Highly Individualized Deep-Brain Stimulation Helps a Patient With Severe Treatment-Resistant Depression
Over the past three years, researchers at The University of California, San Francisco (UCSF)—three of them BBRF grantees and one a member of BBRF’s Scientific Council—have published three papers that have led to an experimental new approach to treating brain and behavior disorders, using deep-brain stimulation (DBS).
All three papers are notable, describing the progression of an idea from laboratory to bedside. The most recent, appearing in Nature Medicine in September 2021, was the subject of a New York Times story. It signaled that a concept which until then had been theoretical had now reached the point of helping a patient: “A ‘Pacemaker for the Brain’: No Treatment Helped Her Depression—Until This.”
The patient who received the new treatment had been depressed since childhood and had not been helped by 20 different combinations of medicines, or by electroconvulsive therapy (ECT) or non-invasive transcranial magnetic stimulation (TMS). “Within a few weeks” of the beginning of her new treatment, she told the Times, “the suicidal thoughts just disappeared. Then it was a gradual process where it was like my lens of the world changed. The device has kept my depression at bay, allowing me to return to my best self and rebuild a life worth living.”
Deep-brain stimulation treatment for patients with severe and unresponsive depression was pioneered on an experimental basis beginning in 2005 by Helen S. Mayberg, M.D., and colleagues. Dr. Mayberg is a BBRF Scientific Council member, three-time BBRF grantee and 2007 winner of BBRF’s Falcone Prize. This story will explain the concept behind a new application of DBS and will explore its potential implications for patients with depression and perhaps other psychiatric illnesses. Although the initial results have been both intriguing and encouraging, researchers involved in designing and delivering the treatment are the first to point out that it has only been tested in a single patient. At this point, they caution, it is impossible to know how it will work in other patients.
Testing of the new treatment was the focus of a 2018 BBRF Young Investigator grant to Katherine W. Scangos, M.D., Ph.D., first author on the new paper. Inklings of the unconventional idea driving the research trace farther back, perhaps most distantly to two BBRF Young Investigator grants awarded in the 1990s to Andrew D. Krystal, M.D., a psychiatrist and expert on brain stimulation and mood disorders. Dr. Krystal is co-leader of the UCSF research team that delivered the treatment, along with Edward F. Chang, M.D., a neurosurgeon and authority on using implantable DBS devices to treat epilepsy.
The team’s September 2021 paper described the application of the new approach in just a single patient. A battery-powered DBS “pacemaker” was surgically implanted within the brain of a 36-year-old woman and programmed to deliver electrical stimulation at specific moments over the course of each day. It was placed in a location where its pulses were expected to help alleviate symptoms of major depression.
Unlike past “open-loop” tests of DBS in treatment-resistant depression, in which stimulation is delivered constantly following implantation of the device, this was a proof-of-concept test for a “closed-loop” approach. The DBS device would be activated intermittently throughout each day for only seconds at a time, and—here is the most notable innovation—only at moments when a sensor placed in another part of the brain detected a specific brain-wave pattern linked in prior tests with the onset of this particular patient’s depressed moods (see illustration).
What is new, then, about the approach is not just that the stimulation is intermittent—and limited to 300 times per day, maximum—but that it is triggered by a signal coming from elsewhere in the brain and relayed to the device. This highly personalized treatment design was not arrived at by guesswork, but only after a 10-day-long brain-mapping process in which brain-wave signals in this patient, as measured by EEG (electroencephalography), were painstakingly correlated with fluctuations in the patient’s moods.
A KEY INITIAL DISCOVERY
When Drs. Krystal, Chang, and Scangos performed their bold clinical test of closed-loop neuromodulation, they were building upon research in which Dr. Chang had previously been involved. Dr. Chang co-led a team with BBRF Scientific Council member Vikaas S. Sohal, M.D., Ph.D., a 2009 BBRF Young Investigator who had trained in the Stanford University lab of BBRF Scientific Council member, two-time grantee and recent Lasker Award winner Dr. Karl Deisseroth.
In November 2018 Drs. Sohal, Chang and colleagues reported in the journal Cell their discovery of a “subnetwork” (or “subnet”) in the brain connecting the amygdala and hippocampus, two areas centrally involved in the processing of emotions. They were surprised to find that recurrent and highly specific variations in EEG signals emanating from this subnetwork were directly correlated with worsening mood in 13 human subjects, attributable by the subjects to the onset of anxiety. The variations occurred in the EEG bandwidth called the beta band, which registers neurons oscillating at between 13 and 30 times per second.
It has long been assumed that human brain networks somehow encode variations in mood, although precisely how they do so remains unknown. The insight provided by Drs. Sohal, Chang, and colleagues was among the fruits of President Barack Obama’s “Brain Initiative” and backing by DARPA, the Defense Advanced Research Projects Agency, and the Dolby family. The researchers made a direct connection between the beta-band signal in the amygdala-hippocampus “subnet” which corresponded directly with a specific change in mood—and not just in one individual but in 13. The team was surprised by this result.
Interestingly, the mood signal was detected in 13 of the study participants, but not in 8 others. All 21 subjects suffered from epilepsy that had resisted treatment with medications. (The study was conducted to learn more about their brain activity prior to brain surgery designed to prevent seizures). But the anxiety signal was seen only in the 13 who had been assessed previously with comparatively high levels of anxiety—none of the others.
Dr. Krystal notes that the “DARPA subnets” study of 2018, as he and others call it, was notable in part because the EEG signal correlated with the presence of anxiety in the 13 subjects. This suggested to Drs. Krystal, Chang, and Scangos that it might be possible to find other biomarker-like signals in specific patients that would signal the onset of other psychiatric symptoms.
Their follow-up work would focus on the 36-year-old patient with childhood-onset treatment-resistant major depression—her name is Sarah—who became the first patient with psychiatric illness to benefit from closed-loop neuromodulation.
PREPARING THE FIRST PATIENT
The work advanced in two major steps. The first step involved implanting electrodes in Sarah’s brain—an invasive procedure requiring surgery—and systematically assessing her response over 10 days when the team applied electrical stimulation across the brain, with particular attention to five brain areas: the subgenual cingulate, amygdala, hippocampus, ventral capsule (part of the striatum), and orbitofrontal cortex. Cautiously, the team delivered stimulation at varying intensity at each of the locations. As they did, they communicated continuously with Sarah, who conveyed what impact each stimulation had on her mood and feelings.
The result is captured in a graphic conceived by Dr. Scangos and which appeared in the team’s January 2021 paper, published in Nature Medicine. It summarizes what Sarah experienced at each step of the experiment—what Dr. Krystal calls “the clinical effects of neurostimulation,” delivered widely across both hemispheres of the brain.
Here, the team observed something that had also been seen in the earlier “DARPA subnets” study. “We saw in that study that you could elicit changes in emotion-related symptoms—quickly and immediately—when applying stimulation at specific locations,” Dr. Krystal explains.
This is exactly what they now saw in Sarah, over the 10-day “stimulation-response” mapping period. “We asked the patient to rate her depression severity and related symptoms as we proceeded,” he says. They focused on Sarah’s mood, gauging fluctuations in her depression, as well as in anxiety and her energy level. Sarah’s responses are summarized in this illustration.
Based on research to this point, the team reported: “We found an elaborate repertoire of distinctive emotional responses that were rapid in onset, reproducible, and context- and state-dependent. These results provide proof of concept for personalized, circuit-specific medicine in psychiatry.” By context- and state-dependent, the team meant that stimulation in certain spots in the brain could generate different responses, which depended in a consistent and predictable way on Sarah’s mood state and level of alterness at the time of the stimulation.
To be clear, the researchers had not yet tried to treat Sarah; they had just completed the essential preliminary step of stimulating her brain in many locations and noting impacts on her mood and feelings. This work was followed by analysis, aided by computer-driven machine learning, of the already recorded EEG data. In this analysis they sought to find a place or places in Sarah’s brain where changes in her mood were directly reflected in distinct brainwave patterns. This was the search for an individualized biomarker of her depression, very much like the biomarker in the “DARPA subnets” study which was associated with anxiety in 13 epilepsy patients.
“We asked: ‘What patterns are present in the EEG signals as the patient rated her depression as worse; and how did that signal differ when the patient was feeling better?’” Dr. Krystal explains. The team and patient were very fortunate. “In this, our first patient, we found that when there was elevated high-frequency activity (neural oscillations in the “gamma band,” 30+ cycles per second) in the amygdala, that’s when she got more depressed. And with her, it was a very strong relationship.”
In Sarah’s case, the team was similarly fortunate to have found a single location—an area called the ventral capsule—where electrical stimulation at levels beneath the threshold of Sarah’s ability to detect it, “took her depression away,” as Dr. Krystal describes the effect.
“What was extremely moving for me,” Dr. Krystal remembers, “and I think it had a big impact on everybody on the team and on our patient, was that in our early attempts to explore levels of stimulation in her ventral capsule,”—this was in 2020, when they mapped Sarah’s responses to many stimulation intensities at many sites over 10 days—”she had a profound and immediate response. She said: ‘I haven’t felt this way in 10 years. I feel like my old self again!’”
The mapping procedure helped the team know which stimulation site or sites was most likely to help the patient without adding to her problems. At one point, Dr. Krystal recalls, “We stimulated in one place and she said, ‘That feels really good—but I wouldn’t want to live that way. It feels like I’m artificially happy and almost like it’s a “high,” which is not where I want to be.’ Then we stimulated at another place and she said, “That feels really good and that’s what I want to feel like.”
IMPLANTING THE DBS DEVICE
In the second step of the process, it was time to implant a DBS neurostimulation device in Sarah’s brain, a second invasive surgical procedure that followed the protocols which Dr. Chang had perfected, using the same DBS device to prevent seizures in epilepsy patients. The device is made by a company called NeuroPace, and the procedure for its use in epilepsy was approved by the FDA in 2019.
The team placed one electrode in Sarah’s amygdala that would detect the biomarker signal—of an impending shift toward depressed mood; and one electrode in her ventral capsule, which would deliver precisely the stimulation that prior experiments had shown would make her feel better, in the way that she preferred.
“I wasn’t sure how it was going to work,” Dr. Krystal says.
What ended up happening, he says, “was the thing I had hoped for but wasn’t really sure was possible. The biomarker we selected is close enough to the drivers of this patient’s depression that she no longer gets depressed. She never even senses it. We are picking up something driving her depression and delivering stimulation before she has any sense of being depressed.”
The device appears to be functioning much like a thermostat, which is a closed-loop system that senses the temperature in a room and then activates heating or cooling systems to keep the temperature in a desired range. In this case, the DBS device when triggered by the biomarker sensor appears to do a very good job keeping Sarah’s depression “at bay,” as she has described it. It’s not that she doesn’t have shifts in mood. “I think it’s very important to convey the idea that there is a difference between feelings like sadness, grief, and irritation, when bad things happen in our lives, and depression,” Dr. Krystal observes. “Sarah tells us, ‘I still have normal ups and downs. When something good happens, I feel good. When something bad happens, I feel bad. What’s different now is that in the past there were all these triggers that would make me feel sad, and then another process where I would then get more and more depressed. That is not happening now.’”
Dr. Krystal makes approving reference to an observation once made by Dr. Helen Mayberg, a DBS pioneer. The purpose of the new treatment, Dr. Krystal says, is that “we’re not trying to make people ‘happy’; rather, we are trying to eliminate their depression.’”
This is what Sarah seems to be reporting, after living for a year with the implanted DBS device. She’s not artificially happy in the sense that she reported when, in the preliminary stimulus-response mapping phase of the research, stimulation at a certain site made her feel “high”—not a feeling she desired and which actually made her feel uncomfortable. According to Dr. Krystal, the amount of stimulation delivered in Sarah’s ventral capsule by the DBS device has gradually dropped over time, although not dramatically so. While the antidepressant effect of the treatment was immediate, her symptoms of depression reached the point of remission about 4 months after the device was activated and she remains in remission at the time of this writing.
Dr. Krystal has always been a strong believer in integrating behavioral therapy (e.g., talk therapy) with therapies like brain stimulation. And in Sarah’s case, to date, there is some evidence, he notes, that the “closed loop” delivering stimulation to her brain is generating modifications in the way she responds to typical triggers or challenges from day to day, for instance in the context of relationships with others. Such a changed response pattern could conceivably generate another kind of “closed loop,” in the register of behavior. Events that formerly were triggers to a downward spiral in mood, leading to deep depression, are still capable of bothering Sarah, he says, “but now she responds differently because she’s not depressed. This has the potential to shift the dynamic in her relationships, because now, people in her life are going to respond differently to her.”
WHAT NEEDS TO BE PROVEN
There are many unknowns, beginning with the observation, in Dr. Krystal’s words, that “we are not sure if we will ever see a response like this again, when we try this in other patients.” It is not that he lacks optimism or enthusiasm; he and the team are simply unable to make projections based on results in a single patient. They were fortunate in finding a single biomarker signal in Sarah’s case which reliably predicted a worsening of her mood; and equally lucky to have found a single spot in her amygdala where delivery of stimulation that she can’t even feel is able to either counteract or cancel out that signal, so that Sarah is no longer depressed.
Among the outstanding questions: whether the region of the brain being stimulated adapts over time, decreasing or increasing the therapeutic effect, or if the relationship between brain activity and depression shifts, making the biomarker less effective. So far, this has not happened. Also, based on results in Sarah’s case and in several other patients with whom the team is now working to deliver the same kind of therapy, it is not yet known if in different subjects—even those with similar symptoms—the stimulation and biomarker sites will be the same, or similar, or entirely different. For this reason, the entire concept remains highly experimental.
Dr. Krystal and colleagues already believe, however, that multiple potential stimulation sites will probably be found in most patients. They say this based on having found several sites in Sarah’s brain which to varying degrees and in the context of her different mood-states, had some beneficial impact on some of her symptoms.
Dr. Krystal makes clear that Sarah’s depression tends to feature low energy and anhedonia (the inability to experience pleasure). Other patients say they are, in contrast, often anxious and hyper-aroused when depressed. Even Sarah has anxious moments, and interestingly, a stimulation site was found to diminish that feeling—but she said such stimulation had minimal impact on her depression.
It’s possible, Dr. Krystal says, that as DBS devices become more sophisticated, they might be programmed to deliver pulses to address multiple symptoms at different moments. The condition for such a multi-faceted impact on symptoms would be identifying reliable biomarkers for each symptom and deploying sensors to detect them. No one at this point knows whether, if such treatment one day becomes technically possible, how addressing one symptom might impact a patient’s other symptoms, in real time.
Is “closed-loop neuromodulation” a breakthrough? Dr. Krystal is clear: “We’ve established a number of proof-of-principles. But it is very important to be circumspect, very cautious, because we are talking about one patient. In 5 years, if I can come back to you and report on experiences with more patients—20, or 100—then I will be in a better position to answer. But for patients and their families, I feel it’s important to be clear that we won’t know if this approach is going to be helpful to people generally at least until we do a careful, randomized, double-blinded placebo-controlled trial.”
If more patients can be helped with the approach, then it is certainly possible that as the number grows, certain patterns could emerge, Dr. Krystal says. There are almost certainly different major subtypes of depression—and other psychiatric illnesses—so knowledge of what works in multiple patients with similar subtypes could reveal important things about where and when to apply therapeutic stimulation in the brain with DBS.
It’s conceivable that the emergence of patterns, if they are robust, could eliminate the laborious and invasive “stimulus-response mapping” that Sarah bravely endured prior to implantation of the DBS device. Highly robust patterns could also conceivably inform the targeting of non-invasive stimulation for depression or other conditions.
While all of these possibilities are no more than matters of speculation at this point, the team is encouraged to see that when stimulation is effectively applied, results can be rapid and can be repeated consistently over time. This may also be the case in other patients and could occur in the application of this approach to the treatment of symptoms in other psychiatric illness.
Most immediately, Drs. Krystal, Chang, Scangos and colleagues are eager to discover the extent to which the method works for others who suffer from treatment-resistant depression, a highly complex illness that varies considerably among individuals. Depending on the results, they and other researchers will be equally curious to test the concept to address symptoms in other psychiatric disorders.
Written By Peter Tarr, Ph.D.