The Emerging Use of Brain Stimulation Treatments for Psychiatric Disorders

Clinical studies of rTMS in depression

Due to the capacity of rTMS to induce changes in brain activity over time, it has been considered in the treatment of conditions where abnormal cortical activity is evident. The therapeutic effects of rTMS have been explored now in a range of neuropsychiatric disorders with the majority of research focused on the treatment of MDD. This application was first proposed in the mid 1990s. The first promising results were obtained when high frequency trains were applied to the left dorsolateral prefrontal cortex (DLPFC) [10,11]. This application was based on the observation that left DLPFC was underactive in patients who were depressed in resting positron emission tomography (PET) studies [12]. Initial clinical trials were of short duration but established that rTMS appeared to have some antidepressant effects. Over 15 years, a large number of sham controlled clinical trials have been conducted. However, many of these were small as there has been very limited industry support available for trials of the magnitude that would be usually conducted for device or medication regulatory approval [13].

Trials investigating the use of high-frequency stimulation applied to left DLPFC have been subject to several substantive meta-analyses. For example, the meta-analysis by Schutter et al. [14] involved 30 trials and 1164 patients. This analysis showed a highly significant effect of active treatment compared to placebo on the average reduction in depression severity scores (p < 0.00001). The effect sizes seen in these analyses are similar to those seen in trials of antidepressant medication, although many of the trials have been conducted exclusively in patients who are treatment resistant [14]. Notably, another meta-analysis has clearly demonstrated that effect sizes seen in more recent studies have been greater than those seen in earlier research, supporting the idea that the increase in rTMS dose seen in more recent trials has resulted in improved treatment outcomes [15].

Two large multisite trials have been conducted to date: one industry sponsored and one independently funded. A privately held company, Neuronetics, sponsored a randomized sham controlled trial involving 300 patients who had failed a least one antidepressant medication trial [16]. The duration of treatment extended up to 6 weeks (daily treatment 5 days per week) followed by a 3-week taper. There was a significant antidepressant effect of active compared to sham treatment on most of the outcome measures, though not all. The improvement was most substantial in patients who had failed only one medication, as compared to those who had failed more. The results of this trial were utilized in an application for device approval that was successful in the USA in 2008. The second trial, funded by the National Institute of Mental Health, involved 199 patients randomized to active or sham treatment for up to 6 weeks [17]. There was a statistical advantage of active stimulation over sham in the percentage of patients achieving remission, although the overall rate was low (14.1 versus 5.1%).

Studies have also been conducted to directly compare high frequency rTMS to ECT [18–23]. The majority of these have found no differences between the treatments although their power to find differences was limited. One study, incorporating patients with psychotic depression, showed greater benefit with ECT in the psychotic group [21], while a second study has reported greater effects of ECT [24]. One substantial issue with these trials is that many of them have compared a fixed course of unilateral rTMS to a flexible course of often uni and bilateral ECT. This presumably biases somewhat towards the likelihood of finding a better outcome with ECT.

Methods of rTMS administration

A number of substantial questions remain in regard to optimal rTMS administration. The dose of stimulation, typically reflected in the number and intensity of pulses applied, has progressively increased over time. Interestingly, pilot data has recently suggested that antidepressant effects might be achieved much more rapidly with very high dose intensive protocols [25]: this requires further evaluation. Conversely, it is possible that less frequent treatment than the typical 5 days per week scheduling may be of similar efficacy [26].

Despite high-frequency left sided rTMS having been the most extensively evaluated approach, it is not yet clear whether this is the optimal method of rTMS delivery. Low frequency rTMS applied to the right DLPFC appears to have similar efficacy, may be better tolerated, and safer [27–29]. Bilateral approaches have also shown promise [30], although some recent studies have cast doubt about whether they will prove more effective than unilateral stimulation [31,32]. In addition, a range of newer forms of rTMS including theta burst stimulation [33] and priming stimulation [34] may prove more effective than the standard left side high frequency approach, although they have yet to be evaluated in substantive trials.

The optimal method for targeting the DLPFC also remains uncertain. Almost all trials have identified and targeted DLPFC by measuring 5 cm anterior to the scalp surface corresponding to motor cortex, localized using single TMS pulses [35]. However, this clearly results in inaccurate targeting in the majority of patients, often with subsequent stimulation being applied to premotor cortex [36]. It is possible that improved targeting of DLPFC utilizing structural MRI may enhance clinical responses [37]. However, several functional imaging-based targeting approaches have not resulted in improved outcomes [38,39]. Imaging may not be required to produce optimal response: better outcomes may be obtained with a more anterior and lateral coil location [40], or potentially through the use of electroencephalography (EEG) coordinates [41].

Safety and tolerability

Generally speaking, rTMS approaches appear to be relatively safe and well tolerated [42,43]. The main side effects are discomfort on the scalp at the stimulation site during treatment, or the development of a post-stimulation headache [43]. These effects are highly variable between subjects, but are seen more commonly with high stimulation frequencies and intensities. Tolerability appears to be greater when stimulation is introduced at a lower intensity and gradually increased over time.

In regard to more severe possible consequences, rTMS treatment in depression does not appear to have any deleterious effects on cognition, including memory [43]. There have been several case reports of the induction of mania in patients with bipolar disorder [44] and an early case report of what appeared to be new onset delusions [45]. The major concern with rTMS has been the possibility of seizure induction [42]. The occurrence of seizures seems to have been dramatically limited by the use of safety guidelines introduced in the late 1990s [46] although there have been occasional reports. Few of these have been in patients with depression treated within established safety guidelines. The induction of a vasovagal episode is another possibility which can confound the interpretation of a loss of consciousness and should be suspected in patients with a history of fainting related to other medical procedures.

Limited data is also accumulating on the safety of the use of rTMS in a variety of special populations. Treatment has been provided in small trials or case studies in pregnancy [47], in adolescent depression [48,49], as well as in patients with a variety of neurological complications such as Parkinson's disease [50–52], stroke [53,54] and traumatic brain injury [55].

Effects over time

Depression is clearly a relapsing disorder and many patients experience multiple episodes despite the efficacy of antidepressant medication in relapse prevention [56]. Unfortunately we continue to lack a comprehensive understanding of the long-term effects of rTMS treatment on the course of depression. A recent study investigated relapse rates from 204 patients treated over a number of years with rTMS [57]. Event-free remission rates were 75.3% at 2 months, 60% at 3 months, 42.7% at 4 months, and 22.6% at 6 months. Several studies have suggested that the reinstitution of rTMS treatment during depressive relapse is successful in many patients [58,59]. Limited research has also suggested that some benefit may be obtained from maintenance rTMS schedules (for example [60,61]) although substantial studies are lacking in this area.

rTMS in depression: summary of status

A substantive body of work has clearly established that rTMS treatment has antidepressant efficacy. This efficacy is likely to be similar to that seen with antidepressant medication. Although beneficial effects with rTMS appear greater in less treatment resistant patients, those with a greater degree of treatment resistance have clearly responded in a substantial number of clinical trials. rTMS appears to be relatively safe and well tolerated. For these reasons, rTMS is being increasingly applied in clinical practice internationally. It is likely to be useful for patients who are not suitable for ECT, or prefer to avoid that treatment due to concerns about side effects or stigma. rTMS is not likely to replace ECT as a rapidly and powerfully effective antidepressant, but is certainly likely to reduce the need for ECT treatment in a substantive number of patients.

rTMS in schizophrenia

A considerable number of trials have investigated the use of rTMS in the treatment of patients with schizophrenia [62,63]. Quite a number of these studies have not had a specific symptoms focus, and have not generated promising results. However, more hypothesis-driven approaches have produced interesting findings. For example, low frequency stimulation applied to temporoparietal cortex has been used in the potential treatment of refractory hallucinations. The majority of trials of this application have demonstrated benefits of active stimulation over sham (e.g. [64–66]) or ongoing medication treatment only [67], although there have been some negative studies (e.g. [68]). The efficacy of this form of stimulation has been suggested by several meta-analyses (e.g. [69]), and its clinical use was advocated in recent revisions of the influential ‘PORT’ clinical guidelines [70]. However, most of the studies to date have been short term; despite some evidence of the persistence of therapeutic benefits over time [66] the long-term impact of this form of treatment on patients' clinical course remains uncertain.

A second approach has been the use of high frequency stimulation applied to left (or bilateral) prefrontal cortex in the treatment of negative symptoms. There have been both positive [71–73] and negative [74–77] studies in this regard; more substantive, larger and longer-term trials are required.

rTMS in other psychiatric disorders

The use of rTMS has also been evaluated in a number of other psychiatric disorders. However, most of the studies have been small, and limited attempts have been made at replication. Several studies have explored the use of rTMS in mania. High frequency frontal stimulation on the right was initially suggested to be superior to left sided stimulation and sham [78]. However, this was not confirmed in a subsequent study of active versus sham right sided stimulation [79].

In obsessive compulsive disorder (OCD) there has been some inconsistency in the stimulation method applied. Very early on, single rTMS sessions at high frequency on the right DLPFC appeared to produce some benefits [80]. These benefits were also seen in a small, early, non-sham controlled trial with both left and right sided stimulation [81]. However, subsequent studies of both right and left sided (high and low frequency) rTMS have not shown therapeutic benefit [82–85].

Benefits have also been seen in post traumatic stress disorder (PTSD) from single sessions of rTMS [86], as well as in a sham controlled trial of high frequency right PFC stimulation [87]. Negative effects were seen with left sided stimulation [88].

In panic disorder there was initial promise in open label data [89,90] but this has not been supported in a small trial with serotonin reuptake inhibitor medication resistant patients [91].

Finally, research is underway to establish if rTMS has therapeutic potential in addictive disorders. Single session studies have demonstrated that prefrontal rTMS can reduce craving in cocaine or nicotine dependent subjects [92,93]. Two more recent double-blind studies have shown positive therapeutic effects of prefrontal rTMS in alcohol dependence and in nicotine dependence [94,95]. Although both these studies involved longer periods of stimulation, they used divergent rTMS methods; high frequency stimulation was applied on the right in one study, and to the left DLPFC in the other.

Low intensity magnetic stimulation: summary of status

It is obviously early days for research into the brain effects of low intensity magnetic stimulation. However, the data gathered to date suggests that this form of stimulation does have brain effects that may be relevant to the modulation of mood. Further research to explore the therapeutic capacity of these systems is justified.

Transcranial direct current stimulation

An alternative, non-invasive way to modulate brain activity is with the application of a low voltage electrical current. Several forms of electrical stimulation have been developed and tested to a greater or lesser degree in psychiatric disorders.

Transcranial direct current stimulation (tDCS) is a technique that involves the application of a low amplitude (1–2 mA) direct current to the brain through two surface electrodes placed on the scalp [102]. Rubber electrode pads covered with sponges are connected to a low voltage stimulation device. The technology for generating a tDCS current is very basic, and the current itself may be generated with devices run by commonly available batteries. Stimulation is usually applied continually for a period of time, commonly between 15 and 20 minutes.

The notion of tDCS is a relatively old one, with researchers proposing the application of this type of technique during the 1960s and 1970s. However, despite initial enthusiasm, interest in the field faded until it was rediscovered about 10 years ago, through the conduct of a series of neurophysiological studies demonstrating the capacity of tDCS to modulate brain activity [103–105]. This research, which has progressively expanded, has now characterized a variety of aspects of the effect of tDCS on the brain. Most importantly, it has become relatively clear that anodal stimulation (stimulation under the positive electrode) produces a localized increase in cortical excitability [106]. In contrast, a localized decrease in cortical excitability is produced under the cathode [106]. Therefore, either uni-modal or bimodal effects may be produced depending on whether both electrodes are placed on the scalp, or if one is placed in a non-cephalic position.

The immediate effect of tDCS is likely to occur through subtle changes in membrane polarization, related to a small degree of the applied current passing into the brain [107]. However, the effects of tDCS have been shown to last for up to 1 hour after of a single session of stimulation; these more persistent effects may well have a more complicated origin in the brain [107]. For example, lasting effects have been shown to be dependent on activity at the NMDA receptor, and to be modulated by a variety of drugs that affect this receptor in addition to calcium channels [107].

These persistent, but temporary tDCS effects are increasingly being used as a way of non-invasively modulating brain function in a variety of cognitive neuroscience experiments. There is also increasing interest in their potential therapeutic capacity. Following on from the prefrontal rTMS model, the main therapeutic possibility assessed to date has been the use of anodal stimulation applied to the left DLPFC in patients with depression. Four out of five patients responded to 1 week of this form of stimulation in the first clinical trial, compared to no responders in a sham group [108]. In a larger follow up study, active left prefrontal tDCS again resulted in a greater clinical response than sham and occipital stimulation [109]. Several other groups have now explored tDCS effects. Ferrucci et al. applied tDCS twice daily (20-minute sessions at 2 mA) in a group of 14 patients with severe MDD referred for ECT in an open label manner. They found substantial antidepressant effects that interestingly appeared to continue to accumulate after the end of the course of stimulation [110]. The same parameters were then used in a larger sample of both unipolar and bipolar depressed subjects with similar results [111]. In contrast, Loo et al. provided a lower dose stimulation (1 mA) over five stimulation sessions to 40 patients in a sham controlled trial and found no antidepressant effects [112]. Several notable tDCS case reports have also been published: in one patient, depression that developed following stroke was treated and showed a substantial antidepressant effect [113]. In a second case report, mania was induced in a patient with bipolar disorder following frontal stimulation with an extra-cephalic cathode [114].

Although no substantial studies have been conducted to date, the safety profile for tDCS looks relatively benign [107,105]. Itching, tingling, headache and a burning sensation are the most commonly reported side effects and appear transient [115,116]. There has been some concern about the possibility of brain stem effects on respiration when non-cephalic electrodes are used; however, a recent study investigating this issue could find no evidence of adverse events in healthy volunteers [117].

Cranial electrical stimulation

Another form of low voltage electrical brain stimulation is cranial electrical stimulation (CES). CES describes a variety of methods of stimulating the brain, typically comprising alternating, low voltage electrical currents. Forms of CES have been applied to altering brain activity for several centuries (see review in Stagg and Nitsche [107]), but the use of a wide variety of stimulation parameters has been included under the banner of CES, including tDCS as described above. This degree of variability confounds interpretation of the studies conducted to date. A considerable degree of the early development of CES techniques occurred in the former USSR, with data not widely available in English [118]. In a number of countries, including the USA and Australia, CES devices have become commercially available and are being marketed for the treatment of a variety of disorders. There is also considerable variability in the stimulation provided by different devices. For example, the ‘Alpha-Stim SCS’ has been marketed for the treatment of conditions including anxiety disorders, depression and insomnia, and supplies stimulation with bipolar rectangular pulses, provided at low or high frequency and adjusted across an amplitude range, through electrode clips placed on the earlobes. Although such CES devices are often marketed for a variety of indications, there is a very limited evidence base for most of these applications. A variety of open label research trials (for example Bystritsky et al. [119]) and studies in mixed samples have been conducted; many have lacked consistent or substantively validated methods. At this time, the majority of the claims made about the effect of this form of stimulation lack the required support of substantive sham controlled trials in well characterized populations.

Electrical stimulation: summary of status

A small but emerging literature suggests that tDCS may have antidepressant activity, although this requires confirmation in more substantial samples. It is a promising technique, as its low cost suggests it could be a useful alternative treatment in developing countries. There is little evidence to recommend the clinical use of CES at this stage.

Electroconvulsive therapy

ECT remains a widely used and highly effective psychiatric treatment. Its main indication continues to be in the treatment of patients with resistant depression or depression requiring a rapid antidepressant response. The induction of cognitive side effects, particularly anterograde and retrograde amnesia, and the considerable stigma associated with the treatment are ongoing issues relating to the use of ECT. Resultant resistance to its use exists within most communities.

ECT evolved out of pharmacological methods of seizure induction, and different forms of ECT have substantially different efficacy/side-effect profiles [120]. It therefore seems likely that the therapeutic effects and cognitive side effects of ECT are potentially dissociable. It may therefore be possible to find a method of seizure induction that produces the therapeutic benefits associated with ECT without the same cognitive side-effect profile. However, this is not inevitable. The therapeutic potency of ECT may not relate just to the induction of a seizure, but may also be dependent on the actual electrical stimulation of brain structures. This question can only be answered through the implementation of trials of alternative seizure induction techniques that do not produce the same degree of direct, widespread electrical stimulation of brain areas. It may be possible to reduce the spread through the brain of electrical activation produced with ECT with the use of more focal methods of electrical stimulation [121]. However, some degree of shunting across the scalp will occur with any directly applied electrical current, reducing the focal extent of the stimulation.

Magnetic seizure therapy

An alternative method of seizure induction without any diffusion of the stimulus is through the use of a high-powered transcranial magnetic stimulation device. Magnetic seizure therapy (MST) uses high frequency and high intensity magnetic fields to generate a seizure, applying a highly focused magnetic field which repeatedly stimulates local cortical neurones until seizure activity is induced [122]. There is a spread of the seizure through the brain, but no spread of the stimulation field. As with ECT, MST is administered under a general anaesthetic and utilizes similar procedures.

Following the initial proposition of the possibility of MST, early studies focused on establishing whether it would have a more advantageous side-effect profile while attempting to understand the stimulation characteristics capable of seizure induction. These studies were limited by the capability of available stimulators which could only provide short stimulus trains at high power at approximately 50 Hz [123]. This equipment was not able to induce seizures in all subjects, and there was limited capacity for stimulating above an individual subject's seizure threshold [123].

Despite these limitations, the initial MST studies provided some important information. In rhesus monkeys, MST was shown to produce no problematic histological changes [124,125] and appeared to have less cognitive side effects than the animal ECT equivalent [126]. Initial human studies also indicated that MST appeared to have a favourable side-effect profile [127,128]. Initial efficacy data indicated MST had antidepressant properties, but that these may be less than those produced with ECT [129].

A second generation of MST studies has now commenced, utilizing newly developed equipment capable of stimulating at higher intensities up to 100 Hz [123]. Primate studies have shown much more reliable seizure induction with high frequency MST than lower frequency stimulation, whilst still demonstrating fewer cognitive side effects than conventional ECT [130]. Initial human data is also emerging. In the first direct 100 Hz MST and ECT comparison study, similar antidepressant effects were seen between MST and right unilateral ECT; MST also appeared to have a favourable side-effect profile [131]. In a separate study, MST was shown to have antidepressant properties, and appeared to be associated with a rapid return of orientation [131].

MST: summary of status

It is clearly too early to make conclusions about the potential role of MST. However, if direct head-to-head trials prove that it has similar efficacy to ECT, MST could be relatively rapidly rolled out into clinical practice; the infrastructure for the provision of MST largely already exists in the form of standard ECT suites. Although MST has many similarities to ECT, the alternative method of seizure induction and lack of a problematic history will most likely result in substantially less stigma being associated with this treatment. As a consequence, there may be greater patient and community acceptance of MST. Although head-to-head MST–ECT studies are already underway, considerable further research is required to define the optimal methods of MST stimulation; half a century of ECT research has yet to allow us to fully understand the best way to provide this treatment. Factors that require exploration include the optimal frequency/intensity combination, the most effective target site and coil type, and whether the optimal characteristics for seizure induction are the same as the optimal characteristics for antidepressant efficacy.

Focal electrically administered convulsive therapy

Since its inception there has been a progressive improvement in the risk–benefit profile achieved with ECT; a long series of studies have refined knowledge in regard to a variety of parameters of ECT application. For example, changes in the type of pulse applied, the electrode placement and more recently the pulse width have improved the application of ECT [120,132]. However, a number of ECT parameters have not been systematically explored, such as the direction of electrical current and the size and shape of stimulation electrodes. In addition, the focality of ECT stimulation remains very poor due to the shunting of current across the skull. Focal electrically administered convulsive therapy (FEAST) has been proposed as an alternative convulsive or non-convulsive therapy with substantially greater capacity for focused brain stimulation [132].

To date, FEAST involves the use of a unidirectional electrical current provided between two electrodes that vary substantially in size [121,133]. The current passes between a small anterior and large posterior electrode, both placed on the same hemisphere. This type of electrical stimulation appears capable of producing local seizure activity that does not generalize into a tonic clonic seizure [132]. At higher stimulation voltages a generalized convulsion may be produced [132]. To date, research has only established the feasibility of this type of stimulation in non-human primates and it is yet to be determined if either convulsive or non-convulsive forms of FEAST have clinical utility.

Vagal nerve stimulation

Vagal nerve stimulation (VNS) involves the surgical implantation of a pulse generator, similar to a pacemaker, in the chest. This is connected to a stimulating electrode which is attached to the vagus nerve in the neck [134,135]. Stimulation is applied to the vagus nerve continuously, although the stimulation parameters may be adjusted. The main existing indication for VNS is in the treatment of refractory epilepsy; VNS stimulation can reduce seizure frequency but does not commonly allow patients to cease anticonvulsant medication treatment [136].

The first potential use of VNS in psychiatry arose from the observation that patients treated with VNS for epilepsy occasionally experienced mood improvement and that VNS produced changes in brain activation in depression relevant brain regions [134,137]. The report from the initial open label trial of VNS for depression involved 30 patients stimulated for 10 weeks [138]. Between 40 and 50% of the patients achieved clinical response criteria and this response appeared to persist or improve during follow up [139]. Results with a larger sample of 59 patients were more modest (30.5% responders after 10 weeks of treatment, 15.3% remitted), and VNS was found to be less successful for patients who had failed a greater number of medication trials [140].

Subsequently, a multicentre randomized trial was conducted with intended device registration. The pivotal D02 trial was a 10 week double-blind trial [140]. The response rate in the double-blind phase was low and not statistically different between the active and sham groups [140]. When all subjects were followed up at 9 months, the response rate was approximately 30%.

In parallel to the double-blind trial, a group of patients receiving treatment as usual were also evaluated over 12 months. In this analysis, a greater proportion of patients receiving VNS (27%) achieved response by 12 months than in the treatment as usual group (13%) [140]. In a more recent analysis of data from the early studies, Sackeim et al. [141] found that patients who achieved clinical response appeared to maintain it without other changes in antidepressant treatment, despite high levels of chronicity and treatment resistance. In a small additional sample of patients, a > 50% response rate was seen (6 of 11) over 12 months [142]. However, treatment resulted in persistent vocal cord palsies in two patients lasting 2 and 6 months respectively, and a non-responding patient committed suicide.

Vocal cord effects are one of the main side effects of VNS. An alteration to voice, neck discomfort, cough, dysphagia and shortness of breath can all occur, with vocal changes potentially persisting over time [139]. However, VNS does not appear to cause cognitive impairment [143].

VNS therapy was approved by the Food and Drug Administration (FDA) in the USA in 2005 for the treatment of depression (uni- or bipolar) which has not responded to at least four medication trials. Since device registration, the VNS device manufacturing company has conducted a double-blind randomized dose study in 331 patients enrolled across 29 centres in the USA. Response to three levels of stimulation dose was compared during a 22 week acute phase and after 1 year of follow up. The results of this study have not been published in the peer-reviewed literature, but company materials describe a 12 month response rate of between 25 and 50% depending on the rating scale used or level of stimulation intensity [144].

In addition to the use of VNS in depression, several other applications have been proposed. A small open label trial has suggested that VNS may have some efficacy in refractory anxiety disorders [145], and its potential use in obesity and pain management have been suggested but not yet evaluated [146,147].

VNS: summary of status

The data collected and published to date supporting the use of VNS in the treatment of depression is quite limited. However, VNS does appear to have some antidepressant effects and the profile of response to this treatment is substantially and promisingly different from that produced with a variety of other treatment techniques. Antidepressant effects appear to accumulate slowly over time and to persist, with little suggestion in the data so far of the problematic relapse rates common after other acute interventions. However, given the relatively low overall response rate, approaches to better define which patients are likely to respond to VNS are urgently required.

Deep brain stimulation

Deep brain stimulation (DBS) is the second surgical intervention for psychiatric disorders that has evolved from a neurological indication. DBS was developed for the treatment of Parkinson's disease; it is now relatively widely used in this illness as well as in dystonia and tremor disorders [148–152]. DBS also involves the implantation of a pulse generator, like a cardiac pacemaker, in the chest. This is connected to stimulation electrodes which are placed in localized brain regions. Standard DBS equipment involves four closely spaced electrodes at the end of electrode wire; it is likely that a greater variety of hardware alternatives will emerge over time. The treating clinician is able to control the electrodes between which the current flows, as well as current parameters such as voltage, frequency and pulse width. The placement of the electrodes determines their effects: in movement disorders implantation is usually in basal ganglia nuclei such as the subthalamic nucleus or the globus pallidus [153]. Although DBS has been widely conceptualized as producing a ‘reversible lesion’, the mechanism of action of DBS continues to remain unclear. It is possible that DBS actually produces functionally relevant changes through synchronization of local and distal activity, rather than just a lesion affect [154,155].

DBS: summary of status

DBS is clearly a treatment that will be reserved for the most refractory patients due to its invasive nature. However, it appears to have significant therapeutic promise. In 2009, the US FDA granted humanitarian device exemption for the use of a DBS stimulation device in the treatment of OCD. This provides access to DBS therapy for patients with OCD without the conduct of a large-scale placebo-controlled trial by the sponsoring company, a development that has been somewhat controversial [171]. Further research is clearly required to understand the optimal targets for DBS stimulation and also to better understand the optimal stimulation profiles, long-term outcomes and whether likely treatment responders can be preselected.

Epidural cortical stimulation

Epidural cortical stimulation (ECS) is a third surgical option, but one with a very limited research base. ECS involves the implantation of a series of electrodes across the cortical surface. It has been investigated for a number of potential indications where modulating cortical activity may have therapeutic benefit, including in movement disorders and chronic pain [172,173]. Very limited research on the therapeutic benefit of prefrontal ECS in refractory depression has been conducted to date. Some beneficial effects for six patients in an industry sponsored trial were reported in abstract form in 2008 [174] but have not yet been published in detail. In a small open label series of five patients, Nahas et al. reported antidepressant effects: three patients achieved remission with persistence of antidepressant response at 7 month follow up [175].