Invasive Brain Stimulation in the Treatment of Psychiatric Illness

Abstract

Background

Drugs, psychotherapy, and other treatment modalities are effective for many patients with mental illness. Nonetheless, many patients do not achieve a total remission with the currently available interventions, and the recurrence rates are high. As part of the ongoing search for further treatment options for refractory disorders, there is renewed interest in focal neuromodulatory techniques, including invasive ones, and deep brain stimulation (DBS) in particular.

Methods

In this review article, a group consisting of neurosurgeons, psychiatrists, and one practicing ethicist/neurologist summarizes the main aspects of the use of DBS to treat mental illness and offers recommendations on its indications and practical implementation.

Results

The efficacy of DBS against mental illness has not been confirmed in the randomized, controlled trials (RCTs) that have been published to date. This may be because the follow-up times were too short. In contrast to the negative RCTs, case series have indeed shown a positive effect of DBS on severe depression, but this effect can only be seen several months after the operation.

Conclusion

DBS may be a therapeutic option for selected patients with otherwise intractable mental illness. Patients should only be treated in the setting of clinical trials. RCTs with longer follow-up times must be conducted in order to substantiate, if possible, the promising evidence that has been found in case series.

Numerous definitions of treatment-resistant depression (TRD) have been proposed over the last 30 years, many of which are similar or overlapping. Although there are still no established, formal diagnostic criteria, general agreement has developed over time concerning the main features of TRD:

• the presence of unipolar depression

• the failure of treatment with antidepressants, defined either in categorical terms (e.g., inefficacy of at least two antidepressants from different pharmacological classes) or as an increasing resistance to previously successful pharmacotherapy

• resistance to established forms of disorder-specific psychotherapy

• the absence of somatic illness or psychosocial disturbances (1).

Thase and Rush (2) count the failure of electroconvulsive therapy (ECT) as a further criterion of TRD. 20–30% of patients with depression benefit only to a small extent, or not at all, from the treatments available to date.

The criteria for treatment resistance are even less precisely defined for obsessive-compulsive disorder (OCD) than for depression. Both of these diseases are characterized by an incomplete understanding of the underlying biological disturbances, along with a wide variety of treatments currently in use. Trial-and-error remains the guiding principle of pharmacotherapy for both.

Focal neuromodulation is an alternative, non-pharmacological, but still biologically-based form of treatment for hitherto intractable depression and OCD. The term “neuromodulation” refers to a direct, reversible alteration of the function of a neural structure in order to change activity in a neural network of brain regions participating in the processing of affective stimuli. The types of invasive neuromodulation now being used to treat mental illness were developed on the basis of knowledge derived, both from functional imaging of network dysfunction in depression and OCD and from earlier clinical experience with lesioning techniques.

Neuromodulatory approaches to the treatment of intractable mental illness are based on a neural network theory in which it is postulated that a certain set of structurally and functionally connected regions of the brain work together to maintain normal regulation of mood (3, 4). Some types of dysfunction in affective disorders can be described as resulting from faulty communication between nodes of this network (5). Such communication takes place both by chemical means (neurotransmitters) and via electrical signaling; there is thus, theoretically, an opportunity for either pharmacological or electrostimulatory intervention, the latter being useful particularly in cases that are resistant to pharmacotherapy.

The efficacy of deep brain stimulation against treatment-resistant depression

Depression, one of the more serious diseases affecting mankind, affects approximately 20% of all individuals at least once over the course of a lifetime (6). According to the World Health Organization (WHO), more than 300 million people suffer from depression at any given moment (point prevalence) (7). Although many patients can benefit from effective treatment, 20–30% suffer from intractable depression. For some of them, deep brain stimulation (DBS) is a possible treatment option.

Among the recent publications on DBS for depression, the earliest appeared at the beginning of the present century. Mayberg and Lozano (8) were the first to report on stimulation of the subgenual cingulate (SCG, cg25) in a series of six patients. Four of them displayed a significant response to treatment at six months, meaning that their symptoms, as rated on a scale such as the Hamilton Depression Rating Scale (HDRS), improved by at least 50%. Another target region for DBS in depression is the nucleus accumbens septi (NAC); successful stimulation of this target was also reported early on, in small case series (the earliest one consisted of 3 patients) (9). Further targets studied included the inferior thalamic peduncle (ITP), the bed nucleus of the stria terminalis (BNST), and the superolateral branch of the medial forebrain bundle (slMFB) (10). Stimulation of the last-named target (in small case series, as with the other targets) yielded a significant antidepressant effect in up to 85% of the patients so treated, in the experience of multiple centers (11– 14); the effect persisted for up to at least 50 months (12). The brain regions most commonly used as targets in uncontrolled trials are currently the subgenual cingulum (SCG), the ventral capsule/ventral striatum (VC/VS), and the nucleus accumbens septi (NAC). Significant treatment responses have been reported in up to 60% of the severely affected patients in these uncontrolled studies and small case series (10).

An overview of the randomized, controlled trials that have been performed to date (13, 15– 20) can be found in the Table, as well as in three recent reviews on the efficacy of deep brain stimulation for depression (10, 21, 22). The results of the uncontrolled case series (etable) could not be replicated in two large-scale, double-blinded, randomized, multicenter trials (18, 19). The authors of the present review would like to point out that an unexpectedly large number of patients displayed improvement of their depressive symptoms only after the end of the blinded phase (23). We suspect the main problem was that the period of observation was too short: four months are not a long time in the life of a patient who has been suffering for decades from a single, unending depressive episode. The two RCTs provide too little information about the potentially beneficial effect of DBS in these target regions for patients with intractable depression, an effect which may yet be real and therapeutically highly relevant. They can perhaps be best understood as a premature attempt to commercialize DBS for depression before there was any real understanding of the specific properties of the different target areas – in contrast to, for example, the successful use of DBS in the treatment of Parkinson’s disease (24).

eTable. An overview of all studies (other than single case reports) on the efficacy of deep brain stimulation against intractable depression.

Author (reference no.)	N	Study duration	Definition of response	Definitionof remission	Design
		Longest follow-up	Response rate (%)	remission rate (%)
Target: subgenual cingulate gyrus (cg 25)
Eitan, 2018 (16)	9	13 months	40% reduction MADRS	not reported	randomized, controlled trial
		13 months	44.4	not reported
Merkl, 2018 (15)	8	24 months	50% reduction HDRS-24	HDRS-24 ≤ 10	randomized, controlled trial
		28 months up to 4 years	6 months: 34 12 months: 25 24 months: 37 28 months: 51 last follow-up: 51	last follow-up: 33.3
Riva-Posse, 2018 (e1)	11	12 months	50% reduction HDRS 17	HDRS-17 ≤ 7	open trial without randomization
		12 months	12 months: 81.8	12 months: 54
Holtzheimer, 2017 (19)	90	12 months	40% reduction MADRS	MADRS ≤ 10	randomized, controlled trial
		24 months	6 months: 20 (active group) 17 (control group) 12 months: 30 (active group) 27 (control group) 24 months: 49 (all patients)	6 months: 5 (active group) 7 (control group) 12 months: 18 (active group) 7 (control group) 24 months: 26 (all patients)
Accolla, 2016 (e2)	5	not reported	50% reduction HDRS-24 and BDI	not reported	open trial without randomization
		6 months	6 months: 20	not reported
Perez-Caballero, 2014 (e3)	8	not reported	50% reduction HDRS-17	HDRS-17 ≤ 8	open trial without randomization
		4 months	4 months: 38	4 months: 38
Ramasubbu, 2013 (e4)	4	9 months	50% reduction HDRS-17	not reported	open trial without randomization, but with blinded optimization phase
		9 months	9 months: 50	not reported
Merkl, 2013 (e5)	6	6 months	50% reduction HDRS-24	HDRS-24 < 10	randomized and controlled for 24 hours; thereafter an open trial without randomization
		6–9 months	6 months: 33.3	6 months: 33.3
Lozano, 2012 (e6)	21	12 months	50% reduction HDRS-17	NA	open trial without randomization
		12 months	1 month: 57 6 months: 48 12 months: 29	NA
Puigdemont, 2012 (e7)	8	12 months	50% reduction HDRS-17	HDRS-17 ≤ 7	open trial without randomization
		12 months	1 week: 87.5 1 month: 37.5 6 months: 87.5 1 year: 62.5	1 week: 50 1 month: 37.5 6 months: 37.5 1 year: 50
Holtzheimer, 2012 (e8)	17	6 months	50% reduction HDRS-17	HDRS-17 ≤ 7	single-blinded open trial without randomization
		24 months	6 months: 41 12 months: 26 24 months: 92	6 months: 18 12 months: 36 24 months: 58
Kennedy, 2011 (e9)	20	3 years	50% reduction HDRS –17	HDRS-17 ≤ 7	open trial without randomization
		6 years	1 year: 62.5 2: years: 46.2 3 years: 75 last follow-up: 64.3	1 year: 18.8 2: years: 15.5 3 years: 50 last follow-up: 42.9
Lozano, 2008 (e10)	20	12 months	50% reduction HDRS-17	HDRS-17 ≤ 7	open trial without randomization
		12 months	1 month: 35 6 months: 60 12 months: 55	1 month: 10 6 months: 35 12 months: 33
Mayberg, 2005 (8)	6	6 months	50% reduction HDRS-17	HDRS-17 ≤ 8	open trial without randomization
		6 months	1 month: 33.3 2 months: 83.3 6 months: 66	1 month: 0  2 months: 0 6 months: 33.3
Target: ventral capsule/ventral striatum
Dougherty, 2015 (18)	30	4 months	50% reduction MADRS	MADRS ≤ 10	randomized, controlled trial
		2 years	4 months:20 (active group) 14.3 (control group) 1 year: 20 (all patients) 2 years: 23.3 (all patients)	1 year:13 (all patients) 2 years: 20 (all patients)
Malone, 2010 (e11)	17	not reported	50% reduction MADRS	MADRS ≤ 11	open trial without randomization
		14–67 months	3 months: 53 last follow-up: 71	last follow-up: 35
Malone, 2009 (e12)	15	not reported	50% reduction HDRS-24	HDRS-24 ≤ 10	open trial without randomization
		6–51 months	3 months: 46.7 6 months: 40 last follow-up: 53.3	3 months: 20 6 months: 20 last follow-up: 40
Target: anterior limb of internal capsule (ALIC)
Raymaekers, 2017 (20)	7	not reported	50% reduction HDRS-17	HDRS-17 ≤ 7	randomized, controlled trial
		36 to 97 months	last follow-up: 71	last follow-up: 28.5
Bergfeld, 2016 (17)	25	max. 64 weeks	50% reduction HDRS-17	HDRS-17 ≤ 7	randomized, controlled trial
		55 weeks	1 year: 40	1 year: 20
Target: nucleus accumbens
Millet, 2014 (e13)	4	6 months	50% reduction HDRS-17	HDRS-17 ≤ 7	open trial without randomization
		15 months	6 months: 0 15 months: 50	6 months: 0 15 months: 25
Bewernick, 2012 (e14)	11	12 months	50% reduction HDRS-28	HDRS-28 < 10	open trial without randomization
		24 to 36 months	12 months: 45.5 last follow-up: 45.5	12 months: 9 last follow-up: 9
Bewernick, 2010 (e15)	10	12 months	50% reduction HDRS-28	HDRS-28 < 10	open trial without randomization
		12 months	12 months: 50	12 months: 30
Target: superolateral branch of the medial forebrain bundle (slMFB)
Coenen, 2019 (13)	16	12 months	50 % reduction MADRS	MADRS < 10	randomized, controlled trial
		12 months	12 months:100 (all patients)	12 months: 50 (all patients)
Bewernick, 2017 (12)	8	12 months	50% reduction MADRS	MADRS < 10	open trial without randomization
		4 years	12 months: 75 last follow-up: 87 (area under the curve [AUC])	12 months: 50
Fenoy, 2016 (e16)	4	12 months	50% reduction MADRS	not reported	open trial without randomization
		12 months	1 week: 75 6 months: 50	not reported
Schlaepfer, 2013 (11)	7	12 weeks	50% reduction MADRS	MADRS < 10	open trial without randomization
		12 to 33 weeks	1 week: 57 12 weeks: 86	1 week: 43 12 weeks: 57

This table is based on the reviews of Dandekar et al. (22), Kisely et al. (21), and Drobisz and Damborská (10), with more recent studies added.

HDRS, Hamilton Depression Rating Scale; MADRS, Montgomery–Åsberg Depression Rating Scale

In summary, the efficacy of DBS against depression has not yet been convincingly demonstrated. This is due, in part, to the fact that a variety of different targets and stimulation frequencies were used, with varying postoperative periods of observation, in small, uncontrolled studies.

The efficacy of deep brain stimulation in treatment-resistant obsessive-compulsive disorder

The long-term effects of stimulation of the ventral capsule and ventral striatum for the treatment of OCD have been described in several studies, among them two double-blinded RCTs. Denys et al. reported the results in 16 patients, all of whom received stimulation without blinding in the first eight months, followed by a crossover phase lasting 2 × 2 weeks and involving periods of stimulation and no stimulation (“stim on/off”) (25). In the open phase, symptoms were reduced by 46%, as measured on the Y-BOCS (Yale-Brown Obsessive Compulsive Scale). In the crossover phase, scores were 25% lower in periods with the stimulation on than in periods with the stimulation off; this difference was statistically significant (p = 0.004). In the study of Luyten et al., 24 patients were enrolled from 1998 to 2010 (26). An unblinded phase lasting several months, in which the stimulation parameters were optimized, was followed by a double-blinded, randomized crossover phase with and without stimulation, lasting 2 × 3 months. Seventeen patients completed the crossover phase. Y-BOCS scores (the primary endpoint) were significantly better with the stimulation on than with the stimulation off (median difference, 37%). The improvement of Y-BOCS compared to each patient’s baseline was significant as well (median improvement, 42%). After four years of stimulation, the mean improvement compared to baseline was 66%.

Intra- and postoperative complications of deep brain stimulation in general

In a study of DBS in 25 patients, one patient reportedly became severely nauseated intraoperatively, four attempted suicide, and two others had suicidal ideation (17). Other documented side effects included agitation (7 patients), disinhibition (e.g., loquacity) (6 patients), and perceived vibrations of the neurostimulator (3 patients). In another study involving seven patients, eleven severe complications were documented, four of which were related to the stimulating electrode; two patients committed suicide (20). It must be borne in mind that patients undergoing DBS for psychiatric indications have been severely mentally ill for many years and are thus at high risk of suicide and, possibly, other comorbidities.

The ethical evaluation of deep brain stimulation in patients with mental illness

The ethical evaluation of DBS for mental illness does not require any specific, additional ethical criteria beyond those that are already customary, well-established, and generally accepted in medical research and treatment, namely: beneficence/utility, non-maleficence, respect for the patient’s autonomy, and equitable access and distribution (27, 28). These four criteria should not be applied to a prospective DBS procedure in isolation, but also to any alternative forms of treatment being considered at the same time, such as psychoactive drugs or psychotherapy, so that the ethical aspects of all potential treatments can be weighed against each other (28). Much useful information can be gained by analogy and extrapolation from the more than 30 years of research and clinical experience on DBS for other indications (e.g., Parkinson’s disease). Research on DBS performed specifically to treat mental illness should be designed to yield data not only on neuropsychiatric outcomes in the narrow sense, but also on further ethically relevant outcomes. Thus, the study and ethical evaluation of the utility of DBS must not solely be based on statistically measurable effects in clinical and performance-based outcome measures; they must also take account of individually meaningful goals that are relevant to the patient’s everyday life (Box). It follows that:

BOX. The authors’ proposals for practical implementation.

1. Deep brain stimulation (DBS) is now being investigated in research studies as a possible treatment for very severe, intractable mental illness. The lack of response to conventional psychotherapeutic and pharmacotherapeutic methods must be established and documented. The indication for surgery need not be confirmed by an independent psychiatrist; the interdisciplinary treatment team assumes responsibility for this determination.

2. Promising initial results have been published for the use of the superolateral medial forebrain bundle (slMFB) as the stereotactic target in the treatment of both obsessive-compulsive disorder and depression, and for the use of the subgenual cingulum (cg 25) and the anterior limb of the internal capsule (ALIC) as targets in depression (figure). Two controlled trials of DBS for obsessive-compulsive disorder (targets: the nucleus accumbens septi (NAc) and the bed nucleus of the stria terminalis (BNST)) have yielded positive findings. In all of the studies published to date, stimulation was at a frequency of 130 Hz, as for neurological indications.

3. No evidence beyond that of individual case reports is yet available on DBS for other psychiatric indications (anorexia nervosa, dementia, anxiety disorders, addictive disorders, obesity, hyperaggressiveness, schizophrenia).

4. The determination of resistance to treatment in a patient with obsessive-compulsive disorder or depression can only be made after the failure of treatment with three clearly different drugs as recommended in the relevant guidelines (an augmentation strategy with lithium, atypical neuroleptic drugs, or thyroid hormones) and the failure of two different psychotherapeutic techniques that were performed according to the relevant guidelines. Until DBS for psychiatric indications becomes clinically established, it is inadvisable to offer the procedure to patients who have comorbid personality disorders in addition.

5. A patient with depression can only be offered the opportunity to participate in a DBS study after the failure of treatment that has been carried out in conformity with the relevant guidelines, including electroconvulsive therapy.

6. The experimental treatment of psychiatric patients with DBS remains a psychiatric treatment, in particular with respect to patient selection and post-procedural care.

7. Patients can only be treated in study centers in which interdisciplinary care (i.e., both psychiatric and neurosurgical care) can definitively be provided over the long term in the years after the procedure.

8. On the neurosurgical side as well, expertise in the surgical treatment of mental illness is needed, both in view of the specific needs of patients with intractable depression and obsessive-compulsive disorder and in view the specific operative methods used, which differ from those used to treat neurological diseases.

9. Given the severity of the underlying disease in any patient with depression or obsessive-compulsive disorder who undergoes treatment with DBS, it is recommended that the patient’s preoperative psycho- and pharmacotherapy should be continued after surgery until a durable response to the overall treatment has been achieved. In the studies published to date, this has taken up to one year after the procedure.

10. 10. Treatment outcomes should be clinically assessed, ideally by the treating psychiatrist. For patients with depression, it is also advisable to use the Montgomery-Åsberg Depression Rating Scale (MADRS) as an instrument for ongoing assessment.

11. 11. Specific international trial registries should be established, in order to collect outcome data systematically, increase overall experience, enable the development of specific therapies, and avoid the problem of reporting bias, particularly in this field, which is highly susceptible to it.

12. 12. Single case reports and small case series of DBS for mental illness should also be standardized and pre-registered in a specific international trial registry (29).

13. 13. Given the current state of the evidence on DBS for intractable OCD and depression, trials must be planned whose design will enable them to provide a basis for future clinical use. We thus advocate restricting psychiatric DBS, for the time being, to the setting of randomized trials.

• data on patient-reported outcomes should be acquired and reported, even in early-phase studies;

• the use of functional surrogate parameters should be viewed critically; and

• before neurotechnology is used to treat a patient, the individual psychosocial context should be examined in detail, and the patient’s psychosocial environment should be included to the greatest possible extent in the processes of preoperative patient information and consent and in the planning and performance of the treatment (27, 28).

The benefit of the operation must be weighed against its risks (bleeding and infection, both of which have a probability of less than 1%), as well as the implantation of a device that will presumably have to remain in place for the rest of the patient’s life. The implant may adversely affect future diagnostic studies of various kinds (e.g., MRI) and even certain kinds of everyday activity (e.g., passing through security checkpoints involving magnetic fields). Patients with mental illness for whom DBS is considered, like patients with advanced Parkinson’s disease, are a highly vulnerable group whose underlying illness can strongly affect their ability to form and express a clear intention about surgery. This fact should be borne in mind during the process of patient evaluation, information, and counseling, which should be conducted by an experienced psychiatrist (or more than one), always in multiple sessions with the patient. For the time being, DBS for psychiatric indications should only be performed in the setting of clinical trials, so that data on its benefits and risks, on patients’ decision-making processes, and on resource utilization can be collected systematically over time. In rare, exceptional cases of patients with especially severe disease and an adequate evidence base for treatment, surgery outside of a clinical trial can be considered, but then only on the basis of an explicit protocol resembling that of a clinical trial (a so-called n-of-1 trial), so that the treatment can be carried out and documented in standardized fashion. Such cases can be entered into an international neuromodulation experience registry for DBS studies in psychiatric research, where the relevant data are systematically recorded (29).

Table. Randomized, controlled trials on the efficacy of DBS against intractable depression.

Author (reference no.)	N	Study duration	Definition of response	Definitionof remission
		Longest follow-up	Response rate (%)	remission rate (%)
Target: subgenual cingulate gyrus (cg 25)
Eitan, 2018 (16)	9	13 months	40% reduction MADRS	not reported
		13 months	44.4	not reported
Merkl, 2018 (15)	8	24 months	50% reduction HDRS-24	HDRS-24 ≤ 10
		28 months up to 4 years	6 months: 34 12 months: 25 24 months: 37 28 months: 51 last follow-up: 51	last follow-up: 33.3
Holtzheimer, 2017 (19)	90	12 months	40% reduction MADRS	MADRS ≤ 10
		24 months	6 months: 20 (active group) 17 (control group) 12 months: 30 (active group) 27 (control group) 24 months: 49 (all patients)	6 months: 5 (active group) 7 (control group) 12 months: 18 (active group) 7 (control group) 24 months: 26 (all patients)
Target: ventral capsule/ventral striatum
Dougherty, 2015 (18)	30	4 months	50% reduction MADRS	MADRS ≤ 10
		2 years	4 months: 20 (active group) 14.3 (control group) 1 year:20 (all patients) 2 years: 23.3 (all patients)	1 year: 13 (all patients) 2 years: 20 (all patients)
Target: anterior limb of internal capsule (ALIC)
Raymaekers, 2017 (20)	7	not reported	50% reduction HDRS-17	HDRS-17 ≤ 7
		36 to 97 months	last follow-up: 71	last follow-up: 28.5
Bergfeld, 2016 (17)	25	max. 64 weeks	50% reduction HDRS-17	HDRS-17 ≤ 7
		55 weeks	1 year: 40	1 year: 20
Target: superolateral branch of the medial forebrain bundle (slMFB)
Coenen, 2019 (13)	16	12 months	50 % reduction MADRS	MADRS < 10
		12 months	12 months: 100 (all patients)	12 months: 50 (all patients)

This table is based on the reviews of Dandekar et al. (22), Kisely et al. (21), and Drobisz and Damborská (10), with more recent studies added.

HDRS, Hamilton Depression Rating Scale; MADRS, Montgomery–Åsberg Depression Rating Scale

Figure.

Targets for DBS in the treatment of intractable depression and obsessive-compulsive disorder

* IHb only in depression

ALIC, anterior limb of the internal capsule; BNST, bed nucleus of the stria terminalis; DBS, deep brain stimulation; IHb, lateral habenula; ITP, inferior thalamic peduncle; NAc, nucleus accumbens septi; SCG, cg 25, subgenual cingulum, cg 25; sIMFB, superolateral branch of the medial forebrain bundle; VC/VS, ventral capsule/ventral striatum