DBS for Depression
Meanwhile, a team at Baylor College of Medicine, in Houston, is pursuing a different approach to customized DBS for depression. The team’s standardized implant consists of two coordinated sets of electrodes: One targets the SGC brain region involved in profound sadness, while the other stimulates a reward-and-motivation hub deep in the brain’s basal ganglia.
The customization happens on the front end during the initial surgical procedure, when clinicians temporarily place another 10 electrodes into the brain that take recordings via electroencephalography (EEG). This method tracks brain waves and, as patients undergo various tests and activities, allows the Baylor team to map relevant neural networks and connections. At the same time, the doctors can fiddle with the amplitude, pulse width, frequency, and shape of the stimulation field.
“Then we can basically design bespoke stimulation parameters for that individual that are going to move that person’s network in the right direction,” explains Sameer Sheth, the neurosurgeon leading the project. Sheth and his colleagues have treated seven people, with promising initial results.
Any of these highly individualized approaches will involve additional surgical procedures and lengthy stays in the hospital. But as Dougherty of Massachusetts General Hospital points out, “We need to do this invasive research first so that we might be able to use noninvasive approaches later.”
He imagines a future in which electrodes on the scalp or advanced imaging techniques could identify optimal targets and guide treatment adjustments. Even then, however, if DBS requires highly personalized programming, it will be challenging to make it accessible to the millions of people worldwide in the throes of depression.
“The question will always be about the scalability of things,” says Volker A. Coenen, a neurosurgeon at the University of Freiburg Medical Center, in Germany. Coenen is therefore focusing his energy on testing a standardized DBS protocol, one that involves implanting the Vercise Gevia system from Boston Scientific into an area of the brain known as the medial forebrain bundle.
In his view, this brain region offers a more direct and efficient pathway to reward systems and emotional-regulation networks. Still, the various brain regions under consideration are all interconnected, which explains why they all seem to offer some degree of therapeutic benefit. “You can perturb the network from different angles,” Coenen says.
The Road Ahead for DBS
So, which site is best? The answer likely depends on the specific symptoms and underlying brain circuits unique to each individual, says Alik Widge, a psychiatrist and biomedical engineer at the University of Minnesota, in Minneapolis.
“There’s no such thing as DBS for depression. There’s DBS for treating specific cognitive-emotional syndromes,” he argues—and different targets will be suited for accessing different aspects of the disorder. Depression can take many forms: Some people experience it as emotional despondency, while others struggle with obsessive thoughts or a loss of pleasure.
The optimal stimulation method may also vary. Continuous stimulation may work best for people whose depression follows a steady, persistent course, while intermittent or responsive stimulation may be more appropriate for those whose symptoms fluctuate with daily ups and downs. “It’s like the difference between weather and climate,” says Riva Posse—some people may need an umbrella for passing showers, while others need to reinforce their homes against rising tides.
Ultimately, whether they’re tweaking stimulation parameters, finding the best brain targets, or making stimulation respond to real-time brain signals, the goal for researchers in the field remains the same: to create a neurologically precise approach to treating depression in people who have found no relief. “There are so many levers we can press here,” says Nir Lipsman, who directs the Harquail Centre for Neuromodulation at Sunnybrook, in Toronto. He’s confident that at least some of these efforts will unlock new therapeutic possibilities.
“The field is experiencing a kind of reset,” Lipsman adds. Now, with neural activity as a guide, the brains of people undergoing DBS should likewise experience a kind of reset as well.
From Your Site Articles
Related Articles Around the Web
