SEARCH

SEARCH BY CITATION

Keywords:

  • affect;
  • delusion;
  • hallucination;
  • metacognition

Abstract

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

Commonly, individuals prone to hallucinations and delusions hold dysfunctional metacognitive beliefs and report higher levels of negative affect, yet, these associations have not been clearly investigated in non-clinical samples due to the failure to control for high intercorrelations between variables. The aim of the current study was to investigate how hallucination and delusion proneness are associated with dysfunctional metacognitions and negative affect. A cross-sectional sample of 715 students free from psychiatric diagnoses (Mage = 28.1 years, SD = 10.9, range 18–65) completed the Launay-Slade Hallucination Scale (LSHS-R); Peters et al. Delusion Inventory (PDI-21); Depression, Anxiety, and Stress Scale (DASS-21); and the Metacognition Questionnaire (MCQ-30). Findings that participants who were prone to both hallucinations and delusions reported elevated levels of negative affect support the need for targeted mental health treatment for individuals who experience psychological distress related to their hallucinatory and delusional experiences. While metacognition beliefs of need to control thoughts and cognitive self-consciousness, along with the anxiety and stress DASS-21 subscales appeared as significant cross-sectional predictors of proneness to hallucinations and delusions, only metacognitions demonstrated any notable predictive value for delusion proneness. This finding questions the role of metacognitions in determining hallucination and delusion proneness in non-clinical samples.

It has been suggested that psychotic symptoms, such as hallucinations and delusions, exist on a continuum that ranges from normality to clinical psychosis (David, 2010; van Os, Linscott, Myin-Germeys, Delespaul, & Krabbendam, 2009). This continuum model is supported by findings that hallucinations, simply described as false perceptions and delusions, commonly described as false beliefs, are prevalent in individuals without clinical psychosis (Choong, Hunter, & Woodruff, 2007; Freeman, 2006). Epidemiological studies, investigating a variety of psychotic symptoms, have found prevalence rates ranging from 4% (Tien, 1991) to 39% (Ohayon, 2000) in the general population. It has been hypothesised that hallucinations and delusions may result from dysfunction in the regulation of cognition (Morrison, 2001). In line with this, a number of researchers have proposed metacognition, a term used to describe the process of ‘thinking about thinking’ (Varese, Barkus, & Bentall, 2011), as an important factor in the development and maintenance of psychotic symptoms (García-Montes, Cangas, Pérez-Álvarez, Fidalgo, & Gutiérrez, 2006; Morrison, Wells, & Nothard, 2000).

The Self-Regulatory Executive Function model (S-REF; Wells & Matthews, 1994) states that a cognitive-attentional syndrome, described as a state of heightened self-focused attention, activation of dysfunctional beliefs, attentional bias, and ruminative processes, creates a vulnerability to psychological dysfunction. The S-REF model has been proposed as a framework for understanding a range of psychological dysfunctions, including rumination, obsessive compulsive disorder, depression, general anxiety disorder, and psychosis (Matthews & Wells, 2004; Morrison & Wells, 2003; Wells & Carter, 2001; Wells & Matthews, 1994). In accordance with the S-REF model (Wells & Matthews, 1994), it has been suggested that dysfunctional metacognitive beliefs regarding worry and intrusive thoughts, are implicated in the development and maintenance of hallucinations and delusions in both clinical and non-clinical samples (Barkus et al., 2010; García-Montes et al., 2006; Larøi & Van der Linden, 2005).

Other cognitive theories also suggest that hallucinations and delusions may results from disturbances in the regulation of cognition (e.g., David, 2004; Morrison, 2001; Varese & Bentall, 2011). High co-occurrence between hallucinations and delusions in both clinical and non-clinical samples further indicate that such experiences may share similarities in their cognitive underpinning (Larøi & Woodward, 2007; Lincoln, 2007). A common thread within cognitive theories is the idea that psychotic symptomatology results from misinterpretations of cognitive intrusions and that hallucinatory and delusional experiences may be due to misattributed intrusive thoughts, images, and impulses (Morrison, 2001). The meaning given to such experience may in turn influence the individual's behavioural, affective, and cognitive responses (Morrison, 2001). It should also be noted that Morrison et al. 's model does infer causality in that dysfunctional metacognitive beliefs lead to misattribution of intrusive thoughts to external sources, thus generating hallucinatory experiences. However, according to recent review (Varese & Bentall, 2011), the role of metacognitions in the genesis and maintenance of hallucinations is not consistently supported by empirical results, therefore requiring further research.

Symptoms of psychosis have been associated also with emotional disturbances (Freeman & Garety, 2003). Researchers have linked the onset, development, maintenance, and appraisal of hallucinations and delusions with negative emotional states in both clinical and non-clinical populations (Cella, Cooper, Dymond, & Reed, 2008; Freeman & Garety, 2003; Paulik, Badcock, & Maybery, 2006). In particular, a number of studies have found associations between increased negative affect, such as anxiety, depression, and stress, with the expression of hallucinatory and delusional symptoms in samples of otherwise healthy individuals (Allen et al., 2005; Paulik et al., 2006). However, contrary findings of Cangas, Errasti, Garcia-Montes, Alvarez, and Ruiz (2006), along with those by Morrison et al. (2000), demonstrate how inconsistent previous research has been in reliably demonstrating such relationships, especially for hallucination.

A key limitation of previous research investigating cognitive and affective associations with hallucination and delusion proneness, potentially explaining inconsistent findings, has been the failure to control for high intercorrelations between variables (Varese & Bentall, 2011). Consequently, the aim of the current study was to investigate whether the combination of hallucinatory and delusional proneness is associated with higher levels of negative affect in a non-clinical sample. Specifically, the first aim was to compare levels of negative affect in non-clinical participants who are prone to both hallucinations and delusions, against those who are less prone to such experiences. The second aim was to clarify the cognitive-affective cross-sectional predictors of hallucination and delusion proneness in a non-clinical sample, while controlling for covariation between the assessed variables. It was predicted that in accordance with the S-REF model that those participants grouped as prone to both hallucinations and delusions would have higher levels of negative affect than those participants grouped as less prone. Furthermore, it was predicted that delusion proneness, dysfunctional metacognitive beliefs, and depression, stress, and anxiety would be significant predictors of hallucination proneness. Similarly, it was also predicted that hallucination proneness, dysfunctional metacognitive beliefs, and depression, anxiety, and stress would be significant predictors of delusion proneness.

Method

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

Participants

Utilising a cross-sectional survey design, approved by the Human Research Ethic Committee at the University of the Sunshine Coast, all students enrolled at the University of the Sunshine Coast, Queensland, Australia were invited via email to participate in an anonymous internet survey. A total of 715 students, with a mean age of 28.1 years (SD = 10.9, range 18–65), 567 females (79.3%) and 148 males (20.7%), completed the survey.

Materials

Participants completed a self-report questionnaire package which consisted of demographic information (including age and gender) as well as four separate measures assessing hallucination proneness, delusion proneness, negative affect, and dysfunctional metacognitions. Participants were advised not to report experiences that occurred under the influence of alcohol or narcotic substances.

Launay-Slade Hallucination Scale—Revised (LSHS-R)

The LSHS-R (Bentall & Slade, 1985) measures hallucinatory experiences in healthy individuals. Participants respond to each of 12 items exploring present and past hallucinatory experiences on a five-point scale, scoring each item as 0 = ‘Certainly does not apply to me’, 1 = ‘Possibly does not apply to me’, 2 = ‘Unsure’, 3 = ‘Possibly applies to me’, and 4 = ‘Certainly applies to me’. Total score range from 0 to 48, where a higher score indicates a greater predisposition to hallucinations. The English version of the LSHS-R has demonstrated good psychometric properties, with Cronbach's alpha coefficient reported as 0.84 (Jones & Fernyhough, 2009), and good test–retest reliability scores (0.84; Bentall & Slade, 1985).

Peters et al. Delusion Inventory (PDI-21)

The PDI-21 (Peters, Joseph, Day, & Garety, 2004) measures delusion proneness in normal populations. The inventory consists of 21 statements covering areas such as suspiciousness, grandiosity, paranormal beliefs, and persecution. For each statement, a ‘Yes’ (scored as 1) or ‘No’ (scored as 0) response is obtained, yielding a delusion score between 0 and 21. For each ‘Yes’ response, three additional dimensions are assessed to evaluate associated distress, preoccupation, and conviction. Each dimension is measured on a five-point scale (distress: ‘Not distressing at all’ to ‘Very distressing’; preoccupation: ‘Hardly ever think about it’ to ‘Think about it all the time’; conviction: ‘Don't believe it's true’ to ‘Believe it is absolutely true’). For each ‘No’ response obtained, an automatic score of 0 is given for all three additional measures. Combining the total scores of each dimension (range 0–105) with the yes/no delusion score provides a total PDI-21 score range of 0 to 336, where a higher score indicates increasing proneness to delusional beliefs (Peters et al., 2004). The PDI-21 has displayed satisfactory internal consistency (Cronbach's alpha coefficients range from 0.77 to 0.82) and test–retest reliability (range from 0.78 to 0.81; Jones & Fernyhough, 2007; Peters et al., 2004).

Metacognitions Questionnaire (MCQ-30)

The MCQ-30 (Wells & Cartwright-Hatton, 2004) measure individual differences in dysfunctional metacognitive beliefs, judgments, and thought monitoring tendencies. The questionnaire contains 30 items across five subscales: MCQ-CC = lack of cognitive confidence; MCQ-PB = positive beliefs about worry; MCQ-CSC = cognitive self-consciousness; MCQ-UDT = negative beliefs about uncontrollability and danger of thoughts; and MCQ-NCT = negative beliefs about the consequences of not controlling thoughts (Wells & Cartwright-Hatton, 2004). Each item is scored on a 4-point scale; 1 = ‘Do not agree’, 2 = ‘Agree slightly’, 3 = ‘Agree moderately’, and 4 = ‘Agree very much’, yielding a summed score for each factor (range 1–24), as well as a total score for the scale (range 30–120), where a higher score indicates more dysfunctional metacognitive beliefs. The MCQ-30 has displayed good to excellent internal consistency for the total score (Cronbach's alpha = 0.93), and the five factors (Cronbach's alpha range = 0.72 to 0.93), and has demonstrated high test–retest stability for the total score (0.75) and for 4 out of the 5 factors individually (with the exception of negative beliefs, with the lowest coefficient = 0.59; Wells & Cartwright-Hatton, 2004).

Depression Anxiety Stress Scale (DASS-21)

The DASS-21 (Lovibond & Lovibond, 1995) was used to assess negative affectivity. The scale consists of 7 items measuring depression, 7 items measuring anxiety, and 7 items measuring stress. Participants indicate how much each statement applied to them over the past week, using a 4-point scale; 0 = ‘Did not apply to me at all’, 1 = ‘Applied to me to some degree, or some of the time’, 2 = ‘Applied to me to a considerable degree, or a good part of the time’, or 3 = ‘Applied to me very much, or most of the time’ (Lovibond & Lovibond, 1995). Combining the scores and multiplying by two yields a total score between 0 and 42 for each scale separately (depression, anxiety, and stress). The DASS-21 has shown adequate internal consistency (Cronbach's alpha = 0.93), discriminant validity, as well as satisfactory convergent validity when compared to other measures of depression and anxiety (Lovibond & Lovibond, 1995).

Procedure

Basic demographic questions, the LSHS-R, PDI-21, MCQ-30, and DASS-21 were combined in Survey Monkey, an online tool for creating and administering surveys. Prior to commencement of the study, the complete survey (consisting of 84 test items) was piloted, with completion times ranging between 8 min and 13 min. Participant recruitment was achieved by an email sent out to all students currently enrolled at the university, inviting them to participate in an online survey. Participants were informed that the purpose of the study was to examine experiences of unusual thoughts and perceptions, negative emotional states, and, how students control, interpret, and modify their own thinking. Participants were also informed that their consent to take part in this research was indicated by their completion of the survey.

Results

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

SPSS (version 19.0; SPSS Inc., Chicago, IL, USA) was employed for statistical analyses. Data screening was conducted to ensure that appropriate assumptions were met. Data from an initial 759 participants were screened and corrected for missing values, resulting in a retained sample of 715. All variables were examined for fit between the distributions and assumptions for parametric tests. To normalise the anxiety and depression subscales of the DASS-21 and all five subscales of the MCQ-30, Log10 transformations were applied. To normalise the DASS-21 total data, a square root transformation was applied.

Descriptive statistics

The internal consistency of all scales and subscales were analysed. In all cases, Cronbach's alpha scores were 0.82 or above, indicating good to excellent reliability, with the exception of the MCQ-30 subscale ‘Need to control thoughts’, with a Cronbach's alpha of 0.69 (cf. 0.65 in Jones & Fernyhough, 2006; 0.72 in Wells & Cartwright-Hatton, 2004). Descriptive statistics are reported in Table 1.

Table 1. Untransformed mean score, standard deviation, range, and Cronbach's α for the LSHS-R, PDI-21, MCQ-30 and subscales, and DASS-21 and subscales (N = 715)
ScaleMSDRangeα
LSHS-R17.218.660–420.82
PDI-21 total50.9432.360–1520.92
MCQ-3058.9313.6131–1070.89
Cognitive confidence10.523.776–230.82
Positive beliefs10.213.806–230.88
Cognitive self-consciousness14.924.656–240.86
Uncontrollability and danger12.234.776–240.87
Need to control thoughts11.403.466–230.69
DASS-21-total28.2922.010–1180.93
 DASS-21-D8.718.830–420.89
 DASS-21-A6.837.200–350.82
 DASS-21-S12.678.750–370.86

Levels of negative affect in non-clinical participants who are prone to both hallucinations and delusions versus those who are not

An independent t-test was conducted to compare levels of negative affect (DASS-21 total) between participants prone to both hallucinations and delusions (high–high; n = 234) and participants not prone to either hallucinations and delusions (low–low; n= 243). Participants were grouped using a median split on the LSHS-R (Mdn = 17) and the PDI-21 total score (Mdn = 46). Participants whose score was equal to the median value were left out of the analysis (23 cases for LSHS-R and 11 cases for PDI-21). The two groups differed significantly, t(458.6) = −12.21, p < .001, with the high–high group (M = 5.89, SD = 2.01) reporting DASS-21 total scores 2.09 higher, 95% CI [–1.76, −2.44], than the low–low group (M = 3.80, SD = 1.72), with the effect size being large, Cohen's d = 1.11.

Cognitive-affective cross-sectional predictors of hallucination and delusion proneness

In order to investigate the independent predictors of hallucination and delusion proneness, two multiple regression analyses were conducted. The first regression explored the independent contribution of delusion proneness, dysfunctional metacognitions, and negative affect in predicting hallucination proneness. The LSHS-R score was used as the dependent variable, and the PDI-21 total, DASS-21-stress, DASS-21-depression and DASS-21-anxiety, and subscales of the MCQ-30 were entered into the regression simultaneously as predictors.

The PDI-21 total, DASS-21-stress, DASS-21-depression and DASS-21-anxiety, and subscales of the MCQ-30 in combination accounted for a significant 32% of the variability in LSHS-R scores, F(9, 703) = 36.74, p < .001, R2 = 0.32, adjusted R2 = 0.31. Three variables made significant independent contributions to hallucination proneness; delusion proneness (PDI-21; β = 0.48, t = 12.38, p < .001), cognitive self-consciousness (MCQ-CSC; β = 0.08, t = 2.28, p = .023), and negative beliefs about uncontrollability and danger of thoughts (MCQ-UDT; β = 0.11, t = 2.46, p = .014; see Table 2). While the effect size of the regression was large, Cohen's f2 = 0.47, unique MCQ-30 individual contributions were small; both CSC and UDT each explained less than 1% variance in LSHS-R scores.

Table 2. Multiple regression summary statistics for LSHS-R scores (N = 713)
VariableB95% [CI]βPartial rtp
  1. aVariable transformed using Log10; MCQ-CC = lack of cognitive confidence; MCQ-PB = positive beliefs about worry; MCQ-CSC = cognitive self-consciousness; MCQ-UDT = negative beliefs about uncontrollability and danger of thoughts; and MCQ-NCT = negative beliefs about the consequences of not controlling thoughts.

PDI-210.13[0.11, 0.15]0.480.4212.38<.001
DASS-21-Da0.96[−0.68, 2.60]0.050.041.15.252
DASS-21-Aa1.55[−0.11, 3.22]0.080.071.83.068
DASS-21-S−0.09[−0.19, 0.01]−0.09−0.07−1.74.082
MCQ-CCa1.26[−2.71, 5.24]0.020.020.62.534
MCQ-PBa−0.40[−4.12, 3.33]−0.01−0.01−0.21.834
MCQ-CSCa4.80[0.67, 8.93]0.080.092.28.023
MCQ-UDTa5.68[1.15, 10.21]0.110.092.46.014
MCQ-NCTa−4.62[−9.53, 0.30]−0.07−0.07−1.84.066

A second regression analysis explored the independent contributions of hallucination proneness, dysfunctional metacognitions, and negative affect in predicting delusion proneness. LSHS-R, DASS-21 stress, DASS-21 depression and DASS-21 anxiety, and subscales of the MCQ-30 entered into the regression simultaneously as predictors in combination accounted for a significant 47% of the variability in PDI-21 scores, F(9, 703) = 68.30, p < .001, R2 = 0.47, adjusted R2 = 0.46. Further examination showed that five variables made significant independent contributions to delusion scores: hallucination proneness (LSHS-R; β = 0.38, t = 12.38, p < .001), stress (DASS-S; β = 0.09, t = 1.97, p = .050), anxiety (DASS-A; β = 0.09, t = 2.41, p = .02), cognitive self-consciousness (MCQ-CSC; β = 0.10, t = 3.15, p = .002), and need to control thoughts (MCQ-NCT; β = 0.22, t = 6.59, p < .001; see Table 3). Similar to the hallucination results, while the effect size of the regression was large = 0.87, unique DASS-21 and MCQ-30 individual contributions were small. Both DASS-21 anxiety and stress both each explained less than 1% variance in PDI-30 scores, while even the strongest MCQ-30 factor (NCT) only explained less than 5% of PDI-30 variance.

Table 3. Multiple regression summary statistics for PDI-21 scores (N = 713)
VariableB95% [CI]βPartial rtp
  1. aVariable transformed using Log10; MCQ-CC = lack of cognitive confidence; MCQ-PB = positive beliefs about worry; MCQ-CSC = cognitive self-consciousness; MCQ-UDT = negative beliefs about uncontrollability and danger of thoughts; and MCQ-NCT = negative beliefs about the consequences of not controlling thoughts.

LSHS-R1.39[1.17, 1.61]0.380.4212.38<.001
DASS-21-Da3.04[−2.38, 8.46]0.040.041.10.271
DASS-21-Aa6.71[1.24, 12.19]0.090.092.41.016
DASS-21-S0.33[0.00, 0.65]0.090.071.97.050
MCQ-CCa12.01[−1.06, 25.06]0.060.071.80.072
MCQ-PBa−6.73[−18.98, 5.53]−0.03−0.04−1.08.281
MCQ-CSCa21.74[8.18, 35.31]0.100.123.15.002
MCQ-UDTa9.78[−5.19, 24.75]0.050.051.28.200
MCQ-NCTa52.88[37.13, 68.62]0.220.246.59<.001

Discussion

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

With regard to the first research aim, participants prone to both hallucinations and delusions reported significantly higher levels of negative affect than participants not prone to either hallucinations and delusions. While positive associations between hallucination proneness, delusion proneness, and negative affect have been found previously (e.g., Cella et al., 2008; Paulik et al., 2006), to the best of the researchers' knowledge, this is the first investigation to show that participants prone to both hallucinatory and delusional experiences report comparatively higher levels of negative affect.

In accordance with the S-REF and Morrison's (2001) model, delusions may result as secondary symptoms from misinterpretations of cognitive intrusions and abnormal perceptual experiences (Krabbendam et al., 2004; Larøi & Woodward, 2007). As according to Garety, Kuipers, Fowler, Freeman, and Bebbington (2001), experiencing negative emotions may further increase the chance of interpreting such symptoms as threatening, and if negative affect is a consequence of hallucinatory and delusional experiences, such symptoms could reinforce each other.

In relation to the second research aim, it was predicted that delusion proneness, dysfunctional metacognitive beliefs, and depression, stress, and anxiety would be significant predictors of hallucination proneness. The results provided some support for this hypothesis, showing that a substantial proportion (32%) of hallucination scores could be accounted for by the predictors. Delusion proneness was found to be by far the strongest predictor, with uncontrollability and danger and cognitive self-consciousness making small additional contributions towards hallucination scores. Furthermore, while initial correlations indicated small to medium positive associations between hallucination proneness and all measures of negative affect (stress, depression, and anxiety), these variables failed to make a significant contribution towards hallucination scores in the regression.

In accordance with the S-REF model of vulnerability to psychological dysfunction (Wells & Matthews, 1994), it has been suggested that a combination of positive and negative metacognitive beliefs are implicated in hallucination proneness (Baker & Morrison, 1998; Larøi & Van der Linden, 2005). Although the current investigation found that negative beliefs about the uncontrollability and danger of thoughts predicted hallucination proneness, these metacognition only explained very small percentages of hallucination proneness variance. These findings, which importantly controlled for intercorrelations between variables, suggest that metacognitions and negative effect are not central constructs in mediating or moderating hallucination proneness.

It was also predicted that hallucination proneness, dysfunctional metacognitive beliefs, and depression, anxiety, and stress would be significant predictors of delusion proneness. Results indicated that 47% of variance in delusion proneness scores could be explained by the predictors, however, hallucination proneness emerged as the strongest predictor of delusion scores. While metacognition beliefs such as need to control thoughts and cognitive self-consciousness, along with the anxiety and stress DASS-21 subscales appeared as significant cross-sectional predictors, only the metacognitions demonstrated any notable predictive value for delusion proneness. Interpretation of the combined positive findings for the role of metacognition in delusion proneness alongside the negative findings for hallucination, is that these results do not fit with the original S-REF influenced account of Morrison of negative metacognitive beliefs leading to intrusive thoughts generating cognitive dissonance which in turn motivates their externalisation (metacognition leading to the formation of psychotic symptoms). Rather it could be speculated that need to control thoughts and cognitive self-consciousness may, more simply, have a role in increasing attention towards unusual thought content and/or increasing distress, which may be a maintaining factor for these unusual thoughts.

The findings of the current investigation highlight the close association between hallucination proneness and delusion proneness in non-clinical samples. It was demonstrated that participants who are prone to both hallucinations and delusions report comparatively higher levels of negative affect. This finding is new in the literature, and may have important implications for mental health service provision. By controlling for covariations between hallucinations, delusions, dysfunctional metacognitions, and negative affect, the current study addressed an important methodological limitation seen in previous research. From these analyses, the cognitive-affective mechanism underlying the expression of hallucinatory and delusional experiences in non-clinical samples was further clarified. These data support the contention that metacognitions have only a limited association with hallucinatory but a larger involvement in the genesis of delusional experiences in non-clinical samples. While this is in opposition to theoretical standpoints such as by Morrison, Haddock, and Tarrier (1995) who propose hallucinatory and delusional experiences may be due to misattributed intrusive thoughts, images, and impulses, as well as Wells & Matthews' S-REF model (1994), these data do further support recent claims by Varese and Bentall (2011): ‘Although the results from the analogue studies included in this review support the existence of robust relationships between hallucination proneness and metacognitive beliefs, these studies suffered from methodological limitations which might have led to inflated estimations of these associations’ (p. 861).

These findings have implication for the management of hallucinations and delusions symptoms in non-clinical populations. The finding that participants prone to both hallucinations and delusions are at risk of elevated levels of negative affect supports the need for targeted mental health treatment for individuals who experience distress associated with such cognitions. Most importantly from this research, the suggestion that dysfunctional metacognitions do not play any meaningful association with the expression of hallucination symptoms in non-clinical samples, questions the evidence base of metacognitive-focussed techniques for the treatment of hallucination conditions in such populations.

References

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References