- Email: [email protected]
Referential delusions of communication and interpretations of gestures
Available online at www.sciencedirect.com
Psychiatry Research 158 (2008) 27 – 34 www.elsevier.com/locate/psychres
Referential delusions of communication and interpretations of gestures Sandra Bucci a , Mike Startup a,⁎, Paula Wynn a , Amanda Baker b , Terry J. Lewin b b
a School of Psychology, University of Newcastle, Callaghan, NSW 2308, Australia Centre for Mental Health Studies, University of Newcastle, James Fletcher Hospital, NSW 2300, Australia
Received 19 December 2006; received in revised form 12 March 2007; accepted 1 July 2007
Abstract Gestures are an important aspect of non-verbal communication, but people with schizophrenia have poor comprehension of them. However, the tests of gesture comprehension that have been used present only scenes in which interpersonal meaning is communicated, though there is evidence that people with psychotic disorders tend to perceive communications where none were intended. Such mistakes about non-verbal behaviour are the hallmark of a subtype of delusions of reference identified as delusions of communication. Thus we hypothesised that patients with delusions of communication would tend to misinterpret incidental movements as gestures and, since delusions are often derogatory to the self, they would also tend to misinterpret gestures as insulting. Patients with acute psychotic symptoms (n = 64) were recruited according to a 2 × 2 design (presence vs. absence of delusions of communication by presence vs. absence of auditory hallucinations). They, and 57 healthy controls, were presented with 20 brief video clips in which an actor either made a well-known gesture or an incidental movement. After each clip, they selected one of four interpretations: a correct interpretation if a gesture had been presented; the interpretation of a different gesture; an insulting interpretation; no gesture intended (correct for incidental movements). The patients made significantly more errors of all kinds than the controls, perceived significantly more of the incidental movements as gestures, and selected significantly more insulting interpretations of the clips. These differences between patients and controls were almost wholly due to patients with delusions of communication. These results suggest that the difficulties that people with delusions of communication experience in understanding gestures can be explained, at least in part, by the misattribution of self-generated internal events to external sources. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Delusions of reference; Psychosis; Non-verbal communication; Social cognition; Self-monitoring
1. Introduction Social competence is often very poor among people with psychotic disorders (Couture et al., 2006). Accumulating evidence suggests that difficulties in social interaction are at least partly due to inaccurate social ⁎ Corresponding author. Tel.: +61 2 49215979; fax: +61 2 49216980. E-mail address: [email protected] (M. Startup).
perceptions (e.g. Addington et al., 2006; Couture et al., 2006). An important part of social perception is the ability to interpret non-verbal communications correctly. People with schizophrenia have been found to have difficulties in interpreting affect from facial expressions (Edwards et al., 2002), from eyes (Bora et al., 2006), and from prosody (Edwards et al., 2002; Kuscharska-Pietura et al., 2005), and tend to misinterpret averted gazes as directed at the self (Hooker and Park, 2005).
0165-1781/$ - see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2007.07.004
28
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
Gestures and postures are important components of non-verbal communication, but their comprehension has rarely been studied among people with schizophrenia. Most of the research that has been conducted has employed the Profile of Non-verbal Sensitivity test (PONS; Rosenthal and Benowitz, 1986). The test yields separate scores for facial expressions, gestures, and voice intonation, and their combinations. People with schizophrenia have been found to perform poorly on this test not only in comparison with healthy controls (Toomey et al., 2002) but also with psychiatric patients without schizophrenia (Rosenthal and Benowitz, 1986), and these differences are also found for scenes in which gestures are the only cues (Monti and Fingeret, 1987). One of the drawbacks of the PONS is that it presents only scenes in which interpersonal meaning is being communicated. Thus it does not assess any tendencies to perceive intentions to communicate where none were actually present. This is a drawback because there is evidence that people with psychotic disorders tend to attribute knowledge and intentions to others to an excessive extent, an impairment that has been described as ‘hyper-theory of mind’ (Abu-Akel, 1999), or ‘overmentalizing’ (Frith, 2004). This would lead one to expect that people with psychotic disorders would tend to interpret actions as communications even if there were no intention to communicate. For example, Russell et al. (2006) presented participants with animated sequences showing simple triangles moving around a screen, and found that participants with schizophrenia who had symptoms of paranoia used mentalising terms to describe animations even when the triangles moved randomly relative to each other. Clearly, social competence is likely to be compromised if one mistakenly interprets incidental movements or self-stimulatory motor behaviour, such as smoothing one's hair or adjusting a shirt collar, as meaningful gestures. People with symptoms of paranoia not only make excessive attributions of knowledge and intentions to others, but what is mistakenly attributed tends to be about the self (Abu-Akel, 1999). Misattributions of intentions to communicate non-verbally are the hallmark of delusions of reference, especially the subtype of delusions of reference that Startup and Startup (2005) have identified as delusions of communication. These concern the mistaken sense that others are communicating with the self by subtle and oblique verbal (e.g. hints, innuendos) or non-verbal means (e.g. gestures, stances, arrangements of objects). These have been found, in two factor analyses (Bucci, 2006), to be statistically independent of other delusions of reference, such as false beliefs that others are surreptitiously observing (e.g.
using surveillance equipment), following, or gossiping about the self (‘delusions of observation’). Startup and Startup (2005) also found that delusions of communication were not associated with either auditory hallucinations or persecutory delusions. Startup and Startup (2005) noted that referential delusions of communication are similar to auditory hallucinations in that what seems to the patient to be communicated concerns the self and originates from the self, though this origin is not recognised but attributed externally. In other words, they suggested that these delusions might derive from impairments of what Frith (1992) described as self-monitoring, though the impairment would be centred on non-verbal channels in delusions of communication as opposed to verbal channels in auditory hallucinations. In order to test this theory, Bucci (2006) recruited patients with acute psychotic symptoms according to a 2 × 2 design: presence vs. absence of delusions of communication crossed with presence vs. absence of auditory hallucinations. The patients were then presented with 100 brief video clips in which an actor either made a well-known gesture or an incidental movement, with the clips being obscured by visual noise. For each clip, the patients indicated how confident they were that a gesture was portrayed. When the data were analysed by the methods of signal detection theory, all groups showed adequate sensitivity and the groups did not differ in sensitivity, but patients with delusions of communication showed a bias to report gestures which was not shown by patients with hallucinations alone. A control group of healthy volunteers showed significantly greater sensitivity than the patients and a more conservative bias than patients with delusions of communication. Bucci (2006) concluded that these results taken together provided evidence that people with delusions of communication have difficulty in monitoring their own thoughts about non-verbal communications and tend to attribute them externally. The foregoing considerations lead us to the following hypotheses. When presented with video clips in which an actor either makes a well-known gesture or an incidental movement, and participants are required to select an interpretation of the clips, patients with psychotic disorders will tend to select the wrong interpretation more often than healthy controls even when the clips are not obscured with noise. This prediction follows from the difficulties that psychotic patients have been found to have in interpreting gestures (Monti and Fingeret, 1987). In addition, patients with delusions of communication, compared with those without such delusions, will tend to misinterpret incidental movements as gestures as a result of their tendency to
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
‘over-mentalise’ about non-verbal communication (Bucci, 2006). Furthermore, since (according to many textbooks of psychiatry, e.g. Gelder et al., 1989) delusions of reference commonly involve communications that are derogatory to the self, patients with delusions of communication will have a stronger tendency than patients without these delusions to choose erroneous interpretations of the clips which are rejecting or insulting. Finally, since Startup and Startup (2005) hypothesised that delusions of communication arise from impairments of self-monitoring in non-verbal channels, as opposed to verbal channels, as in auditory hallucinations, and Bucci (2006) found that auditory hallucinations were not associated with a bias to misperceive movements as gestures, we expect that patients with auditory hallucinations will not differ from those without such hallucinations in the number of erroneous interpretations they make of gestures.
29
Table 1 Key characteristics of the patients (n = 64) and control subjects (n = 57) Patients
Controls
Characteristic
Mean (S.D.) or %
Mean (S.D.) or %
Mean age in years (S.D.) Male Australian born Never married Living arrangements with parents alone other No current employment Mean estimated IQ (S.D.) Diagnoses Schizophrenia spectrum Affective psychoses Other
34.2 (11.7) 62.5% 92.2% 76.6%
39.6 (16.3) 24.6% 84.2% 42.1%
37.5% 39.1% 23.4% 73.4% 99.9 (12.5)
12.3% 12.3% 75.4% 7.0% 111.8 (7.7)
62.5% 29.7% 7.8%
2.2. Materials 2. Method 2.1. Participants Eighty-seven people with positive psychotic symptoms, who had recently been admitted to a psychiatric hospital, were invited to take part in the study once their psychiatrist had declared them capable of providing informed consent. Sixty-four (74%) accepted the invitation and completed the interview and experimental task. For the healthy control sample, 200 people from the volunteer panel of a medical research institute were invited to participate. Fifty-seven (29%) accepted the invitation and completed the questionnaires and experimental task. Only patients over the age of 18 years were invited to participate, and all provided written consent after the purpose and the procedures of the study had been explained. The exclusion criteria common to both samples were (i) evidence of organic brain impairment; (ii) inadequate English fluency; and (iii) visual impairment that could not be corrected to normal. The control participants were required never to have been diagnosed with a mental illness. Additional exclusion criteria for the patients were questionable or mild delusions of communication or auditory hallucinations (to allow for clear separation between those with and without these symptoms); or mild, moderate or severe visual hallucinations (necessary given the visual nature of the experimental task). Characteristics of both the control and patient participants are shown in Table 1. Chart diagnoses of the patients were provided by the treating psychiatrists using DSM-IV criteria.
2.2.1. Clinical measures Auditory hallucinations among the patients were assessed using four items from the Scale for the Assessment of Positive Symptoms (SAPS; Andreasen, 1984): ‘auditory hallucinations’, ‘voices commenting’, ‘voices conversing’ and ‘global rating of the severity of auditory hallucinations’. Participants who scored 4 or 5 (marked or severe) on the global rating, and had experienced the symptoms within a week of assessment, were assigned to the ‘auditory hallucinations present’ group. Visual hallucinations were also measured by the SAPS ‘visual hallucinations’ item. No participant obtained a score of 2 or more on this item. Questions and probes developed by Startup and Startup (2005; Referential Delusions Interview) were used to inquire about five kinds of delusion of communication: whether the participant believed that information was being communicated: (1) verbally (e.g. hints, double meanings); (2) non-verbally (e.g. gestures, stances); (3) by the public media (e.g. TV, radio); (4) by animals; (5) via inanimate objects/processes (e.g. lights flickering, machine noises). Global ratings were made of the severity of these delusions using a scale modelled on the Ideas and Delusions of Reference item of the SAPS. Participants who scored 4 or 5 (marked or severe) on this item and had experienced the symptoms within a week of assessment were assigned to the ‘delusions of communication present’ group. The global cognitive functioning of both patients and controls was assessed using the National Adult Reading Test (NART), which has been found to provide good estimates of premorbid IQ even in acutely ill,
30
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
unmedicated and chronic schizophrenia patients (Crawford et al., 1992). 2.2.2. Gesture interpretation test — development The experimental task used short clips of an actress making movements towards a video camera, with each movement being either a well-known gesture (intending to communicate a meaning) or an incidental movement (not intending to communicate a meaning). Given that an established pool of gestures and incidental movements was not available, a pilot study was conducted to select the movements that would be included in the experimental trials. Initially, 14 gestures and 18 incidental movements made by one female actor were filmed in a well-lit room. Each one had a lead-in and lead-out time of 1 s in which the actor was seated in a neutral and stationary position gazing at the camera. All clips were silent and were similar in duration (mean = 3.7, ± 0.4 s). The 32 clips were shown to 10 male and 10 female colleagues who rated, on 10-point scales, (a) the likelihood that the movement in the clip was a gesture, and (b) the amount of movement. They also provided descriptions of the movements in the clips. The mean ratings of the amount of movement were similar across clips (range 6.1 to 8.6) and all descriptions of the movements coincided with the experimenters' intentions. Movements which were intended to be gestures and which received a mean rating of 7.5 or more on the ‘gesture likelihood’ response scale were retained (n = 10). Similarly, the movements which were intended to be incidental and which received a mean rating of 3.5 or less were retained (n = 10). 2.2.3. Gesture interpretation test — presentation and scoring The 20 clips were displayed at full screen size on a colour monitor (15" diagonal, 1024 × 768 resolution) which was positioned 60 cm from the participant at the optimal angle where the plane of the screen is at right angles to the participant's line of sight. Each clip was followed by four options from which to choose in order to answer the question, What does the actor's movement mean? The four options included, in random order: (i) the intended gesture (a ‘correct’ interpretation if a gesture had been presented); (ii) another gesture from the pool of gestures (‘incorrect’ gesture); (iii) a rejecting or insulting interpretation of the movement (‘rejecting’ interpretation); and (iv) no gesture intended (‘correct’ interpretation if an incidental movement had been presented). For example, when the presented gesture meant “stop”, the four options displayed were: (1) The
actor made a gesture meaning “Stop” (correct interpretation); (2) the actor made a gesture meaning “I can't hear you” (incorrect gesture); (3) the actor made a gesture meaning “I don't care what you have to say” (rejecting interpretation); and (4) the actor did not make a gesture (non-gesture). The gestures, incidental movements and their corresponding rejecting interpretations are shown in Table 2. Participants were given unlimited time to choose interpretations. Participants were required to select interpretations of the video clips. If a clip showed a gesture, their responses would be incorrect if they selected the interpretation of another gesture from the pool of gestures (‘incorrect’ gesture), if they chose a ‘rejecting’ interpretation of the movement, or if they indicated that the clip showed a non-gesture. On the other hand, if the clip showed an incidental movement, their responses would be incorrect if they selected a gesture interpretation (rejecting or nonrejecting). In addition to the total number of errors, particular subtypes of error were of interest. Thus scores were also calculated for the number of incidental movements which were interpreted as gestures, and for Table 2 Movements in the clips and ‘rejecting’ interpretations Movements Gestures Hand in front to indicate ‘Stop’ Hand to ear to indicate ‘I can't hear you’ Hand in front to wave ‘Hello’ Hand to lips to blow a kiss Head shake from side to side to indicate ‘No’ Hand in front beckoning ‘Come here’ Fingers crossed to indicate ‘Good luck’ Shrug shoulders to indicate ‘I don't know’ Finger to lips to indicate be quiet (‘Shh’) Thumbs down to indicate ‘No good’ Incidental movements Chew nails Brush hair with hair brush Tuck hair behind ears Fix an earring Swat a fly in front of face Apply lipstick Scratch neck Answer a mobile phone Tie up shoelace Pull socks up
‘Rejecting’ interpretations You smell bad Don't talk to me You're not making any sense You'll never amount to anything I don't care what you have to say Stay away from me You look terrible today You don't belong here You always get things wrong You don't belong to our group
You smell bad Don't talk to me You're not making any sense You'll never amount to anything I don't care what you have to say Stay away from me You look terrible today You don't belong here You always get things wrong You don't belong to our group
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
the number of clips (gestures or non-gestures) for which a rejecting interpretation was chosen.
31
3. Results 3.1. Inter-rater reliability
2.3. Procedure Participants were informed that the experiment was about how people interpret gestures. The instructions were as follows: There are times when people deliberately communicate some meaning to you with gestures. For example, if you were on the other side of the room and I wanted to speak with you, I might wave my hand like this (researcher demonstrates beckoning) rather than yell across the room. However, there are also times when people make movements that are not meant to communicate anything at all (researcher demonstrates brushing hair away from eyes). These are not gestures. What I would like you to do now is watch the clips on the computer screen and tell me what you think the actor's movements meant. You will be given four possible interpretations of the movements to choose from. Does that make sense? (Instructions were repeated if necessary). The first author conducted the clinical interviews, administered the NART, and then sat with the patients throughout the experimental task to ensure that no trials were missed, to prompt participants to respond if they were having difficulty, and to monitor their mental state throughout the procedure. One participant with auditory hallucinations but no delusions of communication responded to every clip that she was certain there was a gesture, suggesting that she had misunderstood the task requirements. Therefore, her data were removed. The third author administered the NART, and then sat throughout the experimental task with the control participants. 2.4. Data analyses Ideally, the data would have been analysed by 2 × 2 ANCOVAs with demographic variables entered as covariates. However, since the distributions of error scores were truncated and positively skewed, they were not suitable for parametric analyses. Therefore, differences between the four patient groups were analysed by the Kruskall–Wallis test. This was a substitute for the overall test of a model provided by ANOVA. It was followed by comparisons using the Mann–Whitney U test, which were the nonparametric equivalents of tests of main effects. Unfortunately, there is no non-parametric equivalent of a test of an interaction. Comparisons between the patient groups and the controls were analysed in a series of Mann–Whitney U tests.
The interviewer recorded patients' responses verbatim on the interview schedule, and 19 of the interviews were then rated for symptom severity by the second author who was blind to the interviewer's ratings. The inter-rater reliability for all delusions and hallucinations ratings was high (Pearson r = 0.89–1.00) and significant. Therefore, the interviewer's ratings were used in subsequent analyses. The patients were assigned to four groups based on the severity thresholds for the global rating of auditory hallucinations and the global rating of delusions of communication. There was perfect agreement between the interviewer and the independent assessor on the assignment to their respective groups of the 19 participants that they both assessed. 3.2. Interpretations of the clips The patient groups did not differ significantly on any of the categorical variables shown in Table 1, but there was a marginally significant trend for schizophrenia spectrum disorders to be more common in the groups with delusions of communication (χ2(3) = 6.51, P = 0.091). When the continuous variables were analysed in a 2 (delusions of communication; present vs. absent) × 2 (auditory hallucinations; present vs. absent) design, no significant effects were found for age or IQ. 3.2.1. Total errors The mean numbers of incorrect interpretations, of all kinds, for the four patient groups and the healthy controls are shown in Table 3. It can be seen that the great majority of interpretations of the clips that the patients selected were correct; 93.3% of gestures and 84.8% of incidental movements. Nevertheless, the four groups were found to differ significantly, by the Kruskall–Wallis test, χ2(3) = 8.96, P = 0.03, in the number of incorrect interpretations they made. Further analyses of the source of this difference, with Mann–Whitney U tests, showed that people with auditory hallucinations did not differ from those without hallucinations, Z = 0.008, P = 0.99, but there was a highly significant difference between people with delusions of communication and those without them, Z = 2.98, P = 0.003. People with delusions of communication made an average of 2.94 errors (14.7%, S.D. = 0.57) while people without these delusions made an average of only 0.94 errors (4.7%, S.D. = 0.65). Differences in the mean number of errors of all kinds made by the patient sample as a whole and the complete
32
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
Table 3 Characteristics and mean number (S.D.) of misinterpretations of video clips in three categories for four patient groups and healthy controls Delusions of communication Present
Controls
Absent
Auditory hallucinations
Characteristics N Schizophrenia spectrum (%) % male Mean age (years) Estimated IQ Misinterpretations Total incorrect Rejections IM as gestures
Present Absent
Present Absent
17 76
11 45
17 41
73 32.1 98.7
59 40.8 99.2
1.36 (1.91) 0.72 (0.65) 0.91 (0.94)
1.06 (2.19) 0.65 (1.06) 0.59 (1.00)
65 30.1 98.7 3.18 (4.38) 1.94 (2.97) 2.29 (2.92)
18 73 73 31.2 102.1 2.78 (3.12) 1.22 (1.44) 2.00 (2.70)
57 0 25 39.6 111.8 0.40 (0.62) 0.33 (0.51) 0.30 (0.46)
IM: incidental movements.
sample of healthy controls were analysed with the Mann–Whitney U test. The mean for the patients, 2.17 (S.D. = 3.21), was significantly higher than for the controls, 0.40 (S.D. = 0.62), Z = 4.03, P b 0.001. Differences in the number of errors between each of the patient groups and the controls were analysed with a series of Mann–Whitney U tests. These showed the greatest difference was for patients with delusions of communication, who made significantly more errors than the controls, Z = 4.71, P b 0.001, while the difference between patients without delusions of communication and controls was non-significant, Z = 1.81, P = 0.07. 3.2.2. Specific errors Differences on specific kinds of errors between the patient sample as a whole and the controls were also significant for interpretations of rejections, Z = 3.10, P = 0.002, and for misinterpretations of incidental movements as gestures, Z = 3.52, P b 0.001. Differences in the number of these errors between each of the patient groups and the controls were analysed with a series of Mann–Whitney U tests. The greatest differences were for patients with delusions of communication, who made significantly more interpretations of rejections, Z = 3.26, P = 0.001, and significantly more interpretations of incidental movements as gestures, Z = 3.73, P b 0.001. Patients without delusions of communication did not perceive more rejecting messages than controls, Z = 1.84, P = 0.07, though they did consider more of the incidental movements to be gestures, Z = 2.07, P = 0.04.
3.2.3. Matched control sample The healthy controls were significantly older than the patients, t(99.3) = 2.20, P = 0.03, higher in average estimated IQ, t (94.7) = 6.11, P b 0.001, and included a higher ratio of females, χ2(1) = 20.9, P b 0.001. In order to construct a subsample of control participants who were matched more closely to the patients, all 14 of the male controls were retained but only the nine females with the lowest IQs were kept. The gender ratio of the resulting subsample then no longer differed from the controls, χ2(1) = 0.06, P = 0.81, nor did the average age differ, t = 1.24, P = 0.23, but the mean IQ for the controls (mean: 108.0, S.D.: 8.65) was still significantly higher, t = 2.88, P = 0.005. When the analyses reported above were repeated with the subsample of 23 healthy controls who were matched to the patients, the results were substantially the same. The only result that would lead to a different conclusion was that the difference in interpretations of incidental movements as gestures, between the controls and patients who did not have delusions of communication, was reduced to marginal significance, Z = 1.89, P = 0.058. The matched controls had a significantly higher mean IQ than the patients, but the correlations between IQ and errors were all non-significant (Spearman's rho = − 0.08 to − 0.16) when these groups were combined. 4. Discussion In general, the control participants interpreted the meaning of brief video clips of human movements in line with what was intended. On average, there was 98% agreement (average of 0.4 interpretations ‘incorrect’). This provides strong support for the validity of the test. Among the patients also, the great majority of interpretations of the clips were correct (89%). However, this was significantly less than the healthy controls and is consistent with previous research showing poor comprehension of gestures among people with schizophrenia (Monti and Fingeret, 1987). It is also consistent the low sensitivity, compared with healthy controls, observed by Bucci (2006), in psychotic patients asked to distinguish gestures from incidental movements when video clips were obscured with visual noise. Few of the errors made by the patients consisted of selecting the interpretation of a different gesture from the one that was portrayed (mean = 0.13 errors). They were far more likely to misinterpret an incidental movement as a gesture, and they did this significantly more often than the controls. This is consistent with the ‘over-mentalising’ (Frith, 2004) that has been observed
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
when people with paranoia interpret animated sequences of triangles moving around a screen (Russell et al., 2006). However, these errors were mostly made by people with delusions of communication. This is consistent with the bias to perceive incidental movements as gestures that Bucci (2006) found among people with delusions of communication when video clips of movements were obscured with visual noise. Another kind of error made by the patients more often than the controls was to misinterpret movements as having a rejecting or insulting meaning, even though all the movements in the video clips were affectively neutral in content and the actor maintained a neutral expression. Again these errors were mostly made by people with delusions of communication. This is consistent with the interpretation that Bucci (2006) offered for the results of her signal detection analysis of the perception of gestures, that people with delusions of communication are impaired in self-monitoring for non-verbal material. That is, they appear to misattribute their self-generated internal events, especially self-disparaging preoccupations, to an external source. Interpretation of the findings should be made in light of three main methodological limitations. Firstly, according to chart diagnoses, schizophrenia spectrum disorders were more common among people with delusions of communication (though not significantly so). This is not surprising, given that diagnoses are partly based on these symptoms, but it does mean that group membership and diagnosis are confounded to some extent. Unfortunately, it was not possible to control this difference statistically because the distributions of error scores ruled out the use of parametric statistics. Thus we cannot be sure that the results for people with delusions of communication are not better explained by the presence of schizophrenia spectrum disorders. Secondly, the control participants were on average significantly older and higher in IQ than the patients and included a higher ratio of females. However, the results were not substantively altered when the patients were compared with controls matched in age and gender ratio, and IQ was found not to be correlated with the number of errors. Thirdly, since we included only patients with marked or severe symptoms, we cannot say what effects mild or questionable symptoms might have on interpretations of gestures. Difficulties in social cognition among people with schizophrenia are usually explained in terms of deficits in neurocognition, and certainly there is evidence to support this view (Sergi et al., 2007). For example, it has been found that abilities in affect recognition correlate with immediate memory and general intelligence (Pink-
33
ham and Penn, 2006), and with attentional dysfunction and span of apprehension (Nienow et al., 2006). Performance on the Half Profile of Non-verbal Sensitivity has been found to correlate with degree of sensorimotor gating (Wynn et al., 2005). Such results suggest that deficits in social cognition arise to some extent from difficulties in decoding incoming social information. However, the results of the present study, and those of the study by Bucci (2006), suggest that the difficulties that people with delusions of communication experience in understanding gestures can be explained, at least in part, by the misattribution of self-generated internal events to an external source. These misattributions might themselves be explained by neurocognitive deficits, such as poor intentional inhibition of currently irrelevant memories of previous non-verbal communications, and poor contextual memory, as suggested by Bucci (2006). Nevertheless, the route by which such deficits would have their effect on social cognition would be top-down rather than bottom-up. The results presented above suggest that the difficulties that people suffering acute psychotic episodes experience with interpreting gestures is largely confined to people with delusions of communication. This does not diminish the clinical significance, however, because delusions of reference are among the most common positive psychotic symptoms not only among psychotic inpatients (Minas et al., 1992), but also among outpatients with schizophrenia (World Health Organization, 1973). The proportion of stabilized outpatients who have delusions of communication is currently unknown, but the results of the present study suggest that about 55% of acutely psychotic inpatients suffer from them, though the sample is too small for confidence to be placed in this estimate. Gestures, of course, are only one form of non-verbal communication. Others include facial expression, prosody and gaze. It is possible that delusions of communication could also add to our understanding of the difficulties that people with psychotic disorders have in interpreting these forms of communication. For example, it might be that the tendency to misinterpret averted gazes as directed at the self, which is found among people with schizophrenia (Hooker and Park, 2005), is confined to people with delusions of communication. Further research will be required to clarify this. Acknowledgement We wish to acknowledge the support of the Hunter Medical Research Institute Volunteer Register for their assistance with recruitment of control participants.
34
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
References Abu-Akel, A., 1999. Impaired theory of mind in schizophrenia. Pragmatics & Cognition 7, 247–282. Addington, J., Saeedi, H., Addington, D., 2006. Influence of social perception and social knowledge on cognitive and social functioning in early psychosis. British Journal of Psychiatry 189, 373–378. Andreasen, N.C., 1984. Scale for the Assessment of Positive Symptoms (SAPS). University of Iowa, Iowa City. Bora, E., Eryavuz, A., Kayahan, B., Sungu, G., Veznedaroglu, B., 2006. Social functioning, theory of mind and neurocognition in outpatients with schizophrenia: mental state decoding may be a better predictor of social functioning than mental state reasoning. Psychiatry Research 145, 95–103. Bucci, S., 2006. Referential delusions of communication and reality monitoring deficits in psychosis. Unpublished DClinPsych thesis, University of Newcastle, Australia. Couture, S.M., Penn, D.L., Roberts, D.L., 2006. The functional significance of social cognition in schizophrenia: a review. Schizophrenia Bulletin 32 (Suppl. 1), S44–S63. Crawford, J.R., Besson, J.A.O., Bremner, M., Ebmeier, K.P., Cochrane, R.H.B., Kirkwood, K., 1992. Estimation of premorbid intelligence in schizophrenia. British Journal of Psychiatry 161, 69–74. Edwards, J., Jackson, H.J., Pattison, P.E., 2002. Emotion recognition via facial expression and affective prosody in schizophrenia: a methodological review. Clinical Psychology Review 22, 789–832. Frith, C.D., 1992. The Cognitive Neuropsychology of Schizophrenia. Lawrence Erlbaum, Hove (UK). Frith, C.D., 2004. Schizophrenia and theory of mind. Psychological Medicine 34, 385–389. Gelder, M., Gath, D., Mayou, R., 1989. Oxford Textbook of Psychiatry (2nd ed.). Oxford University Press, Oxford. Hooker, C., Park, S., 2005. You must be looking at me: the nature of gaze perception in schizophrenia patients. Cognitive Neuropsychiatry 10, 327–345. Kucharska-Pietura, K., David, A.S., Masiak, M., Phillips, M.L., 2005. Perception of facial and vocal affect by people with schizophrenia in early and late stages of illness. British Journal of Psychiatry 187, 523–528.
Minas, I.H., Stuart, G.W., Klimidis, S., Jackson, H.J., Singh, B.S., Copolov, D.L., 1992. Positive and negative symptoms in the psychoses: multidimensional scaling of SAPS and SANS items. Schizophrenia Research 8, 143–156. Monti, P.M., Fingeret, A.L., 1987. Social perception and communication skills among schizophrenics and nonschizophrenics. Journal of Clinical Psychology 43, 197–205. Nienow, T.M., Docherty, N.M., Cohen, A.S., Dinzeo, T.J., 2006. Attentional dysfunction, social perception, and social competence: what is the nature of the relationship? Journal of Abnormal Psychology 115, 408–417. Pinkham, A.E., Penn, D.L., 2006. Neurocognitive and social cognitive predictors of interpersonal skill in schizophrenia. Psychiatry Research 143, 167–178. Rosenthal, R., Benowitz, L.I., 1986. Sensitivity to nonverbal communications in normal, psychiatric, and brain damaged samples. In: Blanck, P.D., Buck, R., Rosenthal, R. (Eds.), Nonverbal Communication in the Clinical Context. Pennsylvania University Press, Pennsylvania. Russell, T.A., Reynaud, E., Herba, C., Morris, R., Corcoran, R., 2006. Do you see what I see? Interpretations of intentional movement in schizophrenia. Schizophrenia Research 81, 101–111. Sergi, M.J., Rassovsky, Y., Widmark, C., Reist, C., Erhart, S., Braff, D.L., Marder, S.R., Green, M.F., 2007. Social cognition in schizophrenia: relationships with neurocognition and negative symptoms. Schizophrenia Research 90 (1–3), 316–324. Startup, M., Startup, S., 2005. On two kinds of delusion of reference. Psychiatry Research 137, 87–92. Toomey, R., Schuldberg, D., Corrigan, P., Green, M.F., 2002. Nonverbal social perception and symptomatology in schizophrenia. Schizophrenia Research 53, 83–91. Wynn, J.K., Sergi, M.J., Dawson, M.E., Schell, A.M., Grenn, M.F., 2005. Sensorimotor gating, orienting and social perception in schizophrenia. Schizophrenia Research 73, 319–325. World Health Organization, 1973. Report of the International Pilot Study of Schizophrenia, vol. 1. World Health Organization, Geneva.
Psychiatry Research 158 (2008) 27 – 34 www.elsevier.com/locate/psychres
Referential delusions of communication and interpretations of gestures Sandra Bucci a , Mike Startup a,⁎, Paula Wynn a , Amanda Baker b , Terry J. Lewin b b
a School of Psychology, University of Newcastle, Callaghan, NSW 2308, Australia Centre for Mental Health Studies, University of Newcastle, James Fletcher Hospital, NSW 2300, Australia
Received 19 December 2006; received in revised form 12 March 2007; accepted 1 July 2007
Abstract Gestures are an important aspect of non-verbal communication, but people with schizophrenia have poor comprehension of them. However, the tests of gesture comprehension that have been used present only scenes in which interpersonal meaning is communicated, though there is evidence that people with psychotic disorders tend to perceive communications where none were intended. Such mistakes about non-verbal behaviour are the hallmark of a subtype of delusions of reference identified as delusions of communication. Thus we hypothesised that patients with delusions of communication would tend to misinterpret incidental movements as gestures and, since delusions are often derogatory to the self, they would also tend to misinterpret gestures as insulting. Patients with acute psychotic symptoms (n = 64) were recruited according to a 2 × 2 design (presence vs. absence of delusions of communication by presence vs. absence of auditory hallucinations). They, and 57 healthy controls, were presented with 20 brief video clips in which an actor either made a well-known gesture or an incidental movement. After each clip, they selected one of four interpretations: a correct interpretation if a gesture had been presented; the interpretation of a different gesture; an insulting interpretation; no gesture intended (correct for incidental movements). The patients made significantly more errors of all kinds than the controls, perceived significantly more of the incidental movements as gestures, and selected significantly more insulting interpretations of the clips. These differences between patients and controls were almost wholly due to patients with delusions of communication. These results suggest that the difficulties that people with delusions of communication experience in understanding gestures can be explained, at least in part, by the misattribution of self-generated internal events to external sources. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Delusions of reference; Psychosis; Non-verbal communication; Social cognition; Self-monitoring
1. Introduction Social competence is often very poor among people with psychotic disorders (Couture et al., 2006). Accumulating evidence suggests that difficulties in social interaction are at least partly due to inaccurate social ⁎ Corresponding author. Tel.: +61 2 49215979; fax: +61 2 49216980. E-mail address: [email protected] (M. Startup).
perceptions (e.g. Addington et al., 2006; Couture et al., 2006). An important part of social perception is the ability to interpret non-verbal communications correctly. People with schizophrenia have been found to have difficulties in interpreting affect from facial expressions (Edwards et al., 2002), from eyes (Bora et al., 2006), and from prosody (Edwards et al., 2002; Kuscharska-Pietura et al., 2005), and tend to misinterpret averted gazes as directed at the self (Hooker and Park, 2005).
0165-1781/$ - see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2007.07.004
28
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
Gestures and postures are important components of non-verbal communication, but their comprehension has rarely been studied among people with schizophrenia. Most of the research that has been conducted has employed the Profile of Non-verbal Sensitivity test (PONS; Rosenthal and Benowitz, 1986). The test yields separate scores for facial expressions, gestures, and voice intonation, and their combinations. People with schizophrenia have been found to perform poorly on this test not only in comparison with healthy controls (Toomey et al., 2002) but also with psychiatric patients without schizophrenia (Rosenthal and Benowitz, 1986), and these differences are also found for scenes in which gestures are the only cues (Monti and Fingeret, 1987). One of the drawbacks of the PONS is that it presents only scenes in which interpersonal meaning is being communicated. Thus it does not assess any tendencies to perceive intentions to communicate where none were actually present. This is a drawback because there is evidence that people with psychotic disorders tend to attribute knowledge and intentions to others to an excessive extent, an impairment that has been described as ‘hyper-theory of mind’ (Abu-Akel, 1999), or ‘overmentalizing’ (Frith, 2004). This would lead one to expect that people with psychotic disorders would tend to interpret actions as communications even if there were no intention to communicate. For example, Russell et al. (2006) presented participants with animated sequences showing simple triangles moving around a screen, and found that participants with schizophrenia who had symptoms of paranoia used mentalising terms to describe animations even when the triangles moved randomly relative to each other. Clearly, social competence is likely to be compromised if one mistakenly interprets incidental movements or self-stimulatory motor behaviour, such as smoothing one's hair or adjusting a shirt collar, as meaningful gestures. People with symptoms of paranoia not only make excessive attributions of knowledge and intentions to others, but what is mistakenly attributed tends to be about the self (Abu-Akel, 1999). Misattributions of intentions to communicate non-verbally are the hallmark of delusions of reference, especially the subtype of delusions of reference that Startup and Startup (2005) have identified as delusions of communication. These concern the mistaken sense that others are communicating with the self by subtle and oblique verbal (e.g. hints, innuendos) or non-verbal means (e.g. gestures, stances, arrangements of objects). These have been found, in two factor analyses (Bucci, 2006), to be statistically independent of other delusions of reference, such as false beliefs that others are surreptitiously observing (e.g.
using surveillance equipment), following, or gossiping about the self (‘delusions of observation’). Startup and Startup (2005) also found that delusions of communication were not associated with either auditory hallucinations or persecutory delusions. Startup and Startup (2005) noted that referential delusions of communication are similar to auditory hallucinations in that what seems to the patient to be communicated concerns the self and originates from the self, though this origin is not recognised but attributed externally. In other words, they suggested that these delusions might derive from impairments of what Frith (1992) described as self-monitoring, though the impairment would be centred on non-verbal channels in delusions of communication as opposed to verbal channels in auditory hallucinations. In order to test this theory, Bucci (2006) recruited patients with acute psychotic symptoms according to a 2 × 2 design: presence vs. absence of delusions of communication crossed with presence vs. absence of auditory hallucinations. The patients were then presented with 100 brief video clips in which an actor either made a well-known gesture or an incidental movement, with the clips being obscured by visual noise. For each clip, the patients indicated how confident they were that a gesture was portrayed. When the data were analysed by the methods of signal detection theory, all groups showed adequate sensitivity and the groups did not differ in sensitivity, but patients with delusions of communication showed a bias to report gestures which was not shown by patients with hallucinations alone. A control group of healthy volunteers showed significantly greater sensitivity than the patients and a more conservative bias than patients with delusions of communication. Bucci (2006) concluded that these results taken together provided evidence that people with delusions of communication have difficulty in monitoring their own thoughts about non-verbal communications and tend to attribute them externally. The foregoing considerations lead us to the following hypotheses. When presented with video clips in which an actor either makes a well-known gesture or an incidental movement, and participants are required to select an interpretation of the clips, patients with psychotic disorders will tend to select the wrong interpretation more often than healthy controls even when the clips are not obscured with noise. This prediction follows from the difficulties that psychotic patients have been found to have in interpreting gestures (Monti and Fingeret, 1987). In addition, patients with delusions of communication, compared with those without such delusions, will tend to misinterpret incidental movements as gestures as a result of their tendency to
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
‘over-mentalise’ about non-verbal communication (Bucci, 2006). Furthermore, since (according to many textbooks of psychiatry, e.g. Gelder et al., 1989) delusions of reference commonly involve communications that are derogatory to the self, patients with delusions of communication will have a stronger tendency than patients without these delusions to choose erroneous interpretations of the clips which are rejecting or insulting. Finally, since Startup and Startup (2005) hypothesised that delusions of communication arise from impairments of self-monitoring in non-verbal channels, as opposed to verbal channels, as in auditory hallucinations, and Bucci (2006) found that auditory hallucinations were not associated with a bias to misperceive movements as gestures, we expect that patients with auditory hallucinations will not differ from those without such hallucinations in the number of erroneous interpretations they make of gestures.
29
Table 1 Key characteristics of the patients (n = 64) and control subjects (n = 57) Patients
Controls
Characteristic
Mean (S.D.) or %
Mean (S.D.) or %
Mean age in years (S.D.) Male Australian born Never married Living arrangements with parents alone other No current employment Mean estimated IQ (S.D.) Diagnoses Schizophrenia spectrum Affective psychoses Other
34.2 (11.7) 62.5% 92.2% 76.6%
39.6 (16.3) 24.6% 84.2% 42.1%
37.5% 39.1% 23.4% 73.4% 99.9 (12.5)
12.3% 12.3% 75.4% 7.0% 111.8 (7.7)
62.5% 29.7% 7.8%
2.2. Materials 2. Method 2.1. Participants Eighty-seven people with positive psychotic symptoms, who had recently been admitted to a psychiatric hospital, were invited to take part in the study once their psychiatrist had declared them capable of providing informed consent. Sixty-four (74%) accepted the invitation and completed the interview and experimental task. For the healthy control sample, 200 people from the volunteer panel of a medical research institute were invited to participate. Fifty-seven (29%) accepted the invitation and completed the questionnaires and experimental task. Only patients over the age of 18 years were invited to participate, and all provided written consent after the purpose and the procedures of the study had been explained. The exclusion criteria common to both samples were (i) evidence of organic brain impairment; (ii) inadequate English fluency; and (iii) visual impairment that could not be corrected to normal. The control participants were required never to have been diagnosed with a mental illness. Additional exclusion criteria for the patients were questionable or mild delusions of communication or auditory hallucinations (to allow for clear separation between those with and without these symptoms); or mild, moderate or severe visual hallucinations (necessary given the visual nature of the experimental task). Characteristics of both the control and patient participants are shown in Table 1. Chart diagnoses of the patients were provided by the treating psychiatrists using DSM-IV criteria.
2.2.1. Clinical measures Auditory hallucinations among the patients were assessed using four items from the Scale for the Assessment of Positive Symptoms (SAPS; Andreasen, 1984): ‘auditory hallucinations’, ‘voices commenting’, ‘voices conversing’ and ‘global rating of the severity of auditory hallucinations’. Participants who scored 4 or 5 (marked or severe) on the global rating, and had experienced the symptoms within a week of assessment, were assigned to the ‘auditory hallucinations present’ group. Visual hallucinations were also measured by the SAPS ‘visual hallucinations’ item. No participant obtained a score of 2 or more on this item. Questions and probes developed by Startup and Startup (2005; Referential Delusions Interview) were used to inquire about five kinds of delusion of communication: whether the participant believed that information was being communicated: (1) verbally (e.g. hints, double meanings); (2) non-verbally (e.g. gestures, stances); (3) by the public media (e.g. TV, radio); (4) by animals; (5) via inanimate objects/processes (e.g. lights flickering, machine noises). Global ratings were made of the severity of these delusions using a scale modelled on the Ideas and Delusions of Reference item of the SAPS. Participants who scored 4 or 5 (marked or severe) on this item and had experienced the symptoms within a week of assessment were assigned to the ‘delusions of communication present’ group. The global cognitive functioning of both patients and controls was assessed using the National Adult Reading Test (NART), which has been found to provide good estimates of premorbid IQ even in acutely ill,
30
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
unmedicated and chronic schizophrenia patients (Crawford et al., 1992). 2.2.2. Gesture interpretation test — development The experimental task used short clips of an actress making movements towards a video camera, with each movement being either a well-known gesture (intending to communicate a meaning) or an incidental movement (not intending to communicate a meaning). Given that an established pool of gestures and incidental movements was not available, a pilot study was conducted to select the movements that would be included in the experimental trials. Initially, 14 gestures and 18 incidental movements made by one female actor were filmed in a well-lit room. Each one had a lead-in and lead-out time of 1 s in which the actor was seated in a neutral and stationary position gazing at the camera. All clips were silent and were similar in duration (mean = 3.7, ± 0.4 s). The 32 clips were shown to 10 male and 10 female colleagues who rated, on 10-point scales, (a) the likelihood that the movement in the clip was a gesture, and (b) the amount of movement. They also provided descriptions of the movements in the clips. The mean ratings of the amount of movement were similar across clips (range 6.1 to 8.6) and all descriptions of the movements coincided with the experimenters' intentions. Movements which were intended to be gestures and which received a mean rating of 7.5 or more on the ‘gesture likelihood’ response scale were retained (n = 10). Similarly, the movements which were intended to be incidental and which received a mean rating of 3.5 or less were retained (n = 10). 2.2.3. Gesture interpretation test — presentation and scoring The 20 clips were displayed at full screen size on a colour monitor (15" diagonal, 1024 × 768 resolution) which was positioned 60 cm from the participant at the optimal angle where the plane of the screen is at right angles to the participant's line of sight. Each clip was followed by four options from which to choose in order to answer the question, What does the actor's movement mean? The four options included, in random order: (i) the intended gesture (a ‘correct’ interpretation if a gesture had been presented); (ii) another gesture from the pool of gestures (‘incorrect’ gesture); (iii) a rejecting or insulting interpretation of the movement (‘rejecting’ interpretation); and (iv) no gesture intended (‘correct’ interpretation if an incidental movement had been presented). For example, when the presented gesture meant “stop”, the four options displayed were: (1) The
actor made a gesture meaning “Stop” (correct interpretation); (2) the actor made a gesture meaning “I can't hear you” (incorrect gesture); (3) the actor made a gesture meaning “I don't care what you have to say” (rejecting interpretation); and (4) the actor did not make a gesture (non-gesture). The gestures, incidental movements and their corresponding rejecting interpretations are shown in Table 2. Participants were given unlimited time to choose interpretations. Participants were required to select interpretations of the video clips. If a clip showed a gesture, their responses would be incorrect if they selected the interpretation of another gesture from the pool of gestures (‘incorrect’ gesture), if they chose a ‘rejecting’ interpretation of the movement, or if they indicated that the clip showed a non-gesture. On the other hand, if the clip showed an incidental movement, their responses would be incorrect if they selected a gesture interpretation (rejecting or nonrejecting). In addition to the total number of errors, particular subtypes of error were of interest. Thus scores were also calculated for the number of incidental movements which were interpreted as gestures, and for Table 2 Movements in the clips and ‘rejecting’ interpretations Movements Gestures Hand in front to indicate ‘Stop’ Hand to ear to indicate ‘I can't hear you’ Hand in front to wave ‘Hello’ Hand to lips to blow a kiss Head shake from side to side to indicate ‘No’ Hand in front beckoning ‘Come here’ Fingers crossed to indicate ‘Good luck’ Shrug shoulders to indicate ‘I don't know’ Finger to lips to indicate be quiet (‘Shh’) Thumbs down to indicate ‘No good’ Incidental movements Chew nails Brush hair with hair brush Tuck hair behind ears Fix an earring Swat a fly in front of face Apply lipstick Scratch neck Answer a mobile phone Tie up shoelace Pull socks up
‘Rejecting’ interpretations You smell bad Don't talk to me You're not making any sense You'll never amount to anything I don't care what you have to say Stay away from me You look terrible today You don't belong here You always get things wrong You don't belong to our group
You smell bad Don't talk to me You're not making any sense You'll never amount to anything I don't care what you have to say Stay away from me You look terrible today You don't belong here You always get things wrong You don't belong to our group
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
the number of clips (gestures or non-gestures) for which a rejecting interpretation was chosen.
31
3. Results 3.1. Inter-rater reliability
2.3. Procedure Participants were informed that the experiment was about how people interpret gestures. The instructions were as follows: There are times when people deliberately communicate some meaning to you with gestures. For example, if you were on the other side of the room and I wanted to speak with you, I might wave my hand like this (researcher demonstrates beckoning) rather than yell across the room. However, there are also times when people make movements that are not meant to communicate anything at all (researcher demonstrates brushing hair away from eyes). These are not gestures. What I would like you to do now is watch the clips on the computer screen and tell me what you think the actor's movements meant. You will be given four possible interpretations of the movements to choose from. Does that make sense? (Instructions were repeated if necessary). The first author conducted the clinical interviews, administered the NART, and then sat with the patients throughout the experimental task to ensure that no trials were missed, to prompt participants to respond if they were having difficulty, and to monitor their mental state throughout the procedure. One participant with auditory hallucinations but no delusions of communication responded to every clip that she was certain there was a gesture, suggesting that she had misunderstood the task requirements. Therefore, her data were removed. The third author administered the NART, and then sat throughout the experimental task with the control participants. 2.4. Data analyses Ideally, the data would have been analysed by 2 × 2 ANCOVAs with demographic variables entered as covariates. However, since the distributions of error scores were truncated and positively skewed, they were not suitable for parametric analyses. Therefore, differences between the four patient groups were analysed by the Kruskall–Wallis test. This was a substitute for the overall test of a model provided by ANOVA. It was followed by comparisons using the Mann–Whitney U test, which were the nonparametric equivalents of tests of main effects. Unfortunately, there is no non-parametric equivalent of a test of an interaction. Comparisons between the patient groups and the controls were analysed in a series of Mann–Whitney U tests.
The interviewer recorded patients' responses verbatim on the interview schedule, and 19 of the interviews were then rated for symptom severity by the second author who was blind to the interviewer's ratings. The inter-rater reliability for all delusions and hallucinations ratings was high (Pearson r = 0.89–1.00) and significant. Therefore, the interviewer's ratings were used in subsequent analyses. The patients were assigned to four groups based on the severity thresholds for the global rating of auditory hallucinations and the global rating of delusions of communication. There was perfect agreement between the interviewer and the independent assessor on the assignment to their respective groups of the 19 participants that they both assessed. 3.2. Interpretations of the clips The patient groups did not differ significantly on any of the categorical variables shown in Table 1, but there was a marginally significant trend for schizophrenia spectrum disorders to be more common in the groups with delusions of communication (χ2(3) = 6.51, P = 0.091). When the continuous variables were analysed in a 2 (delusions of communication; present vs. absent) × 2 (auditory hallucinations; present vs. absent) design, no significant effects were found for age or IQ. 3.2.1. Total errors The mean numbers of incorrect interpretations, of all kinds, for the four patient groups and the healthy controls are shown in Table 3. It can be seen that the great majority of interpretations of the clips that the patients selected were correct; 93.3% of gestures and 84.8% of incidental movements. Nevertheless, the four groups were found to differ significantly, by the Kruskall–Wallis test, χ2(3) = 8.96, P = 0.03, in the number of incorrect interpretations they made. Further analyses of the source of this difference, with Mann–Whitney U tests, showed that people with auditory hallucinations did not differ from those without hallucinations, Z = 0.008, P = 0.99, but there was a highly significant difference between people with delusions of communication and those without them, Z = 2.98, P = 0.003. People with delusions of communication made an average of 2.94 errors (14.7%, S.D. = 0.57) while people without these delusions made an average of only 0.94 errors (4.7%, S.D. = 0.65). Differences in the mean number of errors of all kinds made by the patient sample as a whole and the complete
32
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
Table 3 Characteristics and mean number (S.D.) of misinterpretations of video clips in three categories for four patient groups and healthy controls Delusions of communication Present
Controls
Absent
Auditory hallucinations
Characteristics N Schizophrenia spectrum (%) % male Mean age (years) Estimated IQ Misinterpretations Total incorrect Rejections IM as gestures
Present Absent
Present Absent
17 76
11 45
17 41
73 32.1 98.7
59 40.8 99.2
1.36 (1.91) 0.72 (0.65) 0.91 (0.94)
1.06 (2.19) 0.65 (1.06) 0.59 (1.00)
65 30.1 98.7 3.18 (4.38) 1.94 (2.97) 2.29 (2.92)
18 73 73 31.2 102.1 2.78 (3.12) 1.22 (1.44) 2.00 (2.70)
57 0 25 39.6 111.8 0.40 (0.62) 0.33 (0.51) 0.30 (0.46)
IM: incidental movements.
sample of healthy controls were analysed with the Mann–Whitney U test. The mean for the patients, 2.17 (S.D. = 3.21), was significantly higher than for the controls, 0.40 (S.D. = 0.62), Z = 4.03, P b 0.001. Differences in the number of errors between each of the patient groups and the controls were analysed with a series of Mann–Whitney U tests. These showed the greatest difference was for patients with delusions of communication, who made significantly more errors than the controls, Z = 4.71, P b 0.001, while the difference between patients without delusions of communication and controls was non-significant, Z = 1.81, P = 0.07. 3.2.2. Specific errors Differences on specific kinds of errors between the patient sample as a whole and the controls were also significant for interpretations of rejections, Z = 3.10, P = 0.002, and for misinterpretations of incidental movements as gestures, Z = 3.52, P b 0.001. Differences in the number of these errors between each of the patient groups and the controls were analysed with a series of Mann–Whitney U tests. The greatest differences were for patients with delusions of communication, who made significantly more interpretations of rejections, Z = 3.26, P = 0.001, and significantly more interpretations of incidental movements as gestures, Z = 3.73, P b 0.001. Patients without delusions of communication did not perceive more rejecting messages than controls, Z = 1.84, P = 0.07, though they did consider more of the incidental movements to be gestures, Z = 2.07, P = 0.04.
3.2.3. Matched control sample The healthy controls were significantly older than the patients, t(99.3) = 2.20, P = 0.03, higher in average estimated IQ, t (94.7) = 6.11, P b 0.001, and included a higher ratio of females, χ2(1) = 20.9, P b 0.001. In order to construct a subsample of control participants who were matched more closely to the patients, all 14 of the male controls were retained but only the nine females with the lowest IQs were kept. The gender ratio of the resulting subsample then no longer differed from the controls, χ2(1) = 0.06, P = 0.81, nor did the average age differ, t = 1.24, P = 0.23, but the mean IQ for the controls (mean: 108.0, S.D.: 8.65) was still significantly higher, t = 2.88, P = 0.005. When the analyses reported above were repeated with the subsample of 23 healthy controls who were matched to the patients, the results were substantially the same. The only result that would lead to a different conclusion was that the difference in interpretations of incidental movements as gestures, between the controls and patients who did not have delusions of communication, was reduced to marginal significance, Z = 1.89, P = 0.058. The matched controls had a significantly higher mean IQ than the patients, but the correlations between IQ and errors were all non-significant (Spearman's rho = − 0.08 to − 0.16) when these groups were combined. 4. Discussion In general, the control participants interpreted the meaning of brief video clips of human movements in line with what was intended. On average, there was 98% agreement (average of 0.4 interpretations ‘incorrect’). This provides strong support for the validity of the test. Among the patients also, the great majority of interpretations of the clips were correct (89%). However, this was significantly less than the healthy controls and is consistent with previous research showing poor comprehension of gestures among people with schizophrenia (Monti and Fingeret, 1987). It is also consistent the low sensitivity, compared with healthy controls, observed by Bucci (2006), in psychotic patients asked to distinguish gestures from incidental movements when video clips were obscured with visual noise. Few of the errors made by the patients consisted of selecting the interpretation of a different gesture from the one that was portrayed (mean = 0.13 errors). They were far more likely to misinterpret an incidental movement as a gesture, and they did this significantly more often than the controls. This is consistent with the ‘over-mentalising’ (Frith, 2004) that has been observed
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
when people with paranoia interpret animated sequences of triangles moving around a screen (Russell et al., 2006). However, these errors were mostly made by people with delusions of communication. This is consistent with the bias to perceive incidental movements as gestures that Bucci (2006) found among people with delusions of communication when video clips of movements were obscured with visual noise. Another kind of error made by the patients more often than the controls was to misinterpret movements as having a rejecting or insulting meaning, even though all the movements in the video clips were affectively neutral in content and the actor maintained a neutral expression. Again these errors were mostly made by people with delusions of communication. This is consistent with the interpretation that Bucci (2006) offered for the results of her signal detection analysis of the perception of gestures, that people with delusions of communication are impaired in self-monitoring for non-verbal material. That is, they appear to misattribute their self-generated internal events, especially self-disparaging preoccupations, to an external source. Interpretation of the findings should be made in light of three main methodological limitations. Firstly, according to chart diagnoses, schizophrenia spectrum disorders were more common among people with delusions of communication (though not significantly so). This is not surprising, given that diagnoses are partly based on these symptoms, but it does mean that group membership and diagnosis are confounded to some extent. Unfortunately, it was not possible to control this difference statistically because the distributions of error scores ruled out the use of parametric statistics. Thus we cannot be sure that the results for people with delusions of communication are not better explained by the presence of schizophrenia spectrum disorders. Secondly, the control participants were on average significantly older and higher in IQ than the patients and included a higher ratio of females. However, the results were not substantively altered when the patients were compared with controls matched in age and gender ratio, and IQ was found not to be correlated with the number of errors. Thirdly, since we included only patients with marked or severe symptoms, we cannot say what effects mild or questionable symptoms might have on interpretations of gestures. Difficulties in social cognition among people with schizophrenia are usually explained in terms of deficits in neurocognition, and certainly there is evidence to support this view (Sergi et al., 2007). For example, it has been found that abilities in affect recognition correlate with immediate memory and general intelligence (Pink-
33
ham and Penn, 2006), and with attentional dysfunction and span of apprehension (Nienow et al., 2006). Performance on the Half Profile of Non-verbal Sensitivity has been found to correlate with degree of sensorimotor gating (Wynn et al., 2005). Such results suggest that deficits in social cognition arise to some extent from difficulties in decoding incoming social information. However, the results of the present study, and those of the study by Bucci (2006), suggest that the difficulties that people with delusions of communication experience in understanding gestures can be explained, at least in part, by the misattribution of self-generated internal events to an external source. These misattributions might themselves be explained by neurocognitive deficits, such as poor intentional inhibition of currently irrelevant memories of previous non-verbal communications, and poor contextual memory, as suggested by Bucci (2006). Nevertheless, the route by which such deficits would have their effect on social cognition would be top-down rather than bottom-up. The results presented above suggest that the difficulties that people suffering acute psychotic episodes experience with interpreting gestures is largely confined to people with delusions of communication. This does not diminish the clinical significance, however, because delusions of reference are among the most common positive psychotic symptoms not only among psychotic inpatients (Minas et al., 1992), but also among outpatients with schizophrenia (World Health Organization, 1973). The proportion of stabilized outpatients who have delusions of communication is currently unknown, but the results of the present study suggest that about 55% of acutely psychotic inpatients suffer from them, though the sample is too small for confidence to be placed in this estimate. Gestures, of course, are only one form of non-verbal communication. Others include facial expression, prosody and gaze. It is possible that delusions of communication could also add to our understanding of the difficulties that people with psychotic disorders have in interpreting these forms of communication. For example, it might be that the tendency to misinterpret averted gazes as directed at the self, which is found among people with schizophrenia (Hooker and Park, 2005), is confined to people with delusions of communication. Further research will be required to clarify this. Acknowledgement We wish to acknowledge the support of the Hunter Medical Research Institute Volunteer Register for their assistance with recruitment of control participants.
34
S. Bucci et al. / Psychiatry Research 158 (2008) 27–34
References Abu-Akel, A., 1999. Impaired theory of mind in schizophrenia. Pragmatics & Cognition 7, 247–282. Addington, J., Saeedi, H., Addington, D., 2006. Influence of social perception and social knowledge on cognitive and social functioning in early psychosis. British Journal of Psychiatry 189, 373–378. Andreasen, N.C., 1984. Scale for the Assessment of Positive Symptoms (SAPS). University of Iowa, Iowa City. Bora, E., Eryavuz, A., Kayahan, B., Sungu, G., Veznedaroglu, B., 2006. Social functioning, theory of mind and neurocognition in outpatients with schizophrenia: mental state decoding may be a better predictor of social functioning than mental state reasoning. Psychiatry Research 145, 95–103. Bucci, S., 2006. Referential delusions of communication and reality monitoring deficits in psychosis. Unpublished DClinPsych thesis, University of Newcastle, Australia. Couture, S.M., Penn, D.L., Roberts, D.L., 2006. The functional significance of social cognition in schizophrenia: a review. Schizophrenia Bulletin 32 (Suppl. 1), S44–S63. Crawford, J.R., Besson, J.A.O., Bremner, M., Ebmeier, K.P., Cochrane, R.H.B., Kirkwood, K., 1992. Estimation of premorbid intelligence in schizophrenia. British Journal of Psychiatry 161, 69–74. Edwards, J., Jackson, H.J., Pattison, P.E., 2002. Emotion recognition via facial expression and affective prosody in schizophrenia: a methodological review. Clinical Psychology Review 22, 789–832. Frith, C.D., 1992. The Cognitive Neuropsychology of Schizophrenia. Lawrence Erlbaum, Hove (UK). Frith, C.D., 2004. Schizophrenia and theory of mind. Psychological Medicine 34, 385–389. Gelder, M., Gath, D., Mayou, R., 1989. Oxford Textbook of Psychiatry (2nd ed.). Oxford University Press, Oxford. Hooker, C., Park, S., 2005. You must be looking at me: the nature of gaze perception in schizophrenia patients. Cognitive Neuropsychiatry 10, 327–345. Kucharska-Pietura, K., David, A.S., Masiak, M., Phillips, M.L., 2005. Perception of facial and vocal affect by people with schizophrenia in early and late stages of illness. British Journal of Psychiatry 187, 523–528.
Minas, I.H., Stuart, G.W., Klimidis, S., Jackson, H.J., Singh, B.S., Copolov, D.L., 1992. Positive and negative symptoms in the psychoses: multidimensional scaling of SAPS and SANS items. Schizophrenia Research 8, 143–156. Monti, P.M., Fingeret, A.L., 1987. Social perception and communication skills among schizophrenics and nonschizophrenics. Journal of Clinical Psychology 43, 197–205. Nienow, T.M., Docherty, N.M., Cohen, A.S., Dinzeo, T.J., 2006. Attentional dysfunction, social perception, and social competence: what is the nature of the relationship? Journal of Abnormal Psychology 115, 408–417. Pinkham, A.E., Penn, D.L., 2006. Neurocognitive and social cognitive predictors of interpersonal skill in schizophrenia. Psychiatry Research 143, 167–178. Rosenthal, R., Benowitz, L.I., 1986. Sensitivity to nonverbal communications in normal, psychiatric, and brain damaged samples. In: Blanck, P.D., Buck, R., Rosenthal, R. (Eds.), Nonverbal Communication in the Clinical Context. Pennsylvania University Press, Pennsylvania. Russell, T.A., Reynaud, E., Herba, C., Morris, R., Corcoran, R., 2006. Do you see what I see? Interpretations of intentional movement in schizophrenia. Schizophrenia Research 81, 101–111. Sergi, M.J., Rassovsky, Y., Widmark, C., Reist, C., Erhart, S., Braff, D.L., Marder, S.R., Green, M.F., 2007. Social cognition in schizophrenia: relationships with neurocognition and negative symptoms. Schizophrenia Research 90 (1–3), 316–324. Startup, M., Startup, S., 2005. On two kinds of delusion of reference. Psychiatry Research 137, 87–92. Toomey, R., Schuldberg, D., Corrigan, P., Green, M.F., 2002. Nonverbal social perception and symptomatology in schizophrenia. Schizophrenia Research 53, 83–91. Wynn, J.K., Sergi, M.J., Dawson, M.E., Schell, A.M., Grenn, M.F., 2005. Sensorimotor gating, orienting and social perception in schizophrenia. Schizophrenia Research 73, 319–325. World Health Organization, 1973. Report of the International Pilot Study of Schizophrenia, vol. 1. World Health Organization, Geneva.