The Escitalopram versus Electric Current Therapy for Treating Depression Clinical Study (ELECT-TDCS): rationale and study design of a non-inferiority, triple-arm, placebo-controlled clinical trial

ABSTRACT CONTEXT AND OBJECTIVE: Major depressive disorder (MDD) is a common psychiatric condition, mostly treated with antidepressant drugs, which are limited due to refractoriness and adverse effects. We describe the study rationale and design of ELECT-TDCS (Escitalopram versus Electric Current Therapy for Treating Depression Clinical Study), which is investigating a non-pharmacological treatment known as transcranial direct current stimulation (tDCS). DESIGN AND SETTING: Phase-III, randomized, non-inferiority, triple-arm, placebo-controlled study, ongoing in São Paulo, Brazil. METHODS: ELECT-TDCS compares the efficacy of active tDCS/placebo pill, sham tDCS/escitalopram 20 mg/day and sham tDCS/placebo pill, for ten weeks, randomizing 240 patients in a 3:3:2 ratio, respectively. Our primary aim is to show that tDCS is not inferior to escitalopram with a non-inferiority margin of at least 50% of the escitalopram effect, in relation to placebo. As secondary aims, we investigate several biomarkers such as genetic polymorphisms, neurotrophin serum markers, motor cortical excitability, heart rate variability and neuroimaging. RESULTS: Proving that tDCS is similarly effective to antidepressants would have a tremendous impact on clinical psychiatry, since tDCS is virtually devoid of adverse effects. Its ease of use, portability and low price are further compelling characteristics for its use in primary and secondary healthcare. Multimodal investigation of biomarkers will also contribute towards understanding the antidepressant mechanisms of action of tDCS. CONCLUSION: Our results have the potential to introduce a novel technique to the therapeutic arsenal of treatments for depression.


INTRODUCTION
Major depressive disorder (MDD) is a psychiatric condition with high prevalence and recurrence worldwide, 1 and it is projected to be the second greatest cause of disability worldwide in 2020. 2 However, antidepressant drugs are only moderately effective for MDD treatment. 3Moreover, common adverse effects can lead to treatment discontinuation and recrudescence of symptoms. 4For these reasons, novel treatment strategies are continuously pursued.
In this context, non-invasive brain stimulation therapies (electroceuticals) have been increasingly investigated as nonpharmacological MDD treatments, such as repetitive transcranial magnetic stimulation (rTMS), a technique that discharges potent electromagnetic fields through a coil placed over the patient's head. 5Repetitive TMS is an effective treatment for MDD, 6,7 although issues such as high cost of application, discomfort and specialized technology limit widespread adoption. 8anscranial direct current stimulation (tDCS) is a neuromodulatory technique that presents low cost, high portability and a benign profile of adverse effects and is relatively simple to use. 9,10It consists of applying a low-intensity electric current across the patient's head using two electrodes on the scalp.In a seminal study in 2000, Nitsche and Paulus demonstrated that anodal and cathodal tDCS applied over the motor cortex respectively increased and decreased motor cortical excitability, as measured by means of TMS motor-evoked potentials, 11 thereby showing that tDCS could alter cortical excitability. 12[15] The use of tDCS for MDD is based on findings from neuroimaging studies (such as Mayberg et al. 16 ), which suggest that the depressive state is associated with hypoactivity of prefrontal areas, particularly the left dorsolateral prefrontal cortex (DLPFC), and that it can be resolved through an increase in the activity of this area.Therefore, treatment strategies that increase left DLPFC activity might treat depressive symptoms.This approach has been adopted and validated in studies using high-frequency (excitatory) rTMS 6 and anodal tDCS over the DLPFC. 17Alternatively, some trials have applied the cathode over the right DLPFC (such as Ferrucci et al. 18 ), or have used low-frequency (inhibitory) rTMS over this region, 19 in accordance with the prefrontal asymmetry theory of depression that states that the right DLPFC presents abnormally high activity during depressive episodes. 20veral non-controlled and controlled studies using tDCS for MDD have been published, from 2006 onwards. 17In the first randomized clinical trial (RCT) in this field, Fregni et al. 21owed the efficacy of active tDCS versus sham for ameliorating depressive symptoms, in a pilot study on 10 patients.More recent RCTs enrolling larger samples 22,23 have shown that active tDCS was an effective treatment for MDD.Its efficacy was corroborated by a recent meta-analysis. 24However, exhaustive RCTs addressing tDCS efficacy are still needed, since some results from RCTs have been non-significant, [25][26][27] the total number of subjects investigated is still low and the optimal parameter protocols for tDCS in MDD remain to be determined.
In a recent factorial, placebo-controlled trial named SELECT-TDCS (Sertraline versus Electric Current Therapy for Treating Depression Clinical Study), we assessed the efficacy of tDCS combined and compared with sertraline 50 mg/ day for treating 120 depressed patients. 28After two weeks of treatment and at the endpoint, the combined treatment was statistically superior to the other groups in terms of depression improvement and response and remission rates.Transcranial DCS was also superior to the other groups at the endpoint and not statistically different from sertraline 23 (Figure 1).The main findings from this study were the synergistic effects of the combination of tDCS and sertraline and the efficacy of tDCS as monotherapy.The x-axis represents depression scores measured using the Montgomery-Asberg depression rating scale (MADRS).The y-axis shows changes in depression scores over time according to the treatment group.Adapted from Brunoni et al. 23  that the third comparison is not strictly necessary if the former two were significant, but in their approach, two or three null hypotheses would have to be independently rejected (thereby decreasing the P value), or the trial would only be valid if all hypotheses were rejected. 30Therefore, we use here an alternative model presented by Pigeot et al. 31 and already used in the literature, 32 in which the study aim can be presented in a single H 0 , which simultaneously tests the non-inferiority between active treatments and their superiority against placebo, thereby decreasing the number of multiple comparisons.

MADRS
Our secondary aims are to explore the clinical improvement in terms of response status (more than 50% of improvement from baseline to endpoint) and remission status (HDRS-17 ≤ 7 at endpoint).We will also explore improvement in depression using the MADRS and the BDI.Finally, we will also investigate whether any early improvement (week 3) was observed between the groups.
We also aim to identify several predictors and mediators of tDCS response, as described below.

Participants
We are recruiting patients of both genders, aged 18 to 75 years Participants will have to be either drug-naïve or drug-free regarding the use of antidepressant drugs (we considered an "antidepressant drug" to be any medication approved for treating MDD by the United States Food and Drug Administration as of the time of onset of the trial).A minimum period of 3 weeks (5 weeks for fluoxetine) will be set aside for drug washout.Benzodiazepine drugs will be allowed, although only at low doses (less than 20 mg/day of diazepam or equivalent).Also, since escitalopram is our active comparator, patients using (or who have used) escitalopram in the current depressive episode will not be included, since they would be escitalopram-resistant.
The adherence strategies for minimizing dropouts include: reimbursement of transportation costs; flexibility in the study schedule; allowance of up to four missed visits during the acute study phase and up to two missed visits during the weekly tDCS phase; reminders for patients regarding their appointments; and offering active tDCS after the endpoint to non-responder patients who received sham tDCS.

Interventions
We are using Soterix Medical tDCS devices specially customized for our study (Soterix Medical, New York, NY, USA, Model 1x1 tDCS-CT).For tDCS, the anode is placed over the left and the cathode over the right dorsolateral prefrontal cortex.The electrode positioning is based on the "Omni-Lateral-Electrode" (OLE) system, which is a simple, reproducible, and practical method for positioning the electrodes using a headband (Figures 2A and 2B).Moreover, the OLE method optimizes electric current densities across the DLPFC, in comparison with other nonneuronavigated methods for tDCS electrode placement, such as the EEG international 10-20 system and the 5-5 cm rule (Seibt, Brunoni, Huang and Bikson, under review), taking into consideration inter-subject variability in head anatomy (Figure 2C).
The stimulation parameters are: current intensity of 2 mA, electrode size of 25 cm 2 ; session duration of 30 minutes (excluding the fade-in and fade-out periods of 15 seconds); and total number of 22 sessions, with 15 sessions applied consecutively once a day (except for weekends), and after that, seven more sessions applied once per week until the study endpoint at week 10.For operational reasons, these sessions are held on either Tuesdays or Thursdays, according to the patient's preference and always respecting the one-week interval.
Sham tDCS is delivered using the same procedure as active tDCS, but using a period of only 30 seconds of active stimulation at 2 mA (or an overall active period of 60 seconds, taking into account the 15 seconds for both the fade-in and fade-out periods), with the stimulator remaining active but not generating current for 30 minutes, for the purpose of double blinding.
This method, adapted from Gandiga et al., 33 has been used in several tDCS protocols 34 and had the same blinding efficacy as the placebo pill in SELECT-TDCS. 35We are using fully automated devices that perform active or sham tDCS according to a randomized stimulation code.
The pharmacological intervention starts simultaneously with tDCS and consists of 10 mg pills of escitalopram oxalate or placebo.Patients receive 10 mg/day of escitalopram/placebo for the first 3 weeks and 20 mg/day for the remaining 7 weeks.
Escitalopram was chosen because it is an effective antidepressant treatment with few adverse effects 36 and its full dose (20 mg/day) is relatively easier to achieve than those of other antidepressant drugs.Thus, maximum dose-up titration can be done in the beginning of the trial.Therefore, we are able to compare tDCS against a full dose of an effective antidepressant without compromising blinding due to adverse effects.
The escitalopram pills are from Libbs (São Paulo, Brazil), a Brazilian pharmaceutical drug company that produces Reconter (escitalopram oxalate), which is a generic drug product comparable to the brand reference Lexapro (Lundbeck Brasil Ltda., Rio de Janeiro, Brazil).In Brazil, all generic drug products have their bioequivalence tested and certified by the Brazilian Health Surveillance Agency (ANVISA).We also independently assessed  1).
To assess adverse effects, we use the Systematic Assessment for Treatment of Emergent Effects (SAFTEE) 37 and the tDCS adverse events questionnaire. 10[40][41] Also, we include neuropsychological assessments in our study, primarily to verify that tDCS will not be associated with cognitive impairment, using the Montreal Cognitive Assessment (MOCA), the Wechsler Adult Intelligence Scale-III digit span (forward and backward) and digit symbol coding subtests, the Verbal Fluency Test (FAS and animal/fruit categories) and the Trail Making test.
The Temperament and Character Inventory (TCI) is being collected at baseline and endpoint to assess the personality traits associated with and modified by the antidepressant treatment.
We also assess blinding performance by asking patients to guess their own treatment allocation group at week 3 and at the endpoint, and ask interviewers to make the same guesses regarding these patients' treatment.
Furthermore, we are investigating several biomarkers, including: a) Heart rate variability (HRV), which becomes lower in cases of depression 42 and possibly reflects the disrupted sympathovagal balance observed in MDD.However, recent evidence suggests that this alteration is primarily driven by the direct effects of antidepressant drug treatment. 43In our SELECT-TDCS trial, we did not observe that either tDCS or sertraline changed HRV levels, which were lower than in matched healthy controls. 44In ELECT-TDCS, we assess HRV values over a larger timeframe (of 10 weeks) and in a larger sample size.
b) Genetic polymorphisms: (i) The serotonin transporter gene (5HTTLPR, SLC6A4) is related to antidepressant response. 45,468][49] In SELECT-TDCS we also found that it was associated with a larger active-sham difference. 50i) The 5HT 2A receptor gene (rs6311 and rs6313) was associated with a larger active-sham difference 51 and antidepressant response.52 Although a meta-analysis was inconclusive regarding its association with antidepressant response, 46 it was identified as associated with citalopram antidepressant response in the STAR*D trial.53  the STAR*D trial. The Aallele was associated with lower synthesis of serotonin 54 and a recent meta-analysis 46 showed that this allele was associated with worse antidepressant response.
(iv) Brain-derived neurotrophic factor (BDNF) polymorphism (rs6265) is a neurotrophin associated with synaptic plasticity. 55,56Two meta-analyses showed that Met/Met and Val/Met were associated with antidepressant response, 46,57 while one rTMS study found that Val/Val determined a better antidepressant response. 58In SELECT-TDCS, this polymorphism was not associated with antidepressant response. 50 Motor cortical excitability for tDCS has been used since the reappraisal of tDCS by Nitsche and Paulus in 2000.11 Currently, one important line of tDCS research involves measurement of motor cortex excitability after use of tDCS in combination with psychoactive drugs.59 This procedure allows indirect measurement of the GABAergic activity (ICI and CSP) and glutamatergic activity (ICF) of the motor system.60 These neurotransmitters are involved in the pathophysiology of depression.61 Recent studies observed that in relation to healthy subjects, the ICI, ICF and CSP indices are altered in depressed subjects, [62][63][64] although their role as predictors of antidepressant response was not sufficiently investigated.erally is lower, [65][66][67] and that there is an increase in gray matter in the left DLPFC after successful antidepressant treatment.68 These brain areas will be explored as predictors of tDCS response. OnI, the fractional anisotropy (FA) of the prefrontal cortex (bilaterally), right temporal lobe and right fusiform gyrus were found to be lower in MDD patients.69 Studies on repetitive TMS found an increase in FA in the left frontal-medial gyrus after successful antidepressant treatment, 70,71 and these regions are explored in our study.For resting-state fMRI, we explore changes in the default-mode network (DMN) and anticorrelated network (AC) after tDCS treatment.Previous studies observed that depressed patients present greater DMN activity and lower AC activity 72 and that antidepressant treatment can change the brain activity of these regions. 73e) BDNF levels in blood are lower in depressed subjects than in healthy subjects and increase after successful pharmacological treatment.74,75 Nonetheless, recent meta-analyses showed that ECT increases BDNF levels in blood, but not rTMS or tDCS.76,77 The possible explanations for this are the low number of rTMS/tDCS studies that used non-optimal treatment protocols and the lower period of observation between measurements. InELECT-TDCS, BDNF levels in blood are assessed at baseline, week 3 and the endpoint.

Design and sample size determination
ELECT-TDCS uses a non-inferiority, triple-arm, placebo-controlled design, with 3:3:2 permuted block randomization in which participants are respectively assigned to escitalopram, tDCS or placebo.Our aim is to prove that tDCS is non-inferior (i.e. that it has similar or superior efficacy) to escitalopram.The placebo group is used for the following purposes.(1) To ensure assay sensitivity.The placebo ascertains that a similar result between escitalopram and tDCS occurred not due to a false-positive finding caused by methodological issues such as insufficient sample size, sample bias, poor blinding, etc., but in fact due to true equivalence between treatments.(2) To allow direct (superiority) comparisons between the pharmacological treatment and placebo, which is critical, since up to 50% of antidepressant drug trials fail to detect superiority between the active treatment and placebo. 78(3) To allow direct comparisons between the experimental treatment and placebo, given that evidence of tDCS efficacy is still being established.( 4) To avoid setting up an arbitrary efficacy margin, which would be a problematic approach because the active comparator in the trial will not necessarily have the same efficacy as previous findings in the literature. 79For these reasons, the non-inferiority design with a placebo group is considered to be the gold standard for this type of trial according to the US Food and Drug Administration (FDA) and the European Medical Agency (EMA). 80e M 2 margin (relative efficacy of the experimental treatment to the active comparator) is based on a fraction retention factor (f) of the M 1 , given that the relative efficacy of the experimental treatment to the active comparator could only be larger than the efficacy with placebo if the active comparator is worse than placebo.The fraction retention varies according to the condition investigated: for instance, in studies focusing on vaccines and oncology, the f value should be close to 1; whereas for chronic and functional disorders, this value ranges between one-half and one-third. 31,79There is no consensus regarding the f value in non-inferiority trials for MDD, although Nutt et al. 81 proposed that for generalized anxiety disorder, the value should be 50%.In this trial, we adopted the f value of 50%, based on Nutt et al. 81 and considering that MDD is comparable to generalized anxiety disorder with regard to methodological aspects (large placebo response, short duration of clinical trials, similar pharmacological treatments, comparable scales etc.).
Therefore, in our trial, M 2 /M 1 = 0.5, although M 1 is based on the difference between the active comparator and placebo and it will be known only when the study has been finished.Although this approach indexes the relative efficacy of tDCS versus escitalopram according to the placebo response of the study, it is theoretically possible that the escitalopram-placebo difference is too large, thus producing a large M 2 value that could favor a finding that comparison between escitalopram and tDCS does not show any significant difference, even if escitalopram clinically outperforms tDCS.To avoid this issue, we will only consider that tDCS was non-inferior to escitalopram if the mean difference between escitalopram and tDCS is less than 3 points on the HDRS-17.The threshold of 3 points was chosen considering the National Institute for Health and Clinical Excellence (NICE) guidelines, which establish that this is a clinically meaningful difference in efficacy. 82 determine the sample size, we first needed to establish the randomization proportions between the groups.Pigeot et al. 31 recommended a 1:1:k p ratio, where k p is determined according to the fraction retention factor, as shown below: According to this formula, for f = 0.5, k p is 0.69.Therefore the optimal proportion for allocation between the groups is 1:1:0.69,or approximately 3:3:2, as used in ELECT-TDCS.
We estimated our sample size in accordance with the suggestion of Pigeot et al. 31 for non-inferiority, triple-arm trials: This formula shows that the sample size is a product of three factors (in parentheses).The factor in the first parentheses take into account the values of α and β, which were, respectively 0.025 (one-tailed) and 0.2.The factor in the second parentheses is the f value (the higher the f value is, the larger the necessary sample size will be), which was determined as 0.5 for our study.The factor in the last parentheses involves the standard deviation and effect sizes of the interventions.We took the value of the standard deviation to be a function of the difference in efficacy between the active comparator and the placebo: The ε value indicates that the standard deviation is associated with the mean difference values observed in the study.It ranges from 0.2 to 2, and values > 1 are considered conservative.In the present study, we took the ε value to be 1.5, i.e. the standard deviation would be 1.5 times the value of the difference in efficacy between the active intervention and the placebo, which is compatible with depression studies in which the variance is usually high.† Therefore, considering an attrition rate of 13% (similar to SELECT-TDCS), β = 0.2, α one-tailed = 0.025, f = 0.5 and ε = 1.5, a total sample of 240 patients will be necessary in order to reject our null hypothesis.For 3:3:2 randomization, this means that in the end, 90, 90 and 60 patients will be respectively allocated to the tDCS, escitalopram and placebo groups.

Statistical analysis
In accordance with the recommendations of Pigeot et al. 31 and Rothmann et al., 79 we use a modified t test to address our primary study hypothesis (i.e. that tDCS is non-inferior to escitalopram).
A t test involves obtaining a T value, which should be higher than This T value should be higher than the critical t, which for a one tailed α of 0.025 and 237 (N-3) degrees of freedom would be 1.97.
We will also perform exploratory analyses to identify whether a significant difference between the groups is observed over time.Therefore, we will perform mixed-model analyses of variance (ANOVAs) with one independent, within-subject variable (time, with three levels: baseline, week 3 and endpoint) and one independent, between-subject variable (group, with three levels).The dependent variables will be the scores from HDRS, MADRS or BDI.To determine whether a statistical significance is observed at a two-tailed p value of 0.05 or less, post-hoc analyses will be performed to analyze the main and interaction effects of our interventions.Exploratory analyses will also be performed to identify the influence of clinical, demographic and biological variables on the outcome.In addition, logistic regressions will be performed using response or remission as dependent variables and the group as an independent variable.Adverse effects will be assessed by counting the number *Where σ is the standard deviation of the study and r = (μ tDCS -μ PLACEBO )/ (μ DRUG -μ PLACEBO ).If μ tDCS = μ DRUG , then r = 1.† For comparative purposes, these values were as follows in SELECT-TDCS: μ tDCS = 7.7, μ PLACEBO = 3.6 and σ = 6.3 for the HDRS-17.Therefore, ε = 6.3/(7.7-3.6), or 1.53 (considering that μ tDCS = μ DRUG ).
of events in each group at each evaluation time.The chi-square test will be used to compare the frequency of adverse effects between the treatment groups.
For non-inferiority trials, it is unclear which approach (intention-to-treat, ITT; or per protocol, PP) should be used.
ITT imputes data from study drop-outs and, although conservative for superiority trials, it can favor the hypothesis of lack of difference between groups in a non-inferiority design and, therefore, benefit (instead of penalizing) a non-inferiority design with methodological flaws.Nonetheless, neither the FDA 83 nor the EMA 84 clearly recommend the optimal approach and they suggest that both ITT and PP should be used.Mulla et al. 85 also suggested that ITT should be used because it has the additional benefit of being more robust in relation to bias, compared with PP analysis.Therefore, missing data will be handled using an ITT approach and the findings will be confirmed through PP analysis.Finally, we will perform additional analyses according to the patients' adherence to the study, using the categories of "fully adherent sample" (patients who missed or rescheduled two visits or fewer) and "completer sample" (patients who completed the study in accordance with the protocol).This approach was also used by George et al. 6

DISCUSSION
ELECT-TDCS will be the largest trial to date assessing the efficacy of tDCS and the first designed to specifically compare the efficacy of tDCS with a full dose of escitalopram.Demonstrating that tDCS presents efficacy similar to that of a pharmacological treatment is important, because antidepressant drugs present several relative and some absolute contraindications against use.
Therefore, tDCS could increase the therapeutic arsenal for use among depressed patients who cannot or are not willing to use antidepressant drugs, e.g.pregnant women, 86 HIV patients 87 and patients presenting clinical conditions in which the pharmacokinetic interactions are problematic. 88r study is also interesting from a cost-efficacy perspective, since tDCS is an affordable, portable and ease-to-use therapy.The cost of antidepressant therapy consisting of tDCS might, in fact, be comparable to that of antidepressant drugs, if it is taken into consideration that one tDCS device can be used to perform several applications per day and that one operator can deliver tDCS to two to three patients simultaneously.Finally, tDCS is a technique with few adverse effects. 10There are at present no reports of seizures or other severe life-threatening events.The most serious adverse effect reported hitherto is skin burn at the application site, and this has been an uncommon finding. 89other key aspect of ELECT-TDCS is that several biomarkers will be assessed in a multimodal approach.tDCS does not seem to induce peripheral effects in cases of depression, such as changes to blood neurotrophic factors. 77,90,91[96] Finally, since the same analyses will be performed in the escitalopram group, we will also explore whether the biological predictors and mediators are different between tDCS and escitalopram responders.
Our study protocol was designed by drawing on our experience from our earlier SELECT-TDCS trial and other advances observed in the field.The choice of using escitalopram at a dose of 20 mg/day represents an advance over the 50 mg/day dose of sertraline used in SELECT-TDCS.The decision to increase the total number of sessions from 12 to 22 was based on findings from a meta-analysis 24 that suggested that a higher tDCS "dose" would be associated with a larger improvement in depression.
Other changes in relation to SELECT-TDCS include: a longer wash-out period to ensure that the subjects were truly "antidepressant drug-free"; an increase in the maximum age of participants in order to improve recruitment; use of automated tDCS devices for sham stimulation, so as to improve blinding; a longer study duration to address the effects of tDCS over a longer time frame; and use of HDRS-17 (instead of MADRS) as the primary outcome measurement, given that in SELECT-TDCS, HDRS-17 presented lower variance and greater psychometric validity (Cronbach's alpha and intraclass correlation) than MADRS, and was also more sensitive towards detecting between-group differences (data not published).
Nonetheless, most aspects of the SELECT-TDCS trial were maintained, such as recruitment of unipolar depressed patients with different degrees of refractoriness in an acute episode of at least moderate severity; enrollment of antidepressant-free patients; inclusion of patients using benzodiazepine drugs, since we considered that exclusion of these patients could harm recruitment and, in fact, be ineffective given that participants can retain information regarding use of these drugs; bilateral (leftright) DLPFC stimulation; and comparison with a placebo arm, which is crucial, given that as the efficacy of active versus sham tDCS is still under investigation.

Limitations
ELECT-TDCS also presents some limitations.First, tDCS is being compared with escitalopram and therefore the results will not be fully generalizable to other antidepressant drugs.
Second, we are not testing other tDCS montages and parameters that have also improved depressive symptoms.Third, we are not testing the combination of tDCS with pharmacotherapy or cognitive-behavioral therapy.Finally, our multimodal analyses are exploratory.

Figure 1 .
Figure 1.Main results of the SELECT-TDCS trial.The figure shows the primary outcome of the Sertraline versus Electric Current Therapy for Treating Depression Clinical Study (SELECT-TDCS).23The x-axis represents depression scores measured using the Montgomery-Asberg depression rating scale (MADRS).The y-axis shows changes in depression scores over time according to the treatment group.Adapted from Brunoni et al.23 who have been diagnosed with major depressive disorder during an acute depressive episode, in accordance with the DSM-5 criteria (Diagnostic and Statistical Manual of Mental Disorders, 5 th edition).The eligibility criteria include the presence of a depressive episode of at least moderate intensity (corresponding to a score ≥ 17 on the 17-item Hamilton Depression Rating Scale, HDRS-17), ability to read and understand Portuguese, at least eight years of schooling and availability to adhere to the study protocol.The exclusion criteria are: 1) other neuropsychiatric conditions, such as bipolar disorder, schizophrenia, substance dependence, dementia, traumatic brain injury, epilepsy and so forth (although participants with anxiety disorders can be included if the primary diagnosis is MDD); 2) high suicide risk (i.e.score > 2 in the Hamilton suicide question); 3) pregnancy; 4) specific contraindications against tDCS, such as electronic or metal implants in the cephalic segment; 5) specific contraindications against escitalopram; 6) severe/life-threatening clinical conditions; or 7) previous participation in other tDCS trials.

whetherFigure 2 .
Figure 2. Electrode positioning in the ELECT-TDCS trial.(A) and (B) The figure shows the placement procedure for the Omni-Lateral-Electrode (OLE) system, which is placed as follows: 1) Select EasyStrap size (small, medium or large); 2) Place the midpoint of the occipital strap over inion (I z ); 3) Position the hinges that link occipital-, electrode-and chin strap over the most dorsal point on the ear; 4) Adjust the angle between the occipital and electrode straps to 165° and the distance across the scalp between the dorsal electrode edges to 10 cm.Adapted from Seibt, Brunoni, Huang and Bikson (under review).(C) Brain current flow produced in one tDCS session in ELECT-TDCS.

d)
Brain imaging by means of MRI will be collected in approximately half of the sample (120 patients) at baseline and endpoint, and this imaging will include voxel-based morphometry (VBM), diffusion tensor imaging (DTI) and resting-sate functional MRI.Regarding VBM, one consistent finding observed in depressed patients is that the volume of the grey matter in the prefrontal and anterior cingulate cortex bilat-

(
in a one-tailed test, as in our study) or different to (in a twotailed test) a critical t value.The T value is obtained by dividing the difference in the means by the mean error (obtained according to the variance and sample size).For this analysis, the T value will be obtained as follows: T = (μ tDCS -0.5μ PLACEBO -0.5μ DRUG )/σ √(1/n TDCS + 0.25/n DRUG + 0.25/n PLACEBO ) or T = 7.44 (μ tDCS -0.5μ PLACEBO -0.5 μ DRUG )/σ (considering that the n values are known)

Table 1 .
Study schedule for the ELECT-TDCS (iii) TPH1 polymorphism was associated with citalopram antidepressant response in "D" and "W" represent interventions performed daily and weekly, respectively.