Mechanisms of attentional biases towards threat in anxiety disorders: An integrative review

Mechanisms of Attentional Biases towards Threat in the AnxietyDisorders: An Integrative Review

Josh M. Cisler1,2,* and Ernst H. W. Koster31University of Arkansas2Medical University of South Carolina3Ghent University, Belgium

AbstractA wealth of research demonstrates attentional biases toward threat in the anxiety disorders. Severalmodels have been advanced to explain these biases in anxiety, yet the mechanisms comprising andmediating the biases remain unclear. In the present article, we review evidence regarding themechanisms of attentional biases through careful examination of the components of attentional bias,the mechanisms underlying these components, and the stage of information processing during whichthe biases occur. Facilitated attention, difficulty in disengagement, and attentional avoidancecomprise the components of attentional bias. A threat detection mechanism likely underlies facilitatedattention, a process that may be neurally centered around the amygdala. Attentional control abilitylikely underlies difficulty in disengagement, emotion regulation goals likely underly attentionalavoidance, and both of these processes may be neurally centered around prefrontal cortexfunctioning. The threat detection mechanism may be a mostly automatic process, attentionalavoidance may be a mostly strategic process, and difficulty in disengagement may be a mixture ofautomatic and strategic processing. Recommendations for future research are discussed.

Keywordsanxiety; attentional bias; information processing bias

Mechanisms of Attentional Biases towards Threat in the Anxiety Disorders:An Integrative Review

An attentional bias towards threat refers to differential attentional allocation towardsthreatening stimuli relative to neutral stimuli (Bar-Haim et al., 2007; MacLeod et al., 1986;Mogg & Bradley, 1998). A wealth of research demonstrates that anxious individuals displayan attentional bias towards threatening sources of information, and this effect is less consistentor typically not observed in non-anxious individuals (Bar-Haim et al., 2007; Mogg & Bradley,1998; Williams et al., 1996). While the attentional bias effect has been systematicallydemonstrated numerous times in anxious populations, the exact mechanisms that comprise and

© 2009 Elsevier Ltd. All rights reserved.*To whom correspondence should be directed: 216 Memorial Hall, Psychology Department,University of Arkansas, 72701,[email protected]'s Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resultingproof before it is published in its final citable form. Please note that during the production process errors may be discovered which couldaffect the content, and all legal disclaimers that apply to the journal pertain.

NIH Public AccessAuthor ManuscriptClin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

Published in final edited form as:Clin Psychol Rev. 2010 March ; 30(2): 203. doi:10.1016/j.cpr.2009.11.003.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

underlie attentional biases remain unexplained. Several theoretical models have been advancedto account for attentional biases towards threat in anxiety; however, these models have notbeen evaluated systematically in light of emerging empirical data. The purpose of this paperis to review the relevant theoretical models and empirical data in order to further illuminatethe mechanisms of attentional biases towards threat in anxiety.

The conceptualization adapted in this paper is that the mechanisms of attentional biases canbest be understood by examining three interrelated aspects of attentional bias: 1) the observedcomponents of attentional bias, 2) the mechanisms that may mediate the expression of thesecomponents, and 3) the stage of information processing during which the mediatingmechanisms operate. The components of attentional bias refer to the observable andmeasurable characteristics of attentional bias (i.e., what does an attentional bias ‘look’ like?).Observable characteristics of attentional bias that have been mentioned in the literature includefacilitated attention (e.g., threat stimuli are detected faster than non-threat stimuli), difficultyin disengagement (e.g., it is harder to disengage attention from a threat stimulus relative to aneutral stimulus), and attentional avoidance (e.g., allocating attention towards locationsopposite the location of threat; Cisler et al., 2009; Fox et al., 2001; 2002; Koster et al.,2004a; 2005; 2006; Mogg et al., 2004). The mediating mechanisms of attentional bias refer tothe underlying mechanisms that may produce the observable characteristics of attentional bias.For example, the amygdala has been implicated as a neural mechanism that may mediateautomatic vigilance for threat (Anderson & Phelps, 2001; Davis & Whalen, 2001; Ohman,1996; 2005; Ohman & Wiens, 2004). Attentional control, the cognitive ability to regulateattentional allocation (Derryberry & Reed, 2002; Eysenck et al., 2007; Posner & Rothbart,2000), has been mentioned as a possible mechanism that may mediate difficulties indisengaging attention from threat (Eysenck et al., 2007). Finally, the components of attentionalbias and their mediating mechanisms may be tied to specific stages of information processing.Information processing is commonly divided into two stages: automatic and strategic(McNally, 1995; 1996; Moors & de Houwer, 2006; Schiffrin & Schneider, 1977), although theboundaries between these two stages are not well defined. Automatic processing generallyrefers to processing that is capacity-free and occurs without intent, control, or awareness,whereas strategic processing generally refers to processing that is intentional, controllable,capacity-limited, and dependent on awareness.

These three different domains are likely interrelated. Stage of information processing (e.g.,automatic versus strategic) may constrain the functioning of an underlying mechanism (e.g.,amygdala/threat detection mechanism may only be observed during automatic stages ofprocessing), and the underlying mechanism may constrain the expression of the observablecharacteristic (e.g., facilitated attention to threat may occur due to activation of the amygdala/threat detection mechanism). Therefore, it is our contention that a theoretical explanation ofattentional biases in anxiety must account for and explain observations in each domain.

We begin this review by summarizing the major sources of empirical evidence for attentionalbiases towards threat in anxiety. We then conduct a more detailed analysis as to the mechanismsthat comprise and underlie this attentional bias effect, paying careful attention to thecomponents, mediating mechanisms, and stages of information processing. This analysisincludes a discussion of previous models, a review and integration of the empirical evidence,an evaluation of the prior models, and proposals for future research based on remainingambiguities.

Brief Overview of Major FindingsAttentional bias towards threat among anxious populations is a relatively robust phenomenon(Bar-Haim et al., 2007; Cisler et al., 2009; Mogg & Bradley, 1998; Williams et al., 1996) ,

Cisler and Koster Page 2

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

with a recent meta-analysis demonstrating an aggregate effect size of d = .45 (Bar-Haim et al.,2007). The following sections review 2 major findings. First, attentional biases are observedacross several different experimental tasks. Second, attentional biases are observed acrossanxiety disorders.

Experimental TasksThe observation that attentional biases have been observed in several different tasks isimportant because it suggests that the phenomenon is not an artifact of particular experimentalprocedures or task confounds. The general concept of an attentional bias predicts that attentionwill be preferentially allocated towards threatening compared to neutral stimuli. Thedemonstration that this effect occurs in several experimental settings suggests generalizabilityand ubiquitousness of the phenomenon. Thus, the wealth of data on attentional biases allowsfor a global prediction that if threatening and neutral stimuli occur together, the attention of ananxious individual will likely be biased towards the threat.

Modified Stroop task—The modified Stroop task (Stroop, 1935) displays different typesof words (e.g., threatening and neutral) in varying colors. The participant is asked to report thecolor while ignoring the semantic content of the word. Heightened response times to report thecolor of threat words compared to neutral words are considered an indication of attentionalbias. Numerous studies have demonstrated attentional biases in the Stroop task in anxiouspopulations (see Bar-Haim et al., 2007; Mogg & Bradley, 1998; Williams et al., 1996). Forexample, McNally and colleagues (1990) compared Stroop response times towards PTSDrelated words, OCD related words, positive words, and neutral words between Vietnam combatveterans with and without PTSD. The results revealed that participants with PTSD had longerresponse times towards PTSD-related words relative to positive, OCD, and neutral words.Participants with PTSD had longer response times towards PTSD words compared toparticipants without PTSD, and the response times of participants without PTSD did not differacross word types.

Dot probe task—Although the modified stroop task is the most commonly used task tomeasure attentional bias in anxiety, it may not be an ideal measure of attention due to severalinterpretational difficulties. For instance, delayed responding to threat words can be due toenhanced attention as well as overall delayed responding to threat (Algom et al., 2004).Moreover, the modified Stroop task does not allow for the measurement of spatial attentionallocation (e.g., MacLeod, Mathews, & Tata, 1986). To address these limitations, other taskshave been used to assess attentional biases towards threat. The dot probe task (MacLeod et al.,1986) displays two words on a computer screen with one at the top and one at the bottom(alternatively, the words may appear on the left and right side of the screen). Following a briefstimulus presentation duration (e.g., 500 ms), the stimuli disappear and a probe appears in alocation previously occupied by one of the stimuli. The participant is asked to press a buttonindicating whether the top or bottom stimulus had been replaced by the probe. Attentionalbiases are inferred from different response times towards probes that replace threatening stimuli(i.e., congruent trials) compared to probes that replace neutral stimuli (i.e. incongruent trials).If an individual's attention is systematically drawn to the threat stimulus, response times willbe shorter for probes that replace threatening stimuli compared to probes that replace neutralstimuli. A wealth of research has demonstrated attentional biases in the dot probe task (e.g.,Bar-Haim et al., 2007; Mogg & Bradley, 1998). For example, Bradley and colleagues (1999)found that individuals with GAD demonstrated faster response times towards probes thatreplaced threatening faces relative to probes that replaced neutral faces, and this effect amongGAD individuals was greater compared to non-anxious controls.

Cisler and Koster Page 3

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Visual search task—The visual search task (e.g., Öhman, Flykt, & Esteves, 2001; Rinck,Becker, Kellermann, & Roth, 2003) is another experimental task of attentional biases thatallows for the assessment of spatial attentional allocation. Participants are asked to detect atarget stimulus that is embedded in a matrix of distracting stimuli. For example, the target word“spider” might be displayed in a matrix (e.g., 3 row × 3 column pattern of stimulus presentation)of neutral distracter words. Conversely, a neutral target word may be embedded in a matrix ofspider related words. Attentional biases are inferred from faster response times to detect athreatening stimulus in a matrix of neutral stimuli relative to response times to detect neutralstimuli in neutral matrices (i.e., the individual's attention is drawn to the threat stimulus).Attentional biases can also be inferred from slower response times to detect neutral stimuli ina matrix of threatening stimuli relative to response times to detect neutral stimuli in a matrixof neutral stimuli (i.e., the individual's attention is captured by the threat stimulus). The visualsearch task has also robustly demonstrated attentional biases (Cisler et al., 2009; Rinck, Becker,Kellermann, & Roth, 2003; Öhman et al., 2001; Miltner, Krieschel, Hecht, Tripp, & Weiss,2004). For example, Rinck and colleagues (2005) found attentional biases towards spiderstimuli relative to beetle and butterfly stimuli among spider fearful individuals in threeexperiments using visual search tasks, and these biases were greater relative to non-anxiouscontrols.

Spatial cueing task—The spatial cueing task (Fox et al., 2001; Posner, 1980) also allowsfor the assessment of spatial attention allocation. Participants focus on a fixation point locatedbetween two rectangles. A cue is then presented (e.g., one of the rectangles brightens or athreatening stimulus appears in one of the rectangles), followed by the appearance of a targetin one of the two rectangles. Participants are asked to press a key indicating the rectangle inwhich the target is located. Some of the trials are valid cues (the cue draws attention to therectangle in which the target is located), some of the trials are invalid cues (the cue drawsattention away from the rectangle in which the target is located). Attentional biases areindicated by faster responses on valid threat-cued trials relative to neutral-cued trials.Attentional biases are also indicated by slower responses on invalidly threat-cued trials relativeto neutral-cued trials. The spatial cueing task has also revealed a robust attentional bias effect(Bar-Haim et al., 2007, Cisler et al., 2009).

Anxiety DiagnosisA second main finding in attentional bias research is that the biases occur in all anxietydisorders. Attentional biases have been found in GAD (Bradley, Mogg, White, Groom, & deBono, 1999; Bradley, Mogg, Millar, & White, 1995; Mogg et al., 1993; Rinck et al., 2003),social phobia (Amir et al., 2003; Becker, Rinck, Margraf, & Roth, 2001), PTSD (Bryant &Harvey, 1995; McNally et al., 1990a), specific phobia (Ohman et al., 2001; Rinck et al.,2005), panic disorder (Buckley, Blanchard, & Hickling, 2002; Horenstein & Segui, 1997), andOCD (Amir et al., in press; Cisler & Olatunji, in press; Foa et al., 1993; Tata et al., 1996). Inthe case of OCD, Moritz and colleagues (Moritz et al., 2008; Moritz & von Muhlenen, 2008)have recently found null results for an attentional bias effect in OCD. However, these nullresults are countered, and perhaps explained, by a recent study by Amir and colleagues (inpress). These authors found that individuals with OCD display an attentional bias towardsideographically displayed stimuli in the dot probe task in the first third of the experimentaltrials. In the remainder of the trials the attentional bias effect diminishes. These findings, ifreplicated, suggest that the bias in OCD can be strategically overridden with sufficient practiceand possibly implicate deficits in inhibition (cf. Chamberlain et al., 2005). In any case, a recentmeta-analysis found that the attentional bias effect appears to occur in equal magnitude in alldisorders (Bar-Haim et al., 2007). Accordingly, attentional biases do not appear to be a featureof any specific anxiety disorders, but likely appear to be a component of elevated trait anxietyper se. This finding is theoretically relevant because it necessitates an explanation as to why

Cisler and Koster Page 4

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

attentional biases are robustly related to elevated trait anxiety. That is, any theory ofpathological anxiety must account for the observation that attentional biases towards threat co-occurs with chronic anxiety. Further, the observation that attentional biases are a componentof elevated trait anxiety per se suggests that understanding attentional biases may strengthenour understanding of the processes of anxiety.

SummaryNumerous data demonstrate that attention is biased towards threatening information amonganxious individuals. Bar-Haim and colleagues (2007) recently concluded from a meta-analysisof attentional biases that “with over 150 studies that have established the existence and typicalmagnitude of the threat-related bias in anxious individuals from different populations and witha variety of experimental conditions, it appears as if little will be gained from additional studiesof threat-related bias unless these are strongly driven by theory” (pg. 18). Accordingly, thenext step in attentional bias research is to elucidate how and why attention is biased towardsthreat in anxious individuals.

Previous Theoretical Models of Attentional Biases towards Threat in AnxietyIn the past 20 years several models have been put forward to account for the mechanismsunderlying attentional bias in anxiety. The review of these models will focus on the mechanismsof attentional bias postulated by the models. It is important to note that these models havemainly focused on individual differences in trait anxiety. Spielberger, Gorsuch, Lushene, Vagg,and Jacobs (1983) described trait anxiety as a personality trait that predisposes an individualto respond with anxiety to stressful and novel situations. Provided that the magnitude ofattentional bias does not differ between high trait anxious and clinically anxious individuals(Bar-Haim et al., 2007) and the observation that trait anxiety is an important predisposition todevelop clinical anxiety (Barlow, 2002), it seems safe to assume that the mechanismspostulated in the context of trait anxiety are also useful in understanding effects related toclinical anxiety.

Beck and Clark's cognitive modelBeck and Clark's (1997) model suggests that anxiety is characterized by biases at (1) the initialregistration of a threat stimulus; (2) the activation of a primal threat mode; and (3) the secondaryactivation of more elaborative and reflective modes of thinking. The first mode is involvedwith the automatic orienting to threat, which is largely stimulus-driven. Recognition ofpersonally relevant, negative information subsequently leads to the activation of cognitive,affective, psycho-physiological, and behavioral responses that are innate and, in the case ofanxiety, function to obtain safety and minimize threat (referred to as a “primal mode”). Theseresponses tend to be rigid and inflexible because they were developed to maximize the chancesof survival. At this stage, attention is captured by threat and actions involve coping with threat.The final stage in this model is the process of secondary elaboration where there is slow,detailed, effortful, and schema-driven processing of threat.

Williams, Watts, MacLeod & Mathews (1988) modelWilliams et al. (1988) presented an influential account of information processing biases inanxiety and depression. In their model attentional biases are postulated at the preconsciouslevel. They postulate that the threat value of incoming stimuli is determined by an affectivedecision mechanism (ADM). This system produces an initial decision whether information ishigh or low threatening and the output of this system is moderated by state anxiety. If stimulusinput is appraised as highly threatening, a resource allocation mechanism (RAM) is activated.When the RAM is triggered, attentional resources will be allocated to threat. When stimulusinput is determined as low threatening, attention will be maintained to the task at hand and the

Cisler and Koster Page 5

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

new stimulus input will not be attended to. According to Williams et al. (1988), trait anxietymodulates the RAM: HTA individuals will allocate attention to threat (which results in anattentional bias), whereas LTA individuals will ignore threatening information. This may leadto the favorable attentive processing of threat at preconscious and conscious levels in HTAindividuals.

The Williams et al. model has inspired thorough empirical investigation into the nature ofattentional bias. However, some of these ideas about threat processing in LTA individuals areproblematic (Mogg & Bradley, 1998). Williams et al. propose that the RAM directs attentionalresources away from threat in LTA individuals. This may be true for minor threat, but it seemsimplausible that severe threat will not attract attention regardless of an individual's anxietylevel. Persistent attentional avoidance from severe threat would obviously interfere withadequate responding to threat. Indeed, Wilson and MacLeod (2003) recently demonstrated thatLTA individuals display attentional biases for severely threatening, but not moderatelythreatening stimuli, whereas HTA individuals display attentional biases for both magnitudesof threat. Despite this limitation, the model has strongly influenced contemporary models ofattentional bias to threat.

Öhman's (1996) feature detection modelThe work on attention to threat by Arne Öhman (1996; 2005; Öhman & Wiens, 2004) hasgenerally focused on attention to threat as an evolutionary adaptive process, with particularemphasis on the unconscious processing of threat. In his account, stimulus input is analyzedin a feature detection system. Biologically prepared or high intense stimuli can exert a directinfluence on the arousal system through this feature detection system without any consciousmediation and facilitate attentional allocation to threat. It is worth noting that the featuredetection system can explain biases towards pictorial stimuli, but not lexical stimuli, which donot have biologically relevant features. According to Öhman's model, when information haspassed the feature detection system, it enters a significance evaluation system. Threatening orrelevant information then enters the conscious perception system, which allows a slower,conscious appraisal of meaning through interaction with the emotional memories stored in theexpectancy system. This slower, conscious processing route can also influence the arousalsystem if information is appraised as threatening. In this model, feedback loops are postulatedbetween the autonomic arousal system and the significance evaluation system: heightenedarousal further sensitizes the significance evaluation. Furthermore, the expectancy system mayalso sensitize the significance evaluation for specific stimuli due to prior learning.

Wells and Matthew's modelA markedly different view on attentional bias has been put forward by Wells and Matthews(1994). They strongly argued against the idea of an automatic attentional bias and the viewthat attentional bias is a computational accident. Wells and Matthews emphasized the role oftop-down processes and framed their model within a general self-regulatory executivefunction model. In their view, attentional bias to threat is related to self-knowledge, withvoluntary goals and beliefs of the individual guiding attention to threat. Wells and Mathewsargue that anxiety is characterized by a consciously perceived threat to self-perseverance andthat this motivational state is associated with the monitoring of threat. They propose that theattentional bias effect observed in cognitive-experimental tasks is caused by the belief that itis important to monitor threat. Indeed, Matthews and Wells (2000) discussed limitations ofstudies demonstrating attentional bias without conscious awareness and speculated that underseveral conditions attentional bias may occur without awareness due to the operations ofvoluntary strategies.

Cisler and Koster Page 6

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Mogg and Bradley's (1998) cognitive-motivational modelIn the cognitive-motivational model, attention to threat is understood as a normal and adaptivemechanism. They draw strongly on the neurobiological work of LeDoux (1996) whodemonstrated that threat can be processed through two neural pathways: (1) a fast and crudeanalysis of stimulus features related to previously encountered threat, and (2) a slower, moredetailed analysis of stimulus input, contextual information, and information stored in memory.In Mogg and Bradley's model, attention to threat is determined by two systems. First, a valenceevaluation system is responsible for the initial, preconscious appraisal of stimuli. Output fromthis system is also dependent upon contextual information, prior learning, and state anxietylevel. Trait anxiety influences the reactivity of the valence evaluation system to threat, with aheightened sensitivity to threat in HTA individuals. In HTA individuals, mild threat cues aremore readily appraised as high threat than in LTA individuals. Second, output from the valenceevaluation system feeds into a goal-engagement system which determines the allocation ofprocessing resources. If a stimulus is tagged as highly threatening, current behavior will beinterrupted and attention will be allocated to the stimulus input. If stimulus input is tagged aslow threatening, further processing of this stimulus will be inhibited, attention will bemaintained at ongoing tasks, and current behavior will not be interrupted. According to thisview, differential attention to threat in HTA and LTA individuals can be expected for mildthreat but not for severe threat. Due to an oversensitive valence evaluation system, HTAindividuals have the tendency to appraise mild and ambiguously threatening information ashighly threatening and will attend to this information. In LTA individuals this information willbe ignored as it will generally be appraised as low threatening. One initial difficulty with thistheory was that there was little evidence for an overly sensitive valence evaluation mechanismapart from the observations the mechanism was used to explain.

Mathews and Mackintosh's (1998) modelAnother model that shares many assumptions with the cognitive-motivational model (Mogg& Bradley, 1998), is the cognitive model developed by Mathews and colleagues (Mathews &Mackintosh, 1998; Mathews, Mackintosh, & Fulcher, 1997). In line with the ideas put forwardby Williams et al. (1997), attentional bias is only predicted when threat has to compete withother stimuli or task-demands. A threat evaluation system (TES) is proposed that shares manycharacteristics with the ADM of the Williams et al (1997) model. Stimulus input isautomatically evaluated and output of this system feeds into a distracter/threat representationsystem. The interference caused by the distraction representation is countered up to a certainlevel by voluntary effort aimed at attending targets from the task at hand and strengtheningtheir representations. In accounting for attentional bias, it is postulated that the output of theTES is strengthened by anxiety level. More specifically, stimulus input needs to exceed acertain threshold before output will flow from the TES into the distraction representation. Aheightened anxiety level lowers the threshold value from the TES and causes an increasedoutput of this system. This model further proposes that strong danger cues will attract attentionin everyone, whereas weak danger cues will only do so in individuals with a heightened anxietylevel.

Eysenck and colleagues (2007) attentional control theoryAttentional control theory posits that anxiety disrupts two central executive functions relatedto attentional control: inhibition and shifting. Inhibition refers to the ability to inhibit or regulatedominant or automatic responses. Shifting refers to the adaptive ability to shift attentionbetween tasks depending on context. Eysenck and colleagues discuss these functions in termsof top-down and bottom-up processing. Anxiety impairs inhibition in that anxiety weakens thedegree to which inhibitory mechanisms can regulate automatic responses; that is, anxietyweakens top-down regulatory control. One manner in which this effect may manifest is in

Cisler and Koster Page 7

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

difficulty disengaging attention from distracting threat stimuli. Anxiety potentiates shifting inthat anxiety heightens the degree to which attention is shifted from one task to another; that is,anxiety increases stimulus-driven bottom-up processing. One manner in which this effectmanifests is in the facilitated detection of threat stimuli.

Bar-Haim and colleagues (2007) modelBar-Haim and colleagues suggest a multidimensional model of attentional biases in whichdysfunction of any sequence in the temporal chain may result in attentional bias. First, apreattentive threat evaluation system evaluates environmental stimuli. Threatening stimulifeeds into a resource allocation system and elicits physiological arousal and allocation ofcognitive resources onto the stimuli. A guided threat evaluation system may then assess thecontext of the threat and assess available coping resources. If this system deems the threat lowin significance, the input of the preattentive threat evaluation system may be overrided by agoal engagement system. If the guided threat evaluation system deems the threat high insignificance, attention may be maintained on threat and a high state of anxiety may result.

SummaryIt is obviously difficulty to evaluate these models in relation to the components of attentionalbias, the mediating mechanisms, and stages of processing, given that the models were not allspecifically developed to explain all of these aspects. Further, some of the models weredeveloped prior to research being developed in certain attentional bias aspects (e.g., beforedifficulty in disengagement had been distinguished from facilitated attention). With thatlimitation explicitly stated, a summary of these models in regards to their likely predictionsabout the attentional bias components, mediating mechanisms, and stage of processing isprovided in Table 1. These models all predict the component of vigilance/facilitated attentiontowards threat. The component of difficulty disengaging attention from threat is less consistentacross the models, with only two models (Beck & Clark, 1997;Eysenck, 2007) accounting forand explaining this phenomenon. Attentional avoidance of threat is similarly less representedin the models and only accounted for and explained by one model (Williams et al., 1988).

In regards to the mechanisms that mediate attentional biases, most models posit some sort ofthreat detection mechanism responsible for detecting and orienting attention towardsthreatening stimuli (Bar-Haim et al., 2007; Beck & Clark, 1997; Eysenck et al., 2007; Ohman,1996; Mogg & Bradley, 1998; Mathews & Mackintosh, 1998; Williams et al., 1988). OnlyWells and Mathews (1994) model does not posit an automatic threat detection mechanism.There is great theoretical discord in regard to other possible mechanisms. Some models posita resource allocation mechanism that directs the use of available cognitive resources (Bar-Haim et al., 2007; Mogg & Bradley, 1998; Williams et al., 1988). Some models posit a threatelaboration mechanism, in which strategic processing evaluates the identified threat as eithermajor or minor (Bar-Haim et al., 2007; Beck & Clark, 1997). Some models posit a strategicgoal engagement mechanism, in which the individuals’ goals, beliefs, voluntary effort, orschematic processing can either maintain or override attention to threat (Bar-Haim et al.,2007; Eysenck et al., 2007; Mathews & Mackintosh, 1998; Wells & Mathews, 1994). The stageof information processing in which attentional biases occur is also inconsistently predictedacross the models, with most models positing both stages (Bar-Haim et al., 2007; Beck & Clark,1997; Ohman, 1996; Mathews & Mackintosh, 1998), some models positing only automaticprocessing (Mogg & Bradley, 1998; Williams et al., 1988), and one model positing onlystrategic processing (Wells & Matthews, 1994).

In regards to interrelations between the different aspects, most models generally link anautomatic threat detection mechanism with facilitated attention for threat (Bar-Haim et al.,2007; Beck & Clark, 1997; Mathews & Mackintosh, 1998; Mogg & Bradley, 1998; Ohman,

Cisler and Koster Page 8

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

1996; Williams et al., 1988). One model links difficulty in disengagement with a strategic threatelaboration mechanism (Beck & Clark, 1997), one model links difficulty in disengagement toa guided threat evaluation system (Bar-Haim et al., 2007), and one model links difficulty indisengagement to disruption of attentional control (Eysenck et al., 2007).

In sum, there is little theoretical agreement as to what components attentional biases have, whatmechanisms mediate these components, and during what stage of processing the mechanismsoperate. The only consistent prediction across the models is that a threat detection mechanismoperates at the automatic stage of processing and underlies facilitated attention to threat. Thisgeneral lack of agreement makes it difficult to understand why attention is biased towardsthreat in anxious individuals.

Review of Empirical Evidence for the Mechanisms of Attentional BiasGiven the theoretical discord, we now review the evidence in regards to the components ofattentional bias, mediating mechanisms, and stages of processing.

Components of Attentional BiasOne of the emerging lines of inquiry in attentional bias in anxiety disorder research is theinvestigation of the components of attentional bias (Cisler et al., 2009; Fox et al., 2002; 2001;Koster et al., 2004a). Attentional biases towards threat may be comprised of facilitated attentionto threat, difficulty disengaging attention away from threat, or attentional avoidance of threat.Facilitated attention refers to the relative ease or speed with which attention is drawn to a threatstimulus (i.e., attentional orienting). Difficulty in disengaging refers to the degree to which athreat stimulus captures attention and impairs switching attention from the threat to anotherstimulus. Attentional avoidance refers to a more recent empirical phenomenon in whichattention is preferentially allocated towards locations opposite the location of the threat cue,thus indicating avoidance of the threat cue (e.g., Koster et al., 2005; Mogg et al., 2004).Measurement of these components necessitates a task that can differentiate the components.Three tasks that have been used in the literature for this purpose include the spatial cueing,visual search, and dot probe tasks.

Delayed disengagement from threat—Research using the spatial cueing task hasinvariably demonstrated difficulty in disengagement among anxious individuals (Amir, Elias,Klumpp, & Przeworski, 2003; Cisler & Olatunji, in press; Fox, Russo, Bowles, & Dutton,2001; Fox, Russo, & Dutton, 2002, experiment 1; Koster, Crombez, Verschuere, Van Damme,& Wiersema, 2006; Koster et al., 2005; 2004; Van Damme et al., 2006; Yiend & Mathews,2001). Research using the visual search task has almost invariably demonstrated difficulty indisengagement among anxious individuals (Byrne & Eysenck, 1995; Gilboa-Schechtman, Foa,& Amir, 1999; Juth, Lundqvist, Karlsson, & Öhman, 2005, experiment 5; Lipp & Waters,2007; Miltner, Krieschel, Hecht, Tripp, & Weiss, 2004; Rinck, Becker, Kellermann, & Roth,2003; Rinck, Reinecke, Ellwart, Heuer, & Becker, 2005), with only one exception (Pflugshauptet al., 2005). The dot probe task methodology has only recently been improved to disentanglethe effects of facilitated attention and difficulty in disengagement (Koster et al., 2004a), andthe evidence thus far has almost invariably demonstrated difficulty in disengagement amonganxious individuals (Koster et al., 2004a; Koster et al., 2006b; Salemnik et al., 2007), with onefailure to find disengagement (Carlson & Reinke, 2008, though this study only used maskedstimuli). Accordingly, there is considerably strong evidence that attentional biases towardsthreat are comprised of a difficulty in disengaging attention from threat stimuli. Mogg andcolleagues (2008) have recently argued that the spatial cueing task may not provideunambiguous evidence for delayed disengagement, as there could be a confound betweendelayed disengagement and a generic slowdown effect caused by presentation of threat (e.g.,Algom et al., 2004). Although this indeed is a problem in this task, the delayed disengagement

Cisler and Koster Page 9

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

hypothesis is also corroborated by results obtained in the dot probe and visual search tasks.Further, Cisler and Olatunji (in press) recently found that the relation between elevatedcontamination fear and difficulty in disengagement in the spatial cueing task remained whenstatistically controlling for generic response slowing, suggesting that this task confound doesnot explain the difficulty in disengagement effect. However, future research in this area isneeded among other anxious populations.

Facilitated attention for threat—The evidence for facilitated attention among anxiousindividuals appears mixed at first glance, but important moderating variables may affect theobservation of facilitated attention. Some studies using the spatial cueing task have failed todemonstrate evidence for facilitated attention among anxious individuals (Amir, Elias,Klumpp, & Przeworski, 2003; Fox, Russo, Bowles, & Dutton, 2001; Fox, Russo, & Dutton,2002, experiment 1; Yiend & Mathews, 2001). However, Koster and colleagues (2006a) foundthat at 100 ms stimulus durations, facilitated attention was found towards highly, but not mildly,threatening pictures among high trait anxious individuals. At longer presentation times therewas again no evidence of facilitated attention. Research using dot probe methodologies thatdisentangle facilitated attention from difficulty in disengagement have only found evidencefor difficulty in disengagement towards supraliminally presented stimuli (Koster et al.,2004a; Koster et al., 2006b; Salemnik et al., 2007), while facilitated attention has been foundtowards subliminally presented stimuli (Carlson & Reinke, 2008). Accordingly, stimulusduration (i.e., quick stimulus presentation) may moderate the occurrence of facilitated attentionamong anxious individuals. One prior study (Fox et al., 2001; experiment 2) failed to findfacilitated attention at 100 ms, but this study did not also manipulate threat intensity. Furtherdemonstrating the importance of stimulus intensity, three studies demonstrate that neutralstimuli paired with aversive stimuli (e.g., loud noise bursts, shock) in classical conditioningparadigms elicit facilitated attention in the spatial cueing task (Koster et al., 2005; 2004b; VanDamme et al., 2006). These latter data suggest that stimuli predicting the occurrence ofimminent threat elicit facilitated attention. Accordingly, it appears to be the case that facilitatedattention towards threat is moderated by threat intensity (i.e., highly threatening stimuli) andstimulus duration (i.e., 100 ms or less).

Some studies using the visual search task have also failed to find evidence for facilitatedattention (Rinck, Becker, Kellermann, & Roth, 2003; experiment 1; Rinck, Reinecke, Ellwart,Heuer, & Becker, 2005; experiment 1). However, these studies are contrasted by several othervisual search studies that have documented facilitated attention among anxious individuals(Byrne & Eysenck, 1995; Gilboa-Schechtman, Foa, & Amir, 1999; Juth, Lundqvist, Karlsson,& Öhman, 2005, experiment 5; Miltner, Krieschel, Hecht, Tripp, & Weiss, 2004; Rinck et al.,2003; experiment 2; Rinck et al., 2005, experiments 2 and 3). Research using visual searchtasks have not yet manipulated stimulus duration or threat intensity, so it remains to be seenwhether these variables also explain the inconsistent results in visual search tasks.

Attentional avoidance—Finally, several studies have documented attentional avoidanceamong anxious individuals. Koster and colleagues (2005) found the standard congruency effectin the dot probe (i.e., reaction times on congruent trials were shorter than on incongruent trials)among high trait anxious participants when threat pictures were presented for 500 ms.However, at stimulus durations of 1250 ms, high trait anxious participants demonstratedattentional avoidance of the threat cues, indicated by longer reaction times on congruent trialscompared to incongruent trials. This basic effect, that attentional avoidance is observed at long,but not short or intermediate stimulus presentation durations, has been replicated (Koster etal., 2006; Mogg et al., 2004). Moreover, attentional avoidance of threat has been found whendirectly measuring eye fixations (Garner et al., 2006; Calvo & Avero, 2005; Rohner, 2002;Pflugshaupt et al., 2005). For example, Pflugshaupt and colleagues (2005) found that spiderphobic individuals initially demonstrated rapid eye movement fixations onto a spider stimulus

Cisler and Koster Page 10

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

(i.e., facilitated attention), but they subsequently demonstrated eye movement fixations awayfrom the spider stimulus (i.e., attentional avoidance). Demonstrating robustness, the attentionalavoidance effect at long stimulus durations has been found in the dot probe task (Garner et al.,2006; Koster et al., 2005; Mogg et al., 2004), exogeneous cueing task (Koster et al., 2006), andvisual search tasks (Pflugshaupt et al., 2005). However, there have been some studies usingsimilar methodologies that have failed to find attentional avoidance among anxious individualsat long (e.g., 1500 ms) stimulus durations (Bradley et al., 1998; Mogg et al., 1997). Futureresearch is still needed in this area to clarify the inconsistencies and investigate potentialmoderators (e.g., threat intensity).

Mediating MechanismsAttentional control—One of the relevant areas of emerging research is investigating therole of attentional control in attentional biases (Derryberry & Reed, 2002; Eysenck et al.,2007; Mathews & Wells, 2000). Attentional control is an individual difference variable thatrefers to individuals’ ability to regulate their attentional allocation. This can be construed as a‘top-down’ regulatory ability (Posner & Rothbart, 2000), such that it inhibits the ‘bottom-up’influence of emotional distracters (Eysenck et al., 2007). Derryberry and Reed (2002) foundthat trait anxious individuals with poor attentional control, which was measured via self-report,displayed difficulty disengaging attention from threat at 250 ms stimulus duration as well as500 ms stimulus durations. In contrast, trait anxious individuals with good attentional controldemonstrated difficulty disengaging from threat at 250 ms stimulus duration but not at 500 msstimulus duration. Peers and Lawrence (2009) also found that individuals with poor attentionalcontrol demonstrated difficulties disengaging attention from rapidly presented emotional faces(i.e., 100 ms) during a rapid serial visual presentation task (see Anderson, 2005; Raymond etal., 1992), whereas individuals with good attentional control displayed no disengagementdifficulties.

Recent research has similarly demonstrated that effortful control moderates the relationbetween negative affectivity and attentional bias towards threat in the dot probe task amongchildren and adolescents (Lonigan & Vasey, 2009), such that only youth high in negative affectand low in effortful control displayed attentional biases towards threat displayed for 1250 ms.However, this study used the original dot probe methodology (Macleod et al., 1986); thus,difficulty in disengagement could not specifically be demonstrated. Effortful control refers toindividual differences in the ability to engage executive processes to override dominantresponses (Posner & Rothbart, 2000), and is theorized to be linked with attentional control(Derryberry & Rothbart, 1997). This body of data converges in demonstrating that the abilityto regulate attention allocation may moderate the degree to which attention can be disengagedfrom threatening stimuli. Accordingly, attentional control may be a higher order regulatorymechanism controlling the characteristics of attentional biases towards threat, specificallydisengagement from threat.

Emotion regulation strategy/goals—Emerging research also demonstrates that emotionregulation strategies may moderate attentional biases towards threat. Emotion regulation refersto “the processes by which individuals influence which emotions they have, when they havethem, and how they experience and express these emotions” (Gross, 1998a, pg. 275).Attentional allocation has been strongly proposed as one mechanism of emotion regulation(Gross, 1998; 2001; 2007; Koole, 2009). For example, an individual may allocate attentiononto a distracting poster on the wall while receiving an injection at the doctor's office in orderto reduce negative affect during the injection. Purposeful attentional allocation towards neutralrelative to unpleasant stimuli is akin to the concept of ‘distraction’ (van Dillen & Koole,2007; Sheppes & Meiran, 2008).

Cisler and Koster Page 11

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Johnson (2009) engaged participants in a dot probe task displaying pairs of angry and happyfaces and instructed participants to either focus attention onto the happy faces or provided noinstruction. Stimuli were presented for 17 ms, 500 ms, and 1250 ms. Participants were alsoengaged in difficult anagram tasks before and after the dot probe task. Results demonstratedthat participants given the goal to attend to happy faces demonstrated attentional biases towardsthe happy faces at only stimulus durations of 1250 ms, whereas this was not true of theparticipants given no instruction. Participants instructed to attend to happy faces demonstratedless frustration during the anagram tasks only after the dot probe tasks relative to participantsgiven no instruction. Further, the bias towards attending to happy faces at 1250 ms predictedhow many seconds participants engaged in the second anagram task before giving up only inparticipants instructed to attend to happy faces. Accordingly, this study demonstrates that agoal of attending to positive stimuli leads to more attention towards positive stimuli duringstrategic stages of processing, less frustration during difficult tasks, and better emotionregulation. Dunning and Hajack (2009) similarly found that participants displayed elevatedlate positive potential, a brain electrical activity indicator of increased attention tomotivationally salient stimuli, while passively viewing unpleasant pictures. In contrast,participants directed to allocate their attention onto less distressing aspects of the unpleasantpictures displayed decreased late positive potential. Distraction has also been found to producereduced memory for the threatening stimulus (Sheppes & Meiran, 2009). These datademonstrate that emotion regulation goals may moderate the components of attentional biasesat late stages of processing and that attentional avoidance can indeed regulate emotion. Thesetwo observations suggest that attentional avoidance among anxious participants (Koster et al.,2005; 2006; Mogg et al., 2004; Pflugshaupt et al., 2005) may occur because these participantsare attempting to strategically regulate negative affect via distraction.

Attentional control ability and emotion regulation strategies appear to both be moderators ofattentional biases; however, they are likely separate mechanisms that moderate attentionalbiases in distinct manners. Attentional control can be construed as a regulatory ability: peoplewith better attentional control can disengage attention from threatening stimuli, whereas peoplewith poor attentional control may demonstrate difficulties disengaging attention fromthreatening stimuli. In contrast, emotion regulation may not be an ability per se, but may alsoreflect the individual's strategy for coping with negative emotion. Although researchdemonstrates that some strategies of emotion regulation (e.g., re-appraisal) are more acutelyeffective than others (e.g., suppression; Feldner et al., 2003; Gross & Levenson, 1993; 1997;Gross, 1998b; Moore et al., 2008), an individual's strategy for coping with negative emotionlikely varies considerably given the context (Gross, 1998a). For example, an individual mayexpress facial expressions of negative emotion while being yelled at by a loved one, but thesame individual may suppress facial expressions of negative emotion while being yelled at bytheir boss. Accordingly, context may moderate an individual's online emotion regulation goals/strategies, which in turn moderates whether and to what degree attention will be deployed toa threatening stimulus. It is important to note that attentional control and emotion regulationare likely related, such that whether one can effectively regulate emotions may depend onhigher-order regulatory mechanisms, such as attentional control. Thus, interactions betweenthese mechanisms might be expected.

Neural mechanisms—A wealth of data demonstrate that the amygdala, a brain structurelocated in the temporal lobes, is critically involved in the processing of fear-related informationand expression of fear-related behavior (Davis, 2006; Davis & Whalen, 2001; LeDoux,2000; Myers & Davis, 2007; Rosen, 2004). Relevant to the current topic, enhanced amygdalaactivity is likely a neural mechanism involved in automatic vigilance/facilitated attention forthreat. Carlson and colleagues (2009) recently demonstrated in the dot probe task that maskedcongruent trials were associated with increased amygdala activity, providing strong evidencethat the amygdala is involved in automatic facilitated attention to threat. Similarly, research

Cisler and Koster Page 12

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

demonstrates correlations between the amygdala and attentional biases towards threat (e.g.,Monk et al., 2004; Monk et al., 2008; van den heuvel et al., 2005; Anderson & Phelps, 2001).For example, van den Heuvel and colleagues (2005) found that attentional biases towardspanic-related words among individuals with panic disorder predicted amygdala activity duringfMRI scanning. Moreover, Anderson and Phelps (2001) found that an individual with bilateralamygdala lesions did not display attentional biases towards threat, whereas individuals withunilateral amygdala lesions did display attentional biases towards threat. Second, researchdemonstrates that masked and/or unattended stimuli elicit amygdala activation (Dolan &Vuilleumier, 2003; Morris et al., 1998; Vuilleumier, Armony, Driver, & Dolan, 2001; Whalenet al., 1998; 2004), which is analogous to the findings that masked stimuli elicit attentionalbiases (Bar-Haim et al., 2007; Mogg et al., 1993). For example, Whalen and colleagues(1998) found greater activation of the amygdala during masked presentations of fearful facescompared to masked happy faces. Further, Whalen and colleagues (2004) later found greateractivation of the amygdala to masked images of fearful eyes (i.e., eyes with an enlargedwhitened area) relative to masked images of normal eyes, demonstrating that the amygdalamay respond automatically to specific fear-relevant features.

These lines of evidence converge in suggesting that the amygdala is critically involved inautomatic vigilance for threat (Davis & Whalen, 2001; Ohman, 2005). However, it is importantto note that automatic activation of the amygdala towards threat may depend on the availabilityof attentional resources (Pessoa, 2005), suggesting that this threat detection mechanism doesnot operate completely automatically and highlighting the fuzzy boundaries separatingautomatic from strategic processing. Further, it is unlikely that only the amygdala underliesthe automatic detection of threat; rather, it is likely most accurate to conceptualize the amygdalaas a central structure in a larger threat detection system.

A wealth of data is beginning to suggest that higher-order cortical structures, such as theprefrontal cortex (PFC) and its subunits and functionally-related structures (e.g., anteriorcingulate cortex [ACC] , orbitofrontal cortex), may be neural mechanisms underlyingdifficulties disengaging attention from threat. These neural structures serve a regulatorypurpose and can down-regulate emotion-relevant limbic structures, thus providing a ‘top-down’ processing influence (Miller & Cohen, 2001). Evidence that prefrontal structures down-regulate sub-cortical emotional systems comes from 1) studies demonstrating that theprefrontal cortex and related structures are critically involved in down-regulating amygdalaprocessing during extinction learning (see Myers & Davis, 2007; Quirk, Garcia, & Gonzalez-Lima, 2006; Quirk et al., 2003; Sotres-Bayon, Cain, & LeDoux, 2006), and 2) studiesdemonstrating that employing the emotion regulation strategy of ‘re-appraisal’ (Gross,1998b) results in increased PFC activity and reduced amygdala activity while watchingaversive films (Eippert et al., 2007; Kim & Hamann, 2007; Urry et al., 2006; Ocshner et al.,2004). These data support the generic effect of top down regulatory control over sub-corticalfear circuits.

Emerging evidence suggests that prefrontal regulatory structures are involved in thedisengagement of attention from threat. Derryberry and Reed (2002) found that trait anxiousindividuals with good attentional control, which is arguably a regulatory skill (also seeDerryberry & Rothbart, 1997; Eysenck et al., 2007; Posner & Rothbart, 2000), are able to shiftattention away from threatening stimuli at 500 ms, whereas trait anxious individuals with poorattentional control are not able to do so. Accordingly, higher-order attentional control abilityappears to determine the degree to which an individual can disengage attention from threat.Bishop and colleagues (2004) found that state anxiety was significantly inversely correlatedwith PFC activity (r = −.60) during a task in which individuals had to ignore pictures of fearfulfacial expressions. These data suggest that regulatory ability (i.e., PFC activity) is reduced inanxious individuals attempting to ignore (i.e., disengage from) threatening information. Bishop

Cisler and Koster Page 13

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

and colleagues (2009) similarly found that elevated trait anxiety is associated with decreasedPFC activity while performing a low demand non-threatening task, suggesting that theattentional system of trait anxious individuals generally (i.e., outside of the context of threat)involves ‘impoverished’ PFC control. Dolcos and McCarthy (2006) found that fear-relateddistracter pictures impaired working memory and that PFC activity was strongly inverselycorrelated with the degree to which the emotional pictures distracted the participants (r = −.74). These data suggest that the degree to which threatening stimuli distract is based on themagnitude of PFC: less PFC activity is linked with greater threat distraction, more PFC islinked with less threat distraction. These sources of data support the hypothesis that prefrontalregulatory structures may underlie difficulty in disengagement from threat. As is the case withthe amygdala, though, it is likely most accurate to conceptualize the PFC as a central structurein a larger attentional control/regulatory system instead of making a 1:1 correspondencebetween PFC function and attentional control.

Stage of Information ProcessingA wealth of data demonstrate that attentional biases are observed at varying stimulus durationpresentations (e.g., 17 ms, 50 ms, 100 ms, 500 ms; Bar-Haim et al., 2007; Mogg et al., 1993;Koster et al., 2006). The observation that attentional biases are observed at varying stimuluspresentation durations is important because it suggests that the effect is not dependent on, oronly found in, certain stages of information processing. However, the nature of attentional biasmay be dependent on the stage of information processing. Information processing is commonlyconceptualized in two stages, automatic and strategic processing stages (Shiffrin & Schneider,1977). Automatic processing generally refers to processing that is effortless, capacity free,unintentional, and outside of conscious control, whereas strategic processing generally refersto processing that is effortful, capacity-limited, intentional, and dependent on conscious control(Shiffrin & Schneider, 1977). Although the boundary conditions between automatic andstrategic processing are blurry at best (see Moors & de Houwer, 2006), conceptualizingattentional biases in terms of automatic and strategic processing is considered theoreticallymeaningful. In particular, McNally (1995) has argued that the criterion of unintentionality isof importance in psychopathology as it relates to limited control over the processing biases.

First, there is substantial empirical evidence to suggest that attentional biases can occur underconditions of limited conscious awareness of the presence of threat. Evidence for this assertioncomes from studies using masked stimuli: stimuli are presented briefly (e.g., 17 ms) andfollowed by a backwards mask that precludes conscious awareness of the stimulus. Forexample, in a Stroop task, the word ‘spider’ may occur for 17 ms and then be immediatelyreplaced by ‘xxxx’ for 483 ms. Attentional biases towards masked stimuli have beendemonstrated in the Stroop task (Bradley, Mogg, Millar, & White, 1995; Harvey, Bryant, &Rapee, 1996; Mogg et al., 1993; van den Hout, Tenney, Huygens, & de Jong, 1997; van Honket al. 2001, experiment 2) and dot probe tasks (Carlson & Reinke, 2008; Mogg, Bradley, &Williams, 1995; Mogg, Bradley, & Hallowell, 1994). The Bar-Haim and colleague (2007)meta-analysis found an aggregate effect size of d = .32 for subliminally presented stimuli. Thisbody of data is theoretically relevant because it suggests automatic pre-conscious biasestowards threat. The observation that non-anxious controls do not tend to display automaticattentional biases (Bar-Haim et al., 2007) suggests that anxious individuals are uniquelycharacterized by an exaggerated threat detection mechanism. However, this research onlymanipulated awareness; thus, this research cannot specifically address intentionality or control.It will be important for future research to manipulate other aspects of automaticity to furthertest the degree to which attentional biases operate automatically.

Second, a substantial amount of empirical evidence also demonstrates an attentional biastowards supraliminally presented stimuli (Bar-Haim et al., 2007; Cisler et al., 2009). For

Cisler and Koster Page 14

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

example, Koster and colleagues (2006) found attentional biases for threat stimuli in the dotprobe task among high trait anxious participants at stimulus presentation durations of 500 ms.It is important to note that an attentional bias towards supraliminally presented stimuli doesnot preclude automatic processing influences. Thus, demonstration of attentional biasestowards supraliminal stimuli is not a “pure” indicator of strategic processing, but is insteadlikely a mixture of both automatic and strategic processing. Supraliminal attentional biaseshave been demonstrated in the Stroop task, dot probe task, spatial cueing task, and visual searchtask (see Bar-Haim et al., 2007; Cisler et al. 2009). The Bar-Haim and colleague (2007) meta-analysis found an aggregate effect size of d = .48 for consciously perceived stimuli, which didnot significantly differ from the effect size of attentional biases towards masked stimuli.Attentional biases may then not only reflect automatic early warning threat detectionmechanisms, but also reflect purposeful attempts to detect and cope with disorder-relevantsources of threat in the environment. These studies, however, only manipulate awareness, andthus cannot speak to other aspects of strategic processing, such as intentionality and control,thus necessitating future research using additional strategic processing manipulations. .

Summary and IntegrationIn regards to the components of attention biases, the evidence demonstrates that attentionalbiases are comprised of facilitated attention to threat at short stimulus durations and high threatintensities, delayed disengagement from threat, and attentional avoidance of threat at late stagesof processing. In regards to the mechanisms that mediate attentional biases, attentional controlappears to mediate difficulty in disengagement from threat, and emotion regulation goals maymediate attentional avoidance. Further, neurobehavioral research is beginning to illuminatethat amygdala activity may mediate facilitated attention to threat. Higher-order corticalstructures centered around PFC activity may mediate delayed disengagement from threat viaindividual differences in the ability to down-regulate the influence of sub-cortical fearstructures and maintain attention on task-relevant stimuli. PFC-centered activity may mediateattentional avoidance, given that emotion regulation goals are linked with attentional avoidanceand PFC-centered activity mediates emotion regulation. In regards to the stage of informationprocessing during which attentional biases emerge, several lines of evidence demonstrate thatattentional biases depend on both automatic and strategic processing.

From this review of the evidence, the postulates from previous models can be evaluated.Generally, no model (see Table 1) predicts all of the findings reviewed above. Morespecifically, difficulty in disengagement and attentional avoidance are particularlyunderrepresented characteristics in the models. Further, emotion regulation and attentionalcontrol are empirically supported underlying mechanisms that are generally unpredicted bythe models. In contrast, several features of attentional biases that are common predictionsacross the models are indeed empirically supported. First, research demonstrates facilitatedattention towards threat, which is consistent with several models. Second, neuroimagingresearch suggests a critical role of the amygdala in the automatic detection of threat, supportingthe postulate of an automatic threat detection mechanism. Third, neuroimaging research alsosuggests a higher-order control mechanism, possibly centered around the PFC and functionallyrelated structures, that regulates attentional allocation to threat. Some emerging research alsosuggests that this control mechanism may be related to difficulty in disengagement andattentional avoidance. Several models posit such a higher-order control mechanism underlyingattentional biases towards threat. Fourth, research suggests that attentional biases are comprisedof both automatic and strategic processing, which is consistent with the predictions of severalmodels.

In a purely derivative manner, the research reviewed above on the mechanisms underlingattentional bias can be arranged into one possible framework of attentional biases towards

Cisler and Koster Page 15

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

threat (see Figure 1). This framework is not intended to be an absolute explanation of attentionalbiases, but instead is a summary of what is currently known. Of particular relevance for thepresent topic are the interrelations between the different aspects implied in this framework.The threat detection mechanism is predicted to operate during automatic processing andunderlie facilitated attention. While this is consistent with several lines of research (e.g., Kosteret al., 2006;Whalen et al., 1998;2004;van Damme et al., 2006), to our knowledge, only twostudies have supported this prediction by using a methodology that concurrently assesses eachof these aspects (Carlson et al., 2009;Monk et al., 2008). It will be important for future researchto employ designs that allow for the concurrent assessment of stage of processing, neuralactivity, and component of attention in order to replicate these findings. Attentional avoidanceand emotion regulation are predicted to operate in the strategic stage of processing and underlieattentional avoidance and difficulty in disengagement. Again, while this postulate is consistentwith several lines of research (e.g., Derryberry & Reed, 2002;Johnson, 2009), to ourknowledge, no study has concurrently demonstrated each of these levels of analyses in onemethodology.

While the interactions between the mediating mechanisms and attentional components seemrelatively well supported (i.e., across the vertical axis in Figure 1), what remains less clear arethe interactions across stage of processing, mediating mechanisms, and attentional components(i.e., across the horizontal axis in Figure 1). For example, in what way are facilitated attentionand difficulty in disengagement related? Research demonstrates that while difficulty indisengagement can be observed independently of facilitated attention (Fox, Russo, Bowles, &Dutton, 2001;Fox, Russo, & Dutton, 2002, experiment 1; Yiend & Mathews, 2001;Amir, Elias,Klumpp, & Przeworski, 2003;Rinck, Becker, Kellermann, & Roth, 2003; experiment 1; Rinck,Reinecke, Ellwart, Heuer, & Becker, 2005; experiment 1), facilitated attention almost neveroccurs without difficulty in disengagement also being observed (Byrne & Eysenck,1995;Gilboa-Schechtman, Foa, & Amir, 1999;Koster et al., 2006a;2005;2004b;Miltner,Krieschel, Hecht, Tripp, & Weiss, 2004;Rinck et al., 2003; experiment 2; Rinck et al., 2005,experiments 2 and 3; Van Damme et al., 2006) with Pflugshaupt et al. (2005), and Carlson andReinke (2008; though this study only used masked stimuli) as the two exceptions. Given theevidence linking facilitated attention with automatic stages of processing (Koster et al.,2006;Morris et al., 1998;Whalen et al., 1998), and difficulty in disengagement with higher-order control mechanisms (e.g., PFC activity; Bishop et al., 2004; 2009), it seems reasonableto conclude that facilitated attention precedes difficulty in disengagement when the two occurtogether. One possible explanation of the observed relation between these two characteristicsmay be that if stimulus properties in a given experiment are sufficient to fire the facilitatedattention mechanism, then a feed-forward process is created whereby bottom-up affectiveprocessing (i.e., threat detection mechanism/amygdala) exerts too strong of an influence forprefrontal regulatory processes to control. Thus, stimuli capable of eliciting facilitated attentionwould necessarily also lead to difficulty in disengagement. Difficulty in disengagement mayoccur independently of facilitated attention because it is not dependent on initial absolute firingof the threat-detection mechanism. Instead, difficulty in disengagement may occur in situationswhere attention is allocated to a stimulus in a non-facilitated manner (e.g., if someone says‘look at that spider!’) because of generally poor regulatory control when attention is allocatedtowards the threatening stimulus. This explanation seems plausible, but future research isnecessary to clarify this seemingly complex relation between these characteristics of attentionalbias.

The vigilance-avoidance hypothesis (e.g., Mogg et al., 2004) posits that anxious individualsfirst demonstrate facilitated attention to threat and then avoid attention away from threat. Thishypothesis suggests a temporal relation between facilitated attention and attentional avoidance,but it is not clear how difficulty in disengagement fits into this description. Indeed, it wouldseem as though difficulty in disengagement and attentional avoidance are competing

Cisler and Koster Page 16

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

mechanisms: the avoidance of attention onto threat necessarily precludes simultaneousdifficulty removing attention from threat. One possible resolution of the apparent conflictbetween these two characteristics comes from Wiereich and colleagues (2008), who argue thatindividuals may overtly avoid allocation attention onto threat, while concurrently covertlymaintaining attention onto threat. For example, a socially phobic man may avert his eyes fromdisapproving faces (overt attentional avoidance), but his cognitive resources may still beallocated onto the disapproving face (covert difficulty disengaging). This explanation suggestsa plausible manner by which these characteristics may co-occur. However, there is littleempirical research investigating the relations between the components of attention, and futureresearch along these lines is necessary to clarify the temporal dynamics of attentional biases.

Similarly, the relations between the candidate mediating mechanisms remain unclear. Forexample, to what degree are the threat detection mechanism and attentional control mechanismrelated? One interesting area for future research to explore is the relative weight of thesemechanisms necessary to produce attentional biases. For example, attentional biases may occurdue to overactivity of the threat detection mechanism, underactivity of the attentional controlmechanism, or a combination of both. Consider the following illustrative example. Theserotonin (5-HT) transporter (5-HTT) facilitates reuptake of 5-HT from the synaptic cleft. Apolymorphism of the 5-HTT gene (5-HTTLPR) results in two variants: a short and long allele.The short allele results in a 50% reduction in 5-HTT availability, resulting in increased synapticlevels of 5-HT (Hariri et al., 2002). The short allele has been linked with attentional biasestowards threat among psychiatric inpatients (Beevers et al., 2007) and healthy women(Osinsky, 2008). Further, the short allele has been linked specifically with difficulty indisengaging attention from threat and not with facilitated attention (Beevers et al., 2009). Theshort allele has also been linked with both amygdala hyperactivity towards threat (Harir et al.,2002; Munafo et al., 2008) and decreased functional connectivity between the amygdala andACC (Pezawas et al., 2005). Accordingly, the route by which the short allele may affectattention for threat may be through enhanced amygdala activity, decreased regulatory controlover the amygdala, or both. As this example illustrates, future research is needed to clarify howthe different underlying mechanisms interact to produce the components of attentional biases.Research along these lines seems particularly important for understanding the processesunderlying anxiety; that is, elucidation of the mechanisms and interactions among mechanismsresponsible for attentional biases may shed light on the cognitive-emotional mechanismsunderlying anxiety. Clarification of this issue may depend on the component of attention beingmeasured (e.g., hyperactive threat detection may affect facilitated attention but not necessarilydifficulty in disengagement) as well as the stage of processing being investigated.

Finally, the interaction between automatic and strategic processing also remains unclear.Whereas there is solid evidence that the amygdala responds to masked threat, suggesting(relative) automaticity, that attentional avoidance is mostly found during strategic processing,and that attentional control moderates difficulty in disengagement during strategic processing,there is little research suggesting that difficulty in disengagement operates mostly during anyone stage. Studies have found difficulty in disengagement during a range of stimuluspresentation durations (e.g., 100-600 ms; Amir et al., 2003; Fox et al., 2001; Koster et al.,2006), suggesting both stages of processing. Difficulty in disengagement's candidateunderlying mechanism, attentional control, was found to affect biases at 500, but not 250 ms,stimulus duration in one study (Derryberry & Reed, 2002), but affect biases towards stimulipresented for 100 ms in another study (Peers & Lawrence, 2009). It is likely most accurate toconceptualize difficulty in disengagement and attentional control as reflecting a mixture ofautomatic and strategic processing. Further clarification of how stage of processing constrainsthe other domains in attentional biases is necessary. This research will likely entail stage ofprocessing manipulations other than simply manipulating stimulus duration in order to assurethat other aspects of processing (e.g., intentionality, control) are also manipulated.

Cisler and Koster Page 17

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

The research on the stage of processing and mediating mechanisms of attentional bias isconsistent with dual process models positing an automatic/associative/intuitive/emotionalprocessing system and a planned/purposeful/verbal/cognitive processing system (Barrett,2004; Carver, Johnson, & Joormann, 2008; Evans, 2008). That is, facilitated attention to threatand the threat detection mechanism seem related to an automatic affective mechanism linkedwith amygdala activity, and attentional avoidance, difficulty in disengagement, attentionalcontrol, and emotion regulation seem related to strategic cognitive-regulatory processes linkedwith higher-order cortical structures. Though a distinction is made between these dual modesof processing both conceptually and in the neural regions involved, the boundaries betweenthese systems remain unclear. Pessoa (2008) recently commented that maintaining a distinctionbetween these emotional and cognitive systems is problematic because the systems areinteractive. Distinguishing between automatic versus strategic, or affective versus cognitive,seems theoretically important in that distinct mechanisms may be involved (see Figure 1), buta distinction between these systems should not entail the functional separation of these systems.That is, just as the keys of a piano are distinct from the piano's strings, neither alone is sufficientfor the piano to function. Similarly, the question should not simply be ‘automatic or strategic?’,or ‘affective or cognitive?’, but a more nuanced investigation of how these systems interact todifferentially impact attentional biases is needed.

Clearly, conceptualizing attentional biases as a dynamic system, in which interrelationsbetween the domains are presumed and investigation of one component necessitatesinvestigation of the other components, is necessary to fully understand relations between thecomponents of attentional biases towards threat. This conceptualization may necessitate morecomplex, but more rigorous, methodological paradigms. Threat intensity may moderatefacilitated attention (Carlson & Rienke, 2008; Koster et al., 2006; van Damme et al., 2006),and thus this manipulation may be necessary to observe facilitated attention. Stimulus durationmanipulations have been a common means to manipulate stage of processing and, presumably,the underlying mechanisms. Another means that has only recently begun to be explored is theuse of cognitive load (Van Dillen & Koole, in press). Embedding a cognitive load task into anattentional task may provide an interesting way to examine the automaticity of attentionalbiases. From the research reviewed above, it would be expected that the depletion of cognitiveresources would potentiate difficulty in disengagement. Given that attentional avoidance maybe linked with purposeful attempts to regulate affect, depletion of cognitive resources mayattenuate attentional avoidance. Given that facilitated attention is presumed to be linked withan automatic threat detection mechanism, it would be expected that depletion of cognitiveresources would have little effect on facilitated attention. Using cognitive load manipulationsappears to be a critical next step in elucidating the mechanisms underlying attentional biastowards threat.

Future DirectionsHow do the biases relate to other aspects of anxious responding?

There is a surprising lack of research investigating how the different components of attentionalbias correlate with other aspects of anxious responding. The research is limited todemonstrating correlations between emotional Stroop biases and self-reported distress duringactual or imagined stressors (MacLeod & Hagan, 1992; Nay, Thorpe, Roberson-Nay, Hecker,& Sigmon, 2004; van den Hout, Tenney, Huygens, Merckelbach, & Kindt, 1995). It remainsunclear whether facilitated attention, difficulty in disengagement, and attentional avoidancesimilarly predict self-reported distress. It is similarly unclear how the specific components ofattentional biases predict avoidance, escape, success during exposure, etc. It is difficulty todetermine how these aspects of attentional biases relate to a larger understanding of anxiety.Moreover, it is difficult to determine which cognitive/affective mechanisms underlie anxietywithout knowing how the mechanisms of attentional bias predict aspects of anxious

Cisler and Koster Page 18

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

responding. Future research linking the specific aspects of attentional biases with aspects ofanxious responding will be critical in elucidating the mechanisms of attentional bias as wellas in understanding the processes underlying anxiety.

How do attention retraining methodologies affect the mechanisms of attentional biases?Numerous studies are beginning to investigate whether training attentional biases affectsanxiety. Studies have shown that training attention to be biased towards threat increases anxiety(MacLeod et al., 2002), and that training attention away from threat can actually reducesymptoms of social phobia (Amir et al., 2008; 2009) and GAD (Schmidt et al., 2009; Hazenet al., 2009). Although these emerging data are exciting, the mechanisms by which the trainingprotocols work are not entirely clear. For example, protocols that train attention away fromthreat ostensibly are increasing attentional avoidance, which from this review would actuallybe contraindicated. However, an alternative explanation is that training attentional avoidanceis necessarily also training increased ability to disengage from threat. As such, these attentiontraining treatment protocols may work through bolstering the ability to disengage. Consistentwith this explanation, the only attention training study that has investigated which componentsof attention bias change after the attention training (Amir, Beard, Taylor, et al., 2009) founddecreases in difficulty in disengagement from pre-post attention training. Further, the currentresearch on attention training has not examined how the training affects any of the candidateunderlying mechanisms. For example, does attention training away from threat increaseattentional control? It is clear that future studies of attention training need to assess how thetraining affects each component and underlying mechanism of attentional bias. Research inthis area will have a large impact on understanding the mechanisms of attentional bias, themechanisms of treatment success during attention retraining, as well as the overall mechanismsof pathological anxiety.

Concurrent assessment of each of the three domains involved in attentional biasesAs noted above, conceptualizing attentional biases as a dynamic system in which these levelsare presumed to interact will necessitate research concurrently measuring and/or manipulatingaspects of each of the three levels. This will likely entail a stage of processing and mediatingmechanism manipulation (e.g., stimulus duration, cognitive load, state anxiety, threat intensity)with concurrent measurement of the components of attentional bias. Translational research, inwhich cognitive, affective, and neuroscience each inform one another is likely to be of greathelp. Future research along these lines will help further elucidate the relations between thesedomains.

ConclusionsThe present paper conceptualized attentional biases in terms of three complimentary aspects:the attentional components, mediating mechanisms, and stages of information processing.Research has identified elements within each domain and suggested specific interrelationsbetween the domains. Though future research is still needed to clarify inconsistencies andambiguities, there has been clear progress in identifying the mechanisms of attentional biases.Future research in this area will hopefully further elucidate the specific processes underlyingattentional bias and clarify how this research informs a larger understanding of the processesof anxiety.

ReferencesAlgom D, Chajut E, Lev S. A rational look at the emotional Stroop paradigm: a generic slowdown, not

a Stroop effect. Journal of Experimental Psychology: General 2004;133:323–338. [PubMed:15355142]

Cisler and Koster Page 19

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Amir N, Beard C, Burns M, Bomyea J. Attention modification program in individuals with generalizedanxiety disorder. Journal of Abnormal Psychology 2009;118:28–33. [PubMed: 19222311]

Amir N, Beard C, Taylor CT, Klumpp H, Elias J, Burns M. Attention training in individuals withgeneralized social phobia: a randomized controlled trial. Journal of Consulting and ClinicalPsychology 2009;77:961–973. [PubMed: 19803575]

Amir N, Weber G, Beard C, Bomyea J, Taylor CT. The effect of a single-session attention modificationprogram on response to a public-speaking challenge in socially anxious individuals. Journal ofAbnormal Psychology 2008;117:860–868. [PubMed: 19025232]

Amir N, Elias J, Klumpp H, Przeworski A. Attentional bias to threat in social phobia: Facilitatedprocessing of threat or difficulty disengaging attention from threat? Behaviour Research and Therapy2003;41:1325–1335. [PubMed: 14527531]

Amir N, Najmi S, Morrison AS. Attenuation of attention bias in obsessive-compulsive disorder.Behaviour Research and Therapy. in press.

Anderson AK. Affective influences on the attentional dynamics supporting awareness. Journal ofExperimental Psychology: General 2005;134:258–281. [PubMed: 15869349]

Anderson AK, Phelps EA. Lesions of the human amygdala impair enhanced perception of emotionallysalient events. Nature 2001;411:305–309. [PubMed: 11357132]

Andreasen NC. A unitary model of schizophrenia: Bleuler's ‘fragmented phrene’ as schizencephaly.Archives of General Psychiatry 1999;56:781–787. [PubMed: 12884883]

Bar-Haim Y, Lamy D, Pergamin L, Bakermans-Kranenburg, van IJzendoorn MH. Threat-relatedattentional bias in anxious and non-anxious individuals: a meta-analytic study. Psychological Bulletin2007;133:1–24. [PubMed: 17201568]

Barlow, DH. Anxiety and its disorders: the nature and treatment of anxiety and panic. Guilford Press;New York: 2002.

Barrett LF, Tugade MM, Engle RW. Individual differences in working memory capacity and dual-processtheories of the mind. Psychological Bulletin 2004;130:553–573. [PubMed: 15250813]

Beck AT, Clark DA. An information processing model of anxiety: Automatic and strategic processes.Behaviour Research and Therapy 1997;35:49–58. [PubMed: 9009043]

Becker ES, Rinck M, Margraf J, Roth WT. The emotional Stroop effect in anxiety disorders: Generalemotionality or disorder specificity? Journal of Anxiety Disorders 2001;15:147–159. [PubMed:11442135]

Beevers CG, Gibb BE, McGeary JE, Miller IW. Serotonin transporter genetic variation and biasedattention for emotional word stimuli among psychiatric inpatients. Journal of Abnormal Psychology2007;116:208–212. [PubMed: 17324031]

Beevers CG, Wells TT, Ellis AJ, McGeary JE. Association of the serotonin transporter gene promoterregion (5-HTTLPR) polymorphism with biased attention for emotional stimuli. Journal of AbnormalPsychology 2009;118:670–681. [PubMed: 19685963]

Bishop SJ. Trait anxiety and impoverished prefrontal control of attention. Nature Neuroscience2009;12:92–98.

Bishop S, Duncan J, Brett M, Lawrence AD. Prefrontal cortical function and anxiety: controlling attentionto threat-related stimuli. Nature Neuroscience 2004;7:184–188.

Bradley BP, Mogg K, Millar N, White J. Selective processing of negative information: effects of clinicalanxiety, concurrent depression, and awareness. Journal of Abnormal Psychology 1995;104:532–536.[PubMed: 7673577]

Bradley BP, Mogg K, Falla SJ, Hamilton LR. Attentional bias for threatening facial expressions inanxiety: manipulation of stimulus duration. Cognition & Emotion 1998;12:737–753.

Bradley BP, Mogg K, White J, Groom C, de Bono J. Attentional bias for emotional faces in generalizedanxiety disorder. The British Journal of Clinical Psychology 1999;38:267–278. [PubMed: 10532148]

Bryant RA, Harvey AG. Processing threatening information in posttraumatic stress disorder. Journal ofAbnormal Psychology 1995;104:537–541. [PubMed: 7673578]

Buckley TC, Blanchard EB, Hickling EJ. Automatic and strategic processing of threat stimuli: acomparison between PTSD, panic disorder, and non-anxiety controls. Cognitive Therapy andResearch 2002;26:97–115.

Cisler and Koster Page 20

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Byrne A, Eysenck MW. Trait anxiety, anxious mood, and threat detection. Cognition and Emotion1995;9:549–562.

Calvo MG, Avero P. Time course of attentional bias to emotional scenes in anxiety: gaze direction andduration. Cognition and Emotion 2005;19:433–451.

Carlson JM, Reinke KS. Masked fearful faces modulate the orienting of covert spatial attention. Emotion2008;8:522–529. [PubMed: 18729582]

Carlson JM, Reinke KS, Habib R. A left amygdala mediated network for rapid orienting to masked fearfulfaces. Neuropsychologia 2009;47:1386–1389. [PubMed: 19428403]

Carver CS, Johnson SL, Joormann J. Serotonergic function, two-mode models of self-regulation, andvulnerability to depression: what depression has in common with impulsive aggression.Psychological Bulletin 2008;134:912–943. [PubMed: 18954161]

Chamberlain SR, Blackwell AD, Fineberg NA, Robbins TW, Sahakian BJ. The neuropsychology ofobsessive compulsive disorder: the importance of failures in cognitive and behavioural inhibition ascandidate endophenotypic markers. Neuroscience and Biobehavioral Reviews 2005;29:399–419.[PubMed: 15820546]

Cisler JM, Bacon AK, Williams NL. Phenomenological characteristics of attentional biases towardsthreat: a critical review. Cognitive Therapy and Research 2009;33:221–234.

Cisler JM, Olatunji BO. Components of attentional biases in contamination fear: evidence for difficultyin disengagement. Behaviour Research and Therapy. in press.

Corbetta M, Shulman GL. Control of goal-directed and stimulus-driven attention in the brain. NatureReviews Neuroscience 2002;3:201–215.

Dalgleish T, Taghavi R, Neshat-Doost H, Moradi A, Canterbury R, Yule W. Patterns of processing biasfor emotional information across clinical disorders: a comparison of attention, memory, andprospective cognition in children and adolescents with depression, generalized anxiety, andposttraumatic stress disorder. Journal of Clinical Child and Adolescent Psychology 2003;32:10–21.[PubMed: 12573928]

Davis M. Neural systems involved in fear and anxiety measured with fear-potentiated startle. AmericanPsychologist 2006;61:741–756. [PubMed: 17115806]

Davis M, Whalen PJ. The amygdala: vigilance and emotion. Molecular Psychiatry 2001;6:13–34.[PubMed: 11244481]

Derryberry D, Reed MA. Anxiety-related attentional biases and their regulation by attentional control.Journal of Abnormal Psychology 2002;111:225–236. [PubMed: 12003445]

Derryberry D, Rothbart MK. Reactive and effortful processes in the organization of temperament.Development and Psychopathology 1997;9:633–652. [PubMed: 9448999]

Dolan RJ, Vuilleumier P. Amygdala automaticity in emotional processing. Annals of the New YorkAcademy of Sciences 2003;985:348–355. [PubMed: 12724170]

Dolcos F, McCarthy G. Brain systems mediating cognitive interference by emotional distraction. Journalof Neuroscience 2006;26:2072–2079. [PubMed: 16481440]

Dunning JP, Hajack G. See no evil: directing visual attention within unpleasant images modulates theelectrocortical response. Psychophysiology 2009;46:28–33. [PubMed: 18992071]

Eippert F, Veit R, Weiskopf N, Erb M, Birbaumer N, Anders S. Regulation of emotional responses elicitedby threat-related stimuli. Human Brain Mapping 2007;28:409–423. [PubMed: 17133391]

Eldar S, Ricon T, Bar-Haim Y. Plasticity in attention: implications for stress response in children.Behaviour Research and Therapy 2008;46:450–461. [PubMed: 18313034]

Evans, J. St. B. T. Dual-processing accounts of reasoning, judgment, and social cognition. Annual Reviewof Psychology 2008;59:255–278.

Eysenck MW, Derakshan N, Santos R, Calvo MG. Anxiety and cognitive performance: attentional controltheory. Emotion 2007;7:336–353. [PubMed: 17516812]

Feldner MT, Zvolensky MJ, Eifert GH, Spira AP. Emotional avoidance: an experimental test of individualdifferences and response suppression using biological challenge. Behaviour Research and Therapy2003;41:403–411. [PubMed: 12643964]

Foa EG, Ilai D, McCarthy PR, Shoyer BS, Murdock T. Information processing in obsessive-compulsivedisorder. Cognitive Therapy and Research 1993;17:173–189.

Cisler and Koster Page 21

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Fox E, Russo R, Bowles R, Dutton K. Do threatening stimuli draw or hold visual attention in subclinicalanxiety? Journal of Experimental Psychology: General 2001;130:681–700. [PubMed: 11757875]

Fox E, Russo R, Dutton K. Attentional bias for threat: evidence for delayed disengagement fromemotional faces. Cognition and Emotion 2002;16:355–379. [PubMed: 18273395]

Garber J, Hollon SD. What can specificity designs say about causality in psychopathology research?Psychological Bulletin 1991;110:129–136. [PubMed: 1891514]

Garner M, Mogg K, Bradley BP. Orienting and maintenance of gaze to facial expressions in social anxiety.Journal of Abnormal Psychology 2006;115:760–770. [PubMed: 17100533]

Gilboa-Schechtman E, Foa EB, Amir N. Attentional biases for facial expressions in social phobia: theface-in-the-crowd paradigm. Cognition and Emotion 1999;13:305–318.

Gotlib IH, Kasch KL, Traill S, Joormann J, Arnow BA, Johnson SL. Coherence and specificity ofinformation-processing biases in depression and social phobia. Journal of Abnormal Psychology2004;386-398

Gross, JJ. Handbook of Emotion Regulation. Guilford Press; New York, NY: 2007.Gross JJ. Emotion regulation in adulthood: timing is everything. Current Directions in Psychological

Science 2001;10:214–219.Gross JJ. The emerging field of emotion regulation: an integrative review. Review of General Psychology

1998a;2:271–299.Gross JJ. Antecedent- and response-focused emotion regulation: divergent consequences for experience,

expression, and physiology. Journal of Personality and Social Psychology 1998b;74:224–237.[PubMed: 9457784]

Gross JJ, Levenson RW. Emotional suppression: physiology, self-report, and expressive behavior.Journal of Personality and Social Psychology 1993;64:970–986. [PubMed: 8326473]

Gross JJ, Levenson RW. Hiding feelings: the acute effects of inhibiting negative positive emotion. Journalof Abnormal Psychology 1997;106:95–103. [PubMed: 9103721]

Hariri AR, Drabant EM, Munoz KE, Kolachana b. S. Mattay VS, Egan MF, et al. A susceptibility genefor affective disorders and the response of the human amygdala. Archives of General Psychiatry2005;62:146–152. [PubMed: 15699291]

Hariri AR, Mattay VS, Tessitore A, Kolachana B, Fera F, Goldman D, et al. Serotonin transporter geneticvariation and the response of the human amygdala. Science 2002;297:400–403. [PubMed: 12130784]

Harvey AG, Bryant RA, Rapee RM. Preconscious processing of threat in posttraumatic stress disorder.Cognitive Therapy and Research 1996;20:613–623.

Hazen RA, Vasey MW, Schmidt NB. Attentional retraining: a randomized clinical trial for pathologicalworry. Journal of Psychiatric Research. in press.

Horenstein M, Segui J. Chronometrics of attentional processes in anxiety disorders. Psychopathology1997;30:25–35. [PubMed: 9042679]

Johnson DR. Goal-directed attentional deployment to emotional faces and individual differences inemotion regulation. Journal of Research in Personality 2009;43:8–13.

Juth P, Lundqvist D, Karlsson A, Öhman A. Looking for foes and friends: perceptual and emotionalfactors when finding a face in the crowd. Emotion 2005;5:379–395. [PubMed: 16366743]

Kim SH, Hamann S. Neural correlates of positive and negative emotion regulation. Journal of CognitiveNeuroscience 2007;19:776–798. [PubMed: 17488204]

Kindt M, Bierman D, Brosschot JF. Stroop versus Stroop: comparison of a card format and a single-trialformat of the standard color-word Stroop task and emotional Stroop task. Personality and IndividualDifferences 1996;21:653–661.

Kindt M, Van Den Hout M. Selective attention and anxiety: a perspective on developmental issues andthe causal status. Journal of Psychopathology and Behavioral Assessment 2001;23:193–202.

Koole SL. The psychology of emotion regulation: an integrative review. Cognition and Emotion2009;23:4–41.

Koster EHW, Crombez G, Van Damme S, Verschuere B, De Houwer J. Does imminent threat captureand hold attention? Emotion 2004b;4:312–317. [PubMed: 15456400]

Cisler and Koster Page 22

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Koster EHW, Crombez G, Van Damme S, Verschuere B, De Houwer J. Signals for threat modulateattentional capture and holding: fear-conditioning and extinction during the exogenous cueing task.Cognition and Emotion 2005;19:771–780.

Koster EHW, Crombez G, Verschuere B, De Houwer J. Attention to threat in anxiety-prone individuals:mechanisms underlying attentional bias. Cognitive Therapy and Research 2006b;30:635–643.

Koster EH, Crombez G, Verschuere B, De Houwer J. Selective attention to threat in the dot probeparadigm: Differentiating vigilance and difficulty to disengage. Behaviour Research and Therapy2004a;42:1183–1192. [PubMed: 15350857]

Koster EHW, Crombez G, Vershuere B, Van Damme S, Wiersema JR. Components of attentional biasto threat in high trait anxiety: facilitated engagement, impaired disengagement, and attentionalavoidance. Behaviour Research and Therapy 2006a;44:1757–1771. [PubMed: 16480943]

Koster EHW, Verschuere B, Crombez G, Van Damme S. Time-course of attention for threatening picturesin high and low trait anxiety. Behaviour Research and Therapy 2005;43:1087–1098. [PubMed:15922291]

Lavy EH, Van den Hout M, Arntz A. Attentional bias and spider phobia: conceptual and clinical issues.Behaviour Research and Therapy 1993;31:17–24. [PubMed: 8093338]

LeDoux, JE. The Emotional Brain. Simon & Schuster; New York: 1996.LeDoux JE. Emotion circuits in the brain. Annual Review of Neuroscience 2000;23:155–184.Li S, Tan J, Qian M, Liu X. Continual training of attentional bias in social anxiety. Behaviour Research

and Therapy 2008;46:905–912. [PubMed: 18538305]Lipp, OV. Human fear learning: contemporary procedures and measurement.. In: Craske, MG.; Hermans,

D.; Vansteenwegen, D., editors. Fear and Learning: from basic processes to clinical implications.American Psychological Association; Washington, DC: 2006. p. 37-52.

Lipp OV, Waters AM. When danger lurks in the background: attentional capture by animal fear-relevantdistracters is specific and selectively enhanced by animal fear. Emotion 2007;7:192–200. [PubMed:17352574]

Lonigan CJ, Vasey MW. Negative affectivity, effortful control, and attention to threat-relevant stimuli.Journal of Abnormal Child Psychology 2009;37:387–399. [PubMed: 19043783]

Macleod, C.; Campbell, L.; Rutherford, E.; Wilson, E. The causal status of anxiety-linked attentional andinterpretive bias.. In: Yiend, J., editor. Cognition, emotion, and psychopathology: Theoretical,empirical and clinical directions. Cambridge University Press; New York, NY: 2004. p. 172-189.

MacLeod C, Hagan R. Individual differences in the selective processing of threatening information, andemotional responses to a stressful life event. Behaviour Research and Therapy 1992;30:151–161.[PubMed: 1567344]

MacLeod C, Mathews A, Tata P. Attentional bias in the emotional disorders. Journal of AbnormalPsychology 1986;95:15–20. [PubMed: 3700842]

MacLeod, C.; Rutherford, E. Information-processing approaches: assessing the selective functioning ofattention, interpretation, and retrieval. In: Heimberg, RG.; Turk, CL., editors. Generalized AnxietyDisorder: Advances in Research and Practice. Guilford Press; New York, NY: 2004. p. 109-142.

MacLeod C, Rutherford E, Campbell L, Ebsworthy G, Lin H. Selective attention and emotionalvulnerability: assessing the causal basis of their association through the experimental manipulationof attentional bias. Journal of Abnormal Psychology 2002;111:107–123. [PubMed: 11866165]

Mathews A, Mackintosh B, Fulcher EP. Cognitive biases in anxiety and attention to threat. Trends inCognitive Sciences 1997;1:340–345.

Mathews A, MacLeod C. Induced processing biases have causal effects on anxiety. Cognition andEmotion 2002;16:331–354.

Mathews A, Mogg K, Kentish J, Eysenck M. Effect of psychological treatment on cognitive bias ingeneralized anxiety disorder. Behaviour Research and Therapy 1995;33:293–303. [PubMed:7726805]

Matthews A, Mackintosh B. A cognitive model of selective processing in anxiety. Cognitive Therapyand Research 1998;22:539–560.

Matthews G, Wells A. Attention, automaticity, and affective disorder. Behavior Modification2000;24:69–93. [PubMed: 10641368]

Cisler and Koster Page 23

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

McNally RJ. Automaticity and the anxiety disorders. Behaviour Research and Therapy 1995;33:747–754. [PubMed: 7677712]

McNally, RJ. Cognitive bias in the anxiety disorders.. In: Hope, DA., editor. Nebraska Symposium onMotivation, 1995: perspectives on anxiety, panic, and fear. University of Nebraska Press; Lincoln,NE: 1996. p. 211-250.

McNally RJ, Kaspi SP, Bradley CR, Zeitlin SB. Selective processing of threat cues in posttraumatic stressdisorder. Journal of Abnormal Psychology 1990;99:398–402. [PubMed: 2266215]

Miller EK, Cohen JD. An integrative theory of prefrontal cortex function. Annual Review ofNeuroscience 2001;24:167–202.

Miltner WHR, Krieschel S, Hecht H, Trippe R, Weiss T. Eye movements and behavioral responses tothreatening and nonthreatening stimuli during visual search in phobic and nonphobic subjects.Emotion 2004;4:323–339. [PubMed: 15571432]

Mogg K, Bradley BP. A cognitive-motivational analysis of anxiety. Behaviour Research and Therapy1998;36:809–848. [PubMed: 9701859]

Mogg K, Bradley BP, Bono JD, Painter M. Time course of attentional bias for threat information in non-clinical anxiety. Behaviour Research and Therapy 1997;35:297303.

Mogg K, Bradley BP, Dixon C, Fisher HT, McWilliams A. Trait anxiety, defensiveness and selectiveprocessing of threat: an investigation using two measures of attentional bias. Personality andIndividual Differences 2000;28:1063–1077.

Mogg K, Bradley BP, Hallowell N. Attentional bias to threat: Roles of trait anxiety, stressful events, andawareness. The Quarterly Journal of Experimental Psychology A: Human Experimental Psychology1994;47:841–864.

Mogg K, Bradley BP, Miles F, Dixon C. Time course of attentional bias for threat scenes: testing thevigilance-avoidance hypothesis. Cognition and Emotion 2004;18:689–700.

Mogg K, Bradley BP, Millar N, White J. A follow-up study of cognitive bias in generalized anxietydisorder. Behaviour Research and Therapy 1995;33:927–935. [PubMed: 7487852]

Mogg K, Bradley BP, Williams R. Attentional bias in anxiety and depression: the role of awareness.British Journal of Clinical Psychology 1995;34:17–36. [PubMed: 7757037]

Mogg K, Bradley BP, Williams R, Mathews A. Subliminal processing of emotional information in anxietyand depression. Journal of Abnormal Psychology 1993;102:304–311. [PubMed: 8315143]

Mogg K, Holmes A, Garner M, Bradley BP. Effects of threat cues on attentional shifting, disengagementand response slowing in anxious individuals. Behaviour Research and Therapy 2008;46:656–667.[PubMed: 18395185]

Monk CS, Nelson EE, Woldehawariat G, Montgomery LA, Zarahn E, McClure EB, et al. Experience-dependent plasticity for attention to threat: behavioral and neurophysiological evidence in humans.Biological Psychiatry 2004;56:607–610. [PubMed: 15476691]

Monk CS, Telzer EH, Mogg K, Bradley BP, Mai X, Louro HMC, et al. Amygdala and ventrolateralprefrontal cortex activation to masked angry faces in children and adolescents with generalizedanxiety disorder. Archives of General Psychiatry 2008;65:568–576. [PubMed: 18458208]

Moore S, Zoellner L, Mollenholt N. Are expressive suppression and cognitive re-appraisal associatedwith stress-related symptoms? Behaviour Research and Therapy 2008;46:993–1000. [PubMed:18687419]

Moors A, De Houwer J. Automaticity: a theoretical and conceptual analysis. Psychological Bulletin2006;132:297–326. [PubMed: 16536645]

Moritz S, Fischer B, Hottenrott B, Kellner M, Fricke S, Randjbar S, et al. Words may not be enough! Noincreased emotional Stroop effect in obsessive-compulsive disorder. Behaviour Research andTherapy 2008;46:1101–1104. [PubMed: 18675953]

Moritz S, von Muhlenen A. Investigation of an attentional bias for fear-related materials in obsessive-compulsive checkers. Depression and Anxiety 2008;25:225–229. [PubMed: 17405162]

Morris J,S, Ohman A, Dolan RJ. Conscious and unconscious emotional learning in the human amygdala.Nature 1998;393:467–470. [PubMed: 9624001]

Munafo MR, Brown SM, Hariri AR. Serotonin transporter (5-HTTLPR) genotype and amygdalaactivation: a meta-analysis. Biological Psychiatry 2008;63:852–857. [PubMed: 17949693]

Cisler and Koster Page 24

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Myers KM, Davis M. Mechanisms of fear extinction. Molecular Psychiatry 2007;12:120–150. [PubMed:17160066]

Nay WT, Thorpe GL, Roberson-Nay R, Hecker JE, Sigmon ST. Attentional bias to threat and emotionalresponse to biological challenge. Journal of Anxiety Disorders 2004;18:609–627. [PubMed:15275942]

Ochsner KN, Ray RD, Cooper JC, Robertson ER, Chopra S, Gabrieli JDE, et al. For better or for worse:neural systems supporting the cognitive down- and up-regulation of negative emotion. Neuroimage2004;23:483–499. [PubMed: 15488398]

Ohman, A. Preferential preattentive processing of threat in anxiety: preparedness and attentional biases..In: Rapee, RM., editor. Current Controversies in the Anxiety Disorders. Guilford Press; New York:1996.

Öhman, A. Fear and anxiety: evolutionary, cognitive, and clinical perspectives.. In: Lewis, M.; Haviland-Jones, JM., editors. Handbook of emotions. 2nd edition. Guilford Press; New York, NY: 2000. p.573-593.

Öhman A. The role of the amygdala in human fear: automatic detection of threat.Psychoneuroendocrinology 2005;30:953–958. [PubMed: 15963650]

Öhman A, Flykt A, Esteves F. Emotion drives attention: detecting the snake in the grass. Journal ofExperimental Psychology: General 2001;130:466–478. [PubMed: 11561921]

Öhman A, Mineka S. Fears, phobias, and preparedness: toward an evolved module of fear and fearlearning. Psychological Review 2001;108:483–522. [PubMed: 11488376]

Öhman A, Wiens S. The concept of an evolved fear module and cognitive theories of anxiety. Feelingsand emotions: the Amsterdam symposium 2004:58–80.

Osinksy R, Reuter M, Kupper Y, Schmitz A, Kozyra E, Alexander N, et al. Variation in the serotonintransporter gene modulates selective attention to threat. Emotion 2008;8:584–588. [PubMed:18729589]

Peers PV, Lawrence AD. Attentional control of emotional distraction in rapid serial visual presentation.Emotion 2009;9:140–145. [PubMed: 19186928]

Pessoa L. On the relationship between emotion and cognition. Nature Reviews Neuroscience 2008;9:148–158.

Pezawas L, Meyer-Lindenberg A, Drabant EM, Verchinski BA, Munoz KE, Kolachana BS, et al. 5-HTTLP polymorphism impacts human cingulate-amygdala interactions: a genetic susceptibilitymechanism for depression. Nature Neuroscience 2005;8:828–834.

Pflugshaupt T, Mosimann UP, von Wartburg R, Schmitt W, Nyffeler T, Muri RM. Hypervigilance-avoidance pattern in spider phobia. Journal of Anxiety Disorders 2005;19:105–116. [PubMed:15488370]

Posner MI. Orienting of attention. Quarterly Journal of Experimental Psychology 1980;32:3–25.[PubMed: 7367577]

Posner MR, Rothbart MK. Developing mechanisms of self-regulation. Development andPsychopathology 2000;12:427–441. [PubMed: 11014746]

Quirk GJ, Garcia R, Gonzalez-Lima F. Prefrontal mechanisms in extinction of conditioned fear.Biological Psychiatry 2006;60:337–343. [PubMed: 16712801]

Quirk GJ, Likhtik E, Pelletier JG, Pare D. Stimulation of medial prefrontal cortex decreases theresponsiveness of central amygdala output neurons. Journal of Neuroscience 2003;23:8800–8807.[PubMed: 14507980]

Ratcliff R. Methods for dealing with reaction time outliers. Psychological Bulletin 1993;114:510–532.[PubMed: 8272468]

Raymond JE, Shapiro KL, Arnell KM. Temporary suppression of visual processing in an RSVP task: Anattentional blink? Journal of Experimental Psychology: Human Perception and Performance1992;18:849–860. [PubMed: 1500880]

Rinck M, Becker ES, Kellermann J, Roth WT. Selective attention in anxiety: distraction and enhancementin visual search. Depression and Anxiety 2003;18:18–28. [PubMed: 12900949]

Rinck M, Reinecke A, Ellwart T, Heuer K, Becker ES. Speeded detection and increased distraction infear of spiders: Evidence from eye movements. Journal of Abnormal Psychology 2005;114:235–248. [PubMed: 15869354]

Cisler and Koster Page 25

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Rohner J-C. The time-course of visual threat processing: high trait anxious individuals eventually averttheir gaze from angry faces. Cognition and Emotion 2002;16:837–844.

Rosen JB. The neurobiology of conditioned and unconditioned fear: a neurobehavioral system analysisof the amygdala. Behavioral and Cognitive Neuroscience Reviews 2004;3:23–41. [PubMed:15191640]

Salemink E, van den Hout MA, Kindt M. Selective attention and threat: quick orienting versus slowdisengagement and two versions of the dot probe task. Behaviour Research and Therapy2007;45:607–615. [PubMed: 16769035]

Schmukle SC. Unreliability of the dot probe task. European Journal of Personality 2005;19:595–605.Sergerie K, Chochol C, Armony JL. The role of the amygdala in emotional processing: a quantitative

meta-analysis of functional neuroimaging studies. Neuroscience & Biobehavioral Reviews2008;32:811–830. [PubMed: 18316124]

Sheppes G, Meiran N. Divergent cognitive costs for online forms of reappraisal and distraction. Emotion2008;8:870–874. [PubMed: 19102598]

Shiffrin RM, Schneider W. Controlled and automatic human information processing: II. Perceptuallearning, automatic attending and a general theory. Psychological Review 1977;84:127–190.

Siegrist M. Test-retest reliability of different versions of the Stroop test. The Journal of Psychology1997;131:299–306.

Sotres-Bayon F, Cain CK, LeDoux JE. Brain mechanisms of fear extinction: historical perspectives onthe contribution of prefrontal cortex. Biological Psychiatry 2006;60:329–336. [PubMed: 16412988]

Spielberger, CD.; Gorsuch, RL.; Lushene, R.; Vagg, PR.; Jacobs, GA. Manual for the State-Trait AnxietyInventory. Consulting Psychologists Press; Palo Alto, CA: 1983.

Stroop JR. Studies of interference in serial verbal reactions. Journal of Experimental Psychology1935;18:643–662. (Reprinted in 1992 in the Journal of Experimental Psychology: General, 121,15-23.

Tata PR, Leibowitz JA, Prunty MJ, Cameron M, et al. Attentional bias in obsessional compulsive disorder.Behaviour Research and Therapy 1996;34:53–60. [PubMed: 8561765]

Urry HI, van Reekum CM, Johnstone T, Kalin NH, Thurow ME, Schaefer HS, et al. Amygdala andventromedial prefrontal cortex are inversely coupled during regulation of negative affect and predictthe diurnal pattern of cortisol secretion among older adults. Journal of Neuroscience 2006;26:4415–4425. [PubMed: 16624961]

Van Damme S, Crombez G, Hermans D, Koster EHW, Eccleston C. The role of extinction andreinstatement in attentional bias to threat: A conditioning approach. Behaviour Research andTherapy 2006;44:1555–1563. [PubMed: 16375852]

van den Heuvel OA, Veltman DJ, Groenewegen HJ, Witter MP, Merkelbach J, Cath DC. Disorder-specific neuroanatomical correlates of attentional bias in obsessive-compulsive disorder, panicdisorder, and hypochondriasis. Archives of General Psychiatry 2005;62:922–933. [PubMed:16061770]

van den Hout M, Tenney N, Huygens K, de Jong P. Preconscious processing of bias in specific phobia.Behaviour Research and Therapy 1997;35:29–34. [PubMed: 9009041]

van den Hout M, Tenney N, Huygens K, Merckelbach H, Kindt M. Responding to subliminal threat cuesis related to trait anxiety and emotional vulnerability: a successful replication of MacLeod andHagan (1992). Behaviour Research and Therapy 1995;33:451–454. [PubMed: 7755532]

Van Dillen LF, Koole SL. How automatic is ’automatic vigilance”? The role of working memory inattentional interference of negative information. Cognition and Emotion. in press.

Van Dillen LF, Koole SL. Clearing the mind: a working memory model of distraction from negativemood. Emotion 2007;7:715–723. [PubMed: 18039038]

van Honk J, Tuiten A, van den Hout M, Putman P, de Haan E, Stam H. Selective attention to unmaskedand masked threatening words: relationships to trait anger and anxiety. Personality and IndividualDifferences 2001;30:711–720.

Vuilleumier P, Armony JL, Driver J, Dolan RJ. Effects of attention and emotion on face processing inthe human brain: an event-related fMRI study. Neuron 2001;30:829–841. [PubMed: 11430815]

Watts FN, McKenna FP, Sharrock R, Trezise L. Colour naming of phobia-related words. British Journalof Psychology 1986;77:97–108. [PubMed: 2869817]

Cisler and Koster Page 26

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Weierich MR, Treat TA, Hollingworth A. Theories and measurement of visual attentional processing inanxiety. Cognition and Emotion 2008;22:985–1018.

Wells, A.; Matthews, G. Attention and emotion: a clinical perspective. Lawrence Erlbaum Associates;Hillsdale, NJ: 1994.

Whalen PJ, Kagan J, Cook RG, Davis FC, Kim H, Polis S, et al. Human amygdala responsivity to maskedfearful eye whites. Science 2004;306:2061. [PubMed: 15604401]

Whalen PJ, Rauch SL, Etcoff NL, McInerney SC, Lee MB, Jenike MA. Masked presentations ofemotional facial expressions modulate amygdala activity without explicit knowledge. Journal ofNeuroscience 1998;18:411–418. [PubMed: 9412517]

Williams JMG, Mathews A, MacLeod C. The emotional Stroop task and psychopathology. PsychologicalBulletin 1996;120:3–24. [PubMed: 8711015]

Williams, JMG.; Watts, FN.; MacLeod, C.; Mathews, A. Cognitive psychology and emotional disorders.Wiley; Chichester: 1988.

Williams, JMG.; Watts, FN.; MacLeod, C.; Mathews, A. Cognitive psychology and emotional disorders.2nd ed.. Wiley; Chistester: 1997.

Wilson E, Macleod M. Contrasting two accounts of anxiety-linked attentional bias: selective attention tovarying levels of stimulus threat intensity. Journal of Abnormal Psychology 2003;112:212–218.[PubMed: 12784830]

Yiend J, Mathews A. Anxiety and attention to threatening pictures. The Quarterly Journal of ExperimentalPsychology 2001;54A:665–681. [PubMed: 11548029]

Cisler and Koster Page 27

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Figure 1.Possible representation of the interrelations between the attentional components, mediatingmechanisms, and stage of information processing in attentional biases towards threat amonganxious individuals.

Cisler and Koster Page 28

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Cisler and Koster Page 29

Table 1

Summary of theoretical models in regards to their predictions of the components, mediating mechanisms, andstage of processing in attentional biases towards threat among anxious individuals.

Model Attention Components Mediating Mechanisms Stage of Processing

Williams etal. 1988

Facilitated attention tothreat in HTA;Attentional avoidance inLTA

Affective DecisionMechanism; ResourceAllocation Mechanism

Automatic

Wells &Mathews1994

Facilitated attention tothreat

Voluntary goals (i.e., beliefthat threat monitoring is

important) underliefacilitated attention to threat

Strategic

Ohman, 1996 Facilitated attention tothreat

Amygdala underlies threatdetection Mechanism, which

underlies facilitatedattention, expectancy system

is involved in consciousappraisal of stimulus

Primarily automatic but alsostrategic feedback loop

Beck & Clark,1997

Facilitated attention tothreat; Difficulty indisengagement

Primal mode constrictsattention onto threat; threat-relevant schemas maintains

attention onto threat

Automatic facilitatedattention; strategic difficulty in

disengagement

Mogg &Bradley 1998

Facilitated attention tothreat

Valence evaluationmechanism appraises stimuli

as threatening; goal-engagement System allocates

attention to stimuli

Automatic

Mathews &mackintosh1998

Facilitated attention tothreat

Threat evaluation systemunderlies facilitated

attention; voluntary effort caninhibit facilitated attention

Automatic facilitated attentioncan be overridden by strategic

effort

Bar-haim etal., 2007

Facilitated attention tothreat; Difficulty indisengagement

Threat detection mechanismappraises valence; resource

allocation mechanismunderlies facilitated

attention; threat elaborationmechanism further appraisesstimulus; Goal engagement

mechanism maintains ordisrupts on-going goal

Automatic threat detection andresource allocation; strategicthreat elaboration and goal

engagement

Eysenck et al.2007

Facilitated attention tothreat; Difficulty indisengagement

Attentional Control underliesfacilitated attention via

augmenting stimulus-drivenattention network,

Attentional control underliesdifficulty in disengagement

via impairing the goal-directed attention network

automatic facilitated attention

Clin Psychol Rev. Author manuscript; available in PMC 2011 March 1.