Research report

Cognitive emotion regulation in euthymic bipolar disorder

Larissa Wolkenstein a,n, Julia C. Zwick a, Martin Hautzinger a, Jutta Joormann b

a Department of Psychology, Clinical Psychology and Psychotherapy, University of Tübingen, Germanyb Department of Psychology, Northwestern University, Evanston, IL, USA

a r t i c l e i n f o

Article history:Received 5 July 2013Received in revised form12 December 2013Accepted 13 December 2013Available online 28 December 2013

Keywords:Affective disordersBipolarDepressionEmotion regulationEuthymicCERQ

a b s t r a c t

Background: Based on findings indicating increased stress reactivity and prolonged stress recovery inindividuals with bipolar disorder (BD), it has been proposed that emotion regulation (ER) deficits lie atthe core of this disorder. Recent studies show an increased use of maladaptive ER strategies anda decreased use of adaptive ER strategies in BD. Whether this pattern is merely a correlate of affectiveepisodes or might be a stable characteristic of BD, however, remains to be explored. In addition, it isunclear whether these deficits in ER are specific to people with a history of BD.Methods: We examined whether euthymic BD individuals differ from healthy controls (HC) andindividuals with a history of Major Depressive Disorder (MDD) with respect to the cognitive ER strategiesthey habitually use (CERQ) in response to negative affect. The sample consisted of 42 bipolar patients,43 patients with MDD and 39 HC.Results: Compared to HC, euthymic BD and MDD individuals reported increased use of rumination,catastrophizing, and self-blame alongside decreased use of positive reappraisal, and putting intoperspective. No differences were found between BD and MDD groups.Limitations: These findings are based on self-reports reflecting the habitual use of ER-strategies. The useof more objective methods and the examination of the spontaneous use of ER-strategies in euthymic BDwould be desirable.Conclusions: Deficits in the habitual use of ER strategies may characterize BD and MDD individuals evenoutside of an acute episode and thereby play a role in the recurrence of affective disorders.

& 2014 Elsevier B.V. All rights reserved.

1. Introduction

According to the World Health Organization, bipolar disorder(BD) is one of the most disabling psychiatric disorders and aleading cause of non-fatal burdens (World Health Organization,2008). BD is characterized by recurrent episodes of depression,periods of sustained and abnormally elevated mood, and/or mixedstates with co-occurring depressive and manic symptoms(American Psychiatric Association, 2000). It has been proposedthat negative life events trigger increases in depressive symptomswhereas certain positive life events trigger (hypo-)manic symp-toms (Johnson, 2005). Research has also shown that individualswith BD display prolonged recovery following a stressful life event(Goplerud and Depue, 1985). In addition, experimental studiessuggest that bipolar spectrum disorders are associated withimpairments in emotional recovery. Subclinical cyclothymic sub-jects compared to healthy controls (HC), for example, exhibitedelevated cortisol levels 3 h after a stressful task indicating pro-longed stress recovery (Depue et al., 1985). Taken together, these

findings suggest deficits in regulatory processes in BD and variousauthors have proposed that difficulties in emotion regulation lie atthe core of BD (Phillips and Vieta, 2007; Johnson et al., 2007;Gruber, 2011).

According to Gross (1998) emotion regulation (ER) refers toprocesses by which individuals influence the appearance of emo-tions and how they experience and express these emotions. Someof these processes are implicit, automatic, and are performedwithout any consciousness or effort, whereas others are explicit,controlled, and are exerted consciously and with effort (Gyuraket al., 2011). Strategies that are performed consciously and effort-fully can be subdivided into behavioral ER strategies (e.g., situationselection, expressive suppression) and cognitive ER strategies (e.g.,cognitive reappraisal, rumination) (Gross and John, 2003;Garnefski and Kraaij, 2007). Although all ER strategies might behelpful in particular situations, studies suggest a general advan-tage of some ER strategies over others (e.g., Garnefski and Kraaij,2007; Nezlek and Kuppens, 2008). Garnefski and Kraaij (2007), forexample, showed that the use of catastrophizing, rumination, andself-blame is positively associated with symptoms of depressionand anxiety whereas the use of positive reappraisal is negativelyassociated with these symptoms. Thus, ER strategies can bedifferentiated into (primarily) adaptive strategies and (primarily)maladaptive strategies.

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/jad

Journal of Affective Disorders

0165-0327/$ - see front matter & 2014 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.jad.2013.12.022

n Correspondence to: Department of Clinical Psychology, University of Tübingen,Schleichstr. 4, 72076 Tübingen, Germany.Tel.: þ49 70 7129 77184; fax: þ49 70 7129 5219.

E-mail address: larissa.wolkenstein@uni-tuebingen.de (L. Wolkenstein).

Journal of Affective Disorders 160 (2014) 92–97

Despite the fact that BD is defined by dysregulated emotionalstates (American Psychiatric Association, 2000), few studies todate have investigated the habitual use of ER strategies in BD. Themajority of these studies used the Response Styles Questionnaire(RSQ; Nolen-Hoeksema et al., 1993) and highlighted that, inresponse to negative life events, remitted BD participants rumi-nate more than HC (Thomas et al., 2007; Van Der Gucht et al.,2009). Moreover, self-reported rumination has been associatedwith higher levels of depression and hypomania in adolescents atrisk for BD (Knowles et al., 2005; Thomas and Bentall, 2002). Theuse of a revised version of the RSQ (Knowles et al., 2005) hasfurther revealed that, in response to negative affect, depressed andremitted BD participants use adaptive coping strategies (i.e.,distraction and problem solving) less frequently compared tomanic BD participants (Thomas et al., 2007). Interestingly, manicBD participants reported to use adaptive coping strategies evenmore frequently than HC (Thomas et al., 2007).

Compared to the RSQ and the revised version of the RSQ theCognitive Emotion Regulation Questionnaire (CERQ; Garnefskiet al., 2001) samples a broader range of cognitive strategies usedto regulate emotions in response to negative events. In addition tothe habitual use of rumination, the CERQ also assesses the use ofself-blame, blaming others, catastrophizing, putting into perspec-tive, positive reappraisal, acceptance, positive refocusing, andrefocus on planning. Within the context of BD, the CERQ has beenpreviously used in two studies. The first study highlighted that, inaddition to the more frequent use of rumination, bipolar indivi-duals compared to HC report a more frequent use of catastrophiz-ing and self-blame in response to negative events (Green et al.,2011). This was confirmed by the second study, that also showed aless frequent use of putting into perspective within the BD group(Rowland et al., 2013). Concerning the other cognitive ER strate-gies assessed by the CERQ, neither Green et al. (2011) nor Rowlandet al. (2013) found any differences between bipolar patients andHC. However, the participation of the BD subjects in both studieswas based on diagnoses given previously to the beginning of thestudy in question (in other studies). In other words, subjects0

current mood states were not assessed. Therefore, it is not unlikelythat the inclusion of symptomatic individuals may have blurreddifferences between euthymic BD patients and HC: First, keepingin mind the symptoms of (hypo)-manic episodes, it is probablethat individuals in a (hypo-) manic mood state are more likely torefocus on positive things, to positively reappraise, and to refocuson planning when being confronted with a negative event. Second,in the study by Thomas et al. (2007) manic BD participantsreported a more frequent use of adaptive coping strategiesas assessed by the RSQ not only compared to depressed andeuthymic BD participants, but also compared to HC.

To summarize, there is preliminary evidence that euthymic BDindividuals show increased use of rumination in response to negativeevents. However, no study thus far has compared euthymic BDparticipants to HC in their habitual use of a broad range of cognitiveER strategies as assessed by the CERQ. That is, thus far we do notknow whether deficits in the use of cognitive ER strategies that havebeen reported in BD (Green et al., 2011; Rowland et al., 2013) aremerely a correlate of acute affective symptoms or are presentindependently of acute symptoms and might thus constitute a riskfactor for future episodes. Studies specifically examining euthymicBD participants are of particular importance, given that recentmodels propose that maladaptive reactions to negative and positiveaffect underlie the downward and upward spirals, respectively,which in turn might result in a depressive or a manic episode inBD (Gruber, 2011; Johnson, 2005). It is therefore the main goal of thisstudy to compare the habitual use of cognitive ER strategies of inter-episode BD and HC. Furthermore, to examine whether our findingsare specific to the bipolar spectrum of affective disorders or

generalize to other affective disorders, we included a sample ofremitted unipolar depressed patients. To our knowledge there is onlyone study thus far that has examined the habitual use of cognitive ERstrategies in remitted depression. In this study, recovered depressionwas associated with an increased use of rumination and catastro-phizing as well as a decreased use of putting into perspectivecompared to HC (Ehring et al., 2008).

Due to the shared risk of remitted BD and remitted MDD fordeveloping a depressive episode following a negative life eventand based on previous studies that have examined the habitualuse of ER in BD and MDD participants (Rowland et al., 2013; Greenet al., 2011; Ehring et al., 2008; Van Der Gucht et al., 2009), wehypothesized that euthymic BD patients as well as remitted MDDpatients display increased use of rumination and catastrophizingand decreased use of positive reappraisal and putting into per-spective. Given that BD has repeatedly been associated with anincreased use of self-blame (Green et al., 2011; Rowland et al.,2013), which has not been found in remitted MDD (Ehring et al.,2008), we further propose that euthymic BD patients, but notremitted MDD patients, display an increased use of self-blaming.

2. Methods

2.1. Participants

One hundred and twenty-four participants were recruitedthrough an outpatient clinic as well as through advertisementsposted on the Internet and within the community. Participantswere invited for an interview if they were deemed eligible basedon screening conducted per telephone. To determine the diag-nostic status of participants, trained interviewers administered theStructured Clinical Interview for DSM-IV (SCID; First et al., 1996).Participants with BD (n¼42) either met the diagnostic criteria forremitted bipolar I disorder (62%) or for remitted bipolar II disorder(38%), based on the diagnostic criteria in the fourth edition of theDiagnostic and Statistical Manual of Mental Disorders (DSM-IV;American Psychiatric Association, 2000). Participants with MDD(n¼43) either met DSM-IV criteria for recurrent MDD in remission(74.42%) or for an MDD single episode in remission (25.58%).Participants in the HC group (n¼39) did not meet diagnosticcriteria for any current or past Axis I disorder. Exclusion criteria forall participants were insufficient knowledge of the German lan-guage and age below 18 or above 69. Exclusion criteria for bothclinical groups were lifetime psychotic symptoms (except mood-congruent delusions within affective episodes), current alcohol orsubstance dependency (if they met the lifetime diagnostic criteriathey had to be abstinent for at least 24 months), current alcohol orsubstance abuse, cluster A personality disorders, borderline per-sonality disorder, antisocial personality disorder, and currentanorexia nervosa (BMIr18 kg/m²). Furthermore, participants inboth clinical groups had to be remitted for at least 8 weeks andwere required to take no medication or to take medication ona stable dosage for at least 4 weeks. Of the 42 BD participants 81%and of the 43 MDD participants 39.5% were on various medica-tions at the time of the study (one participant in the HC group tookantidepressants due to sleep disturbances).

2.2. Assessment of clinical symptoms

To assess self-reported current symptom levels of depression, weused the Quick Inventory for Depressive Symptomatology Self-Report(QIDS-SR; Rush et al., 2003). The QIDS-SR comprises 16 items thatassess presence and severity of 16 depression-related symptoms.It demonstrates high internal consistency with Cronbach0s α¼ .86(Rush et al., 2003) and has proven to be suitable not only for MDD

L. Wolkenstein et al. / Journal of Affective Disorders 160 (2014) 92–97 93

but also for BD (Bernstein et al., 2010). The QIDS-SR demonstratedgood internal consistency in the current sample (Cronbach0s α¼ .80).

As a clinician-rated instrument to assess the severity of currentdepressive symptoms we used the Hamilton Depression RatingScale (HAM-D; Hamilton, 1960), which comprises 21 items askingfor the presence and severity of 20 depression-related symptoms.Good internal consistency of this scale has been proven over years(Trajković et al., 2011). Within the current sample the internalconsistency was acceptable (Cronbach0s α¼ .69). To assure theremitted mood state of the clinical groups, we required a scorebelow 9 on the HAM-D for BD and MDD participants.

Self-reported severity of manic symptoms within the BD groupwas assessed with the self-report manic inventory which com-prises 48 true-false items (SRMI; Shugar et al., 1992). Within thecurrent sample the internal consistency of the SRMI was high(Cronbach0s α¼ .85). To assess clinician-rated symptom levels ofmania in the BD group, we used the Young Mania Rating Scale(YMRS; Young et al., 1978), which comprises 11 items asking forthe presence and severity of different mania-associated symptoms.With Cronbach0s α¼ .45 the internal consistency of this scale waspoor within our sample, which is probably not only due to thesmall sample size the calculation is based on (n¼42) but also tothe fact that all BD participants were remitted at the time of thestudy. To assure remission with respect to mania within the BDgroup, we required a score below 12 on the YMRS.

2.3. Cognitive emotion regulation questionnaire

The CERQ is a 36-item self-report questionnaire which hasbeen developed to measure the habitual use of nine cognitive ERstrategies (Garnefski et al., 2001). The strategies characterize theindividual0s style of responding to negative events. Each of thenine conceptually distinct subscales consists of four items. Internalconsistencies for the subscales have been reported to range fromCronbach0s α¼0.68 to Cronbach0s α¼0.83 (Garnefski et al., 2001).The internal consistencies in our sample where comparable withCronbach0s α ranging from 0.64 to 0.86. As in the study ofGarnefski et al. (2001), only the internal consistency of onesubscale was below 0.70 whereas five of the alphas were 0.80 orhigher. In a next step we looked at each study group separately.Within the BD group Cronbach0s α ranged from 0.71 to 0.90,within the MDD group it ranged from 0.59 to 0.83 and within theHC group Cronbach0s α ranged from 0.56 to 0.85.

2.4. Procedure

The procedures of the current study were approved by the localEthical Committee and are in accordance with the HelsinkiDeclaration.

Participants who were deemed eligible through the phonescreening were sent the SCID II screening questionnaire as wellas the CERQ to complete at home and to bring to their diagnosticassessment session. During the diagnostic assessment sessionsociodemographic data were recorded before the intervieweradministered the SCID-I and SCID-II (the latter if necessary), theYMRS (only BD participants), and the HDRS (only BD and MDDparticipants). Afterwards participants were asked to complete theQIDS-SR and the SRMI (only BD group).

2.5. Statistical analyses

Group means were compared with a multivariate analysis ofvariance (MANOVA) and subsequent univariate analyses of var-iance (ANOVAs). Furthermore, post-hoc t-tests were conducted.

3. Results

Demographic and clinical characteristics of the three groups arepresented in Table 1. The three groups differed significantly onlywith respect to current depressive symptoms as assessed by theQIDS-SR, F(2, 121)¼14.99, po .001, η2¼ .20. Follow-up tests indi-cated that both, the BD group, t(48)¼�5.16, po .001, as well asthe MDD group, t(56)¼�6.07, po .001, had significantly higherQIDS-SR scores compared to HC. The two clinical groups, however,did not differ from each other, t(83)¼0.57, p¼ .568.1 Even thoughboth clinical groups differed from HC with respect to currentdepressive symptoms, the scores of all participants lay clearlybelow the clinical cutoff indicating that all patients fulfilled criteriafor remission. In addition, the two clinical groups did not differfrom each other with respect to their HAM-D scores or age at firsttreatment.

3.1. Group differences in cognitive emotion regulation strategies

Our main hypotheses predicted group differences in the habi-tual use of particular ER strategies. Specifically, we predicted thateuthymic BD patients as well as remitted MDD patients display anincreased use of rumination and catastrophizing, and a decreaseduse of positive reappraisal and putting into perspective. Further-more we expected that euthymic BD patients should display anincreased use of self-blame compared to HC and MDD participants.Means and standard deviations are presented in Table 2. First,we conducted a MANOVA on the total scores of the nine CERQ

Table 1Demographic, neuropsychological and clinical characteristics of participants.

Characteristic Bipolar disorder Major depressive disorder Healthy controls

% M SD % M SD % M SD

Gender (female) 61.90 72.09 58.97University entrance diploma (yes) 69.05 72.09 64.10Age 40.86 12.79 36.91 13.35 42.18 13.27QIDS-SR 4.88 4.49 4.40 3.24 1.15 1.27HAM-D 3.00 2.55 3.58 2.40Age at first treatment 28.41 11.46 28.37 12.37SRMI 4.00 4.83YMRS 1.67 2.07

Note: QIDS-SR¼Quick Inventory of Depressive Symptomatology Self-Report; HAM-D¼Hamilton Depression Rating Scale; SRMI¼Self-Report Manic Inventory; YMRS¼YoungMania Rating Scale.

1 Within the whole sample the QIDS-SR was significantly correlated with allcognitive emotion regulation strategies (all rr� .178 or rZ .254 and all po .05).However, following the recommendations of Miller and Chapman (2001) werefrained from including the QIDS-SR as a covariate.

L. Wolkenstein et al. / Journal of Affective Disorders 160 (2014) 92–9794

subscales which yielded a significant main effect of ‘group’, F(18,226)¼4.07, po .001, η2¼ .245. Subsequent ANOVAs indicated asignificant main effect of ‘group’ for self-blame, F(2, 120)¼12.78,po .001, η2¼ .176, rumination, F(2, 120)¼24.11, po .001, η2¼ .287,catastrophizing, F(2, 120)¼11.37, po .001, η2¼ .159, putting intoperspective, F(2, 120)¼16.91, po .001, η2¼ .220, and positivereappraisal, F(2, 120)¼4.17, p¼ .018, η2¼ .065. There were no groupdifferences with respect to blaming others, F(2, 120)¼1.99, p¼ .141,positive refocusing, F(2, 120)¼2.59, p¼ .079, acceptance, F(2,120)¼0.24, p¼ .788, or refocus on planning, F(2, 120)¼1.84,p¼ .163.

Follow-up t-tests were performed on the five significant CERQsubscales to examine specific group difference in the use of thesestrategies. Compared to HC, BD participants reported more frequentuse of self-blame, t(63)¼�5.04, po .001, rumination, t(74)¼�5.67,po .001, and catastrophizing, t(57)¼�4.77, po .001, as well as lessfrequent use of putting into perspective, t(78)¼5.35, po .001, andpositive reappraisal, t(78)¼2.83, p¼ .006. We found the same patternfor MDD participants. Compared to HC, they too showed an increaseduse of self-blame, t(62)¼�4.95, po .001, rumination, t(79)¼�6.98,po .001, and catastrophizing, t(65)¼�4.56, po .001, alongside adecreased use of putting into perspective, t(79)¼4.49, po .001, andpositive reappraisal, t(79)¼2.23, p¼ .028. We did not find anysignificant differences between BD participants and MDD partici-pants (all pZ .264).

3.2. Association between cognitive emotion regulation strategies andresidual symptoms

We further examined whether the habitual use of cognitive ERstrategies is associated with self-reported residual symptomsseparately for the BD and the MDD subsamples.

Within the BD subsample we found a positive correlationbetween depressive residual symptoms as assessed by the QIDSand the use of blaming others (r¼ .541, po .001). In contrast,acceptance was negatively correlated with depressive residualsymptoms (r¼� .354, p¼ .021). Manic residual symptoms asassessed by the SRMI were positively correlated with self-blame(r¼ .359, p¼ .020).

Within the MDD subsample the QIDS sum score was negativelycorrelated with putting into perspective (r¼� .378, p¼ .012).

3.3. Within-group differences in the BD group with respect tocognitive emotion regulation strategies

As the BD group comprised n¼26 BD-I participants and n¼16BD-II participants, we decided to explore whether BD-I and BD-II

patients differed in their habitual use of cognitive emotionregulation strategies. First, we examined whether BD-I and BD-IIpatients differed in symptom levels. We did not find any within-group differences on the QIDS-SR (t(40)¼�1.50, p¼ .149), HAM-D(t(40)¼�0,87, p¼ .389), SRMI (t(40)¼0.56, p¼ .580), and YMRS(t(40)¼�1.47, p¼ .150). Second, we conducted a MANOVA on thetotal scores of the nine CERQ subscales. The main effect of ‘sub-group’ was not significant, F(9, 32)¼0.22, p¼ .989, indicating thatBD-I and BD-II patients do not differ with respect to their habitualuse of cognitive emotion regulation strategies.

4. Discussion

With the knowledge that dysregulated emotion is a definingcharacteristic of BD, this study examined whether euthymic BDindividuals differ from HC with respect to their habitual use of ERstrategies. To clarify whether deficits in ER are specific to thebipolar spectrum or generalize to affective disorders, we includeda sample of unipolar depressed patients in remission in the studydesign. We hypothesized that compared to HC both, euthymic BDpatients and remitted MDD patients, would display an increaseduse of rumination and catastrophizing as well as a decreased useof positive reappraisal and putting into perspective. Furthermore,we proposed that only remitted BD patients would display anincreased use of self-blaming compared to HC.

In line with our hypothesis, we found that BD as well as MDD,compared to HC, are associated with increased use of rumination andcatastrophizing alongside decreased use of positive reappraisal andputting into perspective. In contrast to our hypotheses, we addition-ally found that both, remitted BD patients as well as remitted MDDpatients, show an increased use of self-blame compared to HC.Indeed, our study yielded no differences between BD and MDDparticipants in self-reported habitual use of ER strategies. Further-more, no group differences were found with respect to the useof blaming others, acceptance, positive refocusing, or refocus onplanning.

A heightened tendency to ruminate in response to negativeevents has repeatedly been reported in BD populations (Greenet al., 2011; Thomas et al., 2007; Van Der Gucht et al., 2009).Studies have also shown that BD participants seem to ruminate tothe same extent as participants suffering from MDD (Kim et al.,2012; Johnson et al., 2008). However, studies addressing the latterissue did not assess the current mood state of their participants.Thus, it remained unclear whether the increased use of ruminationis also present in inter-episode BD or whether this finding ismerely due to a subsample of acutely depressed BD participants.Our results show that euthymic BD participants ruminate to thesame extent as participants with a history of MDD. Moreover, andin line with previous studies, we found that increased use ofcatastrophizing and decreased use of putting into perspective areassociated with a history of both, MDD (Ehring et al., 2008) and BD(Green et al., 2011; Rowland et al., 2013). Notably, this is the firststudy to demonstrate that neither increased use of rumination andcatastrophizing nor decreased use of putting into perspective arespecific to BD or MDD, but rather seem to reflect a general markerof vulnerability to affective disorders.

In addition, our results indicate that increased use of self-blameand decreased use of positive reappraisal are also markers ofaffective disorders, rather than being specific to either BD or MDD.However, this finding is in contrast to the results reported in astudy by Ehring et al. (2008), in which neither increased self-blame nor decreased reappraisal was found in remitted MDDparticipants, as well as to the two studies including BD individuals,which did not find decreased use of positive reappraisal in BD(Green et al., 2011; Rowland et al., 2013).

Table 2Means and standard deviations of all cognitive emotion regulation strategies forthe three groups.

CERQ subscale Bipolardisorder

Major depressivedisorder

Healthycontrols

M SD M SD M SD

Self-blame 10.81 3.71 10.93 4.03 7.53 1.91Blaming others 7.33 2.93 7.19 2.40 6.29 2.10Rumination 11.36 3.79 12.16 3.55 7.21 2.71Catastrophizing 8.14 3.59 7.63 3.01 5.24 1.57Putting into perspective 10.36 3.31 11.14 3.10 14.42 3.48Positive refocusing 8.64 3.60 9.19 3.47 10.42 3.61Positive reappraisal 11.36 3.80 11.77 4.17 13.74 3.71Acceptance 11.74 3.46 12.21 3.11 11.84 3.33Refocus on planning 10.57 3.58 11.21 2.87 11.97 3.32

Note: CERQ¼Cognitive Emotion Regulation Questionnaire. Italic values indicate allvalues that differed significantly between groups.

L. Wolkenstein et al. / Journal of Affective Disorders 160 (2014) 92–97 95

Differences in the samples may be responsible for differences infindings across the studies. Whereas the current study onlyincluded remitted bipolar patients, neither Green et al. (2011)nor Rowland et al. (2013) controlled for current mood state. Thus,while a decreased use of positive reappraisal may be characteristicof euthymic BD, the inclusion of symptomatic BD participants mayhave blurred this finding in previous studies. As outlined pre-viously, it is possible that particularly (hypo-)manic BD partici-pants make use of positive reappraisal, which might have led tonon-significant findings in the studies of Green et al. (2011) andRowland et al. (2013). This finding illustrates the importance ofclearly defining the mood state of BD participants in studieslooking at the habitual use of ER strategies. Whereas a decreaseduse of adaptive cognitive reframing techniques in remitted BDpatients may reflect impaired capacities to set aside the signifi-cance of a negative event and to re-appraise, an increased use ofthese strategies in (hypo-)manic BD individuals may be indicativeof their current symptoms.

Whereas Ehring et al. (2008) did not find differences betweenrecovered depressed participants and HC with respect to self-blame and positive reappraisal, we found increased use of self-blame and decreased use of reappraisal not only in euthymic BDbut also in remitted MDD participants. Here, the difference infindings could be due to differences in depressive residual symp-toms of the remitted MDD groups in the two studies, but it isunlikely. Whereas in the current study MDD differed from HCparticipants with respect to depressive symptoms, no such differ-ence was reported in the study by Ehring et al. (2008). However,with a mean QIDS-SR score of 4.40 within the MDD group, theMDD participants in the current study were clearly remitted, aswere the participants in the Ehring sample. Furthermore, we didnot find any correlations between depressive residual symptomsand self-blame or positive reappraisal within our MDD subsample.Thus, it remains unclear what causes the differences between ourfindings and the findings of Ehring et al. (2008). However, studiesin non-clinical samples consistently indicate significant relationsbetween increased use of self-blame, rumination, and catastro-phizing and decreased use of positive reappraisal and emotionalproblems (e.g., current and future depressive symptoms)(Garnefski and Kraaij, 2007, 2009). Thus, one would expect notonly euthymic BD patients but also remitted MDD patients todisplay increased use of self-blame and decreased use of positivereappraisal, which is in line with our results. Future studies arenecessary to replicate these findings.

In addition to the group differences, correlational analysesrevealed associations between depressive residual symptomswithin the BD group and two strategies that did not differ betweengroups: Blaming others was positively correlated with depressiveresidual symptoms whereas acceptance was negatively correlatedwith depressive symptoms. It is possible that compared to HC, BDpatients show a significantly increased extent of blaming others aswell as a significantly decreased extent of acceptance only duringacute depressive episodes. Future studies are required to directlyinvestigate these hypotheses.

Furthermore, an explorative within-group analysis yielded nosignificant difference between BD-I and BD-II patients withrespect to their habitual use of cognitive emotion regulationstrategies. As this analysis was based on a comparatively smallsample size, replication studies using larger samples will benecessary to provide further support for this result.

Our results demonstrate that it is vital to examine which ERstrategies BD patients use while in a stable euthymic mood state ascompared to their strategy selection in view of first affectivesymptoms of either polarity. It is possible, for example, that thestrategy of positive refocusing is useful inter-episode as well as inthe face of first depressive symptoms, whereas it might not be

useful when experiencing first (hypo-)manic symptoms. Gainingmore insight into these selection processes will be useful for thedevelopment of appropriate intervention strategies. In this contextit is worth mentioning that recent studies have demonstrated thateuthymic BD individuals compared to HC report to expendincreased effort to regulate their emotions independent ofwhether they watch an emotional or a non-emotional film-clip—suggesting that inter-episode BD individuals demonstrate a less-nuanced strategy selection with respect to the circumstances(Gruber et al., 2012).

4.1. Limitations

While the findings of our study contribute significantly to ourunderstanding of the habitual use of ER strategies within thecontext of euthymic BD and remitted MDD, there are a number oflimitations of this study that should be kept in mind. The focus ofthis study was on the regulation of negative affect. Previousresearch, however, has pointed out that individual differences inresponding to positive affect may play a critical role in BD (Johnsonet al., 2008; Gruber et al., 2011; Alloy et al., 2009). Dampening orpositive rumination, for example, have not only been shown to beassociated with depressive symptoms but also with symptomsthat might be of special relevance for elevated mood-states such asmania (Feldman et al., 2008). In contrast to depression, BD is notonly associated with an increased tendency to ruminate inresponse to negative affect but also with an increased self-reported tendency to ruminate in response to positive affect(Johnson et al., 2008; Gruber et al., 2011). Whereas trait-rumination in response to both negative and positive affect hasbeen associated with higher life-time frequency of depression(Alloy et al., 2009; Gruber et al., 2011), only trait-rumination inresponse to positive affect has also been linked to higher lifetime-frequency of mania (Gruber et al., 2011). Furthermore, positiverumination has been found to be associated with symptoms ofhypomania in a college student sample (Raes et al., 2009). Futurestudies should address this issue and broaden our knowledgeabout ER strategy use in euthymic BD patients by examiningresponding to negative as well as positive affect.

Furthermore, even though the CERQ is frequently used toassess the habitual use of cognitive emotion regulation strategies(Ehring et al., 2008; Green et al., 2011), it has to be kept in mindthat it is a self-report measure. Thus, it presumes a certain amountof introspective abilities to answer the questions of the CERQ. It isdebatable whether subjects do have comprehensive access to theirown cognitive processes and are able to give valid answers to suchquestions. There may be memory biases, for example, that distortparticipants0 response patterns. Thus, it will be interesting toextent the measures used to assess cognitive emotion regulationstrategies in euthymic bipolar patients by integrating more objec-tive measures in future studies. Of note, even though there are firststudies demonstrating which brain networks are involved inparticular emotion regulation strategies (McRae et al., 2008), ourknowledge with respect to this question is limited.

4.2. Conclusions

To summarize, euthymic BD as well as remitted MDD, isassociated with an increased use of (primarily) maladaptive ERstrategies and a decreased use of (primarily) adaptive ER strategies.Given that this pattern of ER strategy use is present duringremission, i.e., it is not associated with the presence of acuteaffective symptoms, it may likely be a risk factor for the develop-ment of further affective episodes. Studies are needed that gobeyond the examination of self-reported habitual use of ER strate-gies and assess whether the use of certain ER strategies predicts the

L. Wolkenstein et al. / Journal of Affective Disorders 160 (2014) 92–9796

recurrence of affective episodes and whether BD and MDD partici-pants can be trained to effectively use adaptive ER strategies. In arecent study Gruber et al. (2012) found that bipolar patients reportgreater effort but less success in spontaneously regulating theiremotions which suggests that besides the habitual use of certain ERstrategies there may be other deficits influencing ER ability inpatients suffering from affective disorders. Future studies areneeded to clarify which processes underlie these deficits to success-fully regulate emotions and to develop appropriate interventions.Furthermore, we need to examine whether ER deficits are specifi-cally characteristic for remitted affective disorders or whether theycharacterize individuals with any lifetime psychiatric diagnosis. Thelatter would indicate that deficient ER constitutes a risk factor forthe recurrence of either psychopathology.

Role of funding sourceWe had no funding for this study.

Conflict of interestAll authors declare that they have no conflicts of interest.

AcknowledgementsWe thank Marjorie Kinney, who kindly assisted with the proof-reading of the

manuscript.

References

Alloy, L.B., Abramson, L.Y., Flynn, M., Liu, R.T., Grant, D.A., Jager-Hyman, S., White-house, W.G., 2009. Self-focused cognitive styles and bipolar spectrum dis-orders. Concurrent and prospective associations. Int. J. Cognit. Ther. 2, 354–372.

American Psychiatric Association, 2000. Task Force on DSM-IV. Diagnostic andStatistical Manual of Mental Disorders: DSM-IV-TR. American PsychiatricAssociation, Washington, DC.

Bernstein, I.H., Rush, A.J., Suppes, T., Kyotoku, Y., Warden, D., 2010. The quickinventory of depressive symptomatology (clinician and self-report versions) inpatients with bipolar disorder. CNS Spectr. 15, 367–373.

Depue, R.A., Kleiman, R.M., Davis, P., Hutchinson, M., Krauss, S.P., 1985. Thebehavioral high-risk paradigm and bipolar affective disorder, VIII: serum freecortisol in non-patient cyclothymic subjects selected by the General BehaviorInventory. Am. J. Psychiatry 142, 175–181.

Ehring, T., Fischer, S., Schnülle, J., Bösterling, A., Tuschen-Caffier, B., 2008. Char-acteristics of emotion regulation in recovered depressed versus neverdepressed individuals. Pers. Individ. Differ. 44, 1574–1584.

Feldman, G.C., Joormann, J., Johnson, S.L., 2008. Responses to positive affect: a self-report measure of rumination and dampening. Cognitive Ther. Res. 32,507–525.

First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B., 1996. Structured Clinical Inter-view for DSM-IV. American Psychiatric Association, Washington, DC.

Garnefski, N., Kraaij, V., 2007. The cognitive emotion regulation questionnaire:psychometric features and prospective relationships with depression andanxiety in adults. Eur. J. Psychol. Assess. 23, 141–149.

Garnefski, N., Kraaij, V., 2009. Cognitive coping and psychological adjustment indifferent types of stressful life events. Individ. Differ. Res. 7, 168–181.

Garnefski, N., Kraaij, V., Spinhoven, P., 2001. Negative life events, cognitive emotionregulation and emotional problems. Pers. Individ. Differ. 30, 1311–1327.

Goplerud, E., Depue, R.A., 1985. Behavioral response to naturally occurring stress incyclothymia and dysthymia. J. Abnorm. Psychol. 94, 128–139.

Green, M.J., Lino, B.J., Hwang, E.-J., Sparks, A., James, C., Mitchell, P.B., 2011.Cognitive regulation of emotion in bipolar I disorder and unaffected biologicalrelatives. Acta Psychiatr. Scand. 124, 307–316.

Gross, J.J., 1998. The emerging field of emotion regulation: an integrative review.Rev. Gen. Psychol. 2, 271–299.

Gross, J.J., John, O.P., 2003. Individual differences in two emotion regulationprocesses: implications for affect, relationships, and well-being. J. Pers. Soc.Psychol. 85, 348–362.

Gruber, J., 2011. Can feeling too good be bad? Positive emotion persistence (PEP) inbipolar disorder. Curr. Dir. Psychol. Sci. 20, 217–221.

Gruber, J., Eidelman, P., Johnson, S.L., Smith, B., Harvey, A.G., 2011. Hooked on afeeling: rumination about positive and negative emotion in inter-episodebipolar disorder. J. Abnorm. Psychol. 120, 956–961.

Gruber, J., Harvey, A.G., Gross, J.J., 2012. When trying is not enough: emotionregulation and the effort—success gap in bipolar disorder. Emotion 12,997–1003.

Gyurak, A., Gross, J.J., Etkin, A., 2011. Explicit and implicit emotion regulation: adual-process framework. Cognit. Emot. 25, 400–412.

Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry23, 56–62.

Johnson, S.L., 2005. Life events in bipolar disorder: towards more specific models.Clin. Psychol. Rev. 25, 1008–1027.

Johnson, S.L., Gruber, J., Eisner, L., 2007. Emotion and bipolar disorder. In:Rottenberg, J., Johnson, S.L. (Eds.), Emotion and Psychopathology: BridgingAffective and Clinical Science. American Psychological Association, Washington, DC.

Johnson, S.L., Mckenzie, G., Mcmurrich, S., 2008. Ruminative responses to negativeand positive affect among students diagnosed with bipolar disorder and majordepressive disorder. Cognit. Ther. Res. 32, 702–713.

Kim, S., Yu, B.H., Lee, D.S., Kim, J.-H., 2012. Ruminative response in clinical patientswith major depressive disorder, bipolar disorder, and anxiety disorders.J. Affect. Disord. 136, e77–e81.

Knowles, R., Tai, S., Christensen, I., Bentall, R., 2005. Coping with depression andvulnerability to mania: a factor analytic study of the Nolen-Hoeksema (1991)response styles questionnaire. Br. J. Clin. Psychol. 44, 99–112.

McRae, K., Ochsner, K.N., Mauss, I.B., Gabrieli, J.J.D., Gross, J.J., 2008. Genderdifferences in emotion regulation: an fMRI study of cognitive reappraisal.Group Process. Intergr. Relat. 11, 143–162.

Miller, G.A., Chapman, J.P., 2001. Misunderstanding analysis of covariance.J. Abnorm. Psychol. 110, 40–48.

Nezlek, J.B., Kuppens, P., 2008. Regulating positive and negative emotions in dailylife. J. Pers. 76, 561–580.

Nolen-Hoeksema, S., Morrow, J., Fredrickson, B.L., 1993. Response styles and theduration of episodes of depressed mood. J. Abnorm. Psychol. 102, 20–28.

Phillips, M.L., Vieta, E., 2007. Identifying functional neuroimaging markers ofbipolar disorder: towards DSM-V. Schizophr. Bull. 33, 893–904.

Raes, F., Daems, K., Feldman, G.C., Johnson, S.L., Van Gucht, D., 2009. A psychometricevaluation of the Dutch version of the responses to positive affect question-naire. Psychol. Belg. 49, 293–310.

Rowland, J.E., Hamilton, M.K., Lino, B.J., Ly, P., Denny, K., Hwang, E.-J., Mitchell, P.B.,Carr, V.J., Green, M.J., 2013. Cognitive regulation of negative affect in schizo-phrenia and bipolar disorder. Psychiatry Res. 208, 21–28.

Rush, A.J., Trivedi, M.H., Ibrahim, H.M., Carmody, T.J., Arnow, B., Klein, D.N.,Markowitz, J.C., Ninan, P.T., Kornstein, S., Manber, R., Thase, M.E., Kocsis, J.H.,Keller, M.B., 2003. The 16-item quick inventory of depressive symptomatology(QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometricevaluation in patients with chronic major depression. Biol. Psychiatry 54,573–583.

Shugar, G., Schertzer, S., Toner, B.B., Digasbarro, I., 1992. Development, use, andfactor analysis of a self-report inventory for mania. Compr. Psychiatry 33,325–331.

Thomas, J., Bentall, R.P., 2002. Hypomanic traits and response styles to depression.Br. J. Clin. Psychol. 41, 309–313.

Thomas, J., Knowles, R., Tai, S., Bentall, R.P., 2007. Response styles to depressedmood in bipolar affective disorder. J. Affect. Disord. 100, 249–252.

Trajković, G., Starčević, V., Latas, M., Leštarević, M., Ille, T., Bukumirić, Z., Marinković, J.,2011. Reliability of the Hamilton Rating Scale for depression: a meta-analysis overa period of 49 years. Psychiatry Res. 189, 1–9.

Van Der Gucht, E., Morriss, R., Lancaster, G., Kinderman, P., Bentall, R.P., 2009.Psychological processes in bipolar affective disorder: negative cognitive styleand reward processing. Br. J. Psychiatry 194, 146–151.

World Health Organization, 2008. The Global Burden of Disease: 2004 Update.WHO Press, Geneva, Switzerland.

Young, R.C., Biggs, J.T., Ziegler, V.E., Meyer, D.A., 1978. A rating scale for mania:reliability, validity and sensitivity. Br. J. Psychiatry 133, 429–435.

L. Wolkenstein et al. / Journal of Affective Disorders 160 (2014) 92–97 97