Subliminal attention bias modification training in socially anxious

12 Pages·2013·1.74 MB·English

by  Keren Maoz

ORIGINALRESEARCHARTICLE
HUMAN NEUROSCIENCE published:23July2013
doi:10.3389/fnhum.2013.00389
Subliminal attention bias modification training in socially
anxious individuals
KerenMaoz1*,RanyAbend1,NathanA.Fox2,DanielS.Pine3 andYairBar-Haim1,4
1SchoolofPsychologicalSciences,TelAvivUniversity,Tel-Aviv,Israel
2DepartmentofHumanDevelopment,UniversityofMaryland,CollegePark,MD,USA
3NationalInstituteofMentalHealth,Bethesda,MD,USA
4SagolSchoolofNeuroscience,TelAvivUniversity,Tel-Aviv,Israel
Editedby: Anxious individuals demonstrate threat-related attention biases both when threat stimuli
AlexandreHeeren,Univesité are presented within conscious awareness and when presented below awareness
CatholiquedeLouvain,Belgium threshold. Nevertheless, attention bias modification (ABM) research has rarely utilized
Reviewedby: sub-awareness protocols in an attempt to modify attention patterns and reduce anxiety.
Wolf-GeroLange,Radboud
Exploring the potential of subliminal ABM is of interest, as it may target attention
UniversityNijmegen,Netherlands
JunMoriya,RikkyoUniversity,Japan processes related to anxiety that are distinct from those engaged by supraliminal
CristinaMogoas¸e,Babes¸-Bolyai ABM. Here we examined the effect of a subliminal ABM training protocol on levels of
University,Romania social anxiety and stress vulnerability. Fifty-one socially anxious students were randomly
EvaDeHullu,OpenUniversity,
assigned to either ABMor placebo condition, and completed a pre-training assessment,
Netherlands
four training sessions, a social stressor task, and a post-training assessment. Results
*Correspondence:
KerenMaoz,SchoolofPsychological indicate that the subliminal ABM used here did not induce detectable changes in
Sciences,TelAvivUniversity,Ramat threat-related attention from pre- to post-training as measured by two independent
Aviv,Tel-Aviv69978,Israel attention tasks. Furthermore, the ABM and placebo groups did not differ on either
e-mail:[email protected]
self-reported social anxiety post-training or state anxiety following stress induction.
Post-hoc auxiliary analyses suggest thatABMmaybeassociatedwithsmallerelevations
in state anxiety during the stressor task only for participants who demonstrate attention
biastowardthreatatbaseline.Implicationsandfutureresearchdirections arediscussed.
Keywords:socialanxiety,stressvulnerability,attentionbiasmodification,masking,subliminal
INTRODUCTION thevastmajorityofABMstudiestodatehaveusedsupraliminal
Numerous studies across clinical and sub-clinical populations (i.e.,consciouslyperceived)stimulipresentationstomodifythese
havefoundthatanxious individualsdemonstrate anattentional biases.
bias toward threat-related stimuli (Bar-Haim et al., 2007). This Exploring the potential of subliminal ABM is of interest, as
bias manifests even when threat stimuli are presented below it may target a different layer of attention processes related to
awarenessthresholds(MathewsandMacLeod,1986;Moggetal., anxiety. In line with this idea, brain imaging and psychophysi-
1993; van den Hout et al., 1995; Fox, 2002; Mogg and Bradley, ology studies found distinct responses to readily-identifiable,as
2002). Additional findings further suggest that threat biases opposedtomasked,subliminalthreat-relatedstimuliinanxious
causallyaffectstressvulnerability(MacLeodetal.,2002;Mathews relative to non-anxious individuals (Ohman and Soares, 1994;
and MacLeod, 2002; Eldar et al., 2008). Based on such obser- Etkin et al., 2004; Li et al., 2007; Tsunoda et al., 2008). These
vations, attention bias modification (ABM) treatments have findings suggest that hypersensitivity to threat in anxious indi-
startedtoemerge,exploringthepotentialofcomputerizedtools viduals may occur prior to conscious awareness. For example,
to modify attention patterns and consequently reduce stress- Etkin et al. (2004) demonstrated that supraliminalandsublim-
vulnerability and anxiety (Koster et al., 2009; Bar-Haim, 2010; inalpresentationsofthreatfacesmodulatedneuralactivationin
Hakamataetal.,2010;Beard,2011;HallionandRuscio,2011). distinctregionsoftheamygdala.Specifically,subliminalpresenta-
Clinical ABM trials indicate that training patients to attend tionsmodulatedactivityinthebasolateralregionoftheamygdala,
awayfromthreatreduces self-reported aswell asclinicallyeval- and this activation was positively correlated with trait anxiety.
uatedanxietylevels(Amiretal.,2009,2011;Schmidtetal.,2009; Thus,subliminalABMmayprovideanopportunitytointervene
Eldar et al., 2012). In addition, studies with non-clinical high- with anxiety-maintainingmechanisms thatact very early in the
anxious participants show that ABM training typically reduces processingstream.
stressvulnerabilityinthefaceoflab-induced(Amiretal.,2008; To our knowledge only one study used subliminal presenta-
Bar-Haim et al., 2011) or real-life (See et al., 2009) stressors. tions in the context of ABM. MacLeod et al. (2002; study 1)
However, although many studies have utilized subliminal stim- used a dot-probe task to train non-anxious students to attend
uli to measure preconscious threat-related attention biases and eithertothreatorneutralwords.Asingle-sessionprotocol with
their associations to anxiety and stress vulnerability (MacLeod 576 active training trials was used, in which half of the tri-
and Hagan, 1992; van den Hout et al., 1995; Fox et al., 2010), als were presented subliminally and half were presented well
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|1
Maozetal. SubliminalABMTinsocialanxiety
within consciousawareness. Inaddition,96attention biasmea- METHODS
surement trials were intermixed throughout the active training PARTICIPANTS
trials. Half of these measurement trials were subliminally pre- Sixtysociallyanxiousundergraduatestudentswereinvitedtopar-
sentedandtheotherhalfwerepresentedwithinconsciousaware- ticipate in the study based on their high total scores (>30) on
ness. Following training, lower stress vulnerability was found theLiebowitzSocialAnxietyScale(LSAS,Liebowitz,1987)com-
in the group trained to attend away from threat relative to the pletedinamasssurveyatthebeginningoftheacademicyear.A
group trained to attend toward threat. The results also indi- cutoffscoreof30wasfoundtoprovidethebestbalancebetween
cated that attentional changes following training emerged only false positive and false negative diagnostic errors in classifying
for consciously presented measurement trials and not for sub- individuals with social anxiety disorder (Mennin et al., 2002;
liminalmeasurementtrials.However,becausealltrialtypes(sub- Rytwinskietal.,2009).TheLSASwasagainadministeredtothese
liminal, supraliminal; training, measurement) were presented 60 students in the lab during the pre-assessment session of the
in a mixed fashion, conclusive inference on the specific effect studytoverifyhighsocialanxietylevels.Eightstudentsreported
of subliminal training was complicated. Moreover, as men- lower levels of social anxiety relative to their initial report and
tioned, this study included only non-anxious participants. It no longer met the criterion of LSAS >30. These students were
has been previously demonstrated that non-anxious individu- thus excluded from further participation inthe study. An addi-
als’ attention is less reactive to subliminal threatening stimuli tional student decided not to participate. Thus, 51 participants
as compared to anxious individuals (Mathews and MacLeod, (meanage=22.70years,SD=1.65;41females)wererandomly
1986; Mogg et al., 1993). This could suggest a possible expla- assignedtoeitheranABMgroup(n=24)or aplacebocontrol
nation as to why no change in preconscious attention pro- group (n=27). The mean LSAS score for the final sample was
cesses was demonstrated following training among these non- 55.73(SD=16.90),placingtheirmeanscoremorethan3stan-
anxious subjects. Finally, the study by MacLeod et al. (2002) dard deviations above the mean for individuals with no axis I
usedwordstimulithatmightbelessoptimalthanevolutionary- diagnosis(Frescoetal.,2001). Thegroups didnotdifferinage,
relevant threat, such as faces (Ohman and Mineka, 2001), for gender distribution, baselinethreat biasscores, andmean LSAS
early threat-attention modification processes. Thus, it appears scores,allps>0.15(seeTable1forbaselinemeansandSDsby
that more research is needed to explore the effects of sublim- group, and Figure1 for a CONSORT diagram). The study was
inal ABM on anxiety and stress vulnerability among anxious approvedbytheinstitutionalreviewboard.Participantsprovided
individuals. signedinformedconsent.
The aim of the current study was to examine the efficacy of
a subliminal dot-probe ABM protocol on attention bias, anx- QUESTIONNAIRES
iety levels, and stress vulnerability in a group of undergrad- SocialanxietywasassessedwiththeLSAS(Liebowitz,1987).This
uate students with high levels of self-reported social anxiety. scaleconsistsof24items describingsocialinteractionsandper-
We decided to focus on socially anxious individuals for sev- formance situations. The LSAS possesses strong psychometric
eral reasons: first, for methodological reasons, we wanted to properties(Frescoetal.,2001).TheHebrewversionoftheLSAS
keep our sample ashomogenous as possiblewith respect to the wasfoundvalidandreliable(Levinetal.,2002).Cronbach’salpha
nature of their anxiety. This enabled us to specify the stim- in the current sample was 0.90 and 0.93 for the baseline and
uli and the stressful manipulation to the characteristics of this post-ABM/Placebosessions,respectively.
particular anxiety. Second, we selected this specific population Stateanxietywasmeasuredwiththestatesub-scale(STAI-S)of
becauseprevious findingsindicateABMefficacywithsupralim- theState-TraitAnxietyInventory(STAI,Spielbergeretal.,1983).
inalpresentationsinclinicallydiagnosedSocialAnxietyDisorder The STAI-S consists of 20 items measuring current, situational
patients (Amir et al., 2009; Schmidt et al., 2009; Heeren et al., levelsofanxiety.TheHebrewversionoftheSTAIwasfoundvalid
2011, 2012), as well as in analog samples with moderate to andreliable(Teichman andMelinic,1979). Cronbach’salphain
highself-reportedsocialanxiety(Amiretal.,2008;Klumppand the current sample was 0.89 or higher in each of the STAI-S
Amir, 2010). The subliminal ABM protocol used in the cur- administrations (baseline, pre and during the stressor task, and
rent study followed the parameters from these previous supral- atpostmeasurement).
iminal ABM studies, which have demonstrated positive effects
of ABM in socially anxious populations (Amir et al., 2008, ATTENTIONALBIASASSESSMENT
2009; Schmidt et al., 2009), but with subliminal presentations. Thedotprobetask
We expected the subliminal ABM protocol to change threat- The sequence of events on a dot-probe trial is described in
related attentionpatterns inaccordwith thetrainingcondition. Figure2. Each trial began with the presentation of a fixation
That is, that participants trained to attend away from threats display (500ms; white cross 1×1cm), on which the partici-
would show a reduction in threat-related attention bias follow- pants were requested to focus their gaze. The fixation display
ing ABM. No change in attention pattern was expected in the was followed by a presentation of a pair of faces. Face pairs
placebo control condition which was not intended to manipu- comprisedeitherdisgust-neutralorneutral-neutralfacialexpres-
late attention. We also expected that participants in the ABM sions of the same actor. Pictures of eight different actors were
condition will displayless anxiety andlower stress vulnerability used (four female), taken from a standardized set of emotional
followingtrainingrelativetoparticipantsintheplacebocontrol expressions (Matsumoto and Ekman, 1988). Each face pho-
condition. tograph was placed on a gray square background subtending
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|2
Maozetal. SubliminalABMTinsocialanxiety
Table1|MeansandSDsofbaseline,post-training,pre-stressor,andduring-stressormeasurementsbygroup∗.
Baseline Post-training
ABM Placebo-Control ABM Placebo-Control
Gender(F/M) 20/4 20/6
Age 22.96(1.94) 22.42(1.33)
LSASscore 54.67(14.55) 55.96(18.99) 53.63(16.88) 54.62(21.28)
STAI-Sscore 38.00(10.80) 38.35(7.92) 34.59(9.28) 38.54(9.50)
DOT-PROBE
MeanRT—threat 527(63) 533(47) 478(44) 470(39)
MeanRT—neutral 527(58) 532(50) 479(40) 469(33)
Threatbiasscore 0.10(19) −0.88(18) 1.01(18) −0.97(17)
AFFECTIVESPATIALCUING
MeanRT—threatvalid 585(92) 565(69) 556(79) 542(56)
MeanRT—neutralvalid 582(94) 562(70) 559(76) 542(59)
MeanRT—threatinvalid 659(118) 666(98) 636(106) 646(84)
MeanRT—neutralinvalid 675(146) 664(92) 635(102) 647(89)
Threatengagement −3.06(23) −2.35(22) 3.59(22) −0.49(20)
Threatdisengagement −16.04(54) 1.52(40) 0.74(36) −0.83(37)
Pre-stressor During-stressor
ABM Placebo-Control ABM Placebo-Control
STAI-Sscore 38.29(10.19) 40.27(10.85) 49.84(9.75) 52.73(10.49)
*Nobetween-groupdifferenceswerefoundatbaseline,post-training,pre-stressororduring-stressor,allps>0.10.
LSASisLiebowitzSocialAnxietyScale,STAI-SisSpielbergerState-TraitAnxietyInventory-State.
50mm in width and 37.5mm in height. The face photographs Theaffectivespatialcueingtask
were presented with equal distance from the top and bottom Assessingattentionbiasusingthedot-probetaskfollowingABM
of the fixation cross, with a distance of 15mm between them. may reflect only near transfer of the training effect since it
The top photograph was positioned 30mm from the top edge relies onthe sametask demands andstimuli as the ABMtrain-
of the screen. The faces were displayed for 17ms and were ing itself. To further test for generalizationof potential changes
then masked by a pair of scrambled neutral faces displayed in threat attendance as a function of ABM, we used an affec-
for 68ms (see Mogg and Bradley, 2002 for a similar masking tive variant of Posner’s spatial cueing task (Stormark et al.,
procedure). After the masking disappeared, a target probe con- 1995; Fox et al., 2001, 2002). In Posner’s original task (Posner,
sisting of either the letter E or F (font Arial, size 14, bold) 1980), a cue appears in one of two locations, and is followed
appeared at the location previously occupied by one of the by a target at the cued location on a majority of the trials
masks, and remained on the screen until response. Participants (valid cue) and at the alternative location on a minority of
had to determine which of the letters appeared by pressing the trials (invalid cue). Speeding on valid trials is attributed to
one of two pre-specified buttons on a mouse. The task com- the benefits of attentional engagement with the cued location.
prised128trialsofdisgust-neutralpairsand32trialsofneutral- Slowing on invalid trials is associated with the costs of hav-
neutral pairs, for a total of 160 trials, displayed in a random ing to disengage attention from the cued location. Systematic
order. manipulationoftheemotionalcontentofcuesrevealstheeffectof
The 128 disgust-neutral trials were counterbalanced with cuevalenceonattention.Studiesusingthistasktypicallyreport
regard to actor identity, disgusted face location (top, bottom), increased dwelling time on threat invalid cues relative to neu-
probe location (top, bottom), and probe type (E, F). Attention tral invalid cues, in anxious relative to non-anxious individuals
biasforeachparticipantwascalculatedbysubtractingthemean (Bar-Haim et al., 2007). This is thought to reflect a difficulty
RT of trials in which the target probe appeared at the disgust indisengagingattentionfromthreatamonganxiousindividuals.
face location from the mean RT of trials in which the target To test for far transfer of training effects we used the emo-
appearedattheneutralfacelocation.Positivescoresreflectabias tional spatial cuing task in addition to the dot-probe task and
towardthreat(threatvigilance),whereasnegativescoresreflectan also used different stimuli (angry faces rather than disgusted
attentionalbiasawayfromthreat(threatavoidance).Onneutral– faces).Bothdisgustandangerexpressionsconstitutethreatening
neutraltrialstargetlocationandtypewerefullycounterbalanced cuestosociallyanxiousindividuals,andattentionalvigilancefor
and RTs from these trials were not included in attention bias bothtypesofstimuliwasdemonstratedinthispopulation(e.g.,
calculation. Moggetal.,2004;Pishyaretal.,2004).Here,theemotionalcues
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|3
Maozetal. SubliminalABMTinsocialanxiety
FIGURE1|CONSORTdiagramandsequenceofeventsinthestudy.
the cue was presented either in the left or right box for 17ms,
and immediately masked by a scrambled neutral face displayed
for 68ms. The target arrow then appeared in either the same
box as the cue (valid trials) or the opposite box (invalid trials)
andremainedon thescreen until response. Thetaskcomprised
192trialsofwhich75%werevalidand25%wereinvalid.Within
eachtypeoftrial(valid/invalid),cuetype(neutral/angry),target
location(left/right),andtargettype(pointingup/pointingdown)
were fully counterbalanced. Throughout the task each actor’s
photographsappearedatotalof12times—6timeswithanangry
expressionand6timeswithaneutralexpression.Threatengage-
ment was calculated as mean RT for valid neutral trials minus
meanRTforvalidthreattrials.Positiveengagementscoresreflect
attentionalbiastowardthreat(threatengagement),whereasnega-
FIGURE2|Sequenceofeventsinasubliminaldot-probetrial. tiveengagementscoresreflectanattentionalbiasawayfromthreat
(threatavoidance).Threatdisengagementwascalculatedasmean
RT for invalid threat trials minus mean RT for invalid neutral
consisted of face photographs of 16 different actors (8 females) trials. Positive disengagement scores are considered to reflect a
taken from the NimStim stimulus set (Tottenham et al., 2009). difficultyindisengagingattentionfromthreatstimuli.
Twopicturesofeachactorwereselecteddepictinganangryand
a neutral expression. The target was an arrow pointing either ATTENTIONBIASMODIFICATION(ABM)
up or down. Participants had to determine the arrow’s direc- TheABMversionofthedot-probetaskdisplayedthesamestim-
tion by pressing one of two pre-specified buttons on a mouse. uli as those used for threat bias assessment except that target
Cueandtargetstimuliwerepresentedinsidetwodarkgrayboxes probes (E,F)appeared onlyatthelocationpreviouslyoccupied
(50mm×65mm)whichweredisplayedcontinuouslytotheleft byneutralfaceswiththeaimofimplicitlyestablishingtheseasa
andtherightofthescreencenter.Eachtrialwasinitiatedbyafix- predictivecueforthelocationoftheprobe.Theplacebocontrol
ationcrosspresentedinthecenterofthescreenfor500ms.Then, group received the same number andtype of trials as the ABM
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|4
Maozetal. SubliminalABMTinsocialanxiety
group but in a fully counterbalanced manner as was done dur- andaskedparticipantstocompletetheSTAI-Sagain.Theexper-
ingthreatbias assessment. Thus,noattention modification was imenter made it clear that the speech will be resumed shortly
expectedintheplacebocontrolgroup. after completion of thequestionnaire. Followingoneadditional
minute of speech participants were halted and thanked. Stress
VISUALMASKINGEFFICACYTEST vulnerabilitywasindexedasthechangebetweenpre-andduring-
To ensure that participants were not consciously aware of the stressor STAI-S. The seventh and final session was held in the
emotional valence of the masked faces, an objective detection following week and took place in the same room as the train-
task was used (Merikleet al.,2001). Thistwo-alternative forced ingsessions.Eachparticipantperformedthesamedot-probeand
choice task comprised 32 trials. Stimuli were pairs of identi- affectivespatialcueingtasksasinthebaselinesession.Then,par-
cal face pictures (i.e., two neutral faces or two disgusted faces) ticipantscompletedthetestofvisualmaskingefficacyfollowedby
takenfromthosepresentedintheassessmentversionofthedot- completionoftheLSASandtheSTAI-S.
probe task. In each trial, a face pair was presented and masked
in the same manner as in the dot-probe task. Participants were DATAANALYSIS
told that half of the trials contain a pair of identical faces fea- Trials with RTs shorter than 150ms or longer than 2000ms, or
turinganegative valence,whereas theotherhalfcontainsapair incorrect response were excluded. Then, for each participant,
of identical neutral faces, and had to indicate via button press meanRTpertrialtypewascalculated,andtrialswithRTsdevi-
whether the faces in each trial were “neutral” or “negative.” A atingbymorethan2.5SDsfromthemeanwerefurtherexcluded.
95% confidence interval was calculated to reflect chance level Thisresultedintheremovalofanaverageof6%ofalltrialsper
performance. participant.
The effect of subliminal ABM on attention was assessed for
SOCIALSTRESSINDUCTIONTASK biasscoresonthedot-probetaskandforRTsintheaffectivespa-
The social stress induction task was similar to the one used by tialcueingtask.Dot-probeattentionbiasscoresweresubjectedto
Amiretal.(2008).Participantswereaskedtochooseoneofthree a2×2repeated-measures ANOVAwithGroup(ABM,placebo)
discussion topics (using nuclear energy to produce electricity, as a between-subjects factor and Time (baseline, post-training)
school uniform, or toll roads) and prepare a 5-minute speech asarepeatedwithin-subjectfactor.Responsetimesontheaffec-
concerning claims in favor of and against the selected topic. tivespatialcueingtaskweresubmittedtoa2×2×2×2ANOVA
Participantswereinformedthattheirspeechwouldbevideotaped with Group (ABM, placebo) as a between-subjects factor, and
and later evaluated for quality by the research staff. During the Time(baseline,post-training),CueValidity(valid,invalid),and
speech taskanunfamiliarmaleexperimenter waspresentinthe CueValence(threat,neutral)asrepeatedwithin-subjectfactors.
room,providedinstructions,andoperatedthevideocamera. ToexaminetheeffectofsubliminalABMontraitsocialanxiety
levels, total social anxiety scores from the LSAS were submitted
PROCEDURE to a 2×2 ANOVA with Group (ABM, placebo) as a between-
Over a period of five weeks, participants completed a base- subjects factor and Time (baseline, post-training) as a repeated
lineassessmentsession,fourattentiontraining/placebosessions, within-subject factor. To examine the effect of subliminal ABM
a stress induction session, and a final evaluation session (see onvulnerabilitytosocialstress,arepeated-measuresANOVAwas
CONSORT diagram Figure1). The baseline assessment session conductedonSTAI-Sscoresbeforeandduring-thestressortask.
lasted 25min during which participants completed the STAI-S, Group (ABM, placebo) served as a between-subjects factor and
LSAS, and the dot-probe and affective spatial cueing tasks. All Stressor-Phase(pre-stressor,during-stressor)servedasarepeated
computerizedtaskswereruninadarkenedroom,ona17-inch- within-subjectfactor.
screenlaptopcomputer (LenovoR61i),usingE-Primesoftware. Because recent studies suggest that baseline attention bias
Participantswereseatedataviewingdistanceof80cmfromthe toward threat may predict supraliminal ABM training efficacy
monitor. Following the assessment session participants received (Amiretal.,2011),aswellascognitive-behavioraltreatmenteffi-
fourtrainingsessionsaccordingtotheirgroupassignment(two cacy (Waters et al., 2012), we explored this possibility in the
sessionsonnonconsecutivedaysperweek,overtwoweeks).Each currentsubliminalABMstudy.Weconductedtwopost-hocanaly-
training session lasted approximately 10min. The sixth session, sestotestwhetherbaselinevigilanceoravoidance(attentionbias
conducted4–7daysfollowingthelasttrainingsession,wasded- towardor awayfromthreat)modulated theeffect ofsubliminal
icatedtotestingtheeffects ofABMonstressvulnerabilityusing ABM on social anxiety and stress vulnerability. First, following
thesocialstress inductiontask.Thissessiontook placeinadif- Waters et al. (2012), we divided the participants to two groups
ferent room than the room of the training sessions. The male based on whether they had a bias toward threat (“attenders,”
experimenter administering this session was unfamiliar to the attentionbias>0;n=29)orabiasawayfromthreat(“avoiders,”
participants and blind to all aspects and purposes of the study. attention bias < 0; n=21) at baseline. This new dichotomous
ParticipantscompletedtheSTAI-Sinawaitingroom.Theexper- variablewas entered asanadditionalbetween-subjects factor in
imentertheninvitedparticipantstoenterthetestingroomwhere the above described primary ANOVAs. Second, following Amir
the social stress procedure was conducted. Following 5min of etal.(2011),weregressedbaselineattentionbiasasacontinuous
speechpreparationparticipantswereaskedtostepuptoamarked predictor,alongwithtraininggroup(ABM/Placebo)(step1)and
spotinfrontofthecameraanddelivertheirspeech.Twominutes theirinteraction term(step2)onstateanxiety(STAI-S)change
intotheirspeech, theexperimenter temporarilypaused thetask scorefrompre-toduringthestressinductionepisode.Thesame
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|5
Maozetal. SubliminalABMTinsocialanxiety
regressionmodelwasalsoappliedtochangeintraitsocialanxiety Socialstressvulnerability
(LSAS)frompre-topostABM/Placebo. AmaineffectofStressor-Phasewasfound,F(1,48) =54.86,p<
0.0001, η2 =0.53, demonstrating that the stressor task signifi-
p
RESULTS cantly increased state anxiety levels from pre-stressor (mean =
VISUALMASKINGEFFICACYTEST 39.32, SD=10.48) to during-stressor (mean = 51.34, SD=
All participants but one performed the task at chance level 10.14). No other main or interaction effects reached statistical
(mean=50.12%correct,SD=8.12),indicatingthatthemask- significance,allps>0.32.
ingprocedureswereeffectiveandthatparticipantswereunaware
oftheaffectivevalenceofthefaces.Oneparticipanthadanabove-
SECONDARYPOST-HOCANALYSES—TESTINGTHEEFFECTOF
threshold accuracy performance (72% correct). All the analyses
BASELINEATTENTIONBIAS
reported exclude the data from this participant. When analyses
Baselinebiasasadichotomousfactor
wereconductedincludingthisparticularsubject,nochangeswere
Means and SDs of all baseline and post-training measurements
notedintheresultspattern.
for baseline attenders and avoiders are presented in TableA1
(seeappendix).Theaveragebaselineattentionbiasscoresinthe
BASELINEMEASUREMENTS
attenders (12ms, SD=9) and the avoiders (−18ms, SD=12)
Means and SDs for LSAS, STAI-S, and RTs and bias scores on
the dot-probe and affective spatial cueing tasks at baseline by groups were each significantly different from zero, t(28) =7.44,
training condition are provided in Table1 (left panel). None p<0.0001 and t(20) =−6.63, p<0.0001, respectively. There
were no significant differences in baseline bias scores between
of these measures significantly differed between the ABM and
Placebo-Controlgroups(allps>0.19).Attentionbiasinthedot ABM and control participants neither in the attenders nor in
the avoiders groups (all ps > 0.6, for means and SDs see
probetaskaswellasengagementanddisengagementbiasesinthe
Table2).
affective spatialcueingtaskwere notsignificantlydifferentthan
Adding the baseline attention bias (toward or away from
zero neither in the ABM group nor in the Control group (all
ps>0.15). threat)asafactortoallanalysesproducedtwosignificantinterac-
tion effects: first, when testing for effects of ABM on attention
POST-TRAININGMEASUREMENTS bias using the dot-probe task, a Time-by-Baseline Bias inter-
Means and SDs for LSAS, STAI-S, and RTs and bias scores on action effect emerged, F(1,46) =24.27, p<0.0001, η2p =0.35.
the dot-probe andaffective spatial cueing tasks at post-training Both participants who had an attention bias toward threat and
bytrainingconditionareprovidedinTable1(rightpanel). participantswhohadanattentionbiasawayfromthreatatbase-
line(means=12.31and−17.97,SDs=8.91and12.42,respec-
Changeinattentionthreatbias tively)convergedtowardhavingnobiasfollowingABM/Placebo
Dot-probe. Thisanalysisyieldednosignificantmainorinterac- (means=0.44and−0.66,SDs=18.43and15.13,respectively).
tion effects, indicating no detectable changes in attention bias The second significant interaction was related to the effect
scoresfrompre-topost-training,allps>0.68. of ABM on stress vulnerability. STAI-S means and SDs before
and during the social stressor task are presented in Table2.
Affective spatial cuing task. RTs to invalidly cued trials were There was a significant three-way Stressor-Phase-by-Group-by-
longerthanRTstovalidlycuedtrialsreflectingtheclassicPosner BaselineBiasinteractioneffect,F(1,46) =12.31,p<0.001,η2p =
validityeffect,F(1,48) =187.13,p<0.0001,η2p =0.80.Inaddi- 0.21. To explicate this interaction, two ANOVAs of Group
tion,amaineffectofTimewasfound,reflectingfasteroverallRTs (ABM,placebo)byStressor-Phase(pre-stressor,during-stressor)
followingABM/Placebo,F(1,48)=6.39,p<0.05,η2p =0.12.No were conducted, one for participants who had attention bias
othermainorinteractioneffectsreachedstatisticalsignificance. awayfromthreatatbaseline(avoiders)andoneforparticipants
who had attention bias toward threat at baseline (attenders).
Traitsocialanxiety(LSAS) Forthreatavoiders,thosewhoreceivedABMshowedlargerele-
This analysis yielded no significant main or interaction effects, vation in state anxiety in response to the stressor task relative
indicatingthatsubliminalABMdidnotaffectself-reportedtrait to their counterparts in the placebo training group. However,
socialanxiety(LSAS),allps>0.41. this interaction was non-significant, F(1,19) =3.89, p=0.063.
Table2|MeansandSDsofbaselineattentionbiasandstateanxiety(STAI-S)pre-andduring-stressor,forbaselineattendersandavoidersby
traininggroup.
Baselineattenders Baselineavoiders
ABM(N=15) Placebo-Control(N=14) ABM(N=9) Placebo-Control(N=12)
Baselineattentionbias 12(8) 13(10) −19(16) −17(9)
STAI-S
Pre-Stressor 39.33(7.91) 36.65(8.26) 36.56(13.55) 44.50(12.27)
During-Stressor 47.60(9.72) 54.71(9.93) 53.57(9.12) 50.42(11.07)
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|6
Maozetal. SubliminalABMTinsocialanxiety
For threat attenders, however, a significant Stressor-Phase-by- change. Baseline attention bias predicted state anxiety change
Group interaction effect was found, F(1,27) =10.36, p<0.005, at a trend level of significance with greater baseline threat bias
η2 =0.28.Follow-upcontrastsindicatedthatbothgroups(ABM predicting greater elevations in state anxiety following stress
p
and Placebo) showed significantincreases in state anxiety from induction. Importantly, the interaction term between baseline
before-to-duringstressor, both ps<0.001.Additionalbetween- attention bias and ABM/Placebo group significantly predicted
group contrasts revealed that the two groups did not differ on state anxiety change. Follow-up simple slope analyses demon-
STAI-S scores before stress induction, t(27) =−0.89, p=0.38. strated that for the Placebo-Control group the slope coefficient
Interestingly, the ABM group showed lower STAI-S scores rela- was positive and significantly different from zero, B=0.46,
tive to the placebo group during stress, t(27) =1.95, p=0.062. t(24) =4.55, p<0.0001, suggesting that in this group, individ-
This non-significant trend may suggest that among those who uals with greater attention bias to threat at baseline demon-
attendedtowardthreatatbaseline,thosewhoreceivedABMwere strated larger elevations of anxiety during the stress task. In
lessvulnerabletothestressor(Figure3). contrast,fortheABMgrouptheslopecoefficientwasnotsignif-
icantly different from zero, B=−0.16, t(22) =−1.36, p>0.15
Baselinebiasasacontinuousfactor (Figure4).
The estimated coefficients and significance levels for the two The regression model predicting change in trait social anx-
steps in the regression model are shown in Table3. The overall iety (LSAS) from pre- to post-training was non-significant
regression model significantlyexplained 31 percent of the vari- (p>0.8).
ance in state anxiety change due to stress induction, F(3,46) =
6.80,p<0.001.Thismodelexplainedsignificantlyandsubstan- DISCUSSION
tially more variance in stress-related anxiety change as com- The present study is the first to report a randomized con-
paredtothemodel consideringonlybaselineattentionbiasand trolled ABM trial using subliminally-presented stimuli in high
group as single predictors, without taking into account their socially-anxiousindividuals.Theaimofthestudywastoexam-
interaction.Specifically,whennotconsideringtheinteractionin ine whether subliminal ABM training away from threat faces
the model, ABM/Placebo Group did not predict state anxiety
FIGURE4|Simpleslopeanalyses:theestimatedchangeinstate
FIGURE3|StateAnxietyScores(STAI-S)andstandarderrorbarsfor anxietyfrompretoduringthestressortask,asafunctionofthe
participantswithbaselinethreatvigilancepre-andduringthesocial interactionbetweenbaselineattentionbiasandtraininggroup
stressortaskbyattentiontrainingcondition.∗∗p<0.0001;+p=0.06. (ABM/Placebo-control).
Table3|Estimatedcoefficients,standarderrors,and0.95confidenceintervalsforpredictorsinthetwostepsoftheregressionmodel
predictingstressor-relatedanxietychange.
Predictor B SE t 95%CI R2 (cid:2)R2
Step1 Baselineattentionbias 0.15+ 0.09 1.74 −0.02– 0.33 0.06
Traininggroup −1.06 3.18 −0.34 −7.45– 5.33
Step2 Baselineattentionbias 0.46** 0.11 4.27 0.24– 0.67 0.31 0.25**
Traininggroup −1.30 2.76 −0.47 −6.85– 4.26
Baselineattentionbias×Traininggroup −0.61** 0.15 −4.04 −0.92–−0.31
+p=0.089;**p<0.001.Traininggroup=ABM/Placebo-controlgroups.B=unstandardizedestimatedcoefficient.SE=standarderror.CI=confidenceinterval.
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|7
Maozetal. SubliminalABMTinsocialanxiety
was effective in reducing levels of social anxiety and social change early preconscious processes using subliminal ABM
stress vulnerability in socially-anxious students. In our effort requires more training sessions, more trials per session,
to accomplish this aim, we relied on commonly used ABM different stimuli, or different masking procedure. If so, more
and attention bias measurement methods that have proved experimental research is needed to unveil these param-
effective in supraliminal ABM research in similar populations. eters that individually or together obscure the expected
However, despite this methodological effort and in contrast effects.
to our expectations, the subliminal ABM used in the current Third, it may be considered that the subliminal ABM train-
study did not induce detectable change in threat-related atten- ing did in fact influence attentional patterns, but our mea-
tion, neither on the dot-probe task (near transfer) nor on surement tools failed to detect this change. This could be
the affective spatial cueing task (far transfer). We were also due to the relatively long time interval that elapsed from the
unable to find an effect of ABM on self-reported trait social end of training to the post-training measurement of atten-
anxiety. tion bias, or due to the effects of using subliminal stimuli.
Therearevariouspotentialexplanationsforthenull-findings For example, one may consider the possibility that consciously
which could be broadly classified into four general types: (a) perceived threat is necessary for ABM effects on attention to
that subliminal ABM is inherently ineffective for changing pre- surface in measurement. In supraliminal ABM protocols the
conscious attention patterns in anxious individuals; (b) that presence of threat stimuli is consciously perceived through-
the specific subliminal ABM protocol used here was not effec- out all assessment and practice sessions. In contrast, in the
tive in inducing the expected change in attention patterns; current study conscious perception of threat never occurred
(c) that subliminal ABM did in fact modify attentional pat- neither during training nor during threat bias assessments.
terns, but our measurement precision failed to detect this Future studies may consider measuring change in threat bias
change; and (d) that subliminal attention processes are related usingsupraliminalpresentationsevenwhenABMissubliminally
to anxiety only in a sub-group of socially anxious individ- delivered.
uals, those who demonstrate threat-related attention bias at Finally, the finding of reduced stress vulnerability following
baseline. Because our sample as a whole did not demon- subliminal ABM in participants who demonstrated attention
strate a measureable attention bias toward subliminal threat at bias toward threat at baseline may offer a clue that for a sub-
baseline this possibility was tested in post-hoc analyses look- group of participants there was in fact some effect of sublimi-
ing at the role of baseline threat bias in anxiety reduction nal ABM on stress vulnerability. This finding is in accord with
as a function of ABM. The theoretical and practical implica- recentlyreportedresultsfromsupraliminalABMinpatientsdiag-
tions of each of the above-listed options are distinct in impor- nosed with general social phobia (Amir et al., 2011). In Amir
tant ways. Next we discuss the implications of each of these et al. (2011), patients in the ABM condition who had greater
possibilities. threatbiasatbaselinedisplayedsignificantlylargerreductionsin
Subliminal ABM could be inherently ineffective in inducing clinician-rated social anxiety symptoms relative to their coun-
change in preconscious attention patterns because the relevant terparts in the placebo condition. ABM did not differ from
neuro-cognitive mechanisms supporting this process might be placebo in patients who did not show threat bias at baseline.
less malleable to change by ABM as compared to processes These findings from Amir et al. (2011) and the present finding
occurring within perceptual awareness. Specifically, subliminal are also in line with the basic rationale for ABM procedures.
threat detection typically involves sub-cortical structures such That is, that pre-treatment threat bias is the target for ABM;
as the amygdala, one of the core components in danger detec- hencetheabsenceofsuchbiasmightrenderABMineffective. A
tion andevaluation(Ledoux, 2000; Amaral,2002), whosefunc- similar rationale was offered in a randomized controlled ABM
tion relies, at least in part, on automatic, rapid-responding study with clinically anxious children, which applied a base-
neural architecture (Ohman and Mineka, 2001; Ohman, 2005; linebiastowardthreatasaninclusioncriterionforparticipation
Adolphs and Spezio, 2006). If preconscious attention patterns (Eldar et al., 2012). The current finding along with previous
are indeed less malleable to change, two important conclu- results highlights the possibility that ABM procedures may be
sions may be derived: practically, there may be no reason to beneficial to a specific sub-group of socially-anxious individu-
investfurthereffortsinsubliminalABMmethods.Theoretically, als characterized by attention bias toward threatening cues at
one may speculate that consistent failure to modify atten- baseline. If proved reliable, such specificity in predicting treat-
tional patterns using subliminal ABM is consistent with the ment efficacy may be ultimately applied to personalize anxiety
notion that modification of threat bias by ABM and the asso- treatment. However, it is important to keep in mind that, at
ciated reduction in anxiety are mediated by later processes of least in the present study, this finding was part of a post-hoc
attention control rather than by automatic attention capture exploration and could be merely incidental. It is also impor-
(Browningetal.,2010;EldarandBar-Haim,2010;Heerenetal., tant to note that attention bias in both threat attenders and
2013). threat avoiders converged toward zero at post-training, thus
A second possible explanation for the current null results couldsimplyreflect regression to themean. Futurestudies may
is that the specific subliminal ABM protocol used here was benefit from designs that specifically and a priori hypothesize
not sufficiently effective to induce the expected change in about the role of baseline threat bias in the clinical response
attention patterns. For example, it might be that trying to toABM.
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|8
Maozetal. SubliminalABMTinsocialanxiety
Despite the discouraging findings with subliminal ABM Considering the small-to-medium effect size of the threat bias
thus far, the potential of this intervention should not be phenomenoningeneral,itisnotuncommonthatasinglestudy
dismissed prematurely. If future studies could substantiate willnotbeabletofindanxiety-relatedattentionbias(Bar-Haim
evidence that subliminal training of threat-related attention etal.,2007). Furthermore, theeffectismanytimes reported for
may have anxiolytic effects, it would point to the poten- between group designs comparing anxious individuals to non-
tial of ABM to target components of threat processing that anxious controls, while in the current study there were only
function outside perceptual awareness. Such specific processes high-anxious participants. One way to probe this shortcoming
may occur independently and within different brain net- is to analyze the results referring to participants who actually
works than processes activated by consciously-perceived threats showed threat bias at baseline (as was done here using post-
(Etkin et al., 2004; Li et al., 2007). Future studies may uti- hoc analyses). While these analyses seem to support the notion
lize brain imaging techniques to directly examine and com- that baseline threat bias may be important for ABM success,
pare the underlying neuro-cognitive mechanisms affected by caution should be taken in the interpretation of these post-hoc
subliminal and supraliminal ABM. If indeed subliminal and findings, particularly those based on the dichotomous split to
supraliminal ABM methods influence different neuro-cognitive attenders and avoiders. This particular analysis relied on very
mechanisms related to anxiety, they may possess additive smallgroupsizesandalsosuffersfromthepossibilitythatsome
therapeutic values, and may prove more efficient if deliv- group members may not truly deviate from zero bias. These
ered as a combined treatment procedure. This possibility concerns may be alleviated to some extent by the supportive
should be determined in future research directly comparing findings relyingon baselineattention biasas acontinuous vari-
subliminal ABM, supraliminal ABM, and a combination of ableintheregression analysesandshouldbeexplored infuture
thetwo. research.
Interpretation of the results of the present study should be Inconclusion,thecurrentstudyismainlyofferingnullresults
considered in light of important limitations. First, the partic- of subliminalABM, as we were unableto showdirect effects of
ipants in the present study were not clinically-diagnosed with subliminal ABM training on attention patterns or anxiety lev-
social phobia, but rather represent a sample of undergraduate els. Nevertheless, we think it may be important for the ABM
students who self-reported high levels of social anxiety. While research community to be exposed to these findings so that
the use of analog populations typically provides an opportu- both an open discussion of the issue could be advanced and
nity to test preliminary treatment-related ideas, in the current future studies could use this failure as a stepping stone for
study it might havealsolimited the potential to detect anxiety- their ABM designs. The current study is the first to report null
related effects of subliminal ABM training. Future studies with results for subliminalABM conducted with an anxious popula-
clinically-diagnosed populations could further test the efficacy tion and it corresponds with recently published null effects of
of subliminal ABM. Second, the present sample may not be supraliminal ABM (Carlbring et al., 2012; Bunnell et al., 2013;
large enough to detect existing effects if these are relatively Neubauer et al., 2013). We hope that researchers would con-
small. It should be noted, however, that the detected effect of tinuetosharebothnull-andpositive-findingsconcerningABM
subliminalABMonsocialstress vulnerabilitywithinthethreat- in order to advance understanding and experimentation in this
attenderssub-groupisquiterobustconsideringthesmallsample field.Inthesamevein,wealsothoughtitisworthwhiletoreport
size.Third,thefactthatattention biastowardsubliminalthreat the post-hoc analyses suggesting that subliminal ABM training
was not significantly different from zero in the current sam- may carry some potential to reduce social stress vulnerability,
ple could be considered a limitation that may have hampered and that baseline threat bias may serve as a marker for such
thepossibilitytodetect attentionalandanxiety-relatedchanges. efficacy.
REFERENCES Amir,N.,Taylor,C.T.,andDonohue, Bar-Haim, Y., Lamy, D., Pergamin, Browning, M., Holmes, E. A., and
Adolphs, R., and Spezio, M. (2006). M.C.(2011).Predictorsofresponse L., Bakermans-Kranenburg, M. J., Harmer, C.J. (2010).The modifi-
Roleoftheamygdalainprocessing to an attention modification pro- andVanIjzendoorn,M.H.(2007). cation of attentional bias to emo-
visualsocialstimuli.Prog.BrainRes. gram in generalized social pho- Threat-related attentional bias in tionalinformation:areviewofthe
156, 363–378. doi: 10.1016/S0079- bia. J. Consult. Clin. Psychol. 79, anxious and nonanxious individu- techniques, mechanisms, and rele-
6123(06)56020-0 533–541.doi:10.1037/a0023808 als: a meta-analytic study. Psychol. vancetoemotionaldisorders.Cogn.
Amaral, D. G. (2002). The primate Amir,N.,Weber,G.,Beard,C.,Bomyea, Bull.133,1–24.doi:10.1037/0033- Affect. Behav. Neurosci. 10, 8–20.
amygdala and the neurobiology J., and Taylor, C. T. (2008). The 2909.133.1.1 doi:10.3758/CABN.10.1.8
of social behavior: implications effect of a single-session attention Bar-Haim,Y.,Morag,I.,andGlickman, Bunnell, B. E., Beidel, D. C., and
for understanding social anxiety. modification programonresponse S. (2011). Training anxious chil- Mesa, F. (2013). A random-
Biol. Psychiatry 51, 11–17. doi: to a public-speaking challenge dren to disengage attention from ized trial of attention training
10.1016/S0006-3223(01)01307-5 in socially anxious individuals. threat: a randomized controlled for generalized social phobia:
Amir, N., Beard, C., Taylor, C. T., J. Abnorm. Psychol. 117, 860–868. trial.J.ChildPsychol.Psychiatry52, does attention training change
Klumpp, H., Elias, J., Bums, M., doi:10.1037/a0013445 861–869. doi: 10.1111/j.1469-7610. social behavior? Behav. Ther. doi:
et al. (2009). Attention training Bar-Haim,Y.(2010).Researchreview: 2011.02368.x 10.1016/j.beth.2013.04.010. [Epub
in individuals with generalized attentionbiasmodification(ABM): Beard,C.(2011).Cognitivebiasmod- aheadofprint].
social phobia: a randomized a novel treatment for anxiety dis- ification for anxiety: current evi- Carlbring, P., Apelstrand, M., Sehlin,
controlled trial. J. Consult. Clin. orders. J. Child Psychol. Psychiatry denceandfuturedirections.Expert H., Amir, N., Rousseau, A.,
Psychol.77,961–973.doi:10.1037/ 51, 859–870. doi: 10.1111/j.1469- Rev. Neurother. 11, 299–311. doi: Hofmann, S. G., et al. (2012).
a0016685 7610.2010.02251.x 10.1586/ern.10.194 Internet-delivered attention bias
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|9
Maozetal. SubliminalABMTinsocialanxiety
modificationtraininginindividuals modification treatment: a meta- Liebowitz,M.R.(1987).Socialphobia. Neubauer, K., Von Auer, M., Murray,
withsocialanxietydisorder–adou- analysis toward the establishment Mod.Probl.Pharmacopsychiatry22, E., Petermann, F., Helbig-Lang, S.,
ble blind randomized controlled of novel treatment for anxiety. 141–173. andGerlach,A.L.(2013).Internet-
trial. BMC Psychiatry 12:66. doi: Biol. Psychiatry 68, 982–990. doi: MacLeod, C., and Hagan, R. (1992). delivered attention modification
10.1186/1471-244X-12-66 10.1016/j.biopsych.2010.07.021 Individual-differences in the training as a treatment for social
Eldar, S., Apter, A., Lotan, D., Edgar, Hallion, L. S., and Ruscio, A. M. selective processing of threaten- phobia: a randomized controlled
K. P., Naim, R., Fox, N. A., et al. (2011).Ameta-analysisoftheeffect ing information, and emotional trial. Behav. Res. Ther. 51, 87–97.
(2012).Attentionbiasmodification of cognitive bias modification on responses to a stressful life event. doi:10.1016/j.brat.2012.10.006
treatment for pediatric anxiety anxiety and depression. Psychol. Behav.Res.Ther.30,151–161.doi: Ohman, A. (2005). The rote of
disorders: a randomized con- Bull. 137, 940–958. doi: 10.1037/ 10.1016/0005-7967(92)90138-7 the amygdala in human fear:
trolled trial.Am.J.Psychiatry 169, a0024355 MacLeod,C.,Rutherford,E.,Campbell, automatic detection of threat.
213–220. doi: 10.1176/appi. Heeren, A., De Raedt, R., Koster, E. L., Ebsworthy, G., and Holker, Psychoneuroendocrinology 30,
ajp.2011.11060886 H. W., and Philippot, P. (2013). L. (2002). Selective attention and 953–958.doi:10.1016/j.psyneuen.20
Eldar, S., and Bar-Haim, Y. (2010). The (neuro)cognitive mechanisms emotional vulnerability: assessing 05.03.019
Neural plasticity in response to behind attention bias modifica- the causal basis of their asso- Ohman, A., and Mineka, S. (2001).
attention training in anxiety. tion in anxiety: proposals based ciation through the experimental Fears, phobias, and preparedness:
Psychol. Med. 40, 667–677. doi: on theoretical accounts of atten- manipulation of attentional bias. toward an evolved module of fear
10.1017/S0033291709990766 tional bias. Front. Hum. Neurosci. J. Abnorm. Psychol. 111, 107–123. and fear learning. Psychol. Rev.
Eldar, S., Ricon, T., and Bar-Haim, 7:119. doi: 10.3389/fnhum.2013. doi:10.1037/0021-843X.111.1.107 108, 483–522. doi: 10.1037/0033-
Y. (2008). Plasticity in atten- 00119 Mathews,A.,andMacLeod,C.(1986). 295X.108.3.483
tion: implications for stress Heeren,A.,Lievens,L.,andPhilippot, Discriminationofthreatcueswith- Ohman,A.,andSoares,J.J.F.(1994).
response in children. Behav. P. (2011). How does attention out awareness in anxiety-states. ‘Unconscious anxiety’: phobic
Res. Ther. 46, 450–461. doi: training work in social phobia: J. Abnorm. Psychol. 95, 131–138. responses to masked stimuli.
10.1016/j.brat.2008.01.012 disengagement from threat or doi:10.1037/0021-843X.95.2.131 J. Abnorm. Psychol. 103, 231–240.
Etkin, A., Klemenhagen, K. C., re-engagement to non-threat? Mathews, A., and MacLeod, C. doi:10.1037/0021-843X.103.2.231
Dudman, J. T., Rogan, M. T., J. Anxiety Disord. 25, 1108–1115. (2002). Induced processing biases Pishyar,R.,Harris,L.M.,andMenzies,
Hen, R., Kandel, E. R., et al. doi:10.1016/j.janxdis.2011.08.001 have causal effects on anxiety. R. G. (2004). Attentional bias for
(2004). Individual differences in Heeren,A.,Reese,H.E.,McNally,R.J., Cogn. Emot. 16, 331–354. doi: words and faces in social anxiety.
trait anxiety predict the response andPhilippot,P.(2012).Attention 10.1080/02699930143000518 AnxietyStressCoping17,23–36.doi:
of the basolateral amygdala to training toward and away from Matsumoto,D.,andEkman,P.(1988). 10.1080/10615800310001601458
unconsciously processed fearful threat in social phobia: effects The Japanese and CaucasianFacial Posner,M.I.(1980).Orientingofatten-
faces. Neuron 44, 1043–1055. doi: on subjective, behavioral, and Expressions of Emotion (JACFEE) tion.Q.J.Exp.Psychol.32,3–25.doi:
10.1016/j.neuron.2004.12.006 physiological measures of anxiety. and Neutrals (JACNeuF) [CDs]. 10.1080/00335558008248231
Fox, E. (2002). Processing emotional Behav. Res. Ther. 50, 30–39. doi: San Francisco, CA: Department Rytwinski, N. K., Fresco, D. M.,
facial expressions: the role of anx- 10.1016/j.brat.2011.10.005 of Psychology, San Francisco State Heimberg, R. G., Coles, M. E.,
iety and awareness. Cogn. Affect. Klumpp, H., and Amir, N. (2010). University. Liebowitz, M. R., Cissell, S., et al.
Behav. Neurosci. 2, 52–63. doi: Preliminary study of attention Mennin, D. S., Fresco, D. M., (2009).Screeningforsocialanxiety
10.3758/CABN.2.1.52 training to threat and neutral Heimberg, R. G., Schneier, F. disorderwiththeself-reportversion
Fox, E., Cahill, S., and Zougkou, K. faces on anxious reactivity to a R., Davies, S. O., and Liebowitz, of the Liebowitz Social Anxiety
(2010). Preconscious process- social stressor in social anxiety. M. R. (2002). Screening for social Scale. Depress. Anxiety 26, 34–38.
ing biases predict emotional Cogn. Ther. Res. 34, 263–271. doi: anxiety disorder in the clinical doi:10.1002/da.20503
reactivity to stress. Biol. 10.1007/s10608-009-9251-0 setting: using the Liebowitz Social Schmidt,N.B.,Richey,J.A.,Buckner,
Psychiatry 67, 371–377. doi: Koster,E.H.W.,Fox,E.,andMacLeod, AnxietyScale.J.AnxietyDisord.16, J. D., and Timpano, K. R. (2009).
10.1016/j.biopsych.2009.11.018 C.(2009).Introductiontothespe- 661–673. doi: 10.1016/S0887-6185 Attention training for gener-
Fox, E., Russo, R., Bowles, R., and cialsectiononcognitivebiasmod- (02)00134-2 alized social anxiety disorder.
Dutton,K.(2001).Dothreatening ification in emotional disorders. Merikle, P. M., Smilek, D., and J.Abnorm.Psychol.118,5–14.doi:
stimuli draw or hold visual atten- J. Abnorm. Psychol. 118, 1–4. doi: Eastwood, J.D.(2001).Perception 10.1037/a0013643
tion in subclinical anxiety? J. Exp. 10.1037/a0014379 without awareness: perspectives See, J., MacLeod, C., and Bridle, R.
Psychol. Gen. 130, 681–700. doi: Ledoux, J. E. (2000). Emotion cir- from cognitive psychology. (2009). The reduction of anxiety
10.1037/0096-3445.130.4.681 cuits in the brain. Annu. Rev. Cognition 79, 115–134. doi: vulnerabilitythroughthemodifica-
Fox, E., Russo, R., and Dutton, Neurosci. 23, 155–184. doi: 10.1016/S0010-0277(00)00126-8 tionofattentionalbias:areal-world
K. (2002). Attentional bias for 10.1146/annurev.neuro.23.1.155 Mogg, K., and Bradley, B. P. (2002). study using a home-based cogni-
threat: evidence for delayed disen- Levin, J. B., Maron, S., Gur, S., Selective orienting of attention to tive bias modification procedure.
gagement from emotional faces. Wechter, D., and Hermesh, H. maskedthreatfacesinsocialanxiety. J.Abnorm.Psychol.118,65–75.doi:
Cogn. Emot. 16, 355–379. doi: (2002). Psychometric properties Behav.Res.Ther.40,1403–1414.doi: 10.1037/a0014377
10.1080/02699930143000527 and three proposed subscales of a 10.1016/S0005-7967(02)00017-7 Spielberger, C. D., Gorsuch, R. L.,
Fresco,D.M.,Coles,M.E.,Heimberg, self-reportversionoftheLiebowitz Mogg, K., Bradley, B. P., Williams, Lushene,R.,Vagg,P.R.,andJacobs,
R. G., Liebowitz, M. R., Hami, Social Anxiety Scale translated R., and Mathews, A. (1993). G.A.(1983).ManualfortheState-
S., Stein, M.B., et al. (2001). The into Hebrew. Depress. Anxiety 16, Subliminal processing of emo- Trait Anxiety Inventory. Palo Alto,
Liebowitz Social Anxiety Scale: 143–151.doi:10.1002/da.10064 tional information in anxiety and CA:ConsultingPsychologistsPress.
a comparison of the psychomet- Li, W., Zinbarg, R. E., and Paller, depression.J.Abnorm.Psychol.102, Stormark, K. M., Nordby, H., and
ric properties of self-report and K. A. (2007). Trait anxiety mod- 304–311. doi: 10.1037/0021-843X. Hugdahl, K. (1995). Attentional
clinician-administered formats. ulates supraliminal and sublim- 102.2.304 shifts to emotionally charged
Psychol. Med. 31, 1025–1035. doi: inal threat: brain potential evi- Mogg, K., Philippot, P., and Bradley, cues: behavioral and ERP data.
10.1017/S0033291701004056 dence for early and late process- B. P. (2004). Selective attention to Cogn. Emot. 9, 507–523. doi:
Hakamata,Y.,Lissek,S.,Bar-Haim,Y., inginfluences.Cogn.Affect. Behav. angryfacesinclinicalsocialphobia. 10.1080/02699939508408978
Britton,J.C.,Fox,N.A.,Leibenluft, Neurosci. 7, 25–36. doi: 10.3758/ J. Abnorm. Psychol. 113, 160–165. Teichman,Y.,andMelinic,H.(1979).
E., et al. (2010). Attention bias CABN.7.1.25 doi:10.1037/0021-843X.113.1.160 STAIandSTAIC–HebrewGuidlines
FrontiersinHumanNeuroscience www.frontiersin.org July2013|Volume7|Article389|10