Processing of threat-related affect is delayed in delusion-prone individuals

Processing of threat-related affect is delayed indelusion-prone individuals

M.J. Green*Department of Psychology, University of Sydney, Australia

L.M. WilliamsCognitive Neuroscience Unit, The Brain Dynamics Centre, Westmead Hospital and Department of Psychology,

University of Sydney, Australia

D. J. DavidsonSchool of Psychology, University of New England, Armidale, Australia

Objectives. On the basis of previous reports of an attentional bias for threat-relatedemotional material in deluded schizophrenics (e.g. Bentall & Kaney, 1989), thepresent study examined the proposal that a similar bias would be demonstrated bydelusion-prone individuals, reflected by longer response latencies for the task ofprocessing threat-related facially displayed affects (e.g. anger, fear).

Design. A non-randomized matched group design was employed to examine theperformance of delusion-prone individuals in comparison with a control group.

Methods. 50 psychiatrically healthy participants completed the Peters et al.Delusions Inventory (PDI) as an index of delusional ideation (Peters, Day, & Garety,1996; Peters, Joseph, & Garety, 1999). Subjects were presented with a standard setof facial stimuli depicting happy, sad, neutral, fearful and angry emotion expressions(Mazurski & Bond, 1993). Reaction times for the task of identifying each type ofaffect were compared between groups of high and low scorers on the PDI.

Results. Highly delusion-prone individuals displayed a significant delay inprocessing angry facial expressions in comparison with low scorers on the PDI.

Conclusions. The increased response latency for processing angry expressions wasinterpreted as evidence of attentional bias for material posing a threat to the self,supporting previous cognitive data in relation to deluded patients. Threateningfacial expressions may be regarded with increased significance by delusion-proneindividuals, and it is possible that this bias is involved in the formation of delusionalbeliefs.

* Requests for reprints should be addressed to Melissa J. Green, Department of Psychology, University of Sydney,Camperdown NSW 2006, Australia. (e-mail: [email protected]).

157British Journal of Clinical Psychology (2001), 40, 157–165 Printed in Great Britain

# 2001 The British Psychological Society

The idea that biased selective attention mechanisms may contribute to the formation ofpersecutory delusions is by no means a recent one. For example, Ullmann and Krasner(1969) proposed that a selective attention bias to threatening stimuli may influence thedevelopment of delusional beliefs, such that an abnormal process of data collection couldlead an individual to believe false conclusions about his/her environment, despite logicalreasoning processes in drawing such conclusions. Similarly, Shapiro (1965, p. 67) arguedthat it is the paranoid’s ‘biased seizing of significant clues’, combined with persistentinterpersonal suspiciousness, which characterizes his/her cognitive organization. Giventhat delusional themes commonly reflect concerns about personal safety within the socialmilieu, it is plausible that a selective attention bias for threatening stimuli couldinfluence the perceived significance of this material in relation to the confirmation ofunderlying suspicions about others’ intentions to harm oneself. The present studyinvestigated the presence of such a bias in delusion-prone individuals, to examine itspossible role in the development of delusional beliefs.

A number of recent experiments provide support for the contentions raised above. Forexample, an aberrant style of data gathering has been consistently demonstrated bydeluded individuals during probabilistic reasoning tasks, in which decisions are madeupon the basis of less data in comparison to control groups (Dudley, John, Young, &Over, 1997a; Fear & Healy, 1997; Huq, Garety, & Hemsley, 1988; Garety & Hemsley,1994; Garety, Hemsley, & Wessely, 1991; Peters, Day, & Garety, 1997). Notably, thisresult was exaggerated with the use of emotionally salient material (Dudley, John,Young, & Over, 1997b), which suggests that the perception of interpersonalcommunicative gestures may be particularly vulnerable to selective attention biases.

Other investigations of content-specific information processing in deludedindividuals have revealed evidence of an attentional bias for threat-related emotionwords at the pre-conscious level of processing (Bentall & Kaney, 1989; Fear, Sharp, &Healy, 1996). These studies employed an ‘Emotional Stroop’ task (a variation of theoriginal Stroop, 1935) in which participants were required to name the ink colour ofneutral or affect laden words. Deluded participants displayed increased responselatencies for naming the ink colour of emotion words of a threatening nature incomparison to meaningless letter strings, neutral, anxious, and depressive words. Theincreased latency reflects a reduction in the efficiency of active inhibitory processes thatshould normally allow the focus of attention to remain upon the task of naming the inkcolour of the stimuli. The observed interference (increased latency) suggests thatemotion words of a threatening nature are particularly salient for deluded individuals.

Affect processing deficits in schizophrenia have also been indexed throughbehavioural (see Edwards, Jackson, & Pattison, in press, for a review) andpsychophysiological measures. For example, aberrant visual scan paths have beenassociated with positive schizophrenic symptoms during facial affect encoding tasks(Kee, Kern, & Green, 1998; Loughland, Williams, & Gordon, in press; Manor et al.,1999; Streit, Wolwer, & Gaebel, 1997). Furthermore, schizophrenic patients sufferingpersecutory delusions have been shown to avoid gazing at threatening areas ofambiguous scenes (Phillips, Senior, & David, 2000), and to spend more time viewingphotographs depicting direct and hidden threat, compared to ‘happy’ and ‘potentiallythreatening’ scenes (Freeman, Garety, & Phillips, 2000). Studies of the neural correlatesof threat perception in schizophrenia have also revealed that paranoid schizophrenics

158 M. J. Green et al.

exhibit a differential neural response when viewing threat-related facial expressions(Phillips et al., 1999).

Additionally, a recent investigation of event-related potentials (ERPs) evoked forfacial stimuli showed that schizophrenic participants demonstrate significantly delayedlatencies for all ERP components, with decreased frontal (but enhanced posterior) P200amplitude upon presentation of angry faces (Horley, Gordon, Gonsalvez, Lazzaro, &Bahramali, in press). Since the P200 is considered to reflect decision-making processes,it is possible that the angry face may have activated a conscious process of threat-evaluation in the schizophrenic group.

While the results of these face processing studies support contentions of aberrantthreat-perception in schizophrenia, no studies have investigated hypervigilance forthreatening material in terms of attentional bias for specific facially displayed affects.We are currently undertaking such a project, while the study reported here is the first toexamine the proposal that a similar attentional bias to that found in clinical populationswould be demonstrated by delusion-prone individuals. In the present study it washypothesized that individuals presenting with high levels of delusional ideation woulddemonstrate longer response latencies for the task of identifying threat-related facialexpressions specifically (i.e. anger, fear).

Methods

Participants

The participants were 23 male and 27 female students from undergraduate courses at the University ofNew England who volunteered to take part in the study (N = 50). The age of participants ranged from16 to 46 years (M = 21.76, SD = 5.41), and 90% of participants were aged under 27 years. Allparticipants were paid $5 in compensation for time at the completion of the experimental session.

Materials

IQ. The National Adult Reading Test (NART; Nelson & Willison, 1991) is rapid measure of IQ whichmeasures word-reading skills only. The individual is asked to read a list of 50 words aloud; scores on theNART indicate the number of words that are pronounced incorrectly. This error score can be reliablytranslated into an estimate of WAIS-R IQ (see also Willshire, Kinsella, & Prior, 1991).

Level of delusional ideation. Level of ‘delusional ideation’ was assessed using the Peters et al. DelusionsInventory (PDI-21; Peters et al., 1996). Responses to the PDI-21 are in dichotomous (Yes/No) format.‘Yes’ answers are subsequently assessed on a 5-point Likert scale (numbers 1–5 correspond with increasesin severity) for the dimensions of ‘distress’, ‘preoccupation’ and ‘conviction’. The PDI has demonstratedgood internal consistency and concurrent validity (Peters, Joseph, & Garety, 1999). Criterion validity hasbeen established by persons belonging to religious cults and psychotic in-patients showing significantlyhigher scores on the PDI in comparison to normal controls (Peters, Day, McKenna, & Orbach, 1999).

Facial images. Ten colour slides, including one male (model MA2) and one female (model FE1) face eachdepicting happy, sad, neutral, angry and fearful expressions (Mazurski & Bond, 1993), were presented toparticipants on a VGA monitor in a fixed, random order. The two models from the Mazurski and Bond(1993) series were chosen on the basis of high levels of inter-rater agreement between 100 raters, acrossall expressions during the original standardization procedure. By using only two models we attempted tocontrol for the influence of identity on affect recognition accuracy, and colour slides were considered tobe more ecologically valid than monochrome faces.

159Processing of threat-related affect in delusion-prone individuals

Procedure

Participants were seated approximately 60 cm from the computer monitor, holding the response box.The experimenter controlled the presentation of stimuli manually, with an inter-stimulus interval ofapproximately 5 seconds. Instructions were explained to participants using the following standard rubric:

In a moment you will be presented with a series of faces. Each time a face appears, Iwant you to decide how the person in the photo is feeling—it will be your job tocome up with an emotion word to describe how the person feels. When you havedecided, please press any button on the response box. After you have pressed thebutton, you should say the emotion word out loud so that I can write it down. It ismost important that you press the button before saying the emotion word. Do youunderstand?

These instructions were intended to emphasize accuracy, rather than speed of response. However,participants were not informed that responses could be classed as ‘right’ or ‘wrong’ in order for the task toremain a perceptual, rather than an evaluative one. Reaction times for this task were recorded by havingparticipants press a key when they were ready to identify each affect verbally. No practice trials weregiven.

The actual verbal responses were recorded verbatim by the experimenter. For the entire group, therewere 86 single-word responses that were not identical to the five basic emotion terms corresponding withthe facial stimuli. We implemented a standardized, objective procedure to categorize these ‘variant’responses as one of the basic emotions represented by the stimuli. Variant responses were placed in one ofthe five categories according to consensus ratings across 20 independent volunteers (age range: 20–52;mean age: 28.46) who each completed a questionnaire to indicate which emotion term from the facial set(i.e. happy, sad, neutral, anger, fear) was most closely related in meaning to each variant word.

Results

Delusional ideation

‘Yes’ responses to the PDI formed a measure of delusional ideation (group mean = 9.52;SD = 4.93; range 2–19). Comparison groups were formed on the basis of a median split(median = 9). Participants scoring below the median score on the PDI formed thecontrol group; those scoring on or above the median formed the experimental (delusion-prone) group. Independent group t tests revealed no significant differences between theexperimental groups in terms of age (t(48) = 1.36, p = .18) or scores on the NART(t(46) = 0.02, p = .98) or WAIS-R derivative (t(46) = 0.05, p = .95). Participants’characteristics can be seen in Table 1.

Table 1. Participants’ characteristics (PDI-21)

AgeSex

NARTGroup N (years) M F errors WAIS–R PDI-21

Control 24 22.83 9 15 18.73 107.36 5.04(low PDI) (6.82) (7.28) (9.0) (1.68)Experimental 26 20.77 14 12 18.69 107.23 13.65(high PDI) (3.52) (6.09) (7.59) (2.83)

Note. Means and standard deviations (in parentheses).


Affect recognition

The mean number of correctly identified facial expressions (including neutralexpressions) for the entire sample was 6.74 (SD = 1.35). These data indicate that onthe whole, participants identified facial affects with approximately 50–80% accuracy.An independent samples t test revealed no differences in the number of emotionscorrectly identified by high vs. low scorers on the PDI (t(48) = 1.1, p = .28).

Speed of processing

No differences were found between control and experimental groups in mean reactiontimes (RTs) across the entire series of stimuli. Mean affect-specific RTs for experimentaland control groups are presented in Table 2 (see also Fig. 1). Multivariate analysis of

Table 2. Mean reaction times (msec) for specific affects by experimental and controlgroups

Affect

Group Anger Fear Neutral Happy Sad

Control 3 056.12 3 236.08 3 845.18 2 298.33 4 151.75(low PDI) (1 615.33) (1 570.06) (1 907.86) (1 516.90) (2 111.1)Experimental 4 149.50 3 738.02 4 083.09 2 907.25 4 229.15(high PDI) (1 666.87) (1 370.21) (2 002.90) (1 692.50) (1 596.46)

Note. Standard deviations in parentheses.

Figure 1. Mean reaction times for specific affects in control and experimental groups.


variance was performed on affect-specific RT scores, with ‘PDI group’ (high vs. low) asthe between-participant variable, and ‘affect’ (happy, sad, neutral, anger, fear) as within-participant variables. This revealed a significant main effect for ‘affect ’(F(1,48) = 12.98, p < .000). One-way ANOVAs for each affect revealed that thegroup of highly delusion-prone individuals displayed significantly longer RTs for thetask of naming the angry facial expression in particular (F(1,48) = 5.53, p = .02). Nodifferences were revealed between these groups in the speed of processing any other facialexpressions. Possible effects of ‘age’ (F(1,48) = 2.1, p = .15) and ‘IQ’ (F(1,48) = 0.88,p = .352) on RTs for angry faces were ruled out using MANCOVAs with these variablesincluded as covariates.

Discussion

The group of highly delusion-prone individuals displayed a significant delay inprocessing angry facial expressions in comparison with participants reporting low levelsof delusional ideation. There are several possible explanations of this finding; however,the interpretations offered here must be regarded as speculative given that this is thefirst time this has been demonstrated in delusion-prone subjects. Of particular concernis whether the observed difference for angry faces is due to a relatively shorter latency inthe control group (in comparison to reaction times for other facial expressions) that isnot exhibited by the delusion-prone group (rather than a relatively longer latency in thedelusion-prone group; see Fig. 1). Indeed, the results of a recent study employing visualsearch tasks suggest that non-psychiatric participants process angry faces faster thanother expressions (Fox et al., 2000) in line with an evolutionary view. While thisinterpretation fits well with the present data, it does not preclude the significance of thecomparative delay exhibited by delusion-prone individuals when processing angry faces.

The delayed response of delusion-prone individuals may represent a bias for threat-related material, consistent with previous findings of pre-conscious attentional bias indeluded individuals (Bentall & Kaney, 1989; Fear et al., 1996). However, an importantdifference between the results from emotional Stroop tasks and the present paradigm isthat the former represents a covert (pre-conscious) bias of selective attention, while thepresent finding suggests the presence of a bias occurring later, during overt (conscious)appraisals of the angry face stimuli (such that the stimulus has seemingly ‘captured’one’s attention). Thus, the increased latency for naming angry faces could reflectdifficulty in disengaging this material from consciousness, and possibly increased‘threat’ appraisal, thereby impeding the timeliness of the desired response. Indeed, thismay reflect the salience with which threatening material is perceived by paranoidindividuals.

The apparent convergence of evidence across pre-conscious and conscious levels ofattention purports to the existence of a bias throughout the cognitive system (i.e.including an initial orienting bias for threatening stimuli, and subsequent difficulty inconsciously shifting this material from the focus of attention). It is possible that a pre-attentive bias could unduly influence the perceived significance of stimuli posing athreat to the self-concept; that is, angry faces might be endowed with increasedsignificance simply because they are impervious to selective attention filtering (i.e. ashas been similarly argued by Anscombe, 1987). Alternatively, it is possible that the


overt bias reflects a preoccupation with external stimuli that confirms pre-existingparanoid suspicions. For example, an attentional bias for material related to the theme ofcurrent delusions has been demonstrated by Leafhead, Young, and Szulecka (1996), andRossell, Shapleske, and David (1998) showed that deluded schizophrenics tend to bemore believing of ambiguous or implausible information which is congruent with thecontent of their delusions.

Given the present findings in an ‘at risk’ sample, it could be argued that prodromalprocesses of delusion formation are comparable to the information processing biasesrevealed in clinically deluded groups. That is, pre-conscious selective attentionmechanisms of delusion-prone individuals may be biased toward the confirmation ofquasi-delusional ideas. The present results suggest that this bias is active prior to theonset of dysfunctional delusional beliefs, thereby implicating its role in the formation ofdelusions. An alternative argument is that those individuals scoring highly on the PDIare already mildly deluded, such that the apparent bias may serve only to maintainexisting delusions. Empirical evidence against the latter view can be drawn from reportsindicating that only a small percentage of psychosis-prone individuals actually sufferdeviant psychotic experiences (e.g. Chapman & Chapman, 1987). Moreover, while it isentirely plausible that individuals with high PDI scores could already hold some falsebeliefs, we have argued that the proposed attentional bias may contribute to theformation of maladaptive delusions.

The present results in delusion-prone subjects are also consistent with electro-physiological reports of delayed ERP components in schizophrenia when viewing angryfacial expressions (Horley et al., in press). As noted by these authors, the observed frontaldelay of the N200 component in schizophrenic participants suggests that decision-making processes are employed in relation to the angry face during controlled (later)stages of processing. Additionally, the delayed P300 component suggests that aberrantcontextual information may be associated with angry faces in schizophrenia. Takentogether, a plausible interpretation of these and the present findings is that delusion-prone individuals might exhibit extended processing of angry faces due to context-dependent information associated with threatening stimuli. Thus, it is conceivable thatconscious decision-making processes might be activated in paranoid individuals whenprocessing threatening facial expressions, for the purpose of assessing the severity of thethreat. In support of this contention, Le Doux (1998) suggests that fearful situations areprocessed rapidly via subcortical pathways between the thalamus and amygdala, whileslower parallel processes involving both cortico-thalamic networks and the hippocampalformation provide higher level sensory and ‘contextual’ (sensory-independent)information, respectively. Notably, cell abnormalities (Benes, 1999) and reducedvolume of hippocampal and amygdala regions (Bogerts et al., 1993) have been associatedwith positive symptoms of schizophrenia. It might therefore be prudent to employconcurrent electrophysiological measures during future investigations of threatperception in schizophrenia, in order to gather more definitive temporal informationwith regard to the proposed cognitive bias.

Given that the present result is highly preliminary, the explanations offered hereshould be regarded as tentative. Future attempts to replicate these findings might payparticular attention to the pattern of responses within each group in order to clarifywhether the difference in reaction times for angry faces is due to a shorter latency by the


control group rather than increased latency in the delusion-prone group. This findingwould not be inconsistent with the interpretations presented above, in that a relativelack of speed for processing angry faces by delusion-prone individuals could alsorepresent conscious appraisal of threat that is not undertaken by control subjects. Furtherstudy of this phenomenon in relation to delusion formation is also recommended, withthe use of more complex social stimuli to establish whether delusion-prone and deludedindividuals exhibit an attentional bias for threatening faces within realistic socialcontexts.

References

Anscombe, R. (1987). The disorder of consciousness in schizophrenia. Schizophrenia Bulletin, 11, 241–260.

Benes, F. M. (1999). Evidence for altered trisynaptic circuitry in schizophrenic hippocampus. BiologicalPsychiatry, 46, 589–599.

Bentall, R. P., & Kaney, S. (1989). Content specific information processing and persecutory delusions:An investigation using the emotional Stroop test. British Journal of Medical Psychology, 62, 355–364.

Bogerts, B., Lieberman, J. A., Ashtari, M., Bilder, R. M., DeGreef, G., Lerner, G., Johns, C., & Masiar, S.(1993). Hippocampus-amygdala volumes and psychopathology in chronic schizophrenia. BiologicalPsychiatry, 33, 236–246.

Chapman, L. J., & Chapman, J. P. (1987). The search for symptoms predictive of schizophrenia.Schizophrenia Bulletin, 13, 497–503.

Dudley, R. E. J., John, C. H., Young, A. W., & Over, D. E. (1997a). Normal and abnormal reasoning inpeople with delusions. British Journal of Clinical Psychology, 36, 243–258.

Dudley, R. E. J., John, C. H., Young, A. W., & Over, D. E. (1997b). The effect of self-referent materialon the reasoning of people with delusions. British Journal of Clinical Psychology, 36, 575–584.

Edwards, J., Jackson, H. J., & Pattison, P. E. (in press). Emotion recognition via facial expression andaffective prosody in schizophrenia: A methodological review. Clinical Psychology Review.

Fear, C. F., & Healy, D. (1997). Probabilistic reasoning in obsessive compulsive and delusional disorders.Psychological Medicine, 27, 199–208.

Fear, C., Sharp, H., & Healy, D. (1996). Cognitive processes in delusional disorders. British Journal ofPsychiatry, 168, 1–8.

Fox, E., Lester, V., Russo, R., Bowles, R. J., Pichler, A., & Dutton, K. (2000). Facial expressions ofemotion: Are angry faces detected more efficiently? Cognition and Emotion, 14, 61–92.

Freeman, D., Garety, P. A., & Phillips, M. A. (2000). An examination of hypervigilance for externalthreat in individuals with generalised anxiety disorder and individuals with persecutory delusionsusing visual scanpaths. Quarterly Journal of Experimental Psychology, 53A(2), 549–567.

Garety, P., & Hemsley, D. R. (1994). Delusions: Investigations into the psychology of delusional reasoning.Oxford: Oxford University Press.

Garety, P. A., Hemsley, D. R., & Wessely, S. (1991). Reasoning in deluded schizophrenic and paranoidpatients: Biases in performance on a probabilistic inference task. Journal of Nervous and Mental Disease,179, 194–201.

Horley, K., Gordon, E., Gonsalvez, C., Lazzaro, I., & Bahramali, H. (in press). Event-related potentials tofaces in schizophrenia. International Journal of Neuroscience.

Huq, S. F., Garety, P. A., & Hemsley, D. R. (1988). Probablilistic judgements in deluded and non-deluded subjects. Quarterly Journal of Experimental Psychology, 40A, 801–812.

Kee, K. S., Kern, R. S., & Green, M. F. (1998). Perception of emotion and neurocognitive functioning inschizophrena: What’s the link? Psychiatry Research, 81, 57–65.

Le Doux, J. (1998). The emotional brain. London: Weidenfeld & Nicolson.Leafhead, K. M., Young, A. W., & Szulecka, T. K. (1996). Delusions demand attention. Cognitive

Neuropsychiatry, 1, 5–16.


Loughland, C. M., Williams, L. M., & Gordon, E. (in press). Visual scanpaths to positive and negativefacial emotions in an outpatient schizophrenia sample. Journal of Neuropsychiatry and ClinicalNeurosciences .

Manor, B., Gordon, E., Williams, L., Rennie, C., Bahramali, H., Latimer, C., Barry, R., & Meares, R.(1999). Eye movements reflect impaired face processing in patients with schizophrenia. BiologicalPsychiatry, 46, 963–969.

Mazurski, E. J., & Bond, N. W. (1993). A new series of slides depicting facial expressions of affect: Acomparison with the Pictures of Facial Affect series. Australian Journal of Psychology, 45, 41–47.

Nelson, H. E., & Willison, J. (1991). National Adult Reading Test (NART) (2nd ed.). Windsor: NFER-NELSON.

Peters, E. R., Day, S., & Garety, P. A. (1996). The Peters et al., Delusions Inventory (PDI): New normsfor the 21-item version. Schizophrenia Research, 18, 118.

Peters, E. R., Day, S., & Garety, P. A. (1997). From preconscious to conscious processing—where doesthe abnormality lie in delusions? Schizophrenia Research, 24, 120.

Peters, E. R., Day, S., McKenna, J., & Orbach, G. (1999). Delusional ideation in religious and psychoticpopulations. British Journal of Clinical Psychology, 37, 83–96.

Peters, E. R., Joseph, S., & Garety, P. A. (1999). The measurement of delusional ideation in the normalpopulation: Introducing the PDI (Peters et al., Delusions Inventory). Schizophrenia Bulletin, 25, 553–576.

Phillips, M. L., Senior, C., & David, A. S. (2000). Perception of threat in schizophrenics with persecutorydelusions: An investigation using visual scanpaths. Psychological Medicine, 30, 157–167.

Phillips, M. L., Williams, L. M., Senior, C., Bullmore, E. T., Brammer, M. J., Andrew, C., Williams, S.,& David, A. S. (1999). A differential neural response to threatening and non-threatening negativefacial expressions in paranoid and non-paranoid schizophrenics. Psychiatry Research: NeuroimagingSection, 92, 11–31.

Rossell, S. L., Shapleske, J., & David, A. S. (1998). Sentence verification and delusions: a content specificdeficit. Psychological Medicine, 28, 1189–1198.

Shapiro, D. (1965). Neurotic styles. New York: Basic Books.Streit, M., Wolwer, W., & Gaebel, W. (1997). Facial affect recognition and visual scanning behaviour in

the course of schizophrenia. Schizophrenia Research, 24, 311–317.Stroop, J. R. (1935). Studies of interference in serial verbal relations. Journal of Experimental Psychology,

18, 643–662.Ullmann, L. P., & Krasner, L. (1969). A psychological approach to abnormal behaviour. Englewood Cliffs, NJ:

Prentice Hall.Willshire, D., Kinsella, G., & Prior, M. (1991). Estimating WAIS–R IQ from the National Adult

Reading Test: A cross-validation. Journal of Clinical and Experimental Neuropsychology, 13, 204–216.

Received 1 November 1999; revised version received 1 May 2000


Documents

Processing of threat-related affect is delayed in delusion-prone individuals