Comparison of the continuous performance test with and without working memory demands in healthy controls and patients with schizophrenia

Comparison of the continuous performance test with and without working memory demands in healthy controls and patients with schizophrenia

Schizophrenia Research 48 (2001) 307±316 www.elsevier.com/locate/schres Comparison of the continuous performance test with and without working memor...

154KB Sizes 1 Downloads 16 Views

Schizophrenia Research 48 (2001) 307±316

www.elsevier.com/locate/schres

Comparison of the continuous performance test with and without working memory demands in healthy controls and patients with schizophrenia Matthew M. Kurtz*, J. Daniel Ragland, Warren Bilker, Ruben C. Gur, Raquel E. Gur Brain-Behavior Laboratory, Department of Psychiatry, 10th Floor Gates Building, University of Pennsylvania School of Medicine, 3400 Spruce Street, Philadelphia, PA 19104, USA Received 26 November 1999; accepted 27 March 2000

Abstract The Penn Continuous Performance Test (PCPT), a measure of sustained visual attention developed for use in functional neuroimaging studies, was compared with a standard CPT developed by Gordon Diagnostic Systems (GDS; Vigilance subtest). The PCPT and the GDS CPT were administered with a standard neuropsychological battery to 68 healthy adults to assess reliability and construct validity. The test had adequate internal consistency, and convergent validity was established through signi®cant correlations between measures of ef®ciency on the PCPT and the GDS CPT. With the exception of a signi®cant correlation between ef®ciency measures on the GDS CPT and a measure of auditory sustained attention, neither version of the CPT correlated signi®cantly with other measures in the battery. Factor analysis showed that the PCPT loaded with the GDS CPT. In 39 patients with schizophrenia and 39 matched, healthy controls, equivalent impairment was evident on the two CPT tasks. Neither version correlated signi®cantly with symptom measurements. These results support previous conclusions that sustained visual attention in schizophrenia is a core information processing de®cit, not directly related to symptomatology. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Construct validity; Continuous Performance Test; Schizophrenia; Symptom scales

1. Introduction Attentional dysfunction is one of multiple cognitive domains impaired in schizophrenia (e.g. Blanchard and Neale, 1994; Gur et al., 1997; Heinrichs and Zakzanis, 1998). De®cits in sustained visual attention, as measured by accuracy on the Continuous Performance Test (CPT), were ®rst described in schizophrenia in the 1960s (Orzack and Kornetsky, * Corresponding author. Tel.: 11-215-662-2890; fax: 11-215662-7903. E-mail address: [email protected] (M.M. Kurtz Ph.D.).

1966; Orzack et al., 1967), and subsequent studies have revealed that CPT de®cits are present when patients are actively psychotic and when in remission (Asarnow and MacCrimmon, 1978). The stability of these impairments has led some to hypothesize that impaired sustained visual attention represents a trait or core information processing de®cit, which serves as a mediator of the pathophysiology of schizophrenia (Nuechterlein and Dawson, 1984). Measures of visual sustained attention have been identi®ed as endophenotypic markers in unaffected ®rst-degree relatives of probands with schizophrenia (Cornblatt and Keilp, 1994). Different forms of the CPT have been utilized as

0920-9964/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0920-996 4(00)00060-8

308

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

measures of attentional dysfunction in schizophrenia. The earliest versions of the CPT, developed to assess attentional impairment in brain injury, consisted of an easy `X' version and a harder `AX' version (Rosvold et al., 1956). A sequential series of single letters was visually presented at a rate of one item per second, for a total of 10 min. The subject was asked to respond every time the letter X appeared among a string of random letters. True positives and false positives were recorded. A more dif®cult `A±X' version required the subject to respond as quickly as possible only when an X occurred after an A. In this modi®ed CPT, the subject was required to discriminate the presented stimulus and retain in working memory (Baddeley, 1992) the previously presented stimulus in order to perform correctly (Rosvold et al., 1956). This approach is currently implemented using digits (1 and 9 instead of A and X) in the Gordon Diagnostic Systems (GDS) CPT (Gordon, 1986). Studies using these earlier versions of the CPT were not sensitive to de®cits in genetically predisposed, unaffected ®rst-degree relatives, largely because the tasks were too easy to detect subtle de®cits (Asarnow et al., 1977; Cornblatt and Erlenmeyer-Kimling, 1984). To increase dif®culty level, several variants of the original `X' and `AX' version were developed. The two most widely used are as follows. (1) The degraded stimulus CPT developed by Nuechterlein et al. (1983). In this task, target stimuli are blurred and visual noise is superimposed on targets in the form of a plus sign. This produces a more dif®cult task by increasing perceptual demands. (2) The CPT identical-pairs (CPT-IP) version developed by Cornblatt et al. (1988). In the CPT-IP, the target item is the second stimulus of any pair of identical stimuli. This version is more dif®cult, as stimuli are presented for very brief 50 ms intervals (one stimulus per second) and the subject is asked to maintain previous stimulus presentations actively in working memory in order to detect a `match' (Cornblatt et al., 1988). Subsequent studies using these more dif®cult CPT versions have revealed de®cits in patients and `at-risk' relatives (Cornblatt and Keilp, 1994). Several studies investigated the relationship of impaired CPT performance to symptom expression in schizophrenia. Strauss et al. (1993) examined the relationship between the CPT and the Span of Apprehension Test (SPAN) to clinical symptoms as

measured by the Brief Psychiatric Rating Scale (BPRS; Overall and Gorham, 1980) in 50 outpatients with schizophrenia or schizoaffective disorder. The CPT, but not SPAN performance, was modestly correlated with measures of thought disorder, whereas SPAN but not CPT performance correlated signi®cantly with negative symptoms as measured by BPRS factor two (emotional withdrawal, motor retardation, and blunted affect). This latter ®nding appeared to be mediated by general levels of verbal ability. In a related study, Buchanan et al. (1997) investigated the relationship of CPT and SPAN performance comparing de®cit with non-de®cit subtypes. The study revealed differences in accuracy between 20 de®cit and 56 non-de®cit patients on the CPT, but not the SPAN task. The authors concluded that de®cit patients are uniquely characterized by impaired attention allocation or visuo-perceptual organization. Not all studies, however, have documented a relationship between symptoms and vigilance. Addington and Addington (1998) investigated the relationship of CPT and SPAN task performance 3 months after a psychotic episode to symptom measurements (PANSS; Kay, 1991), 12 months after the same psychotic episode. Results failed to identify a signi®cant relationship between CPT and SPAN performance and either positive or negative symptom scale ratings in 40 patients with schizophrenia. A possible factor that may explain these discrepant ®ndings is that previous efforts to increase CPT dif®culty also introduced additional processing demands. Thus, the Nuechterlein et al. (1986) paradigm increases the perceptual complexity and effort of the stimuli, whereas both the CPT-AX and CPT-IP versions increase demands on working memory. Therefore, de®cits in patients or family members may also re¯ect dif®culties in perceptual processing or working memory. The Penn CPT (PCPT) was designed to be more dif®cult than CPT-X without increasing demands on perceptual systems or working memory. In this task, the subject is asked to respond to a set of vertical and horizontal lines whenever they form a digit. Thus, each judgment is made on the basis of the immediate stimulus, not on the basis of a previously presented stimulus that must be maintained in working memory. The goals of the present study were as follows. (1) To

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

assess construct validity of the PCPT in healthy people. We compared performance on the PCPT and GDS CPT and examined the relationship of these variables to performance in other neurocognitive domains (Study 1). (2) To evaluate the performance of patients with schizophrenia on these two versions of the CPT and to examine whether it correlated with severity of positive or negative symptoms and duration of illness (Study 2).

2. Methods 2.1. Participants Healthy people and patients with schizophrenia were drawn from an on-going longitudinal investigation of behavior and brain function at the Schizophrenia Research Center of the University of Pennsylvania. After informed consent was obtained, participants underwent standard comprehensive screening and assessment procedures (Gur et al., 1991; Shtasel et al., 1991). For healthy people this included administration of the non-patient edition of the Structured Clinical Interview (SCID-NP; Shtasel et al., 1991; Spitzer et al., 1996b). For patients this included the patient edition of the SCID (SCID-P; Spitzer et al., 1996a) for Diagnostic and Statistical Manual of Mental Disorders (DSM-IV). Scales included the Brief Psychiatric rating Scale (BPRS; Overall and Gorham, 1980), the Scale for Assessment of Positive Symptoms (SAPS; Andreasen, 1983), and the Scale for Assessment of Negative Symptoms (SANS; Andreasen, 1984). These were administered by investigators trained to a criterion reliability of 0.90, intraclass correlation (Shtasel et al., 1992). For all participants detailed medical history, physical examination, and laboratory tests were obtained. Entry criteria to the MHCRC for healthy people included: (a) no history of a medical illness that might affect brain function; (b) no history of a DSM-IV Axis I psychiatric disorder (American Psychiatric Association, 1994); (c) no neurological disorder (e.g. epilepsy, migraine, head trauma with loss of consciousness); (d) no ®rst-degree relatives with a diagnosis of schizophrenia or affective illness. For patients, entrance criteria included: (a) a diagnosis of schizophrenia or schizophreniform disorder by DSM-IV criteria; (b) no concomitant Axis I or II disorder including past or present substance

309

abuse or dependence; (c) no neurological disorder (as for controls). The sample included 68 healthy volunteers and 39 patients with schizophrenia. The groups were similar in ethnic diversity to each other and to the community. Study 1 examined the 68 healthy controls (29 males and 39 females). They had a mean (^SD) age of 25.9 ^ 5.6 years and an average of 15.4 ^ 2.1 years of education. For Study 2, a subsample of 39 of the healthy subjects was matched with a sample of 39 patients with schizophrenia by age and sex. As a disease of early onset, schizophrenia may affect educational levels (Resnick, 1992). Thus, patients and healthy people were also matched on parental education (see Table 1). Some 13 of the 39 patients were medicated at the time of testing. For patients previously exposed to medication, mean typical neuroleptic dose was 367.6 ^ 325.5 mg (range: 106.4±1209.6 mg) chlorpromazine equivalent units per day, whereas mean atypical neuroleptic dose was 501.0 ^ 430.4 mg (range: 8.29±1106.6 mg) chlorpromazine units per day. 2.2. Test procedures 2.2.1. GDS CPT Vigilance subtest In the GDS CPT Vigilance task, a series of numbers are shown serially on a front display. Analogous to the original `AX' version of the CPT, in the GDS `1±9' CPT, the participant is asked to respond as quickly as possible when the number `1' is followed by the number `9'. True positives, false alarms and reaction time are recorded as dependent measures. There is a Table 1 Demographic and clinical characteristics of patient and control samples a Controls n 39 Male (n) 30 Female (n) 9 Age (years) 25.6 (5.4) Education (years) 15.2 (1.8) Parental education (years) 13.6 (3.6) Age at onset N/A Number of hospitalizations N/A a

Patients

P value

39 ± 32 ± 7 ± 27.0 (7.4) NS 12.6 (1.6) 0.000 01 13.8 (3.2) NS 22.3 (5.5) ± 3.6 (7.9) ±

Data are mean (^SD). The samples differed on education level (t ˆ 6.44, P , 0.000 01). NS: non-signi®cant at the P . 0.05 level. N/A: not applicable.

310

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

total of 30 target sequences out of a total of 360 trials. Trials are divided into three blocks consisting of 120 stimuli and ten target sequences each. The GDS CPT records the number of correct presses, omission and commission errors for both the total test as well as each of the three blocks of trials. Correct responses represent visual vigilance and memory, whereas false positive responses (commissions) are considered indicative of impulsivity. Speed of information processing is measured by reaction time to target sequences. Numbers are presented at a rate of one per second and are exposed for 200 ms each. The test takes approximately 6 min to complete. Instructions were standardized and administered from the GDS CPT test manual (Gordon, 1986). 2.2.2. PCPT The PCPT is administered by computer using the Power Laboratory platform (Chute and Westall, 1997). In this task, the participant was asked to respond to a set of vertical and horizontal lines _whenever the lines formed a digit. Thus, each judgment is based on the present, rather than the previous, stimulus and hence does not rely on working memory. Correct responses and reaction time to targets were recorded. Stimuli consisted of sevensegment displays. On target trials these segments formed digits 2, 4 and 5, whereas on distractor

Fig. 1. Illustration of two stimuli from the PCPT. The stimulus in the left panel is a target item. The stimulus in the right panel is a distractor.

trials the segment display created a nonsense con®guration (see Fig. 1). There were 36 targets randomly presented among a total of 360 items. Stimuli were presented at a rate of one per second and were exposed for 300 ms each. Test instructions were read to the participant from the computer screen: ªYou will see some lines on the screen. At times, these lines will form complete numbers. Whenever you see a number, press the `1' key as quickly as you can. First we will do some lines for practice.º The task cycles through two targets and several distractor items, before proceeding to the actual task. The PCPT takes 7 min to complete. The order of presentation of the PCPT and GDS CPT was counterbalanced across subjects. The tests were administered with a battery of standardized neuropsychological tests (Saykin et al., 1991; Censits et al., 1997). 2.3. Statistical analysis For all analyses, an ef®ciency measure was used as an index of performance on the two CPT versions. Performance re¯ects a trade-off between accuracy and speed. The advantage of this measure is that it combines both accuracy and speed of responding into a single measure. As the performance of healthy adults on the CPT is highly accurate, the use of an ef®ciency measure further differentiates subject scores by including variance attributable to speed of processing. Including reaction-time measures is of particular importance, as several CPT studies, in other neuropsychiatric disorders, have suggested that slowed reaction time may be more sensitive to de®cits in information processing than measures of accuracy (Zametkin et al, 1990; Holdnack et al, 1995). Ef®ciency was calculated for each subject as a ratio of number of correct responses per unit time by dividing the number of true positives (hits) by average reaction time on correct responses (in milliseconds). Because this variable is a proportion, an arcsine or p angular transformation …angle ˆ arcsin proportion† was performed (Snedecor and Cochran, 1980). 2.3.1. Study 1 Reliability estimates for performance measures on the PCPT were calculated using the Spearman± Brown split-half procedure (Allen and Yen, 1979)

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

311

Table 2 Mean scores and Pearson intercorrelations of neuropsychological test variables in healthy controls a Variables

Mean (SD)

1

2

3

4

5

6

7

8

9

10

11

1. PCPTEFF 2. CPTEFF 3. RHY 4. WCSTCAT 5. WCSTPE 6. TAP 7. JOLO 8. BD 9. STEREO 10. MAEVN 11. VOC

0.27 (0.03) 0.26 (0.02) 27.8 (2.1) 6.5 (2.0) 10.1 (5.0) 44.9 (4.8) 25.4 (3.3) 34.9 (10.4) 7.0 (2.2) 54.3 (4.5) 55.1 (9.8)

± 0.35 2 0.13 0.00 0.03 0.05 2 0.17 2 0.07 0.15 2 0.08 2 0.18

± 0.26 2 0.14 0.09 0.23 0.13 0.05 2 0.01 0.10 0.05

± 2 0.04 0.06 0.12 0.13 0.22 2 0.13 0.11 0.18

± 2 0.86 0.03 2 0.18 0.14 2 0.16 0.15 0.20

± 0.06 0.16 2 0.07 0.22 2 0.21 2 0.24

± 0.05 0.06 2 0.12 0.05 2 0.01

± 0.43 0.05 0.26 0.29

± 2 0.08 0.45 0.55

± 2 0.26 2 0.09

± 0.60

±

a

PCPTTP: Penn Continuous Performance Test, ef®ciency score; CPTTP: Gordon Diagnostics' Continuous Performance Test, ef®ciency score; RHY: Seashore Rhythm Test; WCSTCAT: Wisconsin Card Sorting Test, Categories; WCSTPE: Wisconsin Card Sorting Test Perseverative Errors; TAP: Finger Tapping; JOLO: Judgment of Line Orientation; STEREO: Stereognosis, total time; BD: Wechsler Adult Intelligence Scale-Revised, Block Design subtest; MAEVN: Visual Naming subtest of the Multilingual Aphasia Examination; VOC: Wechsler Adult Intelligence Scale-Revised, Vocabulary subtest. Correlations of P , 0.05 are represented in boldface.

for estimating internal consistency. Construct validity (Cronbach and Meehl, 1955) was evaluated by correlating performance on the PCPT with performance on other tests in the battery. Correlations with other attentional tests (the Vigilance subtest of the GDS CPT and Seashore Rhythm Test) were considered evidence of convergent validity, whereas lower correlations with executive (WCST Categories and Perseverative Errors), visuospatial (Judgment of Line Orientation, Block Design) and sensory-motor (Finger Tapping, Stereognosis) tasks were considered evidence of divergent validity. 2.3.2. Study 2 The statistical analysis comparing patients and healthy controls tested the following hypotheses. (1) Patients with schizophrenia perform more poorly than healthy participants on the Vigilance subtest of the GDS CPT and the PCPT. (2) PCPT and CPT performance correlates with severity of clinical symptoms and duration of illness. To test the ®rst hypothesis a 2 £ 2 mixed-design ANOVA was used with diagnosis as a grouping factor and test version as a withinsubject (repeated measures) factor, and the ef®ciency scores as the dependent measure. The second hypothesis was tested with Pearson correlations between the ef®ciency scores for both CPT measures and indices of clinical symptomatology, including the age of onset, duration of illness, BPRS total score, and

global ratings from the SANS (affective ¯attening, attention, alogia, avolition/apathy, and anhedonia) and SAPS (hallucinations, delusions, bizarre behavior, and formal thought disorder). 3. Results 3.1. Study 1 Performance of healthy participants on the CPT and other tests in the battery is presented in Table 2. The number of variables was limited to 11 to achieve a recommended 5:1 participant-to-variable ratio for factor analysis (Gorsuch, 1974), and to contain Type I (experimenter-wise) error. 3.1.1. Reliability Internal consistency for the PCPT was calculated using the Spearman±Brown split-half reliability formula (Allen and Yen, 1979) for the 68 healthy controls. This yielded a high correlation coef®cient of 0.89, indicating acceptable levels of internal stability. 3.1.2. Correlations between the PCPT, GDS CPT and standard neuropsychological measures PCPT ef®ciency scores were correlated with ef®ciency scores on the Gordon Diagnostic CPT

312

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

Table 3 Principal-component factor analysis of neuropsychological variables in healthy controls: Varimax transformation matrix Test variable a

PCPT CPT RHY WCSTCAT WCSTPE TAP JOLO BD STEREO MAEVN VOC

Factor b 1. Intellectual

2. Executive

3. Attention

0.15 2 0.13 2 0.23 2 0.09 0.11 0.04 2 0.65 2 0.80 0.02 2 0.73 2 0.81

0.08 2 0.15 2 0.15 0.92 2 0.93 2 0.02 2 0.38 0.03 2 0.30 0.21 0.20

2 0.85 2 0.72 0.01 0.01 2 0.03 2 0.24 0.04 2 0.02 2 0.31 0.01 0.07

with any of the measures from the neuropsychological battery. Thus, the pattern of correlations between the PCPT, the GDS CPT and other neuropsychological measures was similar, and both appear to assess speci®c, rather than generalized abilities. This observation was con®rmed with an EQS model to compare differences in the patterns of correlations between ef®ciency measures from the GDS CPT and the Penn CPT and measures from the neuropsychological battery. Maximum likelihood parameter estimation was used. The results indicated that these two correlation matrices did not differ signi®cantly (P , 0.05). It is important to note, however, that the EQS procedure requires large sample sizes, and the power for detecting this difference in correlations was low.

a

PCPT: Penn Continuous Performance Test, ef®ciency score; CPT: Gordon Diagnostics' Continuous Performance Test, ef®ciency score; RHY: Seashore Rhythm Test; WCSTCAT: Wisconsin Card Sorting Test, Categories; WCSTPE: Wisconsin Card Sorting Test Perseverative Errors; TAP: Finger Tapping; JOLO: Judgment of Line Orientation; STEREO: Stereognosis; MAEVN: Visual Naming subtest of the Multilingual Aphasia Examination; VOC: Wechsler Adult Intelligence Scale-Revised, Vocabulary subtest. b Factor loadings of 0.60 or greater appear in boldface type.

(r ˆ 0.35; P , 0.005). This provides some evidence of convergent validity. The small to moderate size of the correlation, however, also suggests that the differing cognitive demands of the two tasks produced somewhat different patterns of performance. Providing some evidence of divergent validity, there were small and non-signi®cant correlations between PCPT ef®ciency scores and tests of visuospatial function (Judgment of Line Orientation and Block Design), and sensory-motor function (Finger Tapping and Stereognosis). The PCPT did not correlate with measures of executive function (WCST Categories and Perseverative Errors) or general verbal intellectual ability (WAIS-R Vocab and MAEVisual Naming). To provide a comparison with the standard Gordon Diagnostics' CPT, the correlations between PCPT and GDS CPT scores with the rest of the neuropsychological battery were compared. As can be seen in Table 2, there was a correlation between ef®ciency measures from the GDS CPT and number correct on the Seashore Rhythm Test (r ˆ 0.26; P , 0.05). Ef®ciency measures from the PCPT and the GDS CPT, while correlated with each other, did not correlate

3.1.3. Factor analysis An exploratory factor analysis was performed to clarify the results of the correlational ®ndings. A principal components analysis was used to extract initial factors from the 55 correlations in Table 2. Scree criteria were used to identify principal components with eigenvalues greater than one. Three principal factors were retained. This unrotated factor pattern matrix was subjected to an orthogonal rotation (Varimax). Loadings of 0.6 or greater from the primary reference structure were used to delineate factors. Results from the orthogonal rotation are presented in Table 3. This model accounted for 53% of the variance. The factors can be labeled as: (a) General Intellectual, de®ned by high loadings for Block Design, WAIS-R Vocabulary, MAE Visual Naming, and Judgment of Line Orientation performance; (b) Executive, consisting of number of categories and perseverative errors on the Wisconsin Card Sorting Test; (c) Visual Vigilance, composed of ef®ciency measures on the GDS and PCPT. 3.2. Study 2 3.2.1. Correlations between PCPT and CPT performance in patients with schizophrenia Consistent with the ®ndings from healthy people in Study 1, ef®ciency measures from the two versions of the CPT were correlated in patients (r ˆ 0.41; P , 0.01).

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

Fig. 2. Ef®ciency scores on the GDS and Penn versions of the CPT in patients with schizophrenia and matched, healthy participants.

3.2.2. PCPT and CPT performance There was a main effect of diagnosis F(1, 74) ˆ 28.13; P , 0.0001; patients had lower ef®ciency scores than healthy participants on both tasks. The main effect of test was also signi®cant F(1, 74) ˆ 4.08; P , 0.05; ef®ciency scores were lower for all subjects on the PCPT (Fig. 2). 3.2.3. Correlations between PCPT, CPT and clinical indices Correlations between attentional measures and clinical measures were essentially nil, and none is signi®cant if correction for multiple comparisons (Bonferroni) is applied. In general, the correlations were in the expected direction of poorer performance associated with greater symptom severity 4. Discussion These results indicate that the PCPT, a new version of the CPT designed to minimize demands on working memory and the perceptual system, is a reliable test that measures sustained visual attention. Reliability data indicate that it is internally consistent. Construct validation analyses provided tentative support for both convergent and divergent validity. With respect to convergence, ef®ciency measures on the PCPT correlated with ef®ciency measures on the standard

313

GDS CPT in both healthy people and patients with schizophrenia. However, the signi®cant correlation was smaller than expected. The divergence in performance between the two task versions most likely re¯ects their differing cognitive demands. In healthy people, with the exception of a correlation between the GDS CPT and the Seashore Rhythm Test, the PCPT correlated with other measures in the neuropsychological battery in a manner nearly identical to that of the standard GDS version of the CPT. With respect to divergence, performance on the PCPT was unrelated to performance on visuospatial and sensorymotor tasks. Results of an exploratory factor analysis suggests that performance on the PCPT loads on the same factor as the GDS CPT. A possible explanation for the observed relationship between the standard GDS CPT, and the Seashore Rhythm Test, is that both tasks include working memory components, in addition to sustained attention. Whereas the GDS CPT requires subjects to remember consecutive integers (a `9' that appears directly after a `1'), the Seashore Rhythm Test requires subjects to remember two sequentially presented auditory stimuli and compare these stimuli in order to make a determination, `same' or `different'. Similarly, the lack of correlation between ef®ciency measured from the PCPT and other measures of sustained attention (e.g. Seashore Rhythm Test) on the test battery most likely re¯ects the absence of other measures on our battery that tap simple, visual sustained attention without working memory demands. It is possible that ef®ciency on the PCPT may be more closely linked to very simple aspects of attention, such as sensitivity to novelty. For example, performance on an `oddball' task, in which a subject's response to a novel sensory stimulus is measured, may be more closely linked to the type of simple, sustained attention measured by the Penn CPT. Study 2 indicated that performance of patients with schizophrenia was equally impaired relative to healthy controls for both versions of the CPT. However, ratings of positive and negative symptoms on the SAPS and SANS were not related to ef®ciency scores for either the Penn or GDS versions of the CPT. These results provide further evidence that impairment in sustained visual attention in schizophrenia represents a core information processing de®cit that is orthogonal to ¯uctuations in patient symptomatology (Nuechterlein

314

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

and Dawson, 1984). These ®ndings are consistent with previous reports that positive psychotic symptoms, while possibly in¯uencing response bias (beta), have little impact on sensitivity (d-prime) to targets on the CPT (Cornblatt et al, 1985). Results are also consistent with several studies showing little relationship between BPRS Factor 1 (conceptual disorganization, hallucinatory behavior and unusual thought content) and positive symptom scale scores from the PANNS and sensitivity to targets on the CPT (Addington and Addington, 1998; Strauss et al., 1993; Nuechterlein et al., 1986). It should be noted that measures of thought disorder from the SAPS were not related to accuracy and reaction time on either version of the CPT. Consistent with this ®nding, Liu et al. (1997) failed to ®nd a signi®cant relationship between measures of thought disorder from the PANSS and accuracy on degraded and undegraded versions of the CPT. These ®ndings, however, are incongruent with several reports, showing negative correlations between ratings of thought disorder and sensitivity measures on the CPT [see Neuchterlein et al. (1986), Strauss et al. (1993) and Pandurangi et al. (1994)]. This discrepancy in the literature may be explained by differences in symptom assessment tools. The present study, as well as the Liu et al. (1997) study, used a limited number of items to assess formal thought disorder that were part of more general scales of positive and negative symptomatology (SAPS and PANSS respectively). In contrast, the Nuechterlein et al. (1986), Strauss et al. (1993) and Pandurangi et al. (1994) studies all used the Thought, Language and Communication (TLC) scale developed by Andreasen (1986), which evaluates thought disorder more comprehensively. The absence of a relationship between negative symptom ratings from the SANS and ef®ciency measures from either version of the CPT is consistent with some reports (Strauss et al., 1993; Addington and Addington, 1998) and discrepant from others (Buchanan et al., 1997; Lees Roitman et al., 1997). These discrepancies most likely represent differences in both sample composition and the tools selected to assess negative symptoms. For example, although Lees Roitman et al. (1997) found a relationship between negative symptomatology as measured by BPRS Factor 2 (emotional withdrawal, blunted affect and motor retardation), and errors of omission and sensitivity on an `A±X' version of the CPT, the authors

studied a chronically hospitalized population from the VA in contrast to the community-dwelling sample utilized in the current study. Likewise, although Buchanan et al. (1997) found a signi®cant relationship between negative symptomatology and performance on a degraded stimulus version of the CPT, the authors used a binary measure of negative symptomatology, the de®cit/non-de®cit classi®cation, rather than the SANS. To help clarify these divergent research ®ndings, future research should be targeted at directly comparing the effects of sample characteristics and type of symptom scale on the relationship between visual sustained attention and symptomatology in schizophrenia. Several caveats to the present ®ndings should be noted. Construct validity for the PCPT would have been strengthened by including additional measures of simple, sustained attention in our neuropsychological test battery. With regard to the failure of correlations between CPT performance and symptom measurements to reach signi®cance, it is noteworthy that all were in the expected direction of greater symptom severity associated with poorer CPT performance. In light of the stringency of the Bonferroni correction, and assuming only a moderate effect size, power for this portion of the study may have been low. Another limitation of this study is that many patients were evaluated while on antipsychotic medication. Nonetheless, it seems unlikely that observed de®cits are attributable to treatment effects alone. Epstein et al (1996), in a study of 38 patients with schizophrenia, investigated this issue directly by comparing medicated and unmedicated patient groups, as well as healthy controls. The results of their study showed a signi®cant effect of diagnosis, but failed to show a signi®cant effect of medication status on sensitivity measures on an `A±X' version of the CPT. Furthermore, in the present sample, there were no signi®cant differences for CPT performance between neuroleptic and non-neuroleptic naive patients. It should be noted that several other studies have reported a relationship between medication and performance on a variety of versions of CPT in patient samples (e.g. Harvey et al., 1990; Serper et al., 1990). These limitations notwithstanding, our results support the construct validity of a CPT measure of attentional vigilance, which does not rely on working memory or perceptual degredation. They also indicate

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

that patients with schizophrenia are impaired on an attentional measure, suggesting attention as a primary area of dysfunction in schizophrenia rather than a byproduct of de®cits in working memory (GoldmanRakic, 1994). The results thus support recent focus on attentive and pre-attentive de®cits as core dysfunction that may re¯ect an endophenotypic vulnerability marker for schizophrenia (Adler et al., 1998). Future studies can examine components of the attentional process that may be differentially impaired in schizophrenia and establish the heritability of such measures. Acknowledgements We thank Travis Turner for his aid in conducting the reliability analysis. This research was funded by grants MH-19112, MH-43880, and MH-01336 from the National Institutes of Health and from the EJLB foundation. References Addington, J., Addington, D., 1998. Visual attention and symptoms in schizophrenia: a 1-year follow-up. Schizophr. Res. 34, 95± 99. Adler, L.E., Freedman, R., Ross, R.G., Olincy, A., Waldo, M.C., 1998. Elementary phenotypes in the neurobiological and genetic study of schizophrenia. Biol. Psychiat. 46, 8±18. Allen, M.J., Yen, W.M., 1979. Introduction to Measurement Theory. Brooks/Cole, Belmont, CA. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders. fourth edition, APA, Washington, DC. Andreasen, N.C., 1983. The Scale for the Assessment of Negative Symptoms. The University of Iowa, Iowa City, IA. Andreasen, N.C., 1984. The Scale for the Assessment of Negative Symptoms. The University of Iowa, Iowa City, IA. Andreasen, N.C., 1986. Scale for the assessment of thought, language and communication (TLC). Schizophr. Bull. 12, 473±482. Asarnow, R.F., MacCrimmon, D.J., 1978. Residual performance de®cit in clinically remitted schizophrenics: a marker of schizophrenia?. J. Abnorm. Psychol. 87, 597±608. Asarnow, R.F., Steffy, R.A., MacCrimmmon, D.J., Cleghorn, J.M., 1977. An attentional assessment of foster children at risk for schizophrenia. J. Abnorm. Psychol. 86, 267±275. Baddeley, A., 1992. Working memory. Science 255, 556±559. Blanchard, J.J., Neale, J.M., 1994. The neuropsychological signature of schizophrenia: generalized or differential de®cit?. Am. J. Psychiatry 151 (1), 40±48.

315

Buchanan, R.W., Strauss, M.E., Brier, A., Kirkpatrick, B., Carpenter, W.T., 1997. Attentional impairment in de®cit and non-de®cit forms of schizophrenia. Am. J. Psychiatry 154, 363±370. Censits, D.M., Ragland, J.D., Gur, R.C., Gur, R.E., 1997. Neuropsychological evidence supporting a neurodevelopmental model of schizophrenia: a longitud study. Schizophr. Res. 24 (3), 289±298. Chute, D.L., Westall, R.F., 1997. PowerLaboratory. MacLaboratory Inc., Devon, PA. Cornblatt, B.A., Erlenmeyer-Kimling, L., 1984. Early attentional predictors of adolescent behavioral disturbance in children at risk for schizophrenia. In: Watt, N.F., Anthony, E.J., Wynne, L.C., Rolfe, J.E. (Eds.), Children at Risk for Schizophrenia: a Longitudinal Perspective. Cambridge University Press, New York, p. 198. Cornblatt, B.A., Keilp, J.G., 1994. Impaired attention, genetics and the pathophysiology of schizophrenia. Schizophr. Bull. 20, 31± 46. Cornblatt, B.A., Lenzenweger, M.R., Dworkin, R.H., ErlenmeyerKimling, L., 1985. Positive and negative schizophrenic symptoms, attention and information processing. Schizophr. Bull. 11, 397±412. Cornblatt, B.A., Risch, N.J., Faris, G., Friedman, D., ErlenmeyerKimling, L., 1988. The Continuous Performance Test, identical pairs version (CPT-IP): I, new ®ndings about sustained attention in normal families. Psychiat. Res. 26, 223±238. Cronbach, L.J., Meehl, P.E., 1955. Construct validity in psychological tests. Psychol. Bull. 52, 281±302. Epstein, J.I., Keefe, R.S., Roitman, S.L., Harvey, P.D., Mohs, R.C., 1996. Impact of neuroleptic medications on Continuous Performance Test measures in schizophrenia. Biol. Psychiatry 39, 902±905. Goldman-Rakic, P.S., 1994. Working memory dysfunction in schizophrenia. J. Neuropsychiatry Clin. Neurosci. 6, 348±357. Gordon, M., 1986. Manual Gordon Diagnostic System. Gordon Diagnostic Systems, New York. Gorsuch, R.L., 1974. Factor Analysis. Saunders, Philadelphia. Gur, R.C., Ragland, J.D., Gur, R.E., 1997. Cognitive changes in schizophrenia Ð a critical look. Int. Rev. Psychiatry 9, 449± 457. Gur, R.E., Mozley, P.D., Resnick, S.M., Levick, S., Erwin, R., Saykin, A.J., Gur, R.C, 1991. Relations among clinical scales in schizophrenia: overlap and subtypes. Am. J. Psychiatry 148, 472±478. Harvey, P.D., Keefe, R.S.E., Moskowitz, J., Putnam, K.M., Mohs, R.C., Davis, K.L., 1990. Attentional markers of vulnerability to schizophrenia: performance of medicated and unmedicated patients and normals. Psychiat. Res. 33, 179±188. Heinrichs, R.W., Zakzanis, K.K., 1998. Neurocognitive de®cit in schizophrenia: a quantitative review of the evidence. Neuropsychology 12, 426±445. Holdnack, J.A., Moberg, P.J., Arnold, S.E., Gur, R.C., Gur, R.E., 1995. Speed of processing and verbal learning de®cits in adults diagnosed with attention de®cit disorder. Neuropsychiatry Neuropsychol. Behav. Neurol. 8 (4), 282±292. Kay, S.R., 1991. Positive and negative syndromes in schizophrenia. Assessment and Research. Brunner/Mazel, New York.

316

M.M. Kurtz et al. / Schizophrenia Research 48 (2001) 307±316

Lees Roitman, S.E., Keefe, R.S.E., Harvey, P.D., Siever, L.J., Mohs, R.C., 1997. Attentional and eye tracking de®cits correlate with negative symptoms in schizophrenia. Schizophr. Res. 26, 139±146. Liu, S.K., Hwu, H.G., Chen, W.J., 1997. Clinical symptom dimensions and de®cits on the Continuous Performance Test in schizophrenia. Schizophr. Res. 25, 211±219. Nuechterlein, K.H., Dawson, M.E., 1984. Information processing and attentional functioning in the developmental course of schizophrenia. Schizophr. Bull. 12, 408±426. Nuechterlein, K.H., Edell, W.S., Norris, M., Dawson, M.E., 1986. Attentional and vulnerability indicators, thought disorder and negative symptoms. Schizophr. Bull. 12, 408±426. Nuechterlein, K.H., Parasvraman, R., Jiang, Q., 1983. Visualk sustained attention: image degradation produces rapid sensitivity decrement over time. Science 220, 327±329. Orzack, M.H., Kornetsky, C., 1966. Attention dysfunction in chronic schizophrenia. Arch. Gen. Psychiatry 14, 323±326. Orzack, M.H., Kornetsky, C., Freeman, H., 1967. The effects of daily carphenazine on attention in the schizophrenic patient. Psychopharmacologia 11, 31±38. Overall, D.R., Gorham, D.R., 1980. The brief psychiatric rating scale. J. Oper. Psychiatry 11, 48±64. Pandurangi, A.K., Sax, K.W., Pelonero, A.L., Goldberg, S.C., 1994. Sustained attention and positive formal thought disorder in schizophrenia. Schizophr. Res. 13, 109±116. Resnick, S.M., 1992. Matching for education in studies of schizophrenia. Arch. Gen. Psychiat. 49, 246. Rosvold, H.E., Mirsky, A., Sarason, I., Bransome, E.D., Beck, L.H.,

1956. A continuous performance test of brain damage. J. Consult. Psychol. 20, 343±350. Saykin, A.J., Gur, R.C., Gur, R.E., Mozley, D., Mozley, L.H., Resnick, S.M., Kester, D.B., Sta®niak, P., 1991. Neuropsychological function in schizophrenia: selective impairment in learning and memory. Arch. Gen. Psychiatry 48, 618±624. Serper, M.R., Bergman, R.L., Harvey, P.D., 1990. Medication may be required for the development of automatic information processing in schizophrenia. Psychiatry Res. 32, 281±288. Shtasel, D.L., Gur, R.E., Mozley, P.D., Richards, J., Taleff, M.M., Heimberg, C., 1991. Volunteers for biomedical research: recruitment and screening of normal controls. Arch. Gen. Psychiatry 48, 1022±1025. Shtasel, D.L., Gur, R.E., Gallacher, F.G., Heimberg, C., Cannon, T.D., Gur, R.C., 1992. Phenomenology and functioning in ®rstepisode schizophrenia. Schizophr. Bull. 18, 449±462. Snedecor, G.W., Cochran, W.G., 1980. Statistical Methods. Iowa State, Ames, IA. Spitzer, R.L., Williams, T.B., Gibbon, M., 1996a. Structured Clinical Interview for the DSM-IV (SCID-P). New York State Psychiatric Institute, New York. Spitzer, R.L., Williams, T.B., Gibbon, M., 1996b. Structured Clinical Interview for the DSM-IV: Non-Patient Version (SCIDNP). New York State Psychiatric Institute, New York. Strauss, M.E., Buchanan, R.W., Hale, J., 1993. Relations between attentional de®cits and clinical symptoms in schizophrenic outpatients. Psychiat. Res. 47, 205±213. Zametkin, A.J., Nordahl, T.E., Gross, M., 1990. Cerebral glucose metabolism in adults with hyperactivity of childhood onset. N. Engl. J. Med. 323 (20), 1361±1367.