The report by Brunelin et al. describes the first-ever transcranial direct-current stimulation (tDCS) clinical trial targeting auditory/verbal hallucinations (AVHs) in schizophrenia, a core positive symptom that is often highly distressing and disabling. Multiple studies of repetitive transcranial magnetic stimulation (rTMS) targeting AVHs have been described with overall promise, but negative studies as well. rTMS, for this application, induces actual neuronal firing at 1 Hertz, producing cortical effects akin to long-term depression. Instead, tDCS achieves physiological results through more subtle, sustained shifts in membrane polarity. The tDCS protocol reported by Brunelin et al. has a sound rationale, with the suppressive cathode positioned over the left temporoparietal junction (where cortical activation generating AVHs is likely to occur), and the activating anodal electrode over a left prefrontal site (possibly boosting deficient corollary discharge postulated to contribute also to AVHs). There is good reason to believe that blindness was maintained for their patients—the somatic sensation of tDCS is minimal and easily replicated in the sham condition. Their effect size contrasting tDCS with sham stimulation in terms of reduced AVHs was most impressive (1.58), with robust durability of effects three months after the intervention (with continued trends in the direction of further improvement), and statistically significant improvements in overall PANSS scores, again with an impressive effect size. This approach, therefore, could reflect a significant advance over current rTMS methodologies for this indication. On the practical side, tDCS systems are less expensive than rTMS systems.

The sample size of the Brunelin study is not large, so results must be viewed as preliminary, and the proof of the pudding will, of course, be to ascertain if other groups can replicate these results. Moreover, a mean 38 percent reduction in AVHs reported at final assessment indicates that there is room for further improvement. The robustness of their effect size for tDCS versus sham stimulation in part reflected minimal symptom improvements detected for the latter rather than overly large improvements in AVHs for the active condition—possibly arising from the fact that the Brunelin protocol spanned only five days—thereby reducing chances of nonspecific effects. We don't know if an extension of tDCS beyond their twice-per-day, five-day protocol would further expand these treatment effects. It remains my hope that brain stimulation protocols specifically targeting AVHs—with further refinements based on a growing body of knowledge regarding underlying pathophysiology—will be able to achieve full remission in the great majority of patients. The exciting results of the study by Brunelin et al. could be an important step forward in this quest.

View all comments by Ralph Hoffman

I was intrigued by the findings reported by Brunelin and colleagues on tDCS as a treatment for hallucinations. There have been increasing reports that this non-invasive neurostimulation technique may be efficacious in the treatment of a range of psychiatric and neurological disorders, including depression, stroke, pain, and epilepsy.

Brunelin and colleagues focused tDCS activity on two cortical "nodes" to provide local and opposing effects on two different brain cortical areas—the left temporo-parietal junction (cathodal inhibition), and the left dorsolateral prefrontal cortex (DLPFC, anodal stimulation). The results showed positive effects of the active tDCS treatment on auditory verbal hallucinations (31 percent hallucination score reduction vs. 8 percent reduction in the sham treatment), which were sustained after a three-month period. There was also a reduction in the negative symptom dimension, though not in the disorganization and grandiosity dimensions.

I would have been interested to see a finer-grained analysis of the findings. Specifically, which dimension of auditory hallucinations showed "positive effects"? Was it the frequency of hallucinatory experiences? Was there a change in negative content? Alternatively, was there reduced interference, or enhanced insight (a domain that is rarely examined in clinical studies of hallucinations)? A closer examination of AHRS scores should provide some of these answers.

A related comment refers to the effects of tDCS on the cognitive mechanisms that underlie auditory hallucinations. Many studies have shown the facilitation of cognitive functions by anodal tDCS and potential for cognitive enhancement. Specifically, anodal tDCS over the left DLPFC has been shown to improve executive abilities in domains of inhibition and working memory performance (e.g., Ditye et al., 2012; Zaehle, 2012). Given the demonstrated link between reduced cognitive inhibition and frequency of auditory hallucinations (e.g., see Waters et al., 2012, for review), could one possible explanation for Brunelin’s finding of improvements in hallucinations be tied, at least in part, to cognitive improvements from the stimulation to the left DLPFC? In other words, did the tDCS anodal effects over the DLPFC improve the performance of patients on executive functions, specifically, inhibitory control? If so, one might speculate that such tDCS-mediated enhanced control abilities over thought and action might reduce the frequency of hallucinations by enabling more effective control over the onset and content of these experiences.

Furthermore, given that broad inhibitory dysfunctions are a common feature of auditory hallucinations in borderline personality disorder (Grootens et al., 2008) and in individuals who are "prone" to auditory hallucinations (as measured with the Launay-Slade Hallucination Scale-LSHS-R), one might anticipate broad beneficial effects of tDCS over hallucinations in a range of conditions. Inhibitory dysfunctions have also been demonstrated in populations with visual hallucinations (e.g., Parkinson’s disease), pointing to cross-modal mechanisms that might be amenable to investigations with tDCS.

By contrast, the tDCS cathode over the left temporo-parietal junction might be more closely linked to this region’s function involved in generating auditory signals. According to this formulation, one (simplistic) explanation might be that the inhibitory functions of the tDCS cathode act to dampen aberrant and hypersalient auditory signals, caused by deviant trigger of activations in language-related areas (see Jardri et al., 2011).

Overall, the combined use of cathode and anode over the temporo-parietal junction and DLPFC, respectively, might improve hallucinations by enhancing control over mental contents, and reducing the salience of aberrant auditory signals.

While the above speculations about the role of the anode or cathode are tentative, they provide one explanation for the beneficial effects of tDCS over brain cortical regions which have been linked to auditory hallucinations.

References:

Ditye T, Jacobson L, Walf V, Lavidor M (2012) Modulating behavioral inhibition by tDCS combined with cognitive training. Experimental Brain Research. Abstract

Jardri R, Pouchet A, Pins D, Thomas P. Cortical activations during auditory verbal hallucinations in schizophrenia: a coordinate-based meta-analysis. Am J Psychiatry. 2011;168: 73–81. Abstract

Grootens K, van Luijtelaar G, Buitelaar J, van der Laan A, Hummelen J, Verkes R (2008) Inhibition errors in borderline personality disorders with psychotic-like symptoms. Progress in NeuroPsychopharmacology and Biology Psychiatry, 32(1), 267-273. Abstract

Waters F, Allen P, Aleman A, Fernyhough C, Woodward T, Badcock J, Barkus E, Johns L, Varese F, Menon M, Vercamen A, Laroi F (2012) Auditory hallucinations in schizophrenia and nonschizophrenia populations: A review and integrated model of cognitive mechanisms. Schizophrenia Bulletin. Abstract

View all comments by Flavie Waters

Brunelin and colleagues present important and exciting findings in this American Journal of Psychiatry report. After an initial enthusiastic surge of positive reports on the efficacy of rTMS in the treatment of auditory-verbal hallucinations several years ago, recent findings have painted a rather less encouraging picture. Larger-scale, well-controlled studies failed to replicate the beneficial effect of low-frequency rTMS of the left temporoparietal regions compared to sham. This suggests that the effects are not as robust as originally believed and/or that earlier positive findings may have been due to small sample-related biases.

As this new report suggests, tDCS may provide an alternative and potentially superior approach to the treatment of psychotic symptoms in medication-resistant schizophrenia. Perhaps the key component here is the use of a dual-target electrode montage, focusing on two sites known to show functional (and structural) alterations in schizophrenia. As the authors suggest, the idea that frontotemporal connectivity may be at the heart of some of the symptoms of schizophrenia has gained some traction in recent times, with several reports of structural changes in white matter tracts connecting the frontal and temporal cortices, as well as aberrant functional connectivity measured during the resting state or during task-related processing. The application of cathodal and anodal stimulation on hyperactivated and hypoactivated sites, respectively, seems to provide a very promising avenue of investigation, and could explain the more generalized effects on negative and general symptoms.

Another advantage of tDCS lies in the availability of a decidedly more reliable placebo condition. Sham and active tDCS produce very similar superficial sensations, and double blinding of experiments is more feasible. In rTMS, on the other hand, the patient may well be aware that she is receiving sham stimulation, particular in the case of crossover designs. Brunelin et al. applied a parallel design, so their assessment of blinding effectiveness could be considered to be somewhat biased, due to a lack of intra-individual comparison. The authors state that: “patients reported they could not tell which group they had been allocated to.” Nevertheless, it would have been interesting to know the percentage of correct “guessed allocations.”

Importantly, however, patients were well matched on relevant characteristics at baseline, which strengthens the conclusion that the observed improvements in symptoms were due to true tDCS effects. Replication in a larger cohort, using a crossover design, will be imperative to further substantiate these promising results.

Although the presented findings are highly encouraging, research targeting clinical improvement through brain stimulation faces a number of important considerations. First, recent reports of symptom improvement following high-frequency rTMS stimulation of the temporal cortex challenge the current understanding of hallucinations in terms of hyperactivated cortex, and/or the neurophysiological mechanisms underlying brain stimulation. Additionally, it seems wise to venture beyond the traditional left hemispheric targets, given evidence of significant interindividual variability in neuronal correlates of symptomatology and the role of right hemispheric "language" centers in the generation of hallucinations. Importantly, we should take into account both local effects on the stimulated sites, and effects on functionally connected brain regions, which may be compensatory in nature. Accepted knowledge on functional and structural connectivity should be used to guide the application of excitatory and inhibitory stimulation protocols. Finally, the treatment duration in the report by Brunelin et al. was short, and full remission was not achieved in any of the patients. As Ralph Hoffman stated in his comment, it will be essential to verify whether additional benefits could be achieved with longer durations of treatment. Given the maintenance of the achieved improvement over three months in the absence of additional stimulation sessions, one could speculate that further treatment—even at a frequency of once a day or less, up to a few weeks—might produce long-lasting and superior effects.

In all, Brunelin and colleagues have blazed a trail for novel applications of tDCS in the management of schizophrenia. Needless to say, further explorations of the possibilities afforded by this technique are required, and independent replication will be imperative to advance along this promising avenue of investigation.

View all comments by Ans Vercammen

Is direct current stimulation a wonder method for schizophrenia?

Written in collaboration with André Aleman, Christina W. Slotema and Dennis J.L.G. Schutter

With great interest, we read the randomized controlled trial (RCT) by Brunelin et al. in advance publication in the American Journal of Psychiatry. In this study, transcranial direct current stimulation (tDCS) was administered over the left temporo-parietal and frontal cortex for the treatment of both auditory verbal hallucinations and negative symptoms in medication-refractory patients with schizophrenia. Fifteen patients received a series of 10 2-mA tDCS treatments, applied twice daily and another 15 patients received a series of 10 sham treatments. A 31% improvement in clinical symptoms was observed in the active treatment group, as compared to an 8% improvement in the placebo treatment group. For the severity of the auditory verbal hallucinations, an acute large effect size of 1.58 was reported (95% CI=0.76–2.40). Indeed, this is a very large effect size, when compared, for example, to the effects of antipsychotic medication that range between 0.4 and 0.6 (Leucht et al., 2012). Furthermore, the effect observed by Brunelin et al lasted for at least 3 months and appeared to become even larger over time (i.e., 38% improvement). For the severity of negative symptoms the effect size was also very large: 1.07 (95% CI=0.30-1.84).

This study is important as the results may open new avenues in the somatic treatment of medication-refractory patients suffering from auditory verbal hallucinations and negative symptoms by exposing the brain to low intensity currents in a safe and non-invasive way. Despite these promising effects, we will argue that additional and larger RCT are needed to rule out publication biases that are commonly observed when new techniques are introduced. In addition, unraveling the working mechanisms by which tDCS establishes its effects in the human brain is crucial for evaluating clinical relevance.

The unknowns surrounding the working mechanisms of tDCS

The biophysical principle by which tDCS establishes its effects involves the polarization of superficial nervous tissue by applying a constant, one directional, flow of electric current between two electrodes attached to the scalp. The administered voltage gradient is believed to have differential effects on level of cortical excitability. The positively charged electrode (anode) causes an inward current and is suggested to increase neuronal excitability, whereas the negatively charged electrode (cathode) induces an inward going current and decreases neural excitability (Nitsche & Paulus, 2000).

Long-term potentiation (LTP) and long-term depression (LTD) have been put forward as possible physiological mechanisms underlying the distinct effects of low current stimulation, but the way these phenomena add to understanding the presently observed large improvements over time remains elusive. Moreover, specific knowledge concerning the structural and functional organization of the cerebral cortex is crucial for understanding the physiological effects of tDCS and its subsequent relation to its potential to improve auditory verbal hallucinations and negative symptoms.

Mechanistic explanations limited to describing non-invasive brain stimulation effects in terms of decreasing and/or increasing assumed hyper- and/or hypo-excitable brain regions may not be the whole story in capturing the brain’s modus operandi. Furthermore, orientation and physiological state of the stimulated nerve tissue have been shown to determine whether neurons are brought into a depolarized or hyperpolarized state (Kabakov et al., 2012).

In addition, the notion that oscillations constitute an imperative property of brain functioning suggests that, in addition to explaining the neurophysiological effects of electric current stimulation mainly from an amplitude modulation perspective, studying tDCS effects in the context of frequency modulation may turn out to be a more fruitful approach (Jacobson et al., 2012).

In any case, the need for understanding the neural processes underlying clinical improvement provides evidence in its own right. Unraveling the mechanisms by which tDCS, and any treatment for that matter, establishes effects in the brain are critical for evaluating its potential suitability as a somatic treatment.

Efficacy of transcranial magnetic stimulation (TMS) for schizophrenia
The first RCT by Hoffman and colleagues (2003) reported a large effect size of 0.8 of TMS as compared to placebo on the severity of hallucinations, which was followed by a second publication of another group finding an even larger effect size of 1.2 (Brunelin et al., 2006). Both studies included only a limited sample size. However, a few years later negative studies also were published. At this time, 17 double blind placebo-controlled TMS studies on hallucinations have been published. The mean weighted effect size of these studies is now 0.3-0.44, depending on the site of stimulation, which is still superior to sham treatment (Slotema et al., in review), but the negative correlation between effect size and year of publication suggests that over time, the mean effect size may become even smaller.

Trends over time in effect sizes of new treatment methods
The trend of effect sizes to decrease over time is by no means specific for TMS or tDCS. It is a very general trend that can be observed whenever new treatment methods are introduced (Emerson et al., 2010) Such trends show that effect sizes decrease with year of publication (Munafo et al., 2010). For example, when selective serotonin reuptake inhibitors (SSRI) were first tested for the treatment of depression, effect sizes of more than 1 as compared to placebo were reported reviewed by (Taylor et al., 2011), which temporarily created its legacy as wonder drug. Over the course of 20 years, the mean weighted effect size of SSRI for depression has gradually decreased to values around 0.2-0.3.

This trend is most likely the result of publication bias, occurring both at the level of the researchers and at the level of the journals (Siontis et al., 2011). A remarkably high effect size is spectacular news and could suggest the discovery of a new wonder treatment. Studies with such findings are therefore easily published in high impact journals (Siontis et al., 2011). By contrast, studies of similar sample sizes with marginal or non-significant findings indicate that the new treatment is not a wonder drug, and provide less spectacular news that may not easily be accepted for publication. Researchers are aware of this and as a result refrain from submitting such findings for publication. Usually, after some years, studies with larger sample sizes become available. When these large studies also have negative findings, the chance for a type II error is smaller and both researchers and journals feel that these findings should be published too. This is when meta-analyses start to notice a decrease in efficacy.

In this view, the study by Brunelin et al. is exemplary of a first placebo-controlled study reporting on a new treatment technique. The study includes a small sample size and finds very large effect sizes. It is published in a high impact journal. It is possible that other groups have undertaken similar small studies which have not yielded such large effect sizes, yet these studies may not have been submitted or not accepted for publication (i.e. file drawer effect). Like Brunelin and colleagues we are very curious to see the first large RCTs in which tDCS treatment is evaluated. We sincerely hope that tDCS is the exception to the rule, as a highly effective, side effect-free, cheap and safe method to treat both refractory hallucinations and negative symptoms certainly is most welcome. However, given the previous observations for other new methods, it could well be the case that ten years from now the mean weighted effect size of tDCS is somewhere between 0.2 and 0.3. Hopefully, uncovering the neural working mechanisms underlying the improvement of refractory hallucinations and negative symptoms with tDCS treatment will prevent this from occurring.

References:

Bech P, Cialdella P, Haugh MC, Birkett MA, Hours A, Boissel JP, Tollefson GD. Meta-analysis of randomised controlled trials of fluoxetine v. placebo and tricyclic antidepressants in the short-term treatment of major depression. Br J Psychiatry. 2000 May;176:421-8

Brunelin J, Poulet E, Bediou B, Kallel L, Dalery J, D'amato T, Saoud M. Low frequency repetitive transcranial magnetic stimulation improves source monitoring deficit in hallucinating patients with schizophrenia. Schizophr Res. 2006 Jan 1;81(1):41-5.

Emerson GB, Warme WJ, Wolf FM, et al. Testing for the presence of positive-outcome bias in peer review: a randomized controlled trial. Arch Intern Med 2010;170: 1934-1939 21098355

Hoffman RE, Hawkins KA, Gueorguieva R, Boutros NN, Rachid F, Carroll K, Krystal JH. Transcranial magnetic stimulation of left temporoparietal cortex and medication-resistant auditory hallucinations. Arch Gen Psychiatry. 2003 Jan;60(1):49-56. 12511172

Jacobson L, Ezra A, Berger U, Lavidor M. Modulating oscillatory brain activity correlates of behavioral inhibition using transcranial direct current stimulation. Clin Neurophysiol. 2012 May;123(5):979-84. 21995999 Kabakov AY, Muller PA, Pascual-Leone A, Jensen FE, Rotenberg A. Contribution of axonal orientation to pathway-dependent modulation of excitatory transmission by direct current stimulation in isolated rat hippocampus. J Neurophysiol. 2012 Apr;107(7):1881-9. 22219028

Munafo MR, Flint J. How reliable are scientific studies? Br J Psychiatry 2010;197: 257-258

Leucht S, Hierl S, Kissling W, Dold M, Davis JM. Putting the efficacy of psychiatric and general medicine medication into perspective: review of meta-analyses. Br J Psychiatry. 2012 Feb;200(2):97-106.

Nitsche MA, Paulus W. Excitability changes induced in the human motor cortex by weak transcranial direct current stimulation. J Physiol 2000;527:633-639.

Siontis KC, Evangelou E, Ioannidis JP. Magnitude of effects in clinical trials published in high-impact general medical journals. Int J Epidemiol. 2011 Oct;40(5):1280-91.

Slotema C.W, Aleman A., Daskalakis Z.J, Sommer I.E Meta-analysis of repetitive transcranial magnetic stimulation targeted at the left temporoparietal area in the treatment of auditory verbal hallucinations: do recent studies report smaller effect sizes? in revision at Schiz Res

Taylor D, Meader N, Bird V, Pilling S, Creed F, Goldberg D; pharmacology subgroup of the National Institute for Health and Clinical Excellence Guideline Development Group for Depression in Chronic Physical Health Problems. Pharmacological interventions for people with depression and chronic physical health problems: systematic review and meta-analyses of safety and efficacy. Br J Psychiatry. 2011 Mar;198(3):179-88.

View all comments by Iris Sommer