Stability in cognitive dysfunctions in schizophrenic patients

ELSEVIER Psychiatry Research 69 (1997) 131-141

PSYCHIATRY

RESEARCH

Stability in cognitive dysfunctions in schizophrenic patients

Bjmn Rishovd Rund”“, Nils Inge Landrs,“, Anne Lill Orbeckb

“Institute of Psychology, University of Oslo, P.O. Box 1094, Blindem, N-031 7 Oslo, Norway hDikemark Hospital, Oslo, Norway

Received 2 May 1996; revised 4 October 1996; accepted 15 November 1996

Abstract

The purpose of the present study was to classify cognitive dysfunctions in schizophrenic subjects according to a trait/state model. A sample of 15 patients was examined on 10 different cognitive measures in two distinctly different phases: an acute psychotic state and partial remission. To determine the degree of dysfunction at the two stages of illness, the schizophrenic patients were also compared to 14 non-psychiatric controls. Six of the 10 measures examined can be classified as cognitive deficits in schizophrenics. Four measures are possibly trait-dependent components: two backward masking scores and two long-term memory measures. A short-term memory measure is the only one that can be classified as an episode-linked factor. The other cognitive deficits found can be characterized as mediating vulnerability factors, i.e. they are more prominent in the acute psychotic state, but do not completely disappear during remission states. 0 1997 Elsevier Science Ireland Ltd.

Keywords: Longitudinal; Information processing; Trait/state; Memory; Backward masking; Concept formation

1. Introduction

The purpose of the present study is to investi- gate which cognitive deficits are episode-linked in schizophrenia and which are independent of clini-

*Corresponding author. Tel.: + 47 22855339; fax: +47

22854968.

cal state, i.e. enduring features which may be indicators of vulnerability.

Information processing models have become

central frames of reference for experimental

schizophrenia research. The principal assumption

underlying these models is that cognitive activities

can be described as a series of stages through

which information is guided. Different test

paradigms capture the subjects’ processing capac-

0165.17X1/97/$17,00 0 1997 Elsevier Science Ireland Ltd. All rights reserved.

PlZSO165-1781(96)03043-6

132 B.R. Rund et al. / Pqchiatly Research 69 (1997) 131-141

ity at different levels. Deficient performance has been demonstrated in schizophrenics in all stages of information processing. However, results have been inconsistent (see Cromwell and Spaulding, 1978; Nuechterlein and Dawson, 1984a; Knight, 1984; Rund and Landro, 1990; Asarnow et al., 1991; Braff et al., 1991; Kietzman, 1991; Saykin et al., 1991; Braff, 1993; Elliot and Sahakian, 1995).

A central hypothesis in this research field is that cognitive deficits play etiologically significant roles in the development of schizophrenia. The most essential issue in determining which cognitive factors play etiological roles and which are the result of the psychosis is to differentiate markers of vulnerability (see Zubin and Spring, 1977; Cromwell and Spaulding, 1978; Nuechter- lein et al., 1992) from episodic markers (also called symptom-linked variables). Deficits that fall into the category of vulnerability markers are stable characteristics that are consistently deviant from normal levels, even during symptomatic remission (Nuechterlein and Dawson, 1984b). Such deficits are detectable among non-symptomatic vulnerable persons and are present prior to onset of the illness. In addition, markers of vulnerability should differentiate schizophrenics from other psychotic patients (Asarnow and Mac- Crimmon, 1981). ‘Vulnerability factors’ may be causal links in the development or expression of schizophrenia).

Episode indicators, on the other hand, are as- sociated with certain features of schizophrenic symptomatology. They deviate from normal levels during symptomatic episodes, but return to normal levels during asymptomatic periods. Episode indicators seem to be synonymous with state-dependent components (Saccuzzo and Braff, 1986).

Nuechterlein and Dawson (1984) also suggest a third factor in their heuristic vulnerability/stress model: mediating vulnerability factors. These are postulated to deviate from normal levels during psychotic and asymptomatic states, but also to co-vary with level of symptomatology. The charac- teristic pattern of these factors is that they are present in asymptomatic states, and are further exacerbated in symptomatic (acute) states.

In order to determine the status of a certain deficit, vulnerability markers need to be demon-

strated in the premorbid, acute psychotic and relatively remitted (post-psychotic) phases of schizophrenia. ‘Trait-dependent components’ (Saccuzzo and Braff, 1986) may also be vulnerability markers, but a trait does not necessarily exist at a very early point in life. Thus, the methodological prerequisite of being present in a premorbid period is not that definite with respect to demonstration of a trait-dependent component as with vulnerability factors. However, the demonstration of a trait-dependent component can be seen as the first step in determining whether a factor is a vulnerability marker.

Using these criteria, several aspects of information processing and attentional functioning are potential vulnerability indicators for schizophrenics. These are deficits in: sustained focused attention in high-processing-load vigilance tasks; rapid read-out from sensory storage in the presence of patterned noise stimuli (backward masking); short-term recall involving active rehearsal espe- cially in the presence of distraction; and eye movement dysfunctions. Some other deficits have proven to be episode indicators, specifically deficient performance on the continuous performance test with low processing load (see Gold- berg et al., 1993; Lewine, 1984a; Nopoulos et al., 1994; Nuechterlein and Dawson, 1984; Rund and Landro, 1990; Sweeney et al., 1991). For other information processing deficits, the vulnerability status is unsettled.

The primary goal of the present study is to classify different categories of cognitive functions in schizophrenics according to the trait/state model outlined above. First, to distinguish between those which are deviant and those which are not deviant in schizophrenics. Second, to examine which are trait-dependent components, i.e. present during floridly psychotic phases as well as partially remitted phases with significant symptom improvement. Third, to find out which are state- dependent components, i.e. are present only during acute psychotic phases. Fourth, to examine which can be classified as mediating factors, i.e. those which are present during acute phases and return to a level between normal subjects and acute schizophrenics’ in the phase of partial remission. It must be possible to identify

B.R. Rund et al. /Psychiatry Research 69 (19971 1.31-141 13.3

schizophrenic patients who have changed from an acute psychotic state to partial remission with significant symptomatic improvement during a certain time period to conduct this research.

2. Methods

2.1. Subjects

The present study is a longitudinal assessment of cognitive functions in a large sample of psychotic subjects (see Rund et al., 1993). Most of the patients were examined at 1 and 2 years following the first (index) assessment. The subjects were 15 schizophrenic patients who clearly changed from an acute psychotic phase with florid symptoms to a partially remitted phase with significant symptom improvement during the period of testing. The schizophrenics were compared to 14 non-psychiatric (normal) controls in order to determine the degree of dysfunction at the two stages of illness.

Clinical state was determined using the expanded version of the Brief Psychiatric Rating Scale (BPRS), according to the principles developed by Lukoff et al. (1986) yet employing more stringent criteria for change from an acute to a remission state. Psychotism was based on three core BPRS symptom scales: unusual thought content, hallucinations, and conceptual disorganization. Clinical remission was determined by a 3- point decrease on any of the three key symptom scales, if a maximum severity score of 4 on that scale occurred, or by a decrease of at least 6 points on the sum score of the three key symptoms. Ten of the patients hada decrease of at least 6 points on the sum. Seven patients had a 3-point decrease on the key score of unusual thought content, six patients had a corresponding decrease on the conceptual disorganization score, and five patients had a 3-point decrease on the hallucination score.

An experienced psychologist carried out all the BPRS ratings. The interrater reliability of his ratings on the three psychotism scores was examined on a sample of 12 psychotic patients. The interrater agreement between the two raters was 0.87 (ICC 1,l; Schrout and Fleiss, 1979).

The sample of 15 patients included eight un- differentiated and seven disorganized schizophrenics assessed at index testing with the criteria from the Diagnostic and Statistical Manual of Men- tal Disorders (rev. 3rd edn; DSM-III-R; Ameri- canpsychiatric Association, 1987). Information for DSM-III-R ratings was obtained from the Struc- tured Clinical Interview for DSM-III (Spitzer and Williams, 1984),a semistructured tape-recorded research instrument. In addition, information was obtained from case records. Diagnostic interrater agreement, based on 30 patients from the original sample (22 schizophrenic and eight affectively disturbed subjects, see Rund et al,, 1993), was satisfactory (K = 0.76).

The patients represent a heterogeneous group in the following respects: (1) age: patients were between 16 and 63 years old at index assessment; (2) previous hospitalization: varying from none (four patients) to 26.3 years: and (3) educational levels: varying from 7 to 17 years of formal education.

The patients were screened for head traumas and other neuropathological conditions as well as alcoholism or drug abuse as a principal diagnosis. The screening was based upon patient reports and case records. Patients’ symptom scores, as assessed by the BPRS, were relatively high at the index testing. with a mean of 2.4 for all of the 24 items summed together. Their present psychosocial functioning measured by the Global Assess- ment Scale (GAS) (Endicott et al., 1076) was in the middle range of hospitalized psychiatric patients, reaching a mean of 38 for the group of 15 patients al the index testing. Premorbid adjustment, assessed by the UCLA Premorbid Adjust- ment Scale (Evans et al., 1972), was in the lower range for this group, averaging 15.3.

Twelve of the 15 patients had received neuroleptic medication in the acute phase, and all patients received such medication in their remission phase. (The reason that three more patients were on medication in the partial remission phase than in the acute phase is that a few patients were tested immediately after the onset of their psychotic episode, before they were admitted to the hospital or given medication.) There was no significant (I’ > 0.05) difference between patients’

134 B.R. Rund et al. /Psychiatry Research 69 (1997) 131-141

neuroleptic medication status in terms of ‘defined daily doses’ (DDD) (Norwegian Medical Depot, 1989) in the acute state (M= 1.57, SD. = 1.48) and the remission state (M= 1.23, S.D. = 0.62). Three patients were on anticholinergic medication in the acute phase, and two patients received anticholinergic medicines during remission.

Seven patients changed from inpatient to outpatient status during the assessment period. Five patients were in an after-care institution at both times of testing. Three patients were inpatients at both assessments. The patients were usually treated with a combination of neuroleptic medication (see above) and supportive psychotherapy.

The group of 14 non-psychiatric controls, se- lected among nursing staff, had no history of psychiatric illness. No significant differences appeared between the schizophrenic and control groups regarding sex, age or education. Table 1 summarizes some characteristics of the schizophrenic patients and the normal controls.

2.2. Measures

2.2.1. Cognitive assessment Cognition is assessed across a variety of tasks

with highly varied processing demands. The battery of cognitive laboratory measures (COGLAB) was mainly developed by W. Spaulding, Ph.D. (Spaulding et al., 1981, 1989; see also Penn et al., 19931 and is supplemented with two memory tasks in the present study. These measures are included

Table 1 Demographic and psychosocial characteristics of groups

in the test battery to achieve breadth of measurement across pre-attentional, attentional, and conceptual levels of cognition. Some test data are excluded because missing data or the presence of too many outliers suggested that technical errors had occurred.

The individual tests, presented in the order of their complexity, are: Backward Masking Task, Short-Term Memory Task, Long-Term Memory Task, and the Card Sorting Task. The test battery is presented on an Apple II E microcomputer with a color video monitor (Model Number M01081, equipped with a subject-response con- sole and special circuitry for millisecond-precision timing of stimuli and response latencies (Spauld- ing et al., 1981, 1989).

2.2.1.1. Backward masking. A standard stimulus duration procedure is used in which pairs of digits (target stimuli) are presented for 16.5 ms on the CRT screen. They are followed by a patterned mask, consisting of x’s, of equal duration. The visual angle for the stimuli is about 0.57” in the vertical dimension and the same in the horizontal dimension. A small arrow presented just above the target location provides a fixation point. A stimulus onset asynchrony (SOAs) procedure is used for target-masking intervals. SOA describes the interval between the relative onsets of the target and the mask. The task consists of 30 stimulus presentations: 10 of 16.5-ms SOA mask (short), 10 of 33-ms SOA mask (long), and 10 of no mask. The no-mask condition presumably

Characteristics Groups

Normal subjects (N = 14)

Schizophrenics (N= 15)

Sex. % male

Age at index, Y Education at index, Y Global psychosocial functioning (GAS) Overall symptom score (BPRS) at time 1 Overall symptom score (BPRS) at time 2 Premorbid adjustment Duration of previous hospitalization, months

57 13 M S.D. M 33.4 10.2 33.3 10.2 2.7 11.2

38 2.4 1.7

15.3 95.4

S.D. 14.8 2.9

12.1 0.4 0.3 5.5

97.9

B.R. Rund et al. /Psychiatry Research 69 (1997) 1.~1-141 135

measures the ability to apprehend and report briefly presented pairs of digits and is thus best characterized as a test of apprehension. Three measures derive from this test: a no-mask condition; a long SOA (33 ms) mask condition; and a short (16.5 ms) mask condition. Apprehension of each digit in the pair is scored separately, so the maximum score is 20 for each of the three conditions. (For more details, see Rund, 1993.)

2.2.1.2. Memory. The separation between short- and long-term memory tests is based on the modal approach proposed by Atkinson and Shiffrin (1968). This theoretical model distinguishes among kinds of memory according to the length of time information is stored. If information is active for a longer period than about 20 s, the long-term memory store is involved.

The aim of the procedure for the Short-Tetm Memory test (STM) is to measure information loss at various intervals (Peterson and Peterson, 1959). In the present procedure, consonant trigrams are presented for 1 s on the screen, and the subject is instructed to read the letters to make sure that the initial registration of the letters is correct. Immediately afterwards a three-digit number is presented, and the subject is instructed to read this number aloud and then count backwards by three. The retention interval is either 3 or 15 s, distributed randomly across the 20 trials. The maximum number of correctly recalled letters is 30 in each condition. Two measures are derived: one for the 3-s retention interval (STM 3), and one for the 15-s retention interval (STM 15). (More detailed information on the STM test is provided by Landro et al., 19931.

The Long-Term Mernoly test (LTM) is the ‘Con- tinuous Distractor Paradigm’developed by Bjork and Whitten (1974). The list of word pairs consists of 30 words divided into five trials, each containing three pairs. The first word pair is presented on the screen for 2 s, and the subject is instructed to read the words aloud. An interven- ing activity of counting backwards by threes is required for 20 s until the second pair is presented, which is also followed by counting backwards for 20 s until the third pair is presented. After the presentation of the third pair, the subject is again instructed to count backwards for 20

s after which the recall of all six words is required. Thus, the time intervals between presentation of stimulus and request for response are 20, 40 and 60 s, respectively. The scores are the total number of correctly recalled words divided by pairs presented as the 1. pairs, 2. pairs, and 3. pairs. (For more details on the LTM test, see Landro et al., 1993.)

2.2.1.3. Card Sorting Task (CST). This is a mod- ified computer version of the frequently used Wisconsin Card Sorting Test, a neuropsychological test of concept formation (Fey, 19511. The test demands that the subject discern three sorting parameters (color, number, and form) through trial and error. Auditory and visual feedback of ‘right’ or ‘wrong’ is provided for each card sorted. After a criterion number of five correct sorts, the relevant sorting parameter changes. The sorts need not be consecutive. (This is the main difference from the Wisconsin Card Sorting Test, where the sorting principle changes after 10 consecutive trials.) The CST yields two scores, random and perseuerative errors.

The latter are number of errors attributable to perseverative sorting by a principle that is no longer the correct one. For example, when the correct parameter changes from color to form, subsequent sorts using color increment the perse- veration score.

2.2.2. Clinical assessments All patients were evaluated for present psy-

chosocial functioning according to the Global As- sessment Scale (GAS) (Endicott et al., 19761, a rating scale with scores from 1 to 100.

Psychiatric symptoms were assessed with the Expanded Brief Psychiatric Rating Scale (Lukoff et al., 19861, a rating scale for 24 different psychiatric symptoms, with individual item scores rang- ing from 1 to 7. The overall symptom score is the total score for all the items that have been assessed divided by the number of items that have been scored.

Patients’ premorbid adjustment was rated on the UCLA Premorbid Adjustment Scale (Evans et al., 1972), which includes an assessment of ability to function within seven social areas.

136 B.R. Rund et al. /Psychiatry Research 69 (1997) 131-141

2.3. Procedure 3. Results

After subjects had provided written informed consent, they were tested individually with the cognitive test battery. The tests were presented in a standardized order: Backward Masking, Card Sorting Task, Short-Term Memory and Long- Term Memory. The test session lasted from 1 h and 20 min to 2 h, and the subjects were allowed to take one or more breaks between the individual tests if they felt tired. All inpatients and most outpatients were tested in a laboratory at Dike- mark Hospital, Oslo, Norway. A few outpatients were tested in an outpatient clinic office. The computer administering the test battery was oper- ated by two different psychologists and two psychology graduate students.

The group means and standard deviations for the 10 cognitive measures are shown in Table 2.

Results were evaluated in Group X Measure X

Time analyses of variance with repeated measures on the last factor. First, an overall analysis was done where all the submeasures of a task were summed in one score. Then, Group X Time ANOVAs with the submeasures treated as dependent variables were conducted. In the cases where a significant Time x Group effect was found, a t-test (planned comparison) was done to compare the schizophrenic and normal subjects at that measure at time 2 to examine if there was still a significant difference between the two groups in the phase of remission.

The diagnostic interview (KID), UCLA Pre- A repeated measures analysis of variance of morbid Adjustment Scale, GAS, and BPRS were the sum score of the Backward Masking measures all presented to the subjects after the cognitive yielded a main effect for Group (F = 4.04, d.f. = assessment, and in the order mentioned. They 1,27, P < 0.09, for Measure (F = 86.96, d.f.= were all done by an experienced psychologist. 2,54, P < O.OOl>, for Group X Measure (F = 3.19,

Table 2 Group means and standard deviations of the cognitive measures at time 1 and time 2 for non-psychiatric controls and schizophrenics

Group

BM. No mask

BM. Long mask

BM. Short mask

STM 3

STM 15

LTM 1. pair

LTM 2. pair

LTM 3. pair

CST. Random errors

CST. Perseverative errors

Mean S.D. Mean S.D. Mean S.D. Mean SD. Mean SD. Mean S.D. Mean S.D. Mean S.D. Mean SD. Mean S.D.

Time 1

Controls (N= 14)

19.71 0.83

11.93 7.29 5.21 4.51

25.43 3.06

16.43 5.42 6.71 1.98 4.43 2.47 5.79 1.97

12.14 11.88 12.79 8.11

Schiz. (N = 1.5)

17.67 4.34 7.40 4.75 4.67 2.94

17.00 7.31

11.93 6.27 4.53 3.54 2.87 2.36 2.80 2.98

31.73 19.10 17.80 9.77

Time 2

Controls (N = 14)

19.93 0.27

12.43 5.96 4.07 3.36

24.79 5.18

15.36 4.86 7.50 2.28 4.71 2.05 6.43 2.17 8.29 5.54

10.50 6.95

Schiz. (N= 15)

19.33 0.98 8.00 5.54 2.73 3.01

22.86 6.06

13.86 5.75 5.71 3.07 3.86 2.21 4.43 2.38

16.40 11.17 12.80 9.59

B.R. Rund et al. /Psychiatry Research 69 (1997) 1.11-141 137

d.f. = 254, P < 0.05>, and for Measure X Time (F = 183.88, d.f. = 2,54, P < 001).

A 2(Group) x 2(Time) ANOVA of the No musk measure revealed a Group effect (F = 4.47, d.f. = 1,27, P < 0.05>, but no Time or Time X Group effect. For the Long mask measure a main effect was found for Group (F = 4.63, d.f. = 1,27, P < 0.05) in the 2(Group) x 2(Time) ANOVA, but no effect for Time or Group X Time. The 2(Group) x 2(Time) ANOVA of the Short musk measure displayed an effect for Time (F = 5.37, d.f. = 1,27, P < O.OS>, but no effect for Group or for Group x

Time. A 2(Group) X ;?(Measure) X 2(Time) ANOVA

of the sum score of the Short-Term Memory data revealed a main effect for Group (F = 7.38, df = 1,26, P < O.Ol), for Time (F = 88.26, d.f. = 1,26, P < 0.0011, and for Group X Measure (F = 5.80, d.f. = 1,26, P < 0.05).

The 2(Group) X 2(Time) ANOVA of the STM 3 data showed a main effect for Group (F = 9.31, d.f. = 1,26, P < 0.005) and for Group X Time (F = 5.95, d.f. = 1,26, P < 0.021, while the effect of Time was at trend level (P = 0.06). When a 2(Group) x 2(Time) ANOVA was done with the STM 15 data, no significant main effects were found, neither for Group nor for Time.

The 2(Group) X 3(Measure) X 2(Time) ANOVA of the sum score of the Long-Term Memory data showed a main effect for Group (F= 7.54, d.f. = 1,26, P < O.Ol), for Time (F = 4.10, d.f. = 1,26, P < 0.05), and for Measure X

Time (F = 16.83, d.f. = 2,52, P < 0.001). A significant main effect for Group (F = 4.80,

d.f. = 1.26, P < 0.051, and an effect of Time at the trend level (P = 0.06) appeared when the LTM 1. pair data were analysed in a 2 (Group) X 2 (Time) ANOVA. The 2 (Group) X 2 (Time) ANOVA of the LTM 2. pair data yielded no significant effects, neither for Group, nor for Time. The 2 (Group) x 2 (Time) ANOVA of the LTM 3. pair showed a significant effect, for Group (F = 10.17, d.f.= 1.26, P< 0.004) and for Time (F= 6.48, d.f. = 1.26, P < 0.05).

Finally, the 2(Group) X 2(Measure) X 2(Time) ANOVA of the sum score of the CST data revealed a main effect for Group (F = 10.55, d.f. = 1,27, P < 0.003), for Measure (F= 7.12, d.f.=

1,27, P < O.Ol>, for Time (F= 6.18, d.f.= 1,27, P < 0.02), for Group X Time (F = 11.92, d.f. = 1,27, P < 0.0021, and for Measure x Time (F = 6,94, d.f. = 1,27, P < 0.01).

The 2(Group) X 2(Time) ANOVA of the CST score of random errors showed a significant effect for Group (F = 13.45, d.f. = 1,27, P < 0.001) and for Time (F = 10.25, d.f. = 1,27, P < 0.003). There was also a nearly significant Time x Group effect found (F = 3.67, d.f.= 2,27, P = 0.066). The 2(Group) x 2 (Time) analysis of variance of the perseueratiue errors CST score showed no significant main effects, neither for Group nor for Time.

The effect of age was controlled by using a covariance procedure. When the effects of age were partialled out in this way, none of the group effects reported above were changed.

4. Discussion

The group effects found in the analyses of variance indicate that all cognitive measures in the present study are cognitive deficits in schizophrenics, except four: the short mask condition on Backward Masking, the Short-Term Me- mory with a 15-s retention interval, the Long- Term Memory test 2. pair, and the CST score of perseverative errors.

The possible candidates for trait-dependent components are the no mask and the long mask condition on the backward masking procedure, and the 1. pair and 3. pair on the Long-Term Memory test. On these measures schizophrenics did significantly worse than the control subjects while there was no time by group interaction effect. At the LTM 3. pair, a time effect was found, which indicates a learning effect.

The data from the no-mask condition must be interpreted with caution since a ‘ceiling effect’ might have influenced the results.

Regarding the long mask condition, results concur with the findings of Spaulding et al. (1989) of deficient performance in schizotypal subjects compared to normal controls on the same backward masking measure. Both findings support the assumption that this is a trait-dependent factor in schizophrenia.

On 1. pair and 3. pair of the Long-Term Me-

138 B.R. Rund et al. / Psychiatv Research 69 (1997) 131-141

mory test, schizophrenics showed a stable performance. When all three long-term measures were combined, a similar pattern appeared. This suggests that a long-term memory deficit is a trait- dependent component in schizophrenia. A time effect in the overall variance analysis indicates that a certain practice effect exists.

Only one cognitive measure in the present study can with a certain amount of reliability be classified as a purely episode-linked factor: the Short- Term Memory @TM) test with a 3-s retention interval. Schizophrenics perform significantly worse than normal subjects, but they improve significantly from Time 1 to Time 2. At Time 2, there is no longer a significant difference between the two groups’ performance levels. On the STM with the long retention interval (15 s), a similar performance pattern appears. However, neither the group effect nor the time effect reaches a level of significance.

Verbal memory deficits in schizophrenia have been attributed to organizational disturbance, re- sulting in ineffective encoding (Koh, 1978). In this way,the observation of verbal recall deficit and associative disturbance have been connected. The present finding of the Short-Term Memory test with 3-s retention interval as an episode-linked factor supports this. The process involved is probably closely related to the efficiency of encoding and thus influenced by psychotic ideation and associative disturbance. The present findings suggest that recall deficits related to longer retention intervals may be more enduring features of an amnesia-like nature in schizophrenia (cf. Saykin et al., 1991; Landra et al., 1993).

Deficits on the two card sorting measures can be characterized as mediating factors; i.e. they are more prominent in the acute psychotic state, but do not completely disappear during remission states. Regarding the score of random errors, the Group by Time effect was nearly significant. The perseverative error score can possibly also be classified as a mediating factor, but the group effect does not reach a level of significance. The lack of a group effect on this measure is a finding that varies from some other reports (Green et al., 1992; Franke et al., 1992). The reason might be that there are procedural variations and to some

extent also different scoring systems in our study as compared to others.

When the two card sorting scores were combined, however, a significant group effect appeared. This combined score is probably the one most comparable to the total error score in the original Wisconsin Card Sorting Test. Thus, our finding concurs with previous reports on this point, including a study by Goldberg et al. (1990). On the two card sorting measures, normal controls improve from Time 1 to Time 2, although not as much as the schizophrenics. Thus, a certain degree of learning effect cannot be ruled out.

The finding that most cognitive deficits are more prominent in the acute psychotic phase yet do not disappear when the symptoms diminish, demonstrates that schizophrenia is an illness that varies in conspicuousness and distinctiveness from one period to another. The schizophrenia illness, even for shorter periods of time, very seldom disappears (Rund, 1990; Rund and Torgalsboen, 1990).

A puzzling finding in the present study is that the schizophrenics perform significantly better in the acute state compared to the partial remission state in the short backward masking condition. On the long backward masking condition, the schizophrenics perform at the same level in the two phases (no significant difference). One explanation is that the type of processing required in a backward masking task occurs faster in the acute phase compared to the remission phase. Most likely this is due to a higher activity in the dopaminergic system during the acute phase. With a long interval between stimulus and mask, the schizophrenics have enough time to process the stimulus adequately in both states. In the short interval condition, however, schizophrenics’ processing is too slow (the rate of conductance is too slow) to process the stimulus adequately when in a state of remission with low dopaminergic activity. However, this is a highly speculative explanation.

Another possible explanation of the backward masking results, not necessarily independent from the one outlined above, is that there is a floor effect for the short SOA masking condition, and that the long SOA condition is a better discrimi-

B.R. Rund et al. /Psychiatry Research 69 (1997) 131-141 139

nator between the two subject groups. If this is the case, the two conditions may simply reflect different psychometric properties, and the findings be explained as psychometric artifacts (see Chapman and Chapman, 1978; Rund, 1983).

It should also be noted that the computerized version of the backward masking procedure suf- fers from a certain lack of precision. The Apple II system (with the time card) only allows presentation of stimuli in 16.5ms increments. These technical shortcomings might also have restricted information concerning the backward masking results (see Rund et al., 1996).

Research on cognitive functions in psychotic patients must consider the possible effects of neuroleptic and anticholinergic drugs. Adequate control of this variable in research is difficult (Chapman and Chapman, 1977; Neale and Olt- manns, 1980; Spohn and Fitzpatric, 1980). The effect of antipsychotic medication was not controlled for in the present study. Previous reports indicate that, generally, medication improves ba- sic information processing, rather than decreases efficiency (see Saccuzzo and Braff, 1986; Spohn and Strauss, 1989). A recent study suggests that cognitive functioning in schizophrenic patients was stable across changes in medication status (Cannon et al., 1994). Our data indicate that on most measures unmedicated patients did a little worse than patients on neuroleptic medication at Time 1.

In cross-sectional comparisons, observed differences between schizophrenics and normal subjects would probably have been strengthened if the patients had been unmedicated.

For a definitive examination of medication effects, medication would have to be varied while keeping clinical state constant. However, the question of primary interest in this study is whether cognitive performance changes from one clinical state to another within the group of schizophrenics. Although medication status and performance are correlated for the three patients not receiving drugs at Time 1, it does not seem reasonable that this sub-sample should account for the general trends. For most patients, medication status did not change significantly during the

period in question. This is also an argument against medication representing a serious artifact.

The present study has other methodological shortcomings besides the possible medication effects. One is the small number of cases, while another is the issue of practice and fatigue effects. Changes in test performance from the first to the second testing might be due to learning effects. This effect first of all applies to the Card Sorting Task (Green et al., 1992). The performance of the two subject groups might have been differently influenced by the exertion (fatigue) of participating in a comprehensive test session.

In summary, schizophrenics’ performances were relatively stable over clinical phases on four of the five cognitive measures. This suggests that they are trait-dependent factors. Only a short- term memory measure was found to be a possible episode-linked factor: schizophrenics improving significantly from the acute to the remission phase. All the other measures examined might best be characterized as mediating factors. Our findings indicate that schizophrenic patients may only re- cover partially in all respects including that of cognitive functions. Future research in this area with larger samples of patients and clearer dis- tinction between clinical phases is warranted.

Acknowledgements

We would like to thank Dag Erik Eilertsen, MA, who has helped in statistical analyses of data.

This research project was supported by Grant 377.91/005 from the Norwegian Research Coun- cil for Science and the Humanities, and by funds from Dikemark Hospital.

References

American Psychiatric Association (1987) Diqqmstic and Stati.c- tical Manual of Mental Disorders, 3rd ed. rev. American

Psychiatric Press, Washington, DC

Asarnow. R.F., Granholm, E. and Sherman, T. (199 1) Span of

apprehension in schizophrenia. In: Steinhauer. S.R, Gruze-

her, J.H. and Zubin, J. (Eds.). Handbook of Schizophrenia, Vol. 5: Neuropsychology, Psychophysiology and Information Processing. Elsevier, Amsterdam, pp. 3X-370.

Asarnow. R.F. and MacCrimmon. D.J. (1981) Span of apprc-

140 B.R. Rund et al. /Psychiatry Research 69 (1997) U-141

hension deficits during postpsychotic stages of schizophre-

nia. Arch Gen Psychiatry 38, 1006-1011. Atkinson, R.G. and Shiffrin, R.M. (1968) Human memory: A

proposed system and its control processes. In: Spence K.W.

(Ed.), 7’he Psychology of Learning and Motivation. Advances in Research and Theory. Vol. 2. Academic Press, New York,

pp. 89-195.

Bjork, R.A. and Whitten, W.B. (1974) Recency-sensitive re-

trieval processes in long-term free recall. Cogn Psychol 6, 173-189.

Braff, D.L. (1993) Information processing and attention dys-

functions in schizophrenia. Schizophr Bull 19, 233-259.

Braff, D.L., Saccuzzo, D.P. and Geyer, M.A. (1991) Informa-

tion processing dysfunctions in schizophrenia: Studies of

visual backward masking, sensorimotor gating, and habitua-

tion. In: Steinhauer, S.R., Gruzelier, J.H. and Zubin J.

(Eds.), Handbook of Schizophenia, Vol. 5: Neuropsychologv, Psychophysiology and Information Processing. Elsevier, Am-

sterdam, pp. 303-334.

Cannon, T.D., Zorilla, L.E., Chtasel, D., Gur, R.E., Gur, R.C.,

Marco, E.J., Moberg, R. and Price R.A. (19941 Neuropsy-

chological functioning in siblings discordant for schizophre-

nia and healthy volunteers. Arch Gen Psychiatry 51, 651-661.

Chapman, L.J. and Chapman, J.P. (1977) Selection of subjects

in studies of schizophrenic cognition. J Abnorrn Psychol 86, 10-15.

Chapman, L.J. and Chapman, J.P. (1978) The measurement of

differential deficit. J Psychiatr Res 14, 303-311.

Cromwell, R. and Spaulding, W. (1978) How schizophrenics

handle information. In: Fann, W.E. and Williams, R.L.

(Eds.), The Phenomenology and Treatment of Schizophrenia. Spectrum Press, New York, pp. 127-162.

Elliot, R. and Sahakian, B.J. (1995) The neuropsychology of

schizophrenia - relations with clinical and neurobiological

dimensions. Psycho1 Med 25, 581-594.

Endicott, J., Spitzer, R.L., Fleiss, J. and Cohen, J. (19761 The

Global Assessment Scale: a procedure for overall severity

of psychiatric disturbance. Arch Gen Psychiatry 33,766-771.

Evans, J., Rodnick, E., Goldstein, M. and Judd, L. (1972)

Premorbid adjustment, phenothiazine treatment and remis-

sion in acute schizophrenics. Arch Gen Psychiatry 27, 486-490.

Fey, E.T. (1951) The performance of young schizophrenics

and young normals in the Wisconsin Card Sorting Test. J

Consult Psycho1 15, 311-319. Franke, P., Maier, W., Hain, C. and Klingler, T. (1992) Wis-

consin Card Sorting Test: an indicator of vulnerability?

Schizophr Res 6, 243-249. Goldberg, T.E., Ragland, J.D., Torrey, E.F., Gold, J.M.,

Bigelow, L.B. and Weinberger, D.R. (19901 Neuropsycho-

logical assessment of monozygotic twins discordant for

schizophrenia. Arch Gen Psychiatry 47, 1066-1072. Goldberg, T.E., Hyde, T.M., Kleinman, J.E. and Weinberger,

D.R. (1993) Course of schizophrenia: Neuropsychological

evidence for a static encephalopathy. Schizophr Bull 19, 798-804.

Green, M., Satz, P., Ganzell, S. and Vaclav, J.F. (1992) Wis-

consin Card Sorting Test performance in schizophrenia:

remediation of a stubborn deficit. Am J Psychiatry 149, 62-67.

Kietzman, M.L. (1991) Information processing and

schizophrenia. In: Steinhauer, S.R., Gruzelier, J.H. and

Zubin, J. (Eds.), Handbook of Schizophrenia, Vol. 5: Neu- ropsychology, Psychophysiology and Information Processing. Elsevier, Amsterdam, pp. 227-239.

Knight, R.A. (1984) Converging models of cognitive deficit in

schizophrenia. Nebraska Symposium on Motivation, Vol.

31. In: Spaulding, W.D. and Cole, J.K. (Eds.), Theories of Schizophrenia and Psychosis. University of Nebraska Press,

Lincoln, pp. 93-156.

Koh, SD. (1978) Remembering of verbal material by young

adults. In: Schwartz, S. (Ed.), Language and Cognition in Schizophrenia. Erlbaum, Hillsdale, NJ, pp. 59-99.

Landrg, N.I., &beck, A.L. and Rund, B.R. (1993) Memory

functioning in chronic and non-chronic schizophrenics, af-

fectively disturbed patients and normal controls. Schizophr Res 10, 85-92.

Lewine, R.R.J. (1984) Stalking the schizophrenia marker: Evi-

dence for a general vulnerability model of psychopathology.

In: Watt, N.F., Anthony, J., Wynne, L.C. and Rolf, J.E.

(Eds.), Children at Risk for Schizophrenia. A Longitudinal Perspective. Cambridge University Press, Cambridge, pp.

545-571.

Lukoff, D., Liberman, R.P. and Nuechterlein, K.H. (1986al

Symptom monitoring in the rehabilitation of schizophrenic

patients, Schizophr Bull 12, 578-559.

Lukoff, D., Nuechterlein, K.H. and Ventura, J. (1986b) Ap-

pendix. Manual for Expanded Brief Psychiatric Rating

Scale. (BPRS). Schizophr Bull 12, 594-602.

Neale, J. and Oltmanns, T.F. (1980) Schizophrenia. Wiley, New

York.

Nopoulos, P., Flashman, L., Flaum, M., Arndt, S. and An-

dreasen, N. (1994) Stability of cognitive functioning early in

the course of schizophrenia. Schizophr Res 14, 29-37.

Nuechterlein, K.H. and Dawson, M.E. (1984a) Information

processing and attentional functioning in the developmen-

tal course of schizophrenic disorders. Schizophr Bull 10,

160-203.

Nuechterlein, K.H. and Dawson, M.E. (1984b) A heuristic

vulnerability/stress model of schizophrenic episodes.

Schizophr Bull 10, 300-312.

Nuechterlein, K.H., Dawson, M.E., Gitlin, M., Ventura, J.,

Goldstein, M.J., Snyder, K.S., Yee, C.M. and Mintz, J.

(1992) Developmental processes in schizophrenic disorders:

longitudinal studies of vulnerability and stress. Schizophr Bull 18, 387-425.

Penn, D.L., van der Does, H.J.W., Spaulding, W.D., Garbin,

C., Linszen, D. and Dingemans, P. (1993) Information

processing and social cognitive problem solving in

schizophrenia: assessment of interrelationships and changes

over time. J Nerv Ment Dis 181, 13-20.

Peterson, L.R. and Peterson, M.J. (1959) Short-term retention

of verbal items. J Exp Psychor 58, 190-198.

B.R. Rund et al. /PTchiatry Research 69 (1997) 131-131 141

Rund, B.R. (1983) The effect of distraction on focal attention

in paranoid and non-paranoid schizophrenic patients com-

pared to normals and non-psychotic psychiatric patients. .I

Psychiatr Res 17, 241-250. Rund. B.R. (1990) Fully recovered schizophrenics: a retrospec-

tive study of some premorbid and treatment factors. Psy- chiarry 53, 126~139.

Rund, B.R. (1993) Backward-masking performance in chronic

and non-chronic schizophrenics, affectively disturbed

patients, and normal control subjects. J Abnorm Psycho/ 102. 74-81.

Rund, B.R. and Landro, N.I. (1990) Information processing: a

new model for understanding cognitive disturbances in

psychiatric patients. Acru Pgxhiatr &and 81, 305-316. Rund. B.R.. Landrca, N.I. and Brbeck, A.L. (1993) Stability in

backward masking performance in schizophrenics, affec-

tively disturbed patients, and normal subjects. J Neru Menr

Dis 181, 233-237.

Rund. B.R. and Torgalsbtien, A-K. (1990) Fully recovered

schizophrenics compared to chronic patients on premorbid

and treatment characteristics. Psychiatry Psychobiol 5, 113~121.

Rund. B.R., 0ie, M. and Sundet, K. (1996) Backward-masking

deticit in adolescents with schizophrenic disorders or atten-

tion deficit hyperactivity disorder. Am J Psychiatry 153, I IS- 1157.

Saccuzzo. D.P. and Braff, D.L. (1986) Information-processing

abnormalities: Trait- and state-dependent components.

S&i-_ophr Bull 12. 477 -4.59.

Saykin, A.J., GUI, R.C., Gur, R.E., Mozley. P.D., Mozley, L.H.,

Resnick, S.M., Kester, D.B. and Stafiniac, P. (1991) Neu-

ropsychological functioning in schizophrenia. Selective im-

pairment in memory and learning. Arch Cm Psychiany 48. 618-624.

Schrout, P.E. and Fleiss, J.C. (1979) Intraclass correlations

uses in assessing rater reliability. Psycho/ Bull 86, 42Ok428. Spaulding, W., Garbin, C. and Dras. S.R. (1989) Cognitive

abnormalities in schizophrenic patients and schizotypal co-

lege students. J Nerv Ment Dis 177, 717--72X.

Spaulding, W., Hargrove, S., Crinean. W. and Martin. T.A.

(1981) A nlicrocomputer-based laboratory for psy-

chopathology research in rural settings. B&L’ Res Mcf Itrsff

13. 616-623.

Spitzer, R.L. and Williams, J.B.W. (1984) Smuturcrl Clinical

Interuiewfor I)SM III (SCID). Biometrics Research Depart-

ment, New York State Psychiatric Institute. New York.

Spohn, H.E. and Fitzpatric, T. (1980) Informed consent and

bias in samples of schizophrenic subjects at risk for drug

withdrawal. J Abnorm Psycho/ 89, 79 -92. Spohn, H.E. and Strauss, M.E. (1989) Relation of neurolcptic

and antichollnergic medication lo cognitive function in

schizophrenia. f Abnorm Pyhol 98. 367 -3X0. Sweeney. J.A., llaas. G.L.. Keilp. J.G. and Long. M. (1991)

Evaluation OI the stability of neuropsychological function-

ing after acute episodes of schizophrenia: one-year fol-

lowup study. Psychiafy Res 3X, 63.-70. Zubin, J. and Spring. B. (1977) Vulnerability ~~ A new view of

xhizophrenia. J Abnorm I?y~c~hol 86. IO+ IX

Stability in cognitive dysfunctions in schizophrenic patients

Documents