The influence of approach–avoidance motivational orientationon conflict adaptation

Maikel Hengstler & Rob W. Holland &

Henk van Steenbergen & Ad van Knippenberg

# Psychonomic Society, Inc. 2014

Abstract To deal effectively with a continuously changingenvironment, our cognitive system adaptively regulates re-source allocation. Earlier findings showed that an avoidanceorientation (induced by arm extension), relative to an ap-proach orientation (induced by arm flexion), enhancedsustained cognitive control. In avoidance conditions,performance on a cognitive control task was enhanced,as indicated by a reduced congruency effect, relative toapproach conditions. Extending these findings, in thepresent behavioral studies we investigated dynamic ad-aptations in cognitive control—that is, conflict adapta-tion. We proposed that an avoidance state recruits moreresources in response to conflicting signals, and therebyincreases conflict adaptation. Conversely, in an approachstate, conflict processing diminishes, which consequent-ly weakens conflict adaptation. As predicted, approachversus avoidance arm movements affected both behav-ioral congruency effects and conflict adaptation: Ascompared to approach, avoidance movements elicitedreduced congruency effects and increased conflict adap-tation. These results are discussed in line with a possi-ble underlying neuropsychological model.

Keywords Approach–avoidancemotivational orientation .

Conflict monitoring . Cognitive control . Conflict adaptation .

Resource allocation

An important mechanism of our cognitive system is theallocation of resources to effectively deal with difficult,problematic, or novel situations, by up-regulating cog-nitive control. In a changing environment, being able toregulate cognitive control in a flexible manner helps usto respond adaptively to task demands. In addition tosustained, or tonic, cognitive control, dynamic adjust-ments of cognitive control consist of trial-by-trial adap-tations of cognitive control as a function of previoustrial difficulty due to conflict. This is called conflictadaptation and is thought to reflect temporary, orphasic, enhancements or reductions of cognitive control(Botvinick, Braver, Barch, Carter, & Cohen, 2001).

Paradigms such as the flanker task (Eriksen &Eriksen, 1974) were designed to examine cognitive con-trol processes. In the flanker task, people have to indi-cate the direction of a central arrow, while adjacentarrows are in the same (congruent) or another(incongruent) direction. The strength of the congruencyeffect—entailing that congruent trials are responded tofaster than incongruent trials—varies as a function ofwhether or not the previous trial contained a conflict:The congruency effect on the current trial is less strongafter an incongruent trial than after a congruent trial(Gratton, Coles, & Donchin, 1992). This effect of theprevious trial on current trial performance is an empir-ical indicator of conflict adaptation. Ample studies havesince shown that conflict adaptation is a robust phenom-enon that occurs across several tasks (Egner, 2008;Gratton et al., 1992; Stürmer, Leuthold, Soetens, Schröter, &Sommer, 2002).

M. Hengstler (*) : R. W. Holland :A. van KnippenbergBehavioural Science Institute, Department of Social and CulturalPsychology, Radboud University Nijmegen, Nijmegen, TheNetherlandse-mail: m.hengstler@psych.ru.nl

H. van SteenbergenLeiden Institute for Brain and Cognition, Department of CognitivePsychology, Leiden University, Leiden, The Netherlands

M. HengstlerBehavioural Science Institute, Department of Social and CulturalPsychology, Radboud University Nijmegen, PO Box 9104, 6500HE Nijmegen, The Netherlands

Cogn Affect Behav NeurosciDOI 10.3758/s13415-014-0295-6

Modulators of cognitive control

Recent research has investigated how affectively valencedstimuli or mood states may modulate conflict adaptation (forreviews, see Chiew & Braver, 2011; Dreisbach & Fischer,2012). For example, positive reward signals have been shownto modulate conflict adaptation (van Steenbergen, Band, &Hommel, 2009), although the direction of this effect seems todepend on performance-contingency and the type of stimulithat is used to signal reward (e.g., Braem et al., 2013; Braem,Verguts, Roggeman, & Notebaert, 2012; Stürmer, Nigbur,Schacht, & Sommer, 2011). Furthermore, disentangling ma-nipulations of affect (positive vs. negative) and arousal (highvs. low), several studies have shown that, irrespective ofarousal, negative states result in enhanced conflict adaptation,as compared to positive states (Kuhbandner & Zehetleitner,2011; van Steenbergen, Band, & Hommel, 2010; cf. vanSteenbergen, Band, Hommel, Rombouts, & Nieuwenhuis,2014; van Steenbergen, Booij, Band, Hommel, & van derDoes, 2012). In some studies, arousal did not have any effect(van Steenbergen et al., 2010), whereas others have shownthat high-arousing pictures, such as mutilated bodies, resultedin slower reaction times and reduced conflict adaptation(Padmala, Bauer, & Pessoa, 2011; cf. Fischer, Dreisbach, &Goschke, 2008). Kuhbandner and Zehetleitner (2011) havedemonstrated independent modulation of affect and arousal ondynamic and sustained cognitive control, respectively, as wasalso confirmed by computer simulations showing independentmodulation of related parameters in Botvinick et al.’s (2001)conflict-monitoring model.

To summarize, research has shown differential effects ofrewards, mood states, and arousal. To date, no studies havebeen conducted on the influence of approach and avoidancemotivational states on conflict adaptation. As we will arguebelow, for theoretical reasons, approach versus avoidancemotivational states constitute an important modulator of bothconflict adaptation and the congruency effect.

Approach–avoidance orientation

Approach and avoidance may be considered the two mostfundamental motivational states (Elliot, 2008). Approach mo-tivation helps to attain essential outcomes in the world includ-ing food, drinks, and partners. On the other hand, avoidancemotivation prevents us from danger and negative outcomes.Their impact is documented across multiple domains. Forinstance, several studies illustrated the interplay between ap-proach and avoidance movements and stimulus evaluation.Pull responses (approach) have been shown to be faster to-ward positive than toward negative stimuli, whereas pushreactions (avoidance) have been shown to be faster for nega-tive than for positive stimuli (Chen & Bargh, 1999). The

reverse causality has also been shown. Stimuli viewed duringan approach arm movement were evaluated more positivelythan stimuli viewed during an avoidance arm movement(Cacioppo, Priester, & Berntson, 1993; cf. Centerbar &Clore, 2006). According to Cacioppo and his colleagues,engaging in an approach arm movement provides body feed-back associated with approaching positive stimuli, whereasengaging in an avoidance arm movement provides body feed-back associated with avoiding negative stimuli (see alsoFörster & Strack, 1997, 1998). This theoretical assumptionis based on the notion that over the course of a lifetime, armflexion is repeatedly associated with acquiring or consumingdesired objects (i.e., approach motivation), whereas arm ex-tension is repeatedly associated with rejecting undesired ob-jects (i.e., avoidance motivation).

However, approach and avoidance effects go beyond stim-ulus evaluations (Barsalou, 2008). For example, approach andavoidance have been argued to influence a correspondingmotivational mindset or regulatory focus. According toHiggins (1999), individuals with a promotion focus are gen-erally motivated by the presence and absence of positiveoutcomes, whereas the behavior of individuals in a preventionfocus is motivated by the presence or absence of negativeoutcomes. These general foci have an influence on the regu-lation of people’s behavior. Friedman and Förster (2000,2005a) argued that approach and avoidance arm movements,triggering their corresponding foci, enhance creativity andanalytic reasoning, respectively. An approach mindset seemsto foster a heuristic processing style, enabling individuals tobe more creative—to “think outside the box”—whereas anavoidance mindset fosters a systematic processing style en-abling individuals to be more analytic (Friedman & Förster,2000, 2005a).

Several other studies have suggested the role of arm move-ments in regulatory focus. Research has suggested that theintensity of pressure exerted through approach or avoidancearm movements functions as an indicator of motivationalstrength (Förster, Higgins, & Idson, 1998). Similarly, approachand avoidance arm movements directly influence motivation:Förster (2003) showed that participants engaging in an ap-proach arm movement consumed more food than did individ-uals engaging in an avoidance arm movement. Taken together,all of these findings combined suggest that the arm movementsactivate their corresponding motivational system: Arm flexionactivates a motivational system concerned with the processingof rewards, whereas arm extension activates a motivationalsystem concerned with the processing of possible threats.

Approach–avoidance and sustained cognitive control

Avoidance bodily actions, as opposed to approach bodilyactions, are habitually performed in situations that call for

Cogn Affect Behav Neurosci

vigilance and controlled action. Therefore, the mere executionof an avoidance motor action may function as a subtle alert fordifficult conditions (see Kahneman, 1973), which leads to anoverall greater recruitment of resources that are usually re-quired in this specific context. In contrast, performing anapproach movement serves as a signal of safety and thereforeindicates the absence of threat (Friedman & Förster, 2000).Mobilization of cognitive resources is then less needed, andcognitive control may be down-regulated on a general, toniclevel.

Indeed, previous research has already shown the differen-tial impacts of approach and avoidance motivational states inregulating resource allocation. In a series of studies, it wasshown that avoidance orientation (induced by arm extension),relative to approach orientation (induced by arm flexion),enhanced sustained cognitive control, as measured by reducedcongruency effects on a Stroop task and a task-switchingparadigm (Koch, Holland, & van Knippenberg, 2008; seealso Koch, Holland, Hengstler, & van Knippenberg, 2009;for alternative findings, see Friedman & Förster, 2005b). Inanother study, Koch and colleagues showed further evidencefor a resources allocation account (Koch, 2008). On an initialcognitive control task, individuals in the avoidance conditionperformed better, but subsequently showed greater indicationsof resource depletion. Importantly, the effect of approach–avoidance state on the initial task was mediated by thedegree of depletion, suggesting that superior performance inthe avoidance condition was achieved by greater expenditureof cognitive resources. In the present research, we aimed toextend the findings of Koch and colleagues (2009; Koch et al.,2008) by showing that approach and avoidance motivationalstates may also modulate dynamic adjustments in cognitivecontrol—that is, conflict adaptation.

Approach–avoidance and dynamic cognitive control

Besides a tonic process (cf. Kuhbandner & Zehetleitner,2011), approach and avoidance states may alter conflict-driven resource allocation. Before explaining how this mightwork, it is important to note that conflict is associated withnegative affect, signaling that conflict is an aversive event(Dreisbach & Fischer, 2012; Fritz & Dreisbach, 2013;Schouppe, de Houwer, Ridderinkhof, & Notebaert, 2012).Previous work has shown that individuals who are in anavoidant mindset are biased to conflicting (aversive) cues,whereas individuals who are in an approach mindset are lessconcerned with conflict and focus more on positive outcomes(Derryberry & Reed, 1994; Gomez & Gomez, 2002; Higgins& Tykocinski, 1992; Scholer, Stroessner, & Higgins, 2008;Shah, Higgins, & Friedman, 1998; Strachman & Gable,2006). For example, Derryberry and Reed found that peoplewith strong approach motivation were biased toward positive

cues in a visual target detection task, whereas people withstrong avoidance motivation were biased toward negativecues. Similarly, Gomez and Gomez showed that approachmotivation (BAS) predicted the processing of positive (butnot negative) emotional information, and avoidance motiva-tion (BIS) predicted the processing of negative (but not pos-itive) emotional information. Furthermore, research suggeststhat in an approach state, individuals are not only biasedtoward positive cues, but also retrieve such cues better; thatis, positive information is processed more thoroughly. In anavoidant state, negative information is processed more thor-oughly (Förster & Strack, 1997, 1998).

If being in an avoidant state biases individuals towardconflicting cues, and this information is processed more thor-oughly, this might result in allocating more resources in re-sponse to conflicting signals—that is, a phasic, conflict-drivenadaptation of cognitive control. Conversely, if in an approachstate conflict processing diminishes, this should consequentlyweaken conflict adaptation.

Present research

Following the reasoning described above, we expectedto find (1) enhanced conflict adaptation in the avoidancecondition and reduced conflict adaptation in the ap-proach condition, and (2) a replication of earlier find-ings (Koch et al., 2009; Koch et al., 2008) showing thatavoidance leads to a smaller congruency effect thandoes approach. These ideas were tested in three studies.In the first study, we compared an approach state withan avoidant state. In the second study, we added acontrol condition to examine the separate contributionsof approach and avoidance motivation to the hypothe-sized effects. Finally, in the third study, we employed aflanker task that allowed us to control for feature repe-titions and contingency confounds (Hommel, Proctor, &Vu, 2004; Mayr, 2004; Nieuwenhuis et al., 2006;Schmidt, 2013; Schmidt & de Houwer, 2011) that couldhave affected the indices of conflict adaptation in theformer studies. In addition, we included some additionalquestions, to measure effort of arm movement andmood. Moreover, we included a scale to measurepromotion–prevention orientation, intended as a manip-ulation check of the arm movement manipulation.Throughout the three studies, we employed a manipula-tion of approach and avoidance states by instructingparticipants to make an approach or avoidance motoraction (Cacioppo et al., 1993; Koch et al., 2008). Morespecifically, participants were asked to engage in eitherarm flexion (approach) or arm extension (avoidance)while performing a flanker task.

Cogn Affect Behav Neurosci

Study 1

Methods

Participants A total of 71 students (61 females, ten males)from Radboud University Nijmegen participated in this study.The mean age of the participants was 19.8 years (range 18 to27 years). In return for credits, participants were assigned toeither the approach condition or the avoidance condition.

Procedure Instructions were given on a computer screen.While participants performed an approach or avoidance motoraction with their nondominant hand, they responded to aflanker task using their dominant hand (Eriksen & Eriksen,1974). Using their nondominant hand, participants in theapproach condition pressed a foam ball against the bottomof the table to elicit and maintain isometric flexor contractionof the arms, whereas in the avoidance condition they pressedthe foam ball against the top of the table to elicit and maintainisometric extensor contraction (cf. Cacioppo et al., 1993;Koch et al., 2008). In Fig. 1, photographs illustrate the armpositions used to manipulate approach and avoidance. Theflanker stimuli consisted of a row of five white arrowspointing to the left or the right against a black background.Participants were instructed to make quick and accurate re-sponses on the keyboard using their dominant hand to indicatethe direction of the central arrow, while the adjacent arrowswere in the same (congruent) or in the opposite (incongruent)direction. Participants performed 12 practice trials. In order toreduce between-subjects error noise due to a randomized trialorder, instead of using a fully randomized order for eachparticipant, we created two pseudorandomized orders of 66test trials that were divided into two blocks of 33 trials (thefirst trial of each block could not be used, because it had noprevious trial).1 Within these two orders, all four trial types(cC, iC, cI, and iI) were randomly distributed over the 66trials, with the restriction that they had to be administeredequally often, and the same trial type should not be presentedmore than three times in a row. Each trial started with afixation cross displayed for 1,000 ms, followed by the flankerstimulus that was displayed until a response was given. In thepractice phase, upon an error, feedback was provided byshowing a red cross for 400 ms.

Data analysis We analyzed the reaction times (RTs) of correctresponses for test trials using a mixed design, with PreviousTrial (c vs. i) and Current Trial (C vs. I) as within-subjectsfactors and Motivational State (approach vs. avoidance) as abetween-subjects factor. Error trials (3.7%) and trials with RTsmore than 3 SDs above or below the individual condition-specific mean (0.9%) were excluded from the analysis. We

1 We refrained from using a higher number of trials because (1) we werenot sure about the duration of the effect of arm movement—since the armmovement had to be performed during the task, one can imagine that theeffect of the arm manipulation might wear off after some time; and (2)some studies have suggested that conflict adaptation effects disappearquickly over time (Mayr & Awh, 2008; van Steenbergen, Haasnoot,Bocanegra, Berretty, & Hommel, 2014 Manuscript submitted forpublication), so using a low number of trials per condition might helpto measure the effect that we aimed to investigate here. A previous studyshowing effects of dysphoric mood on conflict adaptation had also usedonly 64 trials in total (van Steenbergen, Booij, et al., 2012).

Fig. 1 (a) Photograph of the approach induction in Studies 1, 2, and 3.(b) Photograph of the avoidance induction in Studies 1 and 2. (c)Photograph of the avoidance induction in Study 3

Cogn Affect Behav Neurosci

also removed one participant who had a much higher generalRT than the average (6 SDs above the sample mean), and oneparticipant who made many more errors on inconsistent trialsthan the average (6 SDs above the sample mean).2

Results

The results showed a congruency effect: The main effect ofcurrent trial was significant, indicating that performance onthe flanker task was faster on congruent (383 ms) than onincongruent (429 ms) trials, F(1, 69)=271.08, p<.001,η2=.80. The interaction between previous trial and currenttrial was also significant, indicating a conflict adaptationeffect, F(1, 69)=4.33, p=.04, η2=.06. The conflict adaptationeffect entails that the current RT difference (I – C) was smallerwhen the previous trial was incongruent (i) than when it wascongruent (c). Most importantly, both the congruency andconflict adaptation effects interacted with motivational state,as we will describe below. See Table 1 for the full range ofcondition means.

We found a Previous Trial × Current Trial × MotivationalState interaction, F(1, 69)=6.04, p=.02, η2=.08 (see Fig. 2a).To understand the nature of this interaction, we analyzed theinteraction effects of previous trial and current trial separatelyfor approach and avoidant states. This revealed a strong Pre-vious Trial × Current Trial interaction in the avoidance con-dition, F(1, 34)=11.89, p<.01, η2=.26, whereas this effectwas absent for the approach condition, F(1, 35)<1.

We also found a Current Trial × Motivational State inter-action, F(1, 69)=6.58, p=.01, η2=.09 (see Fig. 2b). Althoughthe effect of current trial—that is, the congruency effect—wassignificant in both the approach and avoidance conditions (ps<.001), the interaction meant that the congruency effect wassignificantly smaller in the avoidance condition (39 ms) thanin the approach condition (53 ms).

Notably, we observed no significant correlation betweenthe congruency effect and the conflict adaptation effect,r(71)=–.07, p=.56. We also tested the unique effect of moti-vational state on the congruency effect with the conflict adap-tation effect partialed out, as well as the effect of motivationalstate on conflict adaptation with the congruency effectpartialed out. To do this, we first computed a congruencyscore (I – C) and a conflict adaptation score [(iI – iC) – (cI –cC)]. Subsequently, we conducted an analysis of covariance(ANCOVA), with congruency score as the dependent variable(DV), Motivational State as the between-subjects factor, andconflict adaptation score as a covariate. The results indicatedthat the effect of motivational state on the congruency effect

remained significant in this analysis, F(1, 71)=6.13, p=.02,η2=.08. Similarly, when we switched the roles of congruencyeffect (from DV to covariate) and conflict adaptation effect(from covariate to DV), again the effect of motivational stateremained significant, F(1, 71)=5.60, p=.02, η2=.08.

Study 2

In Study 1, we compared an approach state with an avoidantstate. The results revealed strong conflict adaptation in theavoidance relative to the approach condition. In addition, wereplicated previous research by finding a smaller congruencyeffect for avoidance than for approach (Koch et al., 2008). InStudy 2, we added a control condition, to examine the separate

2 The interaction between previous trial and current trial became nonsig-nificant (p=.204) when these two outliers remained included in the dataset. Also, the three-way interaction between previous trial, current trial,and motivational state became nonsignificant (p=.108).

Table 1 Mean reaction times (RTs, inmilliseconds) and SEs for each trialtype in the three reported studies

Study Condition Congruency(N – 1)

Congruency(N)

MeanRT

SE

1 Approach Congruent Congruent 380.74 6.74

Incongruent 432.91 8.31

Incongruent Congruent 382.58 7.18

Incongruent 436.37 7.80

Avoidance Congruent Congruent 383.69 6.83

Incongruent 432.10 8.43

Incongruent Congruent 386.43 7.28

Incongruent 415.41 7.91

2 Approach Congruent Congruent 385.77 7.93

Incongruent 441.96 9.76

Incongruent Congruent 391.33 7.59

Incongruent 447.33 9.72

Control Congruent Congruent 395.99 8.07

Incongruent 445.51 9.94

Incongruent Congruent 402.63 7.73

Incongruent 441.30 9.89

Avoidance Congruent Congruent 394.86 7.93

Incongruent 441.13 9.76

Incongruent Congruent 398.46 7.59

Incongruent 420.39 9.72

3 Approach Congruent Congruent 478.19 10.90

Incongruent 519.41 9.66

Incongruent Congruent 480.82 10.61

Incongruent 531.82 9.77

Avoidance Congruent Congruent 493.36 10.67

Incongruent 535.89 9.45

Incongruent Congruent 498.95 10.38

Incongruent 516.37 9.56

Cogn Affect Behav Neurosci

contributions of approach and avoidance motivation to theseeffects.

Methods

Participants A group of 86 students (76 females, ten males)from Radboud University Nijmegen participated in this study.The mean age of the participants was 21.7 years (range 18 to35 years). We added a control condition to examine thedifferential effects of approach and avoidance motor actions.In return for credits, participants were assigned to the ap-proach condition, the control condition, or the avoidancecondition.

Procedure The procedure of Study 2 was exactly the same asin Study 1, with the addition of a control condition. In thecontrol condition, people had to rest their nondominant handon their lap (i.e., they did not engage in arm flexion orextension), whereas in the other conditions they had to per-form the respective arm movements.

Data analysis We used the same kind of analyses as in Study1. Error trials (3.4%) and trials with RTs more than 3 SDsabove or below the individual condition-specific mean (0.6%)were excluded from the analysis. The data revealed nobetween-subjects outliers.

Results

The results showed the congruency effect: Performance on theflanker task was faster on congruent (395 ms) than on incon-gruent (439 ms) trials, F(1, 83)=194.08, p<.001, η2=.70. Theinteraction between previous trial and current trial was againsignificant, indicating a conflict adaptation effect, F(1, 83)=13.26, p<.001, η2=.14. See Table 1 for the full range ofcondition means.

Most importantly, the results of Study 2 also showed aPrevious Trial × Current Trial × Motivational State interac-tion, F(2, 83)=4.72, p=.01, η2=.10 (see Fig. 3a). We found aconflict adaptation effect for the avoidance condition, F(1,28)=14.38, p=.001, η2=.34, as well as for the control condi-tion, F(1, 27)=10.84, p<.01, η2=.29. No evidence of such aneffect was observed for the approach condition, F(1, 28)<1.LSD pairwise comparisons on the conflict adaptation scoresshowed that avoidance (24 ms) differed from approach (0 ms),p<.01, 95% CI [8.48, 39.84], and marginally significantlyfrom control (11 ms), p=.09, 95% CI [–2.32, 29.32]. Controldid not differ from approach, p=.18.

A

B

40

30

20

10

0

-10

Con

flic

t ada

patio

n (m

s)

70

60

50

40

30

20

Con

flic

t ada

patio

n (m

s)

approach avoidance

approach avoidance

10

0

Fig. 2 (a) Mean reaction times (in milliseconds) representing the conflictadaptation scores of Study 1, (cI – cC) – (iI – iC). Error bars representSEs. (b) Mean reaction times representing the congruency scores of Study1, I – C. Error bars represent SEs

A

B

40

30

20

10

0

-10

Con

flic

t ada

patio

n (m

s)

70

60

50

40

30

20

Con

flic

t ada

patio

n (m

s)

approach control

control

avoidance

approach avoidance

10

0

Fig. 3 (a) Mean reaction times (in milliseconds) representing the conflictadaptation scores of Study 2, (cI – cC) – (iI – iC). Error bars representSEs. (b) Mean reaction times representing the congruency scores of Study2, I – C. Error bars represent SEs

Cogn Affect Behav Neurosci

We also found a Current Trial × Motivational State inter-action,F(2, 83)=3.96, p=.02, η2=.09 (see Fig. 3b). All simpleeffects were significant for each level of motivational state (ps<.001). LSD pairwise comparisons on the congruency scoresshowed that avoidance (34 ms) differed from approach (56ms), p<.01, 95% CI [–37.56, –6.44]. Control (44 ms) did notdiffer from either approach or avoidance, ps=.13 and .21,respectively.

Again, we found no significant correlation between thecongruency effect and the conflict adaptation effect,r(86)=.03, p=.82. As in Study 1, we tested the unique effectof motivational state on the congruency effect with the conflictadaptation effect partialed out, as well as the effect of motiva-tional state on conflict adaptation with the congruency effectpartialed out. The results of the ANCOVAs, indicated that theeffect of motivational state on the congruency effect remainedsignificant, F(2, 86)=4.67, p=.01, η2=.10, as well as theeffect of motivational state on the conflict adaptation effect,F(2, 86)=5.43, p=.01, η2=.12.

Reanalysis of Studies 1 and 2

In Studies 1 and 2, we consistently found an effect of moti-vational state on both sustained and dynamic cognitive con-trol. Furthermore, motivational state seemed to independentlyinfluence these two forms of control regulations.

An important limitation of Studies 1 and 2, and of manyother studies in the domain of conflict adaptation, was thepotential confound of repetition priming. Some studies haveshown that conflict adaptation effects can be accounted forentirely by repetition-priming effects (e.g., Mayr, 2004;Nieuwenhuis et al., 2006). Repetitions occur on 50% of thecC and iI trials, but on none of the iC or cI trials. The typicaldecreased RTs on cC and iI trials evident in sequential effectsmight thus (partly) be driven by repetition priming rather thanby conflict adaptation.

To address the issue of repetition effects, at least to someextent, we ran an analysis that excluded all complete-repetition trials—that is, all trials in which the stimulus at trialN+1 was exactly the same as the stimulus at trial N (see Mayr,2004). Because removing all complete-repetition trials re-duced the power of the tests, we created a new data fileencompassing the data from Studies 1 and 2 (without thecontrol condition). We subjected these data to a repeatedmeasures analysis of variance (ANOVA) with Previous Trial(c vs. i) and Current Trial (C vs. I) as within-subjects factors,and Motivational State (approach vs. avoidance) and Study (1vs. 2) as between-subjects factors. Crucially, as is shown inTable 2, the influence of motivational state on both the con-gruency effect and the conflict adaptation effect remainedsignificant after excluding all complete stimulus repetitions.

Although this is encouraging, using a design with fourstimuli and two response options did not allow us to exclude

different alternative explanations in terms of feature integra-tion, which predicts that complete alternations (all remainingcC and iI trials) would be faster than partial repetitions (all cIand iC trials), due to feature binding effects (Hommel et al.,2004). Therefore, in Study 3, we used an optimized design (cf.Duthoo, Abrahamse, Braem, Boehler, & Notebaert, 2014Manuscript under review) in which we only presented com-plete alternations for all sequential trial types (cf. Ullsperger,Bylsma, & Botvinick, 2005), while at the same timecorrecting for the more recently discussed confounds in termsof contingency biases (Schmidt, 2013; Schmidt & de Houwer,2011). We also included some additional questions, to mea-sure effort of arm movement and mood. Moreover, we includ-ed a scale to measure promotion–prevention orientation,which was intended as a manipulation check of the armmovements that we manipulated.

Study 3

Method

Participants A total of 96 students (73 females, 23 males)from Radboud University Nijmegen participated in thisstudy. The mean age of participants was 20.9 years (range18 to 29 years). In return for credits, participants wereassigned to either the approach condition or the avoidancecondition.

Procedure Instructions were given on a computer screen.While participants performed an approach or avoidance motoraction with their nondominant hand, they responded to aflanker task using their dominant hand (Eriksen & Eriksen,1974). Similar to Studies 1 and 2, using their nondominanthand, participants in the approach condition pressed a foamball against the bottom of the table to elicit and maintainisometric arm flexor contraction (Fig. 1a). We slightlychanged the position concerning the avoidance condition. In

Table 2 Results of the meta-analysis over the first two studies, examin-ing the influence of motivational orientation on both the congruencyeffect and the conflict adaptation effect, excluding complete-repetitiontrials

Source F p η2

Congruency (N) × Orientation 6.38 .01 .05

Congruency (N) × Congruency (N – 1) × Orientation 4.55 .04 .04

Congruency (N) × Study 0.20 .66 .00

Congruency (N) × Congruency (N – 1) × Study 5.19 .02 .04

Congruency (N) × Study × Orientation 1.09 .30 .01

Congruency (N) × Congruency (N – 1) × Study ×Orientation

1.93 .17 .02

Cogn Affect Behav Neurosci

this condition, also using their nondominant hand, participantspressed a foam ball against the frontal side of the table to elicitand maintain isometric arm extensor contraction (Fig. 1c). Inour view, the latter position even more strongly convergeswith avoidance, since it more clearly reflects a “pushingsomething away” gesture. The flanker stimuli consisted of arow of five black arrows against a white background. Partic-ipants were instructed to make quick and accurate responseson the keyboard using their dominant hand to indicate thedirection of the central arrow, whereas the adjacent arrowswere in the same (congruent) or in another (incongruent)direction. The responses keys were [i], [j], [k], and [m]. Thearrows could point in one of four directions: up, down, left, orright. In order to control for potential contingency bias effects(Schmidt, 2013; Schmidt & de Houwer, 2011), we includedonly four out of the 12 possible combinations of incongruenttrials, so that four trial types were used (equally often) for boththe congruent and incongruent trials (cf. Manuscript underreview).We created two different orders of 101 trials, in whichall four trial types (cC, cI, iC, and iI) were randomizedwith therestriction that they had to be administered equallyoften, and the same trial type was presented not morethan three times in a row. Finally, stimulus and featurerepetitions were controlled by only allowing completefeature alternations. For example, a congruent trial withall of the arrows facing upward could only be followedby a next stimulus that did not include any upward-facing arrows—for example, a stimulus with flankersfacing left and the central target facing down. The trialafter that could only display a stimulus that did nothave any arrows facing left and/or down, and so on.

Participants performed a block of 16 practice trials.The practice block was repeated until the participants’performance accuracy reached 85% correct answers, oruntil the block was run for a maximum number of fourtimes. Subsequently, participants completed 101 experi-mental trials while engaging in an approach or avoid-ance arm movement. Each trial started with a blankscreen for 1,000 ms, followed by the flanker stimulus,which was displayed until a response was given. In thepractice phase, upon an error, feedback was provided byshowing a red cross for 400 ms.

After the experimental phase of the flanker task, partici-pants were asked once more to engage in the condition-specific arm movement, this time while answering a question-naire about proverbs (RFQ-short; van Stekelenburg &Klandermans, 2003). The questionnaire contained 14 items:seven promotion-focused items (e.g., “Nothing ventured,nothing gained”) and seven prevention-focused items (e.g.,“Better safe than sorry”). Participants were asked to answerthe following question for each of the items on a 7-point Likertscale: “To what extent does this saying apply to you as aperson. Don’t think too long about your answer, what comes

to mind first often is the best answer.” In a pilot study, wefound that this scale reliably differentiated betweenmotivationelicited by approach and avoidance arm movements.3

Finally, participants were asked to indicate, on a scale from1 to 100, the amount of effort that they experienced whileengaging in the condition-specific arm movement (1 = noeffort, 100 = very much effort), and their current mood state(1 = negative mood state, 100 = positive mood state).

Data analysis We used the same kind of analyses as in Stud-ies 1 and 2. Error trials (1.6%) and trials with RTs more than 3SDs above or below the individual condition-specific mean(1.8%)were excluded from the analysis.We also removed twoparticipants who had a higher general RT than the average (>3SDs above the sample mean), and another two participantswho made more errors on inconsistent trials than the average(>3 SDs above the sample mean).4

Results

Manipulation check On the basis of the results of a pilot study,factor analysis was used to identify and compute a score forthe factor underlying the short version of the RFQ (see note 3).First, we observed that all items correlated at least .3 with atleast one other item, suggesting reasonable factorability. Sec-ond, the Kaiser–Meyer–Olkin measure of sampling adequacywas .73, which was above the commonly recommended valueof .6, and Bartlett’s test of sphericity was significant [χ2(91)=291.87, p<.001]. The diagonals of the anti-image correlationmatrix were also all over .5. Finally, the communalities wereall above .3, further confirming that each item shared somecommon variance with the other items. Given these overallindicators, factor analysis was deemed to be suitable with all14 items.

3 The RFQ-short was pilot tested in a different session with differentparticipants (N=55). The α for the promotion scale was .57, and that forthe prevention scale was .59. Both αs are lower than the generallyaccepted minimal reliability of .70. Therefore, instead of using these apriori constructs, all items of the RFQ scale were subjected to an explor-atory principal component analysis. The results of this analysis revealed asingle-factor solution (additional factors proved to be uninterpretable),explaining 18% of the variance. This factor was labeled Promotion VersusPrevention, because promotion-focused items loaded positively on thisfactor, whereas prevention-focused items loaded negatively on it. Subse-quently, we subjected this Promotion Versus Prevention factor score to anANOVAwithMotivational State as a between-subjects factor. The resultsrevealed a significant effect of motivational state, F(1, 55)=7.12, p=.01,η2=.12: Participants in the approach condition responded more in linewith a promotion focus (M=.33), whereas those in the avoidance condi-tion responded more in line with a prevention focus (M=–.35), thuscorroborating the intended motivational implications of the approachand avoidance arm movements. We also used this factor-analytic ap-proach for our manipulation check in Study 3.4 The pattern of significant and nonsignificant results did not changewhen these four participants were included in the analysis.

Cogn Affect Behav Neurosci

Principal component analysis with listwise deletion re-vealed a single-factor solution (additional factors proved dif-ficult to interpret), explaining 26% of the variance.Promotion-focused items loaded positively on this factor,whereas prevention-focused items loaded negatively on it.We computed the factor score for this factor and added it tothe data file, so that each participant had such a score. Toattribute meaning to this factor score, a positive score entailedresponding more in line with a promotion focus, whereas anegative score entailed responding more in line with a pre-vention focus. Subsequently, this factor score was subjected toan ANOVA, with Motivational State as a between-subjectsfactor. The results revealed a significant effect of motivationalstate, F(1, 92)=4.88, p=.03, η2=.05: Participants in an ap-proach state responded more in line with a promotion focus(M= .23), whereas participants in an avoidance stateresponded more in line with a prevention focus (M=–.22).

Flanker task The results showed a congruency effect: Themain effect of current trial was significant, indicating thatperformance on the flanker task was faster on congruent(488 ms) than on incongruent (526 ms) trials, F(1, 90)=161.81, p<.001, η2=.64. The interaction between previoustrial and current trial was not significant, p>.13. However,both effects did interact with motivational state, as we willdescribe below. See Table 1 for the full range of conditionmeans.

We found a Previous Trial × Current Trial × MotivationalState interaction, F(1, 90)=11.63, p<.002, η2=.11 (seeFig. 4a). To illustrate the nature of this interaction, we ana-lyzed the interaction effects of previous trial and current trialseparately for the experimental conditions. We found a strongPrevious Trial×Current Trial interaction in the avoidancecondition, F(1, 46)=9.47, p<.01, η2=.17, whereas this effectwas absent for the approach condition, F(1, 44)=2.60, p>.11,η2=.06.

We also found a Current Trial × Motivational State inter-action, F(1, 90)=7.28, p=.008, η2=.08 (see Fig. 4b). Al-though the effect of current trial—that is, the congruencyeffect—was significant in both the approach and avoidanceconditions (ps<.001), the congruency effect was significantlysmaller in the avoidance condition (30 ms) than in the ap-proach condition (46 ms).

In contrast to Studies 1 and 2, the results now showed asignificant negative correlation between the congruency effectand the conflict adaptation effect, r(92)=–.22, p=.03. Weagain tested the unique effect of motivational state on thecongruency effect with the conflict adaptation effect partialedout, as well as the effect of motivational state on conflictadaptation with the congruency effect partialed out. The re-sults of the ANCOVAs indicated that the effect of motivation-al state on the congruency effect remained significant, F(1,92)=4.35, p=.04, η2=.05, as well as the effect of motivational

state on the conflict adaptation effect, F(1, 92)=8.53, p=.004,η2=.09.

No significant differences in mood (Mappr=74,Mavoid=73)or experienced effort of the arm position (Mappr=58, Mavoid=51) were obtained between conditions, and adjusting for thesevariables as covariates in the reported analyses did not changethe pattern of significant and nonsignificant results reportedabove, suggesting that these variables did not mediate theeffects of motivational state on cognitive control.

Discussion

In the present research, we investigated the influence of ap-proach and avoidance body movements on dynamic andsustained control, as measured by the conflict adaptationeffect and the congruency effect, respectively. This researchwas the first to show an effect of motivational orientation onconflict adaptation. In three studies, we showed that both thecongruency effect and the conflict adaptation effect differedbetween an approach motivational state and an avoidancemotivational state. For the congruency effect, avoidance dif-fered from approach, whereas the control condition differedfrom neither approach nor avoidance. Concerning conflictadaptation, avoidance differed from approach and, marginally,from control, whereas the latter two did not differ significantly

A

B

40

20

30

10

0

-10

-20

Con

flic

t ada

patio

n (m

s)C

ongr

uenc

y (m

s)

70

60

50

40

30

20

approach avoidance

approach avoidance

10

0

Fig. 4 (a) Mean reaction times (in milliseconds) representing the conflictadaptation scores of Study 3, (cI – cC) – (iI – iC). Error bars representSEs. (b) Mean reaction times representing the congruency scores of Study3, I – C. Error bars represent SEs

Cogn Affect Behav Neurosci

(even though the conflict adaptation effect was not significantin the approach condition, whereas it was reliable in thecontrol condition). These effects were replicated in Study 3,where we controlled for feature binding effects, contingencybiases, and possible alternative moderators such as effort andmood. Although more research will be necessary to differen-tiate between the different possible processes, we think thatour reasoning, entailing that motivational orientation influ-ences the attention for conflict signals, and thereby influencesthe regulation of control, is the most parsimonious explanationof the present findings.

Our work seems at least partially in line with previous workon the influence of mood or affective states on conflict adap-tation (e.g., van Steenbergen et al., 2009, 2010). The theoret-ical framework for this link between affective states andmodulation of cognitive control is based on the aversivenature of demand (see Botvinick, 2007; Dreisbach &Fischer, 2012; Fritz & Dreisbach, 2013). That is, positiveaffect might undo or reduce the aversive state triggered by ademand, whereas negative mood states might intensify it (seeFredrickson, Mancuso, Branigan, & Tugade, 2000). However,our findings cannot be explained by differences in moodstates, given that the previous research using the manipulationof approach and avoidance motivational orientations (e.g.,Cretenet & Dru, 2004; Förster & Strack, 1998; Friedman &Förster, 2000; Koch, 2008; Nussinson, Häfner, Seibt, Strack,& Trope, 2012; van den Bergh, Schmitt, Dewitte, & Warlop,2009; van Prooijen, Karremans, & van Beest, 2006) has neverconsistently found any relation between the correspondingarm movements and mood states. In Study 3 of the presentarticle, we also directly tested the influence of mood andfound neither a significant effect of mood nor a differentbehavioral pattern when adding mood scores as a covariate.

Still, approach and avoidance orientations could change theimplicit evaluative processing of conflict. Previous work hasshown that approach and avoidance arm movements changethe evaluation of neutral objects that are evaluated at the timeof the movements (Cacioppo et al., 1993). Approach andavoidance motor actions might reduce or increase, respective-ly, the negativity of the demand, and consequently, modulatethe conflict-driven phasic effort mobilization (cf. Braem et al.,2012; van Steenbergen et al., 2009). However, the latterprocess is unlikely, because approach and avoidance armmovements seem to have different impacts on neutral andambiguous objects than on valenced objects. In fact, studieshave shown that the performance of an avoidance arm move-ment, in conjunction with evaluating a mildly negative object,causes that object to become more positive rather than nega-tive (Centerbar & Clore, 2006). The attitudinal impact ofapproach–avoidance action thus reflects its situated meaning,which depends on the valence of the stimuli being approachedor avoided. Given that conflict is often negatively evaluated,this may trigger a more positive affective state, which is

unlikely to enhance conflict adaptation effects. Therefore, thisexplanation is also not compatible with the present results.

The results with regard to the congruency effect are fully inline with previous work on the influence of approach andavoidance motivational orientation on cognitive control(Koch et al., 2009; Koch et al., 2008). Participants in theavoidance condition were faster (or made fewer errors) onincongruent than on congruent trials. Interestingly, in thepresent research we found evidence that motivational orienta-tion seemingly independently influences both sustained cog-nitive control (i.e., the congruency effect) and dynamic cog-nitive control (i.e., the conflict adaptation effect), because botheffects remained significant when controlling for the other.

It is notable that—in the avoidance as compared to theapproach condition—our findings showed reduced congruen-cy effects and increased conflict adaptation effects at the sametime. Given that previous studies using affect inductions haveshown that both effects can be modulated independently, weassume that these effects are driven by two independentsources. As we argued above, conflict adaptation effectsmay be influenced by sensitivity to the aversive nature ofconflict. On the other hand, sustained control might be drivenby the motivational requirements of the situation, which as-sociate avoidance movements with more mobilized cognitiveresources, in general. Such independent modulation is alsoconsistent with simulation work that has shown dissociablemechanisms of sustained versus dynamic control in theconflict-monitoring model (Kuhbandner & Zehetleitner,2011). However, this is not to say that congruency effectsand conflict adaptation effects are always independent, sincestudies have shown that both effects can be increased when acommon source, such as conflict strength, is manipulated(e.g., Forster, Carter, Cohen, & Cho, 2011; Wendt, Kiesel,Geringswald, Purmann, & Fischer, 2013; cf. Scherbaum, Fi-scher, Dshemuchadse, & Goschke, 2011). However, in linewith cumulating evidence (van Steenbergen et al., 2009; vanSteenbergen, Band, & Hommel, 2012; van Steenbergen,Band, et al., 2014), the increased adaptation effects in thecontext of reduced interference effects in our study show thatconflict strength in mere cognitive terms is not sufficient toaccount for all types of modulation, and suggests that theaffective evaluation of conflict is likely an important sourceof control adaptation, as well.

Although the present article presents behavioral studies, itwill be worthwhile to ponder the underlying neuropsycholog-ical mechanism involved. It is conceivable that approach andavoidance systems directly modulate the lateral prefrontalcortex (PFC), explaining the differential effects of approachand avoidance on sustained cognitive control. Avoidanceseems to be associated with right hemispheric PFC activity,whereas approach is associated with left hemispheric PFCactivity (Berkman & Lieberman, 2009; Rutherford &Lindell, 2011). Enhanced activity in the right hemispheric

Cogn Affect Behav Neurosci

PFC is linked with enhanced attention, especially insupporting responsivity to environmental events—that is, con-flicting signals (Pribram &McGuinness, 1975; cf. Heilman &Van Den Abell, 1980).

More recently, several neuroimaging studies have reportedright PFC involvement in conflict adaptation (e.g., Egner, 2011;Egner & Hirsch, 2005; Kerns et al., 2004), providing moresupport for the possibility of modulation of cognitive controlvia motivational states. The idea of conflict-driven resourceallocation in avoidance might be traced back to levels ofanterior cingulate cortex (ACC) activity. If being in an avoidantstate biases individuals toward conflicting cues (e.g., Gomez &Gomez, 2002), and this information is processed more thor-oughly (Förster & Strack, 1997, 1998), this might result inmaintaining higher levels of ACC activity directly after aconflicting event. This reasoning is based on research investi-gating the error-related negativity (ERN), hypothesized to orig-inate from the ACC. It has been shown that negative affect isassociated with amplified ERNs, whereas positive affect isrelated to reduced ERNs (e.g., Luu, Collins, & Tucker, 2000;Wiswede, Münte, Krämer, & Rüsseler, 2009). Similarly, theACC activity during correct conflict trials might drive theregulation of dynamic cognitive control (cf. van Steenbergen,Band, et al., 2012; van Steenbergen, Band, et al., 2014). If theavoidance state indeed amplifies the ACC response to conflict-ing cues, it follows that it should also strengthen behavioraladaptations to cognitive conflict—that is, a phasic, conflict-driven adaptation of cognitive control. Conversely, being inan approach state might inhibit ACC activity in response toconflicting signals, and thereby decrease conflict monitoringdirectly following conflict (cf. Wiswede et al., 2009). If conflictprocessing diminishes in an approach state, this should conse-quently weaken behavioral adaptation.

One limitation of the present research is that our conclu-sions are based on behavioral data. It will therefore be inter-esting to investigate the neural mechanisms underlying thepresent effects in future studies. For example, using electro-encephalography, one could test the differential effects ofmotivational orientation on the strength of conflict-inducedACC activation (see van Steenbergen, Band, et al., 2012).More precisely, ongoing fronto-central theta oscillations in-duced by previous conflict may sustain longer in an avoidancestate than in an approach state, ostensibly reflecting ACCmodulation. Furthermore, using fMRI, future studies shouldclarify which specific brain regions alter cognitive controlprocesses in response to the current motivational state of anindividual, as well as their respective roles. For instance, it isconceivable that avoidance motivational orientation mightincrease activity in the (rostral) ACC (van Steenbergen,Band, et al., 2014) and/or functional connectivity betweenthe ACC and PFC.

To conclude, we have been the first to show the influenceof approach–avoidance body movements on the regulation of

sustained and dynamic adaptation of cognitive control. Tokeep our actions in line with our internal goals and intentions,our cognitive system is regulated by our current motivationalstate while it processes the information of salient environmen-tal cues. As compared to approach, avoidance was associatedwith a reduced congruency effect and an increased conflictadaptation effect. Our results suggest that approach and avoid-ance motivational orientations differently and independentlyinfluence both sustained and dynamic cognitive control.

Author note This research was supported by NWO Grant No. 056-22-012. We are grateful to Wout Duthoo and Senne Braem for their help indesigning the flanker task described in Study 3.

References

Barsalou, L. W. (2008). Grounded cognition. Annual Review ofPsychology, 59, 617–645. doi:10.1146/annurev.psych.59.103006.093639

Berkman, E. T., & Lieberman, M. D. (2009). Approaching the bad andavoiding the good: Lateral prefrontal cortical asymmetry distin-guishes between action and valence. Journal of CognitiveNeuroscience, 22, 1970–1979. doi:10.1162/jocn.2009.21317

Botvinick, M. M. (2007). Conflict monitoring and decision making:Reconciling two perspectives on anterior cingulate function.Cognitive, Affective, & Behavioral Neuroscience, 7, 356–366. doi:10.3758/CABN.7.4.356

Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen, J.D. (2001). Conflict monitoring and cognitive control. PsychologicalReview, 108, 624–652. doi:10.1037/0033-295X.108.3.624

Braem, S., King, J. A., Korb, F. M., Krebs, R. M., Notebaert, W., &Egner, T. (2013). Affective modulation of cognitive control is de-termined by performance-contingency and mediated by ventrome-dial prefrontal and cingulate cortex. Journal of Neuroscience, 33,16961–16970. doi:10.1523/jneurosci.1208-13.2013

Braem, S., Verguts, T., Roggeman, C., & Notebaert, W. (2012). Rewardmodulates adaptations to conflict. Cognition, 125, 324–332. doi:10.1016/j.cognition.2012.07.015

Cacioppo, J. T., Priester, J. R., & Berntson, G. G. (1993). Rudimentarydeterminants of attitudes: II. Arm flexion and extension have differ-ential effects on attitudes. Journal of Personality and SocialPsychology, 65, 5–17. doi:10.1037/0022-3514.65.1.5

Centerbar, D. B., & Clore, G. L. (2006). Do approach–avoidance actionscreate attitudes? Psychological Science, 17, 22–29. doi:10.1111/j.1467-9280.2005.01660.x

Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation:Immediate behavioral predispositions to approach or avoid thestimulus. Personality and Social Psychology Bulletin, 25, 215–224. doi:10.1177/0146167299025002007

Chiew, K. S., & Braver, T. S. (2011). Positive affect versus reward:Emotional and motivational influences on cognitive control.Frontiers in Psychology, 2, 279. doi:10.3389/fpsyg.2011.00279

Cretenet, J., & Dru, V. (2004). The influence of unilateral and bilateralarm flexion versus extension on judgments: An exploratory case ofmotor congruence. Emotion, 4, 282–294. doi:10.1037/1528-3542.4.3.282

Derryberry, D., & Reed, M. A. (1994). Temperament and attention:Orienting toward and away from positive and negative signals.Journal of Personality and Social Psychology, 66, 1128–1139.doi:10.1037/0022-3514.66.6.1128

Cogn Affect Behav Neurosci

Dreisbach, G., & Fischer, R. (2012). The role of affect andreward in the conflict-triggered adjustment of cognitive con-trol. Frontiers in Human Neuroscience, 6, 342. doi:10.3389/fnhum.2012.00342

Egner, T. (2008). Multiple conflict-driven control mechanisms in thehuman brain. Trends in Cognitive Sciences, 12, 374–380. doi:10.1016/j.tics.2008.07.001

Egner, T. (2011). Right ventrolateral prefrontal cortex mediates indi-vidual differences in conflict-driven cognitive control. Journalof Cognitive Neuroscience, 23, 3903–3913. doi:10.1162/jocn_a_00064

Egner, T., & Hirsch, J. (2005). Cognitive control mechanisms resolveconflict through cortical amplification of task-relevant information.Nature Neuroscience, 8, 1784–1790. doi:10.1038/nn1594

Elliot, A. J. (2008). Handbook of approach and avoidance motivation.Mahwah, NJ: Erlbaum.

Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters uponidentification of a target letter in a non- search task. Perception &Psychophysics, 16, 143–149. doi:10.3758/BF03203267

Fischer, R., Dreisbach, G., & Goschke, T. (2008). Context-sensitiveadjustments of cognitive control: Conflict-adaptation effects aremodulated by processing demands of the ongoing task. Journal ofExperimental Psychology: Learning, Memory, and Cognition, 34,712–718. doi:10.1037/0278-7393.34.3.712

Förster, J. (2003). The influence of approach and avoidance motor actionson food intake. European Journal of Social Psychology, 33, 339–350. doi:10.1002/ejsp.150

Förster, J., Higgins, E. T., & Idson, L. C. (1998). Approach and avoidancestrength during goal attainment: Regulatory focus and the “goallooms larger” effect. Journal of Personality and SocialPsychology, 75, 1115–1131. doi:10.1037/0022-3514.75.5.1115

Förster, J., & Strack, F. (1997). Motor actions in retrieval of valencedinformation: A motor congruence effect. Perceptual and MotorSkills, 85, 1419–1427. doi:10.2466/pms.1997.85.3f.1419

Förster, J., & Strack, F. (1998). Motor actions in retrieval of valencedinformation: Ii. Boundary conditions for motor congruence effects.Perceptual and Motor Skills, 86, 1423–1426. doi:10.2466/pms.1998.86.3c.1423

Forster, S. E., Carter, C. S., Cohen, J. D., & Cho, R. Y. (2011). Parametricmanipulation of the conflict signal and control-state adaptation.Journal of Cognitive Neuroscience, 23, 923–935. doi:10.1162/jocn.2010.21458

Fredrickson, B. L., Mancuso, R. A., Branigan, C., & Tugade, M. M.(2000). The undoing effect of positive emotions. Motivation andEmotion, 24, 237–258. doi:10.1023/A:1010796329158

Friedman, R. S., & Förster, J. (2000). The effects of approach andavoidance motor actions on the elements of creative insight.Journal of Personality and Social Psychology, 79, 477–492. doi:10.1037/0022-3514.79.4.477

Friedman, R. S., & Förster, J. (2005a). Effects of motivational cues onperceptual asymmetry: Implications for creativity and analyticalproblem solving. Journal of Personality and Social Psychology,88, 263–275. doi:10.1037/0022-3514.88.2.263

Friedman, R. S., & Förster, J. (2005b). The influence of approach andavoidance cues on attentional flexibility. Motivation and Emotion,29, 69–81. doi:10.1007/s11031-005-7954-4

Fritz, J., & Dreisbach, G. (2013). Conflicts as aversive signals: Conflictpriming increases negative judgments for neutral stimuli. Cognitive,Affective, & Behavioral Neuroscience, 13, 311–317. doi:10.3758/s13415-012-0147-1

Gomez, A., & Gomez, R. (2002). Personality traits of the behavioralapproach and inhibition systems: Associations with processing ofemotional stimuli. Personality and Individual Differences, 32,1299–1316. doi:10.1016/S0191-8869(01)00119-2

Gratton, G., Coles, M. G. H., & Donchin, E. (1992). Optimizing the useof information: Strategic control of activation of responses. Journal

of Experimental Psychology: General, 121, 480–506. doi:10.1037/0096-3445.121.4.480

Heilman, K., & Van Den Abell, T. (1980). Right hemisphere domi-nance for attention: The mechanism underlying hemisphericasymmetry of inattention. Neurology, 30, 327–330. doi:10.1212/WNL.30.3.327

Higgins, E. T. (1999). Promotion and prevention as a motivationalduality. In S. Chaiken & Y. Trope (Eds.), Dual-process theoriesin social psychology (pp. 503–525). New York, NY: GuilfordPress.

Higgins, E. T., & Tykocinski, O. (1992). Self-discrepancies and biograph-ical memory: Personality and cognition at the level of psychologicalsituation. Personality and Social Psychology Bulletin, 18, 527–535.doi:10.1177/0146167292185002

Hommel, B., Proctor, R. W., & Vu, K. L. (2004). A feature-integrationaccount of sequential effects in the simon task. PsychologicalResearch, 68, 1–17. doi:10.1007/s00426-003-0132-y

Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ:Prentice-Hall.

Kerns, J. G., Cohen, J. D., MacDonald, A. W., III, Cho, R. Y., Stenger, V.A., & Carter, C. S. (2004). Anterior cingulate conflict monitoringand adjustments in control. Science, 303, 1023–1026. doi:10.1126/science.1089910

Koch, S. (2008). Cognitive control in motion: The regulatory function ofapproach and avoidance movements. Doctoral dissertation,Radboud University Nijmegen, Nijmegen, The Netherlands.Retrieved from http://hdl.handle.net/2066/76673

Koch, S., Holland, R. W., Hengstler, M., & van Knippenberg, A. (2009).Body locomotion as regulatory process: Stepping backward en-hances cognitive control. Psychological Science, 20, 549–550. doi:10.1111/j.1467-9280.2009.02342.x

Koch, S., Holland, R. W., & van Knippenberg, A. (2008).Regulating cognitive control through approach–avoidance mo-tor actions. Cognition, 109, 133–142. doi:10.1016/j.cognition.2008.07.014

Kuhbandner, C., & Zehetleitner, M. (2011). Dissociable effects of valenceand arousal in adaptive executive control. PLoS ONE, 6, e29287.doi:10.1371/journal.pone.0029287

Luu, P., Collins, P., & Tucker, D. M. (2000). Mood, personality, and self-monitoring: Negative affect and emotionality in relation to frontallobe mechanisms of error monitoring. Journal of ExperimentalPsychology: General, 129, 43–60. doi:10.1037/0096-3445.129.1.43

Mayr, U. (2004). Conflict, consciousness, and control. Trends inCognitive Sciences, 8, 145–148. doi:10.1016/j.tics.2004.02.006

Mayr, U., & Awh, E. (2008). The elusive link between conflict andconflict adaptation. Psychological Research, 73, 794–802. doi:10.1007/s00426-008-0191-1

Nieuwenhuis, S., Stins, J. F., Posthuma, D., Polderman, T. J., Boomsma,D. I., & de Geus, E. J. (2006). Accounting for sequential effects inthe flanker task: Conflict adaptation or associative priming?Memory& Cognition, 34, 1260–1272. doi:10.3758/BF03193270

Nussinson, R., Häfner, M., Seibt, B., Strack, F., & Trope, Y. (2012).Approach/avoidance orientations affect self-construal and identifi-cation with in-group. Self and Identity, 11, 255–272. doi:10.1080/15298868.2011.559044

Padmala, S., Bauer, A., & Pessoa, L. (2011). Negative emotion impairsconflict-driven executive control. Frontiers in Psychology, 2, 192.doi:10.3389/fpsyg.2011.00192

Pribram, K. H., &McGuinness, D. (1975). Arousal, activation, and effortin the control of attention. Psychological Review, 82, 116–149. doi:10.1037/h0076780

Rutherford, H. J. V., & Lindell, A. K. (2011). Thriving and surviving:Approach and avoidance motivation and lateralization. EmotionReview, 3, 333–343. doi:10.1177/1754073911402392

Scherbaum, S., Fischer, R., Dshemuchadse, M., & Goschke, T. (2011).The dynamics of cognitive control: Evidence for within-trial conflict

Cogn Affect Behav Neurosci

adaptation from frequency-tagged EEG. Psychophysiology, 48,591–600. doi:10.1111/j.1469-8986.2010.01137.x

Schmidt, J. R. (2013). Questioning conflict adaptation: Proportion con-gruent and gratton effects reconsidered. Psychonomic Bulletin &Review, 20, 615–630. doi:10.3758/s13423-012-0373-0

Schmidt, J. R., & de Houwer, J. (2011). Now you see it, now you don’t:Controlling for contingencies and stimulus repetitions eliminates thegratton effect. Acta Psychologica, 138, 176–186. doi:10.1016/j.actpsy.2011.06.002

Scholer, A. A., Stroessner, S. J., & Higgins, E. T. (2008). Responding tonegativity: How a risky tactic can serve a vigilant strategy. Journalof Experimental Social Psychology, 44, 767–774. doi:10.1016/j.jesp.2007.06.006

Schouppe, N., de Houwer, J., Ridderinkhof, R. K., & Notebaert, W.(2012). Conflict: Run! Reduced Stroop interference with avoidanceresponses. Quarterly Journal of Experimental Psychology, 65,1052–1058. doi:10.1080/17470218.2012.685080

Shah, J., Higgins, E. T., & Friedman, R. S. (1998). Performance incen-tives and means: How regulatory focus influences goal attainment.Journal of Personality and Social Psychology, 74, 285–293. doi:10.1037/0022-3514.74.2.285

Strachman, A., & Gable, S. L. (2006). What you want (and do not want)affects what you see (and do not see): Avoidance social goals andsocial events. Personality and Social Psychology Bulletin, 32,1446–1458. doi:10.1177/0146167206291007

Stürmer, B., Leuthold, H., Soetens, E., Schröter, H., & Sommer, W.(2002). Control over location-based response activation in the simontask: Behavioral and electrophysiological evidence. Journal ofExperimental Psychology: Human Perception and Performance,28, 1345–1363. doi:10.1037/0096-1523.28.6.1345

Stürmer, B., Nigbur, R., Schacht, A., & Sommer, W. (2011). Reward andpunishment effects on error processing and conflict control.Frontiers in Psychology, 2, 335. doi:10.3389/fpsyg.2011.00335

Ullsperger, M., Bylsma, L. M., & Botvinick, M. M. (2005). The conflictadaptation effect: It’s not just priming. Cognitive, Affective, &Behavioral Neuroscience, 5, 467–472. doi:10.3758/CABN.5.4.467

van den Bergh, B., Schmitt, J., Dewitte, S., &Warlop, L. (2009). Bendingarms, bending discounting functions-how motor actions affect

intertemporal decision-making. Doctoral dissertation, CatcholicUniversity Leuven, Leuven, Belgium. Retrieved from https://lirias.kuleuven.be/

van Prooijen, J., Karremans, J. C., & van Beest, I. (2006). Proceduraljustice and the hedonic principle: How approach versus avoidancemotivation influences the psychology of voice. Journal ofPersonality and Social Psychology, 91, 686–697. doi:10.1037/0022-3514.91.4.686

van Steenbergen, H., Band, G. P. H, & Hommel, B. (2009). Rewardcounteracts conflict adaptation: Evidence for a role of affect inexecutive control. Psychological Science, 20, 1473–1477. doi:10.1111/j.1467-9280.2009.02470.x

van Steenbergen, H., Band, G. P. H., &Hommel, B. (2010). In the mood foradaptation: How affect regulates conflict-driven control. PsychologicalScience, 21, 1629–1634. doi:10.1177/0956797610385951

van Steenbergen, H., Band, G. P. H., & Hommel, B. (2012). Rewardvalence modulates conflict-driven attentional adaptation:Electrophysiological evidence. Biological Psychology, 90, 234–241. doi:10.1016/j.biopsycho.2012.03.018

van Steenbergen, H., Band, G. P. H., Hommel, B., Rombouts, S. A., &Nieuwenhuis, S. (2014). Hedonic hotspots regulate cingulate-drivenadaptation to cognitive demands. Cerebral Cortex. doi:10.1093/cercor/bht416. Advance online publication.

van Steenbergen, H., Booij, L., Band, G. P. H., Hommel, B., & van derDoes, A. J. W. (2012). Affective regulation of cognitive-controladjustments in remitted depressive patients after acute tryptophandepletion. Cognitive, Affective, & Behavioral Neuroscience, 12,280–286. doi:10.3758/s13415-011-0078-2

van Stekelenburg, J., & Klandermans, B. (2003). Regulatory focus metenmet behulp van spreekwoorden. In D. Wigboldus, M. Dechesne, E.Gordijn, & E. Kluwer (Eds.), Jaarboek sociale psychologie 2003(pp. 345–357). Groningen, The Netherlands: ASPO Pers.

Wendt, M., Kiesel, A., Geringswald, F., Purmann, S., & Fischer, R.(2013). Attentional adjustment to conflict strength. ExperimentalPsychology, 61, 55–67. doi:10.1027/1618-3169/a000227

Wiswede, D., Münte, T. F., Krämer, U.M., &Rüsseler, J. (2009). Embodiedemotion modulates neural signature of performance monitoring. PLoSONE, 4, e5754. doi:10.1371/journal.pone.0005754.g001

Cogn Affect Behav Neurosci