Journal of Affective Disorders 125 (2010) 316–322

Contents lists available at ScienceDirect

Journal of Affective Disorders

j ourna l homepage: www.e lsev ie r.com/ locate / j ad

Research report

Attentional bias in complicated grief

Fiona Maccallum, Richard A. Bryant⁎University of New South Wales, Sydney, Australia

a r t i c l e i n f o

⁎ Corresponding author. School of Psychology, UnWales, N.S.W., 2052, Australia. Tel.: +61 2 9385 3640;

E-mail address: r.bryant@unsw.edu.au (R.A. Bryan

0165-0327/$ – see front matter © 2010 Elsevier B.V.doi:10.1016/j.jad.2010.01.070

a b s t r a c t

Article history:Received 29 March 2009Received in revised form 25 January 2010Accepted 25 January 2010Available online 18 May 2010

Background: Complicated Grief (CG) is a debilitating potential consequence of bereavement.Despite the significant health costs associated with CG, relatively little is known about thecognitive processes associated with the condition. This study investigated information processingin CG.Method: Twenty four individuals with CG and 25 bereaved individuals without CG completed amodified emotional Stroop task in which theywere presented with death-related and neutral cuewords. Half of the participants were also given instructions to suppress thoughts of their lovedone's death while completing the task.Results: CG participants were slower to color name death-related words than No-CG participants,and were slower to color name death-related words than neutral words. This pattern of findingssuggests an attentional bias towards loss-related events.Conclusions: This study represents the first demonstration of an information bias within CG.Consistentwith cognitivemodels of CG, it is possible that dysphoricmood and preoccupationwiththe loss are maintained by selectively attending to reminders of the loss.

© 2010 Elsevier B.V. All rights reserved.

Keywords:Complicated griefBereavementAttentional bias

Increasing evidence indicates that at least 10% of bereavedindividuals will develop Complicated Grief (CG) (alternatelyknownasProlongedGrief) (Prigersonet al., 2008; Stroebe et al.,2007). CG is characterized by a persistent sense of yearning forthe deceased, difficulty accepting the loss, bitterness, lack oftrust, and a loss of perceived meaning in life that is ongoing forat least 6 months after the death (Prigerson et al., 1995a; Zhanget al., 2006). CG is associated with distinct negative psycholog-ical and health outcomes (Boelen and van den Bout, 2008;Boelen et al., 2003a; Prigerson et al., 1995a, 2008). Despite thesignificant public health costs associated with CG, there iscurrently insufficient research into the mechanisms underpin-ning CG.

Many bereaved individuals experience intense yearning,intrusive thoughts, and dysphoric emotions in the earlyweeks and months of their bereavement, but these reactionsgradually subside over time. In contrast, individuals with CGappear stuck in a chronic state of mourning, with intense

iversity of New Southfax:+61 2 9385 3641t).

All rights reserved.

.

yearning and longing for the deceased continuing unabated(Prigerson et al., 2008). Emerging models of CG propose thatseparation distress becomes marked and persistent in CG, inpart, due to insufficient emotional processing of the loss,which results in a failure to update attachment schemas(Boelen et al., 2006b; Shear and Shair, 2005). A mismatchdevelops between the bereaved person's mental representa-tions about their self, the deceased, and the world, and thereality of the death. It is proposed that this discrepancybetween established mental representations of attachmentand the actuality of the death leads to greater occurrence ofintrusive thoughts and attention to loss-related events (seealso Dalgleish and Power, 2004). Although there is conver-gent evidence for frequent intrusions of the deceased in CG(Boelen and Huntjens, 2008; Prigerson et al., 2008; Raphaeland Martinek, 1997), no studies have investigated preferen-tial bias to loss-related stimuli in CG.

One potentially useful method to investigate possiblepreferential information processing in CG is the emotionalStroop paradigm. In this paradigm, participants are presentedwith words relevant to their domain of concern along withneutral words; the difference between the time taken to color

317F. Maccallum, R.A. Bryant / Journal of Affective Disorders 125 (2010) 316–322

name target words compared to neutral words is taken as anindication of the degree to which these target words capturethe individual's attention and cause interference in the colornaming task (for review see Harvey et al., 2004; MacLeod,2005). Using this paradigm, studies have reported slowedcolor naming for concern-relevant words in a range ofemotional disorders, including posttraumatic stress disorder(Bryant and Harvey, 1995; Cassiday et al., 1992; Foa et al.,1991; Harvey et al., 1996; Thrasher et al., 1994), generalizedanxiety disorder (Mathews and MacLeod, 1985), problemdrinking (Sharma et al., 2001), and eating disorders (Sackvilleet al., 1998). Adapting this method to include target wordsthat are relevant to CG would provide a means of experi-mentally indexing the salience of this information in CG.

Theorists have also proposed that CG reactions persistwhen individuals engage in avoidance behaviors (e.g., Boelenet al., 2006b; Shear et al., 2007; Shear and Shair, 2005).Avoidance is thought to impede habituation to painfulmemories and interfere with the integration of the loss intopre-existing schemas (Boelen et al., 2006b; Foa and Kozak,1986; Horowitz, 1986; Shear et al., 2007; Shear and Shair,2005; see also Ehlers and Clarke, 2000). In the case ofbereavement, avoidance may involve behavioral avoidance ofreminders of the deceased or cognitive avoidance strategies,such as thought suppression and rumination. Boelen et al.(Boelen et al., 2003c, 2006a) investigated the role ofavoidance in CG using self-report methodology. In supportof theoretical predictions, this study found that severity of CGand depression was related to endorsement of avoidancestrategies. However, in a longitudinal study the relationshipbetween avoidance and the development of CG across time,after controlling for initial symptoms, was less straightfor-ward (Boelen et al., 2006a). Accordingly, there is a need tomore directly examine the role of avoidance strategies in CG.

Thought suppression has been identified as one of themore common forms of cognitive avoidance across a range ofdisorders (Boelen et al., 2003c; Boelen et al., 2006a; Harveyet al., 2004), and has been correlated with grief severity in across sectional study (Boelen et al., 2003b). Numerous studieshave demonstrated that the intentional suppression ofthoughts and memories can produce a paradoxical increasein the frequency of these phenomena (for review see Harveyet al., 2004; Wenzlaff and Wegner, 2000). Thought suppres-sion has also been linked to the development of attentionalbias in nonclinical participants (Lavy and van den Hout,1994). Ironic control theory explains the paradoxical effect ofthought suppression by postulating that thought suppressionis mediated by two simultaneous processes: an operatingprocess and amonitoring process. The operating process aimsto prevent the unwanted thought from entering awarenessby actively searching for distracters. In parallel, the monitor-ing process searches for indications of the failure of mentalcontrol (i.e., presence of the unwanted thought). The theoryholds that the operating process is a more effortful process,and so the search for alternate thoughts is susceptible tobeing disrupted by demands on cognitive resources. Incontrast, the monitoring process requires fewer cognitiveresources, and is less easily disrupted. When mental capacityis diminished by additional cognitive load, such as clinicallevels of anxiety or intrusions, the operating process becomesdisrupted while the monitoring process continues its search.

In this way, themonitoring process can elicit awareness of theunwanted thoughts and the to-be-suppressed material willbe relatively more activated than any competing mentalinformation (Wegner, 1994; Wegner and Erber, 1992;Wegner et al., 1987; Wenzlaff and Wegner, 2000). This, inturn, may lead to an increase in emotional distress.

This study used the emotional Stroop paradigm andthought suppression task (Wegner and Erber, 1992) toexamine the impact of information salience and attemptedthought suppression on information processing in CG.Individuals with and without CG were administered amodified Stroop task that included neutral and CG-relatedwords. Death-related words were selected for use as targetwords to ensure targets were highly salient to the loss. Half ofthe participants in each group were asked to suppressthoughts of their loved one's death while completing thetask. We hypothesized that CG participants would respondmore slowly to death-relatedwords than control participants.Additionally, we expected that participants who were giveninstructions to suppress thoughts of the death would havegreater interference on death-related words than those notgiven the instruction.

1. Method

1.1. Participants

Twenty four individuals whomet diagnostic criteria for CG(3 male, 21 female) and 25 bereaved individuals without CG(5 male, 20 female) participated in this study. Participantswith CG were recruited from consecutive patients attendingthe Traumatic Stress Clinic for treatment of their CG. The No-CG participants responded to an advertisement to participatein a research project investigating grief experiences. Addi-tional exclusion criteria for the No-CG group included acurrent diagnosis of PTSD and major depression.

1.2. Measures

Complicated Grief Assessment (Zhang et al., 2006) is aclinician-administered semi-structured interview for asses-sing CG. The CGA interview is based on the self-reportInventory of Complicated Grief (Prigerson et al., 1995b) andprovides a diagnosis and severity rating for CG symptoms.The interview assesses for the presence of separation distress(Criterion A), difficulty accepting the death, emotionalnumbness, bitterness, difficulty re-engaging in life and asense of purposelessness and meaninglessness (Criterion B).Each symptom is rated on a five-point scale. A diagnosis of CGis given if Criteria A and B have beenmet for at least 6 monthsand there is evidence of serious day to day impairment infunctioning (Criterion C).

Clinical Administered PTSD Scale—2 (CAPS-2;Blake et al.,1995). The CAPS-2 is a structured clinical interview thatindexes the 17 symptoms described in the DSM-IV PTSDcriteria. Each symptom is rated on a five-point scale in termsof severity and frequency of the symptoms in the past month.

Structure Clinical Interview for the DSM-IV (SCID-IV; Firstet al., 2002). The depression model of the SCID was used toassess for the presence of Major Depressive Disorder.

318 F. Maccallum, R.A. Bryant / Journal of Affective Disorders 125 (2010) 316–322

Beck Depression Inventory—II (BDI-II; Beck et al., 1996). TheBDI-II was used to obtain a continuous measure of depressivesymptomatology. This 21-item self-report measure hasdemonstrated good reliability and validity (Beck et al., 1996).

1.3. Emotional Stroop task

Participants completed a four-color Stroop reaction time(RT) task. This task was administered on a 17-inch desktopcomputer screen using Inquisit software (version 2.0).Reaction time was recorded in milliseconds. During the taskparticipants pressed color-coded keys to indicate whether aseries of words presented on the screen were printed in red,blue, yellow, or green. There were 12 target words and 36non-target words selected from Bradley and Lang (1999) andLeech et al. (2001). Target and non-target words werematched for length and frequency of use. The target wordswere funeral, death, grief, morgue, dead, burial, corpse, coffin,cemetery, grave, cremate and die. Examples of non-targetwords included corner, machine, umbrella, museum, cork, hat,appliance, jug, whistle, and scissors. Six additional non-targetwords were used during a practice phase.

Each word was presented in size 32 Arial font andappeared in the middle of the screen. Words remained onthe screen until the participant responded. There was a 2000-millisecond interval between trials. To prompt the participantfor the upcoming word, each trial commenced with thepresentation of an asterisk in the centre of the screen for1000 ms. Prior to the test phase there were 15 practice trials.Participants received feedback after each practice trialconcerning their accuracy. The test phase consisted of 96trials. During the test phase each of the 48 cues was presentedtwice. Order of presentation was randomized across eachpresentation, with the exception that the first 3 test wordswere designated to be non-target words. No feedback wasgiven in the test phase.Words appeared in a different color oneach presentation.1

1.4. Procedure

All participants underwent a clinical assessment thatincluded the CGA interview, the CAPS, the SCID (depressionmodule), and the BDI-II. They returned on average one weeklater to complete the Stroop task. Participants were told thatthey would be completing a task on the computer that wasconcernedwithmeasuring how quickly they could respond todifferent stimuli. They were given the following instructions:

“You will see a series of words on the screen that areprinted either in red, blue, yellow or green. You will alsosee that red, blue, yellow and green colored squares havebeen placed on 4 keys on the key board. When you see aword, your task is to press the key that corresponds to thecolor the word has been printed in. I want you to indicatethe color of each word as quickly and as accurately as youcan. Try not to make mistakes, but try to be fast. Your

1 Four variations of the program were used to ensure that each word waspresented in each color an equal number of times. However, because oexpected associations between particular words and colors several wordswere not presented in each of the 4 colors. Specifically, “die” and “death”were not presented in red and “grief” did not appear in blue.

f

reactions will be timed. An asterisk will appear on thescreen before each word appears to show you where tolook.”

Participants then completed the 15 practice trials. Oncompletion of the practice trials, the experimenter gaveparticipants instructions to either suppress (Suppression) ornot suppress (Mention) thoughts about the death of theirloved one. In the suppression condition, the instructionsincluded the following:

“For the next few minutes, I want you to block allthoughts of [name's] death. Do not think about it. If anythoughts about this event do happen to pop into yourmind, push them away. Suppress them as quickly and asfirmly as you can. Do this in whatever way works, butpush any such thoughts out of your mind. Considerthoughts of [name's] death forbidden.”

The non-suppression (Mention) condition included thefollowing:

“For the next few minutes many different thoughts mayenter you mind, thoughts of what you are doing thisafternoon, what you did yesterday, [name's] death, or anyother thoughts. If you notice such thoughts, don't try tocontrol them, simply let them come and go as they please.Whatever thoughts come to mind, thoughts about whatyou are having for [lunch/dinner], or [name's] death youdon't need to try to control them, let them come and go,no thought is forbidden.”

After a 60-second delay (to allow participants to considerthe instructions), the Stroop task was commenced; theexperimenter repeated the Suppression or Mention instruc-tions prior to the test phase of the task. At the completion ofthe trials, all participants were asked to rate how much theyhad been trying to suppress thoughts of their loved one'sdeath during the task on a 7-point Likert-type scale (1 = notat all, 7 = completely trying to suppress thoughts).

2. Results

2.1. Participant characteristics

Table 1 indicates that the two groups did not differ interms of age or time since the death. As expected, participantsin the CG group scored higher than the No-CG group on theCGA [t(30.27)=−24.76, p<.001], BDI-II [ t(25.03)=−9.68,p<.001] and CAPS-2 [t(23.55)=−12.93, p<.001]. In the CGsample, 8 (33%) met criteria for PTSD on the CAPS-2; theseparticipants were divided evenly between the Suppressionand Control conditions. In addition, 19 (79%) CG participantsmet criteria for Major Depressive Disorder on the SCID-IV. Chisquare analysis revealed no difference between groups interms of their relationship to the deceased or the suddennessof the death. No No-CG participants met criteria for PTSD orMajor Depressive Disorder.

2 Similar results were obtained using a stepwise regression procedure.Probabality of F to enter the analysis was set at <.05 and the probability of Fto remove a predictor was set at >.10. CGA scores were the only variableentered into the equation, F(1,48)=10.02, p<.004. R2=.18, ΔR2=.18.

Table 1Participant characteristics.

CGn=24

No-CGn=25

Age 52.16 (13.98) 47.88 (14.81)CGA interview 32.29 (3.98) 10.06 (1.36)Years since death 3.96 (2.90) 3.51 (2.34)BDI-II 30.30 (11.87) 5.56 (3.24)CAPS-2 50.58 (18.06) 2.64 (2.01)Relationship of the deceased

Partner: 38% 40%Child 17% 8%Parent 37% 40%Sibling 8% 12%

Sudden death 50% 40%

Note: Standard deviations appear in parentheses. CGA = Complicated GriefAssessment; BDI-II = Beck Depression Inventory, Second Edition; CAPS-2 =Clinical Administered PTSD Scale, Version 2.

319F. Maccallum, R.A. Bryant / Journal of Affective Disorders 125 (2010) 316–322

2.2. Manipulation check

Mean subjective suppression ratings are presented inTable 2. A 2 (Participant Group)×2 (Instructions) analysis ofvariance (ANOVA) revealed significantmain effects forGroup [F(1,48)=25.54 p<.001], and Instruction [F(1,48)=10.53p<.003]. CG participants provided higher suppression ratingsthan No-CG participants, and participants in the Suppressioncondition provided higher suppression ratings than partici-pants in the Mention condition. To explore the impact of theinstruction within conditions, planned paired t-tests wereperformed separately for each group. Whereas CG participantsin the Suppression condition gave higher suppression ratingsthan participants in the Mention condition [t(22)=−2.93,p<.009], there was no significant difference between suppres-sion ratings of No-CG participantswho received Suppression orMention instructions.

2.3. Stroop reaction times

Mean reaction times to death and neutral words for CG andNo-CG participants are presented in Table 2. Prior to analysis,response errors and trials with response times of greater than2000 ms were removed from the data. This equated to 1.04% oftrials for the CG group and 1.75% for the No-CG group. A 2(Group)×2 (Instruction)×2 (Word Type) mixed model ANOVAconducted on response times indicated a significant main effectfor Word Type, F(1,45)=4.23, p<.042, and a significantinteraction between Word Type and Group F(1,45)=7.02,p<.012. There was no effect of suppression instructions onreaction times. Follow-up testing indicated that CG participantswere slower to name death words than No-CG participantst(47)=−2.52, p<.024. Reaction times for neutral wordsdid not differ between groups (p<.22). Planned comparisonswere carried out to assess hypothesized differences betweenreaction times to death and neutral words within groups. Aspredicted, participants in the CG group were significantlyslower to color name death-related words than neutral wordst(23)=−3.51, p<.003. The response times between thewordtypes did not differ in theno-CGgroup(p<.73). Toexamine themagnitude of the difference in reaction times to death andneutral words effect sizes were calculated. For the CG group,this effect size was d=.42, and for the No-CG group d=.01.

Although these results support predictions that CGwould beassociated with slower color naming of death-related words, itis possible that the pattern of responding was better accountedfor by the levels of co-morbid PTSD and depression in thesample. To examine this possibillity, a statistical regressionanalysis was conducted. ICG, BDI and CAPS scoreswere selectedas predictors and the difference in reaction times betweenwordtypes was entered as the dependent variable. Followingrecommendations from Field (2005) a backwards deletionprocedure was used to ensure that important variables wouldbe less likely to be excluded from the analysis (see alsoTabachnick and Fidell, 2001).2 The probability of F to remove apredictor was set at p>.10. PTSD and BDI scores were removedfrom the equation on steps 2 and 3 respectively without asignificant change in the amount of variance predicted by theequation. CGA scores were the only significant predictor in thefinal equation (F(1,48)=10.02, p<.004). Results of the finalstep of this analysis are presented in Table 3.

3. Discussion

This study investigated information processing biases inCG using an emotional Stroop task. As predicted, participantswith CG were slower to respond to death-related words thanneutral words. Overall, CG participants were also slower torespond to death-related words than bereaved participantswithout CG. There was no difference between the reactiontimes of CG and No-CG participants to neutral words.Similarly, there was no difference in reaction times to death-related and neutral words for No-CG participants. This patternoffindings is consistentwith previous studies across a range ofemotional disorders that have found a bias in responding todisorder relevant words on the emotional Stroop task (BryantandHarvey, 1995; Foa et al., 1991; for review see Harvey et al.,2004; MacLeod, 2005).

The observation of a bias in responding to death-relatedwords accords with the proposal that the separation distressobserved in CG is associated with a focus on loss-relatedinformation (Boelen et al., 2006a,b; Prigerson et al., 2008),insofar as CG participants in this study demonstrated arelative disruption in task performance when presented withloss-related cues. The current finding extends current modelsof CG because if reminders of the loss have increased saliencein CG, then it is likely that individuals will allocate additionalcognitive resources to the loss as they ruminate on the loss,and the adverse consequences that it involves. There isincreasing evidence that CG is associated with rumination,maladaptive appraisals about the loss, and excessive atten-tion to one's incapacity to cope in the future without thedeceased (Boelen et al., 2003c, 2006a). There is also growingevidence that people with CG preferentially recall personalmemories of the deceased (Golden et al., 2007; Maccallumand Bryant, 2008). It is possible that the salience of loss-related cues initiates an information processing sequence inwhich the individual then engages in rumination, negativemood states, and a focus on distressing memories; this focus

Table 2Mean reaction times (milliseconds) and suppresion ratings on Stroop task.

CG No-CG

Suppressionn=12

Controln=12

Suppressionn=13

Controln=12

Word typeDeath 954.39 (156.49) 935.49 (200.49) 801.84 (134.10) 846.65 (191.05)Control 891.62 (111.33) 869.01 (149.48) 791.26 (117.73) 872.11 (179.94)

Subjective suppression rating a 6.00 (1.04) 4.33 (1.67) 3.46 (1.94) 2.00 (1.91)

Note: Standard deviations appear in parentheses.a Range 1–7.

320 F. Maccallum, R.A. Bryant / Journal of Affective Disorders 125 (2010) 316–322

may then directly contribute to maintenance of cognitivefactors that perpetuate CG.

We recognize several methodological issues, however,that temper this conclusion. First, we found that CGparticipants were slower at color naming concern-relevantwords than neutral words using an unmasked design. Assuch, we cannot draw conclusions about the stage ofprocessing at which interference is occurring. For example,it is possible that the pattern of results reported hererepresents a difficulty disengaging attention from salientinformation rather than an attentional bias towards thisinformation (for review see Harvey et al., 2004; MacLeod,2005). To further our understanding of the current findingfuture studies should employ masked designs, with sublim-inal presentation of target stimuli (e.g., Harvey et al., 1996).This will allow a determination of whether the response isoccurring at a preconscious or conscious level of processing.

Second, the current study compared bereaved participantswith CG and without CG. A large number of the CG samplealso met criteria for major depression, and a lesser number,PTSD. Although CG has been found to be an independentconstruct (Boelen and van den Bout, 2008; Boelen et al.,2003a) co-morbidity with PTSD and depression is common(Bonanno et al., 2007). Also, PTSD has been reliably associatedwith biased responding on the emotional Stroop task (Bryantand Harvey, 1995; McNally et al., 1993), raising the possibilitythat the current findings reflected an aspect of the co-morbidity in the sample rather than CG. Arguing against thisposition, stepwise multiple regression procedures indicatedthat the pattern of results in the present study was bestaccounted for by CG symptoms. Nonetheless, future studieswould benefit from comparing CG participants with andwithout co-morbidity to isolate the impact of CG onresponding. Third, the current study compared respondingbetween neutral and concern-relevant words, but did notinclude control words matched for emotionality. Accordingly,we cannot rule out the possibility that the observedinterference was due to the emotional tone of the wordsrather than the specific word relevance. Future studies would

Table 3Summary of final step of backwards stepwise regression model.

Variable B SE B β t p<

CGA −2.74 .87 −.419 −3.17 .003

Note: CGA = Complicated Grief Assessment. R2=.18, adjusted R2=.16.

benefit from the inclusion of negative and positive wordcategories to control for the emotionality of the death words.

Fourth, death-related words reflect only one aspect of theloss experience. Future researchwould benefit from includingadditional categories of loss-related target words, such asattachment-related words (e.g., loss, separation, or deceased-specific cues). Finally, there is a need to disentangle theextent to which CG participants are avoidant of these stimuli.In anxiety disorders it has been argued that selectiveattention does not result in greater elaboration of the salientmaterial, but rather facilitates early detection of potentialthreat, allowing the individual to abort further processing ofthreatening material (Williams et al., 1988). In the case of CG,it is argued that individuals may avoid reminders of thereality of the death but seek out reminders of the lost person(Boelen et al., 2006b; Prigerson et al., 2008; Shear and Shair,2005). Alternative methodologies to disentangle approachand avoidance tendencies may be employed, such as dotprobe (Pineles and Minkea, 2005) or eye-gaze (Bryant et al.,1995) paradigms, to directly index preferential encodingversus avoidance of grief-related stimuli. Further, recentstudies have begun to examine the neural mechanismsunderlying yearning in normal grief reactions (e.g., Freed etal., 2009; O'Connor et al., 2008). The application of functionalneuro-imaging techniques to investigate attentional bias inCG may shed further light on approach and avoidanceresponses.

Contrary to our hypothesis thought suppression instruc-tions did not result in longer reaction times to color nametarget words compared to control words. Previous studiesusing similar suppression manipulations have found that aninstruction to suppress target thoughts while completing aStroop task was associated with increased reaction times tocolor name target words (Klein, 2007; Wegner and Erber,1992). There are several possible interpretations for theabsence of this effect in the current study. Suppression ratingsdid not differ for No-CG participants receiving Suppressionand Mention instructions. Further, the CG group reportedmore suppression overall than the No-CG group. Thesefindings suggest that the manipulation to enhance suppres-sion in these participants was not fully successful. Otherstudies have also reported a failure in experimental manip-ulation of suppression in clinical samples (Harvey and Bryant,1998; Roemer and Salters, 2004), arguably because clinicalsamples tend to naturally engage in suppression, whichimpedes attempts to experimentally manipulate this strategy.This interpretation is consistent with proposals that thoughtsuppression is a feature of CG (e.g., Boelen et al., 2003b,

321F. Maccallum, R.A. Bryant / Journal of Affective Disorders 125 (2010) 316–322

2006a). In addition, the sample size used in this study maynot have allowed sufficient power to detect a difference inresponding according to suppression condition betweengroups. Future research will benefit from employing a rangeof methodological approaches, including multiple methodsfor assessing the experimental manipulation (e.g., stream ofconsciousness writing task) and indexing the impact ofsuppression (e.g., sentence unscrambling; see Wenzlaff andWegner, 2000) to advance our understanding of thoughtsuppression in CG.

In summary, the current study provides initial evidencethat CG is characterized by preferential processing ofinformation related to the death of a loved one. It is possiblethat this bias to loss-related stimuli may directly contribute tomaintenance of yearning and dysphoria because it results inrepeated attentional focus on events and reminders of theloss. Future research needs to clarify the specific processingmechanisms implicated in CG to understand the subsequentresponses to this initial attentional bias.

Role of funding sourceThe NHMRC had no role in the study design; in the collection, analysis

and interpretation of data; in the writing of the report; and in the decision tosubmit the paper for publication.

Conflict of interestNo authors are declaring a conflict of interest.

Acknowledgement

This research was supported by a National Health andMedical Research Council Project Grant.

References

Beck, A.T., Steer, R.A., Brown, G.K., 1996. Manual for the Beck DepressionInventory Second Edition. Psychological Corporation, San Antonio, TX.

Blake, D.D., Weathers, F.W., Nagy, L.M., Kaloupek, D.G., Gusman, F.D.,Charney, D.S., Keane, T.M., 1995. The development of a clinicianadministered PTSD scale. Journal of Traumatic Stress 8, 75–90.

Boelen, P.A., Huntjens, R.J.C., 2008. Intrusive images in grief: an exploratorystudy. Clinical Psychology and Psychotherapy 15, 217–226.

Boelen, P.A., van den Bout, J., 2008. Complicated grief and uncomplicatedgrief are distinguishable constructs. Psychiatric Research 157, 311–314.

Boelen, P.A., van den Bout, J., de Keijser, J., 2003a. Traumatic grief as adisorder distinct from bereavement-related depression and anxiety: areplication study with bereaved mental health care patients. AmericanJournal of Psychiatry 160, 1339–1341.

Boelen, P.A., van den Bout, J., van den Hout, M.A., 2003b. The role of cognitivevariables in psychological functioning after the death of a first degreerelative. Behaviour Research and Therapy 41, 1123–1136.

Boelen, P.A., van den Bout, J., van den Hout, M.A., 2003c. The role of negativeinterpretations of grief reactions in emotional problems after bereave-ment. Journal of Behavior Therapy and Experimental Psychiatry 34,225–238.

Boelen, P.A., van den Bout, J., van den Hout, M.A., 2006a. Negative cognitionsand avoidance in emotional problems after bereavement: a prospectivestudy. Behaviour Research and Therapy 44, 1657–1672.

Boelen, P.A., van den Hout, M.A., van den Bout, J., 2006b. A cognitive-behavioural conceptualization of complicated grief. Clinical Psychology:Science and Practice 13, 109–128.

Bonanno, G.A., Neria, Y., Mancini, A., Coifman, K.G., Litz, B.T., Insel, B., 2007. Isthere more to complicated grief than depression and posttraumaticstress disorder? Journal of Abnormal Psychology 116, 342–351.

Bradley, M.M., Lang, P.J., 1999. Affective norms for English words (ANEW):instruction manual and affective ratings. Technical report C — 1. In:T.C.f.R.i (Ed.), Psychophysiology. University of Florida.

Bryant, R.A., Harvey, A.G., 1995. Processing threatening information inposttraumatic stress disorder. Journal of Abnormal Psychology 104,537–541.

Bryant, R.A., Harvey, A.J., Gordon, E., Barry, R.J., 1995. Eye movement andelectrodermal responses to threat stimuli in post-traumatic stressdisorder. International Journal of Psychophysiology 20, 209–213.

Cassiday, J.B., McNally, R.J., Zeitlin, S.B., 1992. Cognitive processing of traumacues in rape victims with posttraumatic stress disorder. CognitiveTherapy and Research 16, 283–295.

Dalgleish, T., Power, M.J., 2004. Emotion-specific and emotion-non-specificcomponents of posttraumatic stress disorder (PTSD): implications for ataxonomy of related psychopathology. Behaviour Research and Therapy42, 1069–1088.

Ehlers, A., Clarke, D.A., 2000. A cognitive model of posttraumatic stressdisorder. Behaviour Research and Therapy 38, 319–345.

Field, A., 2005. Discovering Statistics Using SPSS. Sage, London.First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 2002. Structured Clinical

Interview for DSM-IV-TR Axis I Disorders, ResearchVersion, Patient Edition(SCID-I/P). Biometrics Research, New York State Psychiatric Institute.

Foa, E.B., Kozak, M.J., 1986. Emotional processing of fear: exposure tocorrective information. Psychological Bulletin 99, 20–35.

Foa, E.B., Feske, U., Murdock, T.B., Kozak, M.J., MaCarthy, P.R., 1991.Processing of threat related stimuli in rape victims. Journal of AbnormalPsychology 1000, 156–162.

Freed, P.J., Yanagihara, T.K., Hirsch, J., Mann, J.J., 2009. Neural mechanisms ofgrief regulation. Biological Psychiatry 66, 33–40.

Golden, A.-M., Dalgleish, T., Mackintosh, B., 2007. Levels of specificity ofautobiographical memories and of biographical memories of thedeceased in bereaved individuals with and without complicated grief.Journal of Abnormal Psychology 116, 786–795.

Harvey, A.G., Bryant, R.A., 1998. The effect of attempted thought suppressionin acute stress disorder. Behaviour Research and Therapy 36, 583–590.

Harvey, A.G., Bryant, R.A., Rapee, R.M., 1996. Preconscious processing ofthreat in posttraumatic stress disorder. Cognitive Therapy and Research20, 613–623.

Harvey, A.G., Watkins, E., Mansell, W., Shafran, R., 2004. CognitiveBehavioural Processes across Psychological Disorders: a TransdiagnosticApproach to Research and Treatment. Oxford University Press, Oxford.

Horowitz, M.J., 1986. Stress Response Syndromes, 2nd Ed. Aronson, Northvale, NJ.Klein, A.A., 2007. Suppression-induced hyperaccessibility of thoughts in

abstinent alcoholics: a preliminary investigation. Behaviour Researchand Therapy 45, 169–177.

Lavy, E.H., van den Hout, M.A., 1994. Cognitive avoidance and attentionalbias: causal relationships. Cognitive Therapy and Research 18, 179–191.

Leech, G., Rayson, P., Wilson, A., 2001. Word Frequencies in Written andSpoken English. Longman, London.

Maccallum, F., Bryant, R.A., 2008. Self defining memories in complicatedgrief. Behaviour Research and Therapy 46, 1311–1315.

MacLeod,C., 2005. The Stroop task inclinical research. In: Rubin, D.C.,Wenzel, A.(Eds.), Cognitive Methods and Their Application to Clinical Research.American Psychological Association, Washington, DC, pp. 41–62.

Mathews, A., MacLeod, C., 1985. Selective processing of threat cues in anxietystates. Behaviour Research and Therapy 23, 563–569.

McNally, R.J., English, G.E., Lipke, H.J., 1993. Assessment of intrusive cognitionin PTSD: use of the modified Stroop paradigm. Journal of TraumaticStress 16, 33–41.

O'Connor, M.-F., Wellisch, D.K., Stanton, A.L., Eisenberger, N.I., Irwin, M.R.,Lieberman, M.D., 2008. Craving love? Enduring grief activates brain'sreward center. NeuroImage 42, 969–972.

Pineles, S.L., Minkea, S., 2005. Attentional biases to internal and externalsources of potential threat in social anxiety. Journal of AbnormalPsychology 114, 314–318.

Prigerson, H.G., Frank, E., Kasl, S.V., Reynolds, C.F., Anderson, B., Zubenko, G.S.,Houck, P.R., George, C.J., Kupfer, D.J., 1995a. Complicated grief andbereavement-related depression as distinct disorders: preliminaryempirical validation in elderly bereaved spouses. American Journal ofPsychiatry 152, 22–30.

Prigerson, H.G., Maciejewski, P.K., Reynolds, C.F., Bierhals, A.J., Newsom, J.T.,Fasicka, A., Frank, E.,.J,.D.,.M,.M., 1995b. Inventory of complicated grief: ascale to measure maladaptive symptoms of loss. Psychiatric Research 59,65–79.

Prigerson, H.G., Vanderwerker, L.C., Maciejewski, P.K., 2008. Prolonged griefdisorder: a case for inclusion in DSM-V. In: Stroebe, M.S., Hansson, R.O.,Schut, H., Stroebe, W. (Eds.), Handbook of Bereavement Research andPractice: Advances in Theory and Intervention. American PsychologicalAssociation, Washington, DC, pp. 165–186.

Raphael, B., Martinek, N., 1997. Assessing traumatic bereavement andposttraumatic stress disorder. In: Wilson, J.P., Keane, T.M. (Eds.),Assessing Psychological Trauma. Guilford Press, New York, pp. 373–396.

Roemer, L., Salters, K., 2004. A preliminary study of the effects of directedsuppression of rape-related material amoung rape survivors usingunobtrusive measures. Behavioural and Cognitive Psychotherapy 32,149–164.

322 F. Maccallum, R.A. Bryant / Journal of Affective Disorders 125 (2010) 316–322

Sackville, T., Schotte, D.E., Touyz, S.W., Griffiths, R., Beaumont, P.J.V., 1998.Conscious and preconscious processing of food, body weight and shape,and emotion-related words in women with anorexia nervosa. Interna-tional Journal of Eating Disorders 23, 77–82.

Sharma, D., Albery, I.P., Cook, C., 2001. Selective attentional bias to alcohol-related stimuli in problem-drinkers and non-problem drinkers. Addic-tion 96, 285–295.

Shear, K.M., Shair, H., 2005. Attachment, loss, and complicated grief.Developmental Psychobiology 47, 253–267.

Shear, K.M., Monk, T., Houck, P.R., Melhem, N.M., Frank, E., Reynolds, C.F.,Sillowash, R., 2007. An attachment-based model of complicated griefincluding the role of avoidance. European Archives of Psychiatry andClinical Neuroscience 257, 453–461.

Stroebe, M.S., Schut, H., Stroebe, W., 2007. Health outcomes of bereavement.Lancet 370, 1960–1973.

Tabachnick, B.G., Fidell, L.S., 2001. Using Multivariate Statistics. Allyn andBacon, Boston, MA.

Thrasher, S.M., Dalgleish, T., Yule, W., 1994. Information processing inposttraumatic stress disorder. Behaviour Research and Therapy 32,247–254.

Wegner, D.M., 1994. Ironic processes of mental control. Psychological Review101, 34–52.

Wegner, D.M., Erber, R., 1992. The hyperaccessibility of suppressed thoughts.Journal of Personality and Social Psychology 63, 903–912.

Wegner, D.M., Schneider, D.J., Carter, S.R., White, T.L., 1987. Paradoxicaleffects of thought suppression. Journal of Personality and SocialPsychology 53, 5–13.

Wenzlaff, R.M., Wegner, D.M., 2000. Thought suppression. Annual Review ofPsychology 51, 59–91.

Williams, J.M.G., Watts, F.N., MacLeod, C., Mathews, A., 1988. CognitivePsychology and Emotional Disorders. John Wiley & Sons, Oxford.

Zhang, B., El-Jawahri, A., Prigerson, H.G., 2006. Update of bereavementresearch: evidence-based guidelines for the diagnosis and treatment ofcomplicated grief. Journal of Palliative Medicine 9, 1188–1203.