State-Dependent Effects of Prefrontal Repetitive Transcranial Magnetic Stimulation on Emotional Working Memory

Contents lists available at SciVerse ScienceDirect

Brain Stimulation

journal homepage: www.brainst imjrnl .com

Brain Stimulation 6 (2013) 905e912

State-Dependent Effects of Prefrontal Repetitive Transcranial MagneticStimulation on Emotional Working Memory

Anne Weigand a,b,c,*,1, Aline Richtermeier a,c,1, Melanie Feeser a,b,c, Jia Shen Guo a,c,Benny B. Briesemeister a,c, Simone Grimma,b,c,d, Malek Bajbouj a,b,c

a Cluster of Excellence “Languages of Emotion”, Freie Universitaet Berlin, Habelschwerdter Allee 45, 14195 Berlin, GermanybDepartment of Psychiatry, Campus Benjamin Franklin, Charité Berlin, 14050 Berlin, GermanycDahlem Institute for Neuroimaging of Emotion, Freie Universitaet Berlin, 14195 Berlin, GermanydClinic for Affective Disorders and General Psychiatry, Psychiatric University Hospital Zurich, 8032 Zurich, Switzerland

a r t i c l e i n f o

Article history:Received 22 February 2013Received in revised form20 June 2013Accepted 30 June 2013Available online 6 August 2013

Keywords:rTMStDCSEmotionWorking memoryFrontal asymmetry

Conflicts of interest: The authors declare that therrelation to this article.* Corresponding author. Tel.: þ49 (0)30 838 57863;

E-mail addresses: [email protected], atwei1 Anne Weigand and Aline Richtermeier contributed

1935-861X/$ e see front matter � 2013 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.brs.2013.06.004

a b s t r a c t

Background: A growing body of findings illustrates the importance of state-dependency in studies usingbrain stimulation.Objective: We aimed to investigate the effects of tDCS priming followed by rTMS applied over the rightdorsolateral prefrontal cortex (DLPFC) on emotional working memory.Methods: In a randomized single-blind within-subjects design, participants performed an emotional3-back task at baseline and after tDCS priming (anodal, cathodal) and subsequent low-frequency rTMS(active, sham) of the right DLPFC. Stimuli consisted of words related to the distinct emotion categoriesfear and anger as well as neutral words.Results: Task accuracy increased for fear-related words and decreased for neutral words across stimu-lation conditions. No general state-dependent effects of prefrontal rTMS on working memory werefound. We further showed a detrimental effect of negative emotional content on working memoryperformance.Conclusions: Our findings support a hemispheric lateralization of emotion processing by demonstratingthat the withdrawal-related emotion fear is associated with the right DLPFC and contribute to clarifyingthe interaction between working memory and emotion.

� 2013 Elsevier Inc. All rights reserved.

Introduction

Growing evidence indicates that effects of brain stimulationstrongly depend on the level of neuronal activity at the time ofstimulation [1]. However, most studies take no account of the initialstate in the stimulated brain region. It has been argued that this lackof control is likely to be one of the key factors explaining incon-sistencies in reported stimulation effects, including clinical treat-ment trials [2,3].

State-dependent effects of repetitive transcranial magneticstimulation (rTMS) have extensively been explored in the motorcortex by measuring motor-evoked potentials (MEP) [4]. Interest-ingly, it has been demonstrated that priming with transcranial

e is no conflict of interest in

fax: þ49 (0)30 838 [email protected] (A.Weigand).equally to the manuscript.

ll rights reserved.

direct current stimulation (tDCS) reversed the expected effects oflow-frequency rTMS applied to the primary motor hand area (M1):While excitatory anodal tDCS followed by 1 Hz rTMS resulted inreduced cortical excitability, inhibitory cathodal tDCS followed by1 Hz rTMS produced increased cortical excitability [5]. Similarconverse changes in MEP amplitudes after tDCS priming werefound for subsequent high-frequency rTMS [2,6]. Further experi-mental data from the motor cortex demonstrated that high-frequency rTMS can also be used as a priming tool, enhancing thesuppressive effect of subsequent low-frequency rTMS [7].

In agreement with these findings derived from motor cortexphysiology, clinical evidence revealed that preceding high-frequency rTMS improves the efficacy of low-frequency rTMSapplied to the right dorsolateral prefrontal cortex (DLPFC) in thetreatment of depression [8,9]. Cathodal tDCS priming followed byhigh-frequency rTMS applied to the left DLPFC did not result ingreater antidepressant effects, although one patient respondedwith a highly improved clinical outcome [10].

The importance of state-dependent stimulation effects hasadditionally been illustrated in perceptual and cognitive studies

Figure 1. Schematic illustration of the experimental design. Emotional n-back (EMO-BACK) task performance was measured at baseline and after tDCS priming (anodal/cathodal) followed by low-frequency rTMS (active/sham) applied over the right DLPFC.A standardized time interval was implemented between subsequent stimulations.

Table

1Stim

ulusch

aracteristicsusedin

thefourparallelv

ersion

sof

theem

otional

workingmem

orytask.

Set1

Set2

Set3

Set4

Fear

Ange

rNeu

tral

Fear

Ange

rNeu

tral

Fear

Ange

rNeu

tral

Fear

Ange

rNeu

tral

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

M(SD)

Valen

ce�1

.2(1.0)

�1.4

(0.8)

0.0(0.2)

�1.2

(0.9)

�1.1

(0.9)

0.0(0.2)

�1.3

(0.8)

�1.3

(0.9)

0.0(0.3)

�1.2

(0.6)

�1.4

(0.7)

0.1(0.3)

Arousal

3.6(0.5)

3.5(0.5)

2.4(0.4)

3.6(0.7)

3.4(0.6)

2.5(0.5)

3.6(0.6)

3.4(0.7)

2.4(0.4)

3.5(0.6)

3.5(0.5)

2.3(0.4)

Imag

eability

4.7(1.4)

4.3(1.1)

4.3(1.5)

4.3(1.3)

4.3(1.0)

4.1(1.3)

4.8(1.2)

4.3(1.0)

4.5(1.4)

4.6(1.4)

4.1(1.3)

4.4(1.1)

Letters

6.1(1.2)

6.3(1.5)

6.5(1.1)

6.0(1.1)

6.2(1.5)

6.1(1.1)

6.0(1.3)

6.4(1.2)

6.3(1.0)

6.0(1.3)

6.2(1.3)

6.4(1.2)

Sylla

bles

2.0(0.5)

2.1(0.9)

2.2(0.5)

1.9(0.6)

2.0(0.8)

2.1(0.4)

2.1(0.8)

2.2(0.8)

2.2(0.7)

2.0(0.8)

2.0(0.6)

2.1(0.6)

Phon

emes

5.5(1.3)

5.6(1.5)

5.3(0.9)

5.3(1.0)

5.3(1.7)

5.4(1.2)

5.4(1.5)

5.8(1.3)

5.4(1.1)

5.5(1.3)

5.5(1.5)

5.6(1.0)

Freq

uen

cy31

.5(45.5)

16.0

(23.5)

27.7

(51.8)

26.7

(43.8)

37.2

(56.1)

24.4

(42.2)

16.7

(23.8)

32.5

(69.2)

32.9

(46.8)

27.2

(83.2)

19.7

(42.5)

21.2

(31.2)

M¼

mea

n;SD

¼stan

darddev

iation

.

A. Weigand et al. / Brain Stimulation 6 (2013) 905e912906

investigating causal brainebehavior relationships. For example, byusing adaptation to manipulate neural activation states prior to theapplication of TMS, specific and even spatially overlapping neuronalrepresentations were examined when processing numbers [11] andwords [12] as well as during visual motion perception [13].

Based on these considerations, the primary aim of our study wasto assess the effects of tDCS priming followed by rTMS applied overthe DLPFC in healthy participants on the performance in anemotional working memory task.

Working memory refers to the temporary storage and manipu-lation of information that is no longer accessible in the environ-ment [14,15]. Numerous neuroimaging findings demonstrated a keyrole of the DLPFC in working memory processing [16e18], sup-ported by studies using rTMS [19e22] and tDCS [23e27].

The interaction of working memory and emotion has beeninvestigated only in a small number of studies, providing conflictingresults. At the behavioral level, negative stimuli have shown bothimpairing [28e30] and facilitating [31] effects on task performance.Other studies found no behavioral impact of emotional content onworking memory [32,33]. At the neuronal level, no consistentrelationship was found between DLPFC activity and emotionalstimuli [30,32e34]. Our own data [33] demonstrated a strongerrecruitment of the right DLPFC while processing negative stimuli inworking memory, supporting a valence-specific hemisphericlateralization in emotion processing [35,36]. Furthermore, in ourprevious rTMS study [37], we extended our neuroimaging findingsby demonstrating a causal role of the right DLPFC in workingmemory for negative, withdrawal-related words. These findingsprovided further support for the motivational direction model,which states that the right prefrontal cortex is dominant in theprocessing of withdrawal-related emotions, whereas the leftprefrontal cortex is dominant in the processing of approach-relatedemotions [38,39]. In line with this model, the left prefrontal cortexhas been found to process anger, a negatively valenced emotionthat is associated with approach motivation [40].

Therefore, the second aim of our study was to further exploreour previous findings of prefrontal hemispheric lateralization inemotional working memory processing by distinguishing negativestimuli into withdrawal- and approach-related categories.

Participants were presented with an emotional n-back (EMO-BACK) task consisting of words related to the distinct emotioncategories fear and anger as well as neutral words. In a 2 (anodal,cathodal tDCS priming)� 2 (active, sham rTMS) randomized single-blind within-subjects design, task performance was measured atbaseline and after combined tDCS priming (1 mA, 15 min) andsubsequent rTMS (1 Hz, 15 min, 110% MT) of the right DLPFC.

Based on evidence of state-dependency of rTMS-inducedchanges [2,5e9], we hypothesized that the direction of afteref-fects produced by low-frequency rTMS depends on the polarity oftDCS priming. In accordance with the motivational direction model[38,39], we expected that stimulation of the right DLPFC wouldespecially influence task performance for fear-related words, whichare associated with withdrawal motivation. Due to inconsistentfindings of the impact of emotional content on working memoryprocessing, no specific predictions were made about the directionof stimulation-induced changes in task performance.

Figure 2. Emotion-dependent task accuracy at baseline (pre) and after stimulation(post) across conditions.

A. Weigand et al. / Brain Stimulation 6 (2013) 905e912 907

Methods

Sample

Fifteen healthy volunteers (8 females) participated in the study(mean age 25.3 � 4.0, range 19e32 years). All participants wereright-handed [41] native German speakers and naïve to rTMS. Theirintelligence levels were at or above the range of norm [42]. Noparticipants reported any previous or concomitant neurological orpsychiatric condition [43] or any contradictions to rTMS [44,45]. Toexclude epileptiform activity, every participant underwent resting-state electroencephalography (EEG), screened by a trained physi-cian. The study was approved by the Institutional Review Board ofthe Charité University Medicine (Berlin, Germany). All subjects gavetheir written informed consent in accordance with the ethicalguidelines of the Declaration of Helsinki and were financiallycompensated for their participation.

Emotional working memory task

Based on our previous experiments [33,37], we used a verbalemotional n-back task (EMOBACK). Stimuli were German wordstaken from the Discrete Emotion Norms for Nouns e Berlin Affec-tive Word List [46]. For the EMOBACK task, nouns strongly relatedto fear and anger were selected and controlled for valence andarousal (t’s< 1). In addition to these negative stimuli, neutral wordswere included in the task. The three emotion conditions werematched on imageability, number of letters, syllables, phonemes,frequency and orthographic neighbors (F’s > 1). Four parallel sets,each consisting of 20 words per condition, were created with thedescribed stimulus characteristics (t’s < 1). Table 1 provides anoverview of the stimulus characteristics used in the four parallelversions of the EMOBACK task.

Words were presented in a randomized sequence withina 3-back task. Participants had to press a button whenever a pre-sented word was the same as the one presented three trials earlier.A separate button indicated a non-target. Target and non-targetbutton were located on opposite sides of a standard computerkeyboard. Subjects were instructed to respond as quickly and asaccurately as possible by using both index fingers. The words werepresented for 500 ms in uppercase white letters centered on a blackscreen with an interstimulus interval (ISI) of 1500 ms. 18 words ofthe same emotion condition (fear, anger, neutral) were presented ina block, followed by a fixation period of 10e14 s. Participants wereintensively trained in the task prior to the actual experiment.During task performance, they were seated 90 cm in front of themonitor in a silent room and the monitor was placed at eye level.The EMOBACK task was programmed using Presentation software(Version 14.5, Neurobehavioral Systems Inc., San Francisco, CA,USA).

Figure 3. Accuracy differences for anger-related words from baseline for the fourstimulation conditions in the 2 (anodal/cathodal tDCS priming) � 2 (active/sham rTMS)design.

Experimental procedure

In a single-blind randomized within-subjects design, eachparticipant underwent four different stimulation conditions:anodal or cathodal tDCS priming followed by active or sham rTMSover the right DLPFC. In the first part of the experiment, twostimulation conditions were realized in successive sessions on thesame day, separated by a 45-min washout period to avoid carry-over effects [47,48]. Approximately a week later, the correspond-ing second part of the experiment was completed at the same timeof day. In each stimulation condition, 15 randomized blocks of theEMOBACK task were presented at baseline and immediately afterrTMS application (Fig. 1). In total, 2160 stimuli (540 stimuli for each

of the four stimulation conditions) were presented over the courseof the experiment.

tDCS priming

tDCS was delivered with a battery-driven constant currentstimulator (CX-6650 DCS device by Rolf Schneider Electronics,Gleichen, Germany) using two electrodes placed in saline-soakedsponges (5 � 7 cm). Depending on the stimulation condition,either the anode or the cathode was placed on the right DLPFC. Thereference electrode was placed on the left supraorbital area. tDCSwas delivered for 15 min at 1 mA with a ramp time of 30 s.

rTMS

Biphasic magnetic stimulation was delivered with a Medtronicstimulator (MagPro X100 with MagOption, MagVenture, Farum,Denmark) connected to a figure-eight coil (MCF-B65) after a stan-dardized time interval of 10 min [5]. In this period, an adjustablearm was used to allow the precise coil position to be maintainedthroughout the experiment. The optimal stimulation site of theright DLPFC was marked individually using three-dimensionalmagnetic resonance imaging (3D-MRI) [49]. An anatomical high-resolution image was acquired for each participant on a SiemensTrio 3T scanner with conventional parameters (176 T1-weightedimages with a slice thickness of 1 mm) approximately a week priorto the actual experiment. The coil was angled at 45�, with thejunction of the two coil wings placed perpendicular to the markedstimulation point on the skull using a neuronavigation system

Table 2EMOBACK task performance separated for emotion conditions (fear, anger, neutral) at baseline (pre) and after stimulation (post).

Pre Post

Fear Anger Neutral Fear Anger Neutral

M (SE) M (SE) M (SE) M (SE) M (SE) M (SE)

Accuracy (%) Anodal tDCS Active rTMS 82.3 (1.9) 83.3 (1.4) 85.9 (1.7) 85.6 (1.6) 85.0 (1.7) 83.9 (2.0)Sham rTMS 85.9 (1.9) 86.1 (1.5) 87.8 (1.2) 86.6 (2.0) 86.2 (2.3) 85.4 (2.0)

Cathodal tDCS Active rTMS 85.3 (1.8) 85.7 (1.5) 85.4 (1.6) 87.0 (1.8) 85.0 (1.7) 82.7 (2.2)Sham rTMS 85.0 (2.1) 83.9 (1.8) 85.3 (1.4) 87.9 (1.6) 87.1 (2.1) 84.1 (2.0)

Reaction times (ms) Anodal tDCS Active rTMS 527.3 (40.3) 524.1 (38.4) 523.4 (41.3) 530.7 (35.4) 506.8 (37.2) 505.4 (40.6)Sham rTMS 535.1 (45.3) 535.3 (48.0) 538.8 (47.3) 549.1 (41.8) 524.1 (41.2) 523.1 (43.8)

Cathodal tDCS Active rTMS 535.9 (41.5) 515.2 (39.3) 521.3 (41.8) 531.4 (37.9) 503.4 (38.3) 509.2 (37.0)Sham rTMS 533.8 (46.7) 536.4 (46.3) 530.6 (46.1) 551.0 (41.5) 524.5 (40.9) 515.2 (38.8)

A. Weigand et al. / Brain Stimulation 6 (2013) 905e912908

(eXimia Navigated Brain Stimulation (NBS), Nexstim, Helsinki,Finland). A placebo coil (MCF-P-B65) was positioned in the samemanner for sham stimulation. rTMS was applied for 15 min at 1 Hz(900 pulses) with an intensity of 110% of the resting motorthreshold which was defined as the minimum intensity capable ofevoking motor potentials of at least 50 mV recorded from the rightfirst dorsal interosseus (FDI) in 5/10 stimulations.

Additional measures

To control for interindividual differences in characteristics thatmight influence task performance at baseline [39,50e53] or stim-ulation effects [54,55], we assessed the Big Five personalitydomains [56] as well as participants’ predisposition to anxiety [57],anger [58] and approach and withdrawal behavior [59]. Further-more, subjects completed a mood questionnaire [60] prior to theEMOBACK baseline measurement (Tpre) and immediately afterrTMS application (Tpost). After each stimulation condition, valenceand arousal ratings were performed for all words presented in thefour parallel sets of the EMOBACK task.

Statistical analysis

Results were analyzed using PASW (Predictive Analysis SoftWare, Version 18.0, Chicago; SPSS Inc., Illinois, USA). Accuracy wasdefined as the ratio of correct responses (hits and correct rejections)to total number of stimuli. Mean reaction times of correct responseswere additionally analyzed. In order to investigate the impact ofemotional content on task performance independent of stimula-tion, one-way analyses of variance (ANOVAs) for the factor Emotion(fear, anger, neutral) were applied for accuracy and reaction times

Table 3Sample characteristics.

M (SD)

MWT-B IQ 100.9 (6.2)NEO-FFI Neuroticism 16.4 (6.2)

Extraversion 28.7 (4.8)Openness 32.7 (5.7)Agreeableness 33.0 (5.6)Conscientiousness 32.2 (7.1)

STAI-T 34.6 (5.4)STAXI-T 7.9 (2.9)BIS Anxiety 2.2 (0.5)

Frustration 2.2 (0.6)BAS Drive 3.3 (0.4)

Gratification 3.5 (0.3)

M ¼ mean; SD ¼ standard deviation; IQ ¼ intelligence quotient; MWT-B ¼Mehrfachwahl Wortschatz Intelligenz test (verbal intelligence test); NEO-FFI ¼ NEOFive-Factor Inventory; STAI¼ State-Trait Anxiety Inventory; STAXI¼ State-Trait AngerExpression Inventory; BIS ¼ behavioral inhibition system; BAS ¼ behavioral approachsystem.

at baseline. To analyze stimulation effects, repeated measuresANOVAs were applied for accuracy and reaction times with thefactors tDCS priming (anodal, cathodal), rTMS (active, sham),Emotion (fear, anger, neutral) and Time (pre, post). To evaluatechanges in subjects’ current emotional state, repeated measuresANOVAs were applied for all MDBF subscales with the factors tDCSpriming (anodal, cathodal), rTMS (active, sham) and Time (Tpre, Tpost).Furthermore, individual valence and arousal ratings of the stimulusmaterial were analyzed separately in ANOVAs with the factors tDCSpriming (anodal, cathodal), rTMS (active, sham), Emotion (fear,anger, neutral). For all analyses, the two-tailed threshold of signif-icance was set at P < 0.05. If ANOVAs revealed significant main orinteraction effects, further statistical analyses were conductedusing contrasts or simple effects.

Results

Task performance

Baseline measurementA significant main effect of Emotion was found at baseline

[F(2,28) ¼ 3.61, P < 0.05, h2 ¼ 0.21] indicating impaired taskaccuracy for stimuli with negative valence. In comparison to neutralwords, participants performed significantly less accurate for fear-related [F(1,14) ¼ 5.43, P < 0.05, h2 ¼ 0.28] and anger-relatedwords [F(1,14) ¼ 8.67, P < 0.05, h2 ¼ 0.38]. With regard to reac-tion times, no significant effect of emotional content was found.

Stimulation effects

The ANOVA on accuracy revealed a significant Emotion by Timeinteraction [F(2,28) ¼ 10.21, P < 0.001, h2 ¼ 0.42]. As illustrated inFig. 2, task accuracy increased significantly for fear-related words[F(1,14) ¼ 14.24, P < 0.05, h2 ¼ 0.50] and decreased for neutralwords [F(1,14) ¼ 5.34, P < 0.05, h2 ¼ 0.28] across stimulationconditions.

With regard to specific stimulation effects, there was a trend-wise tDCS priming by Emotion interaction [F(2,28) ¼ 2.7, P ¼ 0.085,h2 ¼ 0.16]. To further explore this interaction we conductedANOVAs for each emotion condition separately. For anger-relatedwords, the threefold interaction tDCS priming by rTMS by Timealmost reached significance [F(1,14) ¼ 4.51, P ¼ 0.052, h2 ¼ 0.24].Further analyses of changes in task performance from baselinerevealed converse rTMS effects dependent on the polarity ofpreceding tDCS (Fig. 3). Contrasts showed that this anger-specificeffect was driven by a trend toward higher task accuracy forsubsequent sham rTMS as compared to active rTMS in the cathodalpriming condition [F(1,14) ¼ 4.24, P ¼ 0.059, h2 ¼ 0.23].

The ANOVA on reaction times revealed a significant main effectof Emotion [F(2,28) ¼ 31.85, P < 0.001, h2 ¼ 0.70] and a significant

Table 4Mood assessment before and after stimulation.

Elevated e depressed mood Wakefulness e sleepiness Calmness e restlessness

Pre Post Pre Post Pre Post

M (SD) M (SD) M (SD) M (SD) M (SD) M (SD)

Anodal tDCS Active rTMS 16.5 (2.5) 15.1 (4.3) 14.1 (3.1) 11.9 (4.1) 16.0 (3.1) 15.1 (3.2)Sham rTMS 16.7 (2.1) 15.7 (3.8) 14.6 (3.0) 12.0 (3.4) 16.2 (2.9) 16.3 (2.7)

Cathodal tDCS Active rTMS 16.8 (2.1) 16.7 (2.0) 13.8 (2.7) 13.1 (2.2) 17.1 (2.0) 16.9 (2.4)Sham rTMS 17.2 (2.0) 16.6 (2.6) 14.8 (2.9) 13.0 (3.0) 16.3 (2.8) 16.9 (2.5)

M ¼ mean; SD ¼ standard deviation.

A. Weigand et al. / Brain Stimulation 6 (2013) 905e912 909

Emotion by Time interaction [F(2,28) ¼ 11.60, P < 0.001, h2 ¼ 0.45],indicating differences between emotion conditions across stimu-lation conditions. Contrasts showed that participants respondedgenerally slower to fear-related words than to anger-related[F(2,13) ¼ 64.83, P < 0.001, h2 ¼ 0.91] and neutral words[F(2,13) ¼ 64.83, P < 0.001, h2 ¼ 0.91]. Over time, reaction timesdecreased for neutral words [F(1,14) ¼ 5.12, P < 0.05, h2 ¼ 0.27] andtrendwise for anger-related words [F(1,14) ¼ 4.08, P ¼ 0.063,h2 ¼ 0.23] across stimulation conditions. No effects of specificstimulation conditions were found for reaction times.

Accuracy and reaction times at baseline (pre) and after stimu-lation (post) are reported in Table 2, separated for emotionconditions.

Control variables

No correlations were found between individual task perfor-mance (at baseline and after stimulation) and all assessed person-ality traits (Table 3).

With regard to potential stimulation effects on participants’emotional state, ANOVAs on theMDBF scores revealed amain effectof Time on the wakefulness-sleepiness subscale [F(1,14) ¼ 10.55,P< 0.05, h2¼ 0.43]. Participants felt in general more tired at the endof an experimental session. Furthermore, a main effect of tDCSpriming was found on the calmness-restlessness subscale[F(1,14) ¼ 6.28, P < 0.05, h2 ¼ 0.31], indicating stronger feelings ofrestlessness after anodal tDCS followed by active rTMS. Importantly,there were no mood changes caused by any of the four stimulationconditions. Means and standard deviations of the MDBF scores areshown in Table 4.

Word ratings

ANOVAs on individual word ratings revealed a main effect ofEmotion on both valence [fear: �0.7 � 0.4, anger: �0.9 � 0.1,neutral: 0.2 � 0.3; F(2,28) ¼ 61.78, P < 0.001, h2 ¼ 0.82] and arousal[fear and anger: 2.6 � 0.2; neutral: 1.9 � 0.7; F(2,28) ¼ 25.81,P < 0.001, h2 ¼ 0.65]. Contrasts showed that in comparison toneutral words, both fear- and anger-related words were rated moreunpleasant [F(1,14) ¼ 42.90, P < 0.001, h2 ¼ 0.76 andF(1,14) ¼ 112.96, P < 0.001, h2 ¼ 0.89] and more arousing[F(1,14) ¼ 38.45, P < 0.001, h2 ¼ 0.73 and F(1,14) ¼ 22.30, P < 0.001,h2 ¼ 0.61]. Between fear- and anger-related words, a significantdifference was found for valence [F(1,14) ¼ 10.00, P < 0.05,h2 ¼ 0.42], indicating higher negative ratings for anger-relatedwords. No stimulation effects on valence or arousal ratings werefound.

Discussion

This study aimed to investigate state-dependent prefrontalrTMS effects on the performance in an emotional working memorytask. By including fear- and anger-related words in an n-back

paradigmwewere additionally able to assess the interface betweenworking memory and emotion as well as potential hemisphericlateralization of emotion processing. To our knowledge, this is thefirst study to explore the influence of tDCS priming followed byrTMS on cognitive-affective functions in healthy participants.

First of all, we found a detrimental effect of negative emotionalcontent on working memory performance independent of stimu-lation. At baseline, participants responded less accurately to bothfear- and anger-related words than to neutral words. This finding isin line with previous studies demonstrating that negativeemotional content can impair task performance in n-back para-digms [28e30]. It has been argued that emotionally salient stimulicapture and hold attention, resulting in increased task demands forhigher executive functions such as working memory [29].

After tDCS priming and subsequent rTMS over the right DLPFC,task accuracy increased for fear-related words and decreased forneutral words across stimulation conditions (Fig. 2). No changes inreaction times were found for fear-related words, whereas partic-ipants responded faster to neutral words. A potential mechanismunderlying our results might be seen in an inhibition of the rightDLPFC across stimulation conditions. According to the motivationaldirection model [38,39] and in line with our previous study [37],reduced activation of the right DLPFC might have influenced theprocessing of fear-related words exclusively. Previous findingsindicate that low-frequency rTMS over the right DLPFC reducesvigilant attention to fear-related stimuli [61]. Considering a detri-mental effect of negative emotional content on working memory[28e30], which is supported by our baseline measurement, theinhibition of fear-related stimuli might have enhanced taskperformance in our study through a reduced attentional bias, andconsequently a prioritized processing of task-relevant information.With regard to the stimulation-induced decrease in task accuracyfor neutral words, a potential explanation might be seen in thebilateral involvement of the DLPFC in the central executive of theworking memory system [62]. The n-back paradigm used in ourstudy involves a number of key processes within working memoryincluding monitoring, updating and manipulating the rememberedinformation [63]. Comparable to the neutral condition in ourEMOBACK task, previous research demonstrated that inhibitorystimulation significantly impaired task performance in a non-emotional verbal n-back task when applied to the left or rightDLPFC [21]. Importantly, our result of a speeding of responseaccompanying the decrease in accuracy for neutral words mightappear to reflect a speed-accuracy trade-off induced by the stim-ulation. However, a general speed-accuracy trade-off is unlikelybecause trendwise speeding of response was also found for anger-related words, but without concomitant decrease in accuracy. Inorder to ensure that a speed-accuracy trade-off had not occurred inthe neutral condition, we calculated the correlation betweenaccuracy and RTs. We did not find a correlation, ruling outa speedeaccuracy trade-off that might have accounted for thedescribed rTMS effects. Alternatively, our results showing shorterreaction times despite decreased task accuracy for neutral words

A. Weigand et al. / Brain Stimulation 6 (2013) 905e912910

might be explained by previous findings suggesting that a diminu-tion in cortical excitability impairs the ability to identify targetstimuli, but, once identified, improves the perception due toa focusing effect [64,65].

With regard to task performance for anger-related words, ourresults showed more specific stimulation effects (Fig. 3). Whereasactive rTMS over the right DLPFC increased task accuracy whenapplied after anodal tDCS, it slightly decreased after cathodal tDCS.These converse stimulation effects in dependence on the polarity ofpreceding tDCS are in line with neurophysiological studies from themotor cortex indicating state-dependent effects of rTMS applica-tion [2,5e7]. Further analyses of our data revealed, however, thatthe interaction between tDCS priming and rTMS was driven byenhanced task accuracy of subsequent sham rTMS as compared toactive rTMS in the cathodal priming condition. Thus, participantsresponded to anger-related words more accurately after inhibitorycathodal tDCS applied over the right DLPFC. Based on the motiva-tional direction model [38,39], we would have expected theopposite effect for anger-related words, namely a decrease in taskaccuracy after inhibiting the right DLPFC. Previous findings showedthat low-frequency rTMS applied over the right DLPFC shifts theanterior asymmetry in brain activation to the left througha contralateral excitation [66,67] and increases selective attentionto anger-related stimuli [68]. However, in our previous study [37],no anger-related effects were found in the EMOBACK task after bothleft- or right-sided low-frequency rTMS of the DLPFC.

Importantly, our stimulation effects were not accompanied bysystematic mood changes across stimulation conditions. Thisfinding supports previous research showing no influence on theemotional state in healthy volunteers after prefrontal stimulationusing low-frequency rTMS [69,70] or tDCS [65,71,72]. Our partici-pants, however, reported significant stronger feelings of restless-ness after anodal tDCS followed by active rTMS. This stimulationcondition consists of an active double and direct stimulation of theDLPFC, making the sensation accompanying the stimulation morediscernible than the other stimulation conditions. It can be specu-lated that stronger feelings of restlessness may have influencedotherwise more specific stimulation effects by disturbing taskperformance.

Because emotional valence and arousal may be controlled bydifferent neural systems [73,74], an advantage of this study is thecarefully matched selection of words with equal valence andarousal levels for fear- and anger-related words according to theBAWL norms. We additionally controlled our stimuli sets withregard to lexical and sublexical variables whose relevance has beenshown in tasks relying onwordmaterial [75]. To further control ourstimulusmaterial, we asked participants to rate valence and arousalof all presented words individually. As expected, the results of theword ratings of our sample revealed that participants perceivedanger- and fear-related words as more unpleasant and arousingrelative to neutral words. Fear- and anger-related words did notdiffer in arousal ratings; however, anger-related words were ratedmore negatively than fear-related words. To further explore thisfinding, it would be reasonable to include individual word ratings toassess the distinct emotions fear and anger in future studies. In thepresent study, distinct emotion ratings could have clarifiedwhetherthe selected anger-related words from the DENN-BAWL werepossibly not correctly perceived and therefore, not associated withthe left DLPFC as suggested by previous findings [67,68]. Impor-tantly, our participants rated the stimulus material equally acrossstimulation conditions indicating that the reported stimulationeffects cannot be explained by differences in perceived valence orarousal levels of our stimuli. Moreover, we controlled for person-ality traits that are associated with a potential influence on emotionprocessing [39,50e53] and might influence stimulation effects

[54,55]. However, a number of additional factors are likely toinfluence individual responses to brain stimulation [76] and need tobe further explored in future research.

In summary, we demonstrated that the withdrawal-relatedemotion fear is associated with the right DLPFC and seems to bedifferentially processed as compared to the approach-relatedemotion anger. This finding is in line with our previous study [37]and further supports a frontal asymmetry of emotion processingas proposed by the motivational direction model [38,39]. Nogeneral state-dependent effects of prefrontal rTMS on workingmemory were found. Furthermore, our data contribute to clarifyingthe interaction between working memory and emotion by sup-porting previous findings of a detrimental effect of negativeemotional content [28e30]. Future interleaved TMS/tDCS-fMRIstudies are needed in order to causally investigate neural mecha-nisms underlying lateralized emotion processing and state-dependent prefrontal stimulation effects on higher cognitivefunctions.

Acknowledgments

We would like to thank Philipp Kazzer for his experimentalassistance and Antje Astalosch for her support in acquisition of thedata.

References

[1] Silvanto J, Muggleton N, Walsh V. State-dependency in brain stimulationstudies of perception and cognition. Trends in Cognitive Sciences2008;12(12):447e54.

[2] Lang N, Siebner HR, Ernst D, Nitsche MA, Paulus W, Lemon RN, et al. Pre-conditioning with transcranial direct current stimulation sensitizes the motorcortex to rapid-rate transcranial magnetic stimulation and controls thedirection of after-effects. Biological Psychiatry 2004;56(9):634e9.

[3] Langguth B, Kleinjung T, Frank E, Landgrebe M, Sand P, Dvorakova J, et al.High-frequency priming stimulation does not enhance the effect of low-frequency rTMS in the treatment of tinnitus. Experimental Brain Research2008;184(4):587e91.

[4] Siebner HR, Hartwigsen G, Kassuba T, Rothwell JC. How does transcranialmagnetic stimulation modify neuronal activity in the brain? Implications forstudies of cognition. Cortex 2009;45(9):1035e42.

[5] Siebner HR, Lang N, Rizzo V, Nitsche MA, Paulus W, Lemon RN, et al. Pre-conditioning of low-frequency repetitive transcranial magnetic stimulationwith transcranial direct current stimulation: evidence for homeostatic plas-ticity in the human motor cortex. The Journal of Neuroscience2004;24(13):3379e85.

[6] Cosentino G, Fierro B, Paladino P, Talamanca S, Vigneri S, Palermo A, et al.Transcranial direct current stimulation preconditioning modulates the effectof high-frequency repetitive transcranial magnetic stimulation in the humanmotor cortex. The European Journal of Neuroscience 2012;35(1):119e24.

[7] Iyer MB, Schleper N, Wassermann EM. Priming stimulation enhances thedepressant effect of low-frequency repetitive transcranial magnetic stimula-tion. The Journal of Neuroscience 2003;23(34):10867e72.

[8] Fitzgerald PB, Hoy K, McQueen S, Herring S, Segrave R, Been G, et al. Primingstimulation enhances the effectiveness of low-frequency right prefrontalcortex transcranial magnetic stimulation in major depression. Journal ofClinical Psychopharmacology 2008;28(1):52e8.

[9] Nongpiur A, Sinha VK, Praharaj SK, Goyal N. Theta-patterned, frequency-modulated priming stimulation enhances low-frequency, right prefrontalcortex repetitive transcranial magnetic stimulation (rTMS) in depression:a randomized, sham-controlled study. Journal of Neuropsychiatry and ClinicalNeurosciences 2011;23(3):348e57.

[10] Loo CK, Sachdev P, Martin D, Pigot M, Alonzo A, Malhi GS, et al. A double-blind,sham-controlled trial of transcranial direct current stimulation for the treat-ment of depression. The International Journal of Neuropsychopharmacology2010;13(1):61e9.

[11] Kadosh RC, Muggleton N, Silvanto J, Walsh V. Double dissociation of format-dependent and number-specific neurons in human parietal cortex. CerebralCortex 2010;20:2166e71.

[12] Cattaneo Z, Devlin JT, Salvini F, Vecchi T, Silvanto J. The causal role of category-specific neuronal representations in the left ventral premotor cortex (PMv) insemantic processing. NeuroImage 2010;49:2728e34.

[13] Cattaneo L, Sandrini M, Schwarzbach J. State-dependent TMS reveals a hier-archical representation of observed acts in the temporal, parietal, and pre-motor cortices. Cerebral Cortex 2010;20:2252e8.

A. Weigand et al. / Brain Stimulation 6 (2013) 905e912 911

[14] D’Esposito M, Postle BR, Rypma B. Prefrontal cortical contributions to workingmemory: evidence from event-related fMRI studies. Experimental BrainResearch 2000;133(1):3e11.

[15] Goldman-Rakic PS. Architecture of the prefrontal cortex and the centralexecutive. Annals of the New York Academy of Sciences 1995;769:71e83.

[16] Curtis CE, D’Esposito M. Persistent activity in the prefrontal cortex duringworking memory. Trends in Cognitive Sciences 2003;7(9):415e23.

[17] Braver TS, Barch DM, Kelley WM, Buckner RL, Cohen NJ, Miezin FM, et al.Direct comparison of prefrontal cortex regions engaged by working and long-term memory tasks. NeuroImage 2001;14(1 Pt 1):48e59.

[18] Jonides J, Schumacher EH, Smith EE, Lauber EJ, Awh E, Minoshima S, et al.Verbal working memory load affects regional brain activation as measured byPET. Journal of Cognitive Neuroscience 1997;9:462e75.

[19] Hamidi M, Tononi G, Postle BR. Evaluating the role of prefrontal and parietalcortices in memory-guided response with repetitive transcranial magneticstimulation. Neuropsychologia 2009;47(2):295e302.

[20] Luber B, Kinnunen LH, Rakitin BC, Ellsasser R, Stern Y, Lisanby SH. Facilitationof performance in a working memory task with rTMS stimulation of theprecuneus: frequency- and time-dependent effects. Brain Research2007;1128(1):120e9.

[21] Mottaghy FM, Krause BJ, Kemna LJ, Topper R, Tellmann L, Beu M, et al.Modulation of the neuronal circuitry subserving working memory in healthyhuman subjects by repetitive transcranial magnetic stimulation. NeuroscienceLetters 2000;280(3):167e70.

[22] SandriniM, Rossini PM,Miniussi C. Lateralized contribution of prefrontal cortexin controlling task-irrelevant information during verbal and spatial workingmemory tasks: rTMS evidence. Neuropsychologia 2008;46(7):2056e63.

[23] Fregni F, Boggio PS, Nitsche M, Bermpohl F, Antal A, Feredoes E, et al. Anodaltranscranial direct current stimulation of prefrontal cortex enhances workingmemory. Experimental Brain Research 2005;166(1):23e30.

[24] Ohn SH, Park CI, Yoo WK, Ko MH, Choi KP, Kim GM, et al. Time-dependenteffect of transcranial direct current stimulation on the enhancement ofworking memory. Neuroreport 2008;19(1):43e7.

[25] Mulquiney PG, Hoy KE, Daskalakis ZJ, Fitzgerald PB. Improving workingmemory: exploring the effect of transcranial random noise stimulation andtranscranial direct current stimulation on the dorsolateral prefrontal cortex.Clinical Neurophysiology 2011;122(12):2384e9.

[26] Sandrini M, Fertonani A, Cohen LG, Miniussi C. Double dissociation of workingmemory load effects induced by bilateral parietal modulation. Neuro-psychologia 2012;50(3):396e402.

[27] Zaehle T, Sandmann P, Thorne JD, Jancke L, Herrmann CS. Transcranial directcurrent stimulation of the prefrontal cortex modulates working memoryperformance: combined behavioural and electrophysiological evidence. BMCNeuroscience 2011;12:2.

[28] Kensinger EA, Corkin S. Effect of negative emotional content on workingmemory and long-term memory. Emotion 2003;3(4):378e93.

[29] Lindström BR, Bohlin G. Threat-relevance impairs executive functions: nega-tive impact on working memory and response inhibition. Emotion2012;12(2):384e93.

[30] Perlstein WM, Elbert T, Stenger VA. Dissociation in human prefrontal cortex ofaffective influences on working memory-related activity. Proceedings of theNational Academy of Sciences of the United States of America2002;99(3):1736e41.

[31] Lindström BR, Bohlin G. Emotion processing facilitates working memoryperformance. Cognition and Emotion 2011;25(7):1196e204.

[32] Döhnel K, Sommer M, Ibach B, Rothmayr C, Meinhardt J, Hajak G. Neuralcorrelates of emotional working memory in patients with mild cognitiveimpairment. Neuropsychologia 2008;46(1):37e48.

[33] Grimm S, Weigand A, Kazzer P, Jacobs AM, Bajbouj M. Neural mechanismsunderlying the integration of emotion and working memory. NeuroImage2012;61(4):1188e94.

[34] Neta M, Whalen PJ. Individual differences in neural activity during a facialexpression vs. identity working memory task. NeuroImage 2011;56(3):1685e92.

[35] Davidson RJ, Irwin W. The functional neuroanatomy of emotion and affectivestyle. Trends in Cognitive Sciences 1999;3(1):11e21.

[36] Herrington JD, Heller W, Mohanty A, Engels AS, Banich MT, Webb AG, et al.Localization of asymmetric brain function in emotion and depression.Psychophysiology 2010;47(3):442e54.

[37] Weigand A, Grimm S, Astalosch A, Guo JS, Briesemeister BB, Lisanby SH, et al.Lateralized effects of prefrontal repetitive transcranial magnetic stimulationon emotional working memory. Experimental Brain Research 2013;227(1):43e52.

[38] Harmon-Jones E. Contributions from research on anger and cognitive disso-nance to understanding the motivational functions of asymmetrical frontalbrain activity. Biological Psychology 2004;67(1e2):51e76.

[39] Harmon-Jones E, Gable PA, Peterson CK. The role of asymmetric frontalcortical activity in emotion-related phenomena: a review and update. Bio-logical Psychology 2010;84(3):451e62.

[40] Harmon-Jones E. Early Career Award. Clarifying the emotive functions ofasymmetrical frontal cortical activity. Psychophysiology 2003;40(6):838e48.

[41] Oldfield RC. The assessment and analysis of handedness: the Edinburghinventory. Neuropsychologia 1971;9(1):97e113.

[42] Lehrl S. Mehrfachwahl-Wortschatz-Intelligenztest MWT-B. Balingen: SpittaVerlag; 2005.

[43] Wittchen HU, Zaudig M, Fydrich T. Strukturiertes Klinisches Interview fürDSM-IV. Achse I und II. Handanweisung. Göttingen: Hogrefe; 1997.

[44] Rossi S, Hallett M, Rossini PM, Pascual-Leone A. Safety, ethical considerations,and application guidelines for the use of transcranial magnetic stimulation inclinical practice and research. Clinical Neurophysiology 2009;120(12):2008e39.

[45] Wassermann EM. Risk and safety of repetitive transcranial magnetic stimu-lation: report and suggested guidelines from the International Workshop onthe Safety of Repetitive Transcranial Magnetic Stimulation, June 5-7, 1996.Electroencephalography and Clinical Neurophysiology 1998;108(1):1e16.

[46] Briesemeister BB, Kuchinke L, Jacobs AM. Discrete emotion norms for nouns:Berlin affective word list (DENN-BAWL). Behavior Research Methods2011;43(2):441e8.

[47] Bermpohl F, Fregni F, Boggio PS, Thut G, Northoff G, Otachi PT, et al. Effect oflow-frequency transcranial magnetic stimulation on an affective go/no-gotask in patients with major depression: role of stimulation site and depres-sion severity. Psychiatry Research 2005;141(1):1e13.

[48] Thut G, Pascual-Leone A. A review of combined TMS-EEG studies to charac-terize lasting effects of repetitive TMS and assess their usefulness in cognitiveand clinical neuroscience. Brain Topography 2010;22(4):219e32.

[49] Peleman K, Van Schuerbeek P, Luypaert R, Stadnik T, De Raedt R, De Mey J,et al. Using 3D-MRI to localize the dorsolateral prefrontal cortex in TMSresearch. The World Journal of Biological Psychiatry 2010;11(2 Pt 2):425e30.

[50] Bishop SJ. Trait anxiety and impoverished prefrontal control of attention.Nature Neuroscience 2009;12(1):92e8.

[51] Van Honk J, Tuiten A, de Haan E, van den Hout M, Stam H. Attentional biasesfor angry faces: relationships to trait anger and anxiety. Cognition andEmotion 2001;15(3):279e97.

[52] Shackman AJ, Sarinopoulos I, Maxwell JS, Pizzagalli DA, Lavric A, Davidson RJ.Anxiety selectively disrupts visuospatial working memory. Emotion2006;6(1):40e61.

[53] Balconi M, Mazza G. Lateralisation effect in comprehension of emotional facialexpression: a comparison between EEG alpha band power and behaviouralinhibition (BIS) and activation (BAS) systems. Laterality 2010;15(3):361e84.

[54] Balconi M, Ferrari C. rTMS stimulation on left DLPFC affects emotional cueretrieval as a function of anxiety level and gender. Depression and Anxiety2012;29(11):976e82.

[55] Wassermann EM, Greenberg BD, Nguyen MB, Murphy DL. Motor cortexexcitability correlates with an anxiety-related personality trait. BiologicalPsychiatry 2001;50(5):377e82.

[56] Costa PT, McCrae RR. Revised NEO personality inventory (NEO-PI-R) and NEOfive-factor inventory (NEO-FFI). Professional manual. Odessa, FL: Psycholog-ical Assessment Ressources; 1992.

[57] Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA. Manual for thestate-trait anxiety inventory. Palo Alto, CA: Consulting Psychologists Press;1983.

[58] Spielberger CD. Manual for the state-trait anger expression inventory. Odessa,FL: Psychological Assessment Resources; 1988.

[59] Hartig J, Moosbrugger H. Die “ARES-Skalen” zur Erfassung der individuellenBIS- und BAS-Sensitivität: Entwicklung einer Lang- und einer Kurzfassung.Zeitschrift für Differentielle und Diagnostische Psychologie 2003;24:291e308.

[60] Steyer R, Schwenkmezger P, Notz P, Eid M. Der Mehrdimensionale Befin-dlichkeitsfragebogen (MDBF). Handanweisung. Göttingen: Hogrefe; 1997.

[61] Van Honk J, Schutter DJ, d’Alfonso AA, Kessels RP, de Haan EH. 1 hz rTMS overthe right prefrontal cortex reduces vigilant attention to unmasked but not tomasked fearful faces. Biological Psychiatry 2002;52(4):312e7.

[62] Smith EE, Jonides J. Storage and executive processes in the frontal lobes.Science 1999;283(5408):1657e61.

[63] Owen AM, McMillan KM, Laird AR, Bullmore E. N-back working memoryparadigm: a meta-analysis of normative functional neuroimaging studies.Human Brain Mapping 2005;25(1):46e59.

[64] Antal A, Nitsche MA, Kruse W, Kincses TZ, Hoffmann KP, Paulus W. Directcurrent stimulation over V5 enhances visuomotor coordination by improvingmotion perception in humans. Journal of Cognitive Neuroscience2004;16(4):521e7.

[65] Nitsche MA, Koschack J, Pohlers H, Hullemann S, Paulus W, Happe S. Effects offrontal transcranial direct current stimulation on emotional state and pro-cessing in healthy humans. Frontiers in Psychiatry 2012;3:1e10.

[66] Nahas Z, Lomarev M, Roberts DR, Shastri A, Lorberbaum JP, Teneback C, et al.Unilateral left prefrontal transcranial magnetic stimulation (TMS) producesintensity-dependent bilateral effects as measured by interleaved BOLD fMRI.Biological Psychiatry 2001;50(9):712e20.

[67] Schutter DJ, van Honk J, d’Alfonso AA, Postma A, de Haan EH. Effects of slowrTMS at the right dorsolateral prefrontal cortex on EEG asymmetry and mood.Neuroreport 2001;12(3):445e7.

[68] D’Alfonso AAL, van Honk J, Hermans E, Postma A, de Haan EHF. Lateralityeffects in selective attention to threat after repetitive transcranial magneticstimulation at the prefrontal cortex in female subjects. Neuroscience Letters2000;280:195e8.

[69] Grisaru N, Bruno R, Pridmore S. Effect on the emotions of healthy individualsof slow repetitive transcranial magnetic stimulation applied to the prefrontalcortex. The Journal of ECT 2001;17(3):184e9.

[70] Jenkins J, Shajahan PM, Lappin JM, Ebmeier KP. Right and left prefrontaltranscranial magnetic stimulation at 1 Hz does not affect mood in healthyvolunteers. BMC Psychiatry 2002;2:1.

A. Weigand et al. / Brain Stimulation 6 (2013) 905e912912

[71] Koenigs M, Ukueberuwa D, Campion P, Grafman J, Wassermann E. Bilateralfrontal transcranial direct current stimulation: failure to replicate classicfindings in healthy subjects. Clinical Neurophysiology 2009;120(1):80e4.

[72] Plazier M, Joos K, Vanneste S, Ost J, De Ridder D. Bifrontal and bioccipitaltranscranial direct current stimulation (tDCS) does not induce mood changesin healthy volunteers: a placebo controlled study. Brain Stimulation2012;5(4):454e61.

[73] GaravanH,Pendergrass JC,RossTJ, SteinEA,RisingerRC.Amygdala responsetobothpositively and negatively valenced stimuli. Neuroreport 2001;12(12):2779e83.

[74] Gerber AJ, Posner J, Gorman D, Colibazzi T, Yu S, Wang Z, et al. An affectivecircumplex model of neural systems subserving valence, arousal, and cogni-tive overlay during the appraisal of emotional faces. Neuropsychologia2008;46(8):2129e39.

[75] Graf R, Nagler M, Jacobs AM. Faktorenanalyse von 57 Variablen der visuellenWorterkennung. Zeitschrift für Psychologie 2005;213(4):205e18.

[76] Ridding MC, Ziemann U. Determinants of the induction of cortical plasticity bynon-invasive brain stimulation in healthy subjects. Journal of Physiology2010;588(13):2291e304.