Event-Related Potential Correlates of Acquisition and Retrieval of Verbal Associations in Schizophrenics and Controls

(1)

Lioba Baving et al.: Recognition-Related Cortical Potentials in Schizoph reniaJOP 14 (2), 2000, Hogrefe & Huber Publishers

Event-Related Potential Correlates of Acquisition and Retrieval of Verbal Associations in Schizophrenics and Controls

Lioba Baving

³

, Brigitte Rockstroh

¹

, Patricia Rößner

¹

, Rudolf Cohen

¹

, Thomas Elbert

¹

, and Walton T. Roth

²

1Department of Psychology, University of Konstanz, Germany,

2Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Palo Alto, CA, USA,

3Central Institute of Mental Health, Mannheim, Germany

Keywords:Event-related potentials, positive Slow Wave, schizophrenia, paired-associate learning, recognition, “old/new”

Abstract Reduced amplitudes of event-related potentials (ERP) have often been reported for schizophrenic patients. Positive ERPs were examined in 16 schizophrenic patients and 16 controls in a visual paired-associate learning task, in which successful learning of word pairs was evidenced by recognition of “old” versus “new” word pairs. Patients performed significantly poorer than controls. During acquisition the to-be-associated word pairs evoked a P2 and positive Slow Wave (SW) of similar amplitude in both groups. Although the recognition of items as “old” induced larger positive ERPs in controls, ERPs did not vary with recognition in patients. The presentation of the word pairs (50% old, 50% new combinations) evoked a pronounced posterior positive SW in patients but an anterior negative SW in controls. This pattern is identical to the one obtained from a previous paired-associate learning task with cue-recall and suggests that schizophrenic patients are capable of producing large amplitude positive waves under conditions different from those typical for P300 evocation in normals. The larger positive SW may indicate increased effort to compensate for deficient acquisition or association formation.

Introduction

The notion of smaller event-related potential (ERP) amplitudes in schizophrenics compared to controls has a long tradition (Barrett, McCallum, & Pocock, 1986;

Blackwood et al., 1987; McCarley et al., 1993; Roth &

Cannon, 1972; see also review by Cohen, 1991). The smaller amplitudes of negative components (such as the auditory N1, the visual N1 and N2 (e. g., Bruder et al., 1999), the mismatch negativity, MMN, or the contingent negative variation, CNV) and of the P300 have been interpreted as indicating dysfunction of the cognitive processes associated with them. When the scalp distribution of the P300 is considered, the reduction in amplitude of the P300 over the midsagittal posterior and left tem-

poral lobe areas in first episode schizophrenics (Salis- bury et al., 1998) has been considered an indication of a temporal lobe dysfunction in schizophrenia (see also O’Donnell et al., 1999). Left hemisphere reduction of other positive ERP components has also been observed in schizophrenics, e. g., in a demanding auditory dichotic listening task (Bruder et al., 1998).

However, the nonspecificity of these reductions with respect to components raises the question as to whether they indicate specific dysfunctions or more general effects of, for example, medication or hospitalization, me- diated perhaps through increased distractibility or lack of motivation. Smaller P300 amplitudes, at least, cannot be attributed solely to neuroleptic or anticholinergic medication, as they also have been found in drug-free patients (e. g., Eikmeier, Lodemann, Zerbin, & Gastpar, 1991;

Konstanzer Online-Publikations-System (KOPS) URL: http://www.ub.uni-konstanz.de/kops/volltexte/2008/6382/

URN: http://nbn-resolving.de/urn:nbn:de:bsz:352-opus-63822

(2)

Ford et al., 1994; Hirayasu et al., 1998; Shagass, Josias- sen, Roemer, & Straumanis, 1986).

Furthermore, the literature contains a number of exceptions to the finding of smaller positive ERPs in schizophrenics, exceptions which have been attributed to either experimental conditions or subject characteristics.

For example, P300s of normal size were reported by Roth, Goodale, and Pfefferbaum (1991) in response to auditory stimuli separated by long interstimulus intervals, and by Pfefferbaum, Ford, White, and Roth (1989) and Ford et al. (1994) in response to visual stimuli.

Strandburg and colleagues (1994) reported a larger early frontal P3 and a larger contingent negative variation in schizophrenics in a visual Span of Apprehension task, in which earlier negative components and the later posteri- or P3 were significantly smaller in patients than in controls. Finally, the P2 in a visual modality shift design was found to be larger in schizophrenic patients than in controls (Rist & Cohen, 1987; Roth, Horvath, Pfefferbaum,

& Kopell, 1980). In a previous study, which was not particularly designed to elicit a typical P300, we observed smaller P300 amplitudes but larger late positivities in schizophrenics than controls during acquisition and during recall in a paired-associate task employing the proactive inhibition paradigm (Rößner, Rockstroh, Cohen, Wagner, & Elbert, 1999).

Regarding clinical characteristics, Olichney and colleagues (1998) found auditory P300 amplitude reductions in early onset schizophrenics (see also O’Donnell et al., 1999), while late onset schizophrenics had P300 amplitudes in the normal range. Furthermore, an increase in amplitude of the visual, but not the auditory P300 amplitudes accompanied clinical improvement (Duncan, Morihisa, Fawcett, & Kirch, 1987) at least in subgroups of patients (Krieger, Tegeler, Schweizer, Bertling, & Kal- veram, 1989).

A paired-associate learning and recognition task was chosen in the present study to further examine positive ERPs in schizophrenic patients. Efficient learning has been previously shown to be associated with positive deflections. For instance, Donchin and Fabiani (1991) reported larger P300 during acquisition of items that were better remembered. Similarly, Fernandez et al.

(1998) reported ERPs to be more positive for subse- quently recalled than not recalled words in a word list learning paradigm with free recall. Naumann (1985) found larger P300 and later positivities (P560, P600) during elaborate, intentional semantic encoding. Rugg (1995) noted a frontal P560 to covary in amplitude with the correct recognition of items as “old” (i. e., previously seen) compared to “new.” More pronounced positive deflections following the identification of an item as “old”

compared to “new” have been repeatedly reported (e. g., Friedman, 1990; Rugg & Nagy, 1989; Smith & Halgren,

1989; overview by Rugg, 1995). When comparing explicit and implicit memory, Rugg, Mark, Walla, Schloer- scheidt, Birch, and Allan (1998) distinguished a more pronounced frontal positivity 300–500 ms post-stimulus for recognized items from a more pronounced parietal positivity in response to “old” items regardless of their accurate recognition. Sustained positive-going shifts (beyond 800 ms) with right-frontal predominance were observed by Allan and Rugg (1997) in (word-stem) cued recall and “old/new”-recognition tasks, and by Johnson, Kounios, and Nolde (1996) following the correct

“old/new” identification of a previously seen word or picture.

Verbal learning and retrieval have been found to be impaired in schizophrenic patients by some authors (Hu- ron et al., 1995; Saykin et al., 1991, 1994), but not all (Blanchard & Neale, 1994). However, while there are numerous studies on memory (dys)function in schizophrenia, studies examining electrocortical correlates of acquisition and retrieval are rare and results are inconsis- tent. Andrews and colleagues (1993) found reduced parietal positivity together with poorer recall in schizophrenic patients. Subjects read congruous and incongru- ous sentences in anticipation of a memory test. While patients and controls did not differ in the amplitude of the N400 component, patients exhibited a reduced effect of congruity on the late positive component that followed the N400. In contrast, Kayser, Bruder, Friedman, Tenke, Amador, Clark, Malaspina, and Gorman (1999) found the same “old/new” distinction in the P300 in schizophrenic patients as was described for healthy subjects, although patients displayed poorer performance and reduced N2 amplitudes.

The present study examined ERPs in a paired-associate learning task with “old/new” recognition as indication of retrieval accuracy. We aimed at clarifying to what extent positive ERPs during acquisition indicate retrieval performance and retrieval-related ERPs.If schizophrenic patients manifest impaired acquisition and/or retrieval, performance should be poorer along with reduced positive slow waves in the “old/new” recognition. Al- though P300 is usually reduced more in the auditory oddball task in schizophrenia than in its visual equiva- lent, a visual paired-associate learning task was used in the present experiment with the intention of verifying and extending our previous finding (Rößner et al., 1999).

ERPs have been used to study memory, e. g., the

“old/new” effect, in visual and auditory tasks, so that we expected to gain some insight into the significance of the learning- and retrieval-related positivites.

(3)

Methods

Subjects

Sixteen schizophrenic inpatients (11 male, mean age 32.8 ± 7 years) were recruited from the local state hospital. DSM diagnoses were made by the responsible psy- chiatrist and/or the psychologist on the basis of the Pre- sent State Examination (PSE, Wing, Cooper, & Sarto- rius, 1974), clinical records and extended personal contact with the patient. Fourteen patients were diag- nosed as paranoid (DSM IV 295.3) and two as disorga- nized subtype (DSM IV 295.1). Duration of illness since the first admission varied between 6 months and 20 years (mean 7 years, 6 months). The number of hospital admis- sions ranged from 1 to 17 (mean 4.5 ± 4.4). Patients with a schizoaffective disorder or with an additional diagnosis such as substance dependence were not included. Pa- tients were also excluded if they had a neurological disorder or if EEG- or CT-abnormalities on the routine admission check were detected. The clinical status during the week of the investigation was evaluated by the Brief Psychiatric Rating Scale (BPRS; Overall & Gorham, 1962, as modified by Lukoff, Nuechterlein, & Ventura, 1986)*, the SANS (Scale for the Assessment of Negative Symptoms; Dieterle, Albus, Eben, Ackenheil, & Rock- stroh, 1986) and the PANSS (Positive and Negative Symptom Scale; Kay, Fiszbein, & Opler, 1987). Individ- ual mean scores on the BPRS varied between 24 and 59 around a mean of 40.7 ± 8.4; for the five BPRS subscales, group means were 8.2 ± 3.1 (anxiety/depression), 8.6 ± 3.2 (anergia), 7.7 ± 2.8 (thought disturbance), 3.6 ± 0.9 (activation), and 4.1 ± 1.5 (hostility-suspiciousness). In- dividual scores on the SANS varied between 0 and 26 around a mean of 7.7 ± 5.9, while the respective values were 7–27 (M = 17.6 ± 6.3) for the PANSS-N and 6–22 (M = 12.8 ± 4.7) for PANSS-P. Except for two unmedicated patients, schizophrenic patients were receiving either standard or atypical neuroleptic treatment (mean 262.3 ± 258.0 mg/day, range 0–900 mg, median 200 mg), but no additional anticholinergic medication.

Sixteen healthy subjects (10 male, mean age 29.4 ± 5.8 years) were selected to be comparable to the patient group on age and education. (Years of education were 11.4 ± 3 years in patients and 10.4 ± 5 years in controls.) Subjects were excluded from the study if they reported to have been treated for neurological or psychiatric dis- orders and were currently taking any psychotropic medication.

Three patients and two controls were left-handed ac- cording to a modified version of the Edinburgh handedness questionnaire (Oldfield, 1971). All subjects were informed about the procedures and gave written consent prior to the experimental session. They received a small financial bonus (about US $10) for participation.

Design and Procedure

Twenty-five pairs of four- to seven-letter nouns were presented simultaneously. Word pairs appeared on a TV- monitor for 1 s. The series of 25 word-pairs was repeated, the sequence of word pairs differing between the two consecutive runs.Subjects were instructed to remember the word pairs. Learning was followed by a recognition phase, in which 50 word pairs, the 25 original pairs in- termixed with 25 new combinations of the original words, were presented. Subjects were asked to indicate by pressing response buttons either with the left or with the right hand, whether they recognized the presented word pair as “old” or “new.” Then the series of simultaneous acquisition-recognition was repeated with 25 new word pairs. The correspondence between decision type (old/new) and dominant response hand (left/right) was counterbalanced across subjects.

The same procedure was realized in a second acquisition-recognition series with the only exception that the two words constituting a pair were presented one after the other separated by an interstimulus interval of 1.4 s.

This sequential presentation was introduced as a chal- lenge of working memory capacity. The sequence of “simultaneous” and “sequential” acquisition-recognition was counterbalanced across subjects. Intertrial intervals between word pairs varied between 2.5 and 4.5 s in both series.

Practice trials with different words preceded the start of the experimental series. The latter started only if the subject responded correctly in more than 50% of the practice trials.

Apparatus and Data Acquisition

High frequent, concrete four- to seven-letter, one- to two- syllable German nouns of low associative proximity were selected (Hager & Hasselhorn, 1984; Ruoff, 1981).

Stimuli (generated by the STIM system, NEUROSCAN,

* The expanded BPRS with behavioral anchors (Lukoff et al., 1986) allows the evaluation of the target symptoms on the basis of 24 constructs:

somatic concern, anxiety, depression, guilt, hostility, suspiciousness, unusual thought content, grandiosity, hallucinations, disorientation, conceptual disorganization, excitement, motor retardation, blunted affect, tension, mannerisms and posturing, uncooperativeness, emotional withdrawal, suicidality, self-neglect, bizarre behavior, elated mood, motor hyperactivity, and distractability; the intensity on each construct is rated on a 7-point scale (not observed to extremely severe).

(4)

GENTASK 2.0) were 2 cm high and covered a horizontal extension of 7 cm at most. For the simultaneous presentation, the two words were presented one on top of the other, covering about 5 cm. Given the distance of about 3 m between the monitor and the subject’s eyes, each word subtended a visual angle of about 0.38 × 1.33 degrees, while two words subtended a visual angle of 0.96 degrees in height. Stimuli were presented as white letters on the dark screen background on a 20 cm monitor. Re- sponse buttons were microswitches that subjects held in their left and right hand.

The EEG was recorded with a DC-amplifier (MES, Munich) from electrodes along the midsagittal line (Fz, Cz, Pz), frontal (F3, F4), central (C3, C4), and parietal (P3, P4) locations over both hemispheres and both mastoids (A1, A2). The vertex electrode Cz served as reference electrode, while a ground electrode was attached to the forehead. The vertical and horizontal EOG were recorded for the correction of eye movement artifacts in ERPs (see next section) from electrodes about 1 cm above and below the left eye and on the left and right sides of the eyes on the temples. Nonpolarizable sintered Ag/AgCl electrodes (ZAK), that were sterilized with eth- anol and disinfectant prior to each experiment, were used with Grass EC2 electrolyte as the conducting medium.

The skin under the electrodes was prepared by rubbing with abrasive paste (OMNIPREP). Bandwidth ranged from DC to 30 Hz (6 dB/octave). For statistical analyses, EEG recordings were converted to mastoid reference (mean of both mastoids). All data were digitized at a rate of 100 Hz and were stored for offline analyses.

Data Reduction and Analysis

Indices ofperformancewere the number of errors for the total 100 recognition trials of each recognition condition (recognition following simultaneous and following sequential acquisition) and the median reaction time (RT) of correct responses, determined per condition and subject. Data from those trials in which response latency exceeded 2.5 s were discarded from further analysis.

EEG recordings were first controlled for slow drifts by an automatic drift correction included in the data acquisition program of the amplifier. All EEG data were visually inspected, and those contaminated by other artifacts (muscle potentials, large drifts, etc.) rejected. The accepted EEG epochs of 1.6 s (including 200 ms pre- stimulus and 1400 ms post-stimulus) were corrected for eye movement and blink artifacts following the method suggested by Elbert, Lutzenberger, Rockstroh, and Bir- baumer (1985) and implemented by Berg (1986). On the basis of three linear regressions from the calibration data, each EEG channel is corrected with respect to the voltage

of three EOG channels, the “vertical” (above minus below left eye), the “horizontal” (left minus right temple electrode), and the “radial” (mean of the four EOG electrodes minus mean of A1 and A2).

Further analyses were based on 80.3% ± 10.0 artifact- free epochs with correct responses in the patient group and 83.4% ± 9.0 in the control group (t(30) = –.93, n.s.).

For these epochs, two positive deflections following stimulus-onset were prominent in the averaged ERP and were determined for every subject,acquisition and recognition, simultaneous or sequential acquisition/recognition, and electrode. The amplitude of the first component was calculated as the difference between the 200-ms pre-stimulus baseline and the voltage at the latency of the maximum amplitude between 100 and 400 ms at Pz. The average latency of this component (228 ± 41 ms) lead us to label it P2. The amplitude of the positive Slow Wave (pSW) was calculated as the difference between baseline and the mean voltage between 500 and 800 ms.

Differences in performance measures between groups and series (recognition following simultaneous and sequential acquisition) were evaluated by means of analyses of variance (ANOVA) with the between-subject fac- torGROUP(patients versus controls) and the within-subject factorSERIEScontrasting simultaneous and sequential presentation of word pairs. For ERP data the ANOVAs (that were performed separately for the simultaneous and sequential acquisition series) comprised the within-subject factors LATERALITY (left, midline, and right hemispheric electrodes) and GRADIENT (frontal, central, and parietal electrodes).Pvalues were adjusted for degrees of freedom with the Greenhouse-Geisser epsilon.Post- hoccomparisons were made by ANOVAs on subsets of data or by planned tests of means.

Results

Performance

Patients made significantly more errors (M = 25.8 ± 14.6 for simultaneous, M = 27.1 ± 14.7 for sequential presentation out of the total 100 trials for each condition), i. e., recognized “old” word pairs as “new” and vice versa, than controls (M = 11.5 ± 13.1 for simultaneous, M = 14.7 ± 11.6 for sequential presentation,F(1,30) = 8.7,P

< .01). This group difference was maintained when trials with response latencies exceeding 2.5 s were included (M = 33.4 ± 13.0 for simultaneous and M = 35.7 ± 13.3 for sequential series in patients; M = 14.3 ± 12.7 and M

= 18.5 ± 13.3 in controls;F(1,30) = 18.0,P< .01). Error rates in the two groups did not differ between simultaneous and sequential presentation of word pairs (interaction ofGROUP×SERIESwas not significant,F< 1). That is,

(5)

learning of sequentially presented word pairs was not more difficult for patients than for controls. Within the patient group, error rates increased with duration of illness as indicated by a correlation ofr= .54 (P< .05).

In contrast to the often reported generally slower re- sponding of schizophrenics, groups did not differ in their mean reaction time in the present study. This may be explained by the cutoff criterion of 2.5 s, which led to the exclusion of more trials in patients (M = 7.6) than in controls (M = 2.8, t = 1.86,P< .1). An inclusion of these trials may have produced the expected group difference in response latency. Both groups were slower to recognize word pairs after sequential acquisition (for patients M = 1282.5 ms; for controls M = 1237.5 ms) than after simultaneous acquisition (for patients M = 1240 ms, for controls M = 1208 ms;SERIES:F(1,30) = 4.3,P< .05;GROUP×

SERIES:F< 1), and both groups were faster to recognize

“old” pairs (for patients M = 1227 ms; for controls M = 1166 ms) than “new” ones (for patients M = 1295.5; for controls M = 1280 ms,F(1,30) = 64.1,P< .01).

Figure 1 illustrates the ERPs evoked duringacquisi- tionwhen the word pairs were presented simultaneously.

Presentation of the word pairs during acquisition tended to elicit a P2 at earlier latencies in controls (M = 217.2 ± 39.8 ms) than in patients (M = 240.6 ± 41 ms;F(1,30) = 2.7,P< .1). In both groups, latencies tended to be shorter (P< .1), when the to-be-associated words were presented with a gap of 1.4 s (P2 to the first word: M = 234.7 ± 40 ms in patients, M = 219.0 ± 42.4 ms in controls; P2 to the second word: M = 238.8 ± 40.6 ms in patients, M = 232.7 ± 53.2 ms in controls) than when the pairs were presented simultaneously.

The P2 was predominant at posterior and right-hemisphere electrodes (GRADIENT×LATERALITY:F(4,120) = 6.5, P< .01; main effectsGRADIENT:F(2,60) = 9.0,P< .01;

LATERALITY:F(2,60) = 16.8,P< .01). Although overall P2 amplitude did not differ between groups, there was a trend for the scalp distribution to differ between groups (GROUP

×GRADIENT×LATERALITY:F(4,120) = 2.6,ε= .79,P= .055):

Patients displayed a steeper fronto-parietal gradient (GRA- DIENT: F(1,15) = 18.1, P< .01) than controls (GRADIENT, n.s.;GROUP×GRADIENT:F(2,60) = 5.11,ε= .66,P< .01), while controls displayed the more pronounced central and right-hemispheric dominance (LATERALITY:F(1,15) = 41.9, P< .01 for the left-midline difference,F(1,15) = 4.8,P<

.05 for the left-right asymmetry) than patients (LATERALITY: F(1,15) = 5.5,P< .05 for the difference left-midline/right;

GROUP×LATERALITY:F(2.60) = 6.5,ε=.97,P< .01).

The amplitude of the pSW did not differ between the groups. It was largest at posterior and right-hemisphere electrodes as confirmed by the interaction GRADIENT×

LATERALITY(F(4,120) = 9.2,P< .01;GRADIENT: F(2,60) = 18.1,P< .01).

The ERP during trials with correct recognition as

shown in Figure 2 was characterized by a positive peak around 200 ms and a pSW extending up to the end of the recording epoch. Mean latencies of the P2 were 240.6 ± 45.3 ms following simultaneous, and 246.9 ± 42.8 ms following sequential acquisition in patients; the respective values were 212.2 ± 48.3 and 223.4 ± 50.5 ms in controls. There were no significant differences between groups and conditions for P2 latencies and no overall group difference for P2 amplitude (F< 1). In the patient group and following simultaneous acquisition the P2 was similarly pronounced when trials were averaged for correct “old/new” responses (6.4 ± 2.9 and 6.7 ± 2.4µV), while it was 0.7µV larger in controls in response to items correctly identified as “old” (7.2 ± 3.5µV) compared to

“new” (6.5 ± 3µV;t(15) = 1.8,P< .1;GROUP×OLD/NEW: F(1,30) = 5.15, P< .05; for recognition following sequential acquisition,F< 1).

As reported for acquisition, patients exhibited a steep-

Figure 1 Event-related potentials during 200 ms pre-stimulus and 1000 ms following the onset of a word pair presented simultaneously, averaged separately for a group of control subjects (top) and a group of schizophrenic patients (bottom) at nine electrode sites (negativity up) during the acquisition of the paired associates.

(6)

er fronto-parietal gradient of the P2 amplitude than controls, who exhibited a central maximum (GROUP×GRADI- ENT: F(2,60) = 10.7, ε = .685, P < .001, on retrieval following simultaneous; F(1,30) = 5.61,ε = .685, P<

.001, on retrieval following sequential presentation during acquisition).Post hoctests confirmed significant differences in P2 amplitude between frontal and central and central and parietal electrodes in patients (F(1,15) = 27.4, P< .001, andF(1,15) = 6.55,P< .05, respectively), while in controls only the difference between central and parietal electrodes reached significance (F(1,15) = 2.6,P<

.05). A central maximum of the P2 with smaller left- and right-hemispheric amplitudes in controls, compared to a right-and-central predominance in patients was indicated by the interaction GROUP × GRADIENT × LATERALITY

(F(4,120) = 2.6,P< .05;GRADIENT×LATERALITY:F(4,120)

= 8.7,ε= .65,P< .01).

Following sequential acquisition, smaller left-hemisphere amplitudes and a more pronounced left-right asymmetry in patients (F(1,15) = 10.8,P< .01) and a central and right-hemispheric maximum in controls (F(1,15) = 13.8,P< .01) explains the interactionGROUP

×GRADIENT×LATERALITY(F(4,120) = 3.3,ε= .751,P<

.05;GROUP×LATERALITY:F(2,60) = 3.2,ε= .998,P< .05).

A Slow Wave with frontal negative predominance (nSW) and positive amplitude at Pz characterized the subsequent ERP in controls, while the Slow Wave had positive amplitudes at all electrodes in patients, predom- inating at Cz and Pz. This group-specific distribution of the Slow Wave was confirmed by the main effectGROUP

(F(1,30) = 7.6,P< .01 for recognition following simultaneous;F(1,30) = 4.5,P< .05 for recognition following sequential presentation during acquisition) and the inter- actionGROUP×GRADIENT×LATERALITY(F(4,120) = 2.6,P

< .05 for retrieval following simultaneous;F(4,120) = 2.5,P< .05 following sequential acquisition).Post hoc analyses confirmed a negative SW in controls and a positive SW in patients at F3, Fz, Cz, and C4 (P< .01 for all tests when retrieval followed simultaneous,P< .05 following sequential acquisition), while the pSW was significantly larger in patients than controls only at P4 (P<

.05) following simultaneous acquisition.

Regardless of the preceding acquisition condition and in both groups, the pSW was larger over posterior than anterior sites and larger over the midline than over the two hemispheres (GRADIENT×LATERALITY:F(4,120) = 9.2, ε= .8, P < .001 for retrieval following simultaneous;

F(4,120) = 9.9, ε= .76, P< .001 following sequential acquisition; for the main effectGRADIENT, the respective F-values were F(2,60) = 11.11, ε = .6, P < .001, and F(2,60) = 12.8, ε = .6, P < .001; for the main effect

LATERALITY, the respectiveF-values wereF(2,60) = 4.2,ε

= .88,P< .05, andF(2,60) = 3.9,ε= .98,P< .05).

Controls showed a difference of about 2µV in the SW between words correctly classified as “old” and “new,”

while patients exhibited the same pSW for “old” and

“new” items (GROUP×CONDITIONF(,1,30) = 12.1,P< .01 for simultaneous,F= 3.7,P< .1 for sequential presentation during acquisition; main effectCONDITION:F(1,30) = 17.4,P< .01 for simultaneous,F= 3.8,P< .1 for sequential presentation of word pairs during acquisition).

The difference in the P2-responses between “old” and

“new” items was related to performance, i. e., the number of correct identifications in controls (r = .49, P = .054) with a larger P2-“old/new”-effect covarying with better recognition. (Since patients did not show an

“old/new”-effect, the correlation ofr= –.1 in this group is not particularly meaningful.) No significant correla- tions were disclosed for the pSW. In patients, coefficients remained nonsignificant after medication was considered as covariate.

Figure 2 Event-related potentials during the retrieval condition, averaged for 200 ms pre-stimulus and 1000 ms following the onset of a word pair, which was either identical to pairs learned before (“old”) or not (“new”).

ERPs are averaged separately for controls (top) and schizophrenic patients (bottom) at nine electrode sites (negativity up) for trials in which subjects correctly identified the word pairs as “old” (thick line) and “new” (thin line).

(7)

Discussion

The present results resemble those reported by Rößner et al. (1999) in that schizophrenic patients were less efficient in retrieval of paired-associates than controls; more interestingly, the group-specific ERP patterns during recognition were similar between the studies. Both studies used similar methods, in that the association of word pairs (concrete, high-frequent nouns with low-associative proximity) was required. However, they differed in the testing through retrieval by recognition (as in the present study) or recall (as in Rößner et al., 1999). Thus, group-specific ERP patterns seem to reflect memory retrieval processes independent of the retrieval mode – recognition or recall. The patient samples were similar in the two studies, both coming from the same ward and being alike in diagnostic subcategories for schizophrenic disorder, chronicity, and neuroleptic medication.

Patients made significantly more errors than controls, when distinguishing between “old” and “new” word pairs, the more so with longer duration of illness (r= .54, P< .05). Poorer performance of schizophrenic patients is in line with many other findings of general performance deficits and memory dysfunction, so that the present results may support the assumption of a general deficit. Response latencies were not longer in patients than in controls, possibly because responses with latencies exceeding 2.5 s were rejected from the analysis.

Therefore, the present results do not clarify to what extent the poorer performance indicates specific deficits, deficient memory encoding, deficient retrieval, and/or the contribution of nonspecific processes involved in the task such as attention.

Acquisition conditions elicited positive ERPs, P2 and positive SW, which did not differ between groups. Don- chin and Fabiani (1991), Fernandez et al. (1998), and Rugg (1995) found better retrieval of those items that elicited the larger positive ERP during acquisition. This covariation was not obvious in the present study. How- ever, both positive deflections elicited during recognition distinguished words recognized as “old” and “new”

and correctly recognized words from errors in controls, but not in patients. The lack of such a distinction suggests a relationship between efficiency of retrieval and positive ERPs rather than a relationship between acquisition- related ERPs and retrieval. However, since the

“old/new” effect on P2 covaried with recognition performance in controls but not in patients, poorer performance and a lack of “old/new” distinction in the ERP in the latter group might have been the consequence of a deficit in the stage of processing related to the P2.

Compared to other studies (e. g., Kayser et al., 1999;

Rößner et al., 1999) the present task elicited only very

small negative deflections in the ERP. This might be a consequence of activity at the mastoids, which were used as reference. It might also be attributed to the specificities of the present design, such as a rather long exposure to visual stimuli (words to be read) or the complexity of the task. In the similar task, Rößner et al. (1999) also found transient reductions of positive deflections between P2, P3 and pSW, called N2 and N4, rather than distinct negative deflections. Amplitudes of these negative ERPs tended to be smaller in patients than in controls.

A particularly intriguing finding seems to be the group-specific ERP-pattern during retrieval, which is similar to the results reported by Rößner et al. (1999) despite different retrieval conditions. Patients displayed a large posterior pSW that exceeded the pSW of controls.

In controls the recognition of new items elicited an anterior negative SW, which was not present in patients. ERP studies suggest encoding and verbal memory are related to left-frontal activity, and retrieval (recall and recognition) to right-frontal activity (Allan & Rugg, 1997;

Düzel, Yonelinas, Mangun, Heinze, & Tulving, 1997;

Rösler, Heil, & Röder, 1997; Uhl et al., 1990; see also Fernandez et al., 1998), while memory search processes have been associated with slow event-related frontal negativities (Rösler et al., 1997). Activation of (mainly right) frontal regions during retrieval has also been observed in PET studies (Fletcher, Shallice, Frith, Fracko- wiak, & Dolan, 1998; Nyberg et al., 1996). Task demands and the extent to which subjects engage in recol- lective processing also affect the region of major activation. Nolde, Johnson, and Raye (1998) found left prefrontal activation with increasing demands of the task (compared to old/new recognition), while Fletcher et al.

(1998) demonstrated increased activation in the dorsal region when monitoring demands were emphasized compared to greater activation in the ventral regions when external cueing was emphasized. Posterior positivity in the ERP has been associated with more implicit recognition by Rugg et al. (1998) and Johnson et al.

(1996), while frontal positivity has been interpreted as an indication of conscious recognition (Rugg et al., 1998) and source monitoring (Johnson et al., 1996). Paller, Ku- tas, and McIsaac (1995) observed increasingly positive ERPs between 500 and 900 ms after stimulus onset with increasingly intense stimulus processing. Finally, distinct ERP topographies have been associated with episodic memory or autonoetic awareness and familiarity or noetic awareness in “remember/know” recognition judgments: Düzel et al. (1997) found “know” (i. e., familiarity) judgments to covary with posterior positivity between 190–200 ms, and frontocentral negativity between 300–600 ms, while “remember” judgments covaried with widespread positivity between 600 and 1000 ms following the onset of the test stimulus.

(8)

Considering this evidence, the group-specific distribution of the Slow Wave in the present study might reflect group differences in the activated processes or strategies, as, for instance, a tendency for explicit, conscious remembering of new items or search strategies for new items in controls and a tendency for implicit memory or familiarity judgments or a (more poorly organized) search in their parietal association cortices in patients.

Selectively impaired recognition as measured by “remember” relative to “know” responses have been reported for schizophrenics by Huron et al. (1995). The marked posterior pSW in patients might then point to an increased compensatory effort consequent upon a dysfunc- tional earlier stage of processing.

Lateral asymmetry of ERPs, although small relative to the predominance at midline electrodes, cannot easily be explained by the literature. Both groups displayed a right-posterior preponderance in the distribution of positivities during learning and retrieval. Similar right-hemispheric and posterior predominance of positive ERPs have been described for less demanding or implicit recognition tasks, while more demanding and explicit, conscious retrieval has been associated with left frontal (Johnson et al., 1996; Nolde et al., 1998; Rugg et al., 1998) or left temporoparietal (Düzel et al., 1997) activation. The lack of clear and group-specific hemispheric asymmetry in the present study may also be due to the visual modality, since P300 abnormalities in schizophrenics and P300 asymmetries have been reported for auditory tasks indicating the contribution of temporal lobe structures to the P300 (O’Donnell et al., 1999).

Another intriguing result in this respect is that the pSW in patients did not differ between correctly identified as “old” or “new” identified items. If the late positive ERP with posterior predominance is considered to reflect implicit recognition procedures, while anterior positivity indicates more elaborate processing, as suggested, for instance, by Rugg (1995; Rugg et al., 1998), it is tempt- ing to conclude that those elaborate memory processes are deficient in schizophrenic patients. As a consequence, patients may exhibit poorer performance, but even on trials with correct recognition, the lack of ERP distinction between “old” and “new” items may be an electrocortical indicator of the failure to distinguish between “old” and “new.” The pronounced posterior SW might represent the focus on implicit retrieval or familiarity estimates. However, it should be noted that the present design differed from standard “old/new” recognition tasks; the simultaneous presentation of two words may have increased task complexity and a relatively long interval of 1 s to read the words may have affected the ERPs. It cannot be ruled out that the specificity of this task a) induced different strategies than those reported above for “old/new” recognition or “remember/know”

tasks, and b) affected the ERP pattern (less negative and more pronounced positive deflections).

To conclude, the present findings clearly demonstrate that no generalization can be made that ERPs are always smaller in schizophrenics than controls, regardless of whether this would be interpreted in psychological terms (deficient cognitive process related to the particular ERP component), biological terms (reduced cortical activation), or methodological terms (poor signal-to-noise ra- tio). In paradigms other than the traditional oddball task, schizophrenics produce positive ERPs as large as those in controls. We can only speculate whether these large positivities reflect some kind of compensation for a deficit in the development of search strategies or for a pre- ferred retrieval (implicit, explicit, remember, know) strategy. These generalizations about positivities are lim- ited by the fact that our results come only from a visual task. It is possible that P3 reduction depends on the cognitive task demands, which may vary with stimulus modality. It is possible that subjects have to focus their attention more in a visual than an auditory task, and that differences between schizophrenics and controls are reduced by these increased task requirements.

Acknowledgment

Research was supported by the Deutsche Forschungsge- meinschaft. We would like to thank Drs. H. Watzl, P.

Werle and H.U. Becker for providing diagnosis and assessment of the patients and Dr. P. Berg for support in data analysis.

References

Allan, K., & Rugg, M.D. (1997). An event-related potential study of explicit memory on tests of cued recall and recognition.

Neuropsychologia,35, 387–397.

Andrews, S., Shelley, A.-M., Ward, P., Fox, A., et al., (1993).

Event-related potential indices of semantic processing in schizophrenia.Biological Psychiatry,34, 443–458.

Barrett, K., McCallum, W.C., & Pocock, P.V. (1986). Brain indi- cators of altered attention and information processing in schizophrenic patients.British Journal of Psychiatry,148, 414–420.

Berg, P. (1986). The residual after correcting event-related potentials for blink artifacts.Psychophysiology,23, 354–364.

Blackwood, D.H., Whalley, L.J., Christie, J.E., Blackburn, I.M., St. Clair, D.M., & McInnes, A. (1987). Changes in auditory P3 event-related potential in schizophrenia and depression.British Journal of Psychiatry,150, 154–160.

Blanchard, J.J., & Neale, J.M. (1994). The neuropsychological signature of schizophrenia: Generalized or differential deficit?

American Journal of Psychiatry,151, 40–48.

Bruder, G., Kayser, J., Tenke, C., Amador, X., Friedman, M., Shariff, Z., & Gorman, J. (1999). Left temporal lobe dysfunction in schizophrenia: Event-related potential and behavioral

(9)

evidence from phonetic and tonal dichotic listening tasks.Ar- chives of General Psychiatry,56, 267–276.

Bruder, G., Kayser, J., Tenke, C., Rabinowicz, E., Friedman, M., Amador, X., Sharif, Z., & Gorman, J. (1998) The time course of visuospatial processing deficits in schizophrenia: An event- related brain potential study.Journal of Abnormal Psychology, 107, 399–411.

Cohen, R. (1991). Event-related potentials and cognitive dysfunctions in schizophrenia. In H. Häfner & W.F. Gattaz (Eds.), Search for the causes of schizophrenia(pp. 342–360). Heidel- berg: Springer-Verlag.

Dieterle, D., Albus, M., Eben, E., Ackenheil, M., & Rockstroh, W.

(1986). Preliminary experiences and results with the Munich version of the Andreasen Scale. Pharmacopsychiatry, 19, 96–100.

Donchin, E., & Fabiani, M. (1991). The use of event-related brain potentials in the study of memory: Is P300 a measure of event distinctiveness? In J.R. Jennings & M.G.H. Coles (Eds.), Handbook of cognitive psychophysiology: Central and auto- nomic system approaches(pp. 471–498). Chichester: Wiley.

Düzel, E., Yonelinas, A.P., Mangun, G.R., Heinze, H.-J., & Tul- ving, E. (1997). Event-related brain potential correlates of two states of conscious awareness in memory.Proceedings of the National Academy of Sciences USA,94, 5973–5978.

Duncan, C.C., Mirihisa, J.M., Fawcett, R.W., & Kirch, D.G.

(1987). P300 in schizophrenia: state or trait marker?Psycho- pharmacology Bulletin,23, 339–352.

Eikmeier, G., Lodemann, E., Zerbin, D., & Gastpar, M. (1991).

Ereigniskorrelierte Potentiale bei schizophrenen Patienten im akuten Krankheitsschub und in Remission. Zeitschrift für EEG-EMG,22, 15–20.

Elbert,T., Lutzenberger, W., Rockstroh, B., & Birbaumer, N.

(1985). Removal of ocular artifacts from the EEG – A biophys- ical approach to the EOG.Electroencephalography and Clini- cal Neurophysiology,60, 455–463.

Fernandez, G., Weyerts, H., Tendolkar, I., Smid, H., Scholz, M.,

& Heinze, H.-J. (1998). Event-related potentials of verbal encoding into episodic memory: Dissociation between the effects of subsequent memory performance and distinctiveness.Psy- chophysiology,35, 709–720.

Fletcher, P.C., Shallice, T., Frith, C.D., Frackowiak, R.S., & Do- lan, R.J. (1998). The functional roles of prefrontal cortex in episodic memory.Brain,121, 1249–1256.

Ford, J.M., White, P.M., Csernansky, J.G., Faustman, W.O., Roth, W.T., & Pfefferbaum, A. (1994). ERPs in schizophrenia: Ef- fects of antipsychotic medication.Biological Psychiatry,36, 153–170.

Friedman, D. (1990). ERPs during continuous recognition memory for words.Psychophysiology,27, 136–148.

Hager, W., & Hasselhorn, M. (Eds.) (1994).Handbuch deutsch- sprachiger Wortnormen. Göttingen: Hogrefe.

Hirayasu, Y., Asato, N., Ohta, H., Hokama, H., Arakaki, H. &

Ogura, C. (1998). Abnormalities of auditory event-related potentials in schizophrenia prior to treatment.Biological Psychi- atry,43, 244–253.

Huron, C., Danion, J.-M., Giacomoni, F., Grangé, D., Robert, P.,

& Rizzo, L. (1995). Impairment of recognition memory with, but not without, conscious recollection in schizophrenia.Amer- ican Journal of Psychiatry,152, 1737–1742.

Johnson, M.K., Kuonios, J., & Nolde, S.F. (1996). Electrophysio-

logical brain activity and memory source monitoring.Neuro- Report,7, 2929–2932.

Kay, S.R., Fiszbein, A., & Opler, L.A. (1987). The Positive and Negative Syndrome Scale (PANSS) for schizophrenia.Schizo- phrenia Bulletin, 13, 261–275.

Kayser, J., Bruder, G.E., Friedman, D., Tenke, C.E., Amador, X.F., Clark, S.C., Malaspina, D., & Gorman, J.M. (1999). Brain event-related potentials (ERPs) in schizophrenia during a word recognition memory task. International Journal of Psycho- physiology, 34,249–266.

Krieger, S., Tegeler, J., Schweizer, A., Bertling, R., & Kalveram, K.T. (1989). Ereignisbezogene Potentiale und elementare In- formationsverarbeitungsprozesse bei schizophrenen Patienten.

In K. Heinrich & B. Bogerts (Eds.),Pathophysiologische und pathomorphologische Aspekte bei schizophrenen Psychosen.

Stuttgart: Schattauer.

Lukoff, D., Nuechterlein, K., & Ventura, J. (1986). Appendix A.

Manual for expanded brief psychiatric rating scale (BPRS).

Schizophrenia Bulletin,12, 594–602.

McCarley, R.W., Shenton, M.E., O’Donnell, B.F., Faux, F.S., Kikinis, R., Nestor, P.G., & Jolesz, F.A. (1993). Auditory P300 abnormalities and left posterior superior temporal gyrus vol- ume reduction in schizophrenia.Archives of General Psychia- try,50, 190–197.

Naumann, E. (1985).Ereigniskorrelierte Potentiale und Gedächt- nis. Frankfurt: Peter Lang Verlag.

Nolde, S.F., Johnson, M.K., & Raye, C.I. (1998). The role of prefrontal cortex during tests of episodic memory.Trends in Cognitive Sciences, 2/10, 399–406.

Nyberg, L., McIntosh, A.R., Cabeza, R., Habib, R., Houle, S., &

Tulving, E. (1996). General and specific brain regions involved in encoding and retrieval of events: What, where, and when.

Proceedings of the National Academy of Sciences USA, 93, 11280–11285.

O’Donnell, B.F., McCarley, R.W., Potts, G.F., Salisbury, D.F., Nestor, P.G., Hirayasu, Y., Niznikiewicz, M.A., Barnard, J., Shen, Z., Weinstein, D.M., Bookstein, F.L., & Shenton, M.

(1999). Identification of neural circuits underlying P300 abnormalities in schizophrenia.Psychophysiology,36, 388–398.

Oldfield, R. (1971). The assessment and analysis of handedness:

The Edinburgh Questionnaire.Neuropsychologia,9, 97–113.

Olichey, J.M., Iragui, V.J., Kutas, M., Nowackim, R., Morris, S.,

& Jeste, D.V. (1998). Relationship between auditory P300 amplitude and age of onset of schizophrenia in older patients.

Psychiatry Research,79, 241–254.

Overall, J.E., & Gorham, D.R. (1962). The Brief Psychiatric Rat- ing Scale.Psychological Reports,10, 419–427.

Paller, K.A., Kutas, M., & McIsaac, H.K. (1995) Monitoring conscious recollection via the electrical activity of the brain.Psy- chological Science, 6/2, 107–111.

Pfefferbaum, A., Ford, J.M., White, P.M., & Roth, W.T. (1989). P3 in schizophrenia is affected by stimulus modality, response requirements, medication status, and negative symptoms.Ar- chives of General Psychiatry,46, 1035–1044.

Rist, F., & Cohen, R. (1987). Effects of modality shift on event- related potentials and reaction time of chronic schizophrenics.

In R. Johnson, J. Rohrbaugh, & R. Parasuraman (Eds.),Cur- rent trends in event-related potential research(pp. 738–745).

Amsterdam: Elsevier.

Rösler, F., Heil, M., & Röder, B. (1997). Slow negative brain

(10)

potentials as reflections of specific modular resources of cognition.Biological Psychology,45, 109–141.

Rößner, P., Rockstroh, B., Cohen, R., Wagner, M., & Elbert, T.

(1999) Event-related potential correlates of proactive interference in schizophrenic patients and controls.Psychophysiology, 36, 199–208.

Roth, W.T., & Cannon, E.H. (1972). Some features of the auditory evoked response in schizophrenics.Archives of General Psy- chiatry, 27, 466–471.

Roth, W.T., Goodale, J., & Pfefferbaum, A. (1991). Auditory event-related potentials and electrodermal activity in medicat- ed and unmedicated schizophrenics.Biological Psychiatry, 29, 585–599.

Roth, W.T., Horvath, T.B., Pfefferbaum, A., & Kopell, B.S.

(1980). Event-related potentials in schizophrenics.Electroen- cephalography and Clinical Neurophysiology,48, 127–139.

Rugg, M.D. (1995). Event-related potential studies of human memory. In M.S. Gazzaniga (Ed.),The cognitive neurosciences (pp. 789–801). Cambridge: MIT Press.

Rugg, M.D., Mark, R.E., Walla, P., Schloerscheidt, A.M., Birch, C.S., & Allan, K. (1998). Dissociation of the neural correlates of implicit and explicit memory.Nature,392, 595–598.

Rugg, M.D., & Nagy, M.E. (1989). Event-related potentials and recognition memory for words.Electroencephalography and Clinical Neurophysiology,72, 395–406.

Ruoff, A. (1981).Häufigkeitswörterbuch gesprochener Sprache.

Tübingen: Niemeyer.

Salisbury, D.F., Shenton, M.E., Sherwood, A.R., Fischer, I.A., Urgelun-Todd, D.A., Tohen, M., & McCarley, R.W. (1998).

First-episode schizophrenic psychosis differs from first-epi- sodes affective psychosis and controls in P300 amplitude over left temporal lobe.Archives of General Psychiatry,55, 173–

180.

Saykin, A.J., Gur, R.C., Gur, R.E., Mozley, P.D., Mozley, L.H., Resnick, S.M., Kester, D.B., & Stafiniak, P. (1991). Neuropsy- chological function in schizophrenia.Archives of General Psy- chiatry, 48, 618–624.

Saykin, A.J., Shtasel, D.L., Gur, R.E., Kester, D.B., Mozley, L.H., Stafiniak, P., & Gur, R.C. (1994). Neuropsychological deficits in neuroleptic naive patients with first-episode schizophrenia.

Archives of General Psychiatry,51, 124–131.

Shagass, C., Josiassen, R.C., Roemer, R.A., & Straumanis, J.J.(1986). Do differing intelligence levels account for EEG and evoked potential differences between chronic schizophrenics and non-patient controls? In C. Shagass, R.C. Josiassen, &

R.A. Roemer (Eds.),Brain electrical potentials and psychopa- thology. Amsterdam: Elsevier.

Smith, M.E., & Halgren, E. (1989). Dissociation of recognition memory components following temporal lobe lesions. JEP:

Learning, Memory, Cognition, 15, 50–60.

Strandburg, .J., Marsh, J.T., Brown, W.S., Asarnow, R.F., Guthrie, D., Higa J., Yee-Bradbury, C., & Nuechterlein, K. (1994). Re- duced attention-related negative potentials in schizophrenic adults.Psychophysiology,31, 272–281.

Uhl, F., Franzen, P., Serles, W., Lang, W., Lindinger, G., & Deecke, L. (1990). Anterior frontal cortex and the effect of proactive interference in paired associate learning. A DC potential study.

Journal of Cognitive Neuroscience,2, 373–382.

Wing, J.K., Cooper, J.E., & Sartorius, N. (1974).Measurement and classification of psychiatric symptoms. London: Cam- bridge University Press.

Prof. Dr. Brigitte Rockstroh Fachgruppe Psychologie Universität Konstanz P.O. Box D23 D-78457 Konstanz Germany

Tel. +49 7531 882085 Fax +49 7531 882891

E-mail brigitte.rockstroh@uni-konstanz.de