Psychiatry Research 273 (2019) 343–349

Contents lists available at ScienceDirect

Psychiatry Research
journal homepage: www.elsevier.com/locate/psychres

Effects and potential mechanisms of transcranial direct current stimulation T

(tDCS) on auditory hallucinations: A meta-analysis
⁎ ⁎⁎
Fuyin Yanga,b, Xinyu Fangc, Wei Tangd, Li Huie, Yan Chenc, Chen Zhangc, , Xing Tiana,b,f,
a
Shanghai Key Laboratory of Brain Functional Genomics (Ministry of Education), School of Psychology and Cognitive Science, East China Normal University, Shanghai
200062, China
b
NYU-ECNU Institute of Brain and Cognitive Science at NYU Shanghai, 3663 Zhongshan Road North, Shanghai, 200062, China
c
Schizophrenia Program, Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine, Shanghai 200232, China
d
Wenzhou Kangning Hospital, Wenzhou Medical University, Wenzhou, Zhejiang, China
e
The Affiliated Guangji Hospital of Soochow University, Suzhou, China
f
Division of Arts and Sciences, New York University Shanghai, 1555 Century Avenue, Shanghai, China, 200122

A R T I C LE I N FO A B S T R A C T

Keywords: Auditory hallucinations are the most common psychiatric symptoms of schizophrenia with high recurrence and
Transcranial direct current stimulation refractoriness. Transcranial direct current stimulation (tDCS), a novel, non-invasion and affordable brain sti-
Auditory hallucinations mulation technique, has been recently applying on the schizophrenia patients to treat the auditory hallucina-
Schizophrenia tions. To analyze the efficacy of tDCS treatment on such symptoms and to reveal its potential working me-
Meta-analysis
chanisms, we carried out a structured literature search in PubMed, Embase and Cochrane Library database up to
May 12, 2018. Five studies that met inclusion criteria with a total of 137 patients were included in this meta-
analysis. After pooling all the data, we found that there was no significant effect between active group and sham
group of tDCS (p = 0.18). When we removed one study that did not collaboratively stimulate the frontal-tem-
poral sites, the active tDCS group marks a significant improvement of therapeutic effect compared with sham
group (p = 0.007). Our findings suggested that tDCS could be a promising tool to alleviate auditory halluci-
nations, provided that the simulation sites and protocols are targeting at the sensorimotor frontal-parietal net-
work.

1. Introduction persistent cognitive deficits (Falloon and Talbot, 1981; Upthegrove

et al., 2016). Recent advance in neuroimaging research have showed
Schizophrenia is a mental illness of unknown etiology and mainly that AVH in schizophrenia is associated with abnormal activity in
characterized by positive and negative symptoms as well as cognitive frontal and temporo-parietal areas (Allen et al., 2012; Hoffman et al.,
impairment (Insel, 2010; Sims, 1988). Positive symptoms are divided 2000; Jardri et al., 2011). Increased activation in such regions may be
into delusions, disordered thoughts and speech, and auditory halluci- correlated with the deficits of self-monitoring functions (Allen et al.,
nations (AH) (Liddle, 1987). Auditory verbal hallucinations (AVH) are 2007; Frith and Corcoran, 2009; McGuire, 1995; Tian and Poeppel,
the most pronounced symptom in schizophrenia patients with auditory 2012; Waters, 2012), corollary discharge and motor-to-sensory trans-
hallucinations (Sommer et al., 2012). formation functions (Ford and Mathalon, 2005; Tian and Poeppel,
AVH refer to the symptoms of hearing voices in the absence of the 2012) and sensory gating functions (Bak et al., 2014). The abnormal of
external stimulus (Stephane et al., 2001). It is reported that 50%−70% these functions may be the causes of AVH.
of patients with AVH during the treatment phrase (Meltzer, 1992; Among numerous studies that apply neuromodulation on these
Waters, 2012). Accumulating studies have demonstrated that halluci- frontal and temporo-parietal target regions for interventions of AVH
nations are resistant to antipsychotic treatment in 25%−30% adult (David, 2004), non-invasive neurostimulation techniques are thought to
schizophrenia patients, which resulted in functional disability and be practically useful to alleviate treatment-resistant of auditory

Abbreviations: AH, auditory hallucination; AVH, auditory verbal hallucination; tDCS, transcranial direct current stimulation; RCT, Randomized controlled trial;
AHRS, Auditory Hallucination Rating Scale; TPJ, temporo-parietal junction
⁎
Corresponding author at: Schizophrenia Program, Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine, Shanghai 200232, China.
⁎⁎
Corresponding author at: New York University Shanghai, 1555 Century Avenue, Shanghai, China, 200122
E-mail addresses: zhangchen645@gmail.com (C. Zhang), xing.tian@nyu.edu (X. Tian).

https://doi.org/10.1016/j.psychres.2019.01.059
Received 19 October 2018; Received in revised form 14 January 2019; Accepted 14 January 2019
Available online 15 January 2019
0165-1781/ © 2019 Elsevier B.V. All rights reserved.
F. Yang et al. Psychiatry Research 273 (2019) 343–349

hallucinations in patients with schizophrenia (Brunelin, 2013; Hoffman 2.2. Inclusion and exclusion criteria
et al., 2000). Transcranial Direct Current Stimulation (tDCS) is one of
such non-invasive techniques that are recently introduced to treat AVH. The inclusion criteria for studies were: (a) studies using tDCS in the
In the clinical setting, tDCS could have some advantages over rTMS intervention group, (b) assessment of the severity of auditory halluci-
such as ease of use, lower cost, and well-tolerated (Brunoni et al., 2011; nations in schizophrenia or other type of schizophrenia spectrum dis-
Priori et al., 2009). tDCS applies weak and constant electric current on orders, (c) diagnosis was established using Diagnostic and Statistical
the scalp (Nitsche and Paulus, 2000), which results in weak electric Manual of Mental Disorders (DSM) or International Statistical
field that alters neural activity and modulate cortical connectivity Classification of Diseases and Related Health Problems (ICD), (d)
(Keeser et al., 2011; Lee et al., 2017). The anodal stimulation is positive methodological design based on randomized controlled trials (RCTs),
stimulation that increases the neuronal excitability of the area being (e) data were presented as mean and standard deviation for auditory
stimulated. Cathodal stimulation decreases the neuronal excitability of hallucinations levels.
the area being stimulated (Nitsche et al., 2003). One of the major The exclusion criteria were: (a) studies without a control group, (b)
function advantages of tDCS is that the polarity of current flow can patients with auditory hallucinations presenting neuropsychiatric co-
selectively manipulate the neural activity of excitatory or inhibitory morbidities, (c) studies that use cortical stimulation techniques other
status (Keeser et al., 2011). The long-term effects of tDCS seem to than tDCS.
change the efficacy of GABAergic activity, NMDA receptors, and mod-
ulation the long-term potentiation and depression (Agarwal et al., 2013; 2.3. Data extraction and quality assessment
Koops et al., 2015; Nitsche et al., 2008).
Such tDCS induced modulation can be effective on alleviate AVH. The following data was extracted from identified studies using
For instance, a recent clinical observation showed a significant reduc- standard and comprehensive forms: the name of the first author, year of
tion of auditory hallucinations in schizophrenia after fronto-temporal publication, demographic information of participants, including gender
tDCS with the anode placed over the left dorsolateral prefrontal cortex and average age, sample size and mean and SD, diagnostic criteria of
(DLPFC) and the cathode placed over the left temporo-parietal junction schizophrenia spectrum disorders, type of intervention, outcome mea-
(TPJ) cortex (Brunelin et al., 2012). This phenomenon might be at- sures.
tributed to the fact that placing the cathodic electrode over the Wer- Study quality was evaluated for risk of bias using the Cochrane
nicke region can reduce cerebral blood flow and decrease functional Handbook for Systematic Reviews of Interventions (Higgins, 2011). The
connectivity between left TPJ and inner speech production brain areas tool distinguished studies based on six-item scale: sequence generation,
(Mondino et al., 2016). A recent study showed that tDCS applying on random allocation to groups, blinding, missing data, selective reporting
frontal and temporal-parietal regions modulates the neural signal and other biases. We identified the level of biases into high, medium
transformation from motor to sensory regions (corollary discharge and low based on the Grading of Recommendations Assessment, De-
function) in patients with AVH (Nawani et al., 2014). velopment and Evaluation (GRADE) criteria in each of studies
Increasing number of studies were conducted to explore the effects (Guyatt et al., 2008). Two authors independently assessed the study
of tDCS for auditory hallucinations (Fitzgerald et al., 2014; Frohlich quality.
et al., 2016; Mondino et al., 2016; Smith et al., 2015). However, in-
consistent treatment results were obtained. It is urgent to examine the 2.4. Statistical analysis
validity, as well as to provide insights on the potential neural me-
chanisms of this new type of treatment. In such a case, meta-analysis is The random-effects model was chosen due to more conservative
a superior option as it allows easily reaching a broad set of subjects. than the fixed-effects model, as well as lower type Ⅰ error and wider
Meanwhile, by collaboratively controlling experimental variables, such confidence interval (CI) (Ades et al., 2005; Fang et al., 2018). To pool
as sample size, symptom measurement, and stimulation protocols, different scales, we used the standardized mean difference (SMD) as the
meta-analysis can enhance statistical power, summarize the common summary statistic in this meta-analysis (Vesterinen et al., 2014), as it
and effective practice and provide guidance for future research of tDCS reveals the effect size of the treatment relative to the variability ob-
on AVH. Therefore, this meta-analysis pooled published literature to served in the same study. Heterogeneity of studies was assessed. If p <
validate the efficacy of tDCS for auditory hallucinations. 0.1 and I2 > 50%, the heterogeneity of studies was significant, other-
wise it was not significant. Finally, to assess the stability of the result, a
sensitivity analysis was performed. All the analyses were done with
2. Methods Review Manager 5.3 software.

2.1. Search strategy 3. Result

We carried out an electronically literature search from PubMed, 3.1. Result of the search
EMBASE, and the Cochrane Library published before May 12, 2018. The
search terms were (“Auditory Hallucination” or “Auditory From the electronic search, a total of 85 publications were identi-
Hallucinations” or “Verbal Auditory Hallucinations” or “Phonism” or fied, from which we excluded 28 due to repetition. The titles and ab-
“Voice”) and (“Transcranial Direct Current Stimulation” or “tDCS” or stracts of 57 articles were read, 33 were excluded because they were not
“Transcranial Electrical Stimulation”). All results were limited to consistent with the topic of this meta-analysis. After reading the full
human studies published in English. Reference lists of the included text of the remaining 24 articles, another 18 studies were further ex-
articles were manually scanned to identify further relevant studies. All cluded because they did not meet the inclusion criteria. As a result, only
identified publications were imported into the reference manager 6 studies fulfilled the eligibility criteria and were selected. One study
EndNote X6 and duplicated records were removed. Then articles were was excluded in that the demographic data for each group was not
then manually scanned to select for the meta-analysis. Two authors provided (Fitzgerald et al., 2014). Ultimately, 5 articles were included
further independently assess the relevance of the articles to the topic in this meta-analysis. The selection process was summarized in Fig. 1.
based on their titles and abstracts and full texts (see Inclusion and
Exclusion Criteria for details). Any disagreements were resolved by 3.2. Study characteristics
discussion and consensus with a third person.
All included studies conducted a randomized, double-blind, sham-

344
F. Yang et al. Psychiatry Research 273 (2019) 343–349

Fig. 1. Flow diagram of included/excluded studies.

Table 1
Clinical parameters of each study included on the meta-analysis.
Study Groups Sample size Age (mean) Diagnosis Diagnostic criteria Outcome measures

Bose (2018) tDCS group 12 31.25 ± 8.32 Schizophrenia DSM-IV AHRS

Control group 13 31.38 ± 7.56
Brunelin (2012) tDCS group 15 40.4 ± 9.9 Schizophrenia DSM-IV AHRS
Control group 15 35.1 ± 7.0
Frohlich (2016) tDCS group 13 43.4 ± 12.6 Schizophrenia and schizoaffective disorder DSM-IV AHRS
Control group 13 40.0 ± 10.7
Mondino (2016) tDCS group 11 36.7 ± 9.7 Schizophrenia DSM-IV AHRS
Control group 12 37.3 ± 9.7
Smith (2015) tDCS group 17 46.8 ± 11.1 Schizophrenia and schizoaffective disorder DSM-IV PANSS
Control group 16 44.9 ± 9.2

DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition;

PANSS = Position and Negative Syndrome Scale; AHRS = Auditory Hallucination Rating Scale.

345
F. Yang et al. Psychiatry Research 273 (2019) 343–349

Table 2
tDCS parameters of each study included on the meta-analysis.
Study tDCS device Electrodes location Sessions (n/day) Current (mA) Time (min)

Bose (2018) Neuroconn eldith A: F3 FP1 10(2*/day) 2 20

DC Stimulator Plus C: T3 P3
Brunelin (2012) Neuroconn eldith A: F3 FP1 10(2*/day) 2 20
DC Stimulator Plus C: T3 P3
Frohlich (2016) Neuroconn eldith A: F3 FP1 5(1*/day) 2 20
DC Stimulator Plus C: T3 P3
Return electrode: Cz
Mondino (2016) Neuroconn eldith A: F3 FP1 10(2*/day) 2 20
DC Stimulator Plus C: T3 P3
Smith (2015) Chattanooga Ionto A: F3; 5(1*/day) 2 20
System stimulator C: FP2

F3 FP1 = left dorsolateral prefrontal cortex (DLPFC).

T3 P3 = left temporoparietal junction (TPJ).
Cz = posterior midline.

controlled tDCS trials. Five studies in total involved 137 DSM-IV schi- mean AHRS scores reduced from 28.3 (4.1) to 19.9 (5.8) in the tDCS
zophrenia patients with auditory hallucinations from two groups: 68 in group, significantly larger than that in the sham group [from 27.2 (6.9)
the active (tDCS treatment) group, 69 in the sham (control) group, to 25.1 (7.7)] [(d = 1.58, p < 0.001)]. Interestingly, the tDCS effect
respectively. In the active group, one study placed the anode over the could last beyond the immediate treatment – AHRS score was reduced
LDLPFC (F3) and the cathode over the contralateral supraorbital ridge 36% and 38% for the active tDCS group at 1 month and 3 months post-
(Fp2) (Smith et al., 2015). Another four studies (Bose et al., 2018; tests, respectively. Whereas in the sham group, the score was reduced
Brunelin et al., 2012; Frohlich et al., 2016; Mondino et al., 2016) both 3% at 1 month and 5% at 3 months. These results suggest that tDCS
placed the anode over the F3/FP1 (left dorsolateral prefrontal cortex) treatment effect could be long-lasting.
and the cathode over the T3/P3 (left temporo-parietal junction), but
one study (Frohlich et al., 2016) added a third electrode: a return 3.4. Meta-Analyses
electrode over Cz (posterior midline). The treatments sites were all
focused on the left hemisphere. In all five studies, active tDCS (2 mA for In the present paper, four studies reported the AHRS score and one
20 min) was performed once or twice daily, for 5 to 10 sessions. In the study (Smith et al., 2015) reported the P3 score in PANSS as the out-
sham group, the stimulation was set at 2 mA lasting only 40 s, though come measure of AVH. We pooled the whole data and found no sig-
the electrodes remained in place for 20 min. The clinical and stimula- nificant difference between active tDCS group and sham group
tion parameters performed in each study are shown in Tables 1 and 2, (Z = 1.33, p = 0.18, Fig. 2A). While there was an obvious hetero-
respectively. geneity in effect size estimates (p < 0.01, I2 = 70%, Fig. 2A). After
In Smith et al. (2015), Frohlich et al. (2016), Mondino et al. (2016) performing a sensitivity analysis by removing one study (Smith et al.,
and Brunelin et al. (2012), the AVH symptom severity (AHRS score, or 2015) that did not perform the stimulation on frontal-parietal network,
PANSS score) was not statistically significant different at baseline be- we find a more homogeneous result (p = 0.19, I2 = 37%, Fig. 2B),
tween active tDCS and sham groups. In Bose et al. (2018), no significant yielding a significant statistical difference (p = 0.007).
difference in AVH (AHRS score) was observed at baseline. However,
differences were found between two groups in other positive and ne- 3.5. Risk of biases
gative symptoms. Patients in sham tDCS group had statistically sig-
nificant greater SAPS and SANS scores at baseline. These baseline dif- The risks of biases are reported in Table 3. Random allocation to an
ferences were controlled for their potential confounding effects in active tDCS group and blinded assessment of outcome were described in
further analyses. all the five studies. Brunelin et al. (2012) and Frohlich et al. (2016) did
In Brunelin et al. (2012), symptom assessments were conducted not describe randomization clearly, and Brunelin et al. (2012) did not
immediately before the first tDCS session, and 5 days, 1 and 3 months report the concealment of allocation clearly, therefore the risk for these
after tDCS. In Mondino et al. (2016) and Bose et al. (2018), the severity items was unclear. Based on the GRADE criteria, the results of the
of symptom was assessed at baseline (before application of the first studies were inconsistent. And only 5 studies included in the meta-
stimulation) and re-assessed immediately after 5 days of twice-daily (10 analysis, therefore the publication bias is also unclear.
sessions) tDCS administration. In Frohlich et al. (2016), AVH were as-
sessed before the first tDCS session and immediately after the final tDCS
4. Discussion
session, as well as a follow-up test 30–45 days after completion of sti-
mulation sessions. In Smith et al. (2015), psychiatric symptoms were
The present meta-analysis including five randomized clinical trials
assessed with PANSS scale at baseline and after the 5th tDCS session.
aimed to assess the treatment effects of tDCS on auditory hallucinations
in schizophrenia. Our results with all five studies did not show sig-
3.3. Primary outcomes nificant difference between active group and sham group of tDCS.
However, when only including four studies that all induced stimulation
The primary outcome was change in the severity of the AVH mea- currents in frontal-parietal network with stimulation pairing between
sured by the Auditory Hallucination Rating Scale (AHRS) (Brunelin F3/FP1 and T3/P3 sites, a significant effect was seen on the active tDCS
et al., 2012; Frohlich et al., 2016; Mondino et al., 2016) and P3 of group compared to the sham group. Therefore, considering the limited
PANSS (Smith et al., 2015). available data, tDCS seems to be a promising tool to alleviate auditory
Three studies demonstrated a therapeutic benefit of tDCS as redu- hallucinations. However, more high-quality RCT studies are required to
cing the severity of AVH (Brunelin et al., 2012; Mondino et al., 2016; further validate the efficacy and working mechanisms of tDCS for au-
Bose et al., 2018). Out of these three studies, one study (Brunelin et al., ditory hallucinations in schizophrenia.
2012) provided a longitudinal dataset. Immediately after treatment, the The major finding of this meta-analysis is that the simulation sites

346
F. Yang et al. Psychiatry Research 273 (2019) 343–349

Fig. 2. (A) Result of Meta-analysis of tDCS for Auditory Hallucinations. (B) Result of sensitivity analysis of tDCS for Auditory Hallucinations.

are crucial for obtaining positive treatment effects of tDCS on alle- toward implementing “personalized neuromodulation” (Datta et al.,
viating AVH in schizophrenia. The consistent and effective simulation 2012; Diederen et al., 2013).
sites are pairing between F3/FP1 and T3/P3 sites. Neurophysiological This meta-analysis provides a valuable way to estimate and infer the
studies have shown the importance of the left hemisphere for halluci- potentials of tDCS treatment method. In the included literatures, only
nations in Schizophrenia, for example the hyperactivity of left TPJ areas three studies proved a significant therapeutic effect of tDCS on the AVH
play a crucial role in the presence of positive symptoms (Shergill et al., in schizophrenia patients. In Frohlich et al. (2016) and
2000; Silbersweig et al., 1995). This is consistent with the significant Smith et al. (2015) studies, they did not find a significant difference
AVH symptom reduction when applied the tDCS stimulation over the between active tDCS and sham tDCS. The inconsistent outcomes are not
left hemisphere in the included studies. This stimulation protocol en- enough to draw any conclusions about the treatment effects. However,
ables a potential current flow in the sensorimotor network via the when we pooled data from these studies and considered various con-
connection between frontal and parietal/temporal regions straints based on neuroscientific theories, the meta-analysis results are
(Psomiades et al., 2017). Such current stimulation may affect and im- consistent and reliable.
prove monitoring functions that have been hypothesized as a possible The meta-analysis also provides an opportunity to compare and
neural mechanism for auditory hallucination (Tian and Poeppel, 2012). summarize several factors that could potential influence the tDCS
Interestingly, previous studies using neuroimaging and scalp elec- treatment effects. For example, the treatment effect may limit to par-
trophysiology methods have demonstrated that malfunction in frontal- ticular homogeneous patient groups. Trials that presented the positive
parietal network may mediate the deficits in source-monitoring, dys- effects of tDCS included only schizophrenic patient individuals,
function of corollary discharge, and inefficient gating of sensory in- whereas other studies that show the negative effects of tDCS also in-
formation in patients with AVH (Bak et al., 2014; Ford and Mathalon, cluded schizoaffective disorder patients in their samples. The incon-
2005; Wang et al., 2011). These findings are consistent with our results, sistent outcomes, therefore could be because of specific deficits in dis-
supporting the linking between motor and sensory systems may be the tinct patient groups. This hints that types of patients should be a factor
neurobiological basis for the ameliorative effects of tDCS on AVH. to consider before applying tDCS treatment.
Further research in combination of clinical and neuroimaging methods The amount of treatment could be another important factor. The
is needed to explore the differential contribution from the stimulation total number of daily tDCS sessions presented in the studies of
of brain regions in the frontal-parietal network. With the development Frohlich et al. (2016) and Smith et al. (2015) was less than that in other
of artificial neural network (ANN), the optimized MRI guided stimu- studies (Bose et al., 2018; Brunelin et al., 2012; Mondino et al., 2016).
lation combine with neural computational modeling are practicable In the study of Frohlich et al. (2016) that also applied stimulation over

Table 3
Evaluation of risk of bias in the 5 included studies based on the seven items in the Cochrane Risk of Bias tool.
Study Random sequence Allocation Blinding of participants Blinding of outcome Incomplete outcome Selective Other
generation concealment and providers assessment data reporting biases

Bose (2018) Low Low Low Low Low Low Low

Brunelin (2012) unclear unclear Low Low Low Low Low
Frohlich (2016) Low Low Low Low Low Low Low
Mondino (2016) unclear unclear Low Low Low Low Low
Smith (2015) Low Low Low Low Low Low Low

347
F. Yang et al. Psychiatry Research 273 (2019) 343–349

frontal-parietal network, the effect of active tDCS on AVH did not reach study variability in the tDCS montages. Computational modeling about
significance. The less amount of treatment could be an important factor the spatial distribution of electric fields induced by tDCS would be a
that contributed to the null effect. Compared with other studies that fruitful way forward in linking treatment amount and efficacy. Future
showed significant symptom reduction, this study had only 5 sessions of clinical trials should specify the stimulation parameters (e.g., the
stimulation, significantly less than 20 sessions in other studies. There- electrode/coil configuration, current amplitude, pulse width, fre-
fore, stimulation sessions that were conducted only once a day for 5 quency, number of pulse) and should combine neuroimaging data to
days could be insufficient to obtain an ideal therapeutic effect. Our implement MRI guided stimulation along with computational modeling
results suggest that the effectiveness of current protocols is constrained for optimize tDCS protocols. If effective, tDCS might be compliment to
by ‘dose’. The effect of tDCS could be improved by increasing the or even replace the antipsychotic medication and would be feasible
amount of treatment. Studies with rTMS have shown that the effect of toward ‘personalized neuromodulation’.
stimulation on the motor cortex needs to be maintained for a period of
time, which are capable of exceeding the cycle of the motor cortex 5. Conclusion
(Carpenter et al., 2012). Therefore, the more ideal tDCS protocol about
the amount of stimulation needs to be established for conducting fur- This meta-analysis revealed significant therapeutic effects of tDCS
ther studies in this area. on reducing severity and frequency of AVH in schizophrenia. These
Symptom assessment tools could also influence on the evaluation of significant effects were obtained by limiting the simulation sites on
tDCS treatment. The finally included four studies in this meta-analysis frontal-parietal regions. These results suggest that tDCS is a promising
measured the change of auditory hallucinations using AHRS, while the tool to alleviate AVH, and the possible neural mechanisms of mon-
removed study (Smith et al., 2015) used P3 in the PANSS scale. The itoring functions in sensorimotor frontal-parietal network may con-
previous study pointed out that the item of hallucinations in the PANSS strain the stimulation protocols as one of the crucial factors to obtain
scale is unreliable to measure the severity of auditory hallucinations in the treatment effects.
a detailed manner (Santor et al., 2007). As the AHRS scale presents an
excellent capacity for evaluating the hallucinatory symptoms, it is more Role of funding source
suitable scale to evaluate the hallucinations refractory to treatment in
population samples containing schizophrenia (Haddock et al., 1999). This study is supported by the National Natural Science Foundation
Therefore, consistent and reliable symptom assessment tools should be of China (31871131), the Shanghai Science and Technology
implemented. Commission Foundation (17JC1404104), Program of Introducing
Compared with meta-analysis studies that aim for a summary of Talents of Discipline to Universities, Base B16018, a grant from the New
findings in a relatively established research topic, some reviews focus York University Global Seed Grants for Collaborative Research (85-
on exploring potential questions and mechanisms in a new direction of 65701-G0757-R4551), and the JRI Seed Grants for Research
research with a limited set of available studies (Hirsch et al., 2018; Collaboration from NYU-ECNU Institute of Brain and Cognitive Science
Vacas et al., 2018; Lan et al., 2017; Bala et al., 2013). The goal of this at NYU Shanghai. The funding sources of this study had no role in study
study is consistent with the latter group and aims to explore the relative design, data analysis, decision to publish, or preparation of the article.
new direction of tDCS in treatment for AVH and its potential underlying
mechanisms. Therefore, we included four studies that were available at Role of contributors
the time. Moreover, because of the goal of this study – exploring a
potential treatment effect of a method and its underlying mechanism, F.Y X.T and C.Z designed the study. F.Y and X.F conducted the lit-
rather than summarizing common findings among studies (e.g. shared erature searches and analysis. F.Y and C.Y undertook the statistical
cortical regions among fMRI studies), we pooled data from multiple analysis. W.T and L.H managed the assessment of risk of Bias and
studies based on their common features and manipulations. The pooling GRADE. F.Y wrote the draft manuscript. F.Y, X.F, C.Y, L.H, C. Z, and X.T
yielded a sample size of more than 40 in our study, which is comparable contributed to the final manuscript.
to most of empirical studies. Our preliminary results suggest that tDCS
may be a good treatment for AVH and the potential underlying me- Conflict of interest
chanism could be the motor-to-sensory transformation in self-mon-
itoring and control. More high-quality RCT studies are required to All authors report that they have no conflicts of interest.
further validate the efficacy and the proposed working mechanism of
tDCS for auditory hallucinations in schizophrenia.” References

4.1. Limitations Ades, A.E., Lu, G., Higgins, J.P., 2005. The interpretation of random-effects meta-analysis
in decision models. Med. Decis. Making 25, 646–654.
Agarwal, S.M., Shivakumar, V., Bose, A., Subramaniam, A., Nawani, H., Chhabra, H.,
Several limitations of our meta-analysis should be mentioned here. Kalmady, S.V., Narayanaswamy, J.C., Venkatasubramanian, G., 2013. Transcranial
First, the sample size of this studies included in this meta-analysis was direct current stimulation in schizophrenia. Clin. Psychopharmacol. Neurosci. 11,
small and statistical power could therefore be limited to identifying 118–125.
Allen, P., Aleman, A., McGuire, P.K., 2007. Inner speech models of auditory verbal hal-
changes with large effect sizes. Moreover, patients were on continuous lucinations: evidence from behavioural and neuroimaging studies. Int. Rev.
therapy during the trials still in combination with other psychotropic Psychiatry 19, 407–415.
drugs, the potential confounding effects of different medications may Allen, P., Modinos, G., Hubl, D., Shields, G., Cachia, A., Jardri, R., Thomas, P., Woodward,
T., Shotbolt, P., Plaze, M., Hoffman, R., 2012. Neuroimaging auditory hallucinations
confuse the effect of tDCS. Furthermore, the state of illness and loca-
in schizophrenia: from neuroanatomy to neurochemistry and beyond. Schizophr.
lization of the electrodes might affect the result. Therefore, a better way Bull. 38, 695–703.
to determine the parameters of the stimulus and to evaluate the clinical Bak, N., Rostrup, E., Larsson, H.B., Glenthoj, B.Y., Oranje, B., 2014. Concurrent functional
magnetic resonance imaging and electroencephalography assessment of sensory
outcomes is needed to determine tDCS as a therapeutic possibility and
gating in schizophrenia. Hum. Brain Mapp. 35, 3578–3587.
its potential working mechanisms for auditory hallucinations in schi- Bala, M.M., Riemsma, R.P., Wolff, R., Kleijnen, J., 2013. Microwave coagulation for liver
zophrenia. Last, the amount of tDCS sessions could be another im- metastases. Cochrane Database Syst. Rev. 10, CD010163.
portant factor. Previous studies shown that reduction in AVH may rely Bose, A., Shivakumar, V., Agarwal, S.M., Kalmady, S.V., Shenoy, S., Sreeraj, V.S.,
Narayanaswamy, J.C., Venkatasubramanian, G., 2018. Efficacy of fronto-temporal
on “dose-dependent” hyperpolarization in auditory or speech-related transcranial direct current stimulation for refractory auditory verbal hallucinations in
regions (Krishnan et al., 2011; Lesh et al., 2010). However, it would be schizophrenia: a randomized, double-blind, sham-controlled study. Schizophr. Res.
hard, if not impossible to examine this assumption because of the inter- 195, 475–480.

348
F. Yang et al. Psychiatry Research 273 (2019) 343–349

Brunelin, J., 2013. TDCS: therapeutic applications in schizophrenia. Biol. Psychiatry 73, hallucinations in patients with schizophrenia: A systematic review and meta-analysis
48S. of randomized controlled trials. Schizophr. Bull. 43, S248–S249.
Brunelin, J., Mondino, M., Gassab, L., Haesebaert, F., Gaha, L., Suaud-Chagny, M.F., Lesh, T.A., Niendam, T.A., Minzenberg, M.J., Carter, C.S., 2010. Cognitive control deficits
Saoud, M., Mechri, A., Poulet, E., 2012. Examining transcranial direct-current sti- in schizophrenia: mechanisms and meaning. Neuropsychopharmacology 36, 316.
mulation (tDCS) as a treatment for hallucinations in schizophrenia. Am. J. Psychiatry Liddle, P.F., 1987. The symptoms of chronic schizophrenia. A re-examination of the po-
169, 719–724. sitive-negative dichotomy. Br. J. Psychiatry 151, 145–151.
Brunoni, A.R., Amadera, J., Berbel, B., Volz, M.S., Rizzerio, B.G., Fregni, F., 2011. A McGuire, P.K., 1995. Abnormal monitoring of inner speech: a physiological basis for
systematic review on reporting and assessment of adverse effects associated with auditory hallucinations. Lancet. 346, 596–600.
transcranial direct current stimulation. Int. J. Neuropsychopharmacol. 14, Meltzer, H.Y., 1992. Treatment of the neuroleptic-nonresponsive schizophrenic patient.
1133–1145. Schizophr. Bull. 18, 515–542.
Carpenter, L.L., Janicak, P.G., Aaronson, S.T., Boyadjis, T., Brock, D.G., Cook, I.A., Mondino, M., Jardri, R., Suaud-Chagny, M.F., Saoud, M., Poulet, E., Brunelin, J., 2016.
Dunner, D.L., Lanocha, K., Solvason, H.B., Demitrack, M.A., 2012. Transcranial Effects of fronto-temporal transcranial direct current stimulation on auditory verbal
magnetic stimulation (TMS) for major depression: a multisite, naturalistic, observa- hallucinations and resting-state functional connectivity of the left temporo-parietal
tional study of acute treatment outcomes in clinical practice. Depress Anxiety 29, junction in patients with schizophrenia. Schizophr. Bull. 42, 318–326.
587–596. Nawani, H., Bose, A., Agarwal, S.M., Shivakumar, V., Chhabra, H., Subramaniam, A.,
Datta, A., Truong, D., Minhas, P., Parra, L.C., Bikson, M., 2012. Inter-Individual variation Kalmady, S., Narayanaswamy, J.C., Venkatasubramanian, G., 2014. Modulation of
during transcranial direct current stimulation and normalization of dose using MRI- corollary discharge dysfunction in schizophrenia by tDCS: preliminary evidence.
derived computational models. Front Psychiatry 3, 91. Brain Stimul. 7, 486–488.
David, A.S., 2004. The cognitive neuropsychiatry of auditory verbal hallucinations: an Nitsche, M.A., Cohen, L.G., Wassermann, E.M., Priori, A., Lang, N., Antal, A., Paulus, W.,
overview. Cognit. Neuropsychiatry 9, 107–123. Hummel, F., Boggio, P.S., Fregni, F., Pascual-Leone, A., 2008. Transcranial direct
Diederen, K.M., Charbonnier, L., Neggers, S.F., van Lutterveld, R., Daalman, K., Slotema, current stimulation: state of the art 2008. Brain Stimul. 1, 206–223.
C.W., Kahn, R.S., Sommer, I.E., 2013. Reproducibility of brain activation during Nitsche, M.A., Nitsche, M.S., Klein, C.C., Tergau, F., Rothwell, J.C., Paulus, W., 2003.
auditory verbal hallucinations. Schizophr. Res. 146, 320–325. Level of action of cathodal DC polarisation induced inhibition of the human motor
Falloon, I.R., Talbot, R.E., 1981. Persistent auditory hallucinations: coping mechanisms cortex. Clin. Neurophysiol. 114, 600–604.
and implications for management. Psychol. Med. 11, 329–339. Nitsche, M.A., Paulus, W., 2000. Excitability changes induced in the human motor cortex
Fang, X., Zhang, Y., Fan, W., Tang, W., Zhang, C., 2018. Interleukin-17 alteration in first- by weak transcranial direct current stimulation. J. Physiol. 527, 633–639.
episode psychosis: a meta-analysis. Mol. Neuropsychiatry 3, 135–140. Priori, A., Hallett, M., Rothwell, J.C., 2009. Repetitive transcranial magnetic stimulation
Fitzgerald, P.B., McQueen, S., Daskalakis, Z.J., Hoy, K.E., 2014. A negative pilot study of or transcranial direct current stimulation? Brain Stimul. 2, 241–245.
daily bimodal transcranial direct current stimulation in schizophrenia. Brain Stimul. Psomiades, M., Mondino, M., Poulet, E., Haesebaert, F., Suaud-Chagny, M.F., Brunelin, J.,
7, 813–816. 2017. Fronto-temporal transcranial direct-current stimulation reduces auditory
Ford, J.M., Mathalon, D.H., 2005. Corollary discharge dysfunction in schizophrenia: can verbal hallucinations and nacetylaspartate-glutamate level in the left temporoparietal
it explain auditory hallucinations? Int. J. Psychophysiol. 58, 179–189. junction in patients with schizophrenia. Brain Stimul. 10, 517–518.
Frith, C.D., Corcoran, R., 2009. Exploring ‘theory of mind’ in people with schizophrenia. Santor, D.A., Ascher-Svanum, H., Lindenmayer, J.P., Obenchain, R.L., 2007. Item re-
Psychol. Med. 26, 521. sponse analysis of the positive and negative syndrome scale. BMC Psychiatry 7, 66.
Frohlich, F., Burrello, T.N., Mellin, J.M., Cordle, A.L., Lustenberger, C.M., Gilmore, J.H., Shergill, S.S., Brammer, M.J., Williams, S.R., Murray, R.M., McGuire, P.K., 2000. Mapping
Jarskog, L.F., 2016. Exploratory study of once-daily transcranial direct current sti- auditory hallucinations in schizophrenia using functional magnetic resonance ima-
mulation (tDCS) as a treatment for auditory hallucinations in schizophrenia. Eur. ging. Arch. Gen. Psychiatry 57, 1033–1038.
Psychiatry 33, 54–60. Silbersweig, D.A., Stern, E., Frith, C., Cahill, C., Holmes, A., Grootoonk, S., Seaward, J.,
Guyatt, G., Oxman, A., Vist, G., Kunz, R., 2008. GrADE: an emerging consensus on rating McKenna, P., Chua, S.E., Schnorr, L., Jones, T., Frackowiak, R.S.J., 1995. A functional
quality of evidence and strength of recommendations. BMJ 336, 924–926. neuroanatomy of hallucinations in schizophrenia. Nature 378, 176.
Haddock, G., McCarron, J., Tarrier, N., Faragher, B., 1999. Scales to measure dimensions Sims, A., 1988. Symptoms in the Mind: An Introduction to Descriptive Psychopathology.
of hallucinations and delusions: the psychotic symptom rating scales. Psychol. Med. Bailliere Tindall Publishers, London, England.
29, 879–889. Smith, R.C., Boules, S., Mattiuz, S., Youssef, M., Tobe, R.H., Sershen, H., Lajtha, A., Nolan,
Higgins, J.P., 2011. Cochrane Handbook for Systematic Reviews of Interventions. Version K., Amiaz, R., Davis, J.M., 2015. Effects of transcranial direct current stimulation
5.1.0 [updated March 2011]. The Cochrane Collaboration. (tDCS) on cognition, symptoms, and smoking in schizophrenia: a randomized con-
Hirsch, M.A., van Wegen, E.E.H., Newman, M.A., Heyn, P.C., 2018. Exercise-induced trolled study. Schizophr. Res. 168, 260–266.
increase in brain-derived neurotrophic factor in human Parkinson's disease: a sys- Sommer, I.E., Slotema, C.W., Daskalakis, Z.J., Derks, E.M., Blom, J.D., van der Gaag, M.,
tematic review and meta-analysis. Transl. Neurodegener. 7, 7. 2012. The treatment of hallucinations in schizophrenia spectrum disorders.
Hoffman, R.E., Boutros, N.N., Hu, S., Berman, R.M., Krystal, J.H., Charney, D.S., 2000. Schizophr. Bull. 38, 704–714.
Transcranial magnetic stimulation and auditory hallucinations in schizophrenia. Stephane, M., Barton, S., Boutros, N.N., 2001. Auditory verbal hallucinations and dys-
Lancet 355, 1073–1075. function of the neural substrates of speech. Schizophr. Res. 50, 61–78.
Insel, T.R., 2010. Rethinking schizophrenia. Nature 468, 187–193. Tian, X., Poeppel, D., 2012. Mental imagery of speech: linking motor and perceptual
Jardri, R., Pouchet, A., Pins, D., Thomas, P., 2011. Cortical activations during auditory systems through internal simulation and estimation. Front. Hum. Neurosci. 6, 314.
verbal hallucinations in schizophrenia: a coordinate-based meta-analysis. Am. J. Upthegrove, R., Broome, M.R., Caldwell, K., Ives, J., Oyebode, F., Wood, S.J., 2016.
Psychiatry 168, 73–81. Understanding auditory verbal hallucinations: a systematic review of current evi-
Keeser, D., Meindl, T., Bor, J., Palm, U., Pogarell, O., Mulert, C., Brunelin, J., Moller, H.J., dence. Acta Psychiatr. Scand. 133, 352–367.
Reiser, M., Padberg, F., 2011. Prefrontal transcranial direct current stimulation Vacas, S.M., Stella, F., Loureiro, J.C., do Couto, F.S., Oliveira-Maia, A.J., Forlenza, O.V.,
changes connectivity of resting-state networks during fMRI. J. Neurosci. 31, 2018. Noninvasive brain stimulation for behavioural and psychological symptoms of
15284–15293. dementia: a systematic review and meta-analysis. Int. J. Geriatr. Psychiatry. 33, 1–10.
Koops, S., van den Brink, H., Sommer, I.E., 2015. Transcranial direct current stimulation Vesterinen, H.M., Sena, E.S., Egan, K.J., Hirst, T.C., Churolov, L., Currie, G.L., Antonic, A.,
as a treatment for auditory hallucinations. Front. Psychol. 6, 244. Howells, D.W., Macleod, M.R., 2014. Meta-analysis of data from animal studies: a
Krishnan, R.R., Fivaz, M., Kraus, M.S., Keefe, R.S.E., 2011. Hierarchical temporal pro- practical guide. J. Neurosci. Methods 221, 92–102.
cessing deficit model of reality distortion and psychoses. Mol. Psychiatry 16, 129. Wang, L., Metzak, P.D., Woodward, T.S., 2011. Aberrant connectivity during self-other
Lan, L., Zhang, X., Li, X., Rong, X., Peng, Y., 2017. The efficacy of transcranial magnetic source monitoring in schizophrenia. Schizophr. Res. 125, 136–142.
stimulation on migraine: a meta-analysis of randomized controlled trails. J. Headache Waters, F., 2012. Multidisciplinary approaches to understanding auditory hallucinations
Pain 18, 86. in schizophrenia and nonschizophrenia populations: the International Consortium on
Lee, E., Chan, P.Y., Lin, J., Hui, C.L.M., Chang, W.C., Chan, S.K.W., Che, E.Y.H., 2017. Hallucination Research. Schizophr. Bull. 38, 693–694.
Efficacy of transcranial direct current stimulation (TDCS) as a treatment for persistent

349