Schizophrenia Resrcwch, 7 ( 1992) 169% I76 I:(’ 1992 Elsevier Science Publishers B.V. All rights reserved
SCHIZO
169 0920.9964/92/$05.00
00223
Wisconsin Card Sorting Test deficits in chronic paranoid schizophrenia Evidence for a relatively Robert
W. Butlera,bqC,
Melissa
A. Jenkinsa,
discrete subgroup? Joyce
Sprock”,
and
David
L. Braff”
“Dcjpartmcw~ of P.s.~~chiutry. b’ni~r.si!, of’ Califi~rniu, Sun Diego. CA, USA and hDepartmmis of Pediatrics and Neurology, Memorial Shun-Kettwin,q Cuncer Ccwrw. Nrw York. NY, USA und ‘Departmenr of Psyhiarr~, Cornell Univrrsiry Medical Collqr, NW York, NY, USA (Received
I I April 1991; revision received 4 November
1991; accepted
21 November
1991)
Some, although not all, researchers have reported dramatically increased numbers of perseverative responses on the Wisconsin Card Sorting Test (WCST) in schizophrenic patients compared to normal comparison subjects. The current study was designed to further explore the nature of possible WCST deficits in a group of paranoid schizophrenic patients compared to normal and psychiatric comparison subjects. In the current study, schizophrenic patients had significantly greater numbers of perseverative responses on the WCST than the comparison groups. The sample of patients with schizophrenia appeared to be characterized by a non-Gaussian distribution of perseverative responses on the WCST. WCST-impaired and WCSTnonimpaired schizophrenic subgroups were compared on cognitive and symptom measures, and increased perseverative responding was associated with negative symptoms, slowed reaction time, and more hospitalizations. While additional research is necessary to further investigate hypotheses of frontal versus generalized brain dysfunction in schizophrenic patients, WCST impairment seems to be present in a clinically meaningful subgroup of paranoid schizophrenic patients. K~J, n~ords: Wisconsin
Card
Sorting
Test: Paranoid
schizophrenia;
INTRODUCTION
There has been increasing interest in the neurobiological substrate of schizophrenia over the past decade. While dysfunction involving the mesolimbit dopamine system in schizophrenia has long been suspected, researchers have now begun to accumulate data that implicates frontal-cortical involvement (Bleuler, 1950; Greenblatt and Solomon, 1953). Using brain imaging techniques, studies have demonstrated reduced frontal lobe metabolism in chronic schizophrenic patients when compared to normal control subjects (Buchsbaum
181.
Corres/~orzd~,nce /o R. Butler, Department of Pediatrics, Box MSKCC, 1275 York Avenue. New York, NY 10021, USA.
(Schizophrenia)
and Franzen, 1974). The et al., 1982; Ingvar increased sophistication of brain imaging techniques and neuropsychological measures are allowing researchers to begin empirical efforts at verifying frontal lobe deficits in the group of schizophrenias. Weinberger has hypothesized that many of the negative symptoms of schizophrenia, such as avolition and inertia, are mediated by hypofrontality and that positive symptoms, such as hallucinations and thought disorder, are more subcortically mediated (Weinberger, 1987). Related to this theory, two studies (Berman et al., 1988; Weinberger et al., 1988) documented frontal lobe hypometabolism, using regional cerebral blood flow methodology, in schizophrenic patients during completion of a modified version of the Wisconsin Card Sorting
170
Test (WCST). The WCST is a neuropsychological measure that assesses flexible problem solving and the ability to benefit from minimally directive feedback on performance errors. Past research has suggested that perseverative responding on the WCST, while not pathognomonic, does tend to be associated with frontal brain dysfunction (Milner, 1963; Robinson et al., 1980). Several studies have investigated WCST performance in schizophrenic patients and documented evidence for impaired responding on this test (Fey, 1951; Malmo, 1974; Kolb and Whishaw, 1983; Stuss et al., 1983). Questions remain, however, concerning the possibility that these observed deficits on the WCST might be a function of a generalized performance deficit rather than a relatively specific frontal mediated dysfunction. Directly addressing this issue, Goldberg et al. (1987) attempted to improve WCST performance in a sample of chronic inpatient schizophrenic subjects through intensive training on the task itself. The authors reported that considerable instruction did not result in improved WCST performance; however, training on a control task of verbal learning was facilitated by instruction. Thus, a considerable body of evidence suggests that increased WCST perseverative responding is a characteristic of schizophrenia. It may be important to note, however, that much of this evidence has been collected on hospitalized patients who were moderately to severely ill and, in some cases, WCST administration was modified from the standardized administration method. Braff et al. (199 1) attempted to determine if increased WCST perseverative responding was characteristic of mild to moderately ill, non-hospitalized chronic schizophrenic subjects when compared to normal controls. These researchers administered the WCST to 40 schizophrenic patients and 40 matched controls, and reported that 16 of the 40 schizophrenic patients performed in a range suggestive of significant impairment. Although some of the schizophrenic patients had overall increases in perseverative responses on the WCST, these patients also had many more prominent deficits on an extended Halstead Reitan Battery. Thus, it appears that some patients with chronic schizophrenia exhibit clearly abnormal performance on the WCST that may be secondary to frontal lobe impairment but that many other
neuropsychological deficits are found in these patients. The literature also suggests that not all chronic schizophrenic subjects have difficulty with the WCST and that, in part, increased perseverative responses might be a function of degree of illness and severity of generalized symptomatology. It may be that, in a carefully defined sample of chronic schizophrenic subjects, a subgroup will emerge that exhibits impaired WCST performance. If a subgroup such as this could be identified, it would be important to investigate if WCST impaired schizophrenic subjects differ from their counterparts on demographic, symptom, illnessrelated, and other cognitive and neuropsychological variables. The current study was designed to attempt to identify a WCST impaired subgroup in a carefully defined sample of inpatients with chronic paranoid schizophrenia. The sample was restricted to paranoid schizophrenia in order to reduce diagnostic heterogeneity. If this subgroup could be identified, we hypothesized that it would be characterized by increased negative symptoms, as predicted by Weinberger’s theory (Weinberger, 1987) and our previous results (Braff et al., 1991). We also hypothesized that WCST impaired schizophrenic subjects would show evidence of increased cognitive dysfunction on other neuropsychological measures of frontal brain function.
METHODS Subjects The schizophrenic group consisted of 48 inpatients hospitalized at the UCSD Medical Center. All subjects met the Research Diagnostic Criteria (RDC) for either chronic or subchronic paranoid schizophrenia. Diagnoses were made by the first and fourth authors based on review of medical history and administration of the Schedule for Affective Disorders and Schizophrenia (SADS) (Spitzer and Endicott, 1975). In order to establish initially that the schizophrenic group was characterized by increased perseverative responding on the WCST, data were also collected on groups of psychiatric and normal controls. The psychiatric control group consisted of 18 inpatient subjects sampled from the same hospital as the schizophrenic subjects. Seven of the
171
subjects had unipolar depression and, of these, three subjects had a comorbidity of personality disorder and one had a comorbidity of posttraumatic stress disorder (PTSD). Seven subjects had bipolar-disorder (four were manic, two were depressed, and one was not otherwise specified). Two patients had PTSD (one had a comorbidity of personality disorder) and two patients had another unspecified affective disorder. Data on a total of 27 normal control subjects were obtained as well. These subjects were recruited from the community by newspaper advertisement and posters. Normal control subjects were screened for significant psychopathology by both a SADS interview and a Minnesota Multiphasic Personality Inventory (MMPI) based algorithm designed to identify psychotic symptoms and schizotypy in the normal population (Butler et al., 1989). A summary of the descriptive characteristics of the three groups is presented in Table I. Measures
The WCST was administered to all subjects in the standard, manual-defined manner (Heaton, 1981). In addition to already described measures, the two patient groups were administered the Scales for the Assessment of Negative and Positive Symptoms (SANS, SAPS) (Andreasen, 1984a,b) and the Brief Psychiatric Rating Scale (BPRS). Subjects in the patient groups were also administered measures of verbal fluency (Benton, 1968) and nonverbal TABLE
1
Age Education
Schizophrenic
Psychiutrlc
Normul
patients
contro1.c
coniro1.c
31.5 (7.9) n=48
31.2 (9.2) ?I= 18
33.4 (9.2) n=27
11.8 (1.7) n=48
13.6 (2.0) n= 18
14.7 (2.2) n=27
WAIS-R
vocabulary
8.3 (2.5) n=45
9.9 (3.0) n= 18
11.4 (2.2) n=26
WAIS-R
digit span
7.2 (2.6) n=43
8.3 (3.2) rz= 18
11.3 (2.7) n=27
Duration
of illness
11.7 (6.9) n=46 12.0 (16.8) n=34
8.8 (6.9) n= 17 4.4 (5.5) n= 17
N/A
Number of previous hospitalizations
N/A
fluency (Jones-Gottman and Mimer, 1977) as additional indices of frontal brain integrity. Finally, all subjects completed a reaction time procedure (Zahn et al., 1978) as an index of information processing ability. As indicated in Table 2, not all subjects were able to complete the full battery. General
procedures
Patients who agreed to participate signed informed consent papers and were not reimbursed for participation. Approximately 50% of all patients approached agreed to participate in the study. Normal controls contacted a research assistant by phone and, if interested, were scheduled for testing. Approximately 80% of these subjects completed the test battery. Symptom measures and neuropsychological testing entailed approximately 445 h of subject participation. In most cases, the subjects completed the procedure in one sitting, however, subject fatigue occasionally necessitated more than one testing session. In no case did subject data collection require more than three sessions or extend over a one week period.
RESULTS
One-way ANOVAs with post-hoc testing revealed no significant group differences in age; however, the schizophrenic group did have significantly fewer years of education than the psychiatric control group (t=3.65, df= 1,64, p=O.OOl) and the normal group (t = 6.42, df= 1,73, p
SCHIZO
169 0920.9964/92/$05.00
00223
Wisconsin Card Sorting Test deficits in chronic paranoid schizophrenia Evidence for a relatively Robert
W. Butlera,bqC,
Melissa
A. Jenkinsa,
discrete subgroup? Joyce
Sprock”,
and
David
L. Braff”
“Dcjpartmcw~ of P.s.~~chiutry. b’ni~r.si!, of’ Califi~rniu, Sun Diego. CA, USA and hDepartmmis of Pediatrics and Neurology, Memorial Shun-Kettwin,q Cuncer Ccwrw. Nrw York. NY, USA und ‘Departmenr of Psyhiarr~, Cornell Univrrsiry Medical Collqr, NW York, NY, USA (Received
I I April 1991; revision received 4 November
1991; accepted
21 November
1991)
Some, although not all, researchers have reported dramatically increased numbers of perseverative responses on the Wisconsin Card Sorting Test (WCST) in schizophrenic patients compared to normal comparison subjects. The current study was designed to further explore the nature of possible WCST deficits in a group of paranoid schizophrenic patients compared to normal and psychiatric comparison subjects. In the current study, schizophrenic patients had significantly greater numbers of perseverative responses on the WCST than the comparison groups. The sample of patients with schizophrenia appeared to be characterized by a non-Gaussian distribution of perseverative responses on the WCST. WCST-impaired and WCSTnonimpaired schizophrenic subgroups were compared on cognitive and symptom measures, and increased perseverative responding was associated with negative symptoms, slowed reaction time, and more hospitalizations. While additional research is necessary to further investigate hypotheses of frontal versus generalized brain dysfunction in schizophrenic patients, WCST impairment seems to be present in a clinically meaningful subgroup of paranoid schizophrenic patients. K~J, n~ords: Wisconsin
Card
Sorting
Test: Paranoid
schizophrenia;
INTRODUCTION
There has been increasing interest in the neurobiological substrate of schizophrenia over the past decade. While dysfunction involving the mesolimbit dopamine system in schizophrenia has long been suspected, researchers have now begun to accumulate data that implicates frontal-cortical involvement (Bleuler, 1950; Greenblatt and Solomon, 1953). Using brain imaging techniques, studies have demonstrated reduced frontal lobe metabolism in chronic schizophrenic patients when compared to normal control subjects (Buchsbaum
181.
Corres/~orzd~,nce /o R. Butler, Department of Pediatrics, Box MSKCC, 1275 York Avenue. New York, NY 10021, USA.
(Schizophrenia)
and Franzen, 1974). The et al., 1982; Ingvar increased sophistication of brain imaging techniques and neuropsychological measures are allowing researchers to begin empirical efforts at verifying frontal lobe deficits in the group of schizophrenias. Weinberger has hypothesized that many of the negative symptoms of schizophrenia, such as avolition and inertia, are mediated by hypofrontality and that positive symptoms, such as hallucinations and thought disorder, are more subcortically mediated (Weinberger, 1987). Related to this theory, two studies (Berman et al., 1988; Weinberger et al., 1988) documented frontal lobe hypometabolism, using regional cerebral blood flow methodology, in schizophrenic patients during completion of a modified version of the Wisconsin Card Sorting
170
Test (WCST). The WCST is a neuropsychological measure that assesses flexible problem solving and the ability to benefit from minimally directive feedback on performance errors. Past research has suggested that perseverative responding on the WCST, while not pathognomonic, does tend to be associated with frontal brain dysfunction (Milner, 1963; Robinson et al., 1980). Several studies have investigated WCST performance in schizophrenic patients and documented evidence for impaired responding on this test (Fey, 1951; Malmo, 1974; Kolb and Whishaw, 1983; Stuss et al., 1983). Questions remain, however, concerning the possibility that these observed deficits on the WCST might be a function of a generalized performance deficit rather than a relatively specific frontal mediated dysfunction. Directly addressing this issue, Goldberg et al. (1987) attempted to improve WCST performance in a sample of chronic inpatient schizophrenic subjects through intensive training on the task itself. The authors reported that considerable instruction did not result in improved WCST performance; however, training on a control task of verbal learning was facilitated by instruction. Thus, a considerable body of evidence suggests that increased WCST perseverative responding is a characteristic of schizophrenia. It may be important to note, however, that much of this evidence has been collected on hospitalized patients who were moderately to severely ill and, in some cases, WCST administration was modified from the standardized administration method. Braff et al. (199 1) attempted to determine if increased WCST perseverative responding was characteristic of mild to moderately ill, non-hospitalized chronic schizophrenic subjects when compared to normal controls. These researchers administered the WCST to 40 schizophrenic patients and 40 matched controls, and reported that 16 of the 40 schizophrenic patients performed in a range suggestive of significant impairment. Although some of the schizophrenic patients had overall increases in perseverative responses on the WCST, these patients also had many more prominent deficits on an extended Halstead Reitan Battery. Thus, it appears that some patients with chronic schizophrenia exhibit clearly abnormal performance on the WCST that may be secondary to frontal lobe impairment but that many other
neuropsychological deficits are found in these patients. The literature also suggests that not all chronic schizophrenic subjects have difficulty with the WCST and that, in part, increased perseverative responses might be a function of degree of illness and severity of generalized symptomatology. It may be that, in a carefully defined sample of chronic schizophrenic subjects, a subgroup will emerge that exhibits impaired WCST performance. If a subgroup such as this could be identified, it would be important to investigate if WCST impaired schizophrenic subjects differ from their counterparts on demographic, symptom, illnessrelated, and other cognitive and neuropsychological variables. The current study was designed to attempt to identify a WCST impaired subgroup in a carefully defined sample of inpatients with chronic paranoid schizophrenia. The sample was restricted to paranoid schizophrenia in order to reduce diagnostic heterogeneity. If this subgroup could be identified, we hypothesized that it would be characterized by increased negative symptoms, as predicted by Weinberger’s theory (Weinberger, 1987) and our previous results (Braff et al., 1991). We also hypothesized that WCST impaired schizophrenic subjects would show evidence of increased cognitive dysfunction on other neuropsychological measures of frontal brain function.
METHODS Subjects The schizophrenic group consisted of 48 inpatients hospitalized at the UCSD Medical Center. All subjects met the Research Diagnostic Criteria (RDC) for either chronic or subchronic paranoid schizophrenia. Diagnoses were made by the first and fourth authors based on review of medical history and administration of the Schedule for Affective Disorders and Schizophrenia (SADS) (Spitzer and Endicott, 1975). In order to establish initially that the schizophrenic group was characterized by increased perseverative responding on the WCST, data were also collected on groups of psychiatric and normal controls. The psychiatric control group consisted of 18 inpatient subjects sampled from the same hospital as the schizophrenic subjects. Seven of the
171
subjects had unipolar depression and, of these, three subjects had a comorbidity of personality disorder and one had a comorbidity of posttraumatic stress disorder (PTSD). Seven subjects had bipolar-disorder (four were manic, two were depressed, and one was not otherwise specified). Two patients had PTSD (one had a comorbidity of personality disorder) and two patients had another unspecified affective disorder. Data on a total of 27 normal control subjects were obtained as well. These subjects were recruited from the community by newspaper advertisement and posters. Normal control subjects were screened for significant psychopathology by both a SADS interview and a Minnesota Multiphasic Personality Inventory (MMPI) based algorithm designed to identify psychotic symptoms and schizotypy in the normal population (Butler et al., 1989). A summary of the descriptive characteristics of the three groups is presented in Table I. Measures
The WCST was administered to all subjects in the standard, manual-defined manner (Heaton, 1981). In addition to already described measures, the two patient groups were administered the Scales for the Assessment of Negative and Positive Symptoms (SANS, SAPS) (Andreasen, 1984a,b) and the Brief Psychiatric Rating Scale (BPRS). Subjects in the patient groups were also administered measures of verbal fluency (Benton, 1968) and nonverbal TABLE
1
Age Education
Schizophrenic
Psychiutrlc
Normul
patients
contro1.c
coniro1.c
31.5 (7.9) n=48
31.2 (9.2) ?I= 18
33.4 (9.2) n=27
11.8 (1.7) n=48
13.6 (2.0) n= 18
14.7 (2.2) n=27
WAIS-R
vocabulary
8.3 (2.5) n=45
9.9 (3.0) n= 18
11.4 (2.2) n=26
WAIS-R
digit span
7.2 (2.6) n=43
8.3 (3.2) rz= 18
11.3 (2.7) n=27
Duration
of illness
11.7 (6.9) n=46 12.0 (16.8) n=34
8.8 (6.9) n= 17 4.4 (5.5) n= 17
N/A
Number of previous hospitalizations
N/A
fluency (Jones-Gottman and Mimer, 1977) as additional indices of frontal brain integrity. Finally, all subjects completed a reaction time procedure (Zahn et al., 1978) as an index of information processing ability. As indicated in Table 2, not all subjects were able to complete the full battery. General
procedures
Patients who agreed to participate signed informed consent papers and were not reimbursed for participation. Approximately 50% of all patients approached agreed to participate in the study. Normal controls contacted a research assistant by phone and, if interested, were scheduled for testing. Approximately 80% of these subjects completed the test battery. Symptom measures and neuropsychological testing entailed approximately 445 h of subject participation. In most cases, the subjects completed the procedure in one sitting, however, subject fatigue occasionally necessitated more than one testing session. In no case did subject data collection require more than three sessions or extend over a one week period.
RESULTS
One-way ANOVAs with post-hoc testing revealed no significant group differences in age; however, the schizophrenic group did have significantly fewer years of education than the psychiatric control group (t=3.65, df= 1,64, p=O.OOl) and the normal group (t = 6.42, df= 1,73, p
