1 3

Exp Brain Res (2013) 227:43–52DOI 10.1007/s00221-013-3483-7

RESEARCH ARTICLE

Lateralized effects of prefrontal repetitive transcranial magnetic stimulation on emotional working memory

Anne Weigand · Simone Grimm · Antje Astalosch · Jia Shen Guo · Benny B. Briesemeister · Sarah H. Lisanby · Bruce Luber · Malek Bajbouj

Received: 15 January 2013 / Accepted: 10 March 2013 / Published online: 30 March 2013 © Springer-Verlag Berlin Heidelberg 2013

We found lateralized rTMS effects in the EMOBACK task accuracy for fear-related words, with enhanced performance after rTMS applied over the right DLPFC and impaired per-formance after rTMS applied over the left DLPFC. No sig-nificant stimulation effect could be found for anger-related and neutral words. Our findings are the first to demonstrate a causal role of the right DLPFC in working memory for negative, withdrawal-related words and provide further sup-port for a hemispheric lateralization of emotion processing.

Keywords rTMS · Emotion · Working memory · DLPFC · Frontal asymmetry

Introduction

Working memory is defined as the cognitive function that underlies our capability to form internal representations of information that is no longer accessible in the environ-ment (D’Esposito et al. 2000; Goldman-Rakic 1995). Over the past few decades, numerous neuroimaging studies have shown that the prefrontal cortex, especially the dorsolateral prefrontal cortex (DLPFC), plays a key role in mediating working memory processes (Braver et al. 2001; Curtis and D’Esposito 2003; Jonides et al. 1997).

The impact of emotional content on working memory has been examined only in a small number of studies, pro-viding conflicting results. At the behavioral level, negative stimuli were shown to have both impairing (Kensinger and Corkin 2003; Lindström and Bohlin 2012; Perlstein et al. 2002) and facilitating (Lindström and Bohlin 2011) effects on working memory, while an improved performance was demonstrated for positive stimuli (Lindström and Bohlin 2011; Perlstein et al. 2002). However, other stud-ies observed no impact of emotional content on working

Abstract Little is known about the neural correlates underlying the integration of working memory and emotion processing. We investigated the effects of low-frequency repetitive transcranial magnetic stimulation (rTMS) applied over the left or right dorsolateral prefrontal cortex (DLPFC) on emotional working memory. In a sham-controlled cross-over design, participants performed an emotional 3-back task (EMOBACK) at baseline and after stimulation (1 Hz, 15 min, 110 % of the resting motor threshold) in two subse-quent sessions. Stimuli were words assigned to the distinct emotion categories fear and anger as well as neutral words.

A. Weigand (*) · S. Grimm · A. Astalosch · J. S. Guo · B. B. Briesemeister · M. Bajbouj Cluster of Excellence “Languages of Emotion”, Freie Universität Berlin, Habelschwerdter Allee 45, 14195 Berlin, Germanye-mail: anne.weigand@fu-berlin.de

A. Weigand · S. Grimm · M. Bajbouj Department of Psychiatry, Campus Benjamin Franklin, Charité Berlin, 14050 Berlin, Germany

A. Weigand · S. Grimm · A. Astalosch · J. S. Guo · B. B. Briesemeister · M. Bajbouj Dahlem Institute for Neuroimaging of Emotion, Freie Universität Berlin, 14195 Berlin, Germany

S. Grimm Clinic for Affective Disorders and General Psychiatry, Psychiatric University Hospital Zurich, 8032 Zurich, Switzerland

S. H. Lisanby · B. Luber Department of Psychiatry and Behavioral Sciences, Duke University, Durham, NC 27710, USA

S. H. Lisanby · B. Luber Department of Psychology and Neuroscience, Duke University, Durham, NC 27710, USA

44 Exp Brain Res (2013) 227:43–52

1 3

memory (Döhnel et al. 2008; Grimm et al. 2012). Inter-estingly, neuroimaging findings revealed a relationship between DLPFC activity and emotional stimuli in work-ing memory paradigms: While Neta and Whalen (2011) found DLPFC activity generally increased, Perlstein et al. (2002) reported right DLPFC function to be influenced by emotional valence, with positive stimuli enhancing and negative stimuli decreasing neuronal activity. In contrast, Döhnel et al. (2008) showed no effect of emotional stimuli on DLPFC activity. Our own data (Grimm et al. 2012), however, provide evidence for a stronger recruitment of the left DLPFC during positive and right DLPFC during negative stimuli.

Considering the correlational nature of neuroimaging studies, repetitive transcranial magnetic stimulation (rTMS) provides a unique research tool to additionally investigate causal brain–behavior relationships in controlled experi-mental designs (Walsh and Pascual-Leone 2003). Based on the principle of electromagnetic induction of an electric field in the brain, rTMS can provide insight into separable neural circuits by non-invasive modulation of brain activ-ity. If rTMS is pulsed at a low frequency (~1 Hz), corti-cal excitability generally decreases, while higher frequency rTMS (5–20 Hz) can increase cortical excitability (Pascual-Leone et al. 1994; Chen et al. 1997). One commonly used method has employed trains of 1 Hz rTMS, which, when of sufficient duration (10–20 min), have been described as producing temporary “virtual lesions” in the targeted cortex (Pascual-Leone et al. 2000; Thut and Pascual-Leone 2010).

In working memory research, rTMS has been applied to the DLPFC in order to interfere with working memory function (e.g., Hamidi et al. 2009; Luber et al. 2007; Sandrini et al. 2008; Mottaghy et al. 2000). Investigators have also raised the question of whether specific functional processes of working memory are associated with a differ-ent hemispheric lateralization. For example, a study using single-pulse TMS reported that the left and right DLPFC are associated, respectively, with verbal and non-verbal working memory (Mull and Seyal 2001).

In previous research on emotion using rTMS, the DLPFC has also been stimulated unilaterally in order to investigate hemispheric lateralization in emotion processing. Several rTMS results in healthy subjects found contrasting emo-tional reaction patterns depending on the stimulation site: The left hemisphere was shown to be dominant for posi-tive emotions and the right hemisphere for negative emo-tions (Baeken et al. 2010; Leyman et al. 2008; Padberg et al. 2001; van Honk and Schutter 2006). These findings are in line with the valence hypothesis suggesting a hemi-spheric lateralization in emotion processing in dependence of stimulus valence (Davidson 1992; Davidson et al. 1999) and in contrast to the right hemisphere hypothesis propos-ing a dominant role for the right hemisphere in all emotion

processing (Borod et al. 1998). However, hemispheric lat-eralization of emotion processing may be organized along a dimension different from positive and negative valence, as growing evidence indicates that the left prefrontal cortex also processes anger, a negatively valenced emotion that is associated with approach motivation (Fox 1991; Harmon-Jones 2003). Therefore, the frontal asymmetry of emotion has been related to a motivational direction model with the dimensions of approach and withdrawal rather than positive and negative emotions (Harmon-Jones 2004; Harmon-Jones et al. 2010). In accordance with this model, an rTMS study by d’Alfonso et al. (2000) revealed that low-frequency rTMS over the right DLPFC caused selective attention toward angry faces, whereas low-frequency rTMS over the left DLPFC revealed the opposite effect, namely a selec-tive attention away from angry faces. Thus, the emotion anger has been associated with left prefrontal activity. In contrast, fear, a negative withdrawal-related emotion, has been associated with the right prefrontal activity in an rTMS study by van Honk et al. (2002). They demonstrated that low-frequency rTMS over the right prefrontal cortex sig-nificantly decreased vigilant attention to unmasked but not masked fearful faces. The authors suggested that this effect might have been induced by a more left-sided dominance in prefrontal processing through contralateral excitation after low-frequency rTMS at suprathreshold intensity (Nahas et al. 2001; Schutter et al. 2001).

Due to divergent findings coming from rTMS and neu-roimaging studies, the neural correlates underlying the integration of working memory and emotion processing remain a matter of debate. Based on these considerations, the present rTMS study aimed to assess for the first time the causal role of the DLPFC and potential hemispheric later-alization in working memory for processing approach- and withdrawal-related words.

Participants were presented with an adapted version of an emotional n-back paradigm (Grimm et al. 2012) that requires participants to decide whether the current stimulus matches the one presented n trials earlier. This paradigm involves a number of key processes within working memory includ-ing monitoring, updating and manipulating the remembered information (Owen et al. 2005) and reliably engages differ-ent prefrontal brain regions (Braver et al. 2001). The stimuli of the emotional n-back (EMOBACK) task consist of emo-tional words related to the distinct emotions fear and anger as well as neutral words. Accordingly, the EMOBACK task provides a novel means of studying the interface between working memory and emotion. Participants performed two parallel versions of the EMOBACK task after MRI-guided active and sham low-frequency rTMS applied over either the left or right DLPFC.

Based on the motivational direction model (Harmon-Jones 2004; Harmon-Jones et al. 2010), differential

45Exp Brain Res (2013) 227:43–52

1 3

predictions can be made for the EMOBACK task perfor-mance depending on the side of prefrontal stimulation: We hypothesized that low-frequency rTMS applied over the right DLPFC impairs particularly the processing of fear, whereas low-frequency rTMS applied over the left DLPFC impairs in particular the processing of anger, affect-ing working memory for stimuli related to these distinct emotions. Since previous findings are mixed, no specific predictions were made about the effects of the expected impairment in emotion processing on working memory task performance.

Methods

Subjects

Twenty-eight healthy volunteers (14 females), aged between 20 and 35 years (mean age 25.8 ± 4.5), were recruited to participate in this study. All subjects were native German speakers, had normal or corrected-to-normal vision and were without any neurological or psychiatric condition, as determined by the short version of the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV (SCID; Wittchen et al. 1997). None of the subjects had any contradictions to rTMS (Rossi et al. 2009; Wassermann 1998), and all were naïve to rTMS. They were right-handed as determined by the Edinburgh Handedness Questionnaire (EHI; Oldfield 1971). A verbal intelligence test (Mehrfachwahl-Wortschatz-Intelligenztest, MWT-B; Lehrl 2005) showed that IQ levels of all partici-pants were at or above the range of the norm. The study was approved by the Institutional Review Board of the Charité University Medicine (Berlin, Germany). Written informed consent was obtained in accordance to the ethical guidelines of the Declaration of Helsinki, and all subjects were finan-cially compensated.

Experimental design

We used the EMOBACK task, an adapted version of a working memory task with emotional words, which has been developed in a previous experiment at our laboratory (Grimm et al. 2012).

Stimuli were German nouns taken from the Discrete Emotion Norms for Nouns–Berlin Affective Word List (DENN-BAWL; Briesemeister et al. 2011a). The DENN-BAWL assigns nouns taken from the BAWL-R (Võ et al. 2009) to discrete emotion categories, namely happiness, sadness, fear, anger and disgust. For the EMOBACK task, words strongly related to fear and anger were selected and controlled for valence and arousal (t’s < 1). Additionally, neutral words were included in the task. All three emotion conditions were matched on imageability, number of letters, syllables, phonemes, frequency and orthographic neighbors according to the BAWL-R (F’s > 1). Two parallel sets of words were created with the described stimulus characteris-tics for the active and sham rTMS condition (t’s < 1), each set containing 30 words per condition. Table 1 provides an overview of the stimulus characteristics used in the two par-allel versions of the EMOBACK task.

The stimulus material was presented within a 3-back working memory task (Fig. 1a). Participants viewed sequences of words and had to decide for each word whether it was the same as the one presented three trials earlier. They were instructed to press a button with the same finger (right index or middle finger) as quickly as possible after the presentation of a target word. The words were presented as white uppercase letters in the center of a black screen for 500 ms with an interstimulus interval (ISI) of 1,500 ms. Par-ticipants were seated in a comfortable chair in a silent room, and the monitor was placed at eye level 90 cm in front of the subjects. The task was programmed using Presentation software (Version 14.5, Neurobehavioral Systems Inc., San Francisco, CA, USA).

Table 1 Stimulus characteristics used in the two parallel versions of the EMOBACK task

M mean, SD standard deviation

Word set 1 Word set 2

Fear Anger Neutral Fear Anger Neutral

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

Valence −1.7 (0.5) −1.7 (0.5) 0.0 (0.3) −1.8 (0.4) −1.8 (0.4) 0.0 (0.3)

Arousal 3.9 (0.5) 3.8 (0.3) 2.4 (0.3) 3.9 (0.5) 3.8 (0.4) 2.3 (0.3)

Imageability 4.2 (1.3) 4.0 (1.2) 4.3 (1.5) 4.5 (1.3) 4.3 (0.9) 4.3 (1.2)

Letters 6.3 (1.2) 6.6 (1.5) 6.4 (1.2) 6.3 (1.2) 6.4 (1.3) 6.2 (1.2)

Syllables 2.0 (0.6) 2.1 (0.6) 2.1 (0.6) 2.1 (0.7) 2.1 (0.7) 2.2 (0.5)

Phonemes 5.5 (1.3) 5.8 (1.5) 5.5 (1.3) 5.7 (1.4) 5.6 (1.4) 5.4 (1.0)

Frequency 20.9 (42.0) 16.0 (29.5) 14.7 (31.2) 18.9 (67.8) 12.3 (16.1) 22.0 (63.6)

Orthographic neighbors

0.8 (0.8) 1.0 (1.9) 1.0 (1.9) 1.0 (1.0) 1.2 (1.2) 1.0 (1.0)

46 Exp Brain Res (2013) 227:43–52

1 3

Participants performed a practice run prior to the experi-ment. In the actual experiment (Fig. 1b), 15 blocks were presented at baseline and after rTMS application in two successive sessions (active and sham rTMS), separated by a 45-min washout period to avoid carry-over effects (Berm-pohl et al. 2005; Thut and Pascual-Leone 2010). Each block contained 18 words of the same emotional condition, fol-lowed by a fixation period of 10–14 s. The block order of the three conditions (fear, anger and neutral) was randomized across the task. In total, 1,080 stimuli (360 stimuli per con-dition) were presented over the course of the experiment.

After each rTMS session, participants rated all words of the two parallel sets of the EMOBACK task on a 7-point scale regarding emotional valence and on a 5-point scale regarding emotional arousal according to the BAWL-R norms (Võ et al. 2009).

Repetitive transcranial magnetic stimulation

Biphasic magnetic stimulation was delivered with a Medtronic stimulator (MagPro X100 with MagOption, MagVenture, Farum, Denmark) connected to a figure-eight coil (MCF-B65). Before rTMS application, the optimal stimulation site of the left and right DLPFC (Brodmann area 9/46) was marked individually using magnetic reso-nance imaging (MRI) non-stereotactic guidance (Peleman et al. 2010). For each participant, we acquired an anatomical high-resolution image on a Siemens Trio 3T scanner with conventional parameters (176 T1-weighted images with a slice thickness of 1 mm) approximately a week prior to the rTMS experiment. Neuronavigation (eXimia Navigated Brain Stimulation (NBS), Nexstim, Helsinki, Finland) was used to position the coil with an adjustable arm perpen-dicular to the marked stimulation point on the skull and to

maintain its precise position throughout the experiment. For sham stimulation, a placebo coil (MCF-P-B65), which looks and sounds identical to the real coil without delivering a magnetic field, was positioned in exactly the same manner.

Surface electromyography was measured in the right first dorsal interosseus (FDI) muscle to determine the rest-ing motor threshold as the minimum intensity of stimulator output necessary to elicit motor potentials of at least 50 μV peak-to-peak amplitude in 5/10 stimulations. For this pur-pose, single-pulse TMS was applied with intervals of at least 8 s to avoid carryover effects. TMS intensity was set at 110 % of the resting motor threshold.

Participants received offline a single 15-min train at 1 Hz (900 pulses) over either the left or right DLPFC after the baseline measurement of the EMOBACK task. Half of the subjects received active and sham rTMS over the left, and half received it over the right DLPFC. The sequence of ses-sions (active and sham rTMS) was randomized across sub-jects in a single-blind crossover design. Subjects performed the task immediately after the stimulation in the same labo-ratory room. The rTMS parameters were well within sug-gested safety guidelines (Rossi et al. 2009; Wassermann 1998). During stimulation, all subjects wore earplugs.

Control variables

The Multidimensional Mood Questionnaire (MDBF; Steyer et al. 1997) was applied prior to the EMOBACK baseline measurement (Tpre) and immediately after the stimulation (Tpost). Anxiety levels were measured with the state version of the State-Trait Anxiety Inventory (STAI-S; Spielberger et al. 1983) prior to rTMS applications. Participants’ pre-disposition to anxiety and to anger was assessed by the trait-version of the STAI (STAI-T; Spielberger et al. 1983)

a

b

Fig. 1 a Example of a trial sequence in the emotional 3-back task (EMOBACK). b Schematic illustration of the experiment

47Exp Brain Res (2013) 227:43–52

1 3

and of the State-Trait Anger Expression Inventory (STAXI-T; Spielberger 1988). In addition, the short version of the Action Regulating Emotion Systems questionnaire (ARES; Hartig and Moosbrugger 2003) was completed in order to control individual differences in the Behavioral Inhibition and the Behavioral Activation System scales (BIS/BAS; Carver and White 1994).

Statistical analysis

Accuracy was defined as the ratio of correct responses (hits and correct rejections) to total number of stimuli. Mean reaction times of hits were additionally analyzed. One-way analyses of variance (ANOVAs) for the factor Emo-tion (fear, anger and neutral) were applied for accuracy and reaction times at baseline to investigate the impact of emotional content on working memory. In order to assess rTMS effects, accuracy and reaction time differences from baseline were analyzed separated for stimulation side using repeated measures ANOVAs with a 2 × 3 × 2 facto-rial design: Stimulation (active and sham), Emotion (fear, anger and neutral) as within-subjects factors and Order (first active and first sham rTMS) as between-subjects factor. Initial analyses revealed no effect of gender and this fac-tor was subsequently dropped from the analyses. For the valence and arousal ratings, 2 × 3 ANOVAs with Stimula-tion (active and sham) and Emotion (fear, anger and neutral) were applied. To evaluate changes in subjects’ current emo-tional state, all MDBF subscales were analyzed using 2 × 2 ANOVA’s with Stimulation (active and sham) and Time (Tpre and Tpost) in dependence of the side of rTMS applica-tion. When appropriate, p values were corrected using the Greenhouse-Geisser procedure. Further statistical analysis was done using post hoc t tests. All tests were two-tailed and the significant threshold was set at a probability of p < 0.05. Statistical analyses were carried out using PASW (Predic-tive Analysis SoftWare, Version 18.0, SPSS Inc., Chicago, IL, USA).

Results

Two subjects had to be excluded from the study due to a headache after the motor threshold measurement in one subject and inexact coil positioning in the other. Thus, data from a total of 26 participants (14 females) were included in the analyses. All remaining subjects tolerated the rTMS applications well.

Demographics and control variables

There were no significant group differences between par-ticipants who received rTMS over the left DLPFC compared

to those who received rTMS over the right DLPFC with respect to age, verbal intelligence (MWT-B), anxiety level before active and sham stimulation (STAI-Sactive and STAI-Ssham) and assessed personality traits (STAI-T, STAXI-T, BIS/BAS scales). More detailed sample characteristics are reported in Table 2.

With regard to possible mood effects of rTMS, only the MDBF calmness–restlessness subscale revealed a sig-nificant interaction effect between Stimulation and Time [F(1,11) = 10.24, p < 0.01, η2 = 0.48] after left-sided DLPFC stimulation, indicating higher feelings of restless-ness after the sham as compared to the active rTMS applica-tion. No significant main or interaction effects were found after stimulating the right DLPFC. Importantly, there were no mood changes caused by active low-frequency rTMS applied over the left or right DLPFC. Table 3 provides details of the MDBF results in dependence of the stimulation side.

EMOBACK task performance

Impact of emotional content

The baseline measurement revealed no significant effect of emotional content on accuracy (fear: 93.3 ± 2.6 %, anger:

Table 2 Demographics and individual characteristics

M mean, SD standard deviation, IQ intelligence quotient, MWT-B Mehrfachwahl–Wortschatz-Intelligenztest (verbal intelligence test), STAI State-Trait Anxiety Inventory, STAXI State-Trait Anger Expres-sion Inventory, BIS Behavioral Inhibition System, BAS Behavioral Approach Systema Independent samples t tests, two-tailed

Stimulation side ta p

Left (N = 12) Right (N = 14)

M (SD) M (SD)

Age 26.8 (5.1) 25.1 (4.3) −.960 .347

MWT-B

IQ 109.0 (14.2) 107.4 (7.7) −.357 .724

STAI-S

Sactive 35.4 (7.6) 35.1 (8.2) −.111 .913

Ssham 37.5 (9.5) 33.8 (7.1) −1.043 .309

STAI-T 35.1 (7.9) 33.5 (6.0) −.580 .567

STAXI-T 7.8 (2.7) 8.1 (2.8) .294 .771

BIS

Anxiety 2.0 (0.5) 2.3 (0.5) 1.393 .176

Frustration 2.2 (0.5) 2.1 (0.5) −.445 .661

Total 2.1 (0.5) 2.2 (0.5) .521 .607

BAS

Drive 3.5 (0.4) 3.6 (0.3) 1.308 .203

Gratification 3.3 (0.5) 3.5 (0.4) 1.466 .156

Total 3.4 (0.4) 3.6 (0.3) 1.580 .127

48 Exp Brain Res (2013) 227:43–52

1 3

93.8 ± 2.5 %, neutral: 93.7 ± 2.6 %) and reaction times (fear: 765.2 ± 151.7 ms, anger: 781.2 ± 152.2 ms and neu-tral: 787.6 ± 140.2 ms).

Stimulation effects

The ANOVA on accuracy for right-sided stimulation revealed a significant Stimulation by Emotion interaction [F(2,24) = 3.47, p < 0.05, η2 = 0.22] indicating differential stimulation effects between emotion conditions. Post hoc t tests for the three emotion conditions between the EMO-BACK task performance at baseline and after rTMS showed that accuracy changed only for fear-related words, increas-ing significantly [t(1,13) = −2.24, p < 0.05] after right-sided active DLPFC stimulation. After left-sided active DLPFC stimulation, there was a trend toward a decrease in accuracy for fear-related words [t(1,11) = 2.05, p = 0.065]. Accord-ingly, further analyses comparing each of the emotion conditions across stimulation sides indicated a significant difference only for fear-related words [t(2,24) = −3.01, p < 0.01]. There was no significant order effect for accuracy as tested with the between-subjects factor order for right- and left-sided stimulation. Figure 2 illustrates the accuracy differences from baseline for the three emotion conditions.

The ANOVA on reaction times for right- and left-sided stimulation revealed no main or interaction effects. An order effect could be excluded due to the non-significance of the between-subjects factor order.

Word ratings

The results of the individual word ratings are presented in Table 4. ANOVAs for valence and arousal ratings showed no significant main effect of Stimulation. Consequently, the emotional stimuli for the active and sham condi-tion had been equally rated by the participants. The main

effect of Emotion was significant for both valence ratings [F(2,50) = 409.08, p < 0.001, η2 = 0.94] and arousal ratings [F(2,50) = 134.04, p < 0.001, η2 = 0.84]. Post hoc t tests revealed that fear- and anger-related words were rated more unpleasant [t(1,25) = 23.63, p < 0.001 and t(1,25) = 21.49, p < 0.001] and more arousing [t(1,25) = 13.84, p < 0.001 and t(1,25) = 12.19, p < 0.001] than neutral words. Between fear- and anger-related words, the valence ratings and arousal ratings did not differ significantly.

Discussion

The present study investigated the effects of low-frequency rTMS applied over the DLPFC on the performance in an emotional working memory task. By including emotional words in a n-back paradigm, we were able to explore the interface between working memory and emotion. We were

Fig. 2 Accuracy differences from baseline for the three emotion con-ditions (fear, anger and neutral) after right and left DLPFC stimula-tion. The illustration combines the sham conditions, since no differ-ences were found between stimulation sides

Table 3 Mood assessment before and after rTMS application (pre and post)

M mean, SD standard deviation

Stimulation side

Left (N = 12) Right (N = 14)

Active Sham Active Sham

Pre Post Pre Post Pre Post Pre Post

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

Good versus bad mood

15.4 (2.6) 15.2 (2.9) 15.8 (2.1) 15.8 (2.0) 17.1 (2.0) 17.1 (2.8) 17.4 (2.3) 17.3 (2.1)

Alertness versus tiredness

11.4 (3.3) 12.6 (2.4) 11.6 (2.4) 13.2 (2.4) 11.9 (2.7) 11.4 (2.4) 11.1 (1.5) 11.8 (1.9)

Calmness versus restlessness

16.3 (2.6) 16.7 (2.5) 17.1 (2.7) 16.0 (2.6) 16.8 (2.5) 16.5 (2.1) 17.4 (2.4) 17.1 (2.5)

Total 43.1 (7.6) 44.4 (5.6) 44.5 (5.7) 44.9 (5.2) 45.8 (6.2) 45.0 (5.7) 45.9 (5.1) 46.1 (5.1)

49Exp Brain Res (2013) 227:43–52

1 3

specifically interested to assess for the first time the distinct involvement of right and left DLPFC in working memory processes for fear- and anger-related words.

We found differential emotion-related effects in the EMOBACK task after rTMS application in dependence on the side of prefrontal stimulation. A significant effect of stimulation side, however, was observed for fear-related words exclusively: Participants performed more accurately after low-frequency rTMS applied over the right DLPFC and less accurately after low-frequency applied over the left DLPFC, in both cases without a change in reaction time. With regard to anger-related and neutral words, no signifi-cant stimulation effect in task performance could be found after left- or right-sided rTMS application.

Our findings support to some extent a frontal asymme-try of emotion processing as proposed by the motivational direction model (Harmon-Jones 2004; Harmon-Jones et al. 2010). In line with this model, our study revealed lateral-ized stimulation effects for fear-related words, indicating that negative withdrawal-related emotions are associated with the right DLPFC. It can be assumed that low-frequency rTMS of the DLPFC produces a decrease in neuronal activ-ity in the stimulated area and a contralateral excitation when applied at suprathreshold intensities (Nahas et al. 2001; Pascual-Leone et al. 1998; Schutter et al. 2001). There-fore, with regard to our lateralized findings for fear-related words, right-sided rTMS application might have impaired the emotion processing for fear-related words, whereas left-sided rTMS application might have enhanced the emotion processing for fear-related words through a contralateral disinhibition. Interestingly, working memory performance for fear-related words was found to be improved after stimu-lating the right DLPFC and impaired after stimulating the left DLPFC. A potential mechanism underlying our results might be seen in the involvement of the DLPFC in atten-tional control (Leyman et al. 2008; Sandrini et al. 2008; Vanderhasselt et al. 2006, 2007). More specifically, the amygdalae have been found to be under prefrontal cortical top-down control, allowing for the regulation of emotion, including the inhibition of emotional information (Bae-ken et al. 2010; Hariri et al. 2000; Ochsner et al. 2009). In our study, the stimulation-induced inhibition of the right DLPFC might have reduced the attentional bias toward fear-related words, resulting in enhanced working memory

task performance through a prioritized processing of task-relevant information, whereas the contralateral activation of the right DLPFC might have enhanced the attentional bias toward fear-related words, resulting in impaired working memory task performance through a diminished ability to suppress task-irrelevant emotional information. This poten-tial explanation is supported by previous findings indicat-ing that low-frequency rTMS over the right DLPFC reduces vigilant attention to fear-related stimuli (van Honk et al. 2002).

Furthermore, we did not find evidence of left-hemi-sphere dominance for processing anger-related words as demonstrated in previous research (Harmon-Jones 2003). Whereas low-frequency rTMS over the right DLPFC has been shown to cause selective attention toward angry faces, low-frequency rTMS over the left DLPFC resulted in selective attention away from angry faces (D’Alfonso et al. 2000). Our data, however, indicated that anger- and fear-related words are differentially processed, providing further support for a dissociation of negative approach- and withdrawal-related emotions according to the motivational direction model (Harmon-Jones 2004; Harmon-Jones et al. 2010).

With regard to neutral words, no significant stimulation effects on task performance were found after rTMS appli-cation. In contrast, a previous study using a non-emotional verbal n-back task showed impaired working memory per-formance after inhibitory rTMS applied to either the right or left DLPFC (Mottaghy et al. 2000). The stimuli imple-mented in our EMOBACK working memory task, how-ever, consisted of nouns instead of letters. More research is needed, though, to investigate the neural correlates underly-ing working memory when processing words.

The present finding of a causal role of the right DLPFC during the EMOBACK task for withdrawal-related stimuli extends our previous finding of DLPFC-activation during emotional working memory (Grimm et al. 2012). Our neuro-imaging data revealed a trend toward a stronger recruitment of the left DLPFC during positive and right DLPFC during negative words, supporting an emotion lateralization as sug-gested by the valence hypothesis (Davidson 1992; Davidson et al. 1999). The stimulus material of our previous study, however, did not allow disentangling negative approach-related from negative withdrawal-related emotions.

Table 4 Valence and arousal ratings of the stimulus material

M mean, SD standard deviation

Active Sham

Fear Anger Neutral Fear Anger Neutral

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

Valence −1.4 (0.3) −1.5 (0.3) 0.1 (0.3) −1.4 (0.4) −1.5 (0.4) 0.1 (0.3)

Arousal 3.1 (0.6) 3.1 (0.6) 1.9 (0.7) 3.1 (0.6) 3.1 (0.6) 2.0 (0.7)

50 Exp Brain Res (2013) 227:43–52

1 3

In line with the study by Döhnel et al. (2008) and our previous study (Grimm et al. 2012), we found no impact of emotional content on working memory task performance at baseline.

Furthermore, we found that the current emotional state was not influenced by rTMS as controlled by the MDBF assessed before and after stimulation. This finding supports previous research examining mood effects of one session of low-frequency prefrontal rTMS causing no immediate mood changes in healthy volunteers (Grisaru et al. 2001; Jenkins et al. 2002). It is important to control for mood also with regard to previous research demonstrating that posi-tive affect enhances overall performance on certain types of cognitive tasks (Herrington et al. 2005). For example, Gray (2001) found that participants performed better on a verbal working memory task after watching a pleasant video and, additionally, showed greater left DLPFC activity during task performance (Gray et al. 2002). By controlling poten-tial mood-related facilitation effects in task performance, our results cannot be attributed to different emotional states of the participants. In addition, we controlled for individual anxiety levels prior to stimulation, which might influence rTMS effects (Vanderhasselt et al. 2010) as well as for per-sonality traits associated with a potential influence on emo-tion processing (Balconi and Mazza 2010; Balconi and Fer-rari 2012; Harmon-Jones et al. 2010; Putman et al. 2004).

Because valence and arousal may influence word pro-cessing (Hofmann et al. 2009) and may be controlled by different neural systems (Garavan et al. 2001; Gerber et al. 2008), an advantage of this study is the carefully matched selection of words with equal valence and arousal levels for fear- and anger-related words according to the BAWL norms. Therefore, our findings provide further evidence for discrete emotion effects that cannot be explained by the two parameters valence and arousal alone (Briesemeister et al. 2011b). We additionally controlled our stimuli sets with regard to lexical and sublexical variables whose relevance has been shown in tasks relying on word material (Graf et al. 2005). To further control our stimulus material, we asked participants to rate the valence and arousal levels of all pre-sented words individually. The results of the word ratings of our sample elicited a successful matching of our stimuli sets according to the BAWL norms. Importantly, we also found that the stimulus material for the stimulation conditions had been equally rated by the participants. This indicates that the reported rTMS effects cannot be explained by different emotional content of the stimuli used in the active as com-pared to sham condition.

There are a number of limitations to this study that are important to consider. We did not include an individual word rating to assess the distinct emotions fear and anger. With regard to our finding that anger-related stimuli are not asso-ciated with the left DLPFC, this could have clarified whether

the selected anger-related words from the DENN-BAWL were possibly not perceived as anger- and, thus, approach-related stimuli by our participants. Although the presented study was limited by sample size, an important strength of this experiment was the use of individual anatomical struc-tures to locate the exact stimulation targets within the left or right DLPFC. It has been shown that the effectiveness of rTMS can be optimized by using MRI-guided stimula-tion (Sack et al. 2009; Sparing et al. 2008). Furthermore, we implemented 45 min of washout period between rTMS ses-sions (Bermpohl et al. 2005; Thut and Pascual-Leone 2010), but the stimulation effects might have lasted longer. How-ever, we counterbalanced active and sham rTMS conditions across participants and failed to observe order effects. The counterbalanced design also allowed us to control for poten-tial learning effects as participants performed the EMO-BACK task repeatedly. To thoroughly address this important concern, a practice run was additionally presented prior to the actual experiment.

In conclusion, the present rTMS study extends the find-ings of our previous neuroimaging study (Grimm et al. 2012) investigating neural mechanisms underlying the inte-grating of working memory and emotion. This is the first study to demonstrate a causal role of the right DLPFC in working memory for fear-related words. Our findings pro-vide further support for a hemispheric lateralization of emo-tion processing with regard to negative, withdrawal-related stimuli as proposed by the motivational direction model (Harmon-Jones 2004; Harmon-Jones et al. 2010).

References

Baeken C, De Raedt R, Van Schuerbeek P, Vanderhasselt MA, De Mey J, Bossuyt A, Luypaert R (2010) Right prefrontal HF-rTMS attenuates right amygdala processing of negatively valenced emotional stimuli in healthy females. Behav Brain Res 214: 450–455

Balconi M, Ferrari C (2012) rTMS stimulation on left DLPFC affects emotional cue retrieval as a function of anxiety level and gender. Depress Anxiety 29:976–982

Balconi M, Mazza G (2010) Lateralisation effect in comprehension of emotional facial expression: a comparison between EEG alpha band power and behavioural inhibition (BIS) and activation (BAS) systems. Laterality 15:361–384

Bermpohl F, Fregni F, Boggio PS et al (2005) Effect of low-frequency transcranial magnetic stimulation on an affective go/no-go task in patients with major depression: role of stimulation site and depres-sion severity. Psychiatry Res 141:1–13

Borod JC, Obler KL, Erhan HM et al (1998) Right hemisphere emo-tional perception: evidence across multiple channels. Neuropsy-chology 12:446–458

Braver TS, Barch DM, Kelley WM et al (2001) Direct comparison of prefrontal cortex regions engaged by working and long-term memory tasks. Neuroimage 14:48–59

Briesemeister BB, Kuchinke L, Jacobs AM (2011a) Discrete emo-tion norms for nouns: Berlin affective word list (DENN-BAWL). Behav Res Methods 43:441–448

51Exp Brain Res (2013) 227:43–52

1 3

Briesemeister BB, Kuchinke L, Jacobs AM (2011b) Discrete emotion effects on lexical decision response times. PLoS ONE 6(8):e23743

Carver CS, White TL (1994) Behavioral inhibition, behavioral acti-vation, and affective responses to impending reward and punish-ment: the BIS/BAS scales. J Pers Soc Psychol 67:319–333

Chen R, Gerloff C, Classen J, Wassermann EM, Hallett M, Cohen LG (1997) Safety of different inter-train intervals for repetitive transcranial magnetic stimulation and recommendations for safe ranges of stimulation parameters. Electroencephalogr Clin Neu-rophysiol 105:415–421

Curtis CE, D’Esposito M (2003) Persistent activity in the prefrontal cortex during working memory. Trends Cogn Sci 7:415–423

D’Alfonso AAL, van Honk J, Hermans E, Postma A, de Haan EHF (2000) Laterality effects in selective attention to threat after repeti-tive transcranial magnetic stimulation at the prefrontal cortex in female subjects. Neurosci Lett 280:195–198

Davidson RJ (1992) Anterior cerebral asymmetry and the nature of emotion. Brain Cogn 20:125–151

Davidson RJ, Abercrombie H, Nitschke JB, Putnam K (1999) Regional brain function, emotion and disorders of emotion. Curr Opin Neu-robiol 9:228–234

D’Esposito M, Postle BR, Rypma B (2000) Prefrontal cortical contri-butions to working memory: evidence from event-related fMRI studies. Exp Brain Res 133:3–11

Döhnel K, Sommer M, Ibach B, Rothmayr C, Meinhardt J, Hajak G (2008) Neural correlates of emotional working memory in patients with mild cognitive impairment. Neuropsychologia 46:37–48

Fox NA (1991) If it’s not left, it’s right. Electroencephalograph asym-metry and the development of emotion. Am Psychol 46:863–872

Garavan H, Pendergrass JC, Ross TJ, Stein EA, Risinger RC (2001) Amygdala response to both positively and negatively valenced stimuli. NeuroReport 12:2779–2783

Gerber AJ, Posner J, Gorman D et al (2008) An affective circumplex model of neural systems subserving valence, arousal, and cogni-tive overlay during the appraisal of emotional faces. Neuropsy-chologia 46:2129–2139

Goldman-Rakic PS (1995) Architecture of the prefrontal cortex and the central executive. Ann N Y Acad Sci 769:71–83

Graf R, Nagler M, Jacobs AM (2005) Faktorenanalyse von 57 Vari-ablen der visuellen Worterkennung. Zeitschrift für Psychologie 213:205–218

Gray JR (2001) Emotional modulation of cognitive control: approach-withdrawal states double-dissociate spatial from verbal two-back task performance. J Exp Psychol Gen 130:436–452

Gray JR, Braver TS, Raichle ME (2002) Integration of emotion and cognition in the lateral prefrontal cortex. Proc Natl Acad Sci USA 99:4115–4120

Grimm S, Weigand A, Kazzer P, Jacobs AM, Bajbouj M (2012) Neural mechanisms underlying the integration of emotion and working memory. Neuroimage 61:1188–1194

Grisaru N, Bruno R, Pridmore S (2001) Effect on the emotions of healthy individuals of slow repetitive transcranial magnetic stimu-lation applied to the prefrontal cortex. J ECT 17:184–189

Hamidi M, Tononi G, Postle BR (2009) Evaluating the role of prefron-tal and parietal cortices in memory-guided response with repetitive transcranial magnetic stimulation. Neuropsychologia 47:295–302

Hariri AR, Bookheimer SY, Mazziotta JC (2000) Modulating emo-tional responses: effects of a neocortical network on the limbic system. NeuroReport 11:43–48

Harmon-Jones E (2003) Early career award. Clarifying the emotive functions of asymmetrical frontal cortical activity. Psychophysiol-ogy 40:838–848

Harmon-Jones E (2004) Contributions from research on anger and cognitive dissonance to understanding the motivational functions of asymmetrical frontal brain activity. Biol Psychol 67:51–76

Harmon-Jones E, Gable PA, Peterson CK (2010) The role of asym-metric frontal cortical activity in emotion-related phenomena: a review and update. Biol Psychol 84:451–462

Hartig J, Moosbrugger H (2003) Die “ARES-Skalen” zur Erfassung der individuellen BIS- und BAS-Sensitivität: Entwicklung einer Lang- und einer Kurzfassung. Zeitschrift für Differentielle und Diagnostische Psychologie 24:291–308

Herrington JD, Mohanty A, Koven NS et al (2005) Emotion-modu-lated performance and activity in left dorsolateral prefrontal cor-tex. Emotion 5:200–207

Hofmann MJ, Kuchinke L, Tamm S, Võ ML, Jacobs AM (2009) Affec-tive processing within 1/10th of a second: high arousal is neces-sary for early facilitative processing of negative but not positive words. Cogn Affect Behav Neurosci 9(4):389–397

Jenkins J, Shajahan PM, Lappin JM, Ebmeier KP (2002) Right and left prefrontal transcranial magnetic stimulation at 1 Hz does not affect mood in healthy volunteers. BMC Psychiatry 2:1

Jonides J, Schumacher EH, Smith EE, Lauber EJ, Awh E, Minoshima S, Koeppe RA (1997) Verbal working memory load affects regional brain activation as measured by PET. J Cogn Neurosci 9:462–475

Kensinger EA, Corkin S (2003) Effect of negative emotional content on working memory and long-term memory. Emotion 3:378–393

Lehrl S (2005) Mehrfachwahl-Wortschatz-Intelligenztest MWT-B. Spitta Verlag, Balingen

Leyman L, De Raedt R, Vanderhasselt MA, Baeken C (2008) Influ-ence of high-frequency repetitive transcranial magnetic stimu-lation over the dorsolateral prefrontal cortex on the inhibition of emotional information in healthy volunteers. Psychol Med 39:1019–1028

Lindström BR, Bohlin G (2011) Emotion processing facilitates work-ing memory performance. Cogn Emot 25:1196–1204

Lindström BR, Bohlin G (2012) Threat-relevance impairs executive functions: negative impact on working memory and response inhi-bition. Emotion 12(2):384–393

Luber B, Kinnunen LH, Rakitin BC, Ellsasser R, Stern Y, Lisanby SH (2007) Facilitation of performance in a working memory task with rTMS stimulation of the precuneus: frequency- and time-depend-ent effects. Brain Res 1128:120–129

Mottaghy FM, Krause BJ, Kemna LJ et al (2000) Modulation of the neuronal circuitry subserving working memory in healthy human subjects by repetitive transcranial magnetic stimulation. Neurosci Lett 280(3):167–170

Mull BR, Seyal M (2001) Transcranial magnetic stimulation of left prefrontal cortex impairs working memory. Clin Neurophysiol 112:1672–1675

Nahas Z, Lomarev M, Roberts DR et al (2001) Unilateral left prefron-tal transcranial magnetic stimulation (TMS) produces intensity-dependent bilateral effects as measured by interleaved BOLD fMRI. Biol Psychiatry 50:712–720

Neta M, Whalen PJ (2011) Individual differences in neural activity during a facial expression vs. identity working memory task. Neu-roimage 56:1685–1692

Ochsner KN, Ray RR, Hughes B et al (2009) Bottom-up and top-down processes in emotion generation: common and distinct neural mechanisms. Psychol Sci 20:1322–1331

Oldfield RC (1971) The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9:97–113

Owen AM, McMillan KM, Laird AR, Bullmore E (2005) N-back working memory paradigm: a meta-analysis of normative func-tional neuroimaging studies. Hum Brain Mapp 25:46–59

Padberg F, Juckel G, Prassl A et al (2001) Prefrontal cortex modulation of mood and emotionally induced facial expressions: a transcra-nial magnetic stimulation study. J Neuropsychiatry Clin Neurosci 13:206–212

52 Exp Brain Res (2013) 227:43–52

1 3

Pascual-Leone A, Valls-Sole J, Wassermann EM, Hallett M (1994) Responses to rapid-rate transcranial magnetic stimulation of the human motor cortex. Brain 117:847–858

Pascual-Leone A, Tormos JM, Keenan J, Tarazona F, Canete C, Catala MD (1998) Study and modulation of human cortical excitabil-ity with transcranial magnetic stimulation. J Clin Neurophysiol 15:333–343

Pascual-Leone A, Walsh V, Rothwell J (2000) Transcranial magnetic stimulation in cognitive neuroscience—virtual lesion, chronome-try, and functional connectivity. Curr Opin Neurobiol 10:232–237

Peleman K, Van Schuerbeek P, Luypaert R et al (2010) Using 3D-MRI to localize the dorsolateral prefrontal cortex in TMS research. World J Biol Psychiatry 11:425–430

Perlstein WM, Elbert T, Stenger VA (2002) Dissociation in human pre-frontal cortex of affective influences on working memory-related activity. Proc Natl Acad Sci USA 99:1736–1741

Putman P, Hermans E, van Honk J (2004) Emotional stroop per-formance for masked angry faces: it’s BAS, not BIS. Emotion 4:305–311

Rossi S, Hallett M, Rossini PM, Pascual-Leone A (2009) Safety, ethi-cal considerations, and application guidelines for the use of tran-scranial magnetic stimulation in clinical practice and research. Clin Neurophysiol 120:2008–2039

Sack AT, Cohen Kadosh R, Schuhmann T, Moerel M, Walsh V, Goe-bel R (2009) Optimizing functional accuracy of TMS in cognitive studies: a comparison of methods. J Cogn Neurosci 21:207–221

Sandrini M, Rossini PM, Miniussi C (2008) Lateralized contribution of prefrontal cortex in controlling task-irrelevant information dur-ing verbal and spatial working memory tasks: rTMS evidence. Neuropsychologia 46:2056–2063

Schutter DJ, van Honk J, d’Alfonso AA, Postma A, de Haan EH (2001) Effects of slow rTMS at the right dorsolateral prefrontal cortex on EEG asymmetry and mood. NeuroReport 12:445–447

Sparing R, Buelte D, Meister IG, Paus T, Fink GR (2008) Transcranial magnetic stimulation and the challenge of coil placement: a com-parison of conventional and stereotaxic neuronavigational strate-gies. Hum Brain Mapp 29:82–96

Spielberger CD (1988) Manual for the state-trait anger expression inventory. Consulting Psychologists Press, Palo Alto

Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA (1983) Manual for the state-trait anxiety inventory. Psychological Assess-ment Resources, Odessa

Steyer R, Schwenkmezger P, Notz P, Eid M (1997) Der Mehrdimen-sionale Befindlichkeitsfragebogen (MDBF). Handanweisung. Hogrefe, Göttingen

Thut G, Pascual-Leone A (2010) A review of combined TMS-EEG studies to characterize lasting effects of repetitive TMS and assess their usefulness in cognitive and clinical neuroscience. Brain Topogr 22:219–232

van Honk J, Schutter DJ (2006) From affective valence to motivational direction: the frontal asymmetry of emotion revised. Psychol Sci 17:963–965

van Honk J, Schutter DJ, d’Alfonso AA, Kessels RP, de Haan EH (2002) 1 Hz rTMS over the right prefrontal cortex reduces vigilant attention to unmasked but not to masked fearful faces. Biol Psy-chiatry 52:312–317

Vanderhasselt MA, De Raedt R, Baeken C, Leyman L, D’Haenen H (2006) The influence of rTMS over the left dorsolateral prefrontal cortex on stroop task performance. Exp Brain Res 169:279–282

Vanderhasselt MA, De Raedt R, Baeken C, Leyman L, Clerinx P, D’Haenen H (2007) The influence of rTMS over the right dorso-lateral prefrontal cortex on top-down attentional processes. Brain Res 1137:111–116

Vanderhasselt MA, De Raedt R, Leyman L, Baeken C (2010) Role of the left DLPFC in endogenous task preparation: experimental repetitive transcranial magnetic stimulation study. Neuropsycho-biology 61:162–168

Võ ML, Conrad M, Kuchinke L, Urton K, Hofmann MJ, Jacobs AM (2009) The Berlin Affective Word List Reloaded (BAWL-R). Behav Res Methods 41:534–538

Walsh V, Pascual-Leone A (2003) Transcranial magnetic stimulation: a neurochronometrics of mind. MIT Press, Cambridge

Wassermann EM (1998) Risk and safety of repetitive transcranial magnetic stimulation: report and suggested guidelines from the International Workshop on the Safety of Repetitive Transcranial Magnetic Stimulation, June 5–7, 1996. Electroencephalogr Clin Neurophysiol 108:1–16

Wittchen HU, Zaudig M, Fydrich T (1997) Strukturiertes Klinisches Interview für DSM-IV. Achse I und II. Handanweisung. Hogrefe, Göttingen