19

Psychiatry Research, 36:19-36 Elsevier

The Structure of Schizotypy:

A Pilot Multitrait

Kenneth S. Kendler. Alfred L. Ochs, Anne David E. Ross, and Allan F. Mirsky Received 1990.

M. Gorman,

June 18, 1990; revised version received September

Twin Study John

K. Hewitt,

10, 1990; accepted November

3,

Abstract. This report of a pilot study examines 29 pairs of twins from a population-based registry on whom four domains of schizotypy have been measured: personal interview using the Structured Interview for Schizotypy, selfreport questionnaire formed from eight published self-report scales, attentional battery of eight individual tests, and root mean square error on smooth pursuit eye tracking. Analyzing the twins as individuals revealed two independent dimensions of clinically rated schizotypy (positive symptom schizotypy and negative symptom schizotypy) and two independent dimensions of self-rated schizotypy (positive trait schizotypy and trait anhedonia). Positive symptom schizotypy was highly correlated with positive trait schizotypy, but not with attentional dysfunction or eye-tracking error. By contrast, negative symptom schizotypy was significantly related to trait anhedonia, attentional dysfunction, and eye-tracking error. Correlations in monozygotic and dizygotic twins suggested that genetic factors were important in all four domains of schizotypy. Except for eye-tracking error, the results are more consistent with a dimensional than a “disease” model of schizotypy. Replication of these results with a larger group of subjects is needed. Key Words. Schizotypy, attention, personality disorder, twin studies.

smooth

pursuit

eye movements,

schizotypal

Several lines of evidence suggest that the vulnerability to schizophrenia, which we define (after Rado, 1953) as schizotypy, can manifest itself in ways other than the full syndrome of deteriorating psychotic disorder: 1. Consistent with early clinical observations (Kendler, 1985), five blind, controlled family or adoption studies using DSM-ZZZcriteria found substantially higher rates for schizophrenia-like personality disorders (schizotypal, schizoid, and paranoid personality disorder) in relatives of schizophrenic vs. control probands (Lowing et al., 1983; Kendler and Gruenberg, 1984; Baron et al., 1985; Frangos et al., 1985; Gershon et al., 1988). 2. Two studies report excess rates of schizophrenia in relatives of individuals either

Kenneth S. Kendler, M.D., is Professor of Psychiatry and Human Genetics; Alfred L. Ochs, Ph.D., is Professor of Neurology; Anne M. Gorman, B.A., is Research Associate, Department of Psychiatry; John K. Hewitt, Ph.D., is Associate Professor of Human Genetics; and David E. Ross, M.D., is Resident, Department of Psychiatry, Medical College of Virginia, Richmond, VA. Allan F. Mirsky, Ph.D., is Chief, Laboratory of Psychology and Psychopathology, National Institute of Mental Health, Bethesda, MD. (Reprint requests to Dr. K.S. Kendler, Dept. of Psychiatry, Box 710 MCV Station, Richmond, VA 23298-0710, USA.) 0165-1781/91/$03.50

@ 1991 Elsevier Scientific

Publishers

Ireland

Ltd.

20 meeting DSM-ZZZcriteria for schizotypal personality disorder (Schulz et al., 1989) or having high levels of perceptual aberration, a postulated psychometric index of schizotypy (Lenzenweger and Loranger, 1989). 3. Several neuropsychological and neurophysiological indices that are abnormal in many individuals with schizophrenia are also abnormal in a subset of their clinically unaffected relatives. This is particularly true for eye-movement dysfunction (e.g., Holzman et al., 1974, 1976, 1977, 1984, 1988; Mather, 1985) and abnormalities in attention, as measured by the Continuous Performance Test (Cornblatt and Erlenmeyer-Kimling, 1985; Nuechterlein, 1983, 1985) or the arithmetic and digitspan subtests of the Wechsler Adult Intelligence Scale (WAIS) (Mirsky, 1988). Other potential indices of schizotypy including neurologic soft signs (Kinney et al., 1986), evoked response abnormalities (Siegel et al., 1984), and thought disorder (Shenton et al., 1989) may also be abnormal in unaffected relatives of schizophrenic patients. Given the evidence that clinical, psychometric, and neuropsychological/ neurophysiological indices of schizotypy are abnormal in certain relatives of individuals with schizophrenia, three further questions become critical: (1) Is schizotypy best conceptualized as a normally distributed trait or as a discontinuous “disease state”? (2) Is schizotypy a unidimensional or multidimensional concept? (3) What are the sources of individual differences in these various indices of schizotypy? In particular, do these indices aggregate in families and, if so, is this due to genetic or familial-environmental influences? The present article addresses these questions in a preliminary fashion using results of a pilot multitrait study of schizotypy in unselected twins from a population-based twin registry.

Methods Ascertainment of Twins and Zygosity Determination. Twins were selected from the Virginia Twin Registry, a population-based registry in which twins are ascertained from a screening of birth certificates within the Commonwealth of Virginia. Criteria for entry into this study were as follows: (1) Caucasian, (2) age 18 to 50, (3) both members residing within 30 miles of Richmond, and (4) neither member having any neurological condition that would potentially interfere with the test performance. Sixty-nine eligible twin pairs were contacted, of whom 16 refused participation. Of the remaining 53 pairs, resources were available to study 29 (42%) of them. These twin pairs were invited to the Medical College of Virginia (MCV) and offered travel expenses and a modest honorarium. For same-sex twin pairs, zygosity was determined by a review of photographs and selfreport information about frequency of being mistaken for one another and degree of physical resemblance. This method has been shown to be at least 95% accurate (Eaves et al., 1989). In four pairs, where the review of zygosity information was inconclusive, blood samples were obtained from both members for zygosity testing, involving eight highly polymorphic DNA markers, which, if all concordant, yield a posterior probability of monozygosity of 0.9997. Three of these pairs were monozygotic (MZ) and one dizygotic (DZ). The final sample consisted of 13 complete MZ pairs, all of which were female-female, and 16 complete DZ pairs, 7 of whom were same-sex (4 female-female and 3 male-male) and 9 of whom were opposite-sex. The mean + SD of the age and years of education of the sample were 33.4 * 7.5 and 14.3 + 1.9, respectively. Measures of Schizotypy. The following four domains of schizotypy were assessed: 1. Schizotypal symptoms and signs were measured using the Structured Interview for Schizotypy (SK). The development, reliability, and validity of this instrument are outlined

21

elsewhere (Kendler et al., 1989). The SIS is a highly structured instrument that examines a wide variety of schizotypal signs and symptoms, including all DSM-III-R criteria for schizotypal personality disorder. In this report, we examine global schizotypal symptoms (measured on a 1- to 7-point scale: 1 = marked, 3 = moderate, 5 = mild, and 7 = none) and signs (measured on a 1- to 5-point scale: 1 = very poor, 2 = poor, 3 = fair, 4 = good, and 5 = excellent). The mean duration of the SIS interview was 39.6 f 8.8 min. The SIS was completed on all 58 twins participating in this project. The majority of the SIS interviews were conducted by A.G., who had been trained by the interview’s developer, K.S.K. The remainder of the SIS interviews were conducted by K.S.K. 2. Schizotypal personality traits were measured by a self-report questionnaire (SRQ) for “schizotypy.“This instrument, developed at MCV, contains 244 items and covers a wide range of published scales that attempt to assess schizotypy: Schizotypy-A (STA) scale (Claridge and Broks, 1984), Launay-Slade Hallucination Scale (Launay and Slade, 1981), Minnesota Multiphasic Personality Inventory (MMPI) “schizoid taxon” (Meehl, 1962), and shortened scales empirically developed with the assistance of Loren Chapman from the full-length Social Nonconformity, Physical Anhedonia, Social Anhedonia, Magical Ideation, and Perceptual Aberration scales (Chapman et al., 1976, 1982; Chapman and Chapman, 1987). For our analyses, we used three subscales of the STA scale previously identified by Hewitt and Claridge (1989): magical ideation, perceptual aberration, paranoid ideation/ suspiciousness, and one subscale including the remaining STA items. Most subjects were given the SRQ to complete at home. Completed SRQs were received from only 44 twins and were available on both members of only 8 MZ and 10 DZ pairs. The third major domain of schizophrenia was measured by an attentional battery consisting of eight neuropsychological tests chosen by A.M. to assess attention. In the order administered, these tests were the Continuous Performance Test (CPT) X task, AX task, and degraded stimulus task; the Trail Making Test; the Stroop Color-Word Interference Test; the Wechsler Adult Intelligence Scale-Revised (WAIS-R) digit span, digit symbol substitution, letter cancellation, and arithmetic tasks; and the Wisconsin Card Sorting Test. Schizophrenic patients have been shown to be impaired on many of these tests (Mirsky, 1988, 1989), which were selected from a larger battery currently in use in the Laboratory of Psychology and Psychopathology at the National Institute of Mental Health. A completed attentional battery was available in 57 subjects (in one subject, the computerized CPT apparatus malfunctioned). The attentional battery was administered to all twins by A.G., who had been trained in its use by members of Dr. Mirsky’s laboratory. The fourth domain of schizotypy assessed was smooth pursuit eye movements. A detailed description of the methods used appears in Ross et al. (1988). In brief, infrared oculography was used to measure eye movements in response to a computer-controlled laser stimulus projected onto a viewing screen. The stimulus consisted of a constant velocity ramp to confirm linearity, calibration steps, and six cycles of a 0.3 Hz sine wave subtending 20 o of visual angle around the midline. Data were recorded and analyzed digitally. The root-mean-square (RMS) error was chosen for this analysis as a global measure of tracking errors from all sources, with the exception of blink artifacts, which were removed during the analysis phase. Eye tracking was available on 57 individuals; in one individual, technical problems rendered the eyetracking record uninterpretable. The measurement of smooth pursuit eye movements was carried out on all twins by A.O. and analyzed by D.R. The average duration of the testing performed at MCV was 3 hours and 35 min. This included administrations of subsections of the Structured Clinical Interview for DSM-ZZZ (Spitzer et al., 1987). No twin was known to have had a psychotic illness. Statistical Analysis. Orthogonal and oblique factor analyses were carried out using the VARIMAX and PROMAX routines in SAS (1985). The number of meaningful factors was determined by analysis of the scree plot. Pearson product-moment correlations were used to examine the relationship between putative indices of schizotypy. For these analyses, twins were treated as individuals. Normality was tested by the Kolomogorov D statistic (abbreviated

22 as D) when the sample size was 2 50, and by the Shapiro-Wilk’s statistic (abbreviated as W) when the sample size was smaller. One-tailed p values were used when clear a priori directional hypotheses were present (e.g., that attentional dysfunction and eye-tracking abnormalities would be positively correlated with measures of schizotypy) and when reporting the correlations in twin pairs (where only positive correlations were expected). For all other comparison, two-tailed p values are reported.

Results Frequency of Schizotypal Traits and Personality Disorder. Table 1 shows the distribution of scores for the 58 interviewed twins on the nine DSM-III-R criteria (American Psychiatric Association, 1987) for schizotypal personality disorder (6 symptoms and 3 signs). While mild symptoms were relatively common, symptom scores in the moderate or more severe range were uncommon. To be accepted as present, a sign had to be rated poor or fair, while a symptom had to be scored 4 or lower. On this basis, only three twins (5.2 + SD 2.9%) met DSM-III-R criteria for schizotypal personality disorder (5 of 9 criteria). Four additional twins met four criteria and five met three. We also used a definition of schizotypal traits which required that a symptom be scored 5 or below to be considered present. By this definition, 23 of the twins (39.7 f 6.4%) met 5 of 9 criteria.

Table 1. Distribution of scores on the 9 DSM-III-R oersonalitv disorder Symptoms Ideas of reference Social anxiety Magical

thinking

Illusions Social

isolation

Suspiciousness

Signs Affect Organization Oddness

of speech

53 Moderate

4

criteria for schizotypal

5 Mild

6

7 None

2

14

36

21

12

19

45

17

28 7

2

14

38

17

29

2

9

28

40

22

2

3

21

16

59

3

14

47

28

9

Poor

Fair

Good

0

26

35

14

0

7

47

47

0

10

21

69

Excellent

Structure of Schizotypy. SIS. Table 2 presents the 19 key global schizotypal symptoms and signs from the SIS that were used for factor analysis. The scree plot indicated two significant factors. In addition, Table 2 shows the factor loadings derived from an orthogonal VARIMAX rotation. An oblique rotation using PROMAX was also done. The resulting factors, which correlated with one another only modestly (+0.12), closely resembled those found with the orthogonal rotation. The first factor, termed ‘positive symptom schizotypy,” accounted for 21.8% of the total variance and had highest loadings on the schizotypal symptoms that

23

Table 2. Factor analysis of the 19 key global symptoms and signs from the Structured Interview for Schizotypy VARIMAX rotated factors Positive symptom schizotypy

Negative symptom schizotypy

Ideas of reference

0.77

0.05

Illusions

0.69

0.03

Suspiciousness

0.68

0.39

Magical

0.66

-0.26

thinking

0.62

-0.03

0.60

0.14

instability

0.56

0.03

DerealizatiorVdepersonalization

0.53

0.07

Organization

0.53

0.13

0.47

0.37

-0.04

0.78

-0.14

0.73

0.08

0.67

0.14

0.60

Oddness*

0.11

0.59

Social isolation

0.10

0.41 0.40

Psychotic-like

symptoms

Sensitivity Affective

of speech”

Social anxiety Rapport* Fullness Restricted

of affect’ emotions*

Appropriateness

of affect’

Irritability

0.40

Guardedness*

0.07

0.10

lmoulsivitv

0.33

-0.44

Note. A schizotypal sign rated cmthe basis of the respondent’s behavior during the interview is indicated by an asterisk. Other items are schizotypal symptoms, which were rated on the basis of the respondent’s self-report

describe, in muted form, the positive psychotic symptoms of schizophrenia: ideas magical thinking, and “psychotic-like of reference, illusions, suspiciousness, symptoms.” Lower loadings on this first factor were found on symptoms that and social anxiety). One reflected avoidant personality traits (e.g., sensitivity schizotypal sign that loaded substantially on this factor was organization of speech. The distribution of this factor score in the twins was normal (D = 0.08, p > 0.10). The second factor, termed “negative symptom schizotypy,” accounted for 17.1% of the variance and had highest loadings on schizotypal signs that reflect, in attenuated form, the negative symptoms of schizophrenia: poor rapport, flatness of affect, restricted emotion, inappropriateness of affect, and global oddness. Four schizotypal symptoms had substantial positive loadings on this factor: social isolation, suspiciousness, irritability, and social anxiety: one-impulsivity-had a substantial negative loading. This factor score was normally distributed in the twins (D = 0.07, p > 0.10). SRQ. Table 3 shows the results of an orthogonal VARIMAX rotation of the individual scales from the SRQ. Two factors were identified, termed positive trait schizotypy and trait anhedonia. The positive trait schizotypy factor, which accounted for 47.2% of the total variance, had highest loadings on scales measuring “psychotic-like” experiences: the Launay-Slade Hallucination Scale, the shortened

24 Table 3. Factor analysis of the Self-Report

Questionnaire

for schizotypy

VARIMAX rotated factors Positive trait schirotypy

Trait anhedonia

0.86

-0.18

0.45

0.13

0.83

-0.09

0.83

-0.04

0.11

0.81

-0.17

0.75

0.61

0.01

0.75

0.05

0.85

0.05

Ideation

0.68

0.48

items

0.83

0.11

Scale Launay-Slade MMPI

Hallucination

“schizoid

Scale

taxon”

Shortened scales of Chapman et al. Perceptual Magical

Aberration

Ideation

Social Anhedonia Physical

Anhedonia

Nonconformity Subscales of STA Scale of Claridge and Broks Magical

Ideation

Perceptual Paranoid Remaining

Aberration

Note. MMPI = Minnesota Multiphasic Personality Inventory. STA = Schizotypy-A.

Perceptual Aberration and Magical Ideation scales of Chapman et al., and the derived subscales from Claridge’s STA scale: magical ideation and perceptual aberration. Lower, but still substantial loadings were also found on thepositive trait schizotypy factor for the nonconformity scale of Chapman et al. and the paranoid ideation subscale of the STA. This factor score had a significant positive skew in the twin sample (W = 0.86, p < O.Ol), which was eliminated by a log transformation (W = 0.96, p > 0.10). The trait anhedonia factor, which accounted for 13.7% of the variance, had very high loadings on two scales-the Social Anhedonia and Physical Anhedonia scales of Chapman et al.-and a modest loading on Claridge’s paranoid ideation subscale. This factor was normally distributed in the twin sample (W = 0.95, p > 0.10). Results from the SRQ were also analyzed by an oblique PROMAX rotation. The resulting two factors, which were uncorrelated with one another (r = 0.03) were identical with those found by orthogonal rotation. Attentional battery. The analyses presented here are restricted to the first principal component of the attentional battery, which is the best summary of overall attentional performance. This factor, which had highest loadings on measures of Wisconsin Card Sorting performance, CPT yc commission errors, and WAIS arithmetic score, had a modest, but not statistically significant positive skewness in the twin sample (D = 0.11, p = 0.07). Smooth pursuit eye movements. The distribution of eye-tracking RMS error in the twin sample was significantly nonnormal (D = 0.27, p < 0.01). This was due to bimodality which could not be removed by transformation. Twins were divided into good trackers (n = 52, mean f SD of RMS = 0.54 f 0.22, range 0.25-l .35) and poor Cross-trait trackers (n = 5, RMS scores = 2.54 + 0.42, range 2.14-3.08). relationships with RMS scores were assessed by analysis of covariance, with twins divided into good and poor trackers.

25 Cross-trait correlations. Age was significantly correlated with negative trait schizotypy (r = 0.30, p = 0.02) and trait anhedonia factors (r = 0.43, p = 0.003). Years of education was significantly correlated with attentional dysfunction (r = 0.42, p = O.OOl), positive symptom schizotypy (r = -0.35, p = 0.007), and positive trait schizotypy (r = -0.44,~ = 0.003). Therefore, Table 4 presents intertrait correlations, with the effect of age and years of education removed. In addition, all

Table 4. Cross-trait correlations for the putative indices of schizotypy with effects of aae and vears of education removed 1

2

3

4

SE 1. Positive 2. Negative

symptom symptom

schizotypy schizotypy

-0.02

SRQ 3. Positive-trait

schizotypy’

4. Trait anhedonia

0.834

0.03

0.29

0.453

0.16

-

0.02

0.312

-0.14

0.15

Attentional battery 5. First principal

component

Note. SIS = Structured Interview for Schizotypy. SRQ = Self-Report Questionnaire. Sample size = 43 or 44 for correlations involving the SRQ; otherwise, 56 or 57. 1. Log-transformed. 2. p < 0.05, 1-tailed. 3. p < 0.01, 1-tailed. 4. p < 0.0001, 1-tailed.

factors are scaled so that increasing scores indicate greater pathology. The following results are of particular interest: 1. The positive symptom schizotypy factor from the SIS was highly correlated with the transformed positive trait schizotypy factor from the SRQ (0.83, p < 0.0001). This high intertrait correlation was not the result of one or two outliers (Fig. 1). Positive symptom schizotypy was also correlated with the second self-report factor, trait anhedonia, but the correlation was much more modest. 2. The negative symptom schizotypy factor from the SIS was significantly correlated with the trait anhedonia factor (0.45, p = 0.001) and attentional dysfunction (0.31, p = 0.01). Fig. 2 shows the relationship between negative symptom schizotypy and attentional dysfunction. 3. Neither the positive trait schizotypy nor the trait anhedonia SRQ was significantly correlated with attentional dysfunction.

factor

from the

Eye tracking and other indices of schizotypy. Analysis of covariance, taking into account the impact of age and years of education, indicated that good and poor trackers did not significantly differ on positive symptom schizotypy (t = 0.71, NS), transformed positive trait schizotypy (t = 0.06, NS), trait anhedonia (t = 1.56, NS), or the first principal component of the attentional battery (t = 1.50, NS). However, compared to good eye trackers, poor eye trackers had significantly higher levels of negative symptom schizotypy (t = 2.08, p = 0.04).

26 Fig. 1. Relationship in 58 twins between the positive symptom schizo typy factor and the log-transformed positive trait schizotypy factor P 0 I i 1 i ” e

’ * 3

2

s Y m P

1

s c h i 2

0

6 -1 1 V P I

0.0

0.2

’

I

0.4 Positive

’

I

0.6 Trait

’

I

’

I

0.6 1.0 Schizotypy

r

I

r

I

’

I

1 .6

I .4 1.2 - Tran6fPrned

’

I1.6

The positive symptom schizotypy factor was assessed at personal interview by the Structured Interview for Schizotypy. The log-transformed positive trait schizotypy factor was assessed by self-report questionnaire. The observed correlation was +0.83 @ < 0.0001).

Fig. 2. Relationship in 58 twins between negative symptom attentional function * * ” ‘1

-2

-1 1st

Negative

symptom

schizotypy

Attentional function was &eased was +0.31 fp < 0.05).

P.C.

0 of

2

1 Altentionol

schizotypy

and

3

Battery

was assessed at personal interview by the Structured Interview for Schizotypy. by the first principal component of the attentional battery. The observed correlation

27 Concordances and Correlations. All three twins meeting DSM-III-R criteria for schizotypal personality disorder were MZ, and two came from one pair, for a probandwise concordance of 213 = 66.7%. A more meaningful analysis was possible for schizotypal traits, as defined above, where the probandwise concordance rate in MZ twins (12/ 14 = 85.7%) was much higher than that found in DZ twins (2/9 = 22.2%). Correlations in MZ twins substantially exceeded those in DZ twins for both the positive symptom and negative symptom schizotypy factors from the SIS (Table 5). Twin

Table 5. Twin correlations schizotypy

for the various

putative MZ

SIS Positive Negative

(4 symptom

schizotypy

symptom

schizotypy

SRQ Positive-trait

(n) schizotypyl

Trait anhedonia

Attentional First principal

battery component

(n)

indices DZ

13

16

0.38

0.09

0.61 2

0.462

8

of

10

0.743

-0.03

0.974

-0.18

12

16

0.522

0.05

Note. SIS = Structured Interview for Schizotypy. SRQ = Self-Report Questionnaire. 1. Log-transformed. 2. p < 0.05, 1 -tailed. 3. p < 0.01,1 -tailed. 4. p < 0.0001, 1-tailed.

Of interest, correlations for the negative symptom schizotypy factor were higher in both MZ and DZ twins than those found for the positive symptom factor. Fig. 3 shows the distribution of scores of MZ twin pairs for negative symptom schizotypy. Correlations for both thepositive trait schizotypy and the anhedonia factors in the eight MZ twin pairs in which both twins returned self-report questionnaires were very high. By contrast, the correlations in these factors for the 10 available pairs of DZ twins were low and negative. The correlation in MZ twin pairs for the first principal component of the attentional battery also substantially exceeded that found in DZ twin pairs (Table 5). Of the five twins who had poor eye tracking, four were MZ and came from two concordant pairs. Probandwise concordance for poor eye tracking in MZ twins was therefore 4/4 = 100%. One DZ twin had poor eye tracking, and the co-twin from that pair had good eye tracking (concordance of O/ 1 = 0%). Discussion or Disease? The prevalence of definite schizotypal personality disorder as defined by DSM-III-R in this population-based sample of twins (5.2 f 2.9%) is within the range found in the relatives of controls from previous family and adoption studies that have used instruments specifically designed to assess schizotypy (Kendler and Gruenberg, 1984; Baron et al., 1985). The Nature of Schizotypy-Dimension

28

Fig. 3. Relationship between the negative symptom schizofypy score in both members of 13 monozygotic twin pairs

11

-

Neqotive

The negative symptom schizofypy score was assessed Schizotypy. The observed correlation was +0.61 lp < 0.01).

Synp

Schizolypy

at personal

interview

by the Structured

Interview for

However, a large proportion of this nonclinical sample had broadly defined schizotypal traits. Both factors based on the personal interview were normally distributed in this sample of twins with no evidence of bimodality. Although limited by the modest sample size, the results reported here are more consistent with the dimensional rather than the disease concept of schizotypy. The Structure of Schizotypy-The Personal Interview. Historically, there have been two divergent concepts of schizotypy (Kendler, 1985). Genetically oriented psychiatrists have developed a “familial” perspective on schizotypy, a syndrome that they felt characterized the deviant but nonpsychotic relatives of schizophrenic patients. Psychodynamically oriented psychiatrists developed a “clinical” view of schizotypy characterizing patients who demonstrated latent symptoms of schizophrenia. In the last decade, two groups of investigators have suggested that schizotypy represents two psychopathological dimensions: a “negative” and “positive” dimension (Kendler et al., 1983), or a “cognitive-perceptual” and “social-interpersonal” dimension (Sever and Gunderson, 1983). This study, which represents, to the best of our knowledge, the first examination of the structure of schizotypy as assessed by interview in a nonclinical population, supports the position that schizotypy represents more than one psychopathological dimension. Two independent factors emerged from the sample. The first or positive symptom schizotypy factor had highest loadings on psychotic-like symptoms,

29 organization of speech, and the avoidant traits of interpersonal sensitivity and social anxiety. This factor resembles the “cognitive-perceptual” dimension proposed by Siever and Gunderson (1983) more closely than the “positive symptom” dimension proposed by Kendler et al. (1983), because of the substantial positive loadings for thought disorder and suspiciousness predicted by the former but not by the latter investigators. Contrary to the predictions of both of these groups of authors, however, “interpersonal” avoidant-like symptoms also loaded substantially on this factor. The second or negative symptom schizotypy factor had highest loadings on signs that reflected rapport, modulation of affect, and oddness and symptoms that reflected social isolation, irritability, and low levels of impulsivity. Suspiciousness and social anxiety also loaded substantially on this factor, but at lower levels than on the first factor. While this factor resembles both “negative symptom” (Kendler et al., 1983) and “social-interpersonal” schizotypy (Siever and Gunderson, 1983) it even more closely approximates the historical view of schizotypy as articulated by the “familial tradition” (Kendler, 1985). It is of interest to compare these results with studies that examined the power of the individual schizotypal criterion to discriminate individuals with schizotypal personality disorder from other patient groups (Jacobsberg et al., 1986; Widiger et al., 1987). These studies, which found that both “positive” and “negative” schizotypal features were important for identifying individuals with schizotypal personality disorder in a clinical setting, suggest that current criteria for schizotypal personality disorder identify individuals with high levels of both positive and negative symptom schizotypy. The Structure of Schizotypy-The Self-Report Questionnaire. The last 15 years have seen a proliferation of self-report questionnaires designed to measure various aspects of “schizotypy” or “psychosis-proneness.” Do these various scales measure the same underlying trait? Chapman et al. (1982) found, in a population of college undergraduates, that Perceptual Aberration and Magical Ideation were highly intercorrelated. Both scales correlated substantially (around 0.45) with nonconformity, but they had a low and negative correlation with Physical Anhedonia. In both an unselected population ascertained at a shopping mall and schizophrenic outpatients, Chapman et al. (1976) found high intercorrelations in measures of Physical and Social Anhedonia. Consistent with these results, the first factor found here, termed positive trait schizotypy, had high loadings on “positive schizotypal” or “psychotic-like” scales from three research groups: the Hallucination scale of Launay and Slade (198 l), the shortened Perceptual Aberration and Magical Ideation scales of Chapman et al. (1982, 1987) and the empirically derived magical ideation and perceptual aberration subscales (Hewitt and Claridge, 1989) of the STA scale of Claridge and Broks (1984). Also in accord with the previous findings of Chapman and colleagues (Chapman et al., 1982; Chapman and Chapman, 1987) we found a distinct second trait anhedonia factor that had high loadings on only the Physical and Social Anhedonia scales of Chapman et al. (1976) and a modest loading on the paranoid ideation subscale of the STA. These results are similar to those found using our self-report

30

measure from over 400 twins from the Virginia Registry (Kendler and Hewitt, in review), except that in the larger sample we were able to detect separate factors for physical and social anhedonia. Intertrait Correlations. One goal of this research was to begin to understand the interrelationships between the multiple putative indices of schizotypy. In particular, would the structure of schizotypy, as revealed by the SIS, be validated by the SRQ, attentional battery, and eye-tracking measures? The SE and the SRQ. The most striking finding concerning the relationship between interview-assessed and self-report measures of schizotypy was the very high correlation between the positive symptom schizotypy factor that emerged from the clinical interview and the positive-trait schizotypy factor that emerged from the self-report measure. The results appear to suggest that the more time-consuming personal interview for “positive symptoms” of schizotypy may be replaced by a self-report questionnaire. However, given the small sample size and the relative cultural homogeneity of the sample, such a conclusion would be premature. Comparisons of the interview-based and self-report assessments of schizotypy also revealed a moderate correlation between the negative symptom schizotypy and traitanhedonia factors. This correlation was unexpected; we had assumed that negative symptom schizotypy, formed largely from signs rated by the interviewer, would tap a dimension of schizotypy that could not be measured by self-report. Because social isolation is part of the negative symptom schizotypy factor, and a correlation between social anhedonia and social isolation would be expected and has been reported (Mishlove and Chapman, 1985), it appeared that the observed correlation between the trait anhedonia and negative symptom schizotypy factors might simply reflect the correlation between trait anhedonia and the social isolation symptom on the SIS. Further analysis indicated, however, that this clearly was not the case (results available). The trait anhedoniu factor may reflect, to a moderate extent, the underlying deficit that was detected as negative schizotypal symptoms at personal interview. Measures of schizotypy and the attentional battery. Given that attentional function is disturbed in many individuals with schizophrenia (Nuechterlein and Dawson, 1984) and in a proportion of their unaffected relatives (Nuechterlein, 1983, 1985; Cornblatt and Erlenmeyer-Kimling, 1985; Mirsky, 1988), it is of interest to determine whether attentional function is correlated in the general population with clinical or questionnaire-derived indices of schizotypy. The one previous study that examined this question (Asarnow et al., 1983) found significant relationships between performance on the Span of Apprehension test and several measures of schizotypy. The summary measure of attentional dysfunction was significantly correlated with only one of the four factors from the interview and self-report assessments: the negative symptom schizotypy factor. This result supports the validity of the distinction between negative symptom schizotypy and the other three clinically derived schizotypal factors. Measures of schizotypy and eye tracking. There is now considerable evidence that a substantial proportion of individuals with schizophrenia and their unaffected

31 relatives demonstrate poor accuracy in smooth pursuit eye movements. In this report, we address a related question: in the general population, is poor eye tracking related to interview- or questionnaire-based assessments of schizotypy? Five studies have addressed this question. Iacono and Lykken (1979) found, in 32 pairs of MZ twins, that psychoticism as measured on the Differential Personality Questionnaire correlated significantly with poor tracking. Siever et al. (1982), in two samples from a college population, found a consistent but modest relationship between poor tracking and the social introversion scale of the MMPI. Siever et al. (1984) identified from a large college population subsets of low-accuracy and highaccuracy trackers. Compared to the high-accuracy trackers, the low-accuracy trackers had higher rates of schizotypal personality disorder; showed higher levels of schizotypal traits, including eccentricity and reduced social interactions; and scored higher on Chapman’s Perceptual Aberration Scale (Siever et al., 1984, 1989; Coursey et al., 1989). Simons and Katkin (1985) examined smooth pursuit eye movements in college students selected for high and low scores on the Physical Anhedonia or Perceptual Aberration scales of Chapman and colleagues. In contrast to the other studies, Simons and Katkin.(1985) found no significant differences in the qualitative assessment of smooth pursuit eye movements in the schizotypal versus control groups. The schizotypal groups did, however, have significantly greater variance, and the poorest trackers were nearly all from the schizotypal groups. Finally, Siever et al. (1990) found significantly more abnormalities in smooth pursuit eye movements in outpatients with schizotypal personality disorder than in matched controls. The global index of eye-tracking efficiency in this study was bimodally distributed. Poor trackers differed significantly from good trackers on only one of the factors: negative symptom schizotypy. These findings further validate the dimension of negative symptom schizotypy assessed by the SIS and suggest that it is meaningfully different from the other three schizotypal factors. The examination of twin correlations in this study provides a very rough estimate of the sources of variation for the traits under consideration, as the sample is too small to allow for meaningful model fitting. Instead, the investigation is guided by the rough rule of thumb that when rMz > 2rDZ, this indicates that familial resemblance is due largely or entirely to genetic factors, but then when rMz < 2rDZ, this suggests that familial resemblance results from both genetic and familialenvironmental factors. Heritability (which can be roughly estimated as either 2(rMz - rDZ) or rMZ, whichever is lower) can, with this sample, only be meaningfully divided into high (e.g. > 0.60), moderate (0.30-0.60) and low (< 0.30). Interview- and questionnaire-assessed schizotypy. A large number of twin studies of the personality dimension of “psychoticism” or related constructs consistently find evidence of moderate heritability (Eaves et al., 1989). However, the relationship between the personality trait of psychoticism and current concepts of schizotypy is unclear. Only four twin studies appear to be directly relevant to the concept of schizotypy as used in this article. In a study of personally interviewed twins ascertained through psychiatric facilities, Torgersen (1983) found probandwise concordance for schizotypal personality disorder to be 33% in MZ twins and 4% in Twin Correlations.

32 DZ twins @ < 0.02). Claridge and Hewitt (1987) administered the “Schizotypal Questionnaire” (Claridge and Broks, 1984) to 210 twin pairs. Examining the STA scale, they found the best fitting model included only individual specific environment and additive genetic variation with heritability estimated at 53 + 6.5%. Kendler et al. (1987) examined an empirically derived “suspiciousness” factor from Eysenck’s Personality Questionnaire, given to 3,810 Australian twin pairs, and also found the best fitting model included only additive gene action and specific environment, producing an estimate of heritability of 41%. Finally, in a sample of 133 twin pairs from the Virginia Twin Registry, Kendler and Hewitt (in review) found heritability estimates of over 50% for three schizotypal factors that they called positive trait schizotypy, nonconformity, and social schizotypy. In the present study, MZ twin correlations substantially exceeded DZ twin correlations for all four factors derived from the interview and the questionnaire. However, the pattern of these correlations was relatively different for each factor. For positive symptom schizotypy, the pattern of correlations suggested a trait with moderate heritability, where familial aggregation resulted entirely from genetic factors. For negative symptom schizotypy, the twin correlations also suggested a trait of moderate heritability, but here familial aggregation would appear to result from both genetic and familial-environmental factors. By contrast, for both of the questionnaire-based factors (positive trait schizotypy and trait anhedonia), the particularly small sample suggests a trait of high heritability where familial aggregation is due entirely to genetic factors. The results, therefore, are consistent with previous studies in suggesting that genetic factors play a significant role in the expression of schizotypal characteristics in the general population. The attentional battery. To our knowledge, no twin studies have been done of the kind of measure of attention that has been used in the study of schizophrenic patients and their relatives. However, Cornblatt et al. (1988) recently reported, from intact normal families, that 40-50% of the variance in CPT performance was due to familial factors (e.g., genes, family environment, or a combination). On the basis of the first principal component of the attentional battery, broadly assessed attention in the twin sample appeared to be a trait with moderate heritability in which genetic factors appear to be responsible for familial resemblance. Eye tracking. There has been one study of smooth pursuit eye movements in unselected MZ and DZ twins (Iacono, 1982). In 58 pairs of twins, Iacono found evidence for substantial genetic influences on RMS error (rMz = 0.68, roz = 0.35). These results are broadly consistent with a study of smooth pursuit eye movement in pairs of twins ascertained through a proband affected with schizophrenia (Holzman et al., 1977). In the present sample, RMS error was bimodally distributed. Concordance for poor eye tracking, based on very small samples, was much higher in MZ than DZ twins. When the relationship of RMS values in twin pairs was analyzed using product-moment correlations, the results were even higher than those found by Iacono (1982): TMz = 0.92 (p < 0.00005); rnz = 0.57 (p = 0.01). However, inspection reveals that the high MZ twin correlation results largely from two “outlier” pairs concordant for poor tracking. Although the small sample size makes analysis problematic, the findings are consistent with previous assertions that genetic factors play an important role in the control of smooth pursuit eye movements.

33 Limitations. The interpretation of the results of this study should be tempered by five possible limitations. First, in all instances for the attentional battery and eyetracking assessment, and in most of the instances with the SIS, the evaluation of both members of a twin pair was carried out by the same individual. Since physical similarity is a good guide to zygosity, a bias introduced by the method of assessment could have inflated MZ relative to DZ correlations. It is implausible that such a bias would play a major role in the objective assessments of attention and eye tracking or in the SRQ. The SIS, by contrast, may be most susceptible to bias, although this would be minimized by its highly structured nature. However, the pattern of results is not consistent with what would be expected if rater bias were important. MZ twin correlations were lower on those factors based on the SIS, where bias should be highest, than they were on the self-report measures, where bias should be absent. A second substantial limitation of this study is the small sample size. This is particularly a problem with the twin analysis. At best, results from the analysis of twin correlations can suggest a broad pattern of results and will, it is hoped, stimulate the collection of a larger sample that would allow more definitive conclusions. Caution is also indicated in the interpretation of some of the cross-trait correlations; it is easy to overinterpret modest correlations with these sample sizes. A third potential limitation has to do with the interpretation of significance levels in a report such as this which contains many comparisons. No attempt has been made to correct thep levels for multiple tests because in nearly all instances tests were of specific a priori hypotheses, most of which had been, in one form or another, previously reported in the literature. While differences of opinion remain about how to deal with these sorts of multiple comparisons, we restrict our treatment to this comment of caution. Fourth, the sampling procedure, which accepted all cooperative twin pairs, resulted, for unknown reasons, in a sample of 13 MZ twin pairs, all of which were female-female. By contrast, the 16 DZ pairs were more evenly divided into nine male-female, three male-male, and four female-female pairs. The point-biserial correlation between sex and the six major indices of schizotypy ranged from i-O.24 to -0.15, none of which were statistically significant. In none of the six major indices of schizotypy did the difference in correlations in same-sex versus opposite-sex DZ twins approach statistical significance. While the sample is not ideal, it appears unlikely that the uneven distribution of sex across zygosity groups substantially biased the results. Fifth, in the analyses of the structure of schizotypy and between-trait correlations, results from both members of a twin pair were treated as independent observations. This approach will not bias any of the observed correlations or the factor analyses that are derived therefrom. However, depending on the degree to which the twins are correlated for the traits in question, it will result in an overestimation of the true degrees of freedom and hence an underestimation of the probability of a type 1 error. Given that more than 55% of our sample were DZ twins where trait correlations tended to be quite modest, the underestimation of type I error rates is, for most comparisons, likely to be small. However, this problem provides an additional note of caution regarding possible overinterpretation of the results. While intriguing, the results of this study warrant replication in a larger sample size.

34

Acknowledgments. This work was supported in part by Public Health Service Grant MH41953 and by the Department of Mental Health and Mental Retardation of the Commonwealth of Virginia. Leroy Thacker, M.S., assisted in the data analysis. Drs. Loren and Jean Chapman kindly provided information that enabled us to develop short forms for their schizotypy scales. Dr. G. Claridge provided helpful advice regarding his STA scale.

References American Psychiatric Association. DSM-III-R: Diagnostic and Statistical Manual of Mental Disorders. 3rd ed., rev. Washington, DC: APA, 1987. Asarnow, R.F.; Nuechterlein, K.H.; and Marder, S.R. Span of apprehension performance, neuropsychological functioning and indices of psychosis-proneness. Journal of Nervous and Mental Disease, 171:662-669, 1983. Baron, M.; Gruen, R.; Rainer, J.D.; Kane, J.; Asnis, L.; and Lord, A. A family study of schizophrenic and normal control probands: Implications for the spectrum concept of schizophrenia. American Journal of Psychiatry, 1421447-455, 1985. Chapman, L.J., and Chapman, J.P. The search for symptoms predictive of schizophrenia. Schizophrenia Bulletin, 13:497-503, 1987. Chapman, L.J.; Chapman, J.P.; and Miller E.N. Reliabilities and intercorrelations of eight measures of proneness to psychosis. Journal of Consulting and Clinical Psychology, 50: 187195, 1982. Chapman, L.J.; Chapman, J.P.; and Raulin, M.L. Scales for physical and social anhedonia. Journal of Abnormal Psychology, 85:374-382, 1976. Claridge, G.S., and Broks, P. Schizotypy and hemisphere function: I. Theoretical considerations and the measurement of schizotypy. Personality and Individual Differences, 5:633-648, 1984. Claridge, G.S., and Hewitt, J.K. A biometrical study of schizotypy in a normal population. Personality and Individual Differences, 8:303-312, 1987. Cornblatt, B.A., and Erlenmeyer-Kimling, L. Global attentional deviance as a marker of risk for schizophrenia: Specificity and predictive validity. Journal of Abnormal Psychology, 4:470-486, 1985. Cornblatt, B.A.; Risch, N.J.; Faris, G.; Friedman, D.; and Erlenmeyer-Kimling, L. The continuous performance test, identical pairs version: I. New findings about sustained attention in normal families. Psychiatry Research, 26:223-238, 1988. Coursey, R.D.; Lees, R.W.; and Siever, L.J. The relationship between smooth pursuit eye movement impairment and psychological measures of psychopathology. Psychological Medicine, 19:343-358, 1989. Eaves, L.J.; Eysenck, H.J.; and Martin, N.G. Genes, Culture and Personality: An Empirical Approach. New York: Academic Press, 1989. Frangos, E.; Atanassenas, G.; Tsitourides, S.; Katsanov, N.; and Alexandrakov, P. Prevalence of DSM-III schizophrenia among the first-degree relatives of schizophrenic probands. Acta Psychiatrica Scandinavica, 72:382-386, 1985. Gershon, E.S.; DeLisi, L.E.; Hamovit, J.; Nurnberger, J.I., Jr.; Maxwell, M.D.; Schreiber, J.; Dauphinais, D.; Dingman, C.W.; and Guroff, J.J. A controlled family study of chronic psychoses. Archives of General Psychiatry, 45:328-336, 1988. Hewitt, J.K., and Claridge, G. The factor structure of schizotypy in a normal population. Personality and Individual DifSerences, 10:323-329, 1989. Holzman, P.S.; Kringlen, E.; Levy, D.L.; Proctor, L.R.; Haberman, S.; and Yasillo, N.J. Abnormal pursuit eye movements in schizophrenia: Evidence for a genetic marker. Archives of General Psychiatry, 34:802-805, 1977. Holzman, P.S.; Kringlen, E.; Matthysse, S.; Flanagan, S.D.; Lipton, R.B.; Cramer, G.; Levin, S.; Lange, K.; and Levy, D.L. A single dominant gene can account for eye tracking dysfunction and schizophrenia in offspring of discordant twins, Archives of General Psychiatry, 45:641-647, 1988.

35

Holzman, P.S.; Levy, D.L.; and Proctor, L.R. Smooth pursuit eye movements, attention, and schizophrenia. Archives of General Psychiatry, 33:1415-1420, 1976. Holzman, P.S.; Proctor, L.R.; Levy, D.L.; Yasillo, N.J.; Meltzer, H.Y.; and Hurt, S.W. Eye tracking dysfunction in schizophrenia patients and their relatives. Archives of General Psychiatry, 31: 143-151, 1974. Holzman, P.S.; Solomon, C.M.; Levin, S.; and Waternaux, C.S. Pursuit eye movement dysfunctons in schizophrenia: Family evidence for specificity. Archives of General Psychiatry, 41:136-139, 1984. Iacono, W.G. Eye tracking in normal twins. Behavior Genetics, 12517-526, 1982. Iacono, W.G., and Lykken, D.T. Eye tracking and psychopathology: New procedures applied to a sample of normal monozygotic twins. Archives of General Psychiatry, 36:13611369, 1979. Jacobsberg, L.B.; Hymowitz, P.; Barasch, A.; and Frances, A.J. Symptoms of schizotypal personality disorder. American Journal of Psychiatry, 143:1222-1227, 1986. Kendler, K.S. Diagnostic approaches to schizotypal personality disorder: A historical perspective. Schizophrenia Bulletin, 11539-554, 1985. Kendler, K.S., and Gruenberg, A.M. An independent analysis of the Copenhagen sample of the Danish Adoption Study of Schizophrenia: VI. The pattern of psychiatric illness, as defined by DSM-III, in adoptees and relatives. Archives of General Psychiatry, 41:555-564, 1984. Kendler, K.S.; Gruenberg, A.M.; and Tsuang, M.T. The specificity of DSM-III schizotypal symptoms. Abstracts of the 136th Annual Meeting of the American Psychiatric Association. Washington, DC: APA, 1983. p. 140. Kendler, K.S.; Heath, A.C.; and Martin, N.G. A genetic epidemiologic study of self-report suspiciousness. Comprehensive Psychiatry, 28:187-196, 1987. Kendler, K.S., and Hewitt, J. The structure of self-report schizotypy in twins. In review. Kendler, K.S.; Lieberman, J.A.; and Walsh, D. The Structured Interview for Schizotypy (SIS): A preliminary report. Schizophrenia Bulletin, 15:559-571, 1989. Kinney, D.K.; Woods, B.T.; and Yurgelun-Todd, D. Neurologic abnormalities in schizophrenic patients and their families: II. Neurologic and psychiatric findings in relatives. Archives of General Psychiatry, 43:665-668, 1986. Launay, G., and Slade, P. The measurement of hallucinatory experience in male and female prisoners. Personality and Individual Differences, 2~221-234, 1981. Lenzenweger, M.F., and Loranger, A.W. Detection of familial schizophrenia using a psychometric measure of schizotypy. Archives of General Psychiatry, 46:902-907, 1989. Lowing, P.A.; Mirsky, A.F.; and Pereira, R. The inheritance of schizophrenia spectrum disorders: A reanalysis of the Danish Adoptee Study data. American Journal of Psychiatry, 140:1167-1171, 1983. Mather, J.A. Eye movements of teenage children of schizophrenics: A possible inherited marker of susceptibility to the disease. Journal of Psychiatric Research, 4:523-532, 1985. Meehl, P.E. Schizotaxia, schizotypy, schizophrenia. American Psychologist, 17:827-838, 1962. Mirsky, A.F. The Israeli high-risk study. In: Dunner, D.L.; Gershon, E.S.; and Barrett, J.E., eds. Relatives at Risk for Mental Disorder. New York: Raven Press, 1988. pp. 279-296. Mirsky, A.F. The neuropsychology of attention: Elements of a complex behavior. In: Perecmen, E., ed. Integrating Theory and Practice in Clinical Neuropsychology. Hillside, NJ: Lawrence Erlbaum, 1989. pp. 75-91. Mishlove, M., and Chapman, L.J. Social anhedonia in the prediction of psychosis proneness. Journal of Abnormal Psychology, 94:384-396, 1985. Nuechterlein, K.H. Signal detection in vigilance tasks and behavioral attributes among offspring of schizophrenic mothers and among hyperactive children. Journal of Abnormal Psychology, 92:4-83, 1983. Nuechterlein, K.H. Converging evidence for vigilance deficit as a vulnerability indicator for in Schizophrenia. New York: schizophrenic disorders. In: Alpert, M., ed. Controversies Guilford Press, 1985. pp. 175-198.

36 Nuechterlein, K.H., and Dawson, M.E. Information processing and attentional functioning in the developmental course of schizophrenic disorders. Schizophrenia Bulletin, 10:160-203, 1984. Rado, S. Dynamics and classification of disordered behavior. American Journal of Psychiatry, 110:406-416, 1953. Ross, D.E.; Ochs, A.L.; Hill, M.R.; Goldberg, S.C.; Pandurangi, A.K.; and Winfrey, C.J. Erratic eye tracking in schizophrenic patients as revealed by high resolution techniques. Biological Psychiatry, 24:675-688, 1988. SAS Institute, Inc. SAS User’s Guide: Statistics, Version 5 edition. Cary, NC: SAS Institute, Inc., 1985. Schulz, P.M.; Soloff, P.H.; Kelly, T.; Morgenstern, M.; Di France, R.; and Schulz, S.C. A family history study of borderline subtypes. Journalof Personality Disorder, 3:217-229, 1989. Shenton, M.E.; Solovay, M.R.; Holzman, P.S.; Coleman, M.; and Gale, H.J. Thought disorder in the relatives of psychotic patients. Archives of General Psychiatry, 46:897-901, 1989. Siegel, C.; Waldo, M.; Mizner, G.; Adler, L.E.; and Freedman, R. Deficits in sensory gating in schizophrenic patients and their relatives: Evidence obtained with auditory evoked responses. Archives of General Psychiatry, 41:607-612, 1984. Siever, L.J.; Coursey, R.D.; Alterman, I.S.; Buchsbaum, M.S.; and Murphy, D.L. Impaired smooth pursuit eye movement: Vulnerability marker for schizotypal personality disorder in a normal volunteer population. American Journal of Psychiatry, 14 I : 1560-l 566, 1984. Siever, L.J.; Coursey, R.D.; Alterman, 1,s.; Zahn, T.; Brody, L.; Bernad, P.; Buchsbaum, M.; Lake, C.R.; and Murphy, D.L. Clinical, psychophysiological, and neurological characteristics of volunteers with impaired smooth pursuit eye movements. Biological Psychiatry, 26:33-51, 1989. Siever, L.J., and Gunderson, J. The search for a schizotypal personality disorder: Historical origins and current status. Comprehensive Psychiatry, 24:199-212, 1983. Siever, L.J.; Haier, R.J.; Coursey, R.D.; Sostek, A.J.; Murphy, D.L.; Holzman, P.S.; and Buchsbaum, M.S. Smooth pursuit eye tracking impairment: Relation to other “markers” of schizophrenia and psychologic correlates. Archives of General Psychiatry, 39: 1001-1005, 1982. Siever, L.J.; Keefe, R.; Bernstein, D.P.; Coccaro, E.F.; Klar, H.M.; Zemishlany, Z.; Peterson, A.E.; Davidson, M.; Mahon, T.; Horvath, T.; and Mohs, R. Eye tracking impairment in clinically identified patients with schizotypal personality disorder. American Journal of Psychiatry, 1471740-745, 1990. Simons, R.F., and Katkin, W. Smooth pursuit eye movements in subjects reporting physical anhedonia and perceptual aberrations. Psychiatry Research, 14:275-289, 1985. Spitzer, R.L.; Williams, J.B.W.; and Gibbon, J. Structured Clinical Interviewfor DSM-IIIR - Patient Version (SCID-P, 4/l/87). New York: New York State Psychiatric Institute, 1987. Torgersen, S. Genetic and nosologic aspects of schizotypal and borderline personality disorders: A twin study. Archives of General Psychiatry, 411546-554, 1983. Widiger, T.A.; Frances, A.; and Trull, T.J. A psychometric analysis of the socialinterpersonal and cognitive-perceptual items for schizotypal personality disorder. Archives of General Psychiatry, 441741-745, 1987.