Long-Term Effects of Repeated Prefrontal Cortex Transcranial Direct Current Stimulation (tDCS) on Food Craving in Normal and Overweight Young Adults

ARTICLE IN PRESS Brain Stimulation ■■ (2016) ■■–■■

Contents lists available at ScienceDirect

Brain Stimulation j o u r n a l h o m e p a g e : w w w. b r a i n s t i m j r n l . c o m

Long-Term Effects of Repeated Prefrontal Cortex Transcranial Direct Current Stimulation (tDCS) on Food Craving in Normal and Overweight Young Adults M. Ljubisavljevic a,*, K. Maxood a, J. Bjekic b, J. Oommen a, N. Nagelkerke c,d a

Department of Physiology, College of Medicine and Health Sciences, UAE University, PO Box 17666, Al Ain, United Arab Emirates Department of Neurophysiology, Institute for Medical Research, University of Belgrade, PO Box 124, Belgrade, Serbia c Malawi-Liverpool-Wellcome Trust Clinical Research Programme, College of Medicine, Queen Elizabeth Central Hospital, Malawi d Institute of Public Health, College of Medicine and Health Sciences, PO Box 17666, UAE University, Al Ain, United Arab Emirates b

A R T I C L E

I N F O

Article history: Received 28 November 2015 Received in revised form 28 June 2016 Accepted 9 July 2016 Available online Keywords: Transcranial direct-current stimulation (tDCS) Craving Brain stimulation Brain plasticity Dorsolateral prefrontal cortex (DLPFC)

A B S T R A C T

Background: The dorsolateral prefrontal cortex (DLPFC) plays an important role in the regulation of food intake. Several previous studies demonstrated that a single session of transcranial direct current stimulation (tDCS) of the DLPFC reduces food craving and caloric intake. Objectives: We hypothesized that repeated tDCS of the right DLPFC cortex may exert long-term changes in food craving in young, healthy adults and that these changes may differ between normal and overweight subjects. Methods: Thirty healthy individuals who reported frequent food cravings without a prior history of eating disorders were initially recruited. Subjects were randomized into an ACTIVE group who received 5 days of real tDCS (20 minutes, anode right-cathode left montage, 2 mA with current density kept at 0.06 mA/ cm2, 1 min ramp-up/ramp-down), and a SHAM group, who received one day of real tDCS, on the first day (same parameters), followed by 4 days of sham tDCS. Food craving intensity was examined by Food Craving Questionnaires State and Trait and Food Craving Inventory before, during, (5-days) and one month (30-days) after tDCS. Results: Single session of tDCS significantly reduced the intensity of current food craving (FCQ-S). Five days of active tDCS significantly reduced habitual experiences of food craving (FCQ-T), when compared to baseline pre-stimulation levels. Furthermore, both current (FCQ-S) and habitual craving (FCQ-T) were significantly reduced 30 days after active tDCS, while sham tDCS, i.e. a single tDCS session did not have significant effects. Also, active tDCS significantly decreased craving for fast food and sweets, and to a lesser degree for fat, while it did not have significant effects on craving for carbohydrates (FCI). There were no significant differences between individual FCQ-T subscales (craving dimensions) after 5 or 30 days of either sham or active tDCS. Changes in craving were not significantly associated with the initial weight, or with weight changes 30 days after the stimulation in the subjects. Conclusions: The results confirm earlier findings that single session of tDCS has immediate effects in reducing food craving. They also show that repeated tDCS over the right DLPFC may increase the duration of its effects, which may be present 30 days after the stimulation. These results support further investigation of the use of tDCS in obesity. © 2016 Elsevier Inc. All rights reserved.

Introduction Food craving is defined as an intense desire to consume a particular food that is difficult to resist [1]. Although still controversial, there is evidence pointing to food craving as a factor involved in

* Corresponding author. Fax: +971 3 7671966. E-mail address: [email protected] (M. Ljubisavljevic). 1935-861X/© 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brs.2016.07.002

obesity and eating disorders [2]. For example, food craving was linked to body mass index, consumption of various types of foods such as sweets, high-fat, and fast-food [3], as well as the ability to lose weight [4,5]. At the same time, diverse findings emerging from both animal models and human studies has drawn attention to the role of the brain in the regulation of food intake and the pathogenesis of obesity. Namely, several lines of evidence suggest that obese subject behavior may be related to abnormal brain function, characterized by dysfunctional inhibitory control and decision-making capacities,

ARTICLE IN PRESS 2

M. Ljubisavljevic et al. / Brain Stimulation ■■ (2016) ■■–■■

primarily related to prefrontal cortex function, e.g. [6,7]. Other studies have further highlighted a link between reward and cognition as an essential component in the cognitive control of food intake in humans [8–12]. Overall, this has prompted the right brain hypothesis of obesity [13]. It postulates that enhancing the activity of the right dorsolateral prefrontal cortex (DLPFC) may strengthen inhibitory control, a core component governing executive self-regulatory processes and goal-oriented eating behavior, thus suppressing the reward-related activity in the reward–cognition neural circuits that drive food craving and overeating [13]. Further support for the potential modulation of DLPFC activity for the management of craving for food also emerged from studies demonstrating that noninvasive stimulation of DLPFC, using techniques such as repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS), can effectively alter other forms of craving such as smoking [14–17], alcoholism [18,19], cocaine addiction [20,21], and eating disorders [22]. Currently, there is still limited research on the effects of rTMS on food cravings. The first trial applied excitatory high-frequency rTMS to the left DLPFC and showed that it inhibits the development of craving when subjects are exposed to real food [23]. A subsequent study, which also used excitatory high-frequency rTMS to the left DLPFC, and aimed to control for the effects of rTMS associated discomfort, found no difference between active and sham stimulation on food craving [24]. Application of inhibitory lowfrequency rTMS to the right DLPFC caused a decrease in the values assigned to food stimuli [25]. The first study that explored the effects of tDCS on food craving showed that a single session of anodal (excitatory) tDCS of the right DLPFC induced a significant decrease in craving for viewed food [26]. Other studies that used the same anode right/cathode left DLPFC stimulation showed that single session of tDCS may temporarily increase the ability to resist food [27], reduce caloric intake [28], and decrease craving for sweet but not savory foods [29]. In overweight subjects, single tDCS combined with exercise decreased hunger and increased satiety immediately after exercise and dampened the appetite 30 minutes later [30]. The only study so far that examined the effects of repeated tDCS showed that one week of daily tDCS, with anode over the right DLPC, reduced overall caloric intake and self-reported appetite scores in overweight subjects [31]. The effectiveness of non-invasive stimulation of the DLPFC to decrease craving levels in substance dependence and craving for highpalatable food was also confirmed by a recent meta-analysis [32]. Despite these encouraging results, no study has yet examined the long-term effects of tDCS on craving. It was our hypothesis that the effects of a single tDCS session will not outlast the intervention, whereas the impact of multiple tDCS sessions will persist after the intervention, thus inducing more lasting effects. To test the hypothesis, we assessed food craving in high-craving individuals immediately after single tDCS session, after multiple tDCS sessions and 30 days later. Material and methods Subjects Subjects were recruited via a broadcast email that was sent to all United Arab Emirates University students with an invitation to participate in a research study, without any initial focus on food craving. Students who expressed interest in participating were asked to complete a questionnaire, which was administered using online LimeService. Participants were not informed that the purpose of this initial questionnaire was to serve as a screening tool, to prevent them from biasing their answers. The questionnaire consisted of 3 parts: demographic data (gender, age, height, weight, and contact details);

a 15 item Food Craving State Questionnaire (FCQ-S) [33] (min score of 15, a max score of 75); and a 12 item craving visual analog scale (CVAS) color picture questionnaire, which required participants to rate how much they ‘would like to eat each food right now if it was available’ to them on a 5-point scale ranging from not at all to very strong (min score of 12, a max score of 60). Pictures were selected from the International Affective Picture System (IAPS) databases and were equally distributed between carbohydrates, fast-food fats, high fats, and sweets. The use of two-dimensional food pictures is widely used for investigation of decision-making about food [34], as it was demonstrated that visual processing of food activates similar brain networks involved in food perception as when tasting and smelling food samples [35,36]. The main outcome measure was the sum of FCQ-S & CVAS (with a min score of 27, and a max score of 135). A total of 280 participants completed the initial screening questionnaire, with a mean score of 58 (range 35 to 131). Scores >108 (i.e. 80% of the cumulative max score and double the mean score) were considered as high craving and all subjects scoring 108 and higher were considered for inclusion. The inclusion criteria were as follows: score of 108 or above on the initial screening questionnaire (signifies the presence of food craving), aged 18 years or older, right handed, and with BMI greater or equal to 18.5 (to exclude underweight subjects). Only righthanded subjects were included to increase the homogeneity of the sample and to be consistent with previous studies on tDCS and food craving. Exclusion criteria included a history of eating disorders, participation in weight loss programs or taking medication to lose weight (currently or within last six months), recreational drug use history, and smoking (>one cigarette per day on average). For the history of eating disorders, screening questions were adapted from the Eating Disorder Examination Questionnaire, eliciting their histories of underweight, binge eating, and purging [37]. A total of 30 subjects (19 Male, 11 Female) were initially enrolled in the study (see also results). The sample size was determined to allow to detect a difference, for a continuous outcome variable, of one within group standard deviation with a power of 80%, using a 5% significance level. Standard calculations give a required sample size of 15 subjects per trial arm, which is also consistent with the existing tDCS literature data [26,27,29,30]. However, it should also be noted that subsequent exclusion of 3 subjects (see later) raises threats to external validity. The study was approved by the local Ethics Committee (Protocol No. 14/53) and all subjects gave written informed consent. The study conforms to the guidelines stated in the Declaration of Helsinki. Tests A battery of craving questions (see below) given on each stimulation day before (pre-) and after (post-) stimulation and 30 days later were used to measure craving. The level of depression was estimated using the Center for Epidemiologic Studies Depression Scale, which was given on Day1 before the stimulation, on Day5 after stimulation and 30 days later. The subjects were also asked to indicate the level of boredom on a 0–10 visual-analog scale before each stimulation session, where 0 indicates the absence of boredom and 10 extreme boredom. The boredom scale was introduced to account for the potential confounding the effect of boredom on craving scale responses. Center for Epidemiologic Studies Depression Scale – Revised (CESD-R) Center for Epidemiologic Studies Depression Scale – Revised [38,39] is a well-established and widely used screening test for depression and depressive disorder. The CESD-R measures symptoms defined by the American Psychiatric Association’s Diagnostic and Statistical Manual (DSM-V) for a major depressive episode. The

ARTICLE IN PRESS M. Ljubisavljevic et al. / Brain Stimulation ■■ (2016) ■■–■■

20 item questionnaire consists of statements and answers ranging from “Not at all or less than one day” to “Nearly every day for 2 weeks”. From their answers, a score is calculated, which is classified into one of five categories of depression (no clinical significance, sub threshold depression symptoms, possible major depressive episode, probably major depressive episode, meets criteria for the major depressive episode). Food Craving Questionnaire State (FCQ-S) and Food Craving Questionnaire Trait (FCQ-T) FCQ-S and FCQ-T were designed to measure specific food cravings [40]. The FCQ-S measures cravings occurring “at this moment”. It is scored on a 5 point scale with “strongly disagree” scored as 1, and “strongly agree” scored as 5. It consists of five subscales namely: desire to eat, positive reinforcement from eating, negative reinforcement from eating, lack of control of eating, and craving as a psychological state. The FCQ-S consists of 15 items, and based on the answers the total score ranges from 15 to 75. The FCQ-T measures craving as a trait. It is scored on a 6 point scale with “never” scored as 1, and “always” scored as 6. It consists of nine subscales namely: intention and planning to consume food, positive reinforcement resulting from eating, negative reinforcement resulting from eating, lack of control of eating, preoccupation with foods, craving as a psychological state, emotions caused by cravings, environmentally induced cravings, and guilt-triggered cravings. The FCQ-S consists of 39 items, and based on the answers the total score ranges from 39 to 234. Food Craving Inventory (FCI) Subjects also completed the Food Craving Inventory (FCI) [41]. The FCI measures cravings for specific foods by examining the frequency of cravings for the items listed in the inventory. The items on the inventory are divided into four categories: carbohydrates/ starches, fast-food fats, high fats, and sweets. It is scored on a 5 point scale with “never” scored as 1, and “always/almost every day” scored as 5. The FCI consists of 28 food items, and based on the answers the scores range from 28 to 140. The FCQ-T and FCQ-S questionnaires are among the most extensively validated multidimensional measures of craving. Their psychometric properties have been shown to be extremely reliable by many construct validity tests, with internal consistency reliability for FCQ-T Cronbach’s alpha coefficient typically reported to be >0.90) in both non-clinical and clinical populations [33,42]. The FCI was also reported to exhibit high inter-item reliability with Cronbach’s alpha coefficient for total scale being >0.85 [41,43]. Study design and stimulation protocol The subjects were randomized into the ACTIVE and SHAM stimulation groups. Subjects in each group were asked to come for one session a day, 5 days in a row, at around the same time each day (3–4 hours after a meal). The ACTIVE group received 5 consecutive days of tDCS while the SHAM group received on the first day active stimulation, followed by 4 days of sham stimulation. This protocol was introduced to allow to examine the immediate effects of single ACTIVE stimulation and compare, at the same time, the longterm effects of single and multiple tDCS sessions. Randomization was stratified by BMI to balance the distribution of BMI between the trial arms. On each day, both groups completed a pre- and poststimulation FCQ-S and FCQ-T questionnaires. FCI questioner was administered only on day 1 and day 5 of stimulation. Subjects were then asked to fill the same questionnaires approximately 30 days after their last stimulation session (31 ± 0.4 days). The researcher conducting the stimulation was not blinded to the participants’ group

3

assignment, whereas the researcher evaluating the responses was blinded for the stimulation (i.e. subjects were coded, and data were anonymised) ensuring a double-blind, parallel protocol. The tDCS was applied with the anode placed over the right DLPFC (F4 located using standard 10–20 EEG electrode position nomenclature) and the cathode placed over the left forehead (serving as a return electrode). The electrodes were 35 cm2 (pad size 5 × 7 cm), and were soaked in a standard saline solution (NaCal 0.9%). Electrodes were connected to a direct current stimulator device (Soterix Medical, USA) which delivered 2 mA current, keeping the current density at 0.06 mA/cm2. The duration of stimulation was 20 minutes, with a ramp-up and -down lasting 30 sec in ACTIVE group. The SHAM stimulation group also received 30 seconds of ramp-up and -down current, at the beginning and end of stimulation session, respectively (Soterix Medical sham waveform), and no current during the session, which is accepted routine for ineffective stimulation [44]. Ramp-up and -down period further aided in ensuring that the subjects remained ‘blind’ for the type of stimulation received during the test ensuring a sham control effect [45,46]. At the end of each stimulation session, the subjects were asked whether they felt any discomfort or other unpleasant sensations (i.e. pain) apart from tingling and itching at the beginning of the stimulation. Also, subjects were asked whether they experienced any adverse effects after the previous stimulation session. No adverse effects were recorded. Data analysis and statistics To reduce heteroskedasticity in regression analysis, craving (FCQS, FCQ-T, and FCI) and depression (CESD-R) scores were analyzed on a log scale. We analyzed each craving scale and depression scale separately by examining the effects of stimulation (ACTIVE vs. SHAM) on craving and depression at Day1 Post, Day5 Pre, and Post, and Day 30. All trials were analyzed using analysis of covariance (i.e. regression). For all trials craving and depression scores (log transformed) at given time (Day1 POST, Day5 PRE and POST, and Day 30) were regressed on the categorical variable stimulus (trial arm) and their corresponding “baseline” value at Day1 Pre. This approach has greater statistical power to detect treatment effects than other methods and enables one to estimate effects of group differences while controlling for baseline [47]. Data were analyzed using IBM SPSS v. 21.0 for Windows (IBM Corp.). Other standard statistical tests were used as stated in the results section. Data were summarized as means ± standard deviation (SD). A significance level of p < 0.05 (twotailed) was used throughout. Results Two participants randomized to ACTIVE and one randomized to SHAM were excluded from analysis: one became depressed (Day 30 CESD-R score of 42), one lost 18 kg due to dieting (see inclusion/ exclusion criteria) and one declined to fill out the questionnaire on Day 30. The participants’ demographic data are summarized in Table 1. In brief, after exclusion of subjects the results are based on 27 subjects (13 received active and 14 sham tDCS). There were no statistically significant differences in age, and BMI on Day 1, between ACTIVE and SHAM group (p > 0.05, t-test). Somewhat surprisingly, there was no statistically significant correlation between the BMI and craving on Day1 before the first tDCS session (p > 0.05, Pearson’s correlation test). On Day 5 of stimulation, subjects were asked to guess whether they received real or sham stimulation. Most (11 in the ACTIVE group, and 12 in the SHAM group) believed they had received the real stimulation while 3 subjects were unsure, and 1 thought to have received sham stimulation. Thus, 60% of subjects guessed wrongly, and their guesses did not differ between the two

ARTICLE IN PRESS M. Ljubisavljevic et al. / Brain Stimulation ■■ (2016) ■■–■■

4

Table 1 Demographic characteristics of participants in ACTIVE and SHAM transcranial direct current stimulation (tDCS) groups. Subjects demographics

ACTIVE

SHAM

All

Age Height Weight Day 1 Weight Day 30 BMI Day 1 BMI Day 30 BMI normal BMI BMI Male Female

21 ± 2.1 1.72 ± 0.1 79.2 ± 21 78.9 ± 21.6 26.3 ± 5.1 26.2 ± 5.2 22 ± 1.6 (6) 26.8 ± 1.2 (4) 34.2 ± 2.3 (3) 9 4

21.6 ± 2 1.74 ± 0.1 75.5 ± 15.2 75.5 ± 15.4 24.9 ± 3.6 24.8 ± 3.7 22.2 ± 1.9 (7) 27.6 ± 0.9 (4) 34.1 ± 2.8 (3) 10 4

21.3 ± 2 1.73 ± 0.1 77.3 ± 18.3 77.1 ± 18.7 25.6 ± 4.4 25.5 ± 4.6 22.1 ± 1.7 (13) 27.2 ± 1 (8) 34.1 ± 2.6 (6) 19 8

trial arms (χ2 (2) = 1.34, p = .51). There was no significant variation (p > 0.05, t-test) in boredom over 5 days of stimulation (mean 5.3, range 4.6–5.8) nor there was a significant correlation between the level of boredom and FCQ-T and FCQ-S score on any of the stimulation days (p > 0.05, Pearson’s correlation test). Table 2 summarize changes in FCQs, FCI, and CESD-R depression scale on Days 1, 5 and 30, while Fig. 1 also shows their changes on Days 2, 3 and 4. Also, Table 2 show changes in FCQ-S and FCQ-T subscales on Days 1, 5 and 30. Craving intensity decreased after a single stimulation session on Day 1 in both ACTIVE and SHAM groups, which both on Day 1 received real tDCS (see methodology). Craving intensity further decreased during 5 days of ACTIVE stimulation (Day 1 to Day 5) and remained decreased 30 days later (Day 30) (Table 2 and Fig. 1). In contrast, in the SHAM group the strength of craving remained unchanged after the first day of stimulation until the end of stimulation period (Day 1 to Day 5) and increased 30 days later (Table 2 and Fig. 1). To examine the effects of single-tDCS session on craving dependent variable FCQ-S (measures current craving) on Day 1 POST was regressed on Day 1 PRE baseline values and GROUP as predictors (i.e. independent variables). The overall regression model was statistically significant for Day 1 POST (AdjR2=0.733, F(2,24) = 36.795, p < .01), but with only Day 1 PRE having a significant effect (β = 0.880 p < 0.01), while the GROUP did not have a significant effect (β = 0.037, p > 0.05). This confirms the effectiveness of active stimulation in SHAM group on Day 1. To evaluate changes in craving after 5 days of tDCS, a separate regression anal-

yses with the dependent variable FCQ-T regressed on Day 1 PRE baseline values and GROUP as predictors showed significant Day 5 POST effect (AdjR2=0.470, F(2,24) = 12.512, p < .01), with both predictors having a significant effect (FCQ-T Day 1 PRE: β = 0.752, p < 0.01; GROUP: β = 0.306, p = 0.05), establishing absence of effects in the SHAM group on Day 5. Table 3 summarizes the long-term effects of tDCS on craving 30 days after stimulation. For this, a regression analysis using Day 30 FCQ-S and FCI scales as dependent variables was performed. Craving intensity was significantly reduced in tDCS ACTIVE group 30 days after the stimulation (Day 30), with both current (FCQ-S) and habitual craving (FCQ-T) scores being significantly dependent on baseline values (FCQs at Day 1 PRE) and stimulation (ACTIVE vs. SHAM group). Also, craving related to different classes of food (FCI) was significantly decreased on Day 30, and dependent on both Day 1 PRE baseline values and stimulation group (Table 3). Craving was significantly decreased for fast-food, sweets, and fat while there was no significant change in craving for food rich in carbohydrates (Table 4). Finally, the same analysis conducted on FCQ-T subscales (Table 2) showed no significant effect of subscale on Day 30 when regressed on Day 1 PRE subscale baselines (p > 0.05). Regression analysis of depression (CESD-R) score on Day 30 to Day 1 PRE and the GROUP did not show significant group effect nor significant changes in subject’s depression levels (p > 0.05). Regression analysis of the same dependent variables (FCQs and FCI on Day 30) regressed on corresponding Day1 PRE craving baseline values and BMI, and GENDER as predictors in separate trials did not show a significant effect of BMI nor gender (p > 0.05). Discussion We examined short (1 and 5 days) and long-term (30 days) changes in the intensity of self-reported food craving after one and five tDCS sessions over the prefrontal cortex (anode right) in normal and overweight high-craving healthy subjects. Consistent with earlier findings craving was reduced shortly after a single session of active tDCS [23,26,27]. The present study expands on these results by demonstrating that repeated tDCS for 5 days may induce more persistent (at least for 30 days) decrease in self-reported craving. This effect differed by food type, in that tDCS reduced food craving for sweet,

Table 2 Raw Food-Cravings Questionnaire – State (FCQ-S) and Food-Craving Inventory (FCI) craving scores, and FCI and Food-Cravings Questionnaire – Trait (FCQ-T) subscale scores (mean ± SD), on Day 1 before (PRE), Day 5 after (POST) transcranial direct current stimulation (tDCS) and 30 days after the last tDCS (Day 30), in ACTIVE and SHAM stimulated groups (n = 13 and n = 14, respectively). Day 1 Pre

Day 1 Post

Day 5 Pre

Day 5 Post

Day 30

Scale/subscale

Active

Sham

Active

Sham

Active

Sham

Active

Sham

Active

Sham

FCQ-T FCQ-S FCI CESD-R FCI Carb FCI F Food FCI Fat FCI Sweet FCQ-T Intentions FCQ-T Anticipation of positive reinforcement FCQ-T Anticipation of relief FCQ-T Lack of control FCQ-T Thoughts and preoccupation FCQ-T Craving as a state FCQ-T Emotions FCQ-T Cues FCQ-T Guilt

153.9 ± 30.6 57.5 ± 7.4 90 ± 13.7 13.4 ± 8.4 24.5 ± 5.1 14.7 ± 2.8 23.5 ± 5.7 24.9 ± 6.9 14.1 ± 2.6 23.7 ± 4.1

132 ± 35.8 50.8 ± 11.2 74.8 ± 16.7 9 ± 5.7 19.7 ± 5.0 13.4 ± 3.8 19.1 ± 6.3 22.7 ± 7.4 11.5 ± 3.9 21 ± 5

146.7 ± 34.7 55.5 ± 7.8 89.7 ± 18.1

125.1 ± 42.3 50.8 ± 10.4 76.4 ± 16.4

125.7 ± 39 48 ± 12 75 ± 19.1

19.7 ± 3.7 13.4 ± 4.1 19.5 ± 6.3 23 ± 8.3 11.2 ± 3.7 19.4 ± 4.8

127.6 ± 40.2 50.1 ± 12.3 75.5 ± 18.8 8.2 ± 7 20.3 ± 5.8 12.8 ± 4.2 19.5 ± 6.4 22.5 ± 8.3 12 ± 3.6 19 ± 5

132.6 ± 32.5 46.5 ± 14.4 80.6 ± 22.7

23.3 ± 6.2 15.1 ± 3.5 23 ± 5.9 25 ± 7.8 13.1 ± 2.7 23.4 ± 4

136.2 ± 32.7 47.2 ± 13.6 79.4 ± 20.9 11.7 ± 8.5 21.6 ± 6.8 12.1 ± 4.1 21 ± 6.6 21.4 ± 8.1 12.1 ± 2.9 21.1 ± 5

21.2 ± 6.8 13.1 ± 4.2 20.7 ± 7.6 22.1 ± 8.2 11.3 ± 3.9 21.2 ± 5.9

19.4 ± 6.0 13.4 ± 4.3 19.5 ± 6.4 22.5 ± 8.3 11.7 ± 3.8 18.8 ± 5.4

134.6 ± 39.5 44.7 ± 13.9 71.4 ± 22 14.4 ± 8.2 21.4 ± 6.7 12.6 ± 3.9 20.3 ± 7.5 20.2 ± 6.7 11.4 ± 3.4 19.3 ± 5.9

135.7 ± 35.7 53.8 ± 9.8 84.8 ± 25.3 11.8 ± 10.4 22.6 ± 7.2 13.7 ± 4.0 22.3 ± 7.3 24.7 ± 8.3 11.9 ± 3.3 20 ± 4.7

12.4 ± 3.1 22.5 ± 7.4 22.4 ± 8.6

10.4 ± 2.4 19 ± 7.6 19.1 ± 10.4

11.7 ± 3.5 21 ± 8.4 21.7 ± 7.5

9.3 ± 3.3 17.1 ± 9.1 19 ± 10

10.7 ± 3.9 19.3 ± 7.5 21.4 ± 7.5

9.2 ± 3.5 18.6 ± 8.8 19.9 ± 10.4

10.4 ± 3.9 17.2 ± 8.1 21 ± 7.3

9.2 ± 3.9 18.1 ± 8.5 19.3 ± 9.9

11 ± 4.4 19.4 ± 6.6 21.3 ± 7.9

10.5 ± 3.8 20 ± 7.3 20.9 ± 9.6

18.1 ± 2.3 14.7 ± 6 17.7 ± 3.4 8.7 ± 3.5

16.7 ± 3.6 10.2 ± 4.4 16.5 ± 5 8 ± 4.1

16.9 ± 3.9 13.7 ± 6 16.9 ± 3.8 8.7 ± 3.9

16 ± 4.5 10.2 ± 5.1 15.6 ± 5.4 7.6 ± 4.4

15.5 ± 3.1 12.4 ± 6 16.4 ± 3.9 8.3 ± 3.9

15.3 ± 4.6 10 ± 4.9 15.5 ± 4.6 7.7 ± 4.5

15.3 ± 4.2 11.7 ± 6 15.8 ± 4.5 8.5 ± 4

15 ± 4.6 10.4 ± 4.9 14.6 ± 5.1 7.9 ± 4

15.4 ± 3.3 13 ± 6.1 14.8 ± 4.5 9.4 ± 4.3

17 ± 3.4 11.2 ± 5 15.8 ± 4.2 8.8 ± 4.3

ARTICLE IN PRESS M. Ljubisavljevic et al. / Brain Stimulation ■■ (2016) ■■–■■

5

Figure 1. Changes in Food-Cravings Questionnaires – Trait and State (FCQ-T and FCQ-S, respectively) and Food-Craving Inventory (FCI) scores and depression scale scores (CESD-R), during 5 days stimulation and 30 days later in ACTIVE (upper panel) and SHAM stimulated (lower panel) group.

fast-food and fat but not for carbohydrates. Food cravings were reduced regardless of initial body weight or gender and were not associated with changes in body weight during and after the intervention. Neuroimaging studies over the last decade had provided significant insights into neurobiological underpinning of abnormal craving and eating behavior. Overall they point to dysfunction of a broad

range of areas in the obese brain, including exaggerated responsiveness of motivation–reward areas and diminished responsiveness of striatum to food cues, as well as dysregulation between them [9]. Additionally, obese individuals also show diminished activation of areas that provide cognitive control and control motivated actions like DLPFC [6,8–10]. Indirect support has also emerged from studies on drug cues and food cues, suggesting that overlapping brain

Table 3 Results of regression analysis of craving scores of Food-Cravings Questionnaires – Trait and State (FCQ-T and FCQ-S, respectively) and Food-Craving Inventory (FCI) on Day 30 using baseline scores on Day 1 before (Day1 PRE) transcranial direct current stimulation and the stimulation group (GROUP) as predictors. Dependent variables

Regression analysis summary

FCQ-T Day 30 Predictors

R = 0.724 Group Day1 Pre R = 0.519 Group Day1 Pre R = 0.548 Group Day1 Pre

FCQ-S Day 30 Predictors FCI Day 30 Predictors

R2 = 0.524 β = 0.308 β = 0.768 R2 = 0.269 β = 0.490 β = 0.402 R2 = 0.300 β = 0.505 β = 0.542

adj.R2 = 0.485 B = 0.185 B = 0.894 adj.R2 = 0.209 B = 0.321 B = 0.616 adj.R2 = 0.242 B = 0.358 B = 0.878

F(2,24) = 13.220 SEB = 0.913 SEB = 0.174 F(2,24) = 4.4284 SEB = 0.121 SEB = 0.283 F(2,24) = 5.1424 SEB = 0.136 SEB = 0.310

p < 0.001 p = 0.050 p = 0.001 p < 0.023 p = 0.014 p = 0.039 p < 0.013 p = 0.014 p = 0.009

ARTICLE IN PRESS M. Ljubisavljevic et al. / Brain Stimulation ■■ (2016) ■■–■■

6

Table 4 Results of regression analysis of Food-Craving Inventory (FCI) subscale scores for each food category (carbohydrates, fast food, fats, and sweet) on Day 30 using baseline scores on Day 1 before (Day1 PRE) transcranial direct current stimulation and the stimulation group (GROUP) as predictors. Dependent variables

Regression analysis summary

FCI Day 30 Carb Predictors

R = 0.380 Group Day1 Carb R = 0.697 Group Day1 Fast Food R = 0.478 Group Day1 Fat R = 0.629 Group Day1 Sweet

FCI Day 30 Fast Food Predictors FCI Day 30 Fat Predictors FCI Day 30 Sweet Predictors

R2 = 0.144 β = 0.344 β = 0.405 R2 = 0.485 β = 0.391 β = 0.692 R2 = 0.228 β = 0.412 β = 0.431 R2 = 0.395 β = 0.470 β = 0.514

circuits and their dysfunction may underlie the observed deficits in obesity and addictive disorders [48,49]. Overall, this has prompted the right brain hypothesis of obesity [13], which postulates that enhancing the right DLPFC activity may strengthen inhibitory control, a core component governing executive self-regulatory processes and goal-oriented eating behavior, thus suppressing the rewardrelated activity in the reward–cognition neural circuits that drive food craving and overeating [13,50]. Several studies used non-invasive brain modulation including rTMS and tDCS to test this hypothesis. They targeted either left (rTMS) [23,24] or right DLPFC (rTMS and tDCS) [25–31]. They consistently found an acute decrease in self-reported food craving and appetite on the visual-analog scale, and the ability to resist food and immediate reduction in caloric intake. Our study also targeted the right DLPFC. Similar to earlier studies, the subjects in our study showed an immediate decrease in food craving after a single session of stimulation (Day 1). The average decrease in FCQ-S, which measures immediate, state-dependent craving, was 4% in ACTIVE group and 1% in SHAM group, which also received real tDCS on Day 1. The average decrease in FCQ-T, which assesses characteristic or typical craving patterns as well as other cognitive aspects of craving, was 5% and 6% in ACTIVE and SHAM groups, respectively. The desire for food consumption measured by VAS decreased by 18% after anode right (active), and cathode left compared to that before stimulation [26]. Goldman et al. found an even larger decrease in subjects’ ability to resist food (26% on VAS scale) and how much they would like to eat each food right now if it were available [27]. Although significant, a considerably smaller decrease in immediate craving in our study may be related to the montage used in our study. We used active anode right (over F4) and an indifferent cathode placed over subjects’ forehead, thus lacking an inhibitory effect of the right DLPFC tDCS potentially present in studies that used cathodes over F3 [26]. Other possible explanation may be related to different measures of craving used in our study. Namely, unlike the majority of previous studies, which measured cue-induced craving (i.e. selfreported craving in response to cues) and caloric intake, we used craving scales (FCQ-T and S) that measure tonic craving (i.e. selfreported craving in the absence of external cues). This may be important as it has been repeatedly demonstrated that cue exposure and associated experience of craving significantly influence and contribute to eating behavior and weight gain [36]. It may have been that in studies using food cues, a strongly conditioned food reactivity response was induced potentially allowing for stronger tDCS effects. On the other hand, as we were interested to examine longterm effects of tDCS, we deemed that if self-reported craving (i.e. tonic-craving) is reduced by tDCS, this would support its potential use as a treatment modality in obesity. Also, since it has been shown by a recent meta-analysis that both cue-induced craving and tonic craving are associated with subsequent eating and weight-related

adj.R2 = 0.728 B = 0.858 B = 0.568 adj.R2 = 0.442 B = 0.285 B = 0.883 adj.R2 = 0.164 B = 0.331 B = 0.499 adj.R2 = 0.345 B = 0.374 B = 0.623

F(2,24) = 2.021 SEB = 1.088 SEB = 0.305 F(2,24) = 11.327 SEB = 0.152 SEB = 0.152 F(2,24) = 3.551 SEB = 0.154 SEB = 0.222 F(2,24) = 7.844 SEB = 0.128 SEB = 0.196

p > 0.050 p = 0.438 p = 0.075 p < 0.001 p = 0.017 p = 0.001 p < 0.044 p = 0.042 p = 0.034 p < 0.002 p = 0.007 p = 0.004

outcomes [36], current results that showed significant tDCS effects on tonic craving as well, further argues for the potential use of tDCS in obesity. The only earlier study that examined the effects of repeated tDCS (8 days) showed a 14% decrease in caloric consumption [31]. In our study, the average decrease of FCQ-T was 13% and of FCQ-s was 19% after 5 days of active stimulation. Interestingly, Jauch-Chara et al. did not find significant differences in food intake between conditions (tDCS vs. sham stimulation) after the first day of stimulation [31]. Similar to our study they also used active right DLPFC stimulation, while the cathode electrode was positioned over the left forehead (supraorbital). The potential difference in tDCS effects after the first day of stimulation may be related to stimulation intensity, which was 1 mA vs. 2 mA used in our study. This is the first study that examined long-term changes in craving intensity after repeated tDCS. We found significant decreases in FCQs compared to baseline values at the beginning of stimulation, which were similar to decrease recorded on Day 5 (FCQ-T decreased 13% and FCQ-S 22%). On the other hand, FCI decreased even further to 21% compared to 11% decrease after Day 5. Long-term effects of DLPFC stimulation were demonstrated earlier, for example with rTMS in drug-resistant major depression [51,52], and with tDCS for treatment of chronic pain [53]. However, studies of long-term effects of tDCS are still sparse. Repeated anodal tDCS improves motion perception in subjects with occipital stroke [54] 28 days after the stimulation and improve chronic post-stroke aphasia [55] 21 weeks after the treatment. The current study adds to growing body of evidence that repeated tDCS may be potentially used to establish and/ or aid improved cognitive control of craving. Nevertheless, it should also be noted that our study used a between-subject design with a relatively small sample (N = 27), which decreased the power of the study, raising the potential issue of external validity, and prompting for larger controlled trials to verify the observed effects. Our study was not designed to examine the mechanisms of longterm action of tDCS. Several recent studies point out that tDCS exert its effects through regulation of BDNF. Also, tDCS induced changes in the BDNF levels in different brain regions, reversing, at the same time, neuropathic pain-related behavioral alterations in rats [56]. In mice, a single session of anodal tDCS induced one-week lasting increases in hippocampal LTP, learning and memory via epigenetic remodeling of BDNF expression [57]. In older adults, the BDNF polymorphism seems to play a role in modulating tDCS-induced motor cortex plasticity further emphasizing the potential role of BDNF in shaping the duration of tDCS effects [58]. The mechanisms behind lasting tDCS effects on craving are certainly complex and can be attributed, at least partly, to different BDNF-related cellular signaling activations causing modulation of GABAergic pathways [59], dopaminergic pathways [60,61] and NMDA related plasticity [62,63].

ARTICLE IN PRESS M. Ljubisavljevic et al. / Brain Stimulation ■■ (2016) ■■–■■

Finally, changes in FCI 30 days after the stimulation suggest decreased craving for sweet, fast-food and fat but not for carbohydrates. Goldman et al. [27] showed that active prefrontal tDCS acutely and significantly decrease cravings for sweet food and carbohydrates. Similarly, Jauch-Chara et al. [31] showed diminished food intake of carbohydrates, while Kekic et al. [29] found decreased craving for sweets. One possible explanation for decreased craving for fast- and sweet food in our study may be related to cultural aspects prevalent in the region where the study was conducted. Namely, all of our subjects were young individuals of Arab descent, born and raised in the Gulf region. Furthermore, a considerable number of them live alone, in the student accommodation. Potentially, these factors may have determined their food-cue reactivity, limiting it to fast-food (student lifestyle and dominant fast-food culture in the region) and sweet food (cultural aspect). At the same time, cue-reactivity related to carbohydrates (pasta and rice, potentially more relevant in the far-east, Europe and the US) may have been lesser yielding no tDCS effect. Interestingly, although carbohydrate craving has been a primary focus of craving research, the initial study that used FCI suggested that specific carbohydrate craving may not be as common [41]. Somewhat surprisingly no association between BMI or gender and tDCS effects on food craving was found. One possibility may be related to limited sample examined in the study with the majority of subjects having normal BMI (N = 13) and only 7 and 6 subjects being overweight and obese, respectively. A recent study found that obese subjects who received anodal tDCS of the left DLPFC tended to have lower caloric intake and lose more weight than with cathodal tDCS [64]. This may suggest that the effects of tDCS depend on the extent of abnormal brain signaling. It is worth noting that we tried to address several potential limitations. To exclude all potential biases related to tDCS and craving subjects were blinded to stimulation. Also, both groups received real stimulation on Day 1. Moreover, lastly, the SHAM group received standard “ramp up/down” stimulation during the sham protocol. Altogether these made it more difficult for subjects to guess correctly whether they received real of sham stimulation, with the majority thinking that they received real stimulation. To exclude the potential effect of mood changes subjects were screened for depression. To exclude the possible immediate effect of food intake and satiety subjects were stimulated in an intermediate state, i.e. about 2–4 h after a meal. To control for long-term effects of weight changes and dieting on cortical plasticity and tDCS neuromodulatory effects only subjects with stable body weight and no dietary intervention for the past 6 months were included in the study. Furthermore, subjects were instructed to maintain their normal, usual diet throughout the study and not engage in any dieting regimen or take medication or dieting supplements. To limit the potential effects of cognitive tasks geared toward cravings during the stimulation subjects were asked to get involved in self-selected cognitive activity not related, in any way to food, i.e. reading for the exam, or listening to music. Lastly, none of our subjects had excessive head subcutaneous fat, which could have affected the current distribution and its density [65]. Conclusions The results show that repeated tDCS reduce craving even 30 days after the intervention. Although based on a small sample size, the results strengthen the case for the use of tDCS as an adjuvant intervention for overweight and obesity. Acknowledgments The study was supported by UAE University Grant 31M102.

7

References [1] Wang G-J, Volkow ND, Thanos PK, Fowler JS. Similarity between obesity and drug addiction as assessed by neurofunctional imaging: a concept review. J Addict Dis 2004;23:39–53. doi:10.1300/J069v23n03_04. [2] Potenza MN, Grilo CM. How relevant is food craving to obesity and its treatment? Front Psychiatry 2014;5:164. doi:10.3389/fpsyt.2014.00164. [3] Chao A, Grilo CM, White MA, Sinha R. Food cravings, food intake, and weight status in a community-based sample. Eat Behav 2014;15:478–82. doi:10.1016/ j.eatbeh.2014.06.003. [4] Batra P, Das SK, Salinardi T, Robinson L, Saltzman E, Scott T, et al. Relationship of cravings with weight loss and hunger. Results from a 6 month worksite weight loss intervention. Appetite 2013;69:1–7. doi:10.1016/j.appet.2013.05.002. [5] Meule A, Papies EK, Kübler A. Differentiating between successful and unsuccessful dieters. Validity and reliability of the Perceived Self-Regulatory Success in Dieting Scale. Appetite 2012;58:822–6. doi:10.1016/j.appet.2012. 01.028. [6] Pignatti R, Bertella L, Albani G, Mauro A, Molinari E, Semenza C. Decision-making in obesity: a study using the Gambling Task. Eat Weight Disord 2006;11:126–32. [7] Uher R, Treasure J. Brain lesions and eating disorders. J Neurol Neurosurg Psychiatry 2005;76:852–7. doi:10.1136/jnnp.2004.048819. [8] Brooks SJ, Cedernaes J, Schiö HB. Increased prefrontal and parahippocampal activation with reduced dorsolateral prefrontal and insular cortex activation to food images in obesity: a meta-analysis of fMRI studies. PLoS ONE 2013;8:e60393. doi:10.1371/journal.pone.0060393. [9] Carnell S, Gibson C, Benson L, Ochner CN, Geliebter A. Neuroimaging and obesity: current knowledge and future directions. Obes Rev 2012;13:43–56. doi:10.1111/ j.1467-789X.2011.00927.x. [10] Hollmann M, Hellrung L, Pleger B, Schlögl H, Kabisch S, Stumvoll M, et al. Neural correlates of the volitional regulation of the desire for food. Int J Obes (Lond) 2012;36:648–55. [11] Kober H, Mende-Siedlecki P, Kross EF, Weber J, Mischel W, Hart CL, et al. Prefrontal–striatal pathway underlies cognitive regulation of craving. Proc Natl Acad Sci 2010;107:14811–16. doi:10.1073/pnas.1007779107. [12] Appelhans BM. Neurobehavioral inhibition of reward-driven feeding: implications for dieting and obesity. Obesity (Silver Spring) 2009;17:640–7. doi:10.1038/oby.2008.638. [13] Alonso-Alonso M, Pascual-Leone A. The right brain hypothesis for obesity. JAMA 2007;297:1819–22. doi:10.1001/jama.297.16.1819. [14] Amiaz R, Levy D, Vainiger D, Grunhaus L, Zangen A. Repeated high-frequency transcranial magnetic stimulation over the dorsolateral prefrontal cortex reduces cigarette craving and consumption. Addiction 2009;104:653–60. doi:10.1111/ j.1360-0443.2008.02448.x. [15] Eichhammer P, Johann M, Kharraz A, Binder H, Pittrow D, Wodarz N, et al. High-frequency repetitive transcranial magnetic stimulation decreases cigarette smoking. J Clin Psychiatry 2003;64:951–3. [16] Fregni F, Liguori P, Fecteau S, Nitsche MA, Pascual-Leone A, Boggio PS. Cortical stimulation of the prefrontal cortex with transcranial direct current stimulation reduces cue-provoked smoking craving: a randomized, Sham-controlled study. J Clin Psychiatry 2008;69:32–40. [17] Boggio PS, Liguori P, Sultani N, Rezende L, Fecteau S, Fregni F. Cumulative priming effects of cortical stimulation on smoking cue-induced craving. Neurosci Lett 2009;463:82–6. doi:10.1016/j.neulet.2009.07.041. [18] Boggio PS, Sultani N, Fecteau S, Merabet L, Mecca T, Pascual-Leone A, et al. Prefrontal cortex modulation using transcranial DC stimulation reduces alcohol craving: a double-blind, sham-controlled study. Drug Alcohol Depend 2008;92:55–60. doi:10.1016/j.drugalcdep.2007.06.011. [19] Mishra BR, Nizamie SH, Das B, Praharaj SK. Efficacy of repetitive transcranial magnetic stimulation in alcohol dependence: a sham-controlled study. Addiction 2010;105:49–55. doi:10.1111/j.1360-0443.2009.02777.x. [20] Camprodon JA, Martínez-Raga J, Alonso-Alonso M, Shih M-C, Pascual-Leone A. One session of high frequency repetitive transcranial magnetic stimulation (rTMS) to the right prefrontal cortex transiently reduces cocaine craving. Drug Alcohol Depend 2007;86:91–4. doi:10.1016/j.drugalcdep.2006.06.002. [21] Politi E, Fauci E, Santoro A, Smeraldi E. Daily sessions of transcranial magnetic stimulation to the left prefrontal cortex gradually reduce cocaine craving. Am J Addict 2008;17:345–6. doi:10.1080/10550490802139283. [22] McClelland J, Bozhilova N, Campbell I, Schmidt U. A systematic review of the effects of neuromodulation on eating and body weight: evidence from human and animal studies. Eur Eat Disord Rev 2013;21:436–55. doi:10.1002/erv.2256. [23] Uher R, Yoganathan D, Mogg A, Eranti SV, Treasure J, Campbell IC, et al. Effect of left prefrontal repetitive transcranial magnetic stimulation on food craving. Biol Psychiatry 2005;58:840–2. doi:10.1016/j.biopsych.2005.05.043. [24] Barth KS, Rydin-Gray S, Kose S, Borckardt JJ, O’Neil PM, Shaw D, et al. Food cravings and the effects of left prefrontal repetitive transcranial magnetic stimulation using an improved sham condition. Front Psychiatry 2011;2:9. doi:10.3389/fpsyt.2011.00009. [25] Camus M, Halelamien N, Plassmann H, Shimojo S, O’Doherty J, Camerer C, et al. Repetitive transcranial magnetic stimulation over the right dorsolateral prefrontal cortex decreases valuations during food choices. Eur J Neurosci 2009;30:1980–8. doi:10.1111/j.1460-9568.2009.06991.x. [26] Fregni F, Orsati F, Pedrosa W, Fecteau S, Tome FAM, Nitsche MA, et al. Transcranial direct current stimulation of the prefrontal cortex modulates the desire for specific foods. Appetite 2008;51:34–41. doi:10.1016/j.appet.2007. 09.016.

ARTICLE IN PRESS 8

M. Ljubisavljevic et al. / Brain Stimulation ■■ (2016) ■■–■■

[27] Goldman RL, Borckardt JJ, Frohman HA, O’Neil PM, Madan A, Campbell LK, et al. Prefrontal cortex transcranial direct current stimulation (tDCS) temporarily reduces food cravings and increases the self-reported ability to resist food in adults with frequent food craving. Appetite 2011;56:741–6. doi:10.1016/ j.appet.2011.02.013. [28] Lapenta OM, Sierve KD, de Macedo EC, Fregni F, Boggio PS. Transcranial direct current stimulation modulates ERP-indexed inhibitory control and reduces food consumption. Appetite 2014;83:42–8. doi:10.1016/j.appet.2014.08.005. [29] Kekic M, McClelland J, Campbell I, Nestler S, Rubia K, David AS, et al. The effects of prefrontal cortex transcranial Direct Current Stimulation (tDCS) on food craving and temporal discounting in women with frequent food cravings. Appetite 2014;78:55–62. doi:10.1016/j.appet.2014.03.010. [30] Montenegro RA, Okano AH, Cunha FA, Gurgel JL, Fontes EB, Farinatti PTV. Prefrontal cortex transcranial direct current stimulation associated with aerobic exercise change aspects of appetite sensation in overweight adults. Appetite 2012;58:333–8. doi:10.1016/j.appet.2011.11.008. [31] Jauch-Chara K, Kistenmacher A, Herzog N, Schwarz M, Schweiger U, Oltmanns KM. Repetitive electric brain stimulation reduces food intake in humans. Am J Clin Nutr 2014;100:1003–9. doi:10.3945/ajcn.113.075481. [32] Jansen JM, Daams JG, Koeter MWJ, Veltman DJ, van den Brink W, Goudriaan AE. Effects of non-invasive neurostimulation on craving: a meta-analysis. Neurosci Biobehav Rev 2013;37:2472–80. doi:10.1016/j.neubiorev.2013. 07.009. [33] Cepeda-Benito A, Gleaves DH, Williams TL, Erath SA. The development and validation of the state and trait food-cravings questionnaires. Behav Ther 2000;31:151–73. doi:10.1016/S0005-7894(00)80009-X. [34] Foroni F, Pergola G, Argiris G, Rumiati RI. The FoodCast Research Image Database (FRIDa). Front Hum Neurosci 2013;7:51. doi:10.3389/fnhum.2013.00051. [35] Killgore WD, Young AD, Femia LA, Bogorodzki P, Rogowska J, Yurgelun-Todd DA. Cortical and limbic activation during viewing of high- versus low-calorie foods. Neuroimage 2003;19:1381–94. doi:10.1016/S1053-8119(03)00191-5. [36] Boswell RG, Kober H. Food cue reactivity and craving predict eating and weight gain: a meta-analytic review. Obes Rev 2015;17:159–77. doi:10.1111/obr.12354. [37] Fairburn CG, Beglin SJ. Assessment of eating disorders: interview or self-report questionnaire? Int J Eat Disord 1994;16:363–70. [38] Eaton WW, Ybarra M, Schwab J. The CESD-R is available on the web. Psychiatry Res 2012;196:161. doi:10.1016/j.psychres.2011.08.018. [39] Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1977;1:385–401. doi:10.1177/ 014662167700100306. [40] Cepeda-Benito A, Gleaves DH, Williams TL, Erath SA. The development and validation of the state and trait food-cravings questionnaires. Behav Ther 2000;31:151–73. [41] White MA, Whisenhunt BL, Williamson DA, Greenway FL, Netemeyer RG. Development and validation of the food-craving inventory. Obes Res 2002;10:107–14. doi:10.1038/oby.2002.17. [42] Moreno S, Rodríguez S, Fernandez MC, Tamez J, Cepeda-Benito A. Clinical validation of the trait and state versions of the Food Craving Questionnaire. Assessment 2008;15:375–87. doi:10.1177/1073191107312651. [43] White MA, Grilo CM. Psychometric properties of the Food Craving Inventory among obese patients with binge eating disorder. Eat Behav 2005;6:239–45. doi:10.1016/j.eatbeh.2005.01.001. [44] Gandiga PC, Hummel FC, Cohen LG. Transcranial DC Stimulation (tDCS): a tool for double-blind sham-controlled clinical studies in brain stimulation. Clin Neurophysiol 2006;117:845–50. doi:10.1016/j.clinph.2005.12.003. [45] Boggio PS, Rigonatti SP, Ribeiro RB, Myczkowski ML, Nitsche MA, Pascual-Leone A, et al. A randomized, double-blind clinical trial on the efficacy of cortical direct current stimulation for the treatment of major depression. Int J Neuropsychopharmacol 2008;11:249–54. doi:10.1017/S1461145707007833.A. [46] Nitsche MA, Cohen LG, Wassermann EM, Priori A, Lang N, Antal A, et al. Transcranial direct current stimulation: state of the art 2008. Brain Stimul 2008;1:206–23. doi:10.1016/j.brs.2008.06.004. [47] Vickers AJ, Altman DG. Statistics notes: analysing controlled trials with baseline and follow up measurements. BMJ 2001;323:1123–4. doi:10.1136/ bmj.323.7321.1123.

[48] Volkow ND, Wang G-J, Tomasi D, Baler RD. Obesity and addiction: neurobiological overlaps. Obes Rev 2013;14:2–18. doi:10.1111/j.1467789X.2012.01031.x. [49] Kaye WH, Wierenga CE, Bailer UF, Simmons AN, Wagner A, Bischoff-Grethe A. Does a shared neurobiology for foods and drugs of abuse contribute to extremes of food ingestion in anorexia and bulimia nervosa? Biol Psychiatry 2013;73:836–42. doi:10.1016/j.biopsych.2013.01.002. [50] Alonso-Alonso M. Translating tDCS into the field of obesity: mechanism-driven approaches. Front Hum Neurosci 2013;7:512. doi:10.3389/fnhum.2013.00512. [51] Kedzior KK, Reitz SK, Azorina V, Loo C. Durability of the antidepressant effect of the high-frequency repetitive transcranial magnetic stimulation (rTMS) in the absence of maintenance treatment in major depression: a systematic review and meta-analysis of 16 double-blind, randomized, sham-controlled trials. Depress Anxiety 2015;32:193–203. doi:10.1002/da.22339. [52] Concerto C, Lanza G, Cantone M, Ferri R, Pennisi G, Bella R, et al. Repetitive transcranial magnetic stimulation in patients with drug-resistant major depression: a six-month clinical follow-up study. Int J Psychiatry Clin Pract 2015;19:252–8. doi:10.3109/13651501.2015.1084329. [53] Concerto C, Al Sawah M, Chusid E, Trepal M, Taylor G, Aguglia E, et al. Anodal transcranial direct current stimulation for chronic pain in the elderly: a pilot study. Aging Clin Exp Res 2015;28:231–7. doi:10.1007/s40520-015-0409-1. [54] Olma MC, Dargie RA, Behrens JR, Kraft A, Irlbacher K, Fahle M, et al. Long-term effects of serial anodal tDCS on motion perception in subjects with occipital stroke measured in the unaffected visual Hemifield. Front Hum Neurosci 2013;7:314. doi:10.3389/fnhum.2013.00314. [55] Vestito L, Rosellini S, Mantero M, Bandini F. Long-term effects of transcranial direct-current stimulation in chronic post-stroke aphasia: a pilot study. Front Hum Neurosci 2014;8:785. doi:10.3389/fnhum.2014.00785. [56] Filho PRM, Vercelino R, Cioato SG, Medeiros LF, De Oliveira C, Scarabelot VL, et al. Transcranial direct current stimulation (tDCS) reverts behavioral alterations and brainstem BDNF levels increase induced by neuropathic pain model: long-lasting effect. Prog Neuropsychopharmacol Biol Psychiatry 2015;64:44–51. doi:10.1016/j.pnpbp.2015.06.016. [57] Podda MV, Cocco S, Mastrodonato A, Fusco S, Leone L, Barbati SA, et al. Anodal transcranial direct current stimulation boosts synaptic plasticity and memory in mice via epigenetic regulation of Bdnf expression. Sci Rep 2016;6:22180. doi:10.1038/srep22180. [58] Puri R, Hinder MR, Fujiyama H, Gomez R, Carson RG, Summers JJ. Durationdependent effects of the BDNF Val66Met polymorphism on anodal tDCS induced motor cortex plasticity in older adults: a group and individual perspective. Front Aging Neurosci 2015;7:107. doi:10.3389/fnagi.2015.00107. [59] Addolorato G, Leggio L, Hopf FW, Diana M, Bonci A. Novel therapeutic strategies for alcohol and drug addiction: focus on GABA, ion channels and transcranial magnetic stimulation. Neuropsychopharmacology 2012;37:163–77. doi:10.1038/npp.2011.216. [60] Sauvaget A, Trojak B, Bulteau S, Jiménez-Murcia S, Fernández-Aranda F, Wolz I, et al. Transcranial Direct Current Stimulation (tDCS) in behavioral and food addiction: a systematic review of efficacy, technical, and methodological issues. Front Neurosci 2015;9:349. doi:10.3389/fnins.2015.00349. [61] Fecteau S, Fregni F, Boggio PS, Camprodon JA, Pascual-Leone A. Neuromodulation of decision-making in the addictive brain. Subst Use Misuse 2010;45:1766–86. doi:10.3109/10826084.2010.482434. [62] Kuo M-F, Paulus W, Nitsche MA. Therapeutic effects of non-invasive brain stimulation with direct currents (tDCS) in neuropsychiatric diseases. Neuroimage 2014;85(Pt 3):948–60. doi:10.1016/j.neuroimage.2013.05.117. [63] Paulus W. Outlasting excitability shifts induced by direct current stimulation of the human brain. Suppl Clin Neurophysiol 2004;57:708–14. [64] Gluck ME, Alonso-Alonso M, Piaggi P, Weise CM, Jumpertz-von Schwartzenberg R, Reinhardt M, et al. Neuromodulation targeted to the prefrontal cortex induces changes in energy intake and weight loss in obesity. Obesity (Silver Spring) 2015;23:2149–56. doi:10.1002/oby.21313. [65] Truong DQ, Magerowski G, Blackburn GL, Bikson M, Alonso-Alonso M. Computational modeling of transcranial Direct Current Stimulation (tDCS) in obesity: impact of head fat and dose guidelines. NeuroImage Clin 2013;2:759– 66. doi:10.1016/j.nicl.2013.05.011.