Background: High frequency repetitive transcranial magnetic stimulation (HF-rTMS) has gained interest as a neuromodulation treatment technique for alcohol dependence. Single sessions of HF-rTMS have consistently shown to decrease craving for substances. However, the results of randomized controlled clinical trials investigating the effect of multiple HF-rTMS sessions in alcohol dependence on abstinence rates and craving are inconsistent. Furthermore, they lack information on the effect of HF-rTMS on cognition and brain functioning. Methods: A single center, single blind, randomized controlled trial with 80 abstinent alcohol dependent subjects in treatment randomized (1:1) to either treatment as usual (TAU) plus ten sessions of active HF-rTMS or TAU plus 10 sessions of placebo/ sham HF-rTMS will be performed. The effects of ten HF-rTMS sessions on craving and neurocognitive functions are obtained. In addition a subset of participants will undergo an MR scanning session before the first and after the last HF-rTMS session in order to investigate the effect of ten HF-rTMS sessions on brain functioning. The primary outcome is the continued abstinence rate after the add-on HF-rTMS treatment. Discussion: This study uses a randomized controlled trial to examine the clinical, neurocognitive and brain functioning effects of ten add-on HF-rTMS sessions in alcohol dependent individuals in treatment. If the add-on treatment is effective, this may add to the evidence needed for approval of this additional treatment method for alcohol dependence by regulatory authorities. Trial registration: The Netherlands National Trial Register (NTR), NTR5291, 6-July-2015. Keywords: Alcohol dependence, HF-rTMS, Abstinence, Craving, Neurocognitive, Neurobiological Background alcohol use disorders. However, these treatments are only Substance dependence is characterized by drug seeking and moderately effective and more than 50% of all treated drug use which persists despite negative social and health patients relapses within one year . In an attempt to im- consequences . In the Netherlands approximately 4–5% prove the treatment of substance dependence, non-invasive of the population is suffering from an alcohol use disorder neuromodulation has gained attention as a new potential . In 2014 more than 30.000 individuals were registered at treatment option [5, 6]. addiction treatment centers in the Netherlands with alcohol High-frequency (HF) repetitive transcranial magnetic as primary substance of abuse . Currently only psycho- stimulation (rTMS)  is one of several types of neuro- social and pharmacological treatments are available for modulation techniques. With this method, a magnetic field penetrates through the skull which can inhibit or activate neurons in the cortex. This magnetic field origi- * Correspondence: email@example.com nates from a coil wherein an alternating electric current Department of Psychiatry and Amsterdam Institute for Addiction Research, is running. The alternating current induces high inten- Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands sity magnetic pulses that pass the skull and generate an Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Schluter et al. BMC Psychiatry (2018) 18:169 Page 2 of 12 electric current in the neural tissue. By using this tech- functioning. We expect higher abstinence rates and de- nique the activity of the targeted cortical area is manipu- creased perceived craving together with improved func- lated [5, 8]. HF-rTMS became a popular investigational tioning on the neurocognitive tasks and brain functioning treatment tool for several psychiatric disorders because measures in the active HF-rTMS treated group compared of its non-invasiveness, tolerability and safety . with the sham HF-rTMS treated group. Individuals suffering from alcohol dependence often experience an intense and abnormal desire for alcohol, Aims of the study also known as craving . Furthermore it is known that This study aims to investigate the efficacy and working alcohol dependent individuals show impaired executive mechanisms of 10 add-on HF-rTMS sessions in a functions, such as diminished cognitive control, cogni- treatment-as-usual setup of alcohol dependence. tive flexibility and working memory [11, 12]. It is be- lieved that perceived craving combined with reduced Primary research question: cognitive control leads to problems in managing craving – What is the effect of ten sessions of active HF-rTMS and consequently relapse . on abstinence rates in alcohol dependent individuals On a neurobiological level craving is associated with in treatment, compared with ten sessions of sham heightened striatal activity related to addiction-relevant HF-rTMS? stimuli, whereas diminished cognitive control is associ- Secondary research questions: ated with decreased prefrontal activity . One of the – What is the effect of 10 sessions of active add-on areas involved in cognitive control is the dorsolateral HF-rTMS treatment on total amount of alcohol prefrontal cortex (DLPFC) . By stimulating the consumed after treatment in alcohol dependent indi- DLPFC with a high-frequency (10 Hz) protocol, the viduals, compared with 10 sessions of sham HF- neural activity of this area is enhanced  thereby im- rTMS? proving cognitive control [16, 17]. Furthermore, it is – What is the effect of 10 sessions of active add-on HF- known that the prefrontal cortex has abundant connec- rTMS treatment on days until first relapse after tions with the striatum . Indeed neuromodulation of treatment in alcohol dependent individuals, compared the DLPFC has shown to induce changes in neurotrans- with 10 sessions of sham HF-rTMS? mitter concentrations in the striatum  and can de- – What is the effect of 10 sessions of active HF-rTMS crease feelings of craving with a medium effect size [d = on perceived craving levels in alcohol dependent 0.48] (Jansen et al., 2013). The DLPFC therefore seems individuals in treatment, compared with 10 sessions an excellent target area for treating alcohol dependence of sham HF-rTMS? because it could enhance cognitive control functions Additional research questions: while also influencing striatal functioning and reducing – What is the effect of 10 sessions of active HF-rTMS feelings of craving . However, results of clinical trials on performance on neurocognitive tasks measuring investigating the effect of multiple HF-rTMS sessions on impulsivity, approach avoidance, spatial working craving are very scarce and inconsistent, with one study memory and compulsivity in alcohol dependent stimulating the right DLPFC that reported reduced crav- individuals in treatment, compared with 10 sessions ing  and one study stimulating the left DLPFC that of sham HF-rTMS? failed finding an effect on craving . Furthermore no – What is the effect of 10 sessions of active HF-rTMS studies thus far investigated the effect of multiple on brain functioning measures related to cognitive HF-rTMS sessions on abstinence rates after a longer control and craving in alcohol dependent individuals period of time. in treatment, compared with 10 sessions of sham Altogether the effect of multiple HF-rTMS sessions on HF-rTMS? abstinence rates, perceived craving and the neurocognitive and neurobiological working mechanisms are poorly in- vestigated (see also, ). The current study aims to eluci- Methods/ design date the effect of multiple sessions of HF-rTMS on Study design abstinence, craving, cognition and brain functioning. We The effectiveness of the HF-rTMS add-on treatment will will conduct a single center, single blind, randomized con- be tested in a parallel, single center, single blind trial in ab- trolled trial with 80 abstinent alcohol dependent subjects stinent alcohol dependent subjects, randomized (1:1) to in treatment randomized (1:1) to either treatment as usual either treatment as usual (TAU) plus 10 sessions of active (TAU) plus 10 sessions of active HF-rTMS or TAU plus HF-rTMS or TAU plus 10 sessions of sham HF-rTMS. A 10 sessions of placebo/ sham HF-rTMS. The effect of ac- subset of the participants will undergo a magnetic reson- tive versus sham HF-rTMS treatment will be investigated ance imaging (MRI) scan prior to the first and after the on measures of abstinence, craving, cognition and brain last stimulation session in order to investigate the effects Schluter et al. BMC Psychiatry (2018) 18:169 Page 3 of 12 of HF-rTMS on brain functioning (referred to as neuroim- sessions we will use a 70 mm double air film coil (Mag- aging study part). stim Co., UK) and a Magstim Rapid stimulator (Mag- stim Co., UK). The intervention will be applied by an Ethical considerations rTMS trained researcher. The training exists of brain This study is approved by the Medical Ethical Committee stimulation courses, practical rTMS tutorials, hands on of the Academic Medical Centre Amsterdam (2015_064) training and first aid and emergency response training. and is registered in The Netherlands Trial Register (NTR) During the stimulation session a predefined protocol will with trial number 5291. Written informed consent is be executed. obtained before screening for in and exclusion criteria The HF-rTMS intervention will be added to the treat- takes place. ment as usual (TAU) provided by the Jellinek addiction treatment centre. This treatment consists of an intensive In- and exclusion criteria 3–5 days per week program with group sessions of cog- All participants will be recruited when they are three to nitive behavioural therapy (CBT), emotion regulation four weeks abstinent, and are recruited from the Jellinek training, and motivational enhancement therapy. Besides Addiction Treatment Centre in Amsterdam, The these sessions, every participant has individual sessions Netherlands. Inclusion criteria are a recent (less than with a psychologist and a mentor every week. In the four months after detoxification) DSM-IV diagnosis of session with the psychologist, comorbidities and other alcohol dependence and age between 20 and 65. Exclu- problems of the patients that occur during treatment are sion criteria are (1) insufficient knowledge of the Dutch discussed. During the mentor session supportive CBT language, (2) Montreal Cognitive Assessment (MOCA) focussing on remaining abstinent is given. Finally, some score below 10, (3) current DSM-IV diagnosis of depres- patients receive pharmacotherapy. sion, schizophrenia or another psychotic disorder, (4) current recreational drug use, (5) rTMS contraindica- Procedure and data collection tions (such as a history of epileptic seizures, metal im- Inclusion procedure plants near the head, use of imipramine, amitriptyline, The participants will start their TAU at a clinical or day doxepine, nortriptyline, maprotiline, chlorpromazine, detoxification unit where they stay for 7–10 days, and clozapine, foscarnet, ganciclovir, ritonavir, amphetamines continue their treatment during an intensive 3–5 days ), and if applicable (for the neuroimaging study part) treatment of at least three weeks. During the beginning (6) MRI contraindications (such as metal implants or of their treatment participants are informed of the study claustrophobia). by the researcher. Individuals who are interested in participation are invited for an appointment to provide Intervention them with more information. If they are interested, in- The intervention exists of 10 HF-rTMS sessions of the formed consent is signed and the participant is screened right DLPFC (rDLPFC) on 10 consecutive workdays. for inclusion and exclusion criteria. If a patient meets all The HF-rTMS parameters of the active intervention are inclusion criteria and none of the exclusion criteria he/ 60 10 Hz trains of five seconds at 110% of the motor she will be included in the study by the researcher. After threshold . The coil will be oriented over the inclusion the patients can indicate whether they also rDLPFC with a horizontal angle of 45° relative to the want to participate in the neuroimaging study part. In nasion-inion midline . For the sham stimulation the that case a subsequent MR screening is performed. In stimulator will be set at the same settings, but the coil order to assure concealed randomization, participants will be tilted 90° relative to the skull . The rDLPFC are assigned to the sham or active stimulation group will be located at position F4 using the International 10– after inclusion, based on the stratification factors 20 EEG system . During the stimulation participants anti-craving medication (yes / no) and age (20–40 / 41– are situated on a comfortable chair with extra neck sup- 65) using variable block sizes (4, 6 and 8) of the random- port. All sessions are applied at the Jellinek addiction isation module implemented in the data management treatment centre in Amsterdam. One stimulation session system Castor EDC (Castor Electronic Data Capture, takes approximately 20–30 min. Ciwit BV, Amsterdam, The Netherlands, 2016). After The motor threshold will be determined at rest before randomization participants start with the research pro- the first and sixth stimulation session, using single pulse cedure described below. TMS in combination with Motor Evoked Potentials. The muscular (left abductor pollicis brevis) response will be Overview of study measured by visually observing a thumb muscular ab- For an overview of instruments, order of assessment and duction. Stimulus intensity will be adjusted until there is moment of assessment see Fig. 1 and Table 1. The first an abduction in five out of ten trials. For the rTMS test day takes approximately 4 h, the fifth and tenth test Schluter et al. BMC Psychiatry (2018) 18:169 Page 4 of 12 Fig. 1 Overview of the measures taken on different test days. Sample characteristics (age, handedness, educational level, medication use, WAIS digit span, NLV, MINI, substance use during life, alcohol use history), neurocognitive tasks (GNGT, AAT, DDT, SWMT, SST, IDED), questionnaires (AUDIT, AASE, BDI, BIS, UPPS, BIS/BAS, STAISTATE, PANAS), extensive craving assessment (AUQ, OCDS, VAS), brief craving assessment (VAS). All these measures will be explained in more detail in the outcomes and instruments section. Symbols: X = week before first rTMS session, XX = week after last rTMS session day take three hours and all the other assessments take any psychiatric DSM-IV diagnoses. The substance use 45 min. After three, six and 12 m a 30-min telephone during life questionnaire from the measurement in the follow-up interview will be held. addictions for triage and evaluation (MATE)  will be used to assess lifetime drug use. For a comprehensive as- Neuroimaging study part The outcome measures con- sessment of the alcohol use history an adapted version cerning brain functioning will be obtained in the week be- of the life time drinking history  will be used. fore the first stimulation session (i.e. to measure baseline brain functioning) and in the week after the last stimula- Primary outcome measure tion session (i.e. to measure the effect of HF-rTMS). Be- The primary outcome of the study will be the abstinence fore the participants enter the scanner they perform a rate after the add-on HF-rTMS treatment (in line with the urine alcohol and drug screening and practice the tasks guidelines of the European Medicines Agency ). This that will be conducted in the MRI scanner. will be defined as the number of abstinent days in the 180 days after the last stimulation session measured using Outcomes and instruments the Time Line Follow Back (TLFB)  at six months Sample characteristics follow up. General patient characteristics such as age, handedness, educational level and use of medication will be assessed. Secondary outcome measures Furthermore the digit span of the Wechsler Adult The secondary outcome measures of the study will be Intelligence Scale (WAIS) will be used to determine described below. working memory capacity . In addition the Dutch Version of the Adult Reading Test (NLV) will be used to – Total alcohol consumption after the add-on HF- assess premorbid intellectual functioning . The Mini rTMS treatment. This amount (g) of alcohol will be International Neuropsychiatric Interview (MINI)  calculated using the TLFB at three months and six will be used to determine whether the participant has months follow-up. Schluter et al. BMC Psychiatry (2018) 18:169 Page 5 of 12 Table 1 Overview of measurement instruments and moment of assessment during study. Symbols: X = MR session week before first rTMS session, XX = MR session week after last rTMS session, XXX = three months after last rTMS session, XXXX = six months after last rTMS session, XXXXX = twelve months after last rTMS session Test day X 1 2 3 4 5 6 7 8 9 10 XX XXX XXXX XXXXX WAIS � NLV � MATE-Q � MINI � Alcohol use history � TLFB �� � VAS � �� ��� �� �� � � � AUQ �� � � � � OCDS �� � � � � DDT �� � Go/No-go �� � SST �� � AAT �� � SWMT �� � IDED �� � Urine test � � Cue reactivity � � Stroop � � MIDT � � Resting state � � ASL � � AASE �� � � � � AUDIT � � BDI �� � BIS/BAS �� � BIS �� � UPPS �� � STAI/STATE �� � PANAS �� � Side effects �� ��� �� �� Alcohol use study � �� ��� �� �� – Days until the first relapse after the add-on HF- Additional clinical outcome measures rTMS treatment, wherein relapse is defined as a heavy drinking day. This holds more than 60 g Explorative alcohol use parameters The effect of alcohol per day for men and more than 40 g alcohol HF-rTMS on alcohol consumption or abstinence will be per day for women [31, 33]. This measure will be assessed using the following explorative outcome measures. assessed using the TLFB at three months follow up and six months follow-up. Individuals who do not – Full abstinence rate after the add- on HF-rTMS relapse will receive the highest score of 90 days. treatment, defined as the number of participants – Change in craving levels after the add-on HF-rTMS who did not consume any alcohol after the HF- treatment will be measured using the Alcohol Urge rTMS treatment. This will be assessed using the Questionnaire (AUQ)  assessed after the last ses- TLFB at six months follow-up. sion of the add-on HF-rTMS treatment, at three and – Treatment success based on drinking status at 12 m six months follow-up. follow up. The drinking status will be subdivided Schluter et al. BMC Psychiatry (2018) 18:169 Page 6 of 12 into the following categories: abstinence (defined as whether they were suffering from any side effects after no use of alcohol), non-excessive drinking (defined the previous HF-rTMS session. The reported side effects as 21 drinking days per 30 days, with a maximum of will be listed. 4 glasses per day) and excessive drinking (more than 21 drinking days per 30 days and more than 4 Neurocognitive measures glasses per day) . Participants will be assigned to Despite the role of the prefrontal cortex in higher cognitive one of the categories based on the last 30 days of processes , the effect of HF-rTMS over the rDLPFC on the 12 m follow up TLFB assessment. The categories neurocognitive measures in alcohol dependent individuals abstinence and non-excessive drinking will be is poorly investigated (see also ). Therefore the effect of defined as successful treatment. active HF-rTMS compared with sham HF-rTMS on several neurocognitive tasks will be assessed before the first, after Additional craving assessment The effect of HF-rTMS the fifth and tenth HF-rTMS session. Neurocognitive tasks on craving will also be assessed using the short (5-item) will focus on cognitive control and other relevant processes version of the Obsessive Compulsive Drinking Scale such as approach behavior and working memory. Partici- (OCDS) . This questionnaire will be assessed prior pants willbesituatedinfront of afingeroperatedtouch to the first, and after the fifth and tenth stimulation screen tablet (Hewlett-Packard; Windows 8.1) on which the session and at the 3, 6 and 12 months follow-up. tasks will be conducted. To assess the acute effect of HF-rTMS on craving two 100 mm visual analogue scale (VAS) ranging for 0 (not (1) Go/ No-go Task (GNGT): This task (adapted from at all) to 100 (very much) will be assessed prior and after ) measures the ability to refrain from action every stimulation session. Participants will be asked to initiation . Participants see a number (‘1’, ‘2’, ‘3’, indicate (1) their current desire to consume alcohol: ‘4’, ‘5’, ‘6’, ‘7’, ‘8’, ‘9’) projected on a screen in front of “How much do you feel like drinking alcohol right them. They are instructed to press the spacebar now?” and (2) their current urge to drink alcohol: “How with their right hand as soon as they see a number, strong is your urge to drink alcohol right now?” by but not if it is the number ‘3’ (=no-go trial) (Fig. 2a). drawing a line on a VAS scale. The outcome measure will be the number of commission errors, i.e. a response to a no-go trial, Questionnaires Questionnaires that will be filled in dur- reflecting action impulsivity. The higher the number ing the study are: of commission errors, the more impulsive an indi- vidual is. – Alcohol consumption burden measured with the (2) Approach Avoidance Task (AAT): This task Alcohol Use Disorders Identification Test (AUDIT) . (adapted from ) measures the bias towards – Confidence in abstaining from drinking as measured approaching alcoholic beverages. During this task with the Alcohol Abstinence Self Efficacy (AASE) . pictures from a validated dataset  of alcoholic – Depressive symptoms as measured with The Beck beverages, sodas and neutral objects (for example Depression Inventory (BDI) . scissors) are presented on a screen in front of the – Impulsive behavior will be assessed using Barratt participant. The pictures are rotated 3° towards the Impulsivity Scale-11 (BIS) , UPPS (urgency, pre- left or right, indicating whether a participant has to meditation, perseverance, sensation seeking)  pull or push a picture using a joystick (Fig. 2b). and Behavioral inhibition and Behavioral Activation When a picture is pulled or pushed the size of the systems (BIS/BAS)  questionnaires. picture increases or decreases respectively. The bias – Current state anxiety as measured with the state score per stimulus category [alcohol/ soda/ neutral] trait anxiety inventory (STAI STATE) . will be calculated by subtracting the median – Experience of positive and/ or negative affect reaction time of the approach (pull) trials from the measured with the Positive And Negative Affect median reaction time of the avoid (push) trials. Scale (PANAS) . When the result of this subtraction is positive, this indicates a relative faster approach compared to Supplementary measures In order to get an overview avoid, i.e. an approach bias. When the result of the of the drinking behavior during the testing period (vari- subtraction is negative this indicates a relatively able defined as: alcohol use study) participants are asked faster avoid compared to approach, i.e. avoid- bias. before every HF-rTMS session “whether (1) he/she con- (3) Delay Discounting Task (DDT): This task (based on sumed alcohol in the last 24 hours and if so (2) what ) measures the extent of impulsive decision kind of alcohol, (3) the alcohol percentage and (4) the making. Participants are presented with a choice amount consumed”. Furthermore participants are asked between an immediate (lower) and a delayed Schluter et al. BMC Psychiatry (2018) 18:169 Page 7 of 12 Fig. 2 Graphical representation of the neurocognitive measures. a Go/ No-go Task (GNGT), b Approach Avoidance Task (AAT), c Delay discounting task (DDT), d Stop signal task (SST), e Spatial Working Memory Task (SWMT), f Intra-dimensional/ Extra- dimensional Set Shift (IDED). Fig. E-F are adapted from the CANTAB instruction manual (higher) hypothetical monetary reward (Fig. 2c). the box on the screen. If no token is found the The value of the immediate reward varies across participant must continue its search until a token is the trials in one block, and depends on the found. When there is a blue token inside, the responses that are made . The outcome participant must now touch the black empty space to measure will be the area under the discounting fill up this space with the token. Now the participant curve (AUC), reflecting the degree of discounting has to begin a new search. The next token will only by delay . Impulsive choice behavior is indicated be hidden in a box that so far has been empty. This by a smaller AUC. procedure is repeated until all tokens are found and (4) Spatial Working Memory Task (SWMT): This task the entire empty space is filled with tokens. The task is part of the Cambridge Neuropsychological Test starts with three boxes, and this will increase to four, Automated Battery (CANTAB) test battery, and six and eight boxes. Touching a box where the token measures the ability of a subject to remember had already been found is considered an error. The spatial information and manipulate this using outcome measure will be the number of errors a working memory. During this task participants are subject makes and reflects working memory capacity. presented with a number of colored squares (or The more errors a subject makes, the lower the boxes for the participant) shown on a screen. The working memory capacity. participant is instructed to find a blue token hidden (5) Stop Signal Task (SST): This task (part of the in each box, and use this token to fill up the black CANTAB test battery) measures the ability of an empty space on the right side of the screen (Fig. 2d). individual to inhibit an ongoing action . In order to open a box, and see whether there is a Participants are presented with a white ring on a hidden token inside, the participant needs to touch black screen. In the ring a white arrow pointing Schluter et al. BMC Psychiatry (2018) 18:169 Page 8 of 12 either to the left or to the right appears. The ten blocks, subdivided into five alcohol and five participant needs to press the left button if the neutral blocks, presented alternatingly. Each blocks arrow points to the left and the right button when contains seven trials which each show one picture. the arrow points to the right as fast as possible (go Of these seven trials, six have a relevant content, and trial). During some trials the participant hears an one is a target, namely an animal. The participants auditory signal (beep), after the arrow appears, are instructed to thoroughly look at all the pictures, which indicates they have to stop their response and press the right button when they see an animal. and not press the button (stop trial) (Fig. 2e). The Before and after the task the participant will be asked outcome measure will be the amount of successful to rate “how much do you feel like drinking alcohol stops and reflects the capability of stopping an right now?” on a 10 point scale in which 1 indicates initiated response. The lower the number of “not at all” and 10 indicates “very much” . We successful stops, the more impulsive an individual is. will compare neuronal activity during watching (6) Intra-dimensional/ Extra- dimensional Set Shift alcohol pictures with neuronal activity during (IDED): This task (part from the CANTAB test watching neutral pictures. battery) tests rule acquisition and reversal learning. (2) Stroop task: This task (adapted from ) measures Two stimuli are presented on a screen from which the neural response during watching two types of the participant has to choose one by pressing on stimuli, congruent or incongruent. All trials contain the screen. These stimuli are made up of two one of the following words in Dutch: “red”, “blue”, artificial dimensions: color-filled shapes and/or “yellow”, “green”. In congruent trials the color of the white lines. The stimuli presented can be either word is the same as the content of the word, while simple (just one of the two dimensions) or compound during incongruent trials the color of the word is (stimuli contain both the dimensions). The task different from the content of the word. Participants contains 9 blocks increasing in difficulty (Fig. 2f). are instructed to indicate the color of the word by After pressing the stimulus on the screen the pressing on the representative button . We will computer gives feedback on whether this was the compare neuronal activity during the incongruent correct stimulus. In this way participants can learn trials with the congruent trials. the task rules. After six correct responses (learning (3) Monetary Incentive Delay Task (MIDT): This task criterion) the program changes the rule, and thereby adapted from [55, 56] measures the neural response the participant reaches the next block. If the during the anticipation of a reward. Participants are participant does not reach the learning criterion the presented with cues (blue triangle or blue circle) test terminates after 50 trials. The outcome measures that indicate whether they could earn €0.01 or will be the number of trials needed to reach the next €0.50. The cue is followed by a target (green star). stage (indicating the rule learning capacity), and the The participant is instructed to press the right number of errors made after a rule change (indicating button as soon as possible when they see the target. the capacity of reversal learning) . Higher If the participant is fast enough, he/she earns the numbers indicate lower rule learning and reversal amount of money. This is communicated to the learning capacity. participant through a feedback screen in which both the amount of earned money as well as the Neuroimaging measures total amount of money is presented. We will Because the effect of HF- rTMS over the rDLPFC of alco- compare the neuronal response during the hol dependent individuals on brain functioning is poorly anticipation of the high reward with activity during investigated , several cognitive tasks will be conducted the anticipation of the low reward. while functional Magnetic Resonance Imaging (fMRI) is (4) Resting state task: During this scan (adopted from performed. Furthermore, as a control measure, arterial ) the neuronal activity of the resting state network spin labelling (ASL) MRI will be performed in order to will be measured. Participants are presented with a investigate changes in cerebral blood flow . MRI black screen and instructed to close their eyes, not scanning will be performed on a 3.0-Tesla Intera full-body think of something in particular, just let their minds scanner (Philips Medical Systems, Best, the Netherlands) wander and try not to fall asleep . Functional with a 32 channel sense head coil located near the connectivity of the rDLPFC will be determined. Academic Medical Centre in Amsterdam. Statistical analyses (1) Cue reactivity task: This task (adapted from ) Power analysis measures the neural response during presentation The current study is the first to investigate the effect of of alcohol and neutral pictures. The task consists of multiple HF-rTMS sessions on abstinence rates measured Schluter et al. BMC Psychiatry (2018) 18:169 Page 9 of 12 six months after the last stimulation session. To the best Secondary outcome measures of our knowledge no studies thus far have investigated the An appropriate test for comparison of two groups will effect of rTMS on abstinence after a longer - clinically be used depending on the distribution of the outcome more relevant - period of time. Therefore no effect size is measures ‘days until first relapse’ and ‘total alcohol known or can be calculated, which makes it impossible to consumption’. perform a scientifically correct power analysis. However, Multilevel models (capable of handling missing data to estimate how many participants we would need, we points) with predictors time, treatment and time X treat- based our population estimate on a previous study with a ment interaction will be used to test for the effect of the similar design , which used craving as the outcome of HF-rTMS add-on treatment on craving. The following interest. Although our secondary outcome measure is time points will be included: post HF-rTMS, 3 months craving, we hypothesized that we should at least include a follow up and 6 months follow up. similar number of participants (30 participants in the ac- tive stimulation group) as . Given a drop-out percent- Additional clinical outcome measures age of 10%, and an estimated lower effect on behavioral measures compared to craving, the current study will in- Explorative alcohol use parameters An appropriate clude 38–40 alcohol dependent individuals per group, test for comparison of two groups will be used depend- resulting in a total of approximately 80 participants. ing on the distribution of the outcome measure ‘full ab- stinence rate’. An appropriate test for comparing a categorical variable between two groups will be used for Neuroimaging study part For the neuroimaging the outcome measure ‘treatment success’. sub-study no information about the effect-size is avail- able. However, studies of the statistical properties of one Additional craving assessment Multilevel models (cap- large fMRI cohort found that the sensitivity and repro- able of handling missing data points) with predictors ducibility of group analyses reaches a plateau at N =27 time, treatment and time X treatment interaction will be . With this sample size, the proportion of correct used to test for the effect of the HF-rTMS add-on treat- classification of truly active and inactive voxels corrected ment on craving measured with the OCDS. The follow- for chance is k > 0.75. In line, this sample size is also ing time points will be included: post HF-rTMS, similar to the sample size that is typically used for neu- 3 months follow up and 6 months follow up. Data of the roimaging studies that compare psychiatric patients with VAS scales will be analyzed by calculating difference healthy controls (N =20–30) . Therefore 28 partici- scores between pre and post HF-rTMS. These scores pants of each group will be included in the neuroimag- will be used in a multilevel model (capable of handling ing study part. missing data points) with predictors time, treatment and time X treatment interaction to test for the effect of the Descriptive statistics HF-rTMS add-on treatment. Descriptive analyses will be performed in order to see whether randomization resulted in two research groups Questionnaires Multilevel models (capable of handling with a similar distribution of demographic factors. Appro- missing data points) with predictors time, treatment and priate parametric and non-parametric statistical tests will time X treatment interaction will be used to test for the be used to analyze descriptive statistics, and if required, effect of the HF-rTMS add-on treatment on question- multiple comparison corrections will be performed. naire scores. Total scores of baseline, intermediate and post-rTMS data will be compared. Primary outcome measure Supplementary measures These measures will be com- Data of the primary outcome measure will be analyzed in pared between groups using an appropriate statistical test accordance with the intention-to-treat principle. Missing depending on the distribution of the outcome measure. data of subjects that are randomized and received at least one stimulation session will be imputed in order to achieve complete datasets. Subsequently an appropriate Neurocognitive measures test for comparing two groups will be used depending on Multilevel models (capable of handling missing data the distribution of the outcome measure. Additionally, points) with predictors time, treatment and time X treat- data of the primary outcome measure will be analyzed ment interaction will be used to test for the effect of the with only treatment completers (according to the per HF-rTMS add-on treatment on neurocognitive measures. protocol analysis), again using an appropriate test depend- The following time points will be included: baseline, inter- ing on the distribution of the outcome measure. mediate and post HF-rTMS. Schluter et al. BMC Psychiatry (2018) 18:169 Page 10 of 12 Neuroimaging measures individuals, chances are high that participants will get The software package Statistical Parametric Mapping lost at follow-up . There is a risk that we may not (SPM) (Wellcome Department of Cognitive Neurology, get the primary outcome measure for all participants, London, UK) will be used to analyze the task induced because it is uncertain whether we can reach all partici- (cue reactivity, stroop, MIDT) activation patterns. First, pants at six months follow-up. However, our research images will be preprocessed, including motion correc- group is experienced with a six months follow-up and tion, normalization to correct for individual differences no main problems have been reported in reaching in anatomy and smoothing. Thereafter, individual subject participants. analyses will be performed within the context of the The first strength of this study is that it investigates the General Linear Model (GLM), using delta functions con- effect of multiple HF-rTMS session. The second strength volved with a synthetic hemodynamic response function of this study is that it takes into account several aspects to model events of interest. Contrast images containing that are important in the treatment of alcohol depend- parameter estimates for each comparison of interest will ence. So far studies mainly looked at the effect of be entered into second level analyses to assess baseline HF-rTMS on self-reported craving, although the most versus HF-rTMS effects. For the resting state activation relevant clinical question is whether HF-rTMS has an ef- spatiotemporal independent component analysis will be fect on abstinence . This study will perform follow-up employed. measurements to assess the number of abstinent days in ASL scans will be analyzed using the Explore ASL tool- the six months after the last neurostimulation session. box in Matlab. First, the T1 images are normalized and Furthermore this study will elucidate the underlying segmented into grey and white matter. The probability mechanism by which rTMS may induce its effects in alco- maps and the gray matter tissue probability maps are then hol dependent individuals by investigating several neuro- spatially normalized using the Diffeomorphic Anatomical cognitive as well as neuroimaging measures . Registration analysis using Exponentiated Lie algebra The limitation of this study is that the trial is not (DARTEL) algorithm . Then, for the ASL time series, double blind controlled. However, a double blind con- motion estimation is performed, as well as exclusion of trolled rTMS study with a sham condition similar to this frames with motion spikes. Subsequently, label and con- study is impossible because the researcher needs to tilt trol images are subtracted and corrected for slice gradi- the coil 90° relative to the skull . Because this study ents. After this, the perfusion weighted images are only uses participants self-report measures and the re- registered to the gray matter tissue probability maps of searcher does not score any clinical effects, the outcome each subject using 6 parameter rigid body registrations, measures are not affected by the knowledge of the re- followed by voxel-based outlier rejection. Then, cerebral searcher, and a double blind paradigm is not necessary. blood flow images will be quantified with the following If this study reveals higher abstinence rates and de- parameters: post-labeling delay = 1525 ms, T1 = creased craving in the active stimulation group com- arterial 1650 ms, labeling efficiency α = 0.8, labeling duration τ = pared with the sham stimulation group, and the active 1650 ms . Finally, all transformations will be mathem- HF-rTMS induces negligible side effects, this may lead atically combined in a single B-spline interpolation and to larger clinical trials. If most of these trials find posi- applied to the CBF maps. This results in an average perfu- tive results of multiple HF-rTMS sessions on treatment sion image per participant which will be used for the outcomes, this eventually could result in approval by the assessment of baseline versus HF- rTMS effects. regulatory authorities as additional treatment method for substance dependence, just as for instance the Food Discussion and Drug Administration has approved HF-rTMS for This paper presents a single blind randomized clinical the treatment of depression. trial protocol investigating whether 10 sessions of active Abbreviations HF-rTMS compared with 10 sessions of sham HF-rTMS AAT: Approach avoidance task; ASL: Arterial Spin Labeling; AUDIT: Alcohol improve the treatment outcomes of alcohol dependence. use disorders identification test; AUQ: Alcohol urge questionnaire; BDI: Beck depression inventory; BIS: Behavioral inhibition system; BIS/BAS: Behavioral The aim of this study is to increase abstinence rates, de- inhibition and behavioral activation systems; CANTAB: Cambridge crease craving, and improve neurocognitive and brain neuropsychological test automated battery; CBT: Cognitive behavioral functioning measures relevant for alcohol dependence therapy; DDT: Delay discounting task; DLPFC: Dorsolateral prefrontal cortex; DSM: Diagnostic and statistical manual of mental disorders; fMRI: Functional treatment. magnetic resonance imaging; GNGT: Go/ No-go task; HF: High frequency; The main challenge of this study will be completion of IDED: Intra-dimensional/ extra- dimensional Set Shift; MATE: Measurement in the entire follow-up procedure. Participants will be the addictions for triage and evaluation; MIDT: Monetary incentive delay task; MINI: Mini international neuropsychiatric interview; MOCA: Montreal called three, six and twelve months after finishing the cognitive assessment; MRI: Magnetic resonance imaging; NLV: Dutch version last stimulation session to assess the number of abstin- of the adult reading test; NTR: Netherlands national trial register; ent days. Within the population of alcohol dependent OCDS: Obsessive compulsive drinking scale; PANAS: Positive and negative Schluter et al. BMC Psychiatry (2018) 18:169 Page 11 of 12 affect scale; rDLPFC: Right dorsolateral prefrontal cortex; rTMS: Repetitive 11. Goudriaan AE, Oosterlaan J, De Beurs E, Van Den Brink W. Neurocognitive transcranial magnetic stimulation; SPM8: Statistical parametric mapping 8; functions in pathological gambling: a comparison with alcohol dependence, SPSS: Statistical package for the social sciences; SST: Stop signal task; STAI- Tourette syndrome and normal controls. Addiction. 2006;101:534–47. STATE: State trait anxiety inventory; SWMT: Spatial working memory task; 12. Schulte MHJ, Cousijn J, den Uyl TE, Goudriaan AE, van den Brink W, Veltman TAU: Treatment as usual; TLFB: Time line follow back; UPPS: Urgency, DJ, et al. Recovery of neurocognitive functions following sustained premeditation, perseverance, sensation seeking; VAS: Visual analogue scale; abstinence after substance dependence and implications for treatment. Clin WAIS: Wechsler adult intelligence scale Psychol Rev Elsevier Ltd. 2014;34:531–50. 13. Koob GF, Volkow ND. Neurocircuitry of addiction. Neuropsychopharmacology Nature Publishing Group. 2010;35:217–38. Funding This research is financed by a VIDI grant from the Netherlands Organization for 14. Cole MW, Schneider W. The cognitive control network: integrated cortical Scientific Research (NWO) (grant number 016–136-354 awarded to AEG). The regions with dissociable functions. NeuroImage. 2007;37:343–60. funding body did not play any role in the design of the study, nor will play a 15. Rossi S, Hallett M, Rossini PM, Pascual-Leone A. Safety, ethical role in data collection, analysis, and interpretation of data and in writing of considerations, and application guidelines for the use of transcranial manuscripts. magnetic stimulation in clinical practice and research. Clin Neurophysiol. 2009;120:323–30. 16. Herremans SC. Vanderhasselt M a., De Raedt R, Baeken C. Reduced intra- Authors’ contributions individual reaction time variability during a go-nogo task in detoxified RSS drafted the manuscript, which was modified by RJvH and AEG. All authors alcohol-dependent patients after one right-sided dorsolateral prefrontal HF- participated in the design of the study. All authors read and approved the final rTMS session. Alcohol Alcohol. 2013;48:552–7. manuscript. 17. Del Felice A, Bellamoli E, Formaggio E, Manganotti P, Masiero S, Cuoghi G, et al. Neurophysiological, psychological and behavioural correlates of rTMS Ethics approval and consent to participate treatment in alcohol dependence. Drug Alcohol Depend Elsevier Ireland This study is approved by the Medical Ethical Committee of the Academic Ltd. 2016;158:147–53. Medical Centre Amsterdam (2015_064). Written informed consent is obtained 18. Morein-zamir S, Robbins TW. Fronto-striatal circuits in response-inhibition : before screening for in and exclusion criteria takes place. relevance to addiction. Brain Res Elsevier. 2014;1628:1–13. 19. Strafella AP, Paus T, Barrett J, Dagher A. Repetitive transcranial magnetic Competing interests stimulation of the human prefrontal cortex induces dopamine release in the The authors declare that they have no competing interests. caudate nucleus. J Neurosci. 2001;21:1–4. 20. Mishra BR, Nizamie SH, Das B, Praharaj SK. Efficacy of repetitive transcranial magnetic stimulation in alcohol dependence: a sham-controlled study. Publisher’sNote Addiction. 2010;105:49–55. Springer Nature remains neutral with regard to jurisdictional claims in published 21. Höppner J, Broese T, Wendler L, Berger C, Thome J. Repetitive transcranial maps and institutional affiliations. magnetic stimulation (rTMS) for treatment of alcohol dependence. World J Biol Psychiatry. 2011;12(Suppl 1):57–62. Author details 22. Salling MC, Martinez D. Brain stimulation in addiction. Department of Psychiatry and Amsterdam Institute for Addiction Research, Neuropsychopharmacology. Nature Publishing Group. 2016;41:1–37. Academic Medical Center, University of Amsterdam, Amsterdam, The 23. Jansen JM, Van Wingen G, Van Den Brink W, Goudriaan AE. Resting Netherlands. Donders Institute for Cognition, Brain and Behaviour, Radboud state connectivity in alcohol dependent patients and the effect of University, Kapittelweg 29, 6525 EN Nijmegen, Gelderland, The Netherlands. repetitive transcranial magnetic stimulation. Eur Neuropsychopharmacol Arkin Mental Health Care, Klaprozenweg 111, 1033 NN Amsterdam, Elsevier. 2015;25:1–10. Noord-Holland, The Netherlands. 24. Thomson RH, Cleve TJ, Bailey NW, Rogasch NC, Maller JJ, Daskalakis ZJ, et al. Blood oxygenation changes modulated by coil orientation during prefrontal Received: 8 November 2017 Accepted: 22 May 2018 transcranial magnetic stimulation. Brain Stimul. 2013;6:576–81. 25. Herwig U, Satrapi P, Schönfeldt-Lecuona C. Using the international 10-20 EEG system for positioning of transcranial magnetic stimulation. Brain References Topogr. 2003;16:95–9. 1. Baler RD, Volkow ND. Drug addiction: the neurobiology of disrupted self- 26. Benson N, Hulac DM, Kranzler JH. Independent examination of the Wechsler control. Trends Mol Med. 2006;12:559–66. adult intelligence scale-fourth edition (WAIS-IV): what does the WAIS-IV 2. De Graaf R, Ten Have M, Van Gool C, Van Dorsselaer S. Prevalence of mental measure? Psychol Assess. 2010;22:121–30. disorders and trends from 1996 to 2009. Results from the Netherlands 27. SchmandB,Bakker D,Saan R,LoumanJ. The Dutch reading test for mental health survey and incidence Study-2. Soc Psychiatry Psychiatr adults: a measure of premorbid intelligence level. Tijdschr Gerontol Epidemiol. 2012;47:203–13. Geriatr. 1991;22:15–9. 3. van Laar MW, van Ooyen-Houben MMJ, Cruts AAN, Meijer RF, Croes EA, 28. Sheenhan D, Lecrubier Y, Sheehan H, Amorim P, Janavs J, Weiller E, et al. Ketelaars APM, et al. Nationale drug monitor. In: 2015; 2015. The Mini-International Neuropsychiatric Interview (MINI): The Development 4. Oudejans SCC, Schippers GM, Spits ME, Stollenga M, van den Brink W. Five and Validation of a Structured Diagnostic Psychiatric Interview for DSM-IV years of ROM in substance abuse treatment centres in the Netherlands. and ICD-10; 1998. p. 22–33. Tijdschr Psychiatr. 2012;54:185–90. 29. Schippers GM, Broekman TG, Buchholz A, Koeter MWJ, Van Den Brink W. 5. Bellamoli E, Manganotti P, Schwartz RP, Rimondo C, Gomma M, Serpelloni Measurements in the addictions for triage and evaluation (MATE): an G. rTMS in the treatment of drug addiction: an update about human instrument based on the World Health Organization family of international studies. Behav Neurol. 2014;2014:815215. classifications. Addiction. 2010;105:862–71. 6. Jansen JM, Daams JG, Koeter MWJ, Veltman DJ, Van Den Brink W, Goudriaan 30. H a S, Sheu W. Reliability of alcohol use indices: the lifetime drinking history AE. Effects of non-invasive neurostimulation on craving: a meta-analysis. and the MAST. J Stud Alcohol. 1982;43:1157–70. Neurosci Biobehav Rev Elsevier Ltd. 2013;37:2472–80. 31. European Medicines Agency. Guideline on the development of medicinal 7. Barker AT, Jalinous R, Freeston IL. Non-invasive magnetic stimulation of products for the treatment of alcohol dependence, vol. 2008; 2010. p. 1–17. human motor cortex. Lancet. 1985;325:1106–7. 32. Sobell L, Sobell M. Timeline follow-back: A technique for assessing self- 8. Guse B, Falkai P, Wobrock T. Cognitive effects of high-frequency repetitive reported alcohol consumption. Meas Alcohol Consum. 1992;17:41–72. transcranial magnetic stimulation: a systematic review. J Neural Transm. 2010;117:105–22. 33. Rehm J. How should prevalence of alcohol use disorder be assessed 9. Lisanby SH, Kinnunen LH, Crupain MJ. Applications of TMS to Therapy in globally? Int J Methods Psychiatr Res. 2016;18:69–83. Psychiatry, vol. 19; 2002. p. 344–60. 34. Drummond DC, Phillips TS. Alcohol urges in alcohol-dependent drinkers: 10. Sinha R. The clinical neurobiology of drug craving. Curr Opin Neurobiol further validation of the alcohol urge questionnaire in an untreated Elsevier Ltd. 2013;23:649–54. community clinical population. Addiction. 2002;97:1465–72. Schluter et al. BMC Psychiatry (2018) 18:169 Page 12 of 12 35. de JM WW a, Lehert P, Schippers GM, Nakovics H, Mann K, van den Brink W. 59. G a v W, van Eijndhoven P, Tendolkar I, Buitelaar J, Verkes RJ, Fernández G. Investigating the structure of craving using structural equation modeling in Neural basis of emotion recognition deficits in first-episode major analysis of the obsessive-compulsive drinking scale: a multinational study. depression. Psychol Med. 2011;41:1397–405. Alcohol Clin Exp Res. 2005;29:509–16. 60. Ashburner J. A fast diffeomorphic image registration algorithm. 36. Conigrave KM, Hall WD, Saunders JB. The AUDIT questionnaire: choosing a NeuroImage. 2007;38:95–113. cut-off score. Addiction. 1995;90:1349–56. 61. Alsop DC, Detre JA, Golay X, Matthias G, Hendrikse J, Hernandez-garcia L, et al. Recommended implementation of arterial spin-labeled perfusion MRI for 37. Sherer M, Maddux JE, Mercandante B, Prentice-Dunn S, Jacobs B, Rogers clinical Applications : a consensus of the ISMRM perfusion study group and RW. The self-efficacy scale: construction and Validation1. Psychol. Reports @ the European consortium for ASL in Dementia. Magn Reson Med. 2014;0 Psychol Report. 1982;51:663–71. 62. Hunt GE, Siegfried N, Morley K, Sitharthan T, Cleary M. Psychosocial 38. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for interventions for people with both severe mental illness and substance measuring depression. Arch Gen Psychiatry. 1961;4:561–71. misuse. Schizophr Bull. 2014;40:18–20. 39. Patton JH, Stanford MS, Barratt ES. Factor structure of the barratt impulsiveness scale. J Clin Psychol. 1995;51:768–74. 40. Whiteside SP, Lynam DR, Miller JD, Reynolds SK. Validation of the UPPS impulsive behaviour scale: a four-factor model of impulsivity. Eur J Pers. 2005;19:559–74. 41. Carver CS, White TL. Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: the BIS/BAS scales. 1994. 42. Spielberger CD. State-trait anxiety inventory. Anxiety. 1987;19:2009. 43. Watson D, Clark L a, Tellegen a. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol. 1988;54:1063–70. 44. Miller EK, Cohen JD. An integrative theory of prefrontal cortex function; 2001. p. 167–202. 45. Durston S, Tottenham NT, Thomas KM, Davidson MC, Eigsti IM, Yang Y, et al. Differential patterns of striatal activation in young children with and without ADHD. Biol Psychiatry. 2003;53:871–8. 46. Hamilton KR, Littlefield AK, Anastasio NC, Cunningham KA, Fink LHL, Wing VC, et al. Rapid-response Impulsivity : definitions , Measurement Issues , and Clinical Implications. 2015;6:168–81. 47. Cousijn J, Goudriaan AE, Wiers RW. Reaching out towards cannabis: approach-bias in heavy cannabis users predicts changes in cannabis use. Addiction. 2011;106:1667–74. 48. Pronk T, van Deursen DS, Beraha EM, Larsen H, Wiers RW. Validation of the Amsterdam beverage picture set: a controlled picture set for cognitive Bias measurement and modification paradigms. Alcohol Clin Exp Res. 2015;39:2047–55. 49. Wittmann M, Leland DS, Paulus MP. Time and decision making: differential contribution of the posterior insular cortex and the striatum during a delay discounting task. Exp Brain Res. 2007;179:643–53. 50. Myerson J, Green L, Warusawitharana M. Area under the curve as a measure of discounting. J Exp Anal Behav. 2001;76:235–43. 51. Jazbec S, Pantelis C, Robbins T, Weickert T, Weinberger DR, Goldberg TE. Intra-dimensional/extra-dimensional set-shifting performance in schizophrenia: impact of distractors. Schizophr Res. 2007;89:339–49. 52. Grade M, Hernandez Tamames JA, Pizzini FB, Achten E, Golay X, Smits M. A neuroradiologist’s guide to arterial spin labeling MRI in clinical practice. Neuroradiology. 2015;57:1181–202. 53. Goudriaan AE, Veltman DJ, Van Den Brink W, Dom G, Schmaal L. Neurophysiological effects of modafinil on cue-exposure in cocaine dependence: a randomized placebo-controlled cross-over study using pharmacological fMRI. Addict Behav Elsevier Ltd. 2013;38:1509–17. 54. Schmaal L, Goudriaan AE, Joos L, Krüse AM, Dom G, Van Den Brink W, et al. Modafinil modulates resting-state functional network connectivity and cognitive control in alcohol-dependent patients. Biol Psychiatry. 2013;73:789–95. 55. Knutson B, Adams CM, Fong GW, Hommer D. Anticipation of Increasing Monetary Reward Selectively Recruits Nucleus Accumbens. J Neurosci. 2001; 21:1–5. 56. Ossewaarde L, van Wingen G, Kooijman SC, Backstrom T, Fernandez G, Hermans EJ. Changes in functioning of mesolimbic incentive processing circuits during the premenstrual phase. Soc Cogn Affect Neurosci. 2011;6:612–20. 57. Birn RM, Molloy EK, Patriat R, Parker T, Meier TB, Kirk GR, et al. The effect of scan length on the reliability of resting-state fMRI connectivity estimates. Neuroimage. Elsevier BV. 2013;83:550–8. 58. Thirion B, Pinel P, Mériaux S, Roche A, Dehaene S, Poline JB. Analysis of a large fMRI cohort: statistical and methodological issues for group analyses. NeuroImage. 2007;35:105–20.
– Springer Journals
Published: Jun 4, 2018