- Email: [email protected]
Impaired verbal source monitoring in schizophrenia: An intermediate trait vulnerability marker?
Schizophrenia Research 89 (2007) 287 – 292 www.elsevier.com/locate/schres
Impaired verbal source monitoring in schizophrenia: An intermediate trait vulnerability marker? Jérôme Brunelin ⁎, Thierry d'Amato, Philippe Brun, Benoit Bediou, Lassad Kallel, Muriel Senn, Emmanuel Poulet, Mohamed Saoud EA3092, University Lyon1, Service du Pr Daley, CH le Vinatier, IFR19, France Received 21 April 2006; received in revised form 23 August 2006; accepted 31 August 2006 Available online 9 October 2006
Abstract Patients with schizophrenia, particularly those with positive symptoms show impaired verbal source monitoring. Specific cognitive deficits have been observed during both active and remission phases of the illness as well as in groups of unaffected first degree relatives of patients with schizophrenia. This type of schizophrenia vulnerability marker may precede the onset of frank psychotic symptoms and contribute to their developments. The aim of this study was first to determine if unaffected siblings were impaired in discriminate internal vs. external generated events when compared to their remitted schizophrenics relatives and healthy subjects. Performances of healthy subjects were then compared with results from previous studies with acute hallucinating patients, acute non-hallucinating patients and patients with resistant auditory verbal hallucinations. Compared with healthy subjects, unaffected siblings are impaired (effect size, ES = 0.7), remitted or acute non-hallucinating patients are more impaired than siblings (ES = 1.4); patients with verbal auditory hallucinations (acute or resistant) are even more impaired than non-hallucinating patients (ES = 2.1). Our results suggest that a source monitoring deficit could be considered as an intermediate vulnerability marker of schizophrenia. © 2006 Elsevier B.V. All rights reserved. Keywords: Source monitoring; Schizophrenia; Hallucination; Vulnerability
1. Introduction Patients with schizophrenia, particularly those with specific positive symptoms such as hallucinations and/ or delusions, may have a deficit in monitoring the generation of their thought. This deficit, termed autonoetic agnosia (which means ‘the inability to identify self-generated mental events’; Keefe et al., 2002), may ⁎ Corresponding author. A3092, University Lyon1, Service du Pr Daley, CH le Vinatier, IFR19, 95 boulevard Pinel, 69677 Bron — France. E-mail address: [email protected] (J. Brunelin). 0920-9964/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2006.08.028
result in the conclusion that self-generated thoughts come from an external source which comes from a consequence of a malfunction in either a central monitor (Frith and Done, 1988; Hoffmann, 1986) or in a reality discrimination process (Bentall, 1990). This deficit can be evaluated by various source monitoring tasks, including tasks requiring discrimination between really heard and imagined-hearing words. These tasks developed by Johnson and Raye (see Johnson et al., 1993) require participants to distinguish between memories of their self-generated material and externally generated events. One frequently replicated finding is that compared with psychiatric controls and normals, patients
288
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
with schizophrenia are more likely to confuse externally and internally generated items, and to attribute selfgenerated items to someone else (Bentall et al., 1991; Keefe et al., 1999). Moreover, this deficit is exacerbated in acute patients with certain types of Schneiderian symptoms (i.e., AVH and delusions; defined as Thought insertion, voices arguing or commenting, made feelings, acts or impulses in Keefe et al., 2002) and particularly with auditory verbal hallucinations (AVH) (Garrett and Silva, 2003; Brunelin et al., 2006a). This deficit could also be improved by the pharmacological treatment of those symptoms (Keefe et al., 2003), or by a specific treatment of AVH which supports a specific relation between self recognition and AVH (Brunelin et al., 2006b). Recently, source monitoring deficit has been also observed in healthy individuals with high levels of psychotic-like experiences (Allen et al., 2006). Thus, this deficit could be considered as a trait marker of schizophrenia because it was detected in acute and remission phases of the illness (being exacerbated when AVH occurs) and in healthy volunteers with hallucinatory proneness and delusional ideation (Allen et al., 2006). Numerous other cognitive deficits have been observed regardless with the different phases of the illness (Asarnow and MacCrimmon, 1978; Nuechterlein and Dawson, 1984). In addition, groups of unaffected first degree relatives of probands with schizophrenia have also been found to be impaired neuropsychologically when compared with normal control groups (Faraone et al., 1995; Green et al., 1997; Saoud et al., 2000). Those deficits which could detect ill, stabilized or recovered patients with schizophrenia as well as unaffected individuals at high risk for schizophrenia were termed vulnerability markers (Spring and Zubin, 1978; Nuechterlein et al., 1994). The aim of this study was firstly, to evaluate source monitoring deficit as a trait vulnerability marker of schizophrenia. In this study, we investigate how unaffected siblings were impaired in discriminate internal vs. external generated events when compared to their remitted relatives patients and healthy controls (detect stabilized or recovered patients with schizophrenia as well as unaffected individuals at high risk for schizophrenia).Secondly we then compared these original results with those from previous studies (Brunelin et al., 2006a,b) to assess how AVH paralleled this deficit in acute patients (detect ill).
Recherche Biomédicale, Lyon B, 11/02/2004). All participants were native French speakers, and gave their written informed consent after a detailed description of the study. Fifteen participants with DSM IV diagnosis of schizophrenia, 15 of their unaffected first degree-relative and 15 healthy subjects were included. All patients were remitted patients with schizophrenia, with no hallucination symptoms or delusional ideation (PANSS positive mean± SEM: 12.8 ± 1.2), receiving antipsychotic drugs and were maintained on those drugs with neither change in dose nor any clinical change throughout 12 months. Unaffected first-degree relatives (siblings) were all brother or sister of included patients. Exclusion criteria for every participants were any major medical illness that could affect brain function (including hypertension and cardiac, endocrine, renal, and pulmonary diseases), current substance abuse, neurological conditions, recent history of head injury with loss of consciousness, low IQ (Raven PMC-T 36 b26; RAVEN, 1987) or verbal comprehension deficits (The Token test 36-items version b 28; Renzi and Faglioni, 1975) or an age of less than 18 or greater than 50 years. Further socio-demographical and clinical characteristics of the population are given in Table 1. 2.2. Source monitoring task
2. Methods
In source memory tasks, subjects had to distinguish between 8 silent imagined-hearing, 8 heard and 8 new non-presented words as described in a previous study (Brunelin et al., 2006a). Words were current French words extracted from a verbal fluency task with the same emotional valence and same length. During the test, words were presented for a duration of 3 s on a computer screen preceded by an instruction. Each word had its own instruction i.e. “Hear this word”, or “Imagine-hearing this word”. Immediately after the end of the test, a response grid including the 16 presented words plus 8 new non-presented words (‘distracters’; range 0–8) was given to the patient. For each word, subjects had to report if the word was externally generated or internally generated, or if the word had not appeared on the screen. In accordance with the literature, our judgment criterion to evaluate source monitoring performances was the total number of source memory confusions i.e. discrimination of self-generated words (range 0–16): misattributions of heard/imagined-hearing words (range 0–8) plus misattributions of imagined-hearing/heard (range 0–8) words.
2.1. Subjects
2.3. Analyses
This study was approved by a local ethical committee (Comité Consultatif de Protection des Personnes dans la
In order to compare performances of siblings, remitted patients with schizophrenia, and healthy
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
289
Table 1 Socio-demographical characteristics of included population (mean ± standard deviation)
n Age (years) Years of full time education PANSS Chlorpromazine equivalent (mg/day)
Healthy subjects
Siblings
Remitted patients
Acute HP (Brunelin et al., 2006a)
Acute NHP (Brunelin et al., 2006a)
Resistant HP (Brunelin et al., 2006b)
15 29.1 ± 7.3 12.5 ± 1.8
15 28.5 ± 7.5 11.6 ± 2.8
15 28.6 ± 7.5 11.1 ± 2.4
30 33.3 ± 8.4 10.8 ± 2.7
31 30.2 ± 9.0 10.9 ± 2.9
14 34.9 ± 8 10.6 ± 2
– –
– –
60.3 ± 19.6 629 ± 692
90.9 ± 15.9 510 ± 221
79.3 ± 17.2 563 ± 494
81.4 ± 11.4 614 ± 177
HP: hallucinating patients with schizophrenia; NHP: non-hallucinating patients with schizophrenia; PANSS: Positive and Negative Syndrome Scale for schizophrenia; Medication: in chlorpromazine equivalent (mg/day).
subjects, we used one way ANOVA followed by a 2tailed-T test. 2.4. Data from previous studies To investigate how hallucinations influence this cognitive deficit, using effect size calculation, we compared current results with those from our previous studies: - 30 acute patients with schizophrenia with AVH (HP); 31 acute non-hallucinating patients with schizophrenia (NHP); (Brunelin et al., 2006a). HP and NHP were significantly different for total PANSS scores (90.9 ± 15.9 in HP vs. 79.3 ± 17.2; t = 2.633; p = 0.01), but not for total PANSS score minus P3 item (hallucinatory behavior) PANSS score (85.9 ± 15.7 vs. 77.7 ± 17.1; t = 1.893; ns). - 14 patients with schizophrenia with resistant AVH before and after repeated Transcranial Magnetic Stimulation (rTMS) treatment (Brunelin et al., 2006b); a new therapeutic method which can specifically curtail refractory auditory hallucinations (Hoffman et al., 2003; Poulet et al., 2005). Further characteristics of including patients were given in Table 1 and in corresponding articles. Irrespective of the studies, all participants performed the same source monitoring task with the same words, in the same condition of testing, with the same experimenter (except for resistant hallucinating patients who randomly performed 2 different tasks requiring discrimination between silent- and overt-reading words). To compare performances of siblings, remitted patients with schizophrenia, then performances of acute patients (HP and NHP) and patients with resistant auditory hallucinations (before and after rTMS treatment) with performances of healthy subjects, we calcu-
lated the effect size of performances. Effect sizes were calculated as: [(mean of experimental group performances) minus (mean of control group performances)] divided by standard deviation of control group. The effect size estimate is slightly biased and is therefore corrected using a factor provided by Hedges and Olkin (1985). An effect size is exactly equivalent to a ‘Z-score’ of a standard Normal distribution (Fig. 1). As suggested by Cohen (1969), an effect size of 0.2 could be considered as small, 0.5 as medium and 0.8 as large. 3. Results 3.1. Siblings impairment, original data Performances of healthy volunteers on a source monitoring task measured by the total number of source inversions (i.e. confusions between heard and imaginedhearing words) was 1.06 ± 1.17 (mean ± standard deviation). Concerning the 2 experimental groups, remitted patients with schizophrenia, performances were 3.0 ± 1.73 and 2.0 ± 1.93 for their unaffected first-degree relative non-psychotic subjects. We observed significant differences between the 3 groups (ANOVA, F(2,42) = 2.736, p = 0.01). Differences were significant between remitted patients and controls, ( p b 0.0001) and between siblings and controls ( p = 0.02), but not between siblings and remitted patients ( p = 0.14). In recognition of distractors, the 3 groups performed equally: 7.5 ± 0.9 for healthy subjects; 7.8 ± 0.3 for siblings and 7.5 ± 0.6 for remitted patients with schizophrenia. (F(2,42) = 1.507, p = 0.23; ns) 3.2. Effect size calculation, results from previous study The biggest effect sizes (N2) were observed in HP (resistant or acute). A large effect size (between 1 and 1.5) was observed in NHP (acute or remitted or resistant
290
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
Fig. 1. Estimates of the size [(mean of experimental group performances) minus (mean of control group performances)] divided by standard deviation of control group and then corrected, of the difference between 15 unaffected siblings, 15 remitted patients, 30 acute hallucinating patients (HP), 31 non-hallucinating patients (NHP) and 14 resistant HP before and after rTMS treatment and control groups (n = 15).
after rTMS). A medium effect size was observed in firstrelative non-psychotic subjects (0.7). 4. Discussion First, compared to healthy subjects and remitted patients with schizophrenia, we found that unaffected siblings of patients had significant intermediate performances in recognizing externally vs. internally generated events. This result corroborates previous studies using other neuropsychological tasks which have reported similar intermediate performances in siblings (Faraone et al., 1995; Green et al., 1997; Keefe et al., 1997; Saoud et al., 2000). Given the evidence of a genetically transmitted predisposition to schizophrenia (Gottesman and
Shields, 1982; Gottesman and Bertelsen, 1989), we can hypothesize that the siblings’ poor performances concerning this task could be related to genetic factors. However, it is not possible to actually affirm that the association is solely due to genetic mechanisms because this study neither allowed the definite distinction of genetic transmission from common familial environment nor took into account clinical traits or dimensions of psychosis proneness. This deficit could also be considered as a trait vulnerability marker, or an actual phenotype of schizophrenia. Some methodological issues could account for these observations. It has been reported that aging could influence source monitoring performance (Henkel et al., 1998); however in our sample all the participants had quite similar ages. The broad range of cognitive deficits found in patients with schizophrenia could also partially explain our observed differences. Indeed, it has been suggested that, compared to non-hallucinating, deficits in source monitoring in subjects with auditory hallucinations may be due to differences in IQ and verbal memory (Seal et al., 1997). However, this relationship was not observed in patients with schizophrenia with IQs in the normal range (Vinogradov et al., 1997). Moreover, in a study of schizophrenia patients monitoring their own drawings, patients performed significantly worse than controls even when IQ, visual memory and vigilance performance were used as covariates (Stirling et al., 1998). In our study, there was no difference between groups for IQ, verbal comprehension and recognition of the distractors. Even if a ceiling effect could partially explain that all the subjects performed equally in the recognition of ‘distractors’, one can hypothesize that all participants have understood the task and performed it with the same motivation. Despite the fact that auditory hallucinations are an immediate, ongoing experience rather than a memory, the source monitoring framework merits an exploratory application to psychotic phenomena, because of its potential value in other clinical areas. A variety of information embedded in a memory at the time it occurs later serves as the basis of judgements about the origin of the memory (Johnson et al., 1993). Secondly, as suggested by the literature (Keefe et al., 1999, 2002), in post-hoc comparison, we found that both HP and NHP, acute, resistant or remitted, present a deficit in recognizing their thoughts and real heard events when compared to the performance of healthy subjects (Fig. 1). As previously reported (Keefe et al., 1999), we found that NHP were impaired in source monitoring tasks (effect size 1.4). This deficit is exacerbated both in acute and resistant HP (effect size N 2). When hallucinations are curtailed by low frequency rTMS treatment in
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
resistant HP (Brunelin et al., 2006b) performances of HP return to the level of NHP performances (effect size 1.4). It has already been shown that antipsychotic medication in acute patients with psychotic symptoms could improve source monitoring task performance (Keefe et al., 2003). Compared to healthy subjects, we reported significant lower performances in source monitoring tasks in unaffected siblings (effect size 0.7). Many investigators conceptualise hallucinations and other ‘psychosislike’ experiences as forming a continuum with normal psychological functioning (Verdoux and van Os, 2002). This is consistent with the observation that in addition to patients with clinically recognized psychotic disorders, a significant minority of the general population also have hallucinatory experiences or proneness (Tien, 1991). Deficit in source monitoring could be detected during both acute and remitted phases of the illness and is present in unaffected first degree relatives of probands with schizophrenia and is present in healthy subjects with a high score in hallucinatory proneness and delusional ideation scale (Allen et al., 2006; Laroi et al., 2005). Thus, this deficit is present in patients with schizophrenia with or without AVH although it may be exacerbated when hallucinatory symptoms are present. We might rather consider it as an intermediate vulnerability marker of schizophrenia (Nuechterlein et al., 1994), being both a trait and state marker; a trait marker of schizophrenia or of auditory hallucination predisposition and a state marker of auditory hallucinations, depending on the level of the disturbance. Even if, a specific relationship could be established between AVH and source monitoring deficit (Garrett and Silva, 2003), numerous other studies exist documenting the relationship between source monitoring deficit and certain types of delusions (Moritz et al., 2005; Brebion et al., 2000) or formal thought disorder (Nienow and Docherty, 2004). Further studies are needed to understand the influence of those symptoms on source monitoring deficit in schizophrenia. These results suggest that defective verbal self-monitoring may precede the onset of frank psychotic symptoms and contribute to their developments both in the general population with hallucinatory proneness and delusional ideation (Allen et al., 2006) and in high risk genetic vulnerable subjects. Thus, according to Nuechterlein and Dawson's (1984) definition, source monitoring deficit could be considered as intermediate marker of vulnerability. Acknowledgments The authors thank Dr S. Marsella, Dr M. Combris, and Dr S Higgins for their contributions. This study
291
was supported by a grant from CH Le Viantier (CSR 2003). References Allen, P., Freeman, D., Johns, L., McGuire, P., 2006. Misattribution of self-generated speech in relation to hallucinatory proneness and delusional ideation in healthy volunteers. Schizophr. Res. 84 (2–3), 281–288. Asarnow, R.F., MacCrimmon, D.J., 1978. Residual performance deficit in clinically remitted schizophrenia: a marker of schizophrenia? J. Abnorm. Psychology 87, 597–608. Bentall, R.P., 1990. The illusion of reality: a psychological model of hallucinations. Psychol. Bull. 107, 82–95. Bentall, R.P., Baker, G.A., Havers, S., 1991. Reality monitoring and psychotic hallucinations. Br. J. Clin. Psychol. 30, 213–222. Brebion, G., Amador, X., David, A., Malaspina, D., Sharif, Z., Gorman, J.M., 2000. Positive symptomatology and sourcemonitoring failure in schizophrenia—an analysis of symptomspecific effects. Psychiatry Res. 21;95 (2), 119–131. Brunelin, J., Combris, M., Poulet, E., Kallel, L., d'Amato, T., Dalery, J., Saoud, M., 2006a. Source monitoring deficits in hallucinating compared to non-hallucinating patients with schizophrenia. Eur. Psychiatr. 21 (4), 259–261. Brunelin, J., Poulet, E., Bediou, B., Kallel, L., Dalery, J., d'Amato, T., Saoud, M., 2006b. Low frequency Repetitive Transcranial Magnetic Stimulation improves source monitoring deficit in hallucinating patients with schizophrenia. Schizophr. Res. 81, 41–45. Seal, M.L., Crowe, S.F., Cheung, P., 1997. Deficits in source monitoring in subjects with auditory hallucinations may be due to differences in verbal intelligence and verbal memory. Cogn. Neuropsychiatry 2 (4), 273–290 (18). Cohen, J., 1969. Statistical Power Analysis for the Behavioral Sciences. Academic Press, NY. Faraone, S.V., Seidman, L.J., Kremen, W.S., Pepple, J.R., Lyons, M.J., Tsuang, M.T., 1995. Neuropsychological functioning among the nonpsychotic relatives of schizophrenic patients: a diagnostic efficiency analysis. J. Abnorm. Psychology 104, 286–304. Frith, C.D., Done, C.J., 1988. Towards a neuropsychology of schizophrenia. Br. J. Psychiatry 153, 437–443. Garrett, M., Silva, R., 2003. Auditory hallucinations, source monitoring, and the belief that “voices” are real. Schizophr Bull. 29 (3), 445–457. Gottesman, I.I., Bertelsen, A., 1989. Confirming unexpressed genotypes for schizophrenia. Arch. Gen. Psychiatry 46, 867–872. Gottesman, I.I., Shields, J., 1982. The epigenetic puzzle. In: Gottesman, I.I., Shields, J. (Eds.), Cambridge University Press, Cambridge. Green, M.F., Nuechterlein, K.H., Breitmeyer, B., 1997. Backward masking performance in unaffected siblings of schizophrenics patients. Arch. Gen. Psychiatry 54, 465–472. Hedges, L., Olkin, I., 1985. Statistical Methods for Meta-Analysis. Academic Press, New York, p. 80. Henkel, L.A., Johnson, M.K., De Leonardis, D.M., 1998. Aging and source monitoring: cognitive processes and neuropsychological correlates. J. Exp. Psychol. Gen. 127 (3), 251–268. Hoffmann, R.E., 1986. Verbal hallucinations and language production processes in schizophrenia. Behav. Brain Sci. 9 (3), 503–548. Hoffman, R.E., Hawkins, K.A., Gueorguieva, R., Boutros, N.N., Rachid, F., Carroll, K., Krystal, J.H., 2003. Transcranial magnetic stimulation of left temporoparietal cortex and medication-resistant auditory hallucinations. Arch. Gen. Psychiatry 60, 49–56.
292
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
Johnson, M.K., Hashtroudi, S., Lindsay, D.S., 1993. Source monitoring. Psychol. Bull. 114 (1), 3–28. Keefe, R.S.E., Silverman, J.M., Mohs, R.C., Siever, L.J., Harvey, P.D., Friedman, L., Roitman, S.E.L., DuPre, R.L., Smith, C.J., Schmeidler, J., Davis, K.L., 1997. Eye tracking, attention, and schizotypal symptoms in nonpsychotic relatives of patients with schizophrenia. Arch. Gen. Psychiatry 54, 169–176. Keefe, R.S., Arnold, M.C., Bayen, U.J., Harvey, P.D., 1999. Source monitoring deficits in patients with schizophrenia; a multinomial modelling analysis. Psychol. Med. 29, 903–914. Keefe, R.S., Arnold, M.C., Bayen, U.J., McEvoy, J.P, Wilson, W.H., 2002. Source-monitoring deficits for self-generated stimuli in schizophrenia: multinomial modelling of data from three sources. Schizophr. Res. 57 (1), 51–67. Keefe, R.S., Poe, M.P., McEvoy, J.P., Vaughan, A., 2003. Source monitoring improvement in patients with schizophrenia receiving antipsychotic medications. Psychopharmacology (Berl) 169 (3–4), 383–389. Laroi, F., Collignon, O., Van der Linden, M., 2005. Source monitoring for actions in hallucination proneness. Cogn. Neuropsychiatry 10 (2), 105–123. Moritz, S., Woodward, T.S., Whitman, J.C., Cuttler, C., 2005. Confidence in errors as a possible basis for delusions in schizophrenia. J. Nerv. Ment. Dis. 193 (1), 9–16. Nienow, T.M., Docherty, N.M., 2004. Internal source monitoring and thought disorder in schizophrenia. J. Nerv. Ment. Dis. 192 (10), 696–700. Nuechterlein, K.H., Dawson, M.E., 1984. Information processing dysfunctions in the developmental course of schizophrenic disorders. Schizophr. Bull. 10, 160–203. Nuechterlein, K.H., Dawson, M.E., Green, M.F., 1994. Informationprocessing abnormalities as neuropsychological vulnerability indicators for schizophrenia. Acta Psychiatr. Scand. 90 (suppl 384), 71–79.
Poulet, E., Brunelin, J., Bediou, B., Bation, R., Forgeard, L., Dalery, J., d'amato, T., Saoud, M., 2005. Slow transcranial magnetic stimulation can rapidly reduce resistant auditory hallucinations in schizophrenia. Biol. Psychiatry 57 (2), 188–191. RAVEN, 1987. Raven coloured progressive matrices. Issy-lesMoulineaux: Editions scientifiques et psychotechniques. Renzi, E.D., Faglioni, P., 1975. Verbal comprehension in aphasic and in normal subjects with a shortened version of the token test (author's transl). Riv. Patol. Nerv. Ment. 96 (4), 252–269. Saoud, M., d'Amato, T., Gutknecht, C., Triboulet, P., Bertaud, J.P., Marie-Cardine, M., Dalery, J., Rochet, T., 2000. Neuropsychological deficit in siblings discordant for schizophrenia. Schizophr. Bull. 26 (4), 893–902. Spring, B., Zubin, J., 1978. Attention and information-processing as indicators of vulnerability to schizophrenic episodes. In: Wynne, L.C., Cromwell, R.L., Matthysse, S. (Eds.), The nature of schizophrenia: approaches to research and treatment. Wiley, New York, NY, pp. 366–375. Stirling, J.D., Hellewell, J.S., Quraishi, N., 1998. Self-monitoring dysfunction and the schizophrenic symptoms of alien control. Psychol. Med. 28 (3), 675–683. Tien, A.Y., 1991. Distributions of hallucinations in the population. Soc. Psychiatry Psychiatr. Epidemiol. 26 (6), 287–292. Verdoux, H., van Os, J., 2002. Psychotic symptoms in non-clinical populations and the continuum of psychosis. Schizophr. Res. 54 (1–2), 59–65. Vinogradov, S., Willis-Shore, J., Poole, J.H., Marten, E., Ober, B.A., Shenaut, G.K., 1997. Clinical and neurocognitive aspects of source monitoring errors in schizophrenia. Am. J. Psychiatry 154 (11), 1530–1537.
Impaired verbal source monitoring in schizophrenia: An intermediate trait vulnerability marker? Jérôme Brunelin ⁎, Thierry d'Amato, Philippe Brun, Benoit Bediou, Lassad Kallel, Muriel Senn, Emmanuel Poulet, Mohamed Saoud EA3092, University Lyon1, Service du Pr Daley, CH le Vinatier, IFR19, France Received 21 April 2006; received in revised form 23 August 2006; accepted 31 August 2006 Available online 9 October 2006
Abstract Patients with schizophrenia, particularly those with positive symptoms show impaired verbal source monitoring. Specific cognitive deficits have been observed during both active and remission phases of the illness as well as in groups of unaffected first degree relatives of patients with schizophrenia. This type of schizophrenia vulnerability marker may precede the onset of frank psychotic symptoms and contribute to their developments. The aim of this study was first to determine if unaffected siblings were impaired in discriminate internal vs. external generated events when compared to their remitted schizophrenics relatives and healthy subjects. Performances of healthy subjects were then compared with results from previous studies with acute hallucinating patients, acute non-hallucinating patients and patients with resistant auditory verbal hallucinations. Compared with healthy subjects, unaffected siblings are impaired (effect size, ES = 0.7), remitted or acute non-hallucinating patients are more impaired than siblings (ES = 1.4); patients with verbal auditory hallucinations (acute or resistant) are even more impaired than non-hallucinating patients (ES = 2.1). Our results suggest that a source monitoring deficit could be considered as an intermediate vulnerability marker of schizophrenia. © 2006 Elsevier B.V. All rights reserved. Keywords: Source monitoring; Schizophrenia; Hallucination; Vulnerability
1. Introduction Patients with schizophrenia, particularly those with specific positive symptoms such as hallucinations and/ or delusions, may have a deficit in monitoring the generation of their thought. This deficit, termed autonoetic agnosia (which means ‘the inability to identify self-generated mental events’; Keefe et al., 2002), may ⁎ Corresponding author. A3092, University Lyon1, Service du Pr Daley, CH le Vinatier, IFR19, 95 boulevard Pinel, 69677 Bron — France. E-mail address: [email protected] (J. Brunelin). 0920-9964/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2006.08.028
result in the conclusion that self-generated thoughts come from an external source which comes from a consequence of a malfunction in either a central monitor (Frith and Done, 1988; Hoffmann, 1986) or in a reality discrimination process (Bentall, 1990). This deficit can be evaluated by various source monitoring tasks, including tasks requiring discrimination between really heard and imagined-hearing words. These tasks developed by Johnson and Raye (see Johnson et al., 1993) require participants to distinguish between memories of their self-generated material and externally generated events. One frequently replicated finding is that compared with psychiatric controls and normals, patients
288
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
with schizophrenia are more likely to confuse externally and internally generated items, and to attribute selfgenerated items to someone else (Bentall et al., 1991; Keefe et al., 1999). Moreover, this deficit is exacerbated in acute patients with certain types of Schneiderian symptoms (i.e., AVH and delusions; defined as Thought insertion, voices arguing or commenting, made feelings, acts or impulses in Keefe et al., 2002) and particularly with auditory verbal hallucinations (AVH) (Garrett and Silva, 2003; Brunelin et al., 2006a). This deficit could also be improved by the pharmacological treatment of those symptoms (Keefe et al., 2003), or by a specific treatment of AVH which supports a specific relation between self recognition and AVH (Brunelin et al., 2006b). Recently, source monitoring deficit has been also observed in healthy individuals with high levels of psychotic-like experiences (Allen et al., 2006). Thus, this deficit could be considered as a trait marker of schizophrenia because it was detected in acute and remission phases of the illness (being exacerbated when AVH occurs) and in healthy volunteers with hallucinatory proneness and delusional ideation (Allen et al., 2006). Numerous other cognitive deficits have been observed regardless with the different phases of the illness (Asarnow and MacCrimmon, 1978; Nuechterlein and Dawson, 1984). In addition, groups of unaffected first degree relatives of probands with schizophrenia have also been found to be impaired neuropsychologically when compared with normal control groups (Faraone et al., 1995; Green et al., 1997; Saoud et al., 2000). Those deficits which could detect ill, stabilized or recovered patients with schizophrenia as well as unaffected individuals at high risk for schizophrenia were termed vulnerability markers (Spring and Zubin, 1978; Nuechterlein et al., 1994). The aim of this study was firstly, to evaluate source monitoring deficit as a trait vulnerability marker of schizophrenia. In this study, we investigate how unaffected siblings were impaired in discriminate internal vs. external generated events when compared to their remitted relatives patients and healthy controls (detect stabilized or recovered patients with schizophrenia as well as unaffected individuals at high risk for schizophrenia).Secondly we then compared these original results with those from previous studies (Brunelin et al., 2006a,b) to assess how AVH paralleled this deficit in acute patients (detect ill).
Recherche Biomédicale, Lyon B, 11/02/2004). All participants were native French speakers, and gave their written informed consent after a detailed description of the study. Fifteen participants with DSM IV diagnosis of schizophrenia, 15 of their unaffected first degree-relative and 15 healthy subjects were included. All patients were remitted patients with schizophrenia, with no hallucination symptoms or delusional ideation (PANSS positive mean± SEM: 12.8 ± 1.2), receiving antipsychotic drugs and were maintained on those drugs with neither change in dose nor any clinical change throughout 12 months. Unaffected first-degree relatives (siblings) were all brother or sister of included patients. Exclusion criteria for every participants were any major medical illness that could affect brain function (including hypertension and cardiac, endocrine, renal, and pulmonary diseases), current substance abuse, neurological conditions, recent history of head injury with loss of consciousness, low IQ (Raven PMC-T 36 b26; RAVEN, 1987) or verbal comprehension deficits (The Token test 36-items version b 28; Renzi and Faglioni, 1975) or an age of less than 18 or greater than 50 years. Further socio-demographical and clinical characteristics of the population are given in Table 1. 2.2. Source monitoring task
2. Methods
In source memory tasks, subjects had to distinguish between 8 silent imagined-hearing, 8 heard and 8 new non-presented words as described in a previous study (Brunelin et al., 2006a). Words were current French words extracted from a verbal fluency task with the same emotional valence and same length. During the test, words were presented for a duration of 3 s on a computer screen preceded by an instruction. Each word had its own instruction i.e. “Hear this word”, or “Imagine-hearing this word”. Immediately after the end of the test, a response grid including the 16 presented words plus 8 new non-presented words (‘distracters’; range 0–8) was given to the patient. For each word, subjects had to report if the word was externally generated or internally generated, or if the word had not appeared on the screen. In accordance with the literature, our judgment criterion to evaluate source monitoring performances was the total number of source memory confusions i.e. discrimination of self-generated words (range 0–16): misattributions of heard/imagined-hearing words (range 0–8) plus misattributions of imagined-hearing/heard (range 0–8) words.
2.1. Subjects
2.3. Analyses
This study was approved by a local ethical committee (Comité Consultatif de Protection des Personnes dans la
In order to compare performances of siblings, remitted patients with schizophrenia, and healthy
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
289
Table 1 Socio-demographical characteristics of included population (mean ± standard deviation)
n Age (years) Years of full time education PANSS Chlorpromazine equivalent (mg/day)
Healthy subjects
Siblings
Remitted patients
Acute HP (Brunelin et al., 2006a)
Acute NHP (Brunelin et al., 2006a)
Resistant HP (Brunelin et al., 2006b)
15 29.1 ± 7.3 12.5 ± 1.8
15 28.5 ± 7.5 11.6 ± 2.8
15 28.6 ± 7.5 11.1 ± 2.4
30 33.3 ± 8.4 10.8 ± 2.7
31 30.2 ± 9.0 10.9 ± 2.9
14 34.9 ± 8 10.6 ± 2
– –
– –
60.3 ± 19.6 629 ± 692
90.9 ± 15.9 510 ± 221
79.3 ± 17.2 563 ± 494
81.4 ± 11.4 614 ± 177
HP: hallucinating patients with schizophrenia; NHP: non-hallucinating patients with schizophrenia; PANSS: Positive and Negative Syndrome Scale for schizophrenia; Medication: in chlorpromazine equivalent (mg/day).
subjects, we used one way ANOVA followed by a 2tailed-T test. 2.4. Data from previous studies To investigate how hallucinations influence this cognitive deficit, using effect size calculation, we compared current results with those from our previous studies: - 30 acute patients with schizophrenia with AVH (HP); 31 acute non-hallucinating patients with schizophrenia (NHP); (Brunelin et al., 2006a). HP and NHP were significantly different for total PANSS scores (90.9 ± 15.9 in HP vs. 79.3 ± 17.2; t = 2.633; p = 0.01), but not for total PANSS score minus P3 item (hallucinatory behavior) PANSS score (85.9 ± 15.7 vs. 77.7 ± 17.1; t = 1.893; ns). - 14 patients with schizophrenia with resistant AVH before and after repeated Transcranial Magnetic Stimulation (rTMS) treatment (Brunelin et al., 2006b); a new therapeutic method which can specifically curtail refractory auditory hallucinations (Hoffman et al., 2003; Poulet et al., 2005). Further characteristics of including patients were given in Table 1 and in corresponding articles. Irrespective of the studies, all participants performed the same source monitoring task with the same words, in the same condition of testing, with the same experimenter (except for resistant hallucinating patients who randomly performed 2 different tasks requiring discrimination between silent- and overt-reading words). To compare performances of siblings, remitted patients with schizophrenia, then performances of acute patients (HP and NHP) and patients with resistant auditory hallucinations (before and after rTMS treatment) with performances of healthy subjects, we calcu-
lated the effect size of performances. Effect sizes were calculated as: [(mean of experimental group performances) minus (mean of control group performances)] divided by standard deviation of control group. The effect size estimate is slightly biased and is therefore corrected using a factor provided by Hedges and Olkin (1985). An effect size is exactly equivalent to a ‘Z-score’ of a standard Normal distribution (Fig. 1). As suggested by Cohen (1969), an effect size of 0.2 could be considered as small, 0.5 as medium and 0.8 as large. 3. Results 3.1. Siblings impairment, original data Performances of healthy volunteers on a source monitoring task measured by the total number of source inversions (i.e. confusions between heard and imaginedhearing words) was 1.06 ± 1.17 (mean ± standard deviation). Concerning the 2 experimental groups, remitted patients with schizophrenia, performances were 3.0 ± 1.73 and 2.0 ± 1.93 for their unaffected first-degree relative non-psychotic subjects. We observed significant differences between the 3 groups (ANOVA, F(2,42) = 2.736, p = 0.01). Differences were significant between remitted patients and controls, ( p b 0.0001) and between siblings and controls ( p = 0.02), but not between siblings and remitted patients ( p = 0.14). In recognition of distractors, the 3 groups performed equally: 7.5 ± 0.9 for healthy subjects; 7.8 ± 0.3 for siblings and 7.5 ± 0.6 for remitted patients with schizophrenia. (F(2,42) = 1.507, p = 0.23; ns) 3.2. Effect size calculation, results from previous study The biggest effect sizes (N2) were observed in HP (resistant or acute). A large effect size (between 1 and 1.5) was observed in NHP (acute or remitted or resistant
290
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
Fig. 1. Estimates of the size [(mean of experimental group performances) minus (mean of control group performances)] divided by standard deviation of control group and then corrected, of the difference between 15 unaffected siblings, 15 remitted patients, 30 acute hallucinating patients (HP), 31 non-hallucinating patients (NHP) and 14 resistant HP before and after rTMS treatment and control groups (n = 15).
after rTMS). A medium effect size was observed in firstrelative non-psychotic subjects (0.7). 4. Discussion First, compared to healthy subjects and remitted patients with schizophrenia, we found that unaffected siblings of patients had significant intermediate performances in recognizing externally vs. internally generated events. This result corroborates previous studies using other neuropsychological tasks which have reported similar intermediate performances in siblings (Faraone et al., 1995; Green et al., 1997; Keefe et al., 1997; Saoud et al., 2000). Given the evidence of a genetically transmitted predisposition to schizophrenia (Gottesman and
Shields, 1982; Gottesman and Bertelsen, 1989), we can hypothesize that the siblings’ poor performances concerning this task could be related to genetic factors. However, it is not possible to actually affirm that the association is solely due to genetic mechanisms because this study neither allowed the definite distinction of genetic transmission from common familial environment nor took into account clinical traits or dimensions of psychosis proneness. This deficit could also be considered as a trait vulnerability marker, or an actual phenotype of schizophrenia. Some methodological issues could account for these observations. It has been reported that aging could influence source monitoring performance (Henkel et al., 1998); however in our sample all the participants had quite similar ages. The broad range of cognitive deficits found in patients with schizophrenia could also partially explain our observed differences. Indeed, it has been suggested that, compared to non-hallucinating, deficits in source monitoring in subjects with auditory hallucinations may be due to differences in IQ and verbal memory (Seal et al., 1997). However, this relationship was not observed in patients with schizophrenia with IQs in the normal range (Vinogradov et al., 1997). Moreover, in a study of schizophrenia patients monitoring their own drawings, patients performed significantly worse than controls even when IQ, visual memory and vigilance performance were used as covariates (Stirling et al., 1998). In our study, there was no difference between groups for IQ, verbal comprehension and recognition of the distractors. Even if a ceiling effect could partially explain that all the subjects performed equally in the recognition of ‘distractors’, one can hypothesize that all participants have understood the task and performed it with the same motivation. Despite the fact that auditory hallucinations are an immediate, ongoing experience rather than a memory, the source monitoring framework merits an exploratory application to psychotic phenomena, because of its potential value in other clinical areas. A variety of information embedded in a memory at the time it occurs later serves as the basis of judgements about the origin of the memory (Johnson et al., 1993). Secondly, as suggested by the literature (Keefe et al., 1999, 2002), in post-hoc comparison, we found that both HP and NHP, acute, resistant or remitted, present a deficit in recognizing their thoughts and real heard events when compared to the performance of healthy subjects (Fig. 1). As previously reported (Keefe et al., 1999), we found that NHP were impaired in source monitoring tasks (effect size 1.4). This deficit is exacerbated both in acute and resistant HP (effect size N 2). When hallucinations are curtailed by low frequency rTMS treatment in
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
resistant HP (Brunelin et al., 2006b) performances of HP return to the level of NHP performances (effect size 1.4). It has already been shown that antipsychotic medication in acute patients with psychotic symptoms could improve source monitoring task performance (Keefe et al., 2003). Compared to healthy subjects, we reported significant lower performances in source monitoring tasks in unaffected siblings (effect size 0.7). Many investigators conceptualise hallucinations and other ‘psychosislike’ experiences as forming a continuum with normal psychological functioning (Verdoux and van Os, 2002). This is consistent with the observation that in addition to patients with clinically recognized psychotic disorders, a significant minority of the general population also have hallucinatory experiences or proneness (Tien, 1991). Deficit in source monitoring could be detected during both acute and remitted phases of the illness and is present in unaffected first degree relatives of probands with schizophrenia and is present in healthy subjects with a high score in hallucinatory proneness and delusional ideation scale (Allen et al., 2006; Laroi et al., 2005). Thus, this deficit is present in patients with schizophrenia with or without AVH although it may be exacerbated when hallucinatory symptoms are present. We might rather consider it as an intermediate vulnerability marker of schizophrenia (Nuechterlein et al., 1994), being both a trait and state marker; a trait marker of schizophrenia or of auditory hallucination predisposition and a state marker of auditory hallucinations, depending on the level of the disturbance. Even if, a specific relationship could be established between AVH and source monitoring deficit (Garrett and Silva, 2003), numerous other studies exist documenting the relationship between source monitoring deficit and certain types of delusions (Moritz et al., 2005; Brebion et al., 2000) or formal thought disorder (Nienow and Docherty, 2004). Further studies are needed to understand the influence of those symptoms on source monitoring deficit in schizophrenia. These results suggest that defective verbal self-monitoring may precede the onset of frank psychotic symptoms and contribute to their developments both in the general population with hallucinatory proneness and delusional ideation (Allen et al., 2006) and in high risk genetic vulnerable subjects. Thus, according to Nuechterlein and Dawson's (1984) definition, source monitoring deficit could be considered as intermediate marker of vulnerability. Acknowledgments The authors thank Dr S. Marsella, Dr M. Combris, and Dr S Higgins for their contributions. This study
291
was supported by a grant from CH Le Viantier (CSR 2003). References Allen, P., Freeman, D., Johns, L., McGuire, P., 2006. Misattribution of self-generated speech in relation to hallucinatory proneness and delusional ideation in healthy volunteers. Schizophr. Res. 84 (2–3), 281–288. Asarnow, R.F., MacCrimmon, D.J., 1978. Residual performance deficit in clinically remitted schizophrenia: a marker of schizophrenia? J. Abnorm. Psychology 87, 597–608. Bentall, R.P., 1990. The illusion of reality: a psychological model of hallucinations. Psychol. Bull. 107, 82–95. Bentall, R.P., Baker, G.A., Havers, S., 1991. Reality monitoring and psychotic hallucinations. Br. J. Clin. Psychol. 30, 213–222. Brebion, G., Amador, X., David, A., Malaspina, D., Sharif, Z., Gorman, J.M., 2000. Positive symptomatology and sourcemonitoring failure in schizophrenia—an analysis of symptomspecific effects. Psychiatry Res. 21;95 (2), 119–131. Brunelin, J., Combris, M., Poulet, E., Kallel, L., d'Amato, T., Dalery, J., Saoud, M., 2006a. Source monitoring deficits in hallucinating compared to non-hallucinating patients with schizophrenia. Eur. Psychiatr. 21 (4), 259–261. Brunelin, J., Poulet, E., Bediou, B., Kallel, L., Dalery, J., d'Amato, T., Saoud, M., 2006b. Low frequency Repetitive Transcranial Magnetic Stimulation improves source monitoring deficit in hallucinating patients with schizophrenia. Schizophr. Res. 81, 41–45. Seal, M.L., Crowe, S.F., Cheung, P., 1997. Deficits in source monitoring in subjects with auditory hallucinations may be due to differences in verbal intelligence and verbal memory. Cogn. Neuropsychiatry 2 (4), 273–290 (18). Cohen, J., 1969. Statistical Power Analysis for the Behavioral Sciences. Academic Press, NY. Faraone, S.V., Seidman, L.J., Kremen, W.S., Pepple, J.R., Lyons, M.J., Tsuang, M.T., 1995. Neuropsychological functioning among the nonpsychotic relatives of schizophrenic patients: a diagnostic efficiency analysis. J. Abnorm. Psychology 104, 286–304. Frith, C.D., Done, C.J., 1988. Towards a neuropsychology of schizophrenia. Br. J. Psychiatry 153, 437–443. Garrett, M., Silva, R., 2003. Auditory hallucinations, source monitoring, and the belief that “voices” are real. Schizophr Bull. 29 (3), 445–457. Gottesman, I.I., Bertelsen, A., 1989. Confirming unexpressed genotypes for schizophrenia. Arch. Gen. Psychiatry 46, 867–872. Gottesman, I.I., Shields, J., 1982. The epigenetic puzzle. In: Gottesman, I.I., Shields, J. (Eds.), Cambridge University Press, Cambridge. Green, M.F., Nuechterlein, K.H., Breitmeyer, B., 1997. Backward masking performance in unaffected siblings of schizophrenics patients. Arch. Gen. Psychiatry 54, 465–472. Hedges, L., Olkin, I., 1985. Statistical Methods for Meta-Analysis. Academic Press, New York, p. 80. Henkel, L.A., Johnson, M.K., De Leonardis, D.M., 1998. Aging and source monitoring: cognitive processes and neuropsychological correlates. J. Exp. Psychol. Gen. 127 (3), 251–268. Hoffmann, R.E., 1986. Verbal hallucinations and language production processes in schizophrenia. Behav. Brain Sci. 9 (3), 503–548. Hoffman, R.E., Hawkins, K.A., Gueorguieva, R., Boutros, N.N., Rachid, F., Carroll, K., Krystal, J.H., 2003. Transcranial magnetic stimulation of left temporoparietal cortex and medication-resistant auditory hallucinations. Arch. Gen. Psychiatry 60, 49–56.
292
J. Brunelin et al. / Schizophrenia Research 89 (2007) 287–292
Johnson, M.K., Hashtroudi, S., Lindsay, D.S., 1993. Source monitoring. Psychol. Bull. 114 (1), 3–28. Keefe, R.S.E., Silverman, J.M., Mohs, R.C., Siever, L.J., Harvey, P.D., Friedman, L., Roitman, S.E.L., DuPre, R.L., Smith, C.J., Schmeidler, J., Davis, K.L., 1997. Eye tracking, attention, and schizotypal symptoms in nonpsychotic relatives of patients with schizophrenia. Arch. Gen. Psychiatry 54, 169–176. Keefe, R.S., Arnold, M.C., Bayen, U.J., Harvey, P.D., 1999. Source monitoring deficits in patients with schizophrenia; a multinomial modelling analysis. Psychol. Med. 29, 903–914. Keefe, R.S., Arnold, M.C., Bayen, U.J., McEvoy, J.P, Wilson, W.H., 2002. Source-monitoring deficits for self-generated stimuli in schizophrenia: multinomial modelling of data from three sources. Schizophr. Res. 57 (1), 51–67. Keefe, R.S., Poe, M.P., McEvoy, J.P., Vaughan, A., 2003. Source monitoring improvement in patients with schizophrenia receiving antipsychotic medications. Psychopharmacology (Berl) 169 (3–4), 383–389. Laroi, F., Collignon, O., Van der Linden, M., 2005. Source monitoring for actions in hallucination proneness. Cogn. Neuropsychiatry 10 (2), 105–123. Moritz, S., Woodward, T.S., Whitman, J.C., Cuttler, C., 2005. Confidence in errors as a possible basis for delusions in schizophrenia. J. Nerv. Ment. Dis. 193 (1), 9–16. Nienow, T.M., Docherty, N.M., 2004. Internal source monitoring and thought disorder in schizophrenia. J. Nerv. Ment. Dis. 192 (10), 696–700. Nuechterlein, K.H., Dawson, M.E., 1984. Information processing dysfunctions in the developmental course of schizophrenic disorders. Schizophr. Bull. 10, 160–203. Nuechterlein, K.H., Dawson, M.E., Green, M.F., 1994. Informationprocessing abnormalities as neuropsychological vulnerability indicators for schizophrenia. Acta Psychiatr. Scand. 90 (suppl 384), 71–79.
Poulet, E., Brunelin, J., Bediou, B., Bation, R., Forgeard, L., Dalery, J., d'amato, T., Saoud, M., 2005. Slow transcranial magnetic stimulation can rapidly reduce resistant auditory hallucinations in schizophrenia. Biol. Psychiatry 57 (2), 188–191. RAVEN, 1987. Raven coloured progressive matrices. Issy-lesMoulineaux: Editions scientifiques et psychotechniques. Renzi, E.D., Faglioni, P., 1975. Verbal comprehension in aphasic and in normal subjects with a shortened version of the token test (author's transl). Riv. Patol. Nerv. Ment. 96 (4), 252–269. Saoud, M., d'Amato, T., Gutknecht, C., Triboulet, P., Bertaud, J.P., Marie-Cardine, M., Dalery, J., Rochet, T., 2000. Neuropsychological deficit in siblings discordant for schizophrenia. Schizophr. Bull. 26 (4), 893–902. Spring, B., Zubin, J., 1978. Attention and information-processing as indicators of vulnerability to schizophrenic episodes. In: Wynne, L.C., Cromwell, R.L., Matthysse, S. (Eds.), The nature of schizophrenia: approaches to research and treatment. Wiley, New York, NY, pp. 366–375. Stirling, J.D., Hellewell, J.S., Quraishi, N., 1998. Self-monitoring dysfunction and the schizophrenic symptoms of alien control. Psychol. Med. 28 (3), 675–683. Tien, A.Y., 1991. Distributions of hallucinations in the population. Soc. Psychiatry Psychiatr. Epidemiol. 26 (6), 287–292. Verdoux, H., van Os, J., 2002. Psychotic symptoms in non-clinical populations and the continuum of psychosis. Schizophr. Res. 54 (1–2), 59–65. Vinogradov, S., Willis-Shore, J., Poole, J.H., Marten, E., Ober, B.A., Shenaut, G.K., 1997. Clinical and neurocognitive aspects of source monitoring errors in schizophrenia. Am. J. Psychiatry 154 (11), 1530–1537.