cognitive bias modification approaches to anxiety...cognitive bias modiﬁcation (cbm): direct...

CP08CH08-MacLeod ARI 1 March 2012 8:11

Cognitive Bias ModificationApproaches to AnxietyColin MacLeod1 and Andrew Mathews2

1School of Psychology, The University of Western Australia, Crawley WA 6009 Australia;email: [email protected] of Psychology, University of California, Davis, California 95616, and MRCCognition and Brain Sciences Unit, Kings College, London CB2 2EF United Kingdom;email: [email protected]

Annu. Rev. Clin. Psychol. 2012. 8:189–217

First published online as a Review in Advance onOctober 25, 2011

The Annual Review of Clinical Psychology is onlineat clinpsy.annualreviews.org

This article’s doi:10.1146/annurev-clinpsy-032511-143052

Copyright c© 2012 by Annual Reviews.All rights reserved

1548-5943/12/0427-0189$20.00

Keywords

CBM, CBM-A, CBM-I, attentional bias, interpretive bias

Abstract

Clinical anxiety disorders and elevated levels of anxiety vulnerabilityare characterized by cognitive biases, and this processing selectivity hasbeen implicated in theoretical accounts of these conditions. We re-view research that has sought to evaluate the causal contributions suchbiases make to anxiety dysfunction and to therapeutically alleviate anx-iety using cognitive-bias modification (CBM) procedures. After con-sidering the purpose and nature of CBM methodologies, we show thatvariants designed to modify selective attention (CBM-A) or interpre-tation (CBM-I) have proven capable of reducing anxiety vulnerabilityand ameliorating dysfunctional anxiety. In addition to supporting thecausal role of cognitive bias in anxiety vulnerability and dysfunction andilluminating the mechanisms that underpin such bias, the findings sug-gest that CBM procedures may have therapeutic promise within clinicalsettings. We discuss key issues within this burgeoning field of researchand suggest future directions CBM research should take to maximizeits theoretical and applied value.

189

Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


Contents

INTRODUCTION. . . . . . . . . . . . . . 190BACKGROUND TO

COGNITIVE BIASMODIFICATIONRESEARCH . . . . . . . . . . . . . . . . . . 191Cognitive Bias and Anxiety. . . . . 191The Purpose of Cognitive Bias

Modification Research . . . . . . 193The Nature of Cognitive Bias

ModificationMethodologies . . . . . . . . . . . . . 193

COGNITIVE BIASMODIFICATIONTARGETINGATTENTIONALSELECTIVITY . . . . . . . . . . . . . . 194Development of Attentional

Bias ModificationTechniques . . . . . . . . . . . . . . . . 194

Single-Session Applicationsof Attentional BiasModification . . . . . . . . . . . . . . . 195

Extended Applicationsof Attentional BiasModification . . . . . . . . . . . . . . . 197

COGNITIVE BIASMODIFICATIONTARGETINGINTERPRETIVESELECTIVITY . . . . . . . . . . . . . . 198Development of Interpretive

Bias ModificationTechniques . . . . . . . . . . . . . . . . 198

Single-Session Applicationsof Interpretive BiasModification . . . . . . . . . . . . . . . 199

Extended Applicationsof Interpretive BiasModification . . . . . . . . . . . . . . . 201

COGNITIVE BIASMODIFICATIONTARGETING OTHERTYPES OF PROCESSINGSELECTIVITY . . . . . . . . . . . . . . 202Modification of Memory . . . . . . . 202Modification of Imagery . . . . . . . 202Modification of Appraisal . . . . . . 203

COGNITIVE BIASMODIFICATION ANDCOGNITIVEMECHANISMS . . . . . . . . . . . . . . 204Can CBM Findings Be

Attributed to DemandEffects? . . . . . . . . . . . . . . . . . . . . 204

Do CBM Effects ReflectChange in the IntendedCognitive Process? . . . . . . . . . 205

Is CBM Change Restricted tothe Targeted CognitiveBias? . . . . . . . . . . . . . . . . . . . . . . . 205

FUTURE DIRECTIONS FORCOGNITIVE BIASMODIFICATIONRESEARCH . . . . . . . . . . . . . . . . . . 206Extending the Reach

of CBM Research . . . . . . . . . . 207Enhancing the Efficacy of

Cognitive Bias ModificationProcedures . . . . . . . . . . . . . . . . . 207

Refining the TherapeuticApplication of CBM inClinical Contexts . . . . . . . . . . . 208

CLOSING COMMENTS . . . . . . . 209

INTRODUCTION

Scientific progress is characterized by a closeassociation between advancement of under-standing and increased ability to bring about

change. Deeper understanding of any complexsystem results in a heightened ability to iden-tify the changes to key variables needed to pro-duce desired outcomes. Of equal importance,it is by changing such variables and observing

190 MacLeod · Mathews

Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


the consequences that understanding is deep-ened. Hence, the ability to directly manipulatecomponent elements of a complex system is ofcrucial importance if we are to gain insight intoits nature and effectively exploit the resultingapplied benefits.

Cognitive models of anxiety vulnerabilityand dysfunction have proven highly influentialacross recent decades (cf. Brown & Barlow1994, Clark & Beck 2010). These modelsshare the premise that biased patterns of basicinformation processing, operating early withinthe cognitive system and at a low level thatmay be inaccessible to awareness, play a centralcausal role in vulnerability to experienceunduly intense anxiety symptoms, though thespecific nature of the proposed cognitive biasesvaries from theory to theory (cf. Mathews &MacLeod 2005). Despite the indirect supportsuch accounts have received from confirmationthat such processing biases are indeed charac-teristic of heightened anxiety vulnerability andanxiety pathology (cf. Ouimet et al. 2009), theability of researchers to adequately test the cru-cial idea that cognitive biases causally influenceclinically relevant symptoms has been handi-capped by the lack of established proceduresto directly manipulate these cognitive biases.This also has limited the ability of clinicians todeliver the potential therapeutic benefits suchtheoretical accounts predict should be gainedfrom the direct modification of these biases forindividuals experiencing problematic anxiety.

Hence, there has been considerable interestin recently developed techniques that haveproven capable of directly modifying low-levelcognitive biases implicated in such modelsof anxiety pathology (cf. Bar-Haim 2010,Hakamata et al. 2010, Hallion & Ruskio 2011,Hertel & Mathews 2011, Mathews 2011). Theexponential growth of research employingthese cognitive bias modification (CBM)procedures has been remarkable. Althoughthe seminal studies were conducted around adecade ago, over 70% of the contemporaryCBM literature is composed of publicationsthat have appeared only within the past threeyears. We provide an overview of this rapidly

Anxiety vulnerability:a continuouslydistributed individualdifference variablereflecting tendency toreadily experienceanxiety

Cognitive bias:systematic selectivityin informationprocessing thatoperates to favor onetype of informationover another

Cognitive biasmodification (CBM):direct manipulation ofa target cognitive bias,by extended exposureto task contingenciesthat favorpredeterminedpatterns of processingselectivity

Attentional bias:a commonly studiedform of cognitive biasinvolving preferentialattention to oneparticular type ofinformation

Interpretive bias:a commonly studiedform of cognitive biasinvolving the tendencyto preferentiallyresolve ambiguity inone particular way

Anxiety dysfunction:problematic anxietysymptoms that areunwarranted by thesituation and interferewith adaptivefunctioning

Anxiety disorder: aparticular syndrome ofdysfunctional anxietysymptoms matchingdiagnostic criteria forone of several clinicallyrecognized categoriesof anxiety pathology

developing new field of clinical research,focusing particularly on its contribution to theunderstanding and attenuation of dysfunctionalanxiety. Although the great majority of CBMwork published to date has focused on atten-tional bias and interpretive bias, the principlesunderlying the CBM approach can readily beextended to other types of cognitive biases also.

BACKGROUND TO COGNITIVEBIAS MODIFICATION RESEARCH

Cognitive Bias and Anxiety

Cognitive accounts of anxiety dysfunctionattribute both heightened anxiety vulnerabilityand clinical anxiety to maladaptive patternsof selective information processing. Theyhave been motivated by the observation thatpatients with anxiety disorders commonlyreport experiencing distinctly threateningthoughts of a type that plausibly could elicit,sustain, or intensify their anxiety symptoms(e.g., Ghahramanlou-Holloway et al. 2007).However, theoretical models developed toexplain these idiosyncrasies in thought contentcausally attributed them to systematic biasesin low-level cognitive mechanisms not readilyavailable to introspective assessment. In par-ticular, biases in attention and interpretationthat operate to selectively favor the processingof emotionally negative information havecommonly been implicated in these models ofanxiety, and sometimes biases in memory func-tion also have been thought to play a role (cf.Ouimet et al. 2009). Cognitive-experimentalmethodologies that directly assess selectiveinformation processing have confirmed thepresence of such biases both in people sufferingfrom anxiety disorders and in nonclinicalindividuals with an elevated dispositionalvulnerability to experience anxiety symptoms.

Clinically anxious patients reliably displayan attentional bias toward negative informa-tion, which is also sometimes shown by healthyindividuals who reported elevated levels oftrait anxiety (cf. Bar-Haim et al. 2007). Thisattentional bias has been assessed in a variety of

www.annualreviews.org • Cognitive Bias Modification Approaches to Anxiety 191

Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


ways. Some techniques, such as the emotionalStroop task, require participants to ignoreemotionally toned distracting informationwhile performing a central task, and selectiveattention to emotionally negative distractorsis inferred by measuring the degree to whichtheir presence disproportionately disruptscentral task performance (cf. Williams et al.1996). Other approaches involve search tasks,where participants scan arrays of stimuli, andattentional bias is inferred from the relativespeed with which they can locate targets ofdiffering emotional tone (e.g., Olatunji et al.2010). Perhaps the most widely used methodof assessing anxiety-linked attentional bias hasbeen the visual probe task, in which stimuli thatdiffer in emotional tone are briefly exposed on acomputer screen before a small visual probe ap-pears in the locus where one or other emotionalstimuli were exposed (e.g., Koster et al. 2006,MacLeod et al. 2007). Participants must quicklydiscriminate probe identity, and relative speed-ing to do so when probes appear in the locus ofnegative stimuli provides an index of selectiveattention to such information. Such assessmenttechniques have repeatedly demonstratedattentional bias to negative stimuli in bothclinical and nonclinical manifestations of dys-functional anxiety (cf. Cisler & Koster 2010).Attentional bias to disorder-relevant informa-tion has sometimes been observed in otherconditions, such as depression (e.g., Baert et al.2010).

Interpretive bias, reflecting selective im-position of negative meanings on ambiguity,also has proven characteristic of clinical andsubclinical anxiety dysfunction (cf. Mathews2011). A common assessment approach hasbeen to examine the impact of initial ambiguousinformation on the processing of subsequenttarget information differentially related to al-ternative meanings of the preceding ambiguity.For example, participants have been exposed tonarrative descriptions of ambiguous scenariosand required after each to make a simple judg-ment about a final target word, such as whetherit is grammatically or lexically legitimate. Aninterpretive bias favoring negative resolutions

of the initial ambiguity is revealed by a process-ing advantage for targets consistent with thisparticular meaning (Hirsch & Mathews 1997).Another technique used to assess anxiety-linkedinterpretive bias involves having participantsread descriptions of ambiguous scenarios andthen giving them a recognition memory testthat presents disambiguated versions of thesescenarios so that their interpretations of theinitial ambiguity can be inferred from the sub-jective familiarity of the alternative disambigua-tions encountered in the memory test (Eysencket al. 1991). Using such assessment procedures,it has been clearly demonstrated that partic-ipants with elevated anxiety vulnerability orsuffering from clinical anxiety are dispro-portionately inclined to interpret ambiguityin a negative manner (cf. Richards 2004).Interpretive bias operating to selectivelyresolve ambiguity in a negative manner alsois associated with depressive disposition (e.g.,Ree et al. 2006).

Evidence of an anxiety-linked memory biashas been more mixed (cf. MacLeod & Mathews2004). Such a bias has sometimes been ob-served both in clinical anxiety patients andin nonclinical participants with elevated traitanxiety (e.g., Ghassemzadeh & Baraheni et al.2003), though it is a more robust characteristicof clinical and subclinical depression (cf.Mathews & MacLeod 2005). Anxiety vulner-ability and dysfunction also is associated withcertain patterns of appraisal bias that plausiblymay contribute to anxious symptomatology.For example, the biased appraisal of anxietysymptoms themselves, resulting in their at-tribution to sinister causes and imbuing themwith the capacity to cause harm, is the keycharacteristic of elevated anxiety sensitivity(Reiss et al. 1986), a disposition predictiveof anxiety pathology (Taylor et al. 1992).Similarly, biased appraisal of intrusive negativethoughts, resulting in the assumption ofpersonal responsibility for these mental events,is a characteristic of obsessive-compulsive dis-order, which may contribute to the heightenedcapacity to elicit anxiety in individuals withthis condition (Salkovskis & Forrester 2002).


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


The compelling evidence that dysfunctionalanxiety is characterized by these patterns ofselective information processing lends plausi-bility to theoretical accounts that implicate suchcognitive bias in the etiology of anxiety disor-ders and in the origin of anxiety vulnerability.Nevertheless, these findings cannot serve to de-termine the causal status of cognitive bias in thisobserved association.

The Purpose of Cognitive BiasModification Research

CBM research has not been motivated by a sin-gle purpose. Rather, the following three re-lated objectives have been pursued throughthe development and application of CBMmethodologies:

1. To determine the causal status of cog-nitive bias: Demonstrating an associationbetween a particular cognitive bias and aheightened disposition to experience anx-iety does not permit the conclusion thatthe bias causally contributes to this dispo-sition. A powerful way of determining ifone variable causally influences another isto test whether the direct manipulation ofthe first serves to alter the second. Hence,an early impetus for the development ofCBM methodologies was to enable thedirect manipulation of such biases to testthe veracity of those theoretical modelsof anxiety that attribute causal status tothem.

2. To evaluate the therapeutic potential ofdirect bias modification: Early CBM re-search demonstrating that CBM couldtemporarily alter anxiety vulnerabilitygenerated interest in the possibility thatthese methodologies may have practicalapplication in the therapeutic attenuationof anxiety dysfunction. This has led to re-search designed to evaluate the capacity ofCBM procedures to alleviate problematicanxiety symptoms and to ameliorate anx-iety responses to stressful environments.

3. To illuminate the nature of cognitive biasmechanisms: Developing the capacity

CBT: cognitivebehavior therapy

to manipulate a process brings withit the opportunity to learn about themechanisms underpinning that process.Just as the development and refinementof conditioning techniques to modifybehavior shed much light on the fun-damental learning mechanisms thatgovern behavioral variability, so toohas the development and evaluation ofCBM techniques enabled researchers toilluminate the fundamental information-processing mechanisms that governanxiety-linked patterns of cognitive bias.

The pursuit of these three objectives hasbeen closely intertwined. CBM work designedto advance understanding of causality hasshaped therapeutic applications of CBM deliv-ered to clinically anxious participants, while theoutcomes of this latter work in turn have shedlight on the causal contributions of selective in-formation processing to anxiety disorders. Bothlines of research have served to increase un-derstanding of the mechanisms that underpinanxiety-linked cognitive bias.

The Nature of Cognitive BiasModification Methodologies

The idea that dysfunctional patterns of think-ing may contribute to anxiety pathology hasdriven the development of cognitive behaviortherapy (CBT) for clinical anxiety disorders(cf. Clark & Beck 2010). Conventional CBTinterventions typically aim to (a) provide pa-tients with insight into the roles their thoughtsplay in the generation and maintenance oftheir anxiety symptoms, (b) assist them in iden-tifying unhelpful thoughts that trigger suchsymptoms, and (c) encourage and enable themto challenge these thoughts in ways that reducetheir credibility and attenuate their emotionalinfluence. In contrast, CBM is not designed toalter the manner in which individuals respondto anxiogenic thoughts but rather to directlychange the cognitive processes that giverise to such thinking. Koster and colleagues(2009) identify two key features of such CBMmethodologies. First, each CBM procedure


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


CBM-A: cognitivebias modificationtargeting selectiveattention

is designed to directly alter one specific low-level bias in selective information processing,theoretically implicated in the generation ofdysfunctional anxiety and usually assumed tooperate prior to conscious thought. Second,CBM does not rely on insight, as the targetedbiases need not be introspectively accessible.Rather, it seeks to modify the target biasthrough extended practice on a task configuredto induce such change. In the majority of cases,this is a reconfigured variant of a cognitive-experimental task that previously has beenemployed to assess this specific cognitive biasand that has proven capable of distinguishingparticipants who differ in terms of anxiety vul-nerability or dysfunction. The reconfigurationinvolves introducing a training contingencyinto the task, such that ease of task performancewill be enhanced by acquisition of the intendedbias change. Generally, participants are notinformed of this training contingency, andthey usually are unable to subsequently reportit. Hence, as Beard (2011) observes, thoughcompletion of a CBM task may be volitional,neither the cognitive bias targeted by thisprocedure nor the process through which biaschange is induced is assumed to be undervolitional control.

The precise nature of the CBM dependsupon the particular type of bias that it is in-tended to change. In the following sections, weseparately review the development and appli-cation of CBM procedures designed to modifyattentional and interpretive bias. We also con-sider how the CBM approach is being extendedto target some other types of clinically relevantprocessing selectivity.

COGNITIVE BIASMODIFICATION TARGETINGATTENTIONAL SELECTIVITY

Development of Attentional BiasModification Techniques

Researchers began developing cognitive biasmodification procedures to change attentionalselectivity (CBM-A) around the mid-1990s

(e.g., MacLeod 1995), and Mathews &MacLeod (2002) provide an early review ofthis work. The CBM-A approach that has beenmost frequently employed across recent yearsrepresents a training version of the attentionalprobe task previously used to assess anxiety-linked attentional bias (MacLeod et al. 1986).In the assessment version of this task, probes arepresented equally often in the screen locationswhere either the negative or neutral member ofa stimulus pair just appeared. However, in thebias modification version of the task, the probesalways appear only in the locus of the negativestimuli (attend-negative training) or the neutralstimuli (avoid-negative training). Whetherthe emotional stimuli are words or images,Mathews & MacLeod (2002) report thatextended exposure to these alternative train-ing conditions serves to induce differentialattentional responding to negative infor-mation. For example, in two studies usingword stimuli, MacLeod and colleagues (2002)exposed participants to 576 trials of thisCBM-A task in either of the training con-ditions. When attentional selectivity wassubsequently measured using new word stimuliand the conventional assessment version of theprobe task, participants given the alternativetraining conditions were found to differ inattentional bias. Those who had completedattend-negative training showed a relativespeeding to probes in the locus of negativewords, indicating attentional vigilance fornegative stimuli. In contrast, participants whohad completed avoid-negative training showeddisproportionate slowing to probes in thelocus of negative words, indicating attentionalavoidance of such stimuli.

The majority of attentional bias modifica-tion studies to date have employed variantsof this probe CBM-A approach, and its capacityto modify attentional selectivity is now well es-tablished (cf. Hakamata et al. 2010). However,other approaches also have been developed.For example, Dandeneau & Baldwin (2004)developed a visual search CBM-A procedure,designed to suppress attention to negativestimuli while developing attentional vigilance


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


for positive stimuli, that required participantsto search for a single positive stimulus in a ma-trix otherwise comprising negative distractorstimuli. Compared to a control condition, thishas been found to reduce attentional vigilancefor negative stimuli as assessed using eitherthe emotional Stroop task (Dandeneau &Baldwin 2004, Dandeneau et al. 2007) or theattentional probe task (Dandeneau & Baldwin2009, Dandeneau et al. 2007).

Single-Session Applications ofAttentional Bias Modification

As we have discussed, attentional bias to nega-tive information is associated with heightenedanxiety vulnerability. CBM-A techniques havebeen used within single-session laboratorystudies to evaluate the hypothesis that such at-tentional selectivity causally contributes to thisdisposition by testing whether its modificationaffects readiness to experience anxiety symp-toms. For example, after having successfullyinduced differential attentional response tonegative information in two groups of mid-traitanxious participants using the probe CBM-Aapproach, MacLeod et al. (2002) exposed themto a stressful anagram task. The degree to whichthis stressor elicited anxiety depended on CBM-A condition. Relative to participants exposedto the attend-negative condition, those who re-ceived the avoid-negative condition displayedattenuated anxiety responses to the anagramstressor. Moreover, participants who developedthe most pronounced attention avoidance ofnegative information in response to the CBM-Amanipulation came to display greatest atten-uation of emotional reactivity to the stressor.Eldar and colleagues (2008) obtained similarfindings using a pictorial version of this probeCBM-A procedure in unselected 7- to 12-year-olds. Children exposed to the avoid-negativeCBM-A condition subsequently showed lessattention to negative information than did chil-dren exposed to the attend-negative condition.Most importantly, in response to a subsequentpuzzle task stressor, the latter children reporteda robust elevation of anxiety while the former

children reported no significant elevation ofanxiety. Independent raters confirmed that thechildren who received avoid-negative CBM-Aalso displayed fewer behavioral signs of anxietyduring the problem task.

Modification of attentional bias using thevisual search variant of CBM-A also has beenfound to influence emotional vulnerability.Dandeneau & Baldwin (2009) gave unselectedparticipants from an adult education centereither a single session of this CBM-A task,configured to induce attentional avoidanceof frowning faces, or a control task with noattentional training contingency. The formercondition served to reduce attentional bias tosocial rejection information, as revealed by asubsequent probe assessment procedure. It alsoserved to attenuate the feelings of rejection laterelicited by a simulated social interaction. Thisbeneficial effect of the CBM-A training wasespecially evident for participants who initiallyscored low on measures of self-esteem. Findingsof this type lend clear support to the hypothesisthat biased attentional response to negativeinformation can make a causal contributionto emotional vulnerability. They also suggestthe clinically important possibility that peoplewith an elevated disposition to experience dys-functional symptoms may potentially benefit inpractical ways from CBM procedures. Researchexamining the impact of CBM-A in partici-pants chosen because they exhibit problematicanxiety symptoms has further supported the po-tential therapeutic value of CBM-A approacheswhile lending weight to the hypothesis thatattentional bias causally contributes to anxietydysfunction.

Amir et al. (2008) exposed participants whoreported difficulty with public speaking to asingle session of pictorial probe CBM-A, eitherin the avoid-negative condition or in a controlcondition containing no attentional trainingcontingency. As intended, those in the formercondition came to display greater attentionalavoidance of negative information compared toparticipants in the control condition. Of mostimportance, they also reported lower levelsof state anxiety and were judged by raters to


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


exhibit less behavioral evidence of anxiety whensubsequently delivering a short speech. Theimpact of the CBM-A manipulation on bothmeasures of anxiety was statistically mediatedby its effect on attentional bias. Reductionof attentional bias to negative informationalso reduces negative thought intrusions inpeople who show an excessive tendency toworry. Hayes and colleagues (2010) assignedparticipants scoring above 56 on the Penn StateWorry Questionnaire (Meyer et al. 1990) toeither a control condition or to CBM-A config-ured to elicit attentional avoidance of negativeverbal stimuli. In addition to inducing suchchange in attentional bias, this CBM-A alsoattenuated negative thought intrusions duringa subsequent worry-induction procedure.Hirsch and colleagues (2011) contrasted theimpact of two variants of CBM-A on negativethought intrusions during a subsequent worry-induction task. One variant was designed toinhibit attentional engagement with negativeinformation by reducing the degree to whichparticipants selectively moved attention towardnegative information presented outside initialattentional focus. The other variant was de-signed to facilitate attentional disengagementfrom negative information by increasing thedegree to which participants moved attentionaway from negative information presentedwithin initial attentional focus. Hirsch et al.(2011) found that the former CBM-A pro-cedure was more effective than the latterin attenuating subsequent negative thoughtintrusions. This led them to conclude thatbiased attentional engagement with negativeinformation may make the greater causalcontribution to this type of anxiety symptom.

Najmi & Amir (2010) have reported ben-eficial effects of CBM-A in people displayingsubclinical obsessive-compulsive symptoms.When given a single session of verbal probeCBM-A in the avoid-negative training condi-tion, these participants came to show reducedattention to contamination-related informa-tion compared to participants exposed toa control procedure with no training con-tingency. Furthermore, they subsequently

demonstrated heightened ability to performa behavioral approach task (BAT) involvingexposure to feared contaminants. Their im-proved BAT performance was mediated by theCBM-A-induced change in attentional bias. Ina recent extension of Eldar et al.’s (2008) study,Bar-Haim et al. (2011) investigated whether thebenefits of CBM-A would extend to a sampleof children selected on the basis of exhibitingchronically high anxiety levels on the Screenfor Child Anxiety Related Emotional Disorders(Birmaher et al. 1999). These dispositionallyanxious children did indeed show attenuatedanxiety reactivity to a puzzle stressor follow-ing exposure to an avoid-negative CBM-Aprocedure.

Such findings indicate that the causal influ-ence of attentional bias extends to dysfunctionalmanifestations of anxiety. Of course, it wouldbe imprudent to conclude from this that atten-tional bias causally contributes to all forms ofanxiety dysfunction. The consistent failure toinfluence particular types of anxiety symptomsthrough the use of CBM-A procedures alsomay be theoretically informative by servingto delineate those facets of anxiety that mayowe little to the influence of attentional bias.Specific fear symptoms have proven particu-larly resistant to the influence of attentionalbias modification. Single-session CBM-Aprocedures, successful in inducing attentionalavoidance of spider-related information, donot attenuate self-report, behavioral, or phys-iological indices of spider fear in spider-fearfulparticipants (Harris & Menzies 1998, Reeseet al. 2010, Van Bockstaele et al. 2011). There isno doubt that selective attentional bias towardspider-related stimuli is a reliable feature of spi-der fear (e.g., Kindt & Brosschot 1997, Mogg& Bradley 2006). Nevertheless, the finding thatits modification does not influence spider fearsymptoms suggests that attentional bias doesnot causally contribute to this condition. It maybe that other forms of selective informationdo play a functional role in specific fear, butReese et al. (2010) suggest that attentionalbias to negative information may contributeonly to the pattern of distressing and repetitive


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


negative thinking that characterizes anxiousrumination and worry, which is less evident inanxiety conditions involving specific fear.

As illustrated by studies of this type, thesingle-session application of CBM-A can pow-erfully contribute to the testing of hypothesesconcerning the causal contributions made byattentional bias to differing facets of anxiety.The encouraging results obtained from single-session CBM-A studies have motivated investi-gators to examine the impact of more extendedCBM delivery, and below we consider the con-tribution made by work that has delivered CBMacross multiple sessions and has assessed its im-pact on anxiety experienced outside the labora-tory setting.

Extended Applications of AttentionalBias Modification

The finding that transient modification of at-tentional bias impacts upon clinically relevantanxiety symptoms in a laboratory setting con-firms that this bias can causally contributeto such symptomatology. It also suggests thatCBM-A may be of potential therapeutic benefitin the alleviation of problematic anxiety. It doesnot, however, permit the conclusion that atten-tional bias makes a meaningful contribution toanxiety in the naturalistic setting. Nor does itmean that CBM-A can be delivered in a man-ner that produces meaningful change in nat-urally occurring anxiety symptoms within thereal world. To address these important issues,researchers have sought to evaluate whether ex-tended exposure to CBM-A can induce endur-ing attentional change that affects real-worldexperience. The findings generally support thehypotheses that attentional bias does causallycontribute to anxiety symptomatology beyondthe laboratory context and testify to the likelyvalue of CBM-A in the treatment of anxietydysfunction.

See and colleagues (2009) delivered anonline version of the probe CBM-A procedureto Singaporean high school graduates on a dailybasis for two weeks prior to their emigrationto commence tertiary education overseas.

GAD: generalizedanxiety disorder

Half received the avoid-negative CBM-Acondition while half were exposed to a controlcondition with no training contingency. Theformer participants alone developed a robustattentional bias away from negative stimuliacross the training period. State anxiety scoresrecorded immediately following the transitionevent were significantly attenuated in theseparticipants relative to those in the control con-dition, and trait anxiety scores declined acrossthe period of the study only for participantsgiven the avoid-negative CBM-A training. Theimpact of the CBM-A manipulation on anxietywas mediated by its effect on attentional bias.Extended delivery of the visual search CBM-Aalso has been found to influence responsesto a stressful work environment. Dandeneauet al. (2007) had telemarketers complete eitherthe avoid-negative or control version of thisCBM-A procedure every day for one week. Par-ticipants in the former condition alone reportedincreased self-esteem and reduced perceivedstress. They also displayed lower cortisol re-lease and cortisol reactivity than participants inthe control group, confirming the attenuationof their stress response. Such findings indicatethat attentional bias does have a causal role inshaping the emotional reactions to situationalstress experienced in real-world settings.

Extended CBM-A can also influencesymptoms associated with anxiety dysfunction.Hazen et al. (2009) delivered five daily sessionsof probe CBM-A to a sample of extreme wor-riers, in either the avoid-negative or controlcondition. The former condition alone servedto induce attentional avoidance of negativeinformation and to significantly attenuate neg-ative emotional symptoms. Pathological worryis the hallmark of generalized anxiety disorder(GAD), and Amir and colleagues (2009a)have demonstrated that the symptoms of thisanxiety disorder also can be improved throughthe use of CBM-A. Patients with GAD weregiven eight sessions of probe CBM-A acrossa four-week period, again delivered eitherin the avoid-negative or control condition.Those in the former condition alone evidencedsignificant reduction of worry and anxiety, and


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


GSD: generalizedsocial anxiety disorder

CBM-I: cognitivebias modificationtargeting selectiveinterpretation

such symptom improvement was mediated bythe reduction of attention to negative informa-tion. Only 50% of the participants receivingavoid-negative CBM-A still met diagnosticcriteria for GAD at the end of the four-weekintervention, compared to 87% of the controlgroup. Attentional bias modification also hasbeen found to influence the symptoms ofgeneralized social anxiety disorder (GSD).Using the same schedule employed by Amiret al. (2009a), Schmidt and colleagues (2009)exposed GSD patients to a variant of this probeCBM-A procedure that modified attentionalresponse to faces displaying critical expres-sions. Unlike those in the control condition,participants in the avoid-negative conditionshowed significant reductions in anxiety anddepression symptoms, maintained at fourmonths. Only 38% of these participantscontinued to meet diagnostic criteria for GSDat the end of the CBM-A program compared to89% of the control group. Similarly impressivefindings have been reported by Amir et al.(2009b), who delivered this same CBM-A pro-cedure to individuals with generalized socialphobia. Unlike control participants, those whoreceived avoid-negative training evidenced asignificant reduction of clinical symptoms, fullymaintained at four-month follow-up, and thissymptom improvement was mediated by theCBM-A-induced change in attentional bias.Emotional dysfunction in children and youthsalso may be responsive to this four-week CBM-A program delivered in the avoid-negativecondition. Rozenman et al. (2011) reportedthat 10- to 17-year-old participants sufferingfrom separation anxiety disorder, social phobia,or GAD responded to this with clinically sig-nificant reductions of anxiety symptoms, andonly 25% of these participants met diagnosticcriteria after the CBM-A intervention.

Findings from the studies reviewed in thissection highlight the capacity of CBM-A toproduce clinically relevant symptom change inparticipants suffering from dysfunctional anxi-ety and give grounds for optimism concerningthe future therapeutic potential of extendedCBM-A procedures in the treatment of anxiety

disorders. Of equal importance, these samefindings provide compelling evidence thatbiased patterns of attentional selectivity docausally contribute to the clinical symptoms ofanxiety pathology.

COGNITIVE BIASMODIFICATION TARGETINGINTERPRETIVE SELECTIVITY

Development of Interpretive BiasModification Techniques

In cognitive bias modification procedures thattarget interpretive bias (CBM-I), each trial firstpresents ambiguous information, after whichthe participant must make a decision that shouldbe facilitated by one or other interpretation ofthis ambiguity. Required decisions are struc-tured such that these consistently benefit fromone particular pattern of selective interpreta-tion, in the expectation that participants willcome to favor this interpretive style. In the firstreported CBM-I procedure, Grey & Mathews(2000) presented participants first with a ho-mograph that permitted a negative and a morepositive interpretation, such as “growth,” whichcan be interpreted negatively to mean a bodilylump caused by disease or more innocuously tomean a general increase in size or importance.The homograph was followed on every trialby a word fragment that participants had toquickly complete. Fragment completion alwaysyielded a word related to a meaning of the ini-tial homograph, which consequently could bea useful aid to such completion. In the CBM-Icondition designed to induce negative inter-pretive bias (interpret-negative), the solution tothe fragment was always a word related to thenegative meaning of the homograph. Thus, inthis condition the homograph “growth” couldbe followed by the fragment C-NC-R, whichyields the completion CANCER. In the otherCBM-I condition, designed to induce morepositive interpretive bias (interpret-positive),the solution to the fragment was always aword related to the homograph’s more positivemeaning. For example, in this condition the


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


homograph “growth” could be followed by thefragment GR-AT-R, which yields the comple-tion GREATER. Following up to 240 CBM-Itrials, the induced pattern of interpretiveselectivity was assessed by measuring relativelatency to process target words related to eithermeaning of a preceding homograph. Acrosstwo studies, performance on these assessmenttrials confirmed that participants exposedto the interpret-positive CBM-I conditioncame to show more benign interpretationsof ambiguity than did those exposed to theinterpret-negative CBM-I condition. Subse-quent studies have confirmed that this CBM-Iprocedure modifies the selective interpretationof ambiguity (e.g., Grey & Mathews 2009).

Mathews & Mackintosh (2000) created aCBM-I variant in which each trial begins withthe textual description of an ambiguous situ-ation, and participants must complete a finalword fragment to provide a meaningful ending.In the interpret-negative CBM-I conditions,final fragments can yield only completionsconsistent with the negative interpretationof the preceding ambiguity, whereas in theinterpret-positive condition, they can yieldonly completions consistent with the positiveinterpretation of this ambiguity. Following 100or so trials of this CBM-I procedure, Mathews& Mackintosh assessed interpretations of newambiguous scenarios, using Eysenck et al.’s(1991) recognition memory procedure. Acrossa series of five studies, participants’ familiarityratings for disambiguated versions of these testscenarios confirmed that they had acquired apattern of interpretive bias consistent with thedirection of CBM-I training. The capacity ofthis CBM-I approach to reliably modify inter-pretive bias has been confirmed in subsequentwork (e.g., Salemink et al. 2009).

Single-Session Applications ofInterpretive Bias Modification

Mathews & Mackintosh (2000) found that par-ticipants who completed a single session oftheir CBM-I task in the interpret-positive train-ing condition subsequently reported lower state

anxiety levels than those who completed thetask in the interpret-negative condition. Sub-sequent research confirmed and extended thisfinding, showing that a session of such CBM-I delivered in the interpret-positive condition,rather than the interpret-negative condition,led to significant decline not only in state anxi-ety but in trait anxiety questionnaire scores also(Salemink et al. 2007a, 2009). Salemink et al.(2009) demonstrated that the effect exerted bythe CBM-I procedure on trait anxiety was me-diated by the induced change in interpretivebias. As these investigators conclude, this sug-gests that interpretative bias makes a causal con-tribution to anxiety vulnerability.

To exclude the possibility that CBM-I-induced change in questionnaire measures oftrait anxiety might reflect only the biased inter-pretation of past emotional experience ratherthan genuine change in anxiety vulnerability,researchers have examined whether CBM-I caninfluence subsequently observed emotional re-activity. Salemink et al. (2007b) found no dif-ference in the degree to which a later anagramstressor served to elevate anxiety in participantspreviously exposed to the alternative CBM-Iconditions. However, it should be noted thatthis was the single study in which these inves-tigators also failed to find an impact of CBM-I condition on their questionnaire measure oftrait anxiety, rendering conclusions difficult.Salemink et al. (2007b) also raise the possibil-ity that their anagram stressor may have in-volved insufficient ambiguity for differences ininterpretive bias to influence anxiety responses.Consistent with this possibility, better evidencefor the causal involvement of interpretive biasin anxiety vulnerability has been obtained usinga stressor that more clearly invites alternativeemotional interpretation. After delivering a sin-gle session of Grey & Mathews’ (2000) CBM-Itask to mid-trait anxious students, Wilson andcolleagues (2006) exposed them to brief videoclips of real-life emergency situations in whichthe victim of a near disaster was injured butultimately rescued. Participants who had justcompleted a CBM-I session in the interpret-negative condition demonstrated a pronounced


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


elevation of both state anxiety and depressionin response to these video clips, whereas theclips did not elevate either anxiety or depres-sion for participants who had completed CBM-I in the interpret-positive condition. CBM-Iprocedures also have proven effective in mod-ifying interpretive bias in children and youths(Lothmann et al. 2011). When Muris and col-leagues exposed schoolchildren aged 8–13 toa single session of CBM-I in the interpret-positive condition, these children judged sub-sequently presented ambiguous scenarios to beless threatening than did children who insteadhad been exposed to the interpret-negativeCBM-I condition (Muris et al. 2008, 2009).

The results of these studies, carried outusing unselected participant samples, clearlysupport the idea that interpretive bias causallycontributes to variation in anxiety vulnera-bility, underpinning differential tendencies toexperience elevated anxiety in response to situ-ations that can be interpreted in different ways.However, this need not mean that interpretivebias also is causally implicated in the types ofanomalous experience associated with anxietydysfunction. Support for such a conclusionwould be strengthened by the demonstra-tion that CBM-I can attenuate pre-existingsymptoms of this type. Several investigatorshave reported such evidence. Murphy et al.(2007) selected participants with pre-existinghigh levels of social anxiety and exposed themto a single session of an auditory CBM-Iprocedure based on Mathews & Mackintosh’s(2000) task. Compared to participants givena control condition with no training contin-gency, those given interpret-positive CBM-Itraining subsequently imposed less-negativeinterpretations on ambiguous test scenariosand expressed lower expectations of feelinganxious in future social situations. Steinman& Teachman (2010) selected participants whoscored high on the Anxiety Sensitivity Index(Reiss et al. 1986), a well-established vulner-ability marker for anxiety pathology (Tayloret al. 1992), and gave them a single session ofMathews & Mackintosh’s CBM-I task in eitherthe interpret-positive condition or a control

condition. The former participants, unlike thelatter, came to display more benign interpre-tations of ambiguity, reduced scores on theAnxiety Sensitivity Index, and a trend towardattenuated anxiety responses to a subsequentlydelivered interoceptive exposure challenge.

It is interesting to note that, in keeping withattentional bias modification, CBM-I has notyet been shown to influence spider fear, despitethe fact that spider-fearful individuals do tendto interpret spider-related scenarios in a dispro-portionately negative manner (de Jong & Muris2002). Teachman & Addison (2008) exposedspider-fearful participants to a single session ofMathews & Mackintosh’s (2000) CBM-I proce-dure in alternative conditions respectively de-signed to increase either negative or positive in-terpretations of spider scenarios. Although thissuccessfully induced a group difference in in-terpretive bias, CBM-I condition did not influ-ence the behavioral avoidance or subjective dis-tress elicited by subsequent exposure to a largespider. Although Teachman & Addison (2008)emphasize the need for further research, theirresults suggest that interpretive bias may notcausally contribute to spider fear. Perhaps, asReese et al. (2010) propose may be the case forattentional bias, selective interpretation insteaddrives patterns of distressing negative thinkingmore characteristic of worry than specific fear.

Certainly, the modification of interpretivebias does appear to influence the frequency ofnegative thought intrusions in worry-proneindividuals. A sample of such worriers, selectedon the basis of their high scores on the PennState Worry Questionnaire, was given a singleCBM-I session by Hirsch et al. (2009). Partici-pants who received an interpret-positive CBM-I procedure reported fewer negative thoughtintrusions during a subsequent breathing focustask than did participants who received acontrol procedure. Given that the hallmark ofGAD is an inflated tendency to worry, Hirschet al.’s findings suggest that interpretive biasmay contribute to the symptomatology of thisanxiety disorder. Consistent with this possi-bility, Hayes et al. (2010) have confirmed thatCBM-I can indeed reduce negative thought


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


intrusions in participants who meet diagnosticcriteria for GAD. When performing a breath-ing focus task immediately after a single sessionof the same CBM-I used by Hirsch et al. (2009),GAD patients exposed to the interpret-positivecondition reported significantly lower rates ofnegative thought intrusion than did those whoreceived the control condition.

Thus, studies delivering single sessions ofCBM-I within the laboratory have proven ca-pable of illuminating the causal status of inter-pretive bias. The results generally support thehypothesis that the selective interpretation ofambiguity can contribute to heightened anxi-ety vulnerability and to clinically relevant pat-terns of anxiety symptoms. We now considerfindings from studies that have employed moreextensive CBM-I delivery and examined its im-pact outside the laboratory setting.

Extended Applications of InterpretiveBias Modification

Researchers have only recently begun to in-vestigate whether CBM-I can induce enduringchange in interpretive bias in ways that influ-ence real-world emotional experience. Thiswork has revealed that the interpretive changeinduced by even a single session of CBM-I issurprisingly robust. Salemink & van den Hout(2010) observed that differential interpretivebias, induced by Mathews & Mackintosh’s(2000) CBM-I procedure, was not attenuatedby a mood induction subsequently adminis-tered in the same experimental session. WhenYiend et al. (2005) investigated the temporalpersistence of differential interpretive biasinduced by such CBM-I, they found it toremain evident even after a 24-hour interval,which was the longest delay considered in theirseries of experiments. Mackintosh et al. (2006)further demonstrated that persistence of CBM-I-induced interpretive bias across a 24-hourperiod was unaffected by changing the contextsof the CBM-I training session and the inter-pretive bias assessment session. Mackintoshet al. also were able to show that the inducedgroup difference in interpretative bias observed

24 hours after a single session of CBM-I wasaccompanied by a corresponding group dif-ference in emotional vulnerability at this laterpoint in time, as revealed by participants’ anxi-ety reactions to the video stressor developed byWilson et al. (2006). Therefore, a single sessionof CBM-I can exert a fairly enduring impact onboth interpretive bias and anxiety vulnerability.

Nevertheless, researchers using CBM-I toinvestigate whether selective interpretationcausally contributes to anxiety experience inreal-world settings generally have employedmultiple sessions of CBM-I delivered acrossmore extended periods of time. For example,to determine whether CBM-I could alter hightrait anxious individuals’ emotional symptomsin their natural environment, Mathews et al.(2007) had them complete four CBM-I sessionsacross a two-week period. When assessed oneweek later, participants who received interpret-positive training evidenced reduced negativeinterpretation of ambiguity and reported low-ered trait anxiety scores compared to controlparticipants. The finding that extended CBM-I can reduce trait anxiety in participants witha pre-existing high level of anxiety vulnerabil-ity has proven to be reliable. Salemink et al.(2009) selected high trait anxious participantswho showed a negative interpretive bias andgave them eight consecutive daily sessions ofMathews & Mackintosh’s (2000) CBM-I pro-cedure. Those who received interpret-positiveCBM-I subsequently evidenced more positiveinterpretive bias than did control participants.They also demonstrated a significant decreasein trait anxiety scores and a reduction of scoreson the SCL-90 (Derogatis & Lazarus 1994),which assesses general psychopathology.

Salemink et al. (2009) found no CBM-Iinduced change on a measure of social anxiety,though social anxiety may not have been aprominent symptom in their participants, whowere selected on the basis of elevated traitanxiety alone. When socially anxious partic-ipants have been given extended exposure tointerpret-positive CBM-I training, this hasbeen shown to attenuate their social anxietysymptoms. Beard & Amir (2008) exposed such


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


participants to eight sessions of CBM-I across afour-week period and assessed the impact of thisintervention two or more days later. Partici-pants who had completed the interpret-positiveCBM-I came to display reduced negative in-terpretations of ambiguity, and reduced socialanxiety symptoms on the SPS, compared tocontrol participants. This reduction of so-cial anxiety symptoms was mediated by theobserved change in their interpretive bias.Vassilopoulos et al. (2009) delivered threesessions of the same CBM-I procedure to 10-to 11-year-old socially anxious children acrossa seven-day period. When assessed 3 to 4 dayslater, negative interpretations of ambiguity hadsignificantly declined in the children exposedto the interpret-positive CBM-I relative tocontrol participants. The former children alonealso evidenced a significant decline in theirscores on the Social Anxiety Scale for Children(La Greca & Stone 1993) and expressed areduced expectancy of experiencing anxietyin an anticipated social situation. Again, themagnitude of the CBM-I-induced reductionin negative interpretive bias significantlypredicted the size of the observed reduction insocial anxiety symptoms.

In summary, therefore, the use of extendedCBM-I has served to indicate that interpretivebias can make a causal contribution to dysfunc-tional anxiety symptoms, with the possible ex-ception of specific fear. The findings lend sup-port to cognitive models that implicate biasedinterpretation in the etiology of anxiety pathol-ogy, and they bode well for the possibility thatextended CBM-I may be of future therapeuticvalue in the treatment of anxiety disorders.

COGNITIVE BIASMODIFICATION TARGETINGOTHER TYPES OF PROCESSINGSELECTIVITY

To date, CBM research has mostly focusedon selective attention and interpretation. How-ever, the scope of CBM techniques is beginningto expand as clinical investigators seek ways ofdirectly manipulating other forms of selectivity,

with the dual objectives of testing modelsthat causally implicate these specific types ofcognitive bias and potentially alleviating clinicalsymptomatology through their modification.Although much of this work is still in its infancy,we briefly consider several of these other CBMapproaches to illustrate their diversity and com-municate the likely flavor of things to come.

Modification of Memory

As we noted above, the relationship betweenanxiety and memory bias is presently unclear.Selective memory retrieval, favoring negativeinformation, has sometimes been observed inanxious participants (e.g., Ghassemzadeh et al.2003). However, this is an inconsistent finding(MacLeod & Mathews 2004), and it is unknownwhether this bias in memory makes a contri-bution to anxiety vulnerability or dysfunction.Hence, it would be of value to develop CBMprocedures capable of directly manipulatingselective memory for negative material in orderto test whether the modification of such mem-ory bias exerts an impact on anxiety. Anderson& Green (2001) demonstrated that whenparticipants were repeatedly exposed to cuespreviously associated with target memories,while endeavoring not to think of these targetmemories, then this effort to not think of themcan drive active forgetting of such targets.Joormann and colleagues have adapted andextended this approach to successfully induceforgetting of negative target information( Joormann et al. 2005, 2009). As yet, it isnot known whether successful application ofthis memory modification procedure serves toattenuate dispositional anxiety or dysfunctionalanxiety symptoms. However, this would beexpected if such selective memory bias causallycontributes to these facets of emotion, andso future CBM work on this topic will betheoretically illuminating while also yieldingpotential applied benefits.

Modification of Imagery

Negative mental imagery is a common featureof psychological dysfunction (Hackman &


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


Holmes 2004), which some theorists proposemay make an especially strong contribution toemotional symptoms such as anxiety (Holmes& Mathews 2005). This has led investigatorsto compare the efficacy of CBM-I proceduresdesigned to target either imagery-based pro-cessing or verbally based processing in orderto test the resulting prediction that the formerwill be more potent in influencing such emo-tional experience. For example, Holmes et al.(2006) used a single-session auditory CBM-Iprocedure to increase positive resolutions ofambiguous scenarios but varied whether partic-ipants were instructed to form mental images orverbal representations of these scenarios. Theimagery condition was more effective in induc-ing a benign interpretive bias, and only in thiscondition did the CBM-I procedure serve toattenuate state anxiety. Holmes and colleagues(2009) have replicated this finding while alsoshowing that the imagery variant of CBM-I ledto greater attenuation of negative emotionalresponse to a subsequent mood induction thanwas evident using the verbal variant. Thesefindings are consistent with the premise thatimagery makes a particularly powerful func-tional contribution to emotional experienceand suggest that the clinical benefits of CBMdesigned to attenuate dysfunctional anxietymay be optimized by the use of proceduresthat directly target negative mental imagery.

Modification of Appraisal

Interpretive bias influences how people re-solve the meaning of intrinsically ambiguousinformation. However, even when informationis not itself ambiguous, people still varyin terms of how they appraise it, drawingdiffering inferences about its importance andimplications. Some dimensions of anxietyvulnerability, such as anxiety sensitivity andobsessionality, are characterized by distinctivebiases in such appraisal processes, and theoristshave implicated these idiosyncratic patternsof implicational thinking in the generation ofassociated anxiety symptoms (Reiss et al. 1986,Salkovskis & Forrester 2002). The premise that

dysfunctional emotional experience reflectsmaladaptive appraisal processes has motivatedclinical researchers to seek ways of directlymodifying appraisal styles. In some cases,participants have been exposed to scenariosand explicitly directed to practice appraisingthem in a prespecified manner, and benefitsof such directed appraisal practice have beenreported (e.g., Schartau et al. 2009, Watkinset al. 2009). These procedures involve theintentional practice of an instructed style ofthinking, as in traditional cognitive behaviortherapy, although the use of preconstructedscenarios enables such practice to be deliveredin an intensive and controlled manner. Theydiffer from CBM methodologies in that thetasks themselves contain no intrinsic trainingcontingencies designed to shape acquisition ofthe desired cognitive change.

However, Lang and colleagues (2009) haveadapted the CBM approach to modify selectiveappraisal. Their study tested the hypothesisthat appraising negative intrusive memoriesas signs of weakness and instability serves toincrease their frequency. Lang et al. amendedMathews & Mackintosh’s CBM procedure bypresenting on each trial a sentence communi-cating a possible appraisal of a negative memoryintrusion, the nature of which depended uponthe identity of a word fragment, which partic-ipants were required to quickly complete. Bymanipulating these word fragments such thattheir completion was made easier by the con-sistent adoption of a particular appraisal style,Lang et al. (2009) created appraise-negativeand appraise-positive versions of this CBMprocedure. Thus, for example, participantscould encounter a sentence that began “Havingintrusive memories means that I am coping”and ended with the fragment “b-dly” (yieldingthe completion “badly”) or “we-l” (yielding thecompletion “well”), in the appraise-negativeand appraise-positive versions, respectively.Participants received a single session of thisCBM procedure, in either condition, beforebeing exposed to a distressing film. Across thesubsequent seven days, those who had receivedappraise-positive CBM reported lower levels


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


of negative memory intrusion concerningthe film than did those who had received theappraise-negative CBM. These findings sup-port the hypothesis that the manner in whichnegative memory intrusions are appraisedcausally influences their frequency.

It seems likely that this CBM approachcould be extended to modify patterns of ap-praisal theoretically implicated in anxiety dys-function, such as the negative appraisal of anx-iety symptoms that characterizes heightenedanxiety sensitivity or the biased appraisal of in-trusive negative thoughts that leads individualswith obsessive-compulsive disorder to assumeundue responsibility for them. Extensions ofappraisal bias modification could serve to testhypotheses concerning the causal contributionsmade by such styles of appraisal to these types ofproblematic anxiety while also potentially pro-viding a means of therapeutically attenuatingtheir severity.

COGNITIVE BIASMODIFICATION ANDCOGNITIVE MECHANISMS

CBM research not only provides insight intothe causal role played by different forms ofprocessing selectivity in shaping anxiety vul-nerability and dysfunction but also can extendknowledge of the mechanisms underpinningthese cognitive biases. Developing the capac-ity to modify a cognitive process deepens un-derstanding of mechanisms in two interrelatedways. First, it requires researchers to identifythe mechanisms that underpin the observedchange in the target process. Second, by ob-serving the degree to which the modification ofone particular cognitive process does, or doesnot, exert an impact on other cognitive pro-cesses, it permits researchers to fractionate thecognitive system into its component elements,distinguishing which types of bias plausibly re-sult from selectivity in shared mechanisms andwhich instead appear to reflect selectivity inthe operation of independent mechanisms. Wehere consider several topical questions perti-nent to each of these issues.

Can CBM Findings Be Attributedto Demand Effects?

An elementary issue concerns whether CBMoperates by producing genuine change incognitive bias and symptomatology or whetherobserved findings might instead representdemand effects. We consider demand effectexplanations implausible for at least six relatedreasons. First, the predictions that most CBMstudies are designed to test are not self-evident.Differences between CBM conditions areusually subtle, involving contingencies thatare neither communicated to participants norrelevant to the decisions they are instructedto make. Demand effect accounts require notonly that these contingencies be apprehendedbut also that participants accurately infertheir expected impacts on both cognition andpertinent symptomatology and feel motivatedto simulate these anticipated consequences.Second, even if this were the case, simulatingthe observed cognitive changes often wouldbe extraordinarily difficult, such as requiringparticipants to display speeding of around30 ms in one assessment condition comparedto another (e.g., Eldar et al. 2008, Grey &Mathews 2000). Third, evidence that CBMcan influence the targeted cognitive processesis not restricted to performance on assessmenttasks but is supported by neurocognitivemeasures. For example, exposure to differingCBM-A training conditions modulates activityin neurocognitive systems implicated in at-tentional control (Browning et al. 2009, Eldar& Bar-Haim 2010). Fourth, symptom changeis often highly specific. For example, demandeffects cannot readily accommodate the ob-servation that participants commonly reportno difference in mood state directly followingCBM-A or CBM-I but instead show differencesin emotional reactivity only to a subsequentstressor (MacLeod et al. 2002, Wilson et al.2006). Fifth, CBM-induced symptom changeis observed on psychophysiological indices thatlie beyond intentional control (Dandeneauet al. 2007). Sixth, when participants have beenrequired to report their expectancies, such


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


reports have consistently led researchers toreject the plausibility of demand effect expla-nations of their findings (Hayes et al. 2010, Seeet al. 2009). Of course, although these consider-ations mitigate against demand effect accountsof CBM-A and CBM-I findings, researchersmust continue to guard against their potentialinfluence by employing designs that obscureexperimenter expectations, objectively assessinduced cognitive change, and supplementself-report symptom indices wherever possiblewith behavioral and physiological measures.

Do CBM Effects Reflect Change inthe Intended Cognitive Process?

The fact that CBM procedures induce genuinecognitive change need not mean that they di-rectly modify the intended cognitive processes.Typically, alternative CBM conditions involvedifferential exposure to key categories of in-formation. In most of the described examples,for instance, these conditions plausibly resultin differing exposure to negative information.Might differential exposure be responsible forthe observed effects, without the need to im-plicate CBM-induced changes in interpretiveand attentional bias? We consider two possiblevariants of this account.

The first possibility is that exposure todiscrepant emotional information may in-duce differential mood states, with change inprocessing selectivity reflecting the indirectconsequence of such mood change. However,even when alternative CBM training condi-tions do not directly influence mood state, theystill elicit differential selective interpretationand attention (Hoppit et al. 2010, Wilsonet al. 2006). Also, mood change has beenstatistically excluded as the source of observedcognitive change (Amir et al. 2008, Hirschet al. 2007), and the experimental inductionof differential mood state does not mimic theimpact of CBM (Standage et al. 2010). Cog-nitive change elicited by CBM is unaffected bymood-induction procedures (Salemink et al.2010) and can remain evident long beyond thelikely duration of transient mood states (Hazen

et al. 2009, Mackintosh et al. 2006). Hence it isunlikely that the bias change elicited by CBMis mediated by the impact of CBM procedureson mood (Mathews 2011).

An alternative possibility is that CBM maysemantically prime those categories of infor-mation that participants are exposed to in waysthat influence performance on bias assessmenttasks. Particularly with respect to CBM-I pro-cedures, such priming effects could in principleemulate the intended cognitive change withoutthis change actually occurring. Specifically,participants completing CBM-I proceduresthat repeatedly expose them to either positiveor negative target information related toalternative meanings of prior ambiguity mightbecome better at processing this class of targetinformation as a consequence of semanticpriming without this reflecting changed in-terpretation of ambiguity. However, CBM-Iinduced change in the relative processing speedof discrepantly valenced target information isobserved only when this information is differ-entially related to preceding ambiguity (Grey& Mathews 2009, Hoppitt et al. 2010, Wilsonet al. 2010) and shows far greater temporalendurance than semantic priming effects.

The cognitive change resulting from CBMmost likely reflects transfer of practiced pro-cessing selectivity from the training task tonew situations that invoke this same cognitiveprocess (Hertel & Mathews 2011). Transferof training depends upon there being a closematch between the cognitive processes impli-cated in both the training and transfer task(Blaxton 1989). Therefore, as we discuss be-low in more detail, examining transfer of train-ing effects following CBM may shed light onthe degree to which overlapping processes con-tribute to differing manifestations of cognitivebias.

Is CBM Change Restricted to theTargeted Cognitive Bias?

There is abundant evidence that CBM trainingcommonly transfers to new stimulus materialspresented in assessment versions of the bias


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


modification task. Such transfer is reassuring,but unsurprising, as the same cognitive processclearly operates in both the training andtransfer task. This is an example of what Hertel& Mathews (2011) refer to as “near transfer.”They use the term “far transfer” when CBMtraining influences emotional reactions to un-related stress tasks or clinical symptomatology.Far transfer effects reveal that the same cogni-tive process modified within the CBM trainingsession operates within the assessment settingto influence the dependent measures of interest.

One way in which CBM transfer effects canbe informative about component cognitive pro-cesses is when supposedly near transfer fails tooccur. The implication is that, despite super-ficial similarities, the cognitive process mod-ified in the training task must be indepen-dent of those operating in the test task. Thereare many examples of successful near trans-fer across different tasks intended to measurethe same cognitive process (e.g., Dandeneau &Baldwin 2004, Grey & Mathews 2000, Wilsonet al. 2006). However, there have also been in-teresting and potentially important failures ofseemingly near transfer. For example, Saleminket al. (2010) found that the effects of CBM-I training transferred to a quite different task,implicitly revealing the interpretations imposedon ambiguity. However, it did not influenceperformance on assessment tasks that requiredparticipants to explicitly report their preferredinterpretations of ambiguous scenarios. Thissuggests that the cognitive processes govern-ing self-reported beliefs about interpretive stylemay be unrelated to the cognitive processesthat govern the resolution of ambiguity. Futuremapping of near transfer failures may permitdelineation of the boundaries between appar-ently similar patterns of selectivity that reflectfundamentally different cognitive mechanisms.

Another way in which CBM transfer effectscan illuminate the cognitive mechanisms iswhen unexpected far transfer occurs betweentasks previously thought to reflect differentcognitive processes. Recently observed fartransfer effects of this nature suggest thatboundaries traditionally imposed by theorists

between attentional, interpretive, and memorybias may need to be reconsidered. For example,it recently has been shown that the effect ofprobe CBM-A training designed to directlymodify attentional bias can transfer to influ-ence interpretive bias also (White et al. 2011).Conversely, the effect of CBM-I trainingdesigned to directly influence interpretivebias can transfer to modify attentional biasalso (Amir et al. 2010). The modification ofinterpretive bias using CBM-I also has beenshown to transfer to measures of memory bias(Salemink et al. 2010, Tran et al. 2011).

So CBM transfer of training effects suggeststhat at least some facets of attentional, inter-pretive, and memory bias stem from sharedselective processing mechanisms and that sometasks intended to assess a single bias insteadappear to measure independent cognitivemechanisms. The more systematic study ofsuch transfer effects should assist future re-searchers to categorize and fractionate the arrayof cognitive mechanisms that contribute to thediversity of processing biases associated withclinically relevant dimensions of individual dif-ferences. Hence, in addition to their capacity totest causal hypotheses concerning the cognitivebasis of anxiety and to attenuate dysfunctionalanxiety, CBM methodologies also equipinvestigators with powerful new tools capableof illuminating the nature of anxiety-linkedpatterns of processing selectivity.

FUTURE DIRECTIONS FORCOGNITIVE BIASMODIFICATION RESEARCH

Despite its relatively short history, CBMresearch already has shed light on the con-tributions made by selective informationprocessing mechanisms to anxiety vulnerabilityand dysfunction and has given rise to promis-ing new methods of ameliorating problematicmanifestations of anxiety. Nevertheless, as wediscuss below, there is great scope for furtherprogress to extend the reach of this work,to enhance the efficacy of CBM procedures,


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


and to refine therapeutic application of CBMprocedures within clinical interventions.

Extending the Reachof CBM Research

In this review we have focused on anxiety, andto date the CBM approach has most commonlybeen employed to investigate the contributionsof selective information processing to anxiety-related conditions. However, cognitive biaseshave been implicated in a great many otherforms of psychological dysfunction, such as de-pression (Baert et al. 2010), addiction (Wierset al. 2007), eating disorders (Williamson et al.1999), chronic pain (Eccleston & Crombez1999), aggression (Bond et al. 2004), insomnia(Ree et al. 2006), and even schizophrenia (Beck& Rector 2005). Broadening the application ofCBM research to such conditions would serveto determine which particular biases function-ally contribute to which aspects of psychologi-cal dysfunction while also potentially increasingthe therapeutic scope of future CBM interven-tions. Such work already has commenced withpromising early results. Researchers have re-ported finding that CBM procedures can influ-ence depression (Wells & Beevers 2010), addic-tive consumption (Fadardi & Cox 2009), bodydissatisfaction and eating restriction (Smith &Rieger 2009), and pain perception (McGowanet al. 2009). Of course, such extension of CBMis certain to present new challenges. For ex-ample, when Steel and colleagues (2010) en-deavored to attenuate anxiety in people withschizophrenia using the CBM-I procedure thatHolmes et al. (2006, 2009) employed to mod-ify interpretive bias in nonpsychotic samples,the procedure failed to influence interpretivebias in their schizophrenic sample. Steel et al.(2010) suggest this may reflect the atypical pat-terns of imagery evidenced by these partic-ipants. Hence, existing CBM tasks may notprovide an adequate arsenal to modify the bi-ases encountered in all disorders. Instead, thisis likely to require development of new CBMtechniques tailored to ensure their capacity to

alter target biases in participants suffering fromparticular clinical conditions.

Researchers also should expand the breadthof outcome measures used in CBM studies tomove beyond the present heavy reliance onself-report (MacLeod et al. 2009a). Only afew investigators have yet included behavioralmeasures of anxiety when assessing symptomchange (Amir et al. 2008, Najmi & Amir 2010),and fewer still have considered physiological in-dices of such change (Dandeneau et al. 2007,Van Bockstaele et al. 2011). Their lead shouldbe followed in future CBM work. Extendingthe range of symptoms measured will revealthe contributions made by particular types ofcognitive bias to alternative categories of symp-toms. It also should be possible to illuminate thefunctional relationship between these differingfacets of anxiety symptomatology using CBMdesigns that more carefully delineate, and trackover time, the sequential consequences of CBMacross this broader range of symptom measures.

Enhancing the Efficacy of CognitiveBias Modification Procedures

The more effectively CBM procedures canchange target cognitive biases the more valu-able they will be; thus, future work must seek tooptimize their capacity to alter processing selec-tivity. Researchers are beginning to capitalizeon the Internet to increase the ease with whichparticipants can be exposed to CBM proceduresfor extended periods of time (MacLeod et al.2007, See et al. 2009), and more widespreadadoption of this approach could enhance train-ing efficacy. Enabling CBM to be delivered out-side the laboratory also could enhance trans-fer of training in ways that increase the impactof CBM in real-world settings. Such transfer oftraining might benefit from the future use ofmobile audio devices to deliver CBM in ex-ternal settings where computer access is im-practical. The viability of this approach is sup-ported by the recent success of auditory CBM-Ivariants in reducing anxiety symptoms (Holmeset al. 2006, 2009), and Standage et al. (2009) re-port that auditory and visual versions of CBM-I


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


procedures are equally effective in altering cog-nitive bias.

We can also expect to see further refine-ments of CBM tasks themselves, designed toenhance their capacity to alter cognitive bias.In choosing which refinements to evaluate,researchers will be influenced by their assump-tions concerning the mechanisms throughwhich current CBM procedures bring aboutobserved cognitive change. In turn, determin-ing whether their chosen refinements succeedor fail will test the validity of such assumptions.For example, differing views concerningwhether CBM procedures alter cognitive se-lectivity through implicit or explicit processesseem likely to influence the future developmentof these procedures. In most CBM studies,investigators have found participants unableto report the training contingency, consistentwith the idea that this exerts an implicit influ-ence of processing selectivity. However, therehave been occasional exceptions, where CBMhas been found to exert a stronger influence onthose participants who report awareness of thetraining contingency (Field et al. 2007), sug-gesting that CBM-induced change may resultfrom explicit insight into the training contin-gency. These competing conceptions generatediffering expectations concerning whetherexplicitly informing participants of the trainingcontingency will enhance or impair the efficacyof CBM. As noted by Beard (2011), earlyfindings concerning the impact of such instruc-tional manipulation have been inconclusive.Krebs et al. (2010) found that explicitly commu-nicating the training contingency increased theimpact of CBM-A on worry, whereas MacLeodet al. (2009b) have reported that it eliminatessuch emotional impact. Further research isneeded to resolve this issue, and this futurework should strengthen the power of CBMprocedures while also testing the veracity ofcompeting views concerning the involvementof implicit or explicit processes in CBM.

The practical value of CBM also will beincreased by methodological developmentsthat enhance the generalization and temporal

stability of induced bias change. In the greatmajority of CBM studies, induced change inprocessing selectivity has generalized to newinformation not employed in the training itself.However, in a few CBM-A studies such gener-alization to new stimuli has not been observed(e.g., Field et al. 2007, 2009), possibly becauseof the restricted range of training stimuliemployed. There is a need to identify how bestto construct and employ training stimulus setssuch that generalization to new materials isoptimized. One promising approach that war-rants formal evaluation has involved graduallyexpanding the stimulus training set across thecourse of the CBM procedure such that theneed to generalize past learning to new stimulibecomes an integral feature of the trainingexperience (See et al. 2009). Generalization ofCBM training to new contexts is likely to befacilitated by delivering such training acrossmultiple contexts. However, even generalizedchange in dysfunctional cognitive bias willproduce lasting benefits only if changes in biasendure across time. Hence, identifying reliablemethods of maximizing stability of CBMtraining effects represents another importantobjective for researchers. The literature onmassed versus spaced learning (Cepeda et al.2006) suggests that the retention of CBMtraining effects may be enhanced by increasingthe temporal separation of CBM training ses-sions (Hertel & Mathews 2011, See et al. 2009).Infrequent booster sessions, briefly re-exposingparticipants to the original CBM procedure,also may help to preserve CBM-inducedchanges in processing selectivity (MacLeodet al. 2009a). Experimentation designed toincrease the magnitude, generalization, andstability of induced cognitive change willamplify the future value of CBM procedures.

Refining the Therapeutic Applicationof CBM in Clinical Contexts

Small-scale randomized controlled trials havesupported the therapeutic value of both CBM-A and CBM-I in the treatment of GAD and


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


social anxiety disorders (e.g., Amir et al.2009a,b; Schmidt et al. 2009; Vassilopouloset al. 2009). However, we concur with Beard(2011) that the time is now right for large-scaleformal field trials evaluating the clinical efficacyof CBM interventions across the spectrumof anxiety disorders. Such trials should notonly compare the therapeutic efficacy of CBMagainst that of established interventions, in-cluding CBT and pharmacological approaches,but should also evaluate whether CBM canbe profitably combined with these otherapproaches in anxiety management treatmentpackages. MacLeod et al. (2009a) suggest thattherapeutic synergy could result from deliv-ering CBM in conjunction with conventionalCBT if the former serves to alter low-levelpatterns of processing selectivity underpinningthreatening thinking and the latter to chal-lenge the veracity of these negative thoughts.Likewise, CBM may augment self-deliveredexposure treatments by altering selectivecognition in ways that increase the frequencyand duration of therapeutic exposure. On thebasis of the observation that selective serotoninreuptake inhibitors elicit cognitive changeof a type similar to that produced by CBM,Browning et al. (2010) predict that concurrentdelivery of both interventions may prove moreefficacious than either alone. A closely relatedissue concerns whether clinical benefits couldbe optimized by combining different variants ofCBM. Given the evidence that psychologicaldysfunction often is characterized by thesimultaneous operation of both attentional andinterpretive bias, the simultaneous delivery ofCBM-A and CBM-I could produce greatersymptom improvement than their individualadministration. Brosan and colleagues (2011)have confirmed the capacity of combinedCBM-A and CBM-I to reduce symptomseverity in clinically anxious outpatients butdid not compare the therapeutic impact oftheir combined and individual delivery.

Tailored treatments deliver customizedcollections of therapeutic elements chosen tomatch the specific needs of individual patients.

There is scope for tailoring CBM interventionsto target the specific bias profiles shown byindividual recipients. As yet it is not knownwhether CBM-A most effectively modifiesanxiety symptoms in those who already displaydysfunctional attentional bias while CBM-Idoes so most effectively for those who al-ready display dysfunctional interpretive bias.However, it has been shown that people differsystematically in readiness to change theirpatterns of selective processing in responseto CBM contingencies (Clarke et al. 2008).Therefore, future development of individu-alized assessment techniques to appraise theprofile and malleability of target cognitivebiases may assist in identifying the individualswhose anxiety dysfunction is most likely tobenefit from particular CBM interventions.

Ultimately, to fully exploit the therapeuticpotential of CBM techniques in the treatmentof anxiety dysfunction, the techniques willneed to be incorporated into packages thatare acceptable to the end-user. In a recentstudy investigating clinically anxious patients’satisfaction with multisession CBM-A andCBM-I interventions delivered in a primaryhealth care setting, Beard et al. (2010) obtainedreassurance that the procedures were generallyrated as acceptable. Nevertheless, recipientsexpressed a need to better understand how theprocedures were supposed to help tackle theirsymptoms. Hence, future clinical investigatorsmust successfully confront the challenge ofidentifying and communicating a clear ratio-nale to patients that lends credibility to CBMinterventions procedures without diminishingtheir therapeutic efficacy.

CLOSING COMMENTS

The existence of a robust association betweenselective information processing and anxietyhas been firmly established for three decades.Across this same period of time, researchershave worked diligently to better understandthe nature of this association and, in partic-ular, to determine how biased cognition may


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


contribute to both anxiety vulnerability and tothe types of clinical dysfunction observed inthe anxiety disorders. The advent of cognitive-experimental techniques that have provencapable of directly modifying anxiety-linkedcognitive biases represents a highly significantjuncture in this research journey, and the recentadoption of these CBM approaches has broughtthe field to an exciting new threshold, whichfuture clinical researchers may come to viewas an important watershed. As we have shownin this review, CBM methodologies have nowproven their ability to increase scientific under-standing by revealing the causal contributionsmade by specific types of cognitive bias toanxiety symptomatology and to elicit desirablechange by attenuating the severity of anxietysymptoms. Their dual capacity to advanceunderstanding and to increase our ability tobring about change seems likely to assure suchmethodologies of a central role in theoreticaland applied aspects of future clinical research.

Of course, it must be borne in mind thatCBM research is still in its infancy. Thepresent techniques are limited in scope and

are unlikely to represent the most effectivemethods of modifying the cognitive biasesthey target. As these existing CBM proceduresare refined and strengthened, and as newCBM techniques are developed to extendand augment those presently available, so theinfluence of this research approach is likelyto steadily increase. CBM research findingsalready justify the conclusion that CBMmethodologies represent powerful and valu-able scientific tools. We would caution againstthe premature conclusion that CBM representsan effective stand-alone treatment for clinicalanxiety disorders, as clinical field trials ofappropriate scale have yet to be carried out.Almost certainly, such clinical conditions aremultiply determined, making it unlikely thatany one therapeutic approach will prove to bethe proverbial magic bullet. Nevertheless, onthe basis of the evidence to date, it seems highlyprobable that CBM approaches will prove tobe therapeutically useful components of futuretreatment packages designed to alleviate psy-chological disorders that include, but probablywill not be restricted to, clinical anxiety.

SUMMARY POINTS

1. The development of cognitive bias modification procedures in anxiety has been motivatedby the three objectives of (a) determining whether cognitive biases causally contributeto anxiety vulnerability and dysfunction, (b) evaluating the capacity of bias modificationprocedures to contribute in a therapeutically useful way to the attenuation of anxiety,and (c) illuminating the nature of the mechanisms that underpin anxiety-linked cognitivebias.

2. Existing CBM methodologies have proven capable of directly modifying both selectiveattention to emotional information and selective interpretation of emotional ambiguity.The principles underlying these successful CBM approaches for manipulating selectiveattention and interpretation are being extended to develop new CBM variants designedto tackle other forms of cognitive bias, such as memory bias and appraisal bias.

3. Studies delivering single sessions of CBM within experimental settings have demon-strated that the modification of attentional and interpretive bias can influence anxietyresponses to laboratory stressors both in unselected participants and in those selectedon the basis of experiencing dysfunctional anxiety symptoms. This supports the hy-pothesis that these cognitive biases can causally contribute to anxiety vulnerability anddysfunction.


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


4. Studies delivering multiple sessions of CBM across more extended periods have demon-strated modification of attentional and interpretive bias that persists across time and leadsto attenuation of anxiety responses to naturalistic stressors and to amelioration of clini-cal symptoms in individuals with generalized anxiety disorder, generalized social anxietydisorder, and generalized social phobia. This supports the potential therapeutic value ofCBM in real-world settings.

5. By systematically investigating the transfer of CBM-induced change in targeted cogni-tive bias to assessment tasks that measure supposedly similar or distinct manifestationsof processing selectivity, researchers are using CBM methodologies to categorize andfractionate the cognitive operations that underpin the spectrum of processing biasesassociated with anxiety vulnerability and dysfunction.

6. Though the mechanisms through which CBM procedures give rise to changes in cogni-tive bias are not yet fully understood, this cognitive change and its attendant emotionalconsequences do appear genuine, as the findings are resistant to alternative explanation,such as demand-based accounts.

7. Future research should seek to extend the reach of CBM approaches to the investigationand attenuation of other types of psychological symptoms known to be associated withbiased patterns of information processing and to increase the power of CBM proceduresto elicit robust cognitive change that generalizes to new situations and endures acrosstime.

8. On the basis of the promise they have shown, the time is right for large-scale field tri-als designed to formally evaluate the capacity of CBM procedures to contribute to thetreatment of different anxiety disorders within clinical settings. Such trials could prof-itably investigate how CBT can best be combined with existing therapeutic proceduresto optimize clinical outcomes.

FUTURE ISSUES

1. Cognitive bias modification research to date has focused mostly on the capacity of CBMto influence anxiety-related conditions. Future work should continue extending the reachof the CBM approach to illuminate the causal contributions of information-processingselectivity to other clinically relevant conditions characterized by cognitive bias, such asdepression, addiction, eating disorders, chronic pain, and aggression. In addition to ad-vancing understanding of the role played by cognitive bias in these types of dysfunctions,CBM may also prove capable of contributing to their alleviation.

2. Although there have been important exceptions, most CBM studies have relied heavilyon self-report measures of subjective experience when evaluating the impact of inducedbias change. It will be important to broaden symptom measures to more routinely includebehavioral and physiological indices. Not only will this increase confidence in the veracityof CBM-elicited symptom change, but it also may enable CBM investigators to testhypotheses concerning functional relationships between these differing categories ofsymptoms.


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


3. To increase the magnitude of the cognitive change elicited by CBM procedures, fu-ture researchers should systematically evaluate the relative efficacy of CBM variants thatdiffer in the degree to which they exploit the change mechanisms theorists propose tobe implicated in CBM. Among other possibilities, it will be important to determinehow variations in the types of verbal instruction given to participants completing CBMprocedures influence their capacity to alter selective information processing.

4. Future researchers should identify the methods of CBM delivery that maximize thegeneralization of induced changes in cognitive bias. It seems likely that maximizing gen-eralization of bias change to new stimuli will require that the use of large stimulus sets inCBM training procedures, maximizing generalization of bias change to new assessmenttasks, will be assisted by the use of multiple CBM training tasks, and maximizing gener-alization of bias change to new contexts will benefit from delivering the CBM trainingin diverse contexts.

5. The practical benefits of CBM will be greatest when the induced bias change enduresacross time. Further research is needed to identify the methodological procedures thatyield the most stable changes in cognitive bias. The potential benefits of employing spacedlearning procedures to extend retention of training should be systematically evaluated,and investigators also should examine how the use of infrequent top-up sessions can bestcontribute to the maintenance of initial bias change.

6. There is a pressing need for large-scale field trials employing formal protocols to evaluatethe clinical efficacy of CBM in the treatment of the anxiety disorders and other forms ofpsychological dysfunction characterized by patterns of processing selectivity amenableto alteration using available CBM procedures.

7. Treatment trials should not only compare the efficacy of CBM against other establishedinterventions for anxiety, including CBT and pharmacological approaches, but shouldalso investigate whether the inclusion of CBM can augment the efficacy of these existingapproaches. It is likely that CBM will make its most valuable contribution through itsinclusion in multimodal treatment packages, and understanding how best to capitalizeon the potential synergies between CBM and other therapeutic elements will ensure theoptimal structuring of these packages.

8. It will be necessary to develop a framework for CBM delivery that end-users, includingpatients and therapists, find acceptable within the clinical settings. Although the efficacyof CBM may not rely on participant insight, the commitment to persevere with CBMprocedures will depend upon recipients’ appraising such techniques as potentially valu-able. Hence, clinical researchers must identify the rationale that can be communicatedto recipients and that serves most effectively to sustain their motivation to engage withCBM procedures, without compromising therapeutic efficacy.

DISCLOSURE STATEMENT

The authors are unaware of any affiliation, funding, or financial holdings that might be perceivedas affecting the objectivity of this review.


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


ACKNOWLEDGMENTS

The preparation of this paper was supported in part by Australian Research Council grantDP0879589. We thank Patrick Clarke for his invaluable assistance in collating the relevantliterature.

LITERATURE CITED

Amir N, Beard C, Burns M, Bomyea J. 2009a. Attention modification program in individuals with generalizedanxiety disorder. J. Abnorm. Psychol. 118:28–33

Amir N, Beard C, Taylor CT, Klumpp H, Elias J, et al. 2009b. Attention training in individuals with generalizedsocial phobia: a randomized controlled trial. J. Consult. Clin. Psychol. 77:961–73

Amir N, Bomyea J, Beard C. 2010. The effect of single-session interpretation modification on attention biasin socially anxious individuals. J. Anxiety Disord. 24:178–82

Amir N, Weber G, Beard C, Bomyea J, Taylor CT. 2008. The effect of a single-session attention modificationprogram on response to a public-speaking challenge in socially anxious individuals. J. Abnorm. Psychol.117:860–68

Anderson MC, Green C. 2001. Suppressing unwanted memories by executive control. Nature 410:366–69Baert S, De Raedt R, Koster EH. 2010. Depression-related attentional bias: the influence of symptom severity

and symptom specificity. Cogn. Emot. 24:1044–52Bar-Haim Y. 2010. Research review: attention bias modification (ADM): a novel treatment for anxiety disor-

ders. J. Child Psychol. Psychiatry 51:859–70Bar-Haim Y, Lamy D, Pergamin L, Bakermans-Kranenburg MJ, van Ijzendoorn MH. 2007. Threat-related

attentional bias in anxious and nonanxious individuals: a meta-analytic study. Psychol. Bull. 133:1–24Bar-Haim Y, Morag I, Glickman S. 2011. Training anxious children to disengage attention from threat:

a randomized controlled trial. J. Child Psychol. Psychiatry 52:861–69Beard C. 2011. Cognitive bias modification for anxiety: current evidence and future directions. Expert Rev.

Neurother. 11:299–311Beard C, Amir N. 2008. A multi-session interpretation modification program: changes in interpretation and

social anxiety symptoms. Behav. Res. Ther. 46:1135–41Beard C, Weisberg RB, Perry A, Schofield C, Amir N. 2010. Feasibility and acceptability of CBM in primary

care settings. In Cognitive Bias Modification in Anxiety: Targeting Multiple Disorders, Biases, and Settings,C Beard, chair. Symp. presented at World Congr. Behav. Cogn. Ther., Boston, MA

Beck AT, Rector NA. 2005. Cognitive approaches to schizophrenia: theory and therapy. Annu. Rev. Clin.Psychol. 1:577–606

Birmaher B, Brent DA, Chiappetta L, Bridge J, Monga S, Baugher M. 1999. Psychometric properties of theScreen for Child Anxiety Related Emotional Disorders (SCARED): a replication study. J. Am. Acad. ChildAdolesc. Psychiatry 38:1230–36

Blaxton TA. 1989. Investigating dissociations among memory measures: support for a transfer-appropriateprocessing framework. J. Exp. Psychol.: Learn. Mem. Cogn. 15:657–68

Bond AJ, Verheyden SL, Wingrove J, Curran H. 2004. Angry cognitive bias, trait aggression and impulsivityin substance users. Psychopharmacology 171:331–39

Brosan L, Hoppitt L, Shelfer L, Sillence A, Mackintosh B. 2011. Cognitive bias modification for attentionand interpretation reduces trait and state anxiety in anxious patients referred to an out-patient service:results from a pilot study. J. Behav. Ther. Exp. Psychiatry 42:258–64

Brown TA, Barlow DH. 1994. Cognitive models of anxiety. In PsycCRITIQUES, pp. 170–71. Washington,DC: Am. Psychol. Assoc.

Browning M, Holmes EA, Harmer CJ. 2010. The modification of attentional bias to emotional information: areview of the techniques, mechanisms, and relevance to emotional disorders. Cogn. Affect. Behav. Neurosci.10:8–20

Browning M, Holmes EA, Murphy SE, Goodwin GM, Harmer CJ. 2009. Lateral prefrontal cortex mediatesthe cognitive modification of attentional bias. Biol. Psychiatry 67:919–25


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


Cepeda NJ, Pashler H, Vul E, Wixted JT, Rohrer D. 2006. Distributed practice in verbal recall tasks: a reviewand quantitative synthesis. Psychol. Bull. 132:354–80

Cisler JM, Koster EH. 2010. Mechanisms of attentional biases towards threat in anxiety disorders: an inte-grative review. Clin. Psychol. Rev. 30:203–16

Clark DA, Beck AT. 2010. Cognitive Therapy of Anxiety Disorders: Science and Practice. New York: GuilfordClarke P, MacLeod C, Shirazee N. 2008. Prepared for the worst: readiness to acquire threat bias and suscep-

tibility to elevate trait anxiety. Emotion 8:47–57Dandeneau SD, Baldwin MW. 2004. The inhibition of socially rejecting information among people with high

versus low self-esteem: the role of attentional bias and the effects of bias reduction training. J. Soc. Clin.Psychol. 23:584–602

Dandeneau SD, Baldwin MW. 2009. The buffering effects of rejection-inhibiting attentional training on socialand performance threat among adult students. Contemp. Educ. Psychol. 34:42–50

Dandeneau SD, Baldwin MW, Baccus JR, Sakellaropoulo M, Pruessner JC. 2007. Cutting stress off at thepass: reducing vigilance and responsiveness to social threat by manipulating attention. J. Personal. Soc.Psychol. 93:651–66

de Jong PJ, Muris P. 2002. Spider phobia: interaction of disgust and perceived likelihood of involuntaryphysical contact. J. Anx. Disord. 16:51–65

Derogatis LR, Lazarus L. 1994. SCL-90–R, Brief Symptom Inventory, and matching clinical rating scales.In The Use of Psychological Testing for Treatment Planning and Outcome Assessment, ed. LR Derogatis,KL Savitz, ME Maruish, pp. 217–48. Hillsdale, NJ: Erlbaum

Eccleston C, Crombez G. 1999. Pain demands attention: a cognitive-affective model of the interruptivefunction of pain. Psychol. Bull. 125:356–66

Eldar S, Ricon T, Bar-Haim Y. 2008. Plasticity in attention: implications for stress response in children. Behav.Res. Ther. 46:450–61

Eysenck MW, Mogg K, May J, Richards A, Mathews A. 1991. Bias in interpretation of ambiguous sentencesrelated to threat in anxiety. J. Abnorm. Psychol. 100:144–50

Fadardi JS, Cox WM. 2009. Reversing the sequence: reducing alcohol consumption by overcoming alcoholattentional bias. Drug Alcohol. Depend. 101:137–45

Field M, Duka T, Eastwood B, Child R, Santarcangelo M, Gayton M. 2007. Experimental manipulation ofattentional biases in heavy drinkers: Do the effects generalise? Psychopharmacology 192:593–608

Field M, Duka T, Tyler E, Schoenmakers T. 2009. Attentional bias modification in tobacco smokers. NicotineTob. Res. 11:812–22

Ghahramanlou-Holloway M, Wenzel A, Lou K, Beck AT. 2007. Differentiating cognitive content betweendepressed and anxious outpatients. Cogn. Behav. Ther. 36:170–78

Ghassemzadeh H, Lzadikhah Z, Baraheni M. 2003. Memory bias in trait anxiety and obsessive compulsivedisorder. Boletin Psicol. 79:21–30

Grey S, Mathews A. 2000. Effects of training on interpretation of emotional ambiguity. Q. J. Exp. Psychol. A53A:1143–62

Grey SJ, Mathews A. 2009. Cognitive bias modification—priming with an ambiguous homograph is necessaryto detect an interpretation training effect. J. Behav. Ther. Exp. Psychiatry 40:338–43

Hackmann A, Holmes EA. 2004. Reflecting on imagery: a clinical perspective and overview of the special issueof memory on mental imagery and memory in psychopathology. Memory 12:389–402

Hakamata Y, Lissek S, Bar-Haim Y, Britton JC, Fox NA, et al. 2010. Attention bias modification treatment:a meta-analysis toward the establishment of novel treatment for anxiety. Biol. Psychiatry 68:982–90

Hallion LS, Ruscio AM. 2011. A meta-analysis of the effect of cognitive bias modification on anxiety anddepression. Psychol. Bull. 137:940–58

Harris LM, Menzies RG. 1998. Changing attentional bias: Can it effect self-reported anxiety? Anxiety StressCoping 11:167–79

Hayes S, Hirsch CR, Mathews A. 2010. Facilitating a benign attentional bias reduces negative thought intru-sions. J. Abnorm. Psychol. 119:235–40

Hazen RA, Vasey MW, Schmidt NB. 2009. Attentional retraining: a randomized clinical trial for pathologicalworry. J. Psychiatr. Res. 43:627–33


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


Hertel PT, Mathews A. 2011. Cognitive bias modification: past perspectives, current findings, and futureapplications. Perspect. Psychol. Sci. In press

Hirsch CR, Hayes S, Mathews A. 2009. Looking on the bright side: Accessing benign meanings reduces worry.J. Abnorm. Psychol. 118:44–54

Hirsch CR, MacLeod C, Mathews A, Sandher O, Siyani A, Hayes S. 2011. The contribution of attentionalbias to worry: distinguishing the roles of selective engagement and disengagement. J. Anxiety Disord.25:272–77

Hirsch CR, Mathews A. 1997. Interpretive inferences when reading about emotional events. Behav. Res. Ther.35:1123–32

Hirsch CR, Mathews A, Clark D. 2007. Inducing an interpretation bias changes self-imagery: a preliminaryinvestigation. Behav. Res. Ther. 45:2173–81

Holmes EA, Lang TJ, Shah DM. 2009. Developing interpretation bias modification as a “cognitive vaccine”for depressed mood: imagining positive events makes you feel better than thinking about them verbally.J. Abnorm. Psychol. 118:76–88

Holmes EA, Mathews A. 2005. Mental imagery and emotion: a special relationship? Emotion 5:489–97Holmes EA, Mathews A, Dalgleish T, Mackintosh B. 2006. Positive interpretation training: effects of mental

imagery versus verbal training on positive mood. Behav. Ther. 37:237–47Hoppitt L, Mathews A, Yiend J, Mackintosh B. 2010. Cognitive mechanisms underlying the emotional effects

of bias modification. Appl. Cogn. Psychol. 24:312–25Joormann J, Hertel PT, Brozovich F, Gotlib IH. 2005. Remembering the good, forgetting the bad: intentional

forgetting of emotional material in depression. J. Abnorm. Psychol. 114:640–48Joormann J, Hertel PT, LeMoult J, Gotlib IH. 2009. Training forgetting of negative material in depression.

J. Abnorm. Psychol. 118:34–43Kindt M, Brosschot JF. 1997. Phobia-related cognitive bias for pictorial and linguistic stimuli. J. Abnorm.

Psychol. 106:644–48Koster EH, Fox E, MacLeod C. 2009. Introduction to the special section on cognitive bias modification in

emotional disorders. J. Abnorm. Psychol. 118:1–4Koster HW, Crombez G, Verschuere B, De Houwer J. 2006. Attentional bias to threat in anxiety prone

individuals: mechanisms underlying attentional bias. Cogn. Ther. Res. 30:635–43Krebs G, Hirsch CR, Mathews A. 2010. The effect of attention modification with explicit versus minimal

instructions on worry. Behav. Res. Ther. 48:251–56La Greca AM, Stone WL. 1993. Social Anxiety Scale for Children-Revised: factor structure and concurrent

validity. J. Clin. Child Psychol. 22:17–27Lang TJ, Moulds ML, Holmes EA. 2009. Reducing depressive intrusions via a computerized cognitive bias

modification of appraisals task: developing a cognitive vaccine. Behav. Res. Ther. 47:139–45Lothmann C, Holmes EA, Chan SW, Lau JY. 2011. Cognitive bias modification training in adolescents:

effects on interpretation biases and mood. J. Child Psychol. Psychiatry 52:24–32Mackintosh B, Mathews A, Yiend J, Ridgeway V, Cook E. 2006. Induced biases in emotional interpretation

influence stress vulnerability and endure despite changes in context. Behav. Ther. 37:209–22MacLeod C. 1995. Training selective attention: a cognitive-experimental technique for reducing anxiety

vulnerability? World Congr. Behav. Cogn. Ther. Abstr. 118MacLeod C, Koster EH, Fox E. 2009a. Whither cognitive bias modification research? Commentary on the

special section articles. J. Abnorm. Psychol. 118:89–99MacLeod C, Mackintosh B, Vujic T. 2009b. Does explicit communication of the training contingency enhance

the efficacy of CBM? Expert Meet. Cogn. Bias Modific. Techniques, Ghent, BelgiumMacLeod C, Mathews A. 2004. Selective memory effects in anxiety disorders: an overview of research findings

and their implications. In Memory and Emotion, ed. D Reisberg, P Hertel, pp. 155–85. New York: OxfordUniv. Press

MacLeod C, Mathews A, Tata P. 1986. Attentional bias in emotional disorders. J. Abnorm. Psychol. 95:15–20MacLeod C, Rutherford E, Campbell L, Ebsworthy G, Holker L. 2002. Selective attention and emotional

vulnerability: assessing the causal basis of their association through the experimental manipulation ofattentional bias. J. Abnorm. Psychol. 111:107–23


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


MacLeod C, Soong LY, Rutherford EM, Campbell LW. 2007. Internet-delivered assessment and manipu-lation of anxiety-linked attentional bias: validation of a free-access attentional probe software package.Behav. Res. Methods 39:533–38

Mathews A. 2011. Effects of modifying the interpretation of emotional ambiguity. J. Cogn. Psychol. In pressMathews A, Mackintosh B. 2000. Induced emotional interpretation bias and anxiety. J. Abnorm. Psychol.

109:602–15Mathews A, MacLeod C. 2002. Induced processing biases have causal effects on anxiety. Cogn. Emot. 16:331–54Mathews A, MacLeod C. 2005. Cognitive vulnerability to emotional disorders. Annu. Rev. Clin. Psychol. 1:167–

95Mathews A, Ridgeway V, Cook E, Yiend J. 2007. Inducing a benign interpretational bias reduces trait anxiety.

J. Behav. Ther. Exp. Psychiatry 38:225–36McGowan N, Sharpe L, Refshauge K, Nicholas M. 2009. The effect of attentional re-training and threat

expectancy in response to acute pain. Pain 142:101–7Meyer T, Miller M, Metzger R, Borkovec TD, et al. 1990. Development and validation of the Penn State

Worry Questionnaire. Behav. Res. Ther. 28:487–95Mogg K, Bradley B. 2006. Time course of attentional bias for fear relevant pictures in spider fearful individuals.

Behav. Res. Ther. 44:1241–50Muris P, Huijding J, Mayer B, Hameetman M. 2008. A space odyssey: experimental manipulation of threat

perception and anxiety-related interpretation bias in children. Child Psychiatry Hum. Dev. 39:469–80Muris P, Huijding J, Mayer B, Remmerswaal D, Vreden S. 2009. Ground control to Major Tom: experimental

manipulation of anxiety-related interpretation bias by means of the “space odyssey” paradigm and effectson avoidance tendencies in children. J. Anxiety Disord. 23:333–40

Murphy R, Hirsch CR, Mathews A, Smith K, Clark DM. 2007. Facilitating a benign interpretation bias in ahigh socially anxious population. Behav. Res. Ther. 45:1517–29

Najmi S, Amir N. 2010. The effect of attention training on a behavioral test of contamination fears in individualswith subclinical obsessive-compulsive symptoms. J. Abnorm. Psychol. 119:136–42

Olatunji BO, Cisler JM, Deacon BJ. 2010. Efficacy of cognitive behavioral therapy for anxiety disorders: areview of meta-analytic findings. Psychiatr. Clin. North Am. 33:557–77

Ouimet AJ, Gawronski B, Dozois DJ. 2009. Cognitive vulnerability to anxiety: a review and an integrativemodel. Clin. Psychol. Rev. 29:459–70

Ree MJ, Pollitt A, Harvey AG. 2006. An investigation of interpretive bias in insomnia: an analog studycomparing normal and poor sleepers. Sleep 29:1359–62

Reese HE, McNally RJ, Najmi S, Amir N. 2010. Attention training for reducing spider fear in spider-fearfulindividuals. J. Anxiety Disord. 24:657–62

Reiss S, Peterson RA, Gursky DM, McNally RJ. 1986. Anxiety sensitivity, anxiety frequency and the predictionsof fearfulness. Behav. Res. Ther. 24:1–8

Richards A. 2004. Anxiety and the resolution of ambiguity. In Cognition, Emotion and Psychopathology: Theoretical,Empirical and Clinical Directions, ed. J Yiend, pp. 130–48. New York: Cambridge Univ. Press

Rozenman M, Weersing V, Amir N. 2011. A case series of attention modification in clinically anxious youths.Behav. Res. Ther. 49:324–30

Salemink E, Hertel P, Mackintosh B. 2010. Interpretation training influences memory for prior interpretations.Emotion 10:903–7

Salemink E, van den Hout M. 2010. Trained interpretive bias survives mood change. J. Behav. Ther. Exp.Psychiatry 41:310–15

Salemink E, van den Hout M, Kindt M. 2007a. Trained interpretive bias and anxiety. Behav. Res. Ther.45:329–40

Salemink E, van den Hout M, Kindt M. 2007b. Trained interpretive bias: validity and effects on anxiety.J. Behav. Ther. Exp. Psychiatry 38:212–24

Salemink E, van den Hout M, Kindt M. 2009. Effects of positive interpretive bias modification in highlyanxious individuals. J. Anxiety Disord. 23:676–83

Salkovskis PM, Forrester E. 2002. Responsibility. In Cognitive Approaches to Obsessions and Compulsions: Theory,Assessment, and Treatment, ed. RO Frost, G Steketee, pp. 45–61. Amsterdam: Pergamon/Elsevier Sci.


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.


Schartau PE, Dalgleish T, Dunn BD. 2009. Seeing the bigger picture: Training in perspective broadeningreduces self-reported affect and psychophysiological response to distressing films and autobiographicalmemories. J. Abnorm. Psychol. 118:15–27

Schmidt NB, Richey J, Buckner JD, Timpano KR. 2009. Attention training for generalized social anxietydisorder. J. Abnorm. Psychol. 118:5–14

See J, MacLeod C, Bridle R. 2009. The reduction of anxiety vulnerability through the modification of at-tentional bias: a real-world study using a home-based cognitive bias modification procedure. J. Abnorm.Psychol. 118:65–75

Smith E, Rieger E. 2009. The effect of attentional training on body dissatisfaction and dietary restriction.Eur. Eat. Disord. Rev. 17:169–76

Standage H, Ashwin C, Fox E. 2009. Comparing visual and auditory presentation for the modification ofinterpretation bias. J. Behav. Ther. Exp. Psychiatry 40:558–70

Standage H, Ashwin C, Fox E. 2010. Is manipulation of mood a critical component of cognitive bias modifi-cation procedures? Behav. Res. Ther. 48:4–10

Steel C, Wykes T, Ruddle A, Smith G, Shah DM, Holmes EA. 2010. Can we harness computerised cognitivebias modification to treat anxiety in schizophrenia? A first step highlighting the role of mental imagery.Psychiatr. Res. 178:451–55

Steinman SA, Teachman BA. 2010. Modifying interpretations among individuals high in anxiety sensitivity.J. Anxiety Disord. 24:71–78

Taylor S, Koch WJ, McNally RJ. 1992. How does anxiety sensitivity vary across the anxiety disorders?J. Anxiety Disord. 6:249–59

Teachman BA, Addison LM. 2008. Training non-threatening interpretations in spider fear. Cogn. Ther. Res.32:448–59

Tran TB, Hertel PT, Joormann J. 2011. Cognitive bias modification: induced interpretive biases affect mem-ory. Emotion 11:145–52

Van Bockstaele B, Verschuere B, Koster EH, Tibboel H, De Houwer J, Crombez G. 2011. Effects of attentiontraining on self-reported, implicit, physiological and behavioural measures of spider fear. J. Behav. Ther.Exp. Psychiatry 42:211–18

Vassilopoulos SP, Banerjee R, Prantzalou C. 2009. Experimental modification of interpretation bias in so-cially anxious children: changes in interpretation, anticipated interpersonal anxiety, and social anxietysymptoms. Behav. Res. Ther. 47:1085–89

Watkins ER, Baeyens CB, Read R. 2009. Concreteness training reduces dysphoria: proof-of-principle forrepeated cognitive bias modification in depression. J. Abnorm. Psychol. 118:55–64

Wells TT, Beevers CG. 2010. Biased attention and dysphoria: Manipulating selective attention reduces sub-sequent depressive symptoms. Cogn. Emot. 24:719–28

White LK, Suway JG, Pine DS, Bar-Haim Y, Fox NA. 2011. Cascading effects: the influence of attention biasto threat on the interpretation of ambiguous information. Behav. Res. Ther. 49:244–51

Wiers RW, Bartholow BD, van den Wildenberg E, Thush C, Engels RC, et al. 2007. Automatic and controlledprocesses and the development of addictive behaviors in adolescents: a review and a model. Pharmacol.Biochem. Behav. 86:263–83

Williams J, Mathews A, MacLeod C. 1996. The emotional Stroop task and psychopathology. Psychol. Bull.120:3–24

Williamson DA, Muller SL, Reas DL, Thaw JM. 1999. Cognitive bias in eating disorders: implications fortheory and treatment. Behav. Modif. 23:556–77

Wilson EJ, MacLeod C, Mathews A, Rutherford EM. 2006. The causal role of interpretive bias in anxietyreactivity. J. Abnorm. Psychol. 115:103–11

Yiend J, Mackintosh B, Mathews A. 2005. Enduring consequences of experimentally induced biases in inter-pretation. Behav. Res. Ther. 43:779–97


Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.

CP08-FrontMatter ARI 2 March 2012 11:20

Annual Review ofClinical Psychology

Volume 8, 2012 Contents

On the History and Future Study of Personality and Its DisordersTheodore Millon � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

A “SMART” Design for Building Individualized Treatment SequencesH. Lei, I. Nahum-Shani, K. Lynch, D. Oslin, and S.A. Murphy � � � � � � � � � � � � � � � � � � � � � � � � �21

Default Mode Network Activity and Connectivity in PsychopathologySusan Whitfield-Gabrieli and Judith M. Ford � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �49

Current Issues in the Diagnosis of Attention Deficit HyperactivityDisorder, Oppositional Defiant Disorder, and Conduct DisorderPaul J. Frick and Joel T. Nigg � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �77

Psychiatric Diagnosis: Lessons from the DSM-IV Pastand Cautions for the DSM-5 FutureAllen J. Frances and Thomas Widiger � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 109

American Indian and Alaska Native Mental Health:Diverse Perspectives on Enduring DisparitiesJoseph P. Gone and Joseph E. Trimble � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 131

Emotion Regulation and Psychopathology: The Role of GenderSusan Nolen-Hoeksema � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 161

Cognitive Bias Modification Approaches to AnxietyColin MacLeod and Andrew Mathews � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 189

Diagnosis and Assessment of Hoarding DisorderRandy O. Frost, Gail Steketee, and David F. Tolin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 219

The Behavioral Activation System and ManiaSheri L. Johnson, Michael D. Edge, M. Kathleen Holmes,

and Charles S. Carver � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 243

Prediction and Prevention of Psychosis in Youth at Clinical High RiskJean Addington and Robert Heinssen � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 269

From Dysfunction to Adaptation: An InteractionistModel of DependencyRobert F. Bornstein � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 291

vi

Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.

CP08-FrontMatter ARI 2 March 2012 11:20

Personality Disorders in DSM-5Andrew E. Skodol � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 317

Development, Evaluation, and Multinational Disseminationof the Triple P-Positive Parenting ProgramMatthew R. Sanders � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 345

Empirical Classification of Eating DisordersPamela K. Keel, Tiffany A. Brown, Lauren A. Holland, and Lindsay P. Bodell � � � � � � � � 381

Obesity and Public PolicyAshley N. Gearhardt, Marie A. Bragg, Rebecca L. Pearl, Natasha A. Schvey,

Christina A. Roberto, and Kelly D. Brownell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 405

Cognition in the Vegetative StateMartin M. Monti � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 431

Coping with Chronic Illness in Childhood and AdolescenceBruce E. Compas, Sarah S. Jaser, Madeleine J. Dunn, and Erin M. Rodriguez � � � � � � � 455

Indexes

Cumulative Index of Contributing Authors, Volumes 1–8 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 481

Cumulative Index of Chapter Titles, Volumes 1–8 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 484

Errata

An online log of corrections to Annual Review of Clinical Psychology articles may befound at http://clinpsy.annualreviews.org

Contents vii

Ann

u. R

ev. C

lin. P

sych

ol. 2

012.

8:18

9-21

7. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Iow

a on

12/

04/1

3. F

or p

erso

nal u

se o

nly.

cognitive bias modification approaches to anxiety...cognitive bias modiﬁcation (cbm): direct...

Documents