Introduction
Although auditory verbal hallucinations (AVH) are a characteristic symptom of psychotic disorders, they are also found in healthy individuals in the general population, with a prevalence of approximately 10–15% (Tien, Reference Tien1991; Van Os et al. Reference Van Os, Hanssen, Bijl and Ravelli2000; Verdoux & van Os, Reference Verdoux and van Os2002; Johns et al. Reference Johns, Cannon, Singleton, Murray, Farrell, Brugha, Bebbington, Jenkins and Meltzer2004; for a review, see Beavan et al. Reference Beavan, Read and Cartwright2011). This includes individuals who report hearing voices quite regularly as well as those who report hearing a voice once in their lifetime. Therefore, AVH have been proposed to form a continuum, ranging from rare occurrences in healthy individuals at one end, through individuals high on ‘schizotypal’ traits, to psychotic patients with frequent occurrence at the other end. Since Romme & Escher's (Reference Romme and Escher1993) seminal work, which first challenged the view of voices necessarily being characteristic of psychiatric illness, a growing body of work has been devoted to the study of hallucinations across this continuum (Larøi et al. Reference Larøi, Sommer, Blom, Fernyhough, Ffytche, Hughdahl, Johns, McCarthy-Jones, Preti, Raballo, Slotema, Stephane and Waters2012). While AVH in these groups show considerable overlap in characteristics such as loudness, number of voices, perceived location of voices and personification (Daalman et al. Reference Daalman, Boks, Diederen, de Weijer, Blom, Kahn and Sommer2011), as well as in brain activity (Diederen et al. Reference Diederen, Daalman, de Weijer, Neggers, van Gastel, Blom, Kahn and Sommer2011), it remains unclear why some voice-hearers remain psychologically healthy (and even perceive their lives to be enriched by their experiences) while others suffer considerable distress and make a transition to psychosis. While it is becoming increasingly accepted that there is a causal link between trauma and psychosis, and hearing voices in particular (Bebbington et al. Reference Bebbington, Jonas, Kuipers, King, Cooper, Brugha, Meltzer, McManus and Jenkins2011; Varese et al. Reference Varese, Smeets, Drukker, Lieverse, Lataster, Viechtbauer, Read, van Os and Bentall2012), both clinical and non-clinical samples with AVH report equally high rates of trauma (Andrew et al. Reference Andrew, Gray and Snowden2008; Lovatt et al. Reference Lovatt, Mason, Brett and Peters2010), and childhood trauma was not found to be predictive of need for care status or emotional valence of the voices content in a large study comparing healthy and distressed voice-hearers (Daalman et al. Reference Daalman, Diederen, Derks, van Lutterveld, Kahn and Sommer2012).
Cognitive models of psychosis suggest that it is not solely the presence of anomalous experiences, such as AVH, which lead to full-blown psychotic symptoms, but rather the appraisals that individuals hold about these experiences (Chadwick & Birchwood, Reference Chadwick and Birchwood1994; Birchwood & Chadwick, Reference Birchwood and Chadwick1997; Morrison, Reference Morrison2001; Garety et al. Reference Garety, Kuipers, Fowler, Freeman and Bebbington2001, Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007; Barkus et al. Reference Barkus, Stirling, French, Morrison, Bentall and Lewis2010). For example, an AVH might be considered as a phenomenon originating from one's own brain, as a message coming from the benevolent spirit of a deceased grandparent, or it might be perceived as an evil force from another dimension. These appraisals are likely to lead to different emotional consequences, with the more malign interpretations being associated with fear and distress, which in turn renders a person more vulnerable to developing psychosis and a ‘need for care’ (Van Os et al. Reference Van Os, Linscott, Myin-Germeys, Delespaul and Krabbendam2009). There is emerging evidence to support this view; for instance, Escher et al. (Reference Escher, Romme, Buiks, Delespaul and van Os2002) found that persistence of voices in a sample of children over a 3-year period was predicted by negative voice appraisals and associated anxiety and depression.
Cognitive biases have been proposed to be instrumental in shaping these appraisals (Garety et al. Reference Garety, Freeman, Jolley, Dunn, Bebbington, Fowler, Kuipers and Dudley2005; Freeman et al. Reference Freeman, Dunn, Fowler, Bebbington, Kuipers, Emsley, Jolley and Garety2012). A cognitive bias is a way in which an individual habitually interprets his or her experiences, gathers information about the world, and develops and maintains beliefs. Typical biases commonly observed in people with emotional disorders include ‘jumping to conclusions’, emotionally based reasoning, and dichotomous thinking (the so-called ‘Beckian’ biases; Beck, Reference Beck1979), which are also present in people with psychosis (Peters et al. Reference Peters, Moritz, Wiseman, Greenwood, Kuipers, Schwannauer, Donaldson, Klinge, Ross, Ison, Williams, Scott, Beck and Garety2013). For example, the typical bias of ‘dichotomous thinking’ leads to an absolute or black-and-white view of things: when a small mistake is made, someone might judge himself or herself as totally useless and worthless. There is now a large body of work demonstrating that cognitive biases, especially jumping to conclusions, play a key role in the formation and maintenance of delusions (for a review, see So et al. Reference So, Garety, Peters and Kapur2010), but few studies have investigated their role in AVH, or, more precisely, the beliefs people hold about their voices. It remains unclear whether cognitive biases differ between individuals experiencing AVH with and without a ‘need for care’. Ascertaining the extent to which cognitive biases are present in healthy voice-hearers may further clarify the relationship between biases, appraisals about voices and the transition to psychosis, as proposed by cognitive models of positive symptoms. Potentially the absence of cognitive biases in the healthy group may prevent the formation of malign appraisals, in turn reducing the chances that hearing voices becomes problematic for the individual and leads to a ‘need for care’.
The Cognitive Biases Questionnaire for psychosis (CBQp) was recently developed to assess cognitive biases in patients with a psychotic disorder (Peters et al. Reference Peters, Moritz, Wiseman, Greenwood, Kuipers, Schwannauer, Donaldson, Klinge, Ross, Ison, Williams, Scott, Beck and Garety2013). In all, five types of cognitive biases, all believed to be important in psychosis, were incorporated in the questionnaire: jumping to conclusions, intentionalizing, catastrophizing, emotional reasoning and dichotomous thinking. The aim of the current study was to compare the presence of cognitive biases with the CBQp in three groups: patients with AVH who had been diagnosed with a psychotic disorder, healthy voice-hearers, and healthy controls. Based on the cognitive model of psychosis, we hypothesized that the healthy voice-hearers would not show the cognitive biases found in the clinical group.
Method
Participants
A total of 72 patients with AVH and a diagnosis of a psychotic disorder, 72 healthy voice-hearers and 72 healthy controls without AVH were included. The healthy voice-hearers did not meet criteria for a DSM-IV diagnosis, as defined by a psychiatrist using the Comprehensive Assessment of Symptoms and History (CASH) interview (Andreasen et al. Reference Andreasen, Flaum and Arndt1992) and the Structured Clinical Interview for Personality Disorder (SCID-II; First et al. Reference First, Spitzer, Gibbon and Williams1995). Depressive disorder in complete remission was not an exclusion criterion.
An exclusion criterion for all groups was alcohol and drug abuse. The healthy controls and voice-hearers were screened for alcohol abuse (more than 20 units per week) and drug abuse (using cannabis more than once a month and/or the use of other illicit substances) by telephone and later with the help of urine samples. In the patient group, alcohol and drug abuse was screened for by an independent psychiatrist with the help of the CASH interview.
For the healthy voice-hearers, the minimum frequency to experience AVH for inclusion in the study was once every 3 months and the minimum duration since onset of AVH was 1 year.
Both the controls and healthy voice-hearers were recruited with the help of a Dutch website called ‘explore your mind’ (www.verkenuwgeest.nl) and selected on the basis of low and high scores, respectively, in the items of the Launay and Slade Hallucinations scale (Larøi et al. Reference Larøi, Marczewski and Van der Linden2004) tapping into AVH. For more details about the selection and assessment procedure, see Sommer et al. (Reference Sommer, Daalman, Rietkerk, Diederen, Bakker, Wijkstra and Boks2010) and Daalman et al. (Reference Daalman, Boks, Diederen, de Weijer, Blom, Kahn and Sommer2011).
The patients with a psychotic disorder were all out-patients from the Voices Clinic of the University Medical Center Utrecht. These patients visited our clinic for regular treatment for psychosis or as a second opinion for intractable psychosis. In this group, clinical diagnoses were confirmed by an independent psychiatrist using the CASH interview. A total of 42 patients (58.3%) were diagnosed with paranoid schizophrenia, 18 (25%) with psychosis not otherwise specified, 10 (13.9%) with schizo-affective disorder and two (2.8%) with disorganized schizophrenia. Demographic and clinical details are provided in Table 1.
Table 1. Demographic and clinical characteristics of participants: clinical and healthy voice-hearers and healthy controls
df, Degrees of freedom; s.d., standard deviation.
a Number of cases with missing data: three in the group of clinical voice-hearers.
The study was approved by the Human Ethics Committee of the University Medical Center Utrecht. After a complete description of the study was provided to the participants, written informed consent was obtained.
Measurements
The CBQp was developed by Peters et al. (Reference Peters, Moritz, Wiseman, Greenwood, Kuipers, Schwannauer, Donaldson, Klinge, Ross, Ison, Williams, Scott, Beck and Garetyin press) to assess cognitive biases relevant to patients with psychosis. It consists of 30 vignettes grouped under two themes: ‘anomalous perception’ (e.g. ‘Imagine that you are walking down the street when you hear your name being called, but when you look around you don't see anybody’) and ‘threatening events’ (e.g. ‘Imagine you receive a letter and you notice it is not sealed’). In all, five types of cognitive biases are assessed: ‘intentionalizing’, ‘catastrophizing’, ‘dichotomous thinking’, ‘jumping to conclusions’ and ‘emotional reasoning’. There are three vignettes per bias for each theme (six vignettes per bias in total). Each vignette is rated on a three-point scale ranging from 1 to 3 (1 = absence of bias; 2 = presence of bias with some qualification; and 3 = presence of bias). The maximum total score for each bias is 18, and for each theme 45, with a total overall score of 90. The minimum total overall score is 30. The scale has good psychometric properties; both internal consistency and test–retest reliability are high (Peters et al. Reference Peters, Moritz, Wiseman, Greenwood, Kuipers, Schwannauer, Donaldson, Klinge, Ross, Ison, Williams, Scott, Beck and Garety2013). In addition to the total score, all subscales as well as both themes were compared between the three groups. A Dutch translation by T. Bastiaens et al. (unpublished) was used for this study.
The Psychotic Symptom Rating Scales (PSYRATS) auditory hallucination rating subscale (AHRS; Haddock et al. Reference Haddock, McCarron, Tarrier and Faragher1999) was used to map the phenomenological characteristics of the AVH. This questionnaire describes 11 characteristics of AVH. Each item is evaluated on a five-point Likert scale ranging from 0 to 4. For the use of this questionnaire in healthy voice-hearers, the range of the frequency scale was extended to 0–6 (also covering options ‘at least once every month’ and ‘at least once every three months’ since AVH are experienced less often than once per week, the original minimum score of this item). This questionnaire was administered by trained psychologists.
The items of the PSYRATS AHRS can be extrapolated into three dimensional subscales (Haddock et al. Reference Haddock, McCarron, Tarrier and Faragher1999; Morrison et al. Reference Morrison, Nothard, Bowe and Wells2004): (1) an emotional characteristics factor (i.e. amount and intensity of distress, amount and degree of negative content items); (2) a physical characteristics factor (i.e. descriptions of the voice: items frequency, duration, location and loudness); and (3) a cognitive interpretation factor (i.e. beliefs regarding the origin and attributions of control: items origin, disruption and control).
Statistics
The primary outcome measure was the total score on the CBQp. Between-group comparison (i.e. psychotic patients with AVH, healthy voice-hearers, and controls) of this measure was achieved through univariate analysis of covariance (ANCOVA), applying a general linear model procedure. In addition, the five cognitive biases subscales of the CBQp (intentionalizing, catastrophizing, dichotomous thinking, jumping to conclusions and emotional reasoning) as well as both themes (anomalous perceptions and threatening events) were analysed in multivariate ANCOVA. Age was entered as a covariate since this variable differed significantly between the three groups and showed small but significant relationships to both intentionalizing (r −0.195, p = 0.004) and catastrophizing (r −0.234, p = 0.001). Gender, total years of education, past history of depression, being married, being divorced and ethnicity were not associated with the CBQp total score or the subscales.
The relationship between cognitive biases and AVH characteristics was investigated with a hierarchical regression analysis. The total CBQp score was the dependent variable. The three PSYRATS factors (cognitive, emotional, and physical) and group membership (i.e. patients versus healthy voice-hearers) were entered stepwise as predictors.
All data were analysed with SPSS (IBM, USA).
Missing values
In the complete sample of 216 participants, four items of the CBQp were missing, and a multiple imputation procedure based on linear regression was used to estimate these values based on the other observed variables.
Results
Description of AVH characteristics
Table 2 illustrates the characteristics of the voices in both AVH groups and the total scores on the three subscales of the PSYRATS. Mean scores are given as well as the description of its closest anchor.
Table 2. Characteristics of auditory verbal hallucinations in healthy and clinical groups
s.d., Standard deviation.
Differences in cognitive biases between the three groups
Total score of the CBQp
The total score of the CBQp differed significantly between the groups (F 2,213 = 37.51, p < 0.001). Pairwise comparisons (mean difference significant at p < 0.05, Bonferroni adjusted) showed that the healthy controls scored significantly lower than both the healthy voice-hearers (p < 0.01) and the clinical group (p < 0.001). Both AVH groups also differed significantly from each other (p < 0.001), with lower scores in the healthy voice-hearers.
The mean total and theme scores on the CBQp in the three groups are illustrated in Fig. 1.
Fig. 1. Total scores on the Cognitive Biases Questionnaire for psychosis (CBQp) (potential range of scores: 30–90), and on the threatening events and anomalous perception themes (potential range of scores: 15–45) in the three groups. Values are means, with their standard errors represented by vertical bars. AVH, Auditory verbal hallucinations.
Individual cognitive biases
There was a statistically significant difference between the three groups on the combined dependent variables (F 10,418 = 11.94, p < 0.001; Pillai's trace 0.44), after correcting for age. When the results for the dependent variables were considered separately the three groups differed significantly on all subscales of the CBQp after correction for multiple testing (0.05/5 = 0.01): intentionalizing (F 2,213 = 14.32, p < 0.001), catastrophizing (F 2,213 = 28.06, p < 0.001), dichotomous thinking (F 2,213 = 20.74, p < 0.001), jumping to conclusions (F 2,213 = 36.26, p < 0.001) and emotional reasoning (F 2,213 = 21.89, p < 0.001).
Further analyses (again, mean difference significant at p < 0.05, Bonferroni adjusted) revealed that healthy controls and healthy voice-hearers scored significantly lower than patients with AVH on four out of the five subscales: intentionalizing (p < 0.001 and p < 0.001, respectively), catastrophizing (p < 0.001 and p < 0.001), dichotomous thinking (p < 0.001 and p < 0.001) and jumping to conclusions (p < 0.001 and p < 0.001), but did not differ significantly from each other. However, on emotional reasoning the healthy controls scored significantly lower than both AVH groups (p < 0.001 and p < 0.001 respectively), which did not differ from each other. The mean scores of the cognitive biases in the three groups are presented in Fig. 2.
Fig. 2. Cognitive Biases Questionnaire for psychosis (CBQp) scores of the three groups on the five cognitive bias subscales (potential range of scores: 6–18). Values are means, with their standard errors represented by vertical bars. AVH, Auditory verbal hallucinations.
Themes of the CBQp
There was a statistically significant difference between the groups on the combined dependent variables (F 4,424 = 26.07, p < 0.001; Pillai's trace 0.40), after correcting for age. When the results for the dependent variables were considered separately the three groups differed significantly on both themes after correction for multiple testing. The total score on threatening events showed a significant main effect for group, after correcting for age (F 2,213 = 41.80, p < 0.001). Pairwise comparisons (mean differences significant at p < 0.05, Bonferroni adjusted) showed that the patients with a psychotic disorder and AVH scored significantly higher than both the healthy controls (p < 0.001) and healthy voice-hearers (p < 0.001). No difference was observed between healthy voice-hearers and healthy controls (p = 1.000).
The total score on anomalous perception also showed a significant main effect for group, after correcting for age (F 2,213 = 24.98, p < 0.001). Pairwise comparisons (mean differences significant at p < 0.05, Bonferroni adjusted) showed that the healthy controls scored significantly lower than both the healthy voice-hearers (p < 0.001) and the clinical group (p < 0.001). Unlike the threatening event theme, no difference was observed between the two AVH groups (p = 0.075).
Mean scores on the threatening events and anomalous perception themes in the three groups are presented in Fig. 1.
Relationships between AVH characteristics and the presence of cognitive biases
A regression analysis showed that both the emotional (i.e. high and intense distress, and negative emotional valence of the voices) and the cognitive interpretation (i.e. belief in the external origin of voices, having little control over the voices and high disruption to life) factors were significant predictors of the presence of cognitive biases. The physical factor was not a significant predictor of CBQp total scores. The results are presented in Table 3.
Table 3. Multiple regression model: predicting the presence of cognitive biases (CBQ total)
CBQ, Cognitive Biases Questionnaire; s.e., standard error; CI, confidence interval.
Discussion
This study investigated the differences in the presence of cognitive biases in voice-hearers with and without a ‘need for care’ and healthy controls. Healthy voice-hearers obtained intermediate total scores between controls and the clinical group on the CBQp (Peters et al. Reference Peters, Moritz, Wiseman, Greenwood, Kuipers, Schwannauer, Donaldson, Klinge, Ross, Ison, Williams, Scott, Beck and Garety2013), measuring several cognitive biases prevalent in psychosis. Although these results suggest that the healthy voice-hearers lie mid-way on the psychosis continuum, further analyses of the CBQp subscales showed that the healthy voice-hearers had the same profile as the healthy controls on four out of the five cognitive biases, differing significantly from the controls on only one subscale that accounted for most of the difference between these two groups on the total CBQp score. These results are consistent with Lawrence & Peters (Reference Lawrence and Peters2004), who found that reasoning biases were limited to people who reported a belief in, rather than experience of, paranormal phenomena.
Emotional reasoning was the only bias where scores were comparable in both AVH groups. This cognitive style, i.e. reasoning based on emotions, feeling or instinct instead of ‘logic’, therefore is the only bias that appears to be related to the presence of, or vulnerability to experience, AVH, rather than to a ‘need for care’. An emotional reasoning bias is perhaps the least psychosis specific, and is not only highly prevalent in other psychiatric disorders, such as anxiety disorders (Clark, Reference Clark1999), but is also arguably the most culturally accepted of the five biases assessed by the CBQp. The remaining biases, namely intentionalizing, catastrophizing, dichotomous thinking and jumping to conclusions, were all significantly higher in the clinical group than in the healthy voice-hearers, suggesting that they may potentially be instrumental in developing unhelpful and distressing appraisals of their AVH, as proposed by cognitive models of psychosis (Chadwick & Birchwood, Reference Chadwick and Birchwood1994; Morrison, Reference Morrison2001; Garety et al. Reference Garety, Kuipers, Fowler, Freeman and Bebbington2001, Reference Garety, Bebbington, Fowler, Freeman and Kuipers2007).
The findings comparing the groups on the two themes of the case vignettes of the CBQp (‘anomalous perceptions’ and ‘threatening events’) were intriguing, and suggest that cognitive biases, similarly to broad reasoning style, are domain specific (Evans et al. Reference Evans, Newstead and Byrne1993; see also Lawrence & Peters, Reference Lawrence and Peters2004). Both healthy and clinical voice-hearers scored higher than healthy controls on the ‘anomalous perceptions’ theme: biases were therefore more pronounced in both the AVH groups when they were presented with information related to unusual perceptual experiences. In contrast, healthy voice-hearers and healthy controls scored lower than the clinical group on the ‘threatening events’ theme: healthy voice-hearers had comparable scores to controls when presented with potential threatening scenarios. These results are in line with previous findings that paranoid (but not necessarily external) appraisals differentiate individuals with psychotic experiences with and without a ‘need for care’ (Brett et al. Reference Brett, Peters, Johns, Tabraham, Valmaggia and McGuire2007; Lovatt et al. Reference Lovatt, Mason, Brett and Peters2010), and that healthy voice-hearers do not show delusional symptoms (Sommer et al. Reference Sommer, Daalman, Rietkerk, Diederen, Bakker, Wijkstra and Boks2010). Indeed, healthy voice-hearers tend to hold appraisals about their voices that are non-threatening, such as benign spiritual explanations (Cottam et al. Reference Cottam, Paul, Doughty, Carpenter, Al-Mousawi, Karvounis and Done2011; Daalman et al. Reference Daalman, Boks, Diederen, de Weijer, Blom, Kahn and Sommer2011), potentially protecting them from developing a ‘need for care’.
The relationship between cognitive biases and AVH factors
Cognitive and emotional, but not physical, characteristics of AVH were found to be related to CBQp scores. The presence of cognitive biases was associated with higher distress and negative emotional valence of voice content, as well as appraisals of the voices as external in origin, of having little control over the voices, and of a high disruption to life. These relationships also provide further support for cognitive models of psychosis, which emphasize the strong links between emotional processes and thinking biases in shaping maladaptive appraisals of psychotic phenomena (Garety et al. Reference Garety, Kuipers, Fowler, Freeman and Bebbington2001; Freeman & Garety, Reference Freeman and Garety2003; Freeman et al. Reference Freeman, Dunn, Fowler, Bebbington, Kuipers, Emsley, Jolley and Garety2012).
Limitations
The healthy voice-hearers and controls in our study were recruited with the help of a website. This sampling strategy could have led to a selection bias, as suspicious individuals may not have completed the questionnaires on the website or rejected our invitation to visit our research laboratory. In addition, the healthy voice-hearers who participated may represent a specific subgroup within the voice-hearing population, i.e. those who are not distressed or whose functioning is not affected upon by additional (sub)clinical symptoms. However, since we set out to investigate possible protective factors in voice-hearers who are not in need for care, we aimed to specifically include this subgroup. The potential biases inherent in our recruitment strategy is therefore not a serious limitation for the research questions posed in this particular study, although we cannot conclude that our sample is representative of voice-hearers in the general population. Nevertheless, anomalous experiences without distress are twice as common (prevalence of approximately 8%) than those with distress (prevalence of approximately 4%) (van Os et al. Reference Van Os, Linscott, Myin-Germeys, Delespaul and Krabbendam2009).
We have tentatively suggested that these findings support cognitive models of psychosis, which propose that cognitive biases may be causally implicated in the formation of appraisals, which in turn may determine the trajectory to health or ill-health. However, no causal claims can be made from our data, since it is also possible that appraisals are affected by a need-for-care status (including the impact of receiving a psychosis diagnosis) rather than the other way round. The relationship between the specific content of voices and thinking biases was not investigated, although a negative emotional valence of content was associated with the presence of cognitive biases.
Other relevant factors, such as experience of trauma, were not included in this study. For instance, a number of studies have suggested that there is a cognitive route between interpersonal trauma and psychosis, i.e. the link between abuse and psychotic experiences may be mediated by appraisals (Gracie et al. Reference Gracie, Freeman, Green, Garety, Kuipers, Hardy, Ray, Dunn, Bebbington and Fowler2007; Lovatt et al. Reference Lovatt, Mason, Brett and Peters2010). It would be interesting to determine whether cognitive biases, especially in relation to appraisals of threat, are related to traumatic experiences earlier in life, especially abuse and other interpersonal traumas such as discrimination (Janssen et al. Reference Janssen, Hanssen, Bak, Bijl, de Graaf, Vollebergh, McKenzie and van Os2003), and how this may make an impact on the content and appraisals of voices in both healthy and clinical voice-hearers.
Clinical implications
The assessment of cognitive biases in individuals with AVH may shed more light on individuals’ vulnerability to make the transition to full-blown psychosis. Furthermore, it can help identify the tendency to make unhelpful appraisals, which can then be targeted in cognitive behavioural therapy (CBT), in turn alleviating the accompanying distress. New adjunctive interventions to CBT have also recently been developed to target reasoning processes specifically, for instance, metacognitive training (Moritz & Woodward, Reference Moritz and Woodward2007; Moritz et al. Reference Moritz, Verckenstedt, Randjbar, Vitzthum and Woodward2011) or the Maudsley Review Training Programme (Waller et al. Reference Waller, Freeman, Jolley, Dunn and Garety2011), which focuses specifically on ‘jumping to conclusions’ and belief flexibility. The results of this study support this recent trend in focusing explicitly on cognitive and reasoning biases, rather than the anomalous experiences themselves.
In conclusion, most cognitive biases associated with psychosis, particularly with themes of threatening events, were absent in healthy voice-hearers, with the exception of emotional reasoning. Cognitive biases were associated with both emotional and cognitive characteristics of voices. These findings overall are consistent with the cognitive model of psychosis, which proposes a central role for appraisals of psychotic experiences. The absence of cognitive biases may therefore prevent the formation of malign appraisals and delusions in healthy voice-hearers, keeping them on the safe end of the psychosis continuum.
Acknowledgements
This work was supported by the Dutch Scientific Research Organization (Nederlandse Wetenschappelijke Organisatie) (grant nos. 916.56.172 and 017.106.301); the Dutch Scientific Research Organization had no further role in study design; in the collection analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. The authors wish to thank Sanne Verkooijen for her time and effort in helping to collect the data and Elemi Breetvelt for assistance with the multiple imputation procedure.
Declaration of Interest
None.
