Schizophrenia Research 133 (2011) 199–204

Contents lists available at SciVerse ScienceDirect

Schizophrenia Research

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

Jumping to conclusions in psychosis: A faulty appraisal

José Luis Rubio a,b, Miguel Ruiz-Veguilla a,⁎, Laureno Hernández a,b, María Luisa Barrigón a,c,María Dolores Salcedo a, Josefa María Moreno a, Emilio Gómez b, Steffen Moritz d, Maite Ferrín e,f

a Developmental Neuropsychiatry Research Unit, Hospital Virgen del Rocio de Sevilla. Servicio de Psiquiatría, Spainb University of Granada, Faculty of Psychology, Department of Experimental Psychology, Granada, Spainc Psychiatry Service, Santa Ana Hospital, Motril, Granada, Spaind University Medical Center Hamburg-Eppendorf, Department of Psychiatry and Psychotherapy, Hamburg, Germanye Department of Child and Adolescent Psychiatry, Institute of Psychiatry, London, UKf Centro de Salud Mental de Estrella, Hospital Garcia Orcoyen, Camino de Logroño 4, 31200. Estela. Navarra, Spain

Abbreviations: JTC, jumping to conclusions; DTDplausibility rating at first stage.⁎ Corresponding author at: Developmental Neuropsy

Virgen del Rocio de Sevilla. Servicio de Psiquiatría, A41013, Spain.

E-mail address: miguel.ruiz.veguilla.sspa@juntadean

0920-9964/$ – see front matter © 2011 Elsevier B.V. Aldoi:10.1016/j.schres.2011.08.008

a b s t r a c t

a r t i c l e i n f o

Article history:Received 2 April 2011Received in revised form 31 July 2011Accepted 15 August 2011Available online 9 September 2011

Keywords:SchizophreniaJumping to conclusionsPictures to decision taskFaulty appraisal

Schizophrenia patients, particularly those with current delusions, show a cognitive bias known as jumping toconclusions, defined as a decision made quickly on the basis of little evidence. The aim of this work was toexamine the underlying mechanisms of this cognitive bias by means of the Picture To Decision Task, whichallows one to analyse the effect of the context on decisions made. We compared the performance of this taskby 42 psychotic patients, 21 siblings of these patients and 77 controls. The results of the current study suggestthat, relative to siblings and controls, patients display a general tendency to jump to conclusions,characterised by overestimating the conviction in their choices at the beginning of the decision processand by a lowered threshold for making decisions in ambiguous contexts, where a greater amount ofinformation is required. These results are interpreted in terms of faulty appraisal, which would be the firstmechanism responsible for the Jumping To Conclusions bias. Theoretical and clinical implications arediscussed.

, drawing to decision; PR-1,

chiatry Research Unit, Hospitalvda.Manuel Siurot s/n. Sevilla

dalucia.es (M. Ruiz-Veguilla).

l rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Several studies have demonstrated that patients with schizophre-nia show a reasoning bias known as jumping to conclusions (JTC),defined as a decision made quickly on the basis of little evidence.Schizophrenia patients, particularly thosewith current delusions, mayoverestimate and use less information to arrive at a decision in tasksthat require them to integrate information tomake a response (Huq etal., 1988; Moritz and Woodward, 2005; Moritz et al., 2007; Speechleyet al., 2010). Similarly, JTC bias has been reported in close relatives ofschizophrenia patients (van Dael et al., 2006) and individuals at a highclinical risk of psychosis (Broome et al., 2007), and it may beassociated with higher levels of conviction in paranoid thoughtswithin the general population (Freeman et al., 2008; Lincoln et al.,2010).

Although there is no unified explanation of the origin of thiscognitive bias, two specific formal hypotheses might be considered(Averbeck et al., 2011). The first hypothesis is that patients over-

estimate the conviction in their choices at the beginning of thedecision-making process (Huq et al., 1988; Lincoln et al., 2010;Speechley et al., 2010). The second hypothesis is that they may have alowered threshold for making decisions, and thus use less informationin arriving at a decision, which is consistent with the so-called liberalacceptance account (Moritz et al., 2009; Veckenstedt et al., 2011).

The principal aim of this work is to contrast the two hypothesescited above bymeans of a new version of the drawing to decision task.This task has been used previously in the study of another cognitivebias related to JTC called “bias against disconfirmatory evidence”(Moritz and Woodward, 2006) and comprises the metacognitivetraining program for schizophrenia patients (Moritz et al., 2011). Likethe beads task (Huq et al., 1988), which is the task most used in thestudy of JTC, the principal dependent measures are the plausibilityrating of each stimulus presented and the amount of informationneeded to reach a final decision about the identity of the depiction.These two measures are analysed in two kinds of trial (“cued” and“uncued”; that is, with and without interpretative cues). This is aspecific characteristic of the task, allowing us to analyse the effect ofthe context in which the decisions are made.

Exploration of both hypotheses through the same task cancontribute to extending the previous results about JTC bias in twoways. Firstly, they provide a unified explanation of themany proposedcauses at the origin of this bias. Secondly, an analysis of the contextwill allow us to discover if this bias is only present when subjects have

200 J.L. Rubio et al. / Schizophrenia Research 133 (2011) 199–204

been instructed to derive their own interpretations about reality, orwhen the context of a decision has been previously defined byinterpretative cues. Moreover, the results from all three groups(schizophrenic patients, their siblings and controls) can be comparedto reflect the hypothesised order of liability to psychosis according tothe studies cited above.

A second aim of this work is to explore associations between theJumping To Conclusions parameters of our task, psychotic symptoms,executive functioning and theory of mind. Jumping To Conclusions andtheory of mind are typically found to be associated with positivesymptoms, and executive functions with negative symptoms. However,empirical evidence for these associations is often not well founded.Recently, Woodward et al. (2009) applied a multivariate approach toassess this pattern of associations. Their results suggest that the JTC biasis related to executive functioning andmay be independent of theory ofmind deficit and positive symptoms. In short, this is an open questiontowards which our study may provide new insights.

1.1. Participants

Overall, 140 subjects took part in the study. The clinical group wasmadeup of 42 consecutive subjects attended in the in-patients unitwhopresented psychosis symptoms (DSM-IV: 295-297-298, and 296 withpsychotic codes) (American Psychiatric Association, 2000). The siblinggroup comprised 21 subjects, while the control group consisted of 77healthy subjects. None of the patients had been diagnosed more thanfive years earlier, a datum corroborated by the patients' clinical historyand informationprovidedby their relatives. Allwere inpatients from thepsychiatry area of the “Complejo Hospitalario de Jaén” (Spain). All theparticipants also met the following exclusion criteria: absence ofcerebral damage and no clinical evidence of drug abuse during thecourse of the study.

1.2. Procedure

Assessment of the patients' psychotic symptoms was carried outon their arrival at the Hospital. Each patient underwent a semi-structured interview that included the modules of psychotic symp-toms and mood state of the Structure Clinical Interview for DSM-IV(First et al., 1997).

The presence and intensity of psychotic symptoms were assessedby means of the PANSS scale at admission to hospital (Kay et al.,1987). The PANSS has been validated in a Spanish population ofschizophrenic patients (Peralta and Cuesta, 1994). Sibling and controlgroups were screened through the Mini International Neuropsychi-atry Interview (MINI; Sheehan et al., 1998). We applied the five factorPANSS model described by van der Gaag et al.(2006).

All patients carried out the different experimental tasks on thedischarge day. The Pictures Decision Task and The Degraded FacialAffect Recognition task were applied to the three groups (control,sibling and psychosis groups). The Hinting Task and The AttentionalNetwork Task were applied only to the psychosis group.

1.2.1. Cannabis useThe consumption of cannabis was recorded using the L section of

the International Diagnostic Interview (Robins, et al., 1988). Weclassified subjects as “heavy users of cannabis”when the frequency ofuse during the period of maximum consumption was daily or nearlydaily for at least a month (Ruiz-Veguilla et al., 2009).

1.2.2. The Pictures Decision TaskOur experimental task is a version of the picture task created by

Moritz et al. (2007). Six experimental trials, following two practicetrials, were presented. Every trial consisted of a sequence of eightstages, each showing a common object that was increasinglydisambiguated by decreasing degrees of visual fragmentation: new

object features were added to each new picture until, eventually, theentire object was displayed in the final stage. The objects weredepicted as post-edit simple black and white drawings. Instructionsand trials were presented using a Microsoft computer. The trials wererun in a fixed order: half the trials (1st, 3rd and 5th) wereaccompanied by six interpretative cues about the identity of theobject displayed over the eight stages; we call these “cued trials”. Inthese trials, participants chose one of the eight cues and their plausibilitywas then ratedusing afive-point Likert scale (1=dismissed, 2=unlikely,3=possible, 4=likely, 5=positive decision). In the remaining trials(2nd, 4th and 6th), no interpretative cues were provided (uncued trials)and the participants were instructed to derive their own interpretations,whichwere subsequently rated for plausibility in the sameway as for thecued trials. Once a decision was made that met with the highestplausibility rating (pressing F5; positive decision), that trial ended and anewtrialwaspresented. Examples of the task canbe seen inAppendices 1(cued trial) and 2 (uncued trial).

In this task, different parameters could be calculated and then usedto provide further insight into the underlying mechanisms of JTC bias.Specifically, five parameters were calculated: Jumping To Conclusionsat first stage (JTC-1), Plausibility Rating at first stage (PR-1), Draws ToDecision (DTD), Time Response at first stage (TR-1) and TimeResponse for Draw To Decision (TR-DTD).

Jumping To Conclusions at first stage (JTC-1) was defined in atleast one of the six experimental trials, with only the first stage beingneeded to decide with absolute certainty the identity of the particularobject (by pressing “F5”=positive decision). This cut-off was adoptedbecause it was considered to be the most definite expression of such areasoning bias and because it is very similar to the parameters used inthe Beads Task.

Plausibility Rating at first stage (PR-1) was defined as the meanplausibility rating at the first stage for cued and uncued trials (range 1to 5). Hence, this parameter serves as a measure of the level ofconviction of beliefs when there is only a little information, whichserves a measure of the first hypothesis proposed in the introductionof this work.

Draws To Decision (DTD) was defined as the mean number ofstages for cued and uncued trials necessary for the participant to reacha final decision about the identity of the objects with absolutecertainty (range 1 to 8; the total number of stages per trial), whichserves as a measure of the second hypothesis proposed. Finally, TimeResponse analyses (TR-PR1 and TR-DTD) were conducted in order toexplore whether patients were faster than siblings and controls,which might reveal differences between groups for time responseparameters. The results for the remaining experimental parameterscould then be better explained in terms of impulsivity.

1.2.3. The Attentional Network TaskThe Attentional Network Task (ANT) was used to assess the

functioning of three attentional networks of the Posner's attentionalmodel (alertness, executive control and orientation) (Posner andPetersen, 1990). The function of the executive control network isprocessing task-relevant information, and it is intimately associatedwith executive functions (Posner and Fan, 2005). The function of theorientation network is to select sensory stimuli, whereas the functionof alertness is to obtain and maintain a state of vigilance. The task wasto identify, as soon as possible, the direction in which an arrowappearing in the center of the screen was pointing (left or right). Theefficiency of the three attentional networks was calculated from thelatency of responses in the different experimental conditions. Eachexperimental session involved a practice block of 24 trials, and threeexperimental blocks comprising 96 trials each.

1.2.4. The Hinting taskTheory of mind was assessed with the Hinting task as described

by Corcoran et al., in which an individual is required to infer real

201J.L. Rubio et al. / Schizophrenia Research 133 (2011) 199–204

intentions behind indirect speech (Corcoran et al., 1995). The taskentails 10 short passages presenting an interaction between twocharacters ending with one of the characters dropping an obvioushint. The subject is then asked what the character really meant whenhe/she said this. If the subjects failed to give the correct response, aneven more obvious hint is added to the story. A correct response istherefore scored as 2 or 1 depending on when the response wasgiven.

1.2.5. The Degraded Facial Affect RecognitionDegraded Facial Affect Recognition is a forced-choice facial affect

labeling task of degraded faces (Van 't Wout et al., 2004) in which allphotographs of faces were passed through a filter that reduced visualcontrast by 30%. Photographs of 4 different actors, 2 male and 2female, were used from the set of faces developed and used by Frigerioet al. (2002). Sixty-four trials were presented on a computer screenand consisted of 16 face presentations in each of 4 conditions: angry,happy, fearful, and neutral. Subjects were asked to indicate theexpression of each face by forced-choice with button press (F1–F4).The labels “angry”, “happy”, “fearful” and “neutral” were alsopresented at the bottom of the screen to remind subjects of thedifferent categories. Subjects were asked to work as accurately aspossible; no time limit was given.

1.2.6. Statistical analysesStatistical analyses were carried out with SPSS Version 15. For all

analyses, the scores of the five experimental parameters of the pictureto decision task were compared between the three groups (patients,siblings and controls), reflecting the hypothesised order of liability forpsychosis. Cannabis consumption and number of total errors inDegraded Facial Affect Recognition were introduced as covariables inall analyses. Attentional Network Task and Hinting Task scores werenot included because they were only done by the patients group. Achi-square test was used to examine JTC-1, so that it constituted acategorical variable; for the other four parameters (PR-1, DTD, TR-1and TR-DTD), multivariate analysis of variance was conducted withgroup as the inter-subject variable and trial type (cued and uncued) asthe intra-subject variable. Post hoc analysis of these four parameterswas carried out in order to examine separately at which levels of eachtrial type (cued and uncued) differences were seen among the threegroups. Finally, correlational analyses were conducted in the patientsgroup in order to examine the relationships among the differentparameters with the tasks described previously: Attentional NetworkTask and Hinting Task.

Table 1Sociodemographic variables.

Patients (N=42) Siblings

N (%)Gender (males) 22 (52%) 11 (52%Heavy cannabis consumer 15 (36%) 6 (28%)

Median (SD)Age (years) 30.33 (10.14) 28.65 (9Years of formal education⁎ 8.57(3.59) 10.80 (3Errors in facial affect recognition⁎ 23.65 (5.87) 18.28 (5Five factor PANSS⁎⁎

Positive 20.88 (4.15)Negative 17.36 (9.48)Disorganisation 18.67 (6.44)Excitement 11.67 (3.51)Emotional distress 7.19 (3.01)

⁎Differences Significant at pb0.05.⁎⁎The five factor PANSS model was taken from van der Gaag et al.(2006).

2. Results

2.1. Sociodemographic and psychopathological variables

The mean age of patients was slightly higher than that of the othergroups; however, no significant differences regarding age, gender orcannabis consumption were found. In terms of variables, “years offormal education” and “errors in facial affect recognition” showedsignificant differences (pb0.05). There were no differences incannabis use among the three groups. Fifteen (36%) of patientswere heavy cannabis users, as were 6 siblings (28%) and 35 (45%) ofcontrols (p=0.06; see Table 1).

In the group of patients, thirty three (79%) had been diagnosed witha first episode of psychosis and themean duration of illness in the othernine (21%) was 2 years (SD=0.4). Seven (16%) were diagnosed asbipolar disorder with psychosis symptoms (mania), 8 (19%) werediagnosed with acute psychosis, 3 (7%) with schizoaffective and 26(60%) with schizophrenia according to DSM-IV criteria.

2.2. Jumping to conclusions, plausibility rating at first stage and draws todecision results

Out of the whole sample, 19 (13.6%) of the participants showed aJTC-1 (in at least one of the 6 trials, needing only the first stage todecide with absolutely certainty), which breaks down to 16 out of the42 in the patients group (38.1%), 0 of 21 siblings (0%), and 3 of 77participants in the control group (3.9%). Results were different bygroup, χ2(2.13)=28.20, pb0.0001). After adjustment by cannabisconsumption and errors in facial affect recognition, results remainedstatistically significant (p=0.003).

Post hoc analysis shows significant differences between patientsand siblings (p=0.001) and between patients and controls(p=0.0001), but no significant difference between siblings andcontrols (p=0.789). Results of JTC-1 suggest that patients display atendency to jump to conclusions on the basis of little evidence, that is,after just one stage.

These differences in JTC-1 were explored carefully by means of theplausibility rating at the first stage (PR-1) so that this parameterwould reflect the intensity of beliefs at the first stage. As can be seen inFig. 1, PR-1 was higher for the three groups in the cued trials than inthe uncued trials; moreover, the patients showed the highest PR-1 forboth types of trial (cued and uncued). Results were different by group(F (2.13)=8.28, pb0.0001). After adjustment by cannabis consump-tion and errors in facial affect recognition results remained statisti-cally significant for cued (p=0.001) and for uncued trials (p=0.005).

(N=21) Controls (N=77)

) 40 (52%) χ²(2.138)=3.56; p=0.17235 (45%) χ²(2.138)=5.60; p=0.061

.57) 28.21 (5.68) F(2.138)=0.83; p=0.439

.69) 12.59 (3.61) F(2.138)=39.40; p=0.001

.50) 19.57 (5.30) F(2.138)=7.52; p=0.001

Fig. 1. Plausibility rating at first stage (PR-1) for the three groups. a) Significantdifferences between patients and siblings. b) Significant differences between patientsand controls. Vertical axis reflects the range of this parameter (1–5, the five points ofplausibility rating scale).

202 J.L. Rubio et al. / Schizophrenia Research 133 (2011) 199–204

Post hoc analysis shows significant differences between patientsand siblings for cued trials (Mean=3.06, SD=1.01 vs. Mean=2.31,SD=0.86; p=0.002) and for uncued trials (Mean=2.77, SD=0.96vs. Mean=1.96, SD=0.78; p=0.004). Likewise, there were signifi-cant differences between patients and controls for cued trials(Mean=3.06, SD=1.01 vs. Mean=2.62, SD=0.77; p=0,009) andfor uncued trials (Mean=2.77, SD=0.96 vs. Mean=2.62, SD=0.14;p=0.003), yet not between siblings and controls for cued (p=0.145)or for uncued trials (p=0.245). PR-1 data analysis shows that thepatients group displays a high level of conviction at the first stage,whereas in siblings and controls, this tendency is not present. Thisresult suggests, therefore, that the first hypothesis proposed (anoverestimate of the information shown) is involved in JTC for bothcued and uncued trials.

On the other hand, with respect to Draws To Decision (DTD),results show that all three groups needed a greater number of stagesin order to make a final decision about the identity of objects inuncued trials (see Fig. 2). Results were different by group only foruncued trials F(2,138)=11.425, p=0.001. After adjustment bycannabis consumption and facial affect recognition, results remainedstatistically significant (p=0.005).

Post hoc analysis revealed significant differences between patientsand siblings for uncued trials (Mean=5.53, SD=0.20 vs.Mean=7.04,SD=0.28; p=0.001), unlike the cued trials (Mean=3.78, SD=0.19 vs.Mean=4.15, SD=0.27; p=0.275). There were also significant differ-ences between patients and controls for uncued trials (Mean=5.53,SD=0.20 vs. Mean=6.83 SD=0.14; p=0.001) but not for cued trials(Mean=3.78, SD=0.19 vs.Mean=3.19, SD=0.14; p=0.458). In turn,there were no significant differences between siblings and controlsfor cued trials (Mean=4.15, SD=0.27 vs. Mean=3.19, SD=0.14;

Fig. 2. Number of stages necessary to take a final decision (DTD) for the three groups. a)Significant differences between patients and siblings. b) Significant differencesbetween patients and control. Vertical axis reflects the range of this parameter (1–8,the total number of stages per trial).

p=0.543) or for uncued trials (Mean=7.04, SD=0.28 vs.Mean=6.83SD=0.14; p=0.511).

DTD results show that in comparison with siblings and controls,patients needed fewer stages to arrive at a final decision in the uncuedtrials. Interesting indeed was the finding that controls and siblingsneeded approximately three more stages to reach a decision inuncued trials than in cued trials (see Fig. 2). However, among patientsthis tendency was clearly diminished, involving approximately oneand a half stages more.

In order to explore whether the psychosis group was moreimpulsive than the control and sibling groups, we compared theresults of Time Response at first stage (TR-1) and Time Response untila Final Decision (TR-DTD) among the three groups (Table 2). Timeresponses proved to be very similar for TR-1 in the three groups, incued and uncued trials. Time responses for TR-DTD in the three groupswere higher for uncued trials than for cued ones, indicating that thedecision process is slower with this condition, and participants mustapparently create their own interpretations through sequences ofeight stages. No significant association was found between TR-1 andgroup (F(2.13)=0.25; p=0.77) or between TR-DTD and group(F(2.13)=0.97; p=0.37). For TR-1, the post hoc analysis did notshow any significant differences among the three groups for cuedtrials (patients/siblings, p=0.720; patients/controls, p=0.129;siblings/controls, p=0.470); and the same was true for uncued trials(patients/siblings, p=0.138; patients/controls, p=0.732; siblings/controls, p=0.193). Similarly, post hoc analysis for TR-DTD gave nosignificant differences among groups for cued trials (patients/siblings,p=0.159; patients/controls, p=0.648; siblings/controls, p=0.258.However, in uncued trials, the TR-DTD was significantly differentbetween the patient and the control group (p=0.002) yet not for theother two comparisons (patients/siblings, p=0.870; siblings/control,p=0.900).

The results of TR-1 suggest that the effect of PR-1 (plausibilityratings at first stage) is not conditioned by impulsivity, for whichreason time responses are very similar for the three groups. In the caseof TR-DTD, patients are a little quicker than controls for uncued trials,a difference which may perhaps be influenced by impulsivity. Thiswould appear less likely if we bear in mind that patients needapproximately three stages fewer to reach a final decision.

2.3. Correlational analyses

Finally, in order to explore the pattern of associations among thepicture to decision task parameters with executive function, theory ofmind and psychotic symptoms, correlational analyseswere conductedin the patients group. Analyses specifically involved JTC-1, PR-1 andDTD parameters, the control executive network of the AttentionalNetwork Task, the Hinting Task and PANSS scores (Table 3).

Results showed positive significant associations between JTC-1and PR-1 for cued trials (r=.712; p=0.0001) and for uncued trials(r=.486; p=0.001) and negative associations between PR-1 andDTD for cued trials (r=−.621; p=0.0001) and for uncued trials (r=−.721; p=0.0001). That is to say, when a strong conviction exists atthe beginning of the decision-making process (the first hypothesis),reflected by a high PR-1, fewer stages are necessary to arrive at a finaldecision, reflected by a low DTD (the second hypothesis). There was

Table 2Mean and standard deviation, in parentheses, for time response at first stage (TR-1) andtime response until a final decision was made (TR-DTD) in seconds.

Parameter Psychosis group Siblings Healthy controls

TR-PR1 cued trials 28.12 (1.72) 22.60 (2.50) 24.69 (1.43)TR-PR1 uncued trials 25.48 (1.73) 25.15 (2.51) 24.23 (1.44)TR-DTD cued trials 70.17 (5.19) 57.22 (7.52) 67.08 (4.31)TR-DTD uncued trials⁎ 119.43 (9.05) 146.91 (13.12) 145.00 (7.51)

⁎Significant differences between patients and control (pb0.05).

203J.L. Rubio et al. / Schizophrenia Research 133 (2011) 199–204

no significant association regarding these parameters and PANSSscores. On the other hand, there were positive significant associationsbetween control executive network and the emotional distress factorof the PANSS (r=.523; p=0.015) and between control executivenetwork and PR1 for cued trials (r=.437; p=0.048). This resultmeans that the decrease in cognitive control (an increased responsetime) is related with the belief intensity at the beginning of thedecision-making process. Finally, there were significant negativeassociation between the Hinting Task with regard to PANSS negativescore (r=.419; p=0.030), with PANSS negative factor (r=.0.426;p=0.027) and with DTD for cued trials (r=−.389; p=0.048), asshown in Table 3.

3. Discussion

The results of the current study suggest that, relative to siblingsand controls, patients display a general tendency to jump toconclusions, characterised by making prompt decisions with a highlevel of conviction, and involving a lowered threshold underambiguous circumstances (uncued trials), in which a greater amountof information is required to make a decision. Moreover, the inverserelationship between the plausibility rating at the first stage anddraws to decision may be considered as evidence of a reasoning bias,involving a failure to integrate new evidence when strong initialbeliefs exist. An interpretation of this result could be that the secondhypothesis proposed (a lowered threshold for making decisions) isinfluenced by the first (an overestimate in the conviction in theirchoices at the beginning of the decision-making process). Theseresults cannot be explained by impulsivity, as there was no significantdifference between groups in the reaction time analysis. On the otherhand, the absence of differences between the healthy controls and thesiblings that we observed would be inconsistent with previous results(van Dael et al., 2006), which suggest that there is a dose–responserelationship between JTC and psychosis liability that is not present inour study. This matter remains unanswered in view of the limitationsof our study: our sample size was small and a different experimentaltask was used. Another limitation of the study is that subjects werenot drug tested. In order to minimize this limitation, we used clinicalevidence to determine if the subject showed clinical symptoms ofacute intoxication. In addition, all patients were evaluated ondischarge day, and the use of drugs was not allowed in the in-patientunit.

In view of the hypersalience hypothesis (Kapur, 2003), psychosispatients show dysregulated dopamine transmission. This anomalymay, depending on context, give rise to an inappropriate salienceattribution. Along this line, our research shows how an ambiguouscontext, such as the uncued task, could strengthen the salienceattribution. Again, overestimated conviction behind choices at thebeginning of the decision process would affect the subsequentchoices. These effects could be explained by the neurotransmissionof dopamine in the ventral striatum pathway, which reinforcesstimulus-response, and has the capacity to reinforce patterns ofcerebral activity in relation with a particular mental event (Speechleyet al., 2010). Accordingly, we speculate that our results could be

Table 3Correlational analyses between the main parameters in the drawing to decision task in the

JTC-1 cued JTC-1 uncued PR-1 c

JTC-1 cued 1 .382(*) .757(*JTC-1 uncued .382(*) 1 .369(*PR-1 cued .757(**) .369(*) 1PR-1 uncued .245 .633(**) .577(*DTD cued −.616(**) −.385(*) −.537DTD uncued −.544(**) −.694(**) −.559

⁎Significant differences between patients and control (pb0.05).⁎⁎Significant differences between patients and control (pb0.005).

explained under the hypothesis that dysregulated dopamine trans-mission in the ventral striatum dopamine pathway would be moresevere in ambiguous contexts, such as the uncued task (Morrison andMurray, 2009).

Our correlation analyses showed a pattern of associations betweenjumping to conclusions and theory of mind on the one hand, andbetween JTC and executive function on the other hand, very similar tovery similar to the study of Woodward et al. (2009) cited in theintroduction. Altogether, these results suggest that JTC bias is relatedto a lack of cognitive flexibility and a higher sensitivity to stress. Theseconcepts represent the two sides of a same coin and could be linked tothe notion of faulty appraisal (Broome et al., 2007).

Mata et al. (2008) showed that cannabis use in schizophrenicpatients was associated with decision-making impairment. In order tocontrol for this potential confounding variable, we adjusted by heavycannabis use and the result was not modified.

Similarly, there is evidence that theory of mind could explain theimpairment in JTC (Woodward et al., 2009). We adjusted the JTCresults by The Degraded Facial Affect Recognition, a task thatmeasures some dimensions of theory of mind. Yet after adjustment,the difference in JTC remained.

Further studies will be necessary to undertake a comprehensiveassessment of different aspects in order to determine what mecha-nisms underlie JTC bias. Although these mechanisms are unclear todate, previous research points to the role of faulty appraisal (Broomeet al., 2007) and to an inability to tolerate ambiguity (McKay et al.,2006). In this sense, our results may be interpreted in terms of faultyappraisal characterised by a high level of conviction at the beginningof the decision process and by an inability to tolerate the presence ofany ambiguity when there is a need to collect information in order toconfigure interpretations for decision-making. In short, our results canbe taken as evidence for a disorganised inference process inschizophrenia, in which the first mechanism responsible is a faultyappraisal. This suggests that the origin of the JTC bias is related, at leastpartly, to emotional factors. To this respect, there is some evidencethat JTC bias is significantly more pronounced under stress conditionsand that this tendency is not present in healthy subjects (Moritz et al.,2009). Perhaps this higher sensitivity to stress in schizophreniapatients explains why, according to our results, JTC bias wassignificantly more pronounced in ambiguous contexts. Some otherstudies have found elevated hippocampal and amygdaloid activityduring emotional intensity judgements (Rădulescu and Mujica-Parodi, 2008). The elevated activity of these regions could well berelated to faulty appraisal.

Our findings on the type of reasoning present in schizophreniamay prove useful in developing specific cognitive-behavioural in-terventions, in which patients would be instructed to create their owninterpretations of reality, and to better determine what kind ofsituation requires the integration of new evidence when strong initialbeliefs exist. Exposure to social situations using virtual reality has thepotential to be incorporated into cognitive behavioural interventionsfor paranoia (Fornells-Ambrojo et al., 2008; Ross et al., 2010). A bettercomprehension of JTC bias would no doubt be helpful in thedevelopment of cognitive-behavioural interventions in psychosis. At

schizophrenia group.

ued PR-1 uncued DTD cued DTD uncued

*) .245 −.616(**) −.544(**)) .633(**) −.385(*) −.694(**)

.577(**) −.537(**) −.559(**)*) 1 −.234 −.521(**)(**) −.234 1 .714(**)(**) −.521(**) .714(**) 1

204 J.L. Rubio et al. / Schizophrenia Research 133 (2011) 199–204

present, the metacognitive training program provides correctiveexperiences for this type of reasoning bias and others, and there isevidence to support the feasibility and efficacy of this approach(Moritz et al., 2011).

Supplementarymaterials related to this article can be found onlineat doi:10.1016/j.schres.2011.08.008.

Role of Funding SourceThis study was supported by grant GI8374199 (Ayudas económicas para el

desarrollo de proyectos de investigación sobre drogodependencias 2007, BOE 263, 2November 2007); grant Fundación Alicia Koplowitz, BAE 09/90088; Estancia Formativade la Junta de Andalucia 2010; Investigacion Biomedica y en Ciencias de Salud enAndalucia 2007; and by a grant from the Spanish Ministry of Education and ScienceCognitive Flexibility in Synaesthesia and Cognitive Rehabilitation (ref.:PSI2009-11789)to E.G. Milán, grant No. 379/05 of the Health Council of the Regional Government ofAndalusia.

They had no further role in study design; in the collection, analysis andinterpretation of data; in the writing of the report; or in the decision to submit thepaper for publication.

ContributorsJOSÉ LUIS RUBIO, participated in the design of the study, the collection of data, the

interpretation of data and drafting of the article; he has approved the final version ofthe manuscript.

LAUREN HERNÁNDEZ, participated in the design of the study, conducted thestatistical analysis and participated in the interpretation of the data and in drafting thearticle; he has approved the final version of the manuscript.

MARIA LUISA BARRIGÓN, participated in the design of the study, the collection ofdata and the interpretation of the data; she has approved the final version of themanuscript.

MAITE FERRÍN, participated in the design of the study, the collection of data and theinterpretation of the data; she has approved the final version of the manuscript.

MARÍA DOLORES SALCEDO, participated in the design of the study, the collection ofdata and their interpretation; she has approved the final version of the manuscript.

JOSEFA MARÍA MORENO, participated in the design of the study and theinterpretation of data; she has approved the final version of the manuscript.

EMILIO GÓMEZ, participated in the design of the study and the interpretation of thedata; he has approved the final version of the manuscript.

STEFFEN MORITZ, participated in the design of the study and the interpretation ofdata; he has approved the final version of the manuscript.

MIGUEL RUIZ-VEGUILLA participated in the design of the study and theinterpretation of data; he has approved the final version of the manuscript.

Conflict of interestThe authors declare that they have no conflicts of interest regarding this study.

AcknowledgementsThis study was supported by grant GI8374199 (Ayudas económicas para el

desarrollo de proyectos de investigación sobre drogodependencias 2007, BOE 263, 2November 2007),grant Fundación Alicia Koplowitz, BAE 09/90088, Estancia Formativade la Junta de Andalucia 2010, Investigacion Biomedica y en Ciencias de Salud enAndalucia 2007, and a grant from the Spanish Ministry of Education and ScienceCognitive Flexibility in Synaesthesia and Cognitive Rehabilitation (ref.: PSI2009-11789)to E. G. Milán, and by grant no. 379/05 of the Health Council of the RegionalGovernment of Andalusia.

They had no further role in study design; in the collection, analysis andinterpretation of data; in the writing of the report; or in the decision to submit thepaper for publication.

References

American Psychiatric Association, 2000. Diagnostic and Statistical Manual of MentalDisorders, fourth ed. rev. American Psychiatric Association, Washington D.C.

Averbeck, B.B., Evans, S., Chouhan, V., Bristow, E., Shergill, S.S., 2011. Probabilisticlearning and inference in schizophrenia. Schizophr. Res. 127 (1–3), 115–122.

Broome, M.R., Johns, L.C., Valli, I., Woolley, J.B., Tabraham, P., Brett, C., Valmaggia, L.,Peters, E., Garety, P.A., McGuire, P.K., 2007. Delusion formation and reasoningbiases in those at clinical high risk for psychosis. Br. J. Psychiatry 191 (Suppl. 51),38–42.

Corcoran, R., Mercer, G., Frith, C.D., 1995. Schizophrenia, symptomatology and socialinference: investigating “theory of mind” in people with schizophrenia. Schizophr.Res. 17 (1), 5–13.

First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 1997. Structured Clinical Interviewfor DSMIV Axis I Disorders (Clinician version), first edition. American PsychiatricAssociation, Washington D.C.

Fornells-Ambrojo, M., Barker, C., Swapp, D., Slater, M., Antley, A., Freeman, D., 2008.Virtual reality and persecutory delusions: safety and feasibility. Schizophr. Res. 104(1–3), 228–236.

Freeman, D., Pugh, K., Garety, P., 2008. Jumping to conclusions and paranoid ideation inthe general population. Schizophr. Res. 102 (1–3), 254–260.

Frigerio, E., Burt, D.M., Montagne, B., Murray, L.K., Perrett, D.I., 2002. Facial affectperception in alcoholics. Psychiatry Res. 113 (1–2), 161–171.

Huq, S.F., Garety, P.A., Hemsley, D.R., 1988. Probabilistic judgements in deluded andnon-deluded subjects. Q. J. Exp. Psychol. 40 (4), 801–812.

Kapur, S., 2003. Psychosis as a state of aberrant salience: a framework linking biology,phenomenology, and pharmacology in schizophrenia. Am. J. Psychiatry 160 (1),13–23.

Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale(PANSS) for schizophrenia. Schizophr. Bull. 23 (1), 99–110.

Lincoln, T.M., Ziegler, M., Mehl, S., Rief, W., 2010. The jumping to conclusions bias indelusions: specificity and changeability. J. Abnorm. Psychol. 119 (1), 40–49.

Mata, I., Rodríguez-Sánchez, J.M., Pelayo-Terán, J.M., Pérez-Iglesias, R., González-Blanch, C., Ramírez-Bonilla, M., Martínez-García, O., Vázquez-Barquero, J.L., Crespo-Facorro, B., 2008. Cannabis abuse is associated with decision-making impairmentamong first-episode patients with schizophrenia-spectrum psychosis. Psychol.Med. 38 (9), 1257–1266.

McKay, R., Langdon, R., Coltheart, M., 2006. Need for closure, jumping to conclusions,and decisiveness in delusion-prone individuals. J. Nerv. Ment. Dis. 194 (6),422–426.

Moritz, S., Woodward, T.S., 2005. Jumping to conclusions in delusional and non-delusional schizophrenic patients. Br. J. Clin. Psychol. 44 (2), 193–207.

Moritz, S.,Woodward, T.S., 2006. A generalized bias against disconfirmatory evidence inschizophrenia. Psychiatry Res. 142 (2–3), 157–165.

Moritz, S., Woodward, T.S., Lambert, M., 2007. Under what circumstances do patientswith schizophrenia jump to conclusions? A liberal acceptance account. Br. J. Clin.Psychol. 46 (2), 127–137.

Moritz, S., Veckenstedt, R., Randjbar, S., Hottenrott, B., Woodward, T.S., von Eckstaedt,F.V., Schmidt, C., Jelinek, L., Lincoln, T.M., 2009. Decision making under uncertaintyand mood induction: further evidence for liberal acceptance in schizophrenia.Psychol. Med. 39 (11), 1821–1829.

Moritz, S., Kerstan, A., Veckenstedt, R., Randjbar, S., Vitzthum, F., Schmidt, C., Heise, M.,Woodward, T.S., 2011. Further evidence for the efficacy of a metacognitive grouptraining in schizophrenia. Behav. Res. Ther. 49 (3), 151–157.

Morrison, P.D., Murray, R.M., 2009. From real-world events to psychosis: the emergingneuropharmacology of delusions. Schizophr. Bull. 35 (49), 668–674.

Peralta, V., Cuesta, M.J., 1994. Psychometric properties of the Positive and NegativeSyndrome Scale (PANSS) in schizophrenia. Psychiatry Res. 53 (1), 31–40.

Posner, M.I., Fan, J., 2005. Attention as an organ system. In: Pormerantz, J. (Ed.),Neurobiology of Perception and Communication: From Synapse to Society. The IVthde Lange Conference. Cambridge University Press, Cambridge, U.K.

Posner, M.I., Petersen, S.E., 1990. The attention system of the human brain. Annu. Rev.Neurosci. 13, 25–42.

Rădulescu, A.R., Mujica-Parodi, L.R., 2008. A systems approach to prefrontal-limbicdysregulation in schizophrenia. Neuropsychobiology 57 (4), 206–216.

Robins, L.N., Wing, J., Wittchen, H.U., Helzer, J.E., Babor, T.F., Burke, J., Farmer, A.,Jablenski, A., Pickens, R., Regier, D.A., Sartorius, N., Towle, M.S., 1988. The CompositeInternational Diagnostic Interview. An epidemiologic instrument suitable for use inconjunction with different diagnostic systems and in different cultures. Arch. Gen.Psychiatry 45 (12), 1069–1077.

Ross, K., Freeman, D., Dunn, G., Garety, P., 2010. A randomized experimentalinvestigation of reasoning training for people with delusions. Schizophr. Bull. 37(2), 324–333.

Ruiz-Veguilla, M., Gurpegui, M., Barrigón, M.L., Ferrín, M., Marín, E., Rubio, J.L.,Gutiérrez, B., Cervilla, J., 2009. Fewer neurological soft signs among first episodepsychosis patients with heavy cannabis use. Schizophr. Res. 107 (2–3), 158–164.

Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J., Weiller, E., Hergueta,T., Baker, R., Dunbar, G.C., 1998. The Mini-International Neuropsychiatric Interview(M.I.N.I.): the development and validation of a structured diagnostic psychiatricinterview for DSM-IV and ICD-10. J. Clin. Psychiatry 59 (22–33), 34–57.

Speechley, W.J., Whitman, J.C., Woodward, T.S., 2010. The contribution of hypersalienceto the “jumping to conclusions” bias associated with delusions in schizophrenia. J.Psychiatry Neurosci. 35 (1), 7–17.

Van 'tWout,M., Aleman, A., Kessels, R.P.C., Laroi, F., Kahn, R.S., 2004. Emotional processingin a non-clinical psychosis-prone sample. Schizophr. Res. 68 (2–3), 271–281.

Van Dael, F., Versmissen, D., Janssen, I., Myin-Germeys, I., van Os, J., Krabbendam, L.,2006. Data gathering: biased in psychosis? Schizophr. Bull. 32 (2), 341–351.

Van der Gaag, M., Hoffman, T., Remijsen, M., Hijman, R., de Haan, L., van Meijel, B., vanHarten, P.N., Valmaggia, L., de Hert, M., Cuijpers, A., Wiersma, D., 2006. The five-factor model of the Positive and Negative Syndrome Scale II: a ten-fold cross-validation of a revised model. Schizophr. Res. 85 (1–3), 280–287.

Veckenstedt, R., Randjbar, S., Vitzthum, F., Hottenrott, B., Woodward, T.S., Moritz, S.,2011. Incorrigibility, jumping to conclusions, and decision threshold in schizo-phrenia. Cogn. Neuropsychiatry 16 (2), 174–192.

Woodward, T.S., Mizrahi, R., Menon, M., Christensen, B.K., 2009. Correspondencesbetween theory of mind, jumping to conclusions, neuropsychological measuresand the symptoms of schizophrenia. Psychiatry Res. 170 (2–3), 119–123.