Schizophrenia Research 75 (2005) 147 – 157 www.elsevier.com/locate/schres
Is unawareness of psychotic disorder a neurocognitive or psychological defensiveness problem? Kenneth L. Subotnika,*, Keith H. Nuechterleina,b, Victoria Irzhevskya, Christina M. Kitchena, Stephanie M. Wooc, Jim Mintza a
Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, 300 UCLA Medical Plaza, Room 2240, Los Angeles, CA 90095-6968, USA b Department of Psychology, University of California, Los Angeles, USA c Graduate School of Education and Psychology, Pepperdine University, USA Received 19 April 2004; received in revised form 30 November 2004; accepted 3 December 2004 Available online 20 January 2005
Abstract We examined whether deficits in attention and perceptual encoding as well as psychological defensiveness were associated with impaired awareness of disorder in schizophrenia. The Scale for Unawareness of Mental Disorder (SUMD) was administered to 52 outpatients with a recent onset of schizophrenia approximately 1–2 months following hospital discharge. Two versions of the Continuous Performance Test (CPT) were used to measure attentional impairment—the Degraded Stimulus CPT (DS-CPT) and a memory-load version (3-7 CPT). Three scales from the Minnesota Multiphasic Personality Inventory were used as indicators of psychological defensiveness: Scales L (Lie), K (Correction), and R (Repression). The Classification and Regression Tree (CART) program, a nonparametric statistical method, was used to identify relationships among multiple predictor variables and to provide optimal splitting scores for each predictor variable. Different combinations of poor target discrimination (dV) on the 3-7 CPT and a cautious response style on the DS-CPT were associated with the three levels of overall unawareness of having a mental disorder. For nonpsychotic patients, better target discrimination (dV) on the 3-7 CPT tended to be associated with better awareness of having a mental disorder. In contrast, unawareness among the patients who were psychotic at the time of the SUMD administration was not discriminated by attentional measures, but was associated with a combination of two measures of psychological defensiveness from the MMPI reflecting guardedness, psychological suppression, attempting to present oneself in a socially desirable light, and social acquiescence. Generally similar associations were found for two other dimensions of poor insight: unawareness of the beneficial effects of antipsychotic medication, and inability to attribute unusual thoughts and hallucinatory experiences to a mental disorder. D 2004 Elsevier B.V. All rights reserved. Keywords: Schizophrenia; Insight; Unawareness of mental disorder; Psychosis; CART; Neurocognition; MMPI; Psychosis
* Corresponding author. Tel.: +1 3108250334; fax: +1 3102063651. E-mail address:
[email protected] (K.L. Subotnik). 0920-9964/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2004.12.005
148
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
1. Introduction Lack of insight that one has a mental disorder has been identified as one of the most characteristic signs of schizophrenia, present in 85% of patients, and poor insight was identified by a large W.H.O. study as the most discriminating symptom between schizophrenia and affective disorder (Carpenter et al., 1973). Interestingly, poor insight is not a criterion for the diagnosis of schizophrenia (DSM-IV; American Psychiatric Association, 1987. This may be because poor insight has been understudied and because it is difficult to operationally define. Recent attempts to better define poor insight (Birchwood et al., 1994; David, 1990; Amador et al., 1991; Marks et al., 2000) have led to new research interest in this topic. The term bunawarenessQ is a descriptive term without etiological implications that has been introduced to separate this concept from the psychoanalytic concept of insight (Amador et al., 1991). The term also highlights the similarity to the unawareness of functional deficits that is often seen in persons with brain injury, typically involving the frontal lobe and/or parietal lobes. We will use the terms bunawarenessQ and bpoor awarenessQ interchangeably, and will use the term bpoor insightQ to represent the broader construct that has been proposed to encompass unawareness of the disorder, the effects of treatment, social consequences of the disorder, the presence of various psychiatric symptoms, as well as difficulty correctly relabeling individual psychiatric symptoms as due to a mental disorder. A number of studies has supported the concept that poor insight is associated with deficits in executive functioning in schizophrenia patients (Drake and Lewis, 2003; Lysaker et al., 2002, 1998; McEvoy et al., 1996; Lysaker and Bell, 1995; Rossell et al., 2003; Smith et al., 2000; Voruganti et al., 1997; Young et al., 1993; Young et al., 1998; Laroi et al., 2000). However, other studies have failed to find this relationship, casting doubt on the conceptualization of poor insight as secondary to executive functioning deficits (Cuesta et al., 1995; Dickerson et al., 1997; Sanz et al., 1998, Goldberg et al., 2001; Collins et al., 1997). Examination of the relationship between poor insight and another core deficit in schizophrenia, memory impairment, has also produced mixed results, with a positive finding reported by (Smith et al., 2000) for immediate
memory impairment, and negative findings reported by (Aleman et al., 2002) for working memory (Carroll et al., 1999; Rossell et al., 2003) for episodic long term memory, and (Kim et al., 2003) for immediate memory. Previous studies have attempted to link unawareness of illness to other cognitive processes. Given the prominence of attentional impairments in schizophrenia (Nuechterlein and Dawson, 1984; Spring et al., 1991; Cornblatt and Keilp, 1994; Nuechterlein et al., 1998) and links to frontal lobe dysfunctions (Buchsbaum et al., 1990; Cornblatt and Keilp, 1994; Mesulam, 2000), it seems reasonable to believe that poor insight may be associated with attentional deficits. Three studies have found a relationship between measures of attention and poor insight (Lysaker and Bell, 1995; Voruganti et al., 1997; Walker and Rossiter, 1989), and three have failed to detect this relationship (Dickerson et al., 1997; Kim et al., 2003; Rossell et al., 2003). However, it seems unlikely that neurocognitive factors account for poor insight to the exclusion of psychological coping factors. Lysaker et al., 2003, provided evidence to support the model by Startup (1996) wherein patients with neurocognitive deficits will have only average insight, and patients without neurocognitive deficits will either have good insight, or have poor insight which is secondary to psychological factors. This was demonstrated using a coping style inventory—patients with good neurocognitive functioning and poor insight tended to emotionally distance themselves from the psychological impact of life events. This is similar to the use of the MMPI Lie scale and a separate measure of denial of everyday problems as indicators of defensiveness (Young, 1998). It is unclear whether the lack of consistency among the published reports of the relationship between neurocognition and unawareness is because most studies did not examine neurocognitive deficits that are stable vulnerability factors for schizophrenia, such as certain attentional and early perceptual processing deficits (Nuechterlein et al., 1992). Alternatively, the relationship between neurocognition and unawareness of mental disorder might be nonlinear and involve a complex interplay with psychological defensiveness (Startup, 1996). In the current study, we examined whether a combination of neurocognitive and psychological factors were tied to poor insight. Specifically,
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
we examined whether deficits in attention and perceptual encoding as well as psychological defensiveness were associated with impaired overall awareness of having a mental disorder. Classification trees were used to identify relationships among the neurocognitive and psychometric predictor variables. We believe that patients weigh the costs and benefits of taking medication, which makes a better understanding of the correlates of the unawareness of the effects of medication vital to efforts to improve medication compliance. Therefore, we also examined whether these predictor variables were related to unawareness of the achieved effects of antipsychotic medication. We also examined two additional items from the SUMD regarding relabeling of symptoms as due to a mental disorder (i.e., the correctness of attributions regarding unusual thoughts and regarding hallucinatory experiences). Together, these SUMD items roughly correspond to the three dimensions of poor insight derived by Neumann et al. (2003), which were identified for two other insight measures. The three factors extracted by Neumann et al. (2003) were boverall unawarenessQ, bunawareness of the need for treatmentQ, and bability to relabel pathological symptomsQ.
2. Methods 2.1. Participants Participants were 52 recent-onset schizophrenia patients in the bDevelopmental Processes in the Early Course of IllnessQ longitudinal study of the Developmental Processes in Schizophrenic Disorders project at UCLA (P.I., Keith Nuechterlein, Ph.D.; Nuechterlein et al., 1992). These participants were recruited from 1995 to 2000 from local inpatient and outpatient facilities in the Los Angeles area or were directly referred to the UCLA Aftercare Research Program. All participants were presented with oral and written information about the research procedures involved in the study, and gave their informed consent. The demographic characteristics of the sample are presented in Table 1. All patients were within two years of the onset of their first psychotic episode, as required for entrance to this project. Other inclusion criteria for the study were a diagnosis by Research Diagnostic Criteria
149
Table 1 Participant demographic characteristics Gender Age Highest grade completed Race/Ethnicity
37 males, 15 females 24.7 years
S.D.=5.3
13.0 years
S.D.=18
39% Caucasian 13% Hispanic
13% Asian 23% AfricanAmerican
range 18–40 years range 8–17 years 12% Other/ Mixed
of schizophrenia (N=42) or schizoaffective disorder, mainly schizophrenic subtype (N=10); age of 18–45 years at the beginning of the study, and sufficient fluency in the English language to avoid invalidating research measures of thought and language processes. All participants met criteria from the Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSMIV: American Psychiatric Association, 1987 for schizophrenia (N=34), schizophreniform (N=8), or schizoaffective disorder (N=9). Exclusion criteria were evidence of a known neurological disorder, evidence of significant and habitual drug abuse or alcoholism in the six months prior to hospitalization, or premorbid mental retardation. The probands were treated with risperidone, psychoeducation, group skills training, and individual case management at the UCLA Aftercare Research Program. The mean daily dosage of risperidone at the time of testing was 4.3 (S.D.=2.2) mg/day. 2.2. Insight and symptom measures 2.2.1. The scale to assess unawareness of mental disorder (SUMD; Amador et al., 1993) The scale to assess unawareness of mental disorder (SUMD; Amador et al., 1993) was used to assess poor insight in all participants. For this report we examined three items that roughly correspond to the three factors of poor insight identified by Neumann et al. (2003). Specifically, we examined: (1) overall unawareness of having a mental disorder; (2) unawareness of the response to medication (if there has been a favorable response); (3) and the correctness of attributions regarding unusual thoughts and of hallucinatory experiences. The patient was asked to reflect on the time period of their most recent psychotic episode, and these ratings were based on
150
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
that episode of psychosis. The attributions regarding unusual thoughts was only rated if the BPRS Unusual Thought Content item had been rated a b3Q or above for that previous psychotic episode, which was the case for 51 subjects. All SUMD ratings were trichotomized as bGoodQ (rating of 1), bMediumQ (ratings of 2 or 3), or bPoorQ (ratings of 4 or 5). We also present the cross-tabulation of these items with the overall unawareness item to show that these do represent separable ratings within our sample (see Table 2). 2.2.2. Brief psychiatric rating scale (BPRS) Psychiatric symptoms were assessed using the expanded version of the Brief Psychiatric Rating Scale (BPRS; Ventura et al., 1993). A psychotic state was defined as a rating of 4 or greater (on a 1–7 scale) on one or more of the three BPRS psychotic items, Unusual Thought Content, Hallucinations, Conceptual Disorganization. In contrast, remission of psychosis was defined as ratings in the nonpathological range (3 or less) on all three of these BPRS items. By these criteria, 23 patients were considered psychotic and 29 were in remission (mean scores of the three positive
symptom items were 2.5 (S.D.=2.9) and 1.1 (S.D.=1.3), respectively). The psychotic and nonpsychotic patients were not significantly different on any of the basic demographic variables or on the primary measures of signal/noise discrimination (dV) for either the DS-CPT or the 3-7 CPT (see measures description below). 2.3. Neurocognitive measures 2.3.1. Degraded stimulus CPT (DS-CPT) Sustained, focused attention involving perceptual analysis and discrimination processes during vigilance was measured with a computerized version of the Degraded Stimulus CPT (DS-CPT: Nuechterlein and Asarnow, 1999a). This visual vigilance task with highly blurred single numerals requires the subject to detect occurrences of b0Q during an 8-minute period of single digits presented tachistoscopically at a pace of 1/sec. For this task, perceptually degraded single-digit numbers between 0 and 9 are presented in quasirandom order on a computer monitor for 33 ms each. A black/white reversal of a random 40% of pixels was used to produce degradation of the stimulus as well as
Table 2 Comparison of ratings of overall unawareness of having a mental disorder with other SUMD items Unawareness of the beneficial effects of antipsychotic medication Overall unawareness
Good Medium Good 10 5 Medium 5 7 High 0 2 Total 15 14 v 2(4)=35.0, pb.0001. 2 subjects did not have ratings on unawareness of medication effects.
High 1 2 18 21
Total 16 14 20 50
High 3 8 17 28
Total 16 15 17 48
High 2 2 11 15
Total 11 13 11 35
Relabeling of unusual thoughts Overall unawareness
Good Medium Good 6 7 Medium 0 7 High 0 0 Total 6 14 v 2(4)=29.1214, pb.0001. 4 subjects did not have ratings on relabeling of unusual thoughts. Relabeling of hallucinatory experiences Overall unawareness
Good Medium Good 4 5 Medium 1 10 High 0 0 Total 5 15 v 2(4)=26.2, pb.0001. 17 subjects did not have ratings on relabeling of hallucinatory experiences.
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
the background to create visual noise. Subjects were seated with their eyes 1 m from the computer screen, and were asked to press a response button when they saw the target stimulus. Twenty-five percent of the 480 stimuli were targets. A measure of signal/noise discrimination, dV, calculated from the proportion of hits and false alarms, was the primary performance variable. Higher dV represents greater sensitivity in distinguishing targets from non-targets. The secondary measure, beta (b), indicates the response criterion used by a subject. To normalize the somewhat skewed distribution of b, we used the natural log of b for the analyses. Low b scores indicate a bliberalQ response tendency, whereas higher scores indicate a bconservativeQ bias (an emphasis on the avoidance of incorrect responses). 2.3.2. Memory load CPT (3-7 CPT) This refinement of the original A-X CPT (Rosvold et al., 1956) measures focused, sustained attention under immediate memory load conditions (Nuechterlein and Asarnow, 1999b). Specifically, the subject was asked to press a button each time that a b3Q was followed by a b7Q within a quasi-random series of single, clearly focused numeral stimuli presented for 33 ms each at a rate of 1 per s. A computer program (Nuechterlein and Asarnow, 1999b) was used to present the 640 stimuli, 25% of which were 3-7 sequences (12.5% were 7’s that required a response). As with the DS-CPT, the primary measure was an index of sensitivity (dV) and the secondary measure was beta (b), a measure of response criterion that is moderately correlated with DS-CPT b. 2.4. Measures of psychological defensiveness We hypothesized that Minnesota Multiphasic Personality Inventory (MMPI: (Hathaway and McKinley, 1951) scales presumed to be indicative of psychological defensiveness and denial of problems, L, K, and R, would be related to unawareness of illness. Scales L and K are two of the validity scales and suggestive of test taking defensiveness, specifically a denial of universally common personal failings (L) and an intolerance of unconventional attitudes and an attempt to present oneself in a socially desirable light (K). Scale R is a factor analytically derived scale for which higher scores are associated with psycho-
151
logical suppression, social acquiescence, guardedness, sidestepping trouble and disagreeable situations, being painstakingly slow and thorough, and resistant to psychological explanations of symptoms (Caldwell, 1988; Dahlstrom et al., 1975). The SUMD was typically administered within the first month after study entry. On average, the DS-CPT was administered within 34.1 (S.D.=24.6) days from the time of the SUMD interview. Similarly, the 3-7 CPT was administered within 32.4 (S.D.=23.8) days, the MMPI within 28.3 (S.D.=20.7) days, and the BPRS within 5.1 (S.D.=6.1) days from the SUMD interview. 2.5. Data analyses The primary analyses were conducted using the Classification and Regression Trees (Breiman et al., 1984) using CART 5.0 (Salford Systems, San Diego, CA). CART is a nonparametric statistical methodology that creates binary decision trees. CART examines all possible interactions among the predictor variables in order to identify the optimal combination of predictor cutting scores, and presents to the findings in a user friendly decision tree format. A within sample cross-validation approach is used to find the optimal tree. Because the CART method may be new to many readers, we also conducted multinomial logistic regression with the four neurocognitive and three MMPI variables used to predict the unawareness of mental disorder scores. As with CART, multinomial logistic regression also is capable of examining categorical outcome variables with more than two levels, but is much more limited than CART in that complex non-linear relationships cannot be readily identified.
3. Results 3.1. Classification and regression tree (CART) analyses 3.1.1. Overall unawareness of mental disorder The CART analyses were successful in identifying the variables that, in combination, predicted the three levels of overall unawareness of illness. Cognitive
152
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
variables were the optimal predictors. As seen in Fig. 1, a combination of poor target discrimination (dV) on the 3-7 CPT (less than 3.75) and a cautious response style on the DS-CPT (a natural log of b that was greater than 0.73) tended to be associated with the greatest overall unawareness of having a mental disorder. Low 3-7 CPT target discrimination associated with less cautious response style tended to be associated with a moderate level of overall unawareness. Higher levels of 3-7 CPT target discrimination tended to be associated with better overall awareness. Because of our interest in the correlates of unawareness that persist when the acute psychosis has cleared, we examined a subset of 31 of patients who were not psychotic at the time of the SUMD administration. For these nonpsychotic patients, poorer target discrimination on the 3-7 CPT again tended to be associated with greater unawareness of having a mental disorder (see Fig. 2). In contrast, unawareness among the patients who were psychotic at the time of the SUMD administration was not discriminated by cognitive measures, but was associated with a combination of two measures of psychological defensiveness from the MMPI (see Fig. 3). Unawareness was best predicted by a combination of higher scores on the MMPI Scale K (a relatively sophisticated type of denial of socially unacceptable problems) and the MMPI Repression scale, whereas the moderate as well as lowest levels of unawareness were associated with lower Scale K (open and not defensive). Psychotic patients with a combination of high Scale K and low Scale R were
3-7 CPT d' Poor Discrimination (3-7 CPT d' ≤ 3.9) Good 11.1% Medium 33.3% Poor 55.6%
Fig. 2. CART tree for overall awareness of mental illness: nonpsychotic patients (N=29).
fairly evenly distributed among the three levels of unawareness. 3.1.2. Unawareness of achieved effects of antipsychotic medication The CART tree (not displayed here because of print space considerations) for unawareness of the effects of medication revealed a subtle interplay of two of the predictor variables, 3-7 CPT log b and 3-7 CPT dV, that was somewhat similar to the tree for overall unawareness. A cautious response tendency (3-7 CPT log b) combined with poorer discrimination (3-7 CPT dV) was associated with the poorest level of awareness. The prediction of unawareness of medication effects was not as interpretable for the nonpsychotic patients, although a simple tree showed that higher levels of defensiveness (MMPI Scale K) tended to be associated with poorer awareness. The tree for psychotic patients revealed that a combination of cautiousness (3-7 CPT log b greater than 2.2) and an MMPI indicator of unsophisticated defensiveness (Scale L T-score of 55 or greater) optimally discriminated the three levels of unawareness. Briefly,
3-7 CPT d'
Poor Discrimination (3-7 CPT d' ≤ 3.75)
High Beta: Cautious Responding (DS-CPT Beta (ln) > 0.73) Good 21.1% Medium 10.5% Poor 68.4%
Good Discrimination (3-7 CPT d' > 3.9) Good 72.7% Medium 9.1% Poor 18.2%
Good Discrimination (3-7 CPT d' > 3.75) Good 50.08% Medium 22.7% Poor 27.3%
Low Beta: Liberal Responding (DS-CPT Beta (ln) ≤ 0.73) Good 9.1% Medium 72.7% Poor 18.2%
Fig. 1. CART tree for overall awareness of mental illness (all 52 patients).
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
153
MMPI Scale K (Correction) T-Score
Above Average Defensiveness (MMPI Scale K T-Score > 47)
High Scale R (higher repression: T-Score > 59) Good 0.0% Medium 25.0% Poor 75.0%
Low Defensiveness MMPI Scale K T-Score ≤ 47 Good 40.0% Medium 60.0% Poor 0.0%
Low Scale R (lower repression: T-Score ≤ 59) Good 40.0% Medium 30.0% Poor 30.0%
Fig. 3. CART tree for overall awareness of mental illness: psychotic patients (N=23).
cautious and defensive psychotic patients had higher levels of unawareness of antipsychotic medication effects, and cautious, nondefensive patients tended to have lower levels of unawareness. More liberal responders (low 3-7 CPT log b) tended to have moderate levels of unawareness of antipsychotic medication effects. 3.1.3. Relabeling of symptoms as due to a mental disorder (i.e., incorrect attributions) The CART tree (not displayed) for the incorrectness of attributions regarding the reasons for having unusual thoughts during the psychotic episode at entry revealed that a low level of 3-7 CPT target discrimination (dV less than 3.87) tended to be associated with more inaccurate relabeling of unusual thoughts. This was the case for all subjects as well as for nonpsychotic patients. No tree could be grown for the psychotic patients for this SUMD item. Fewer participants had hallucinatory experiences during their psychotic episode at entry, so incorrectness of attributions about hallucinatory experiences could only be explored for a subset of 36 patients. Poorer target discrimination (d V) on the DS-CPT (less than 2.7) tended to be associated with incorrect relabeling of hallucinatory experiences. For patients who were not psychotic at the time of interview, poor DS-CPT d V (less than 2.1) tended to be associated with incorrect relabeling of hallucinatory experiences. Incorrect relabeling of hallucinatory experiences among psychotic patients was discriminated by MMPI Scale R (greater than T-score of 56), not by the DSCPT dV.
3.2. Multinomial logistic regression analysis 3.2.1. Overall unawareness of mental disorder The neurocognitive and three MMPI variables were entered into a multinomial logistic regression analysis with the SUMD overall unawareness of mental disorder score as the outcome. Our initial use of continuous variables as predictors did not result in a significant predictive model. Thus, to parallel the selection and dichotomization of predictor variables by CART, we also tested a predictive model that used the same predictor variables identified by CART and the variable cutoffs selected by CART. The best model was selected using the Akiake Information Criteria (AIC). Using all patients, lower scores on the 3-7 CPT measure of target discrimination (dV) (Beta= 1.2, S.E.=0.39, Wald v 2=10.3, pb.002) and a cautious response style on the DS-CPT (natural log of b) (Beta=2.1, S.E.=0.75, Wald v 2=7.6, pb.006) predicted overall unawareness of having a mental disorder. For the nonpsychotic patients, low values of the 3-7 CPT measure of target discrimination (dV) (Beta= 1.2, S.E.=0.45, Wald v 2=7.1, pb.008) predicted overall unawareness of having a mental disorder. For the psychotic patients, higher values of the MMPI Scale R (Beta=1.2, S.E.=0.56, Wald v 2=4.6, pb.04) predicted overall unawareness of having a mental disorder. 3.2.2. Unawareness of achieved effects of antipsychotic medication Similarly, the neurocognitive and MMPI variables were entered into a multinomial logistic regression analyses to predict unawareness of antipsychotic
154
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
medication effects using a model that corresponded to the optimal CART predictor variables and cutoff values for the discrimination of levels of unawareness of the effects of medication. None of the neurocognitive or MMPI scores significantly predicted unawareness for the entire sample, or for the nonpsychotic patients examined separately, although lower values of 3-7 CPT d V showed some tendency to predict greater unawareness of the achieved effects of medication (Beta=0.79, S.E.=0.48, Wald v 2=2.7, pb.10). For the psychotic patients, higher values of the MMPI Scale L (Beta=1.0, S.E.= 0.50, Wald v 2=3.9, pb.05) predicted greater unawareness of the achieved effects of antipsychotic medication. 3.2.3. Relabeling of symptoms as due to a mental disorder Again, the neurocognitive and MMPI variables were entered into a multinomial logistic regression, this time to predict inability to relabel unusual thoughts as a symptom of a mental disorder (i.e., inaccurate attributions) using the optimal CART predictor variables and cutoff values for those variables. None of the neurocognitive or MMPI scores significantly predicted incorrect relabeling of unusual thoughts for the entire sample, or for the psychotic patients examined separately. However, for the nonpsychotic patients, lower values of the 3-7 CPT d V predicted poorer ability to relabel unusual thoughts as a symptom of a mental disorder (Beta= 0.80, S.E.=0.40, Wald v 2 =3.9, pb.05). A separate multinomial logistic regression model using the CART predictor variables and cutoffs for the discrimination of ability to relabel hallucinatory experiences as a symptom of having a mental disorder found that lower DS-CPT dV significantly predicted poorer ability to relabel such experiences disorder (Beta= 1.26, S.E.=0.52, Wald v 2=5.9, pb.02) for all patients, and nonsignificantly so for the nonpsychotic patients (Beta= 0.95, S.E.=0.53, Waldv 2=3.2, pb.08).
4. Discussion Most previous studies of the neurocognitive basis for unawareness of having schizophrenia have implicated deficits in executive functioning or immediate memory, but only five studies have examined deficits in attention and perceptual encoding. Because deficits
in attention and perceptual processing, as well as impairment in insight, are recognized as core features of schizophrenia, we hypothesized that they would be interrelated. Our CART findings confirm that deficits in focused, sustained attention and target discrimination may limit patients’ capacities to recognize that they are suffering from a mental disorder. Poorer awareness of having a mental disorder tended to be associated with poorer target discrimination (dV) on a focused, sustained attention task that requires immediate or working memory (3-7 CPT). A response style measure (log b) from a perceptual load version of the continuous performance test (DS-CPT) further differentiated the patients with poor sustained target discrimination on the 3-7 CPT, from patients who were cautious and tended to under-respond being more likely to have the poorest level of insight, and the less cautious, more liberal responders having moderate insight. Response style did not differentiate levels of insight for patients with better focused, sustained attention. The multinomial logistic regression analyses were consistent with some of the key predictors identified through the CART procedure. CART has the advantage of being more easily interpreted compared to multinomial logistic regression. Some studies have implicated working memory as a deficit associated with poor insight. In our study, for most dimensions of poor insight target discrimination in the memory load CPT was a better discriminator of SUMD items than was the target discrimination in the Degraded Stimulus CPT. The DS-CPT target discrimination index did discriminate the ability to relabel hallucinatory experiences as a symptom of a mental disorder. It is not clear whether this difference reflects a meaningful relationship between impaired visual perceptual discrimination and incorrect understanding of the causes of hallucinatory experiences, or is merely an artifact of the smaller number of subjects who had had hallucinatory experiences. In general, our findings are consistent with Kinsbourne’s (Kinsbourne, 1998) suggestion that deficits in the distribution of attentional resources could result in impaired insight. It is unclear why patients with poor focused, sustained attention who were also under-responsive to CPT stimuli had the greatest unawareness. Underresponding may be an indicator of an impairment in decision making that is separable from target discrimination in the 3-7 CPT.
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
We have been intrigued by our clinical observation that many patients will achieve relative remission of psychotic symptoms without coming to recognize that they have had an episode of psychosis. This study found that the factors associated with deficits in awareness were different during a psychotic episode than during a period of remission shortly following a psychotic episode. Unawareness of having a mental disorder that persists during periods of remission from psychosis was associated with poor target discrimination within our memoryload attentional task. We have previously shown that this measure reflects a mediating vulnerability factor for schizophrenia (Nuechterlein et al., 1992). That is, although patients improve on this measure during remission, they remain impaired relative to normal controls. Thus, unawareness that persists during relative remission of psychosis was associated with a neurocognitive deficit that also tends to persist during remission, and not with psychological defensiveness. In general, the findings for the unawareness of the effects of medication and incorrect attributions regarding symptoms paralleled those for overall unawareness. These patterns of findings suggest that a psychological explanation of poor insight, such as a bsealing overQ (McGlashan et al., 1975), cannot fully account for poor insight during psychotic remission. In contrast, in patients who were psychotic at the time of assessment, unawareness of having a mental disorder, unawareness of the effects of antipsychotic medication, and inability to attribute symptoms to a mental disorder were in general associated with psychological defensiveness, not neurocognitive deficits. In particular, psychotic patients who attempted to present themselves in a socially desirable manner, and tended to be guarded, painstakingly slow and thorough, and to acquiesce in order to avoid social conflict were the most likely to have the poorest overall unawareness of having a mental disorder and poorest ability to attribute hallucinatory experiences to a mental disorder. Psychotic patients who tended to be cautious, inflexible in solving problems, and denied common personal failings tended to have less awareness of the beneficial effects of antipsychotic medication. It is possible that any influence of neurocognitive deficits on insight for these psychotic patients was obscured by the influence of psychological defensiveness. Our results are con-
155
sistent with the finding of Young et al. (1998) that unawareness of having a mental disorder was associated with greater denial of common personal failings. However, because these findings only characterized patients who were psychotic at the time of testing, we cannot say that these dimensions of poor insight, in general, can be attributable to psychological defensiveness. These findings are consistent with a report that poor insight was only modestly associated with neurocognitive impairment during inpatient hospitalization (Kemp and David, 1996). The authors suggested that the relationship between insight and neurocognitive functioning might be obscured during acute psychosis (Kemp and David, 1996). Our MMPI findings further clarify that not only is the influence of neurocognitive deficits on poor insight less during acute psychosis, but the role of psychological defensiveness is greater during such periods. The more unaware a patient is of having a mental disorder, and the more unaware the patient is of the benefits of antipsychotic medication, the less likely he or she will adhere to treatment (McEvoy et al., 1996). Our findings suggest that neurocognitive deficits that tend to be relatively enduring contribute to poor insight, particularly in clinically stable patients. Poor insight patients with attention and immediate memory deficits are at heightened risk for treatment nonadherence. For such patients, external supports for medication adherence, such as family monitoring, are likely to be needed to foster adherence. In summary, deficits in focused, sustained attention may limit patients’ capacities to recognize that they are suffering from a mental disorder. This effect was particularly evident after their psychotic symptoms have resolved. This association was obscured for patients who were psychotic at the assessment point, during which time psychological characteristics such as denying common personal failings, avoiding interpersonal conflict, attempting to present oneself in socially desirable manner, being cautious, guarded, painstakingly slow and thorough, and inflexible in solving problems were more salient than the neurocognitive predictors of insight. The examination of other neurocognitive deficits that have been found to be associated with poor insight in other studies, such as executive functioning and immediate verbal memory deficits, was beyond the scope of the current
156
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157
study. However, it is possible that examining these other neurocognitive deficits using CART methodology might shed some light on the repeated failures to replicate the relationship between executive functioning impairment and poor insight. Similarly, given our demonstration of the importance of taking the psychotic state of patients into consideration, doing so in future studies might serve to clarify the relationship between executive functioning deficits and poor insight. Given the smaller numbers of subjects involved when the psychotic and nonpsychotic subsamples were examined in the current study, this pattern of findings requires replication in future studies. In addition, because some of the discriminators of dimensions of poor insight identified here, especially begin cautious in responding during the neurocognitive tests, painstakingly slow and thorough, and inflexible in solving problems, might be also associated with the presence of negative symptoms, future studies of personality characteristics, neurocognitive impairment, and insight should consider the influence of negative symptoms.
References Aleman, A., de Haan, E.H.F., Kahn, R., 2002. Insight and neurocognitive function in schizophrenia. J. Neuropsychiatry Clin. Neurosci. 14, 241 – 242. Amador, X., Strauss, D.H., Yale, S., Gorman, J.M., 1991. Awareness of illness in schizophrenia. Schizophr. Bull. 17, 113 – 132. Amador, X.F., Strauss, D., Yale, S., Flaum, M.M., 1993. Assessment of insight in psychosis. Am. J. Psychiatry 150, 173 – 179. American Psychiatric Association, 1987. Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R). American Psychiatric Association, Washington, DC. Birchwood, M., et al., 1994. A self-report insight scale for psychosis: reliability, validity and sensitivity to change. Acta Psychiatr. Scand. 89, 62 – 67. Breiman, L., Friedman, J.H., Olshen, R.A., Stone, C.J., 1984. Classification and Regression Trees (CART). Wadsworth, Monterey, CA. Buchsbaum, M.S., et al., 1990. Glucose metabolic rate in normals and schizophrenics during the continuous performance test assessed by positron emission tomography. Br. J. Psychiatry 156, 216 – 227. Caldwell, A.B., 1988. MMPI supplemental scale manual. Caldwell Report Publishing, Los Angeles, CA. Carpenter, W., Strauss, J., Bartko, J., 1973. Flexible system for the diagnosis of schizophrenia: report from the WHO international pilot study of schizophrenia. Science 182, 1275 – 1277.
Carroll, A., et al., 1999. Correlates of insight and insight change in schizophrenia. Schizophr. Res. 35, 247 – 253. Collins, A.A., Remington, G.J., Coulter, K., Birkett, K., 1997. Insight, neurocognitive function and symptom clusters in chronic schizophrenia. Schizophr. Res. 27, 37 – 44. Cornblatt, B.A., Keilp, J.G., 1994. Impaired attention, genetics, and the pathophysiology of schizophrenia. Schizophr. Bull. 20, 31 – 46. Cuesta, M.J., Peralta, V., Caro, F., Leon, J., 1995. Is poor insight in psychotic disorders associated with poor performance on the Wisconsin Card Sorting Test? Am. J. Psychiatry 152, 1380 – 1382. Dahlstrom, W.G., Welsh, G.S., Dahlstrom, L.E., 1975. An MMPI Handbook: I. Clinical Interpretation. Univ. of Minnesota Press, Minneapolis, MN. David, A.S., 1990. Insight and psychosis. Br. J. Psychiatry 156, 798 – 808. Dickerson, F., Boronow, J., Ringel, N., Parente, F., 1997. Lack of insight among outpatients with schizophrenia. Psychiatr. Serv. 48, 195 – 199. Drake, R., Lewis, S.W., 2003. Insight and neurocognition in schizophrenia. Schizophr. Res. 62, 165 – 173. Goldberg, R.W., Green-Paden, L.D., Lehman, A.F., Gold, J.M., 2001. Correlates of insight in serious mental illness. J. Nerv. Ment. Dis., March 189, 137 – 145. Hathaway, S.R., McKinley, J.C., 1951. Minnesota Multiphasic Personality Inventory. Psychological Corporation, New York. Kemp, R., David, A.S., 1996. Psychological predictors of insight and compliance in psychotic patients. Br. J. Psychiatry 169, 444 – 450. Kim, C.H., Jayathilake, K., Meltzer, H.Y., 2003. Hopelessness, neurocognitive function, and insight in schizophrenia: relationship to suicidal behavior. Schizophr. Res. 60, 71 – 80. Kinsbourne, M, 1998. Representations in consciousness and the neuropsychology of insight. In: Amador, X.F., David, A.S. (Eds.), Insight and Psychosis. Oxford University Press, New York, pp. 174 – 192. Laroi, F., et al., 2000. Unawareness of illness in chronic schizophrenia and its relationship to structural brain measures and neuropsychological tests. Psychiatry Res. Neuroimaging 100, 49 – 58. Lysaker, P., Bell, M., 1995. Work rehabilitation and improvements in insight in schizophrenia. J. Nerv. Ment. Dis., February 183, 103 – 106. Lysaker, P., Bell, M.D., Bryson, G., Kaplan, E., 1998. Neurocognitive function and insight in schizophrenia: support for and association with impairments in executive function but not with impairments in global function. Acta Psychiatr. Scand. 97, 297 – 301. Lysaker, P.H., Bryson, G.J., Bell, M.D., 2002. Insight and work performance in schizophrenia. J. Nerv. Ment. Dis. 190, 142 – 146. Lysaker, P.H., Bryson, G.J., Lancaster, R.S., Evans, J.D., Bell, M.D., 2003. Insight in schizophrenia: associations with executive function and coping style. Schizophr. Res. 59, 41 – 47. Marks, K.A., Fastenau, P.S., Lysaker, P.H., Bond, G.R., 2000. SelfAppraisal of Illness Questionnaire (SAIQ): relationship to
K.L. Subotnik et al. / Schizophrenia Research 75 (2005) 147–157 researcher-rated insight and neuropsychological function in schizophrenia. Schizophr. Res. 45, 203 – 211. McEvoy, J., Hartman, M., Gottlieb, D., Godwon, S., Apperson, L.J., Wilson, W., 1996. Common sense, insight and neuropsychological test performance in schizophrenia patients. Schizophr. Bull. 22, 635 – 641. McGlashan, T.H., Levy, S.T., Carpenter Jr., W.T., 1975. Integration and sealing over: clinically distinct recovery styles from schizophrenia. Arch. Gen. Psychiatry 32, 1269 – 1272. Mesulam, M.-M., 2000. Behavioral neuroanatomy: large-scale networks, association cortex, frontal syndromes, the limbic system, and hemispheric specializations. In: Mesulam, M.-M. (Ed.), Principles of Behavioral and Cognitive Neurology. Oxford University Press, London, pp. 1 – 120. Neumann, C., Gonterman, A., Reynolds, F., Pugsley, C., Tekell, J., 2003. The structure of insight in patients with psychotic disorders. Society for Research in Psychopathology, Toronto, Canada. Nuechterlein, K.H., Asarnow, R.F., 1999a. Degraded Stimulus Continuous Performance Test. Los Angeles. Nuechterlein, K.H., Asarnow, R.F., 1999b. Memory-load Continuous Performance Test (3-7 CPT) program for IBM-Compatible Microcomputers. Authors, Los Angeles. Nuechterlein, K.H., Dawson, M.E., 1984. Information processing and attentional functioning in the developmental course of schizophrenic disorders. Schizophr. Bull. 10, 160 – 203. Nuechterlein, K.H., et al., 1992. Developmental processes in schizophrenic disorders: longitudinal studies of vulnerability and stress. Schizophr. Bull. 18, 387 – 425. Nuechterlein, K.H., et al., 1998. Neurocognitive vulnerability factors for schizophrenia: convergence across genetic risk studies and longitudinal trait/state studies. In: Lenzenweger, M.F., Dworkin, R.H. (Eds.), Origins and Development of Schizophrenia: Advances in Experimental Psychopathology. American Psychological Association, Washington, DC, pp. 299 – 327. Rossell, S.L., Coakes, J., Shapleske, J., Woodruff, P.W.R., David, A.S., 2003. Insight: its relationship with cognitive function,
157
brain volume and symptoms in schizophrenia. Psychol. Med. 33, 111 – 119. Rosvold, H.E., Mirsky, A.F., Sarason, I., Bransome Jr., E.D., Beck, L.H., 1956. A continuous performance test of brain damage. J. Consult. Psychol. 20, 343 – 350. Sanz, M., Constable, I., Kemp, R., David, A.S., 1998. A comparative study of insight scales and their relationship to psychopathological and clinical variables. Psychol. Med. 28, 437 – 446. Smith, T.E., Hull, J.M., Israel, L.M., Wilson, D.F., 2000. Insight, symptoms and neurocognitive functioning in schizophrenia and schizoaffective disorder. Schizophr. Bull. 26, 193 – 200. Spring, B., Weinstein, L., Freeman, R., Thompson, S., 1991. Selective attention in schizophrenia. In: Steinhauer, S.R., Gruzelier, J.H., et al., (Eds.), Handbook of Schizophrenia: Neuropsychology, Psychophysiology, and Information Processing. Elsevier, New York, NY, pp. 371 – 396. Startup, M., 1996. Insight and cognitive deficits in schizophrenia: evidence for a curvilinear relationship. Psychol. Med. 26, 1277 – 1281. Ventura, J., et al., 1993. Brief Psychiatric Rating Scale (BPRS) expanded version: scales, anchor points, and administration manual. Int. J. Methods Psychiatr. Res. 3, 227 – 243. Voruganti, L., Heslegrave, R.J., Awad, A.G., 1997. Neurocognitive correlates of positive and negative syndromes in schizophrenia. Can. J. Psychiatry 42, 1066 – 1071. Walker, E., Rossiter, J., 1989. Schizophrenic patients’ self-perceptions: legal and clinical implications. J. Psych. Law, 55 – 73. Young, D., Davila, R., Scher, H., 1993. Unawareness of illness and neuropsychological performance in chronic schizophrenia. Schizophr. Res. 10, 117 – 124. Young, D., Zakzanis, K.K., Bailey, C., Davila, R., Griese, J., Sartory, G., Thom, A., 1998. Further parameters of insight and neuropsychological deficit in schizophrenia and other chronic mental disease. J. Nerv. Ment. Dis. 186, 44 – 50.
