Deep Brain Stimulation for Intractable Obsessive-Compulsive Disorder and Treatment-Resistant Depression

In Lisbon in 1935, neurologist Egas Moniz developed and, with neurosurgeon Almeida Lima, participated in the first leucotomy on a human, after learning of similar work done with chimpanzees by John F. Fulton that same year (1). This initial procedure involved creating two burr holes in the upper frontal region of the skull and injecting alcohol into the white matter of the frontal lobe to destroy the connections to the rest of the brain (2). Moniz reported on his work at the Paris Neurology Society in 1936 (3), where he indicated improvement in 70% of patients (14 of 20) and remission in 35% of patients (7 of 20) and shared his observations that this procedure was more successful among those with mood disorders than among those with schizophrenia. This work ultimately won Moniz a Nobel prize in 1949. Because of limited existing treatments for psychiatric disorders and Moniz’s notoriety after developing cerebral angiography, this procedure was quickly adopted and refined by practitioners in Brazil, Italy, and the United States (4).

Walter Freeman and James Watts brought the leucotomy to the United States. Although initially using the same technique as Moniz, they ultimately created a method that would more completely sever the connection between the frontal lobe and deep brain structures, which they named a “frontal lobotomy” (2). They further developed this procedure and, in the process, made important observations, such as that ablating fibers of the cingulate gyrus resulted in a loss of anxiety for patients (5). Freeman, however, began looking for ways to simplify this technique to meet the needs of growing referrals amid overcrowded U.S. asylums and reports of detrimental conditions in psychiatric hospitals (6). This search led to the adoption of the supraorbital lobotomy, which would be overzealously used to treat patients without trained surgeons, proper sanitation methods, and outside the operating room—such practices were unacceptable to Watts, who stopped collaborating with Freeman.

Unpredictable changes to a patient’s personality following leucotomy or lobotomy, and other complications resulting from these procedures, led neurosurgeons to investigate more precise surgical methods. Through the pursuit of improved precision, neurostimulation research, and animal research, the limbic loops began to be better understood, which further encouraged neurosurgeons and psychiatrists to investigate new ways to deal with psychiatric conditions with fewer undesired outcomes. One especially important procedure to the discussion of deep brain stimulation (DBS) is the anterior cingulotomy.

In 1948, Cairns, Scoville, and Le Beau developed the cingulectomy (the procedure’s original name) with the hope of treating agitation and violence (7–9). This ablative procedure targeted the anterior cingulate gyrus. As early as 1937, researchers including James W. Papez described the role that the anterior cingulate gyrus plays in the expression of emotion (10), with several physiologists finding that area had some regulatory control over the cerebral cortex (11–13). Scoville published the first clinical series of patients receiving this procedure and emphasized the procedure’s surgical difficulty (9).

In a similar pursuit, Spiegel and Wycis (14) chose to target the mediodorsal area of the thalamus to destroy a portion of the same thalamo-frontal circuit that leucotomies had been designed to disconnect. To do this, they used a stereotactic frame while defining internal cerebral landmarks and using a cartesian coordinate system to map this deep brain target. The apparatus they used was originally designed by Victor A. H. Horsley and Robert H. Clarke decades earlier (15); the real innovation was the ability to use radiography, which had improved enough for clinical use during surgeries, and the use of air ventriculography, in which air replaced ventricular fluid to allow for clear visualization under X-ray (16). In 1947, they bolstered the field of stereotactic neurosurgery by presenting their clinical applications of stereotactic frames and the use of internal landmarks for measurement and placement of these devices (16). The development of stereotaxis would not just redefine psychiatric neurosurgery and procedures such as the anterior cingulotomy but would also come to be vital for modern neurosurgery.

Two early proponents of stereotactic neurosurgery were Talairach and Leksell (17), who used stereotactic techniques in targeting the anterior limb of the internal capsule (ALIC) in 1949. This target was chosen following a growing understanding that the prefrontal cortex was connected to the thalamus and basal ganglia through the fibers of the ALIC. Not only did Talairach and Leksell pioneer one of the only surviving psychosurgical procedures, the anterior capsulotomy, but both actively improved stereotaxis. Leksell developed an improved stereotactic frame (18), and Talairach developed a system of air contrast and X-ray imaging to identify deep brain structures and targets for neurosurgery (19).

The development of stereotactic neurosurgery opened the door for investigations of deep brain structures, which could now be safely navigated. It has been reported that the mortality rate associated with neurosurgery decreased from 15% to 1% through the adoption of stereotactic neurosurgery (20). Clinicians performing capsulotomy, as well as the previously mentioned cingulotomy, were now taking advantage of stereotaxis, with Foltz and White (21) performing stereotactic cingulotomy for the treatment of intractable pain in 1962. The anterior cingulotomy and anterior capsulotomy remain useful in the treatment of intractable obsessive-compulsive disorder (OCD) and treatment-resistant depression and have been instrumental in our current understanding of brain targets for psychiatric DBS.

Not only did stereotaxis allow for the success of these aforementioned procedures, but it also allowed for targeted electrical stimulation in the brain. Electrical stimulation was used to test different regions during surgery and to ablate if the desired response was identified. This process could now be used in subcortical regions safely and with precision by using stereotactic methods. Spiegel and Wycis, too, began performing electrical stimulation and recordings of relevant anatomical regions to confirm proper placement before using electrothermal coagulation to create lesions at the desired targets (20, 22, 23).

Electrical stimulation of deep brain regions was becoming a viable tool as it was incorporated into these stereotactic procedures, and its benefits would slowly become apparent. As surgeons continued to search for and refine methods to ablate areas that could treat psychiatric conditions while minimizing poor outcomes, further knowledge was gained that would be essential for the development of DBS. It was learned that low-frequency stimulation activated areas and led to worsening symptoms, but higher frequency stimulation had the opposite response and reduced symptom presentation (24, 25). Additionally, Nashold and Slaughter (26) and Sem-Jacobsen (27) were the first to leave electrodes implanted for several weeks in order to ablate incrementally.

The first electrodes for stimulation and treatment (and not for targeting prior to ablative procedures) were implanted by J. Lawrence Pool in 1948 to treat an elderly woman with anorexia and depression (28). Jose Delgado, a professor of neurophysiology at Yale University, reportedly continued to explore electrical stimulation to assess treatment options for patients (29). This exploration included stimulation of the dorsolateral nucleus of the thalamus and long-term implantation of electrodes. During this work, Delgado and colleagues noticed that patients seemed to improve during stimulation that lasted across several days, and they realized that electrical stimulation could have results similar to psychiatric neurosurgery while being a more conservative treatment approach. Another advocate of electrostimulation, Robert Heath, a psychiatrist at Tulane University, implanted multiple electrodes in different subcortical nuclei in the 1950s. During this time, he studied the treatment of pain by stimulation of the septal area as well as the general use of targeted electrical stimulation in the brain (30–32). Although his work represents some of the earliest use of DBS and contributed to some knowledge, including the use of chronic stimulation in psychiatric disorders, it was significantly marred by the unethical practices he used in his studies.

There were opponents to psychosurgery nearly from the moment it was developed, but by the 1950s, the ethics of such practices were starting to be questioned by the general public. The erratic and excessive use of lobotomy, as encouraged by Freeman, and improved understanding of its complications, led prominent psychiatrists and neurosurgeons to speak out against these procedures. These disputes were at times made public as well, such as in the publication of a Newsweek article explaining a prominent psychiatrist’s disagreement with Freeman at an American Psychiatric Society symposium (33). Popular culture similarly began to take notice of such questionable practices, as evident by and possibly partly because of popular novels, such as One Flew Over the Cuckoo's Nest (1962), or several of Tennessee William’s plays, which drew inspiration from his real-life experiences interacting with his sister who was diagnosed as having schizophrenia and subsequently underwent a lobotomy (2). Legislation was passed to prevent these operations in certain states, and special councils were created to examine these surgical practices (34).

Near the same time, the development of neuroleptics by Jean Delay and Pierre Deniker in 1952 (35) showed a promising approach to treating patients experiencing symptoms of psychosis with measures that were easily reversible and did not require potentially risky surgical procedures. Chlorpromazine gained approval by the Food and Drug Administration (FDA) just 2 years later, and its success and popularity led to research into other medications, such as haloperidol in 1959 (36) and levodopa in 1961 (37). A strong and growing public negative sentiment toward lobotomy and, by association, other ablative and stereotactic procedures, as well as the growing knowledge and use of pharmacotherapies for psychiatric conditions, drew the medical community away from these procedures (38).

This initial period of growth for psychiatric neurosurgery waned, but the insights into stereotaxis and neurophysiology were paramount during the development of DBS. Similarly, surgical procedures, such as anterior cingulotomy and anterior capsulotomy, continue to be used to help treat patients diagnosed as having intractable OCD and/or treatment-resistant depression and have presented potential targets for DBS. In addition to the growing knowledge of limbic circuits that was developed during this time, the use of electrical stimulation in the brain furthered understanding of the impact that electrical frequency has on neurological function at a stimulation site. These insights were the early precursors of developments that would be born out of necessity and alongside other medical innovations.

Around this same time, uses of electrostimulation were being studied in other medical disciplines. In 1957, Earl Bakken engineered the first battery-operated wearable pacemaker and went on to cofound Medtronic (40). This development of the pacemaker became massively successful which not only added legitimacy to the implantation and use of electrostimulation for the treatment of medical conditions, but the other medical uses for a refined and tested electrostimulation device also became readily apparent. Medtronic began to work on other applications for the technology, as did other companies that began creating electrostimulation devices. Clinical research soon took advantage of this technology. The first implementation of pacemaker technology for neurostimulation was in 1967 to treat pain by the placement of electrodes in the spinal cord (41). Thus, Medtronic made its first neurostimulator in 1968, intended for use in spinal cord procedures (42). The potential for this technology was again explored for the treatment of pain in brain regions in 1973 (43). Similarly, neurosurgeons began to use these neurostimulators as an alternative to ablative procedures. The spinal cord and cerebral cortex became interesting targets to treat neurological conditions. Still, only the institutions that had continued stereotactic neurosurgery in the decades prior were able to use this experience to treat and study deep brain structures (38).

Exploration into the use of neurostimulation was halted again in 1976. Although progress toward the use of neurostimulation to treat neuropsychiatric conditions was slowed, the FDA’s new authority over medical devices helped ensure that safe, effective practices would continue in place of the ethically ambiguous work that had characterized some of the neurostimulations in the past. Subjective findings from neurosurgeons and psychiatrists regarding improvements in individual patients no longer would be sufficient for approval of a medical device; objective measures with controlled trials would be needed. This change posed challenges in treating psychiatric conditions, such as OCD and depression, which can have complex and varied presentations. This requirement continues to affect the success of treating these heterogeneous diseases.

One condition that showed progress in treatment and stood out because of its ability to provide quantifiable measures was Parkinson’s disease. This motor disorder had a more direct and consistent symptom presentation and, as such, could be better quantified with a standardized metric. In 1987, the International Parkinson and Movement Disorder Society was established, and the Unified Parkinson’s Disease Rating Scale (UPDRS) was created, replacing all other Parkinson’s disease scales and helping to standardize a quantifiable metric for the disease (44) That same year, Alim Louis Benabid, a neurosurgeon who had continued to practice stereotactic neurosurgery and neurostimulation in Grenoble, France, published findings (45) on the use of DBS for Parkinson’s disease, targeting the thalamic nucleus ventralis intermedius. At the time, this was a common target for thalamotomy used to treat various kinds of tremors. Benabid and his colleagues performed radiofrequency thalamotomy on the side of the brain corresponding to the most prominent tremor, while implanting a deep brain electrode in the same area but on the less affected side. It was found that, similar to past work that had identified that high-frequency stimulation decreased activity with results comparable to ablation, DBS was able to decrease tremors, although not reaching the same effectiveness of traditional thalamotomy (45). Benabid et al. also began to note the importance of frequency as they believed the optimal stimulation frequency would be 200 Hz, although the capabilities of DBS limited them to 130 Hz. Benabid et al. (46) continued to study DBS for Parkinson’s disease and eventually changed the target to the subthalamic nucleus. Although this was not the first use of DBS-related technology and the findings were consistent with previous work, it came at a time when a combination of technological advances, changes in regulation, and appreciation for the need for DBS made this work more tractable.

DBS for Parkinson’s disease targeting the ventral intermediate nucleus of the thalamus was ultimately approved by the FDA in 1997, following the work of Benabid et al. and the development of the UPDRS, which was the primary measure used in clinical trials for DBS for Parkinson’s disease. The initial approval of DBS for Parkinson’s disease in 1997 allowed for unilateral DBS in patients with severe tremors, before expanding the use to treatment for general Parkinson’s disease in 2002 (47).

When Ben Greenberg and his team at Brown University pursued DBS for intractable OCD following the promising results of Nuttin et al., they used the same target as in their previous work with anterior capsulotomy (50), which had indicated improvement in patients’ condition when targeting the ventral portion of the ALIC where it meets the ventral striatum (VC-VS) (51). The rationale for targeting this more posterior region was that there is a denser presence of thalamocortical fibers there. In 2006, Greenberg et al. (51) reported on 10 patients who had received DBS for OCD and had been followed for 3 years. Of the original 10, eight completed 36 months of data collection. The average YBOCS score decreased from 34.6 to 22.3 at 36 months, with four of eight patients experiencing a decrease of ≥35% (remission) and two additional patients with decreases between 25% and 35% (responders). Notably, there was also a decrease in anxiety and depression symptoms, as measured by the Hamilton Anxiety Rating Scale and the Hamilton Depression Rating Scale (HDRS), and an improvement in global functioning, as measured by the Global Assessment of Functioning (51).

Greenberg et al.’s long-term follow-up, published in 2010 (52), examined the findings of four collaborative groups that used DBS, with similar targets and inclusion criteria, to treat intractable OCD. This included the work of Nuttin at Leuven, Belgium; Greenberg at Brown University; and collaborators at the Cleveland Clinic and the University of Florida. Through an 8-year collaboration, they reported on the treatment of 26 patients and found a gradual improvement in OCD symptoms as measured by YBOCS; at the last follow up, 16 of 26 patients were in remission, and three more had responded to the treatment. The gradual improvement in YBOCS over 3 years, with many patients seeing the most improvement after 3 months, was an important distinction from the result found for DBS for Parkinson’s disease, which shows improvements more immediately. Throughout the 26 reported patients, the original ALIC target migrated as more patients were implanted more posteriorly as clinicians noticed improved responses in intraoperative test stimulation. This finding was validated through neuroimaging and animal model results, which indicated more efficient stimulation with a smaller required electric field when the treatment migrated to a more compact connection of the cortico-striato-thalamo-cortical loop (52). In this landmark article (52), patients were divided into three groups, depending on the timing of their implantation during the study and thus how much the target had migrated posteriorly; the latter two groups had greater decreases in YBOCS, with approximately 75% reaching remission. Furthermore, at the 36-month follow-up, the average HDRS score decreased by 43.2%, with 14 patients reaching remission (HDRS score<7).

DBS for intractable OCD targeting the VC-VS was ultimately approved by the FDA in 2009 through a Humanitarian Device Exemption that was based on results from the open-label trial. This is a path for approval for treatment of conditions that affect less than 200,000 U.S. citizens per year, thus providing a challenge to gathering enough clinical evidence to meet the FDA’s typical standard of reasonable assurance of safety and effectiveness, and therefore not requiring RCT results.

In 2002, Luc Mallet and his colleagues reported surprising and pronounced benefits for co-occurring OCD and Parkinsonism among two patients who had received DBS in the subthalamic nucleus (STN) at a high-frequency setting. These two individuals with histories of OCD experienced 54% and 68% improvement in YBOCS scores after receiving stimulation between the anteromedial portion of the subthalamic nucleus and the zona incerta (53). An RCT was then conducted, in which 16 patients received continuous stimulation at the STN for 3 months and sham stimulation for an additional 3 months, with half of the patients receiving active stimulation first and the others experiencing sham stimulation first. Average YBOCS scores were 19 after active stimulation and 28 after sham stimulation. Among the first group receiving active stimulation, 75% (6 of 8) met response criteria on the YBOCS, whereas only 38% (3 of 8) in the sham condition met criteria for response (54). These limited results hold promise for the targeting of the STN to treat intractable OCD, but more research needs to be conducted to receive FDA approval for STN DBS.

Prior to the use of DBS, anterior capsulotomy had been used to treat major depressive disorder as well as OCD (55). As more patients were treated with DBS at the VC-VS target for OCD symptoms, investigators noticed that mood symptoms improved before OCD symptoms (52, 56). This observation, the prior use of anterior capsulotomy for treatment-resistant depression, and neuroimaging studies that indicated the involvement of the VC-VS region in major depressive disorder led multiple groups of researchers to examine open-label use of DBS targeted to the VC-VS for treatment-resistant depression (57–59). These studies demonstrated a response rate of approximately 50%.

Promising findings from these open-label trials inspired RCTs of DBS targeted on the VC-VS for treatment-resistant depression. The first of such was a U.S.-based Medtronic-sponsored trial examining the use of the Reclaim DBS system. This RCT involved bilateral implantation of DBS leads at the VC-VS and a relatively brief initial optimization period for all patients before a 16-week blinded portion, in which half the patients had their devices deactivated as a sham condition and the other half received active stimulation (60). Thirty patients were initially screened and found to meet the study’s inclusion criteria. After the 1-month optimization period, 15 patients received the active treatment and 14 received the sham treatment. Ultimately, the study failed to demonstrate efficacy for DBS targeted to the VC-VS for treatment-resistant depression, as determined through the primary measures of the Montgomery-Asberg Depression Rating Scale (MADRS). Of the patients receiving active stimulation, 20% (3 of 15) responded (with a 50% decrease or greater on the MADRS), whereas a comparable 14.3% receiving sham stimulation (2 of 14) responded. The active group had a mean decrease on the MADRS of 8.0, whereas the sham group had a mean decrease of 9.1. During the open-label continuation phase, the response rate improved slightly, to 26.7% at 18 months and 23.3% at 24 months, but remained much lower than in the previous open-label trials conducted by the same research groups (58).

During this U.S.-based RCT, a similar RCT was being conducted in Europe with a nearly identical target but an altered protocol design. In this European trial, a DBS device was implanted in the ventral anterior limb of the internal capsule of 25 patients who had treatment-resistant depression (61). In contrast to the U.S.-based trial, this RCT began with an open-label phase that lasted 52 weeks before random assignment and double-blind discontinuation of stimulation occurred. Possibly as a result, this trial had a higher percentage of responders in the open-label portion of the study, with 40% reaching a 50% or greater decrease on the HDRS. During this open-label period, mean MADRS scores for responders dropped from 31.2 to 11.8 at the time point when response versus nonresponse was determined, whereas mean MADRS scores for nonresponders stayed relatively the same, with a decrease of only 35.9 to 32.3. After a year, 16 patients (nine responders and seven nonresponders) began the double-blind discontinuation period, in which nine of the patients first received the active stimulation and then the sham condition, whereas seven received the sham condition followed by a return of active stimulation. The DBS discontinuation worsened depression among responders, with a mean HDRS score of 23.1 compared with a mean of 9.4 with the active stimulation. Further supporting the effects of this DBS treatment, the mean HDRS for nonresponders was 23.0 during discontinuation and 19.0 during active stimulation. This RCT remains the only study of DBS for treatment-resistant depression to achieve its criterion for success based on its primary outcome measures.

In 2005, Helen Mayberg and colleagues published the first report (62) of DBS targeting the sgACC white matter adjacent to Brodmann’s area 25 for treatment-resistant depression. This targeting was based on neuroimaging performed by Mayberg and colleagues, which indicated increased activity of this area on positron emission tomography and functional MRI when patients were experiencing sadness, both naturally for individuals with depression and induced for healthy volunteers (63). In this 2005 article (62), 66% of patients (4 of 6) met the criteria for response (a ≥50% decrease in HDRS).

Following these findings, several other open-label trials were conducted examining the application of DBS at this target, also with strong response rates (64–67). After repeated positive outcomes and promising results of these open-label trials, a large multisite RCT of sgACC DBS for treatment-resistant depression was conducted to demonstrate efficacy for FDA approval of this DBS treatment (68). This study involved 90 patients who were randomly assigned to receive active stimulation or sham stimulation. Of the patients in the active stimulation group, 20% met the criteria for response (≥40% reduction in MADRS and no worsening of the Global Assessment of Functioning score compared with baseline), whereas 17% of patients in the sham stimulation group also met the criteria for response. In examining these results with response criteria of 50% or greater reduction in MADRS, as several of the other RCTs had done, the percentages shifted to 12% for the active stimulation group and 10% for the sham stimulation group. However, there was an increase in response rates during an open-label phase from 6 months to the last-reported follow-up at 30 months. At the 30-month follow-up, 36 patients had responded according to the study criteria, and 16 had achieved remission.

DBS of the VC-VS and STN have both shown promising results for the treatment of intractable OCD, but comparisons of the two have required analyses based on separate studies. However, Tyagi et al. (69) conducted a study with DBS implantation in both targets for six participants to more effectively compare the efficacy associated with the two targets and to examine whether any type of synergistic effect might be present. Ultimately, Tyagi et al. found a mean improvement in YBOCS of 53% when targeting the VC-VS, a 45% improvement when targeting the STN, and a 60% improvement with combined targets. Targeting of the STN alone resulted in 50% of patients responding, targeting of the VC-VS alone resulted in 83% of patients responding, and subsequent combination of the two resulted in no new responders.

Unlike DBS for OCD, DBS has not been approved for use by the FDA for treatment-resistant depression. This lack of approval has resulted in the studying of several additional targets that have yet to undergo considerable RCTs. For example, the nucleus accumbens is the location in which the tip of the electrode is placed for VC-VS DBS and thus has been investigated in open-label trials to examine what impact this nearby target may have on outcomes for patients with treatment-resistant depression (57, 70). Similarly, targeting of the inferior thalamic peduncle has also occurred for one patient with major depressive disorder, with a dramatic reduction in HDRS score and lasting results (71).

In 2013, Schlaepfer and colleagues (72) reported on their efforts in targeting the superolateral medial forebrain bundle in DBS treatment for treatment-resistant depression. Patient response was rapid compared with other DBS targets, with five of seven patients qualifying as responders after 1 week of DBS stimulation and three patients achieving remission. In 2019, Coenen et al. (73), as part of an effort to gain insight for a larger trial, reported an RCT evaluating the safety and efficacy of DBS of the medial forebrain bundle for 16 patients. Ultimately, the sham versus active stimulation distinction was reportedly blurred by microlesion effects and/or placebo effects. However, after 12 months, 100% of patients met response criteria, with many seeing such improvements within a week. Additionally, half of the patients had remission of treatment-resistant depression after 1 year of stimulation.

Although current targeting primarily relies on using the same stereotactic mapping targets in terms of x, y, z coordinates across individuals, there is now investigation into more patient-specific targeting, which takes individual anatomic variability into account. The use of preoperative tractography could potentially improve individual outcomes. Various studies (74–76) have analyzed how anatomical variations can result in sizable distances between the intended target and the actual target hit. Initial studies of sgACC targeting with preoperative tractography have shown potentially higher responder rates than with standard x, y, z targeting (74).

A central component of DBS for intractable OCD and treatment-resistant depression is the programming of stimulation parameters, which can be refined by adjusting different components of stimulation and waiting to see if patients improve. This is a methodical process that can take weeks to months to determine the optimal stimulation parameters for each individual. Although some consistency has been identified, such as common frequency, pulse width, and amplitude ranges, it is difficult to ascertain how the brain may respond to stimulation on an individual level. One approach that has the potential to improve outcomes for these conditions that involve complex circuits and heterogeneity is the use of a closed-loop system that can both record neural activity and stimulate. This application requires the identification of relatively precise biomarkers, which is challenging. However, there is potential for systems that can adjust stimulation parameters based on real-time electrophysiological recording to individualize treatment. Early work using beta frequency signals in Parkinson’s disease has been promising, and this approach is now being explored in OCD and treatment-resistant depression. For an in-depth look at this technology, please refer to Widge et al. (77).