Despite its chequered past, deep-brain electrical stimulation is finally showing some signs of success. By Liam Drew
Neurosurgeon Sameer Sheth positions an electrode for deep-brain stimulation.Credit: Sameer Anil Sheth, Baylor College of Medicine
In October 2019, a 36-year-old woman with severe depression entered the University of California, San Francisco (UCSF) Medical Center. Her illness had begun at the age of 11. By 2019, she was going through a four-year episode of depression characterized by low motivation, an inability to experience pleasure, and suicidal ideation. Her depression was unresponsive to all available treatment options and so severe that she was no longer able to work or even drive a car.
At UCSF, neurosurgeons implanted ten electrodes in the woman’s brain, in areas that regulate mood. Every electrode had 16 active sites, meaning the researchers had access to 160 locations. And each site could either zap current into her brain — a technique called deep-brain stimulation (DBS) — or record electrical signals resulting from her neural activity.
Part of Nature Outlook: Depression
For ten days, a team of clinicians recorded and stimulated her brain, while carefully charting the woman’s mood. They observed how her neural activity changed when her mood shifted up and down, seeking patterns at one or more recording sites that correlated with her mental state. And they looked for sites that, when stimulated, improved her mood.
As described in a paper1 published in October 2021, she finished this brain-mapping exercise with two electrodes permanently implanted in brain centres crucial to her depression: one for recording, one for stimulating. When the recording electrode detects telltale signs of a low mood, it automatically triggers the stimulating one to deliver 6 seconds of mood-elevating current pulses.
Within two weeks of activating this ‘closed-loop’ system, the woman’s scores on a depression-rating scale had dropped by more than 50%. And psychiatrist Katherine Scangos, who led the study at UCSF, says that after a few months, the participant entered continuing remission, with depression ratings consistently in the healthy range.
Katherine Scangos (left), a psychiatrist at the University of California, San Francisco, conducts an appointment with her clinical-trial patient Sarah.Credit: Maurice Ramirez for UCSF
The study is an attempt to rebuild the prospects of DBS as a treatment for depression, which took a huge hit when the first generation of approaches failed to show efficacy in two high-profile, industry-sponsored studies2,3 in the 2010s. In those first efforts, researchers only stimulated the brain; they did not record its activity.
“We can’t claim closed loop as effective off one patient,” says Scangos, who now consults at UCSF while working for the biotechnology firm Neumora in South San Francisco, California. But this work and approaches by others are substantiating the idea that listening to the brain can shed light on the nature of depression4 — and provide options for people whose depression has resisted all other treatments.
Starting over
Modern DBS was developed in the late 1980s to treat the debilitating tremor that defines advanced Parkinson’s disease. Permanently implanting electrodes in a specific brain area — chosen on the basis of the known pathophysiology of Parkinson’s — and stimulating it continuously with high-frequency electrical pulses (typically 130 hertz) transformed many people’s lives. Now, almost 300,000 people have received this treatment, and more than 92% report being happy with its effects5.
Depression was one of numerous other brain disorders to which researchers attempted to apply this template. In the early 2000s, two areas of the brain rose to prominence as potential targets, each championed by different research groups. One was the ventral capsule/ventral striatum (VC/VS), a region associated with reward processing. The other — selected by Helen Mayberg, a neurologist now at the Icahn School of Medicine at Mount Sinai in New York City — was the subcallosal cingulate (SCC), a hub in emotion-regulating circuitry associated with the expression of negative mood.
Initial small and non-placebo-controlled studies suggested that stimulating these areas helped most participants with treatment-resistant depression. Quickly, two medical-device companies took a target area each and organized the two ill-fated large, double-blind trials — each with a control arm in which participants were implanted with electrodes that remained switched off.
In 2016, after neither trial showed any benefit from the intervention, prominent DBS researchers convened a meeting to discuss what had gone wrong6. Mayberg, convinced by the successes she had observed in her initial studies, argued that the trial evaluating SCC stimulation had several flaws. She said there had been problems with electrode placement and participant selection. Moreover, she said, the trials were too short, arguing that people can take longer than the observed six months to see improvements. “I know this works,” she says now, referring to her own earlier work. “My job is to reverse engineer what went right.”
It’s a controversial stance. “If you’re going to believe in anything, it’s randomized clinical trials,” says Stanley Caroff, a psychiatrist at the University of Pennsylvania in Philadelphia.
Also present at the 2016 meeting was Sameer Sheth, a neurosurgeon at Baylor College of Medicine in Houston, Texas. Sheth had seen improvements in his own patients when using the original SCC stimulation procedure and had been further convinced of its efficacy by seeing people relapse when the batteries in their devices ran out. Sheth thought there was enough evidence to persist with existing continuous-stimulation protocols. But moving forwards, he wanted to individualize people’s treatment by characterizing their mood-related brain activity, then using that to guide where and what type of electrical stimulation should be delivered. He is now testing this proposition in a small clinical trial.
Sheth and Scangos’s work is united by a desire to confront depression’s interpersonal variations. Scangos criticizes the first-generation studies for ignoring this heterogeneity. “Every patient was treated the same,” she says. “Everyone was stimulated in one brain region. And there was no feedback from the brain.”
In July, the US Food and Drug Administration granted breakthrough device designation to Abbott Laboratories in Abbott Park, Illinois, to restart DBS work. An Abbott spokesperson says the company is planning trials that will use an adjustable DBS system to target the SCC with continuous stimulation. But, for many academic researchers the use of brain recordings is essential for pushing this technology forward.
Mayberg has since incorporated brain recordings into her latest studies to monitor the long-term effect of stimulation. But the intensive brain mapping that Scangos and Sheth use goes much further in seeking to adaptively individualize people’s treatment and to reveal more about the nature of depression and its diversity. “The primary goal is for the patients to get better,” Sheth says. “But along the way, we’re also going to learn a heck of a lot about the brain.”
Listening in
An X-ray showing implanted DBS electrodes.Credit: Zephyr/SPL
Confident that continuous SCC stimulation can be effective, Mayberg ran an unblinded study using improved neuroimaging methods to better position the electrodes7. Nine of the 11 participants benefited after one year of treatment. Such imaging, Mayberg says, should be a prerequisite for future trials. She also hopes that new neural recordings adjacent to the stimulation site will convince her critics, mentioning that unpublished data indicate brain activity changes in a way that predicts recovery.
In 2021, Sheth reported results from the first participant in his trial8 — a 37-year-old man with treatment-resistant major depressive disorder. The man was implanted with stimulating electrodes in both his VC/VS and SCC and an array of electrodes was placed on his cerebral cortex for an initial nine-day inpatient characterization of his brain activity. Sheth’s team found patterns of cortical activity that occurred when the man was in a positive mental state. Then, the researchers changed the ratio of VC/VS to SCC stimulation and altered the stimulation properties of each electrode until they empirically found parameters that evoked matching patterns. The monitoring electrodes were then removed, and these parameters were used for long-term, simultaneous VC/VS and SCC stimulation.
After 18 weeks of stimulation, the man’s depression entered remission. More than a year later, he was depression-free. Around four months after he had entered remission, however, Sheth and his team spent roughly seven weeks gradually dialling down the DBS without the man knowing. As they did so, his depression worsened. When the stimulation was resumed, his mood improved again — strong evidence that the DBS was working.
Sheth’s clinical result, like Scangos’s, is preliminary. Each trial is designed to test 12 participants. So far, Sheth has implanted devices in four people, and Scangos’s UCSF colleagues have implanted three. Scangos and Sheth each anticipate that their trials will take several years to complete.
In Scangos’s first participant, she used 6-second bursts of VC/VS stimulation to elevate the person’s mood — other people might find that different targets work better. And the biomarker the team detected was surprisingly simple: an increase in high-frequency oscillations in the amygdala, a brain region associated with feeling fear. “My hypothesis is that in different patients we’ll find different sites and different biomarkers,” she says.
One unanswered question with this closed-loop stimulation is whether it operates in a fundamentally different way from continuous stimulation. The slow improvements resulting from conventional DBS suggest that it treats depression by gradually changing the brain. Closed-loop stimulation might not do this, but instead continuously controls a person’s mood. “Our model is that by treating the mood state itself, you can treat the disorder,” Scangos says.
There are also substantial concerns about the scalability of approaches that require intensive in-patient screening. One possibility is that neural biomarkers will correlate robustly with specific clinical symptoms. But, Sheth says, “We’re hoping that we’ll learn some things that will generalize to information that can be detected non-invasively.” For example, the use of scalp-based electroencephalograms, which are much easier to use than are systems based on implanted electrodes.
What’s key for Sheth is that DBS is informed by neural data. With the original approaches, he says, improvement in a person’s symptoms was “a tick in our favour”. But if there was no improvement, “we had no idea why, because there was no feedback from the brain”. More data-driven DBS treatment designs might enable the technology to fare better in future large-scale, placebo-controlled trials. If it’s ever to serve large numbers of people, those are the tests that must one day be passed.
Nature 608, S46-S47 (2022)
doi: https://doi.org/10.1038/d41586-022-02209-6
This article is part of Nature Outlook: Depression, an editorially independent supplement produced with the financial support of third parties. About this content.
References
Scangos, K. W. et al. Nature Med. 27, 1696–1700 (2021).
Dougherty, D. D. et al. Biol. Psychiatry 78, 240–248 (2015).
Holtzheimer, P. E. et al. Lancet Psychiatry 4, 839–849 (2017).
Scangos, K. W. et al. Front. Hum. Neurosci. 15, 746499 (2021).
Hitti, F. L. et al. J. Neurosurg. 132, 205–210 (2019).
Bari, A. A. et al. J. Neurol. Neurosurg. Psychiatry 89, 886–896 (2018).
Riva-Posse, P. et al. Mol. Psychiatry 23, 843–849 (2018).
Sheth, S. A. et al. Biol. Psychiatry 92, 246–251 (2022).