Schizophrenia Spectrum Disorders Research Paper

Issues in Diagnosis

Chronicity, Negative Symptoms, and Affective Symptoms

Considerable debate has centered on the importance of a chronic course for a diagnosis of schizophrenia. Almost everyone agrees that individuals who follow a classic chronic course and manifest other symptoms of schizophrenia are true schizophrenia patients. Disagreement centers on those with an episodic course—especially with a small number of episodes. Since 1980 in the United States, the Diagnostic and Statistical Manual of Mental Disorders Third Edition, Third Edition–Revised, and Fourth Edition (DSM-III, DSM-III-R, and DSM-IV, respectively) by theAmerican PsychiatricAssociation (APA; 1980, 1987, 1994) have emphasized chronicity by requiring a 6-month duration for the diagnosis of schizophrenia. In DSM-IV, schizophreniform disorder applies to patients with episodes lasting between 1 and 6 months, whereas brief psychotic disorder applies to episodes of less than 1 month, and it is considered unclear whether either diagnosis is related to schizophrenia (Siris & Lavin, 1995).

In recent decades, extensive research has focused on negative symptoms (the absence or insufficiency of normal behavior), such as poverty of thought or speech, flat or blunted affect, apathy or anhedonia, and avolition or social withdrawal (Andreasen, 2001; Sommers, 1985). These negative symptoms are contrasted with positive symptoms (the presence of abnormal functioning), which may include hallucinations, delusions, thought disorder, disorganized behavior, and inappropriate emotions (Andreasen, 2001). Some theorists view negative symptoms as the fundamental deficiency of schizophrenia; others see them as a defect or residual state resulting from and following an active psychotic state (i.e., positive symptoms). A third perspective conceptualizes positive and negative symptoms as more or less equally important semi-independent processes (Lewine, 1985). The 6-month duration requirement for a diagnosis of schizophrenia that began with DSM-III made negative symptoms much more salient because the active phase symptoms often do not last for 6 months—placing the burden on negative symptoms to demonstrate chronicity. Negative symptoms are conceptualized as prodromal (preceding the onset of the active phase of psychotic symptoms) or residual (continuing after the active phase of psychotic symptoms) and either can be used to satisfy the duration requirement.

The separation of schizophrenia from the affective disorders (depression and mania) constitutes a third major diagnostic issue. The presence of many patients with both schizophrenic and affective symptoms (a schizoaffective clinical picture) challenges any attempt to dichotomize the distinction between schizophrenia and affective disorders (e.g., Kendell, 1982; Meltzer, 1984; Siris & Lavin, 1995). Decisions on this issue can profoundly affect the diagnosis of schizophrenia, and the presence of large numbers of patients with schizoaffective symptoms constitutes a major problem in conceptualizing schizophrenia.

A Brief History of the Conceptualization of Schizophrenia

Many discussions of the concept of schizophrenia begin with Kraepelin’s fundamental contributions toward the end of the nineteenth century and continuing into the early twentieth century (Johnstone, 1999c). Kraepelin grouped together syndromes with disparate clinical pictures on the basis of similarities in course, outcome, and age of onset, applying the name dementia praecox to this new diagnostic category. He viewed the disorder as having a youthful onset, intellectual and volitional disturbances, and a chronic course with intellectual deterioration as the outcome. For Kraepelin, these characteristics were a direct manifestation of an underlying organic disturbance (i.e., psychological or psychosocial factors were not emphasized). As valuable as Kraepelin’s contribution was, three issues created tension regarding his position. First, the chronic, deteriorating course was a defining feature of dementia praecox (Andreasen & Carpenter, 1993), implying that treatment was impossible. Impossibility of treatment should not be a matter of definition of a disorder, but rather an empirical correlate (Chapman & Chapman, 1973). Second, a nonnegligible minority (about 13%) of his dementia praecox patients failed to run a chronic, deteriorating course (Johnstone, 1999a). Third, onset of the disorder was not limited to the youthful onset indicated by the term praecox (Andreasen, 2001).

In response to these problems, Bleuler (1911/1950) offered a very different conceptualization of the disorder, which he renamed schizophrenia. In place of course and outcome emphasized by Kraepelin, Bleuler emphasized signs and symptoms (Andreasen & Carpenter, 1993) and distinguished between fundamental symptoms (somewhat similar to negative symptoms), which he viewed as specific to schizophrenia and as permanent or chronic features of the disorder, and accessory symptoms (somewhat similar to positive symptoms), which may be completely absent during part or all of the course of the disorder or may be very prominent (Bleuler, 1911/1950). Like Kraepelin, Bleuler hypothesized a chronic underlying physical disease process that can progress on its own to produce the full schizophrenia syndrome, but this process was associated with fundamental symptoms. Influenced strongly by Freud, Bleuler proposed that stressful events and other psychological processes could substantially influence the course of the accessory symptoms—an early statement of the popular diathesis-stress model in which a genetic diathesis or vulnerability responds to psychosocial stress by producing symptoms of psychopathology (Rosenthal, 1970; Walker & Diforio, 1997).

Bleuler’s conceptualization was much broader than Kraepelin’s. First, the prototypical clinical picture included more than patients with a chronic, deteriorating course. Second, Bleuler described a continuum of severity, with milder cases blending into the normal range of personality variation. Third, clinical judgment as to the presence of a splitting in the basic functions of the mind replaced a more easily observed chronic deterioration as the basis for diagnosis, making it possible to perceive schizophrenic processes in a very large number of patients.

The emphasis on signs and symptoms introduced by Bleuler combined with the difficulty of making clinical judgments as to what signs and symptoms suffice for a diagnosis of schizophrenia inevitably led to attempts to more precisely identify symptoms specific to schizophrenia. In one particularly influential attempt to achieve this goal, Schneider developed a list of first rank symptoms he believed to be specific to schizophrenia (Andreasen & Carpenter, 1993). The resultant list was restricted to unusual and bizarre hallucinations (e.g., hearing voices speak one’s thoughts aloud, discuss one in the third person, or describe one’s actions) or delusions (e.g., believing that thoughts are inserted by an external force, that one’s thoughts are broadcast to the outside world and heard by others, or that one’s own actions are imposed by an outside force; Johnstone, 1999a). Schneider’s approach, however, has not proven to predict outcome (i.e., chronicity) in schizophrenia or to be specific to schizophrenia (Andreasen & Carpenter, 1993). Similarly, genetic studies have yielded little support for the validity of first rank symptoms when such symptoms are employed as the sole basis for the diagnosis of schizophrenia (Gottesman, McGuffin, & Farmer, 1987; McGuffin, Farmer, Gottesman, Murray, & Reveley, 1984). Nevertheless, first rank symptoms have been a prominent part of the literature on schizophrenia and sometimes are incorporated into the list of active phase symptoms in other diagnostic approaches.

Diagnostic Approaches

A number of well-known approaches to the diagnosis of schizophrenia have been developed. In the United States, the approach since DSM-III is especially noteworthy for requiring a 6-month duration before schizophrenia can be diagnosed, thereby opting for the view that true schizophrenia runs a relatively chronic course. DSM-III was strongly influenced by what are known as the Feighner criteria (Feighner et al., 1972) from the Washington University group and by the research diagnostic criteria (RDC; Spitzer, Endicott, & Robins, 1978) from the New York State Psychiatric Institute group. The Feighner criteria strongly emphasize a chronic course by requiring a 6-month duration, an absence of affective disorder diagnosis, and such predictors of chronicity as being single, showing poor premorbid adjustment, and having a family history of schizophrenia, while requiring only one positive or active symptom. In contrast, the RDC require two active-phase symptoms and only 2 weeks duration, aiming for a concept broader than that captured by chronic schizophrenia. At the same time, the RDC also exclude patients who meet criteria for an affective disorder. As is discussed later in this research paper, in first-admission samples this exclusion can be quite restrictive.

Whereas the aforementioned diagnostic systems were created by consensus of committees attempting to interpret the clinical and research literature, other approaches have selected signs and symptoms that predict existing clinical diagnosis as the criterion. Further, one also hopes that by identifying features common to many diagnosticians, the essence or core features of the diagnosis will be preserved while eliminating more idiosyncratic aspects (Carpenter, Strauss, & Bartko, 1973)—although, of course, this approach depends on the wisdom of current clinical diagnosis. One such system was the New Haven Index (Astrachan et al., 1972), developed during an era of a very broad concept of schizophrenia. The New Haven Index was able to predict clinicians’ diagnoses in New Haven hospitals by using only signs and symptoms without mention of affective symptoms or a minimal duration. Delusions, hallucinations, and thought disorder were emphasized by requiring at least one of them and making any two of them sufficient for a diagnosis.

A second system, sponsored by the World Health Organization, came from the International Pilot Study of Schizophrenia (IPSS; Carpenter et al., 1973) involving 1,202 patients from nine countries. Stepwise discriminant function analysis was used to identify 12 items that best discriminated between patients with hospital diagnoses of schizophrenia or not schizophrenia in half the sample and cross-validated on the other half. These items included delusions (four items reflecting different types), thought disorder, restricted or flat affect, poor insight, poor rapport during the interview, unreliable information during the interview, and the absence of three indications of affective disorders (elation, depressed facial expressions, waking early). One point was awarded for the presence of each of the nine positive indicators and for the absence of each of the three affective symptoms. A table indicating the agreement with hospital diagnoses of schizophrenia and not schizophrenia (in the two samples) as a function of the number of points required to make a diagnosis showed that requiring five or more points resulted in a relatively broad concept of schizophrenia—detecting 80–81% of hospital diagnoses of schizophrenia, while diagnosing as schizophrenic 13–22% of patients not diagnosed as having schizophrenia by the hospital. With the more restrictive criterion of six or more points, these numbers changed to 63–66% and 4–6%, respectively. The authors suggested choosing the criterion (e.g., five or six points) for a diagnosis of schizophrenia depending on the application—hence the term Flexible System, which has been applied to this approach.

If the different diagnostic systems all assessed the same clinical phenomenon but differed only in the threshold for making a diagnosis, they would show a hierarchical relationship: Broader systems, which diagnose more patients as having schizophrenia, would include all patients diagnosed as having schizophrenia by the narrower systems. Unfortunately, that is not the case, and even the narrow systems did not show high agreement (Gottesman et al., 1987). For example, Strauss and Gift (1977) applied seven diagnostic systems to 272 patients in their first hospitalization and found that 122 were diagnosed as schizophrenic by at least one system. The two broadest systems were the New Haven Index and the (IPSS) Flexible System-5 with 68 and 57 diagnoses, respectively. Even though they are very broad, these systems missed about half the 122 patients diagnosed as having schizophrenia by another system. The three narrowest systems were the RDC, Feighner criteria, and Flexible System-6 with 4, 9, and 34 diagnoses, respectively. To illustrate poor agreement even among narrower systems, the Feighner criteria and the RDC agreed on a schizophrenia diagnosis for only two patients, and the Flexible System-6 diagnosed as schizophrenic only 44% of those called schizophrenic by the Feighner criteria. For the broader systems, the Flexible System-5 diagnosed as schizophrenic only 65% of the New Haven Index cases. The authors noted that the exclusion criteria for affective symptoms were responsible for the extreme narrowness of the Feighner criteria and the RDC because affective symptoms are prominent in first-admission samples of hospitalized patients.

Although these diagnostic systems show greater agreement and a more hierarchical structure with more chronic hospitalized patients (Endicott et al., 1982), these data serve to illustrate significant diagnostic disagreement and the importance of chronicity and affective symptoms in making a diagnosis. Strauss and Gift (1977) make another important point about the implications of attempts to narrow the diagnosis of schizophrenia to a chronic, Kraepelinian view: Such attempts shift the problem rather than solving it, creating large numbers of undiagnosed (Feighner criteria) or schizoaffective patients (RDC). Logically, if such large groups of patients are not to be considered as having schizophrenia, then presumably alternative etiologic hypotheses need to be developed and evidence marshaled of discriminant validity visà-vis schizophrenia, which has not been done.

This underscoring of some uncertainties in diagnosis should not be taken to imply a chaotic situation. Quite to the contrary, even with our imperfect conceptualizations of schizophrenia, many reliable findings have been reported, especially in research on genetic influences. Indeed, the genetic data constitute some of the most compelling findings in all of psychopathology, and those findings in turn can inform diagnostic choices.

The Contribution of Genetics

Genetic Studies of Schizophrenia

The genetics of schizophrenia is one of the great stories in research on psychopathology. This research demonstrated unequivocallythatgeneticfactorsareimportanttotheetiologyof schizophrenia (e.g., Bassett, Chow, O’Neill, & Brzustowicz, 2001;Gottesman,1991;Gottesman&Moldin,1998;Kendler, 1999; Kendler & Diehl, 1995; Tsuang & Faraone, 1995). Having established that point, this research additionally provides a foundation for examining the validity of concepts of schizophrenia—thatis, which ways of diagnosing schizophrenia produce the strongest findings from a genetic perspective?

The story begins with family risk for schizophrenia: It has long been known that schizophrenia runs in families (Kendler & Diehl, 1995). In these studies, one starts with a series of index cases or probands with a diagnosis of schizophrenia and then ascertains the risk of schizophrenia among their relatives. Risk for schizophrenia increases with genetic similarity. For example, using pooled European studies from 1920 to 1978, Gottesman et al. (1987) calculated the following risks of (definite) schizophrenia: first-degree relatives 7.3–9.35%, second-degree relatives 2.65–2.94%, and third-degree relatives 1.56%. Similarly, for seven more modern studies with improved methodology, Kendler and Diehl (1993) calculated an estimated morbid risk of schizophrenia among first-degree relatives to be 4.8% for schizophrenia probands compared with 0.5% for control probands, confirming a roughly tenfold increased risk reported by earlier family studies. Of course, in family studies environmental similarity is confounded with genetic similarity, creating a need for twin and adoption studies to clarify interpretation of family risk data.

Twin studies exploit an experiment of nature that creates monozygotic (MZ) and same-sex dizygotic (DZ) twins with 100% (exactly) and 50% (on average) of their genes in common, respectively. Because many aspects of the environment are equally similar for MZ and DZ twins (see Kendler & Diehl, 1995, for a discussion of this assumption), comparison of risk for schizophrenia in the cotwins of schizophrenia probands offers a strong test of the genetic hypothesis. As with family studies, twin studies from the first half of the twentieth century had reported results supportive of the genetic hypothesis: MZ cotwins of a schizophrenia proband have a higher probability of schizophrenia than do DZ cotwins.

These results seemed strongly supportive of the genetic hypothesis, but critics questioned them on methodological grounds. They argued that knowledge of zygosity contaminated the (nonblindfolded) diagnoses of schizophrenia, allowing the investigator’s beliefs about heredity to bias the outcome (Gottesman, 1991). However, in a sophisticated twin study that laid to rest methodological criticisms, Gottesman and Shields (1966) reported 50% concordance in MZ twins compared with only 12% for DZ twins. Other modern twin studies yielded similar results. Gottesman (1991) further reported that a weighted average of the four recent (since 1963) European twin studies showed 48% MZ and 17% DZ twin concordance, a more recent British study showed 40.8% and 5.3% concordance (Cardno et al., 1999), and a summary of four European studies (including Cardno et al., 1999) and one Japanese twin study conducted in the 1990s reported heritabilities of the order of 80% (Cardno & Gottesman, 2000). Thus, twin studies strongly support the genetic hypothesis (Kendler & Diehl, 1995).

Adoption studies completed the final step, in which genetic and environmental similarity are unconfounded. Heston and Denney (1968) followed up the adopted-away but nowadult offspring of mothers hospitalized for schizophrenia in Oregon. At an average age of approximately 36 years, 5 of 47 experimental group (mother schizophrenic) adoptees had developed schizophrenia compared with none out of 50 matched control adoptees (mother not schizophrenic), strongly supporting the genetic explanation for the initial observation that schizophrenia runs in families.

A second, larger adoption study conducted in Denmark exploited excellent government records listing births (including adoption information), psychiatric diagnoses, and current addresses for everyone in the country. In the initial report on the Copenhagen sample, Kety and his colleagues (Kety, Rosenthal, Wender, & Schulsinger, 1968; Kety, Rosenthal, Wender, Schulsinger, & Jacobsen, 1978) selected adopted childrenwholaterdevelopedschizophrenia,alongwithacontrol group with no psychiatric history; using hospital records, they ascertained the rate of schizophrenic spectrum disorders (a broad concept, intended to miss no one with possible schizophrenia) among the roughly 150 biological and 80 adoptive relatives of these index cases. Consistent with the genetic hypothesis, there was an elevated rate of schizophrenia among the biological relatives (8.7%) of the experimental index cases compared with biological relatives of control index cases (1.9%) and with the adoptive relatives of both the experimental group (2.7%) and the control group (3.6%). These results were confirmed in a later report based on interviewbased diagnoses and data from all of Denmark (Kety et al., 1994), as well as a reanalysis of the Danish interview-based data employing DSM-III diagnoses (Kendler, Gruenberg, & Kinney, 1994) and an independent adoption study in Finland (Tienari, 1991).

To summarize, the finding that schizophrenia runs in families could possibly be due to either genetic or environmental influences, but replicated findings from twin and adoption studies support the genetic interpretation. This conclusion raises two important questions: (a) What genetic model applies to schizophrenia and (b) what concepts of schizophrenia are supported by these data?

Genetic Models

Single-Gene Model

A single-gene model is attractive because it would be possible to locate the gene, study its properties, and ultimately understand its effects on brain structure or function that lead to schizophrenia. It would also be possible to identify nonschizophrenic carriers of the gene and to study those individuals to determine what environmental variables trigger the development of schizophrenia. Unfortunately, single-gene models have such a poor match to existing data that they can be rejected—at least for a large majority of cases of schizophrenia (Andreasen, 2001; Bassett et al., 2001; Faraone, Green, Seidman, & Tsuang, 2001; Gottesman & Moldin, 1998; Kendler, 1999; Kendler & Diehl, 1993, 1995; McGue & Gottesman, 1989; Tsuang & Faraone, 1997). Furthermore, attempts to identify a single gene through linkage with genetic markers in family pedigrees has been disappointing (Gottesman & Moldin, 1998; Kendler, 1999; Tsuang & Faraone, 1997).

Multifactorial Polygenic Model

Gottesman and Shields (1967) were the first to apply the multifactorial polygenic (MFP) model to schizophrenia, in which a large number of genes—each often assumed to be of small effect—along with environmental influences contribute in an additive fashion to the overall liability for schizophrenia. The model assumes a threshold such that schizophrenia develops when total liability exceeds the threshold. In general, the genetic data are consistent with MFP models with a threshold (e.g., Andreasen, 2001; Faraone & Tsuang, 1985; Gottesman et al., 1987; Gottesman & Moldin, 1998; Kendler & Diehl, 1993; Tsuang & Faraone, 1997).

Parenthetically, there always has been some theoretical tension between the assumption of additivity in the MFP and the implication of an interaction effect in the diathesis-stress model—that is, stressors producepathology only in those with a diathesis (Gottesman, 1991). The additivity assumption in the MFP is appropriate because most of the extant data fit this more parsimonious model, but interaction effects are to be expected, in which genetic vulnerability increases one’s sensitivity to environmental influences that tend to induce schizophrenia (Gottesman & Moldin, 1998; Gottesman & Shields, 1972; Meehl, 1962). In recent years, evidence has accumulated of such greater effects of stressors among genetically vulnerable individuals (e.g., children or siblings of schizophrenia probands) for obstetric complications (Tsuang & Faraone, 1995), disturbed communication in adoptive parents (Gottesman & Moldin, 1998), and for inferred but unspecified stressors among second-generation African-Caribbean immigrants to the United Kingdom (Moldin & Gottesman, 1997). Additionally, there may well be interaction effects among genes, called epistasis (Bassett et al., 2001; Gottesman & Moldin, 1998).

Mixed Models

Although the MFP model might be considered the default hypothesis because of the good fit with genetic data, intermediate and mixed models cannot be excluded and, in fact, are likely. In the limited-loci-polygenic model (Faraone & Tsuang, 1985) a small number of loci (e.g., two loci with two alleles each) contribute to schizophrenia. In order to fit the empirical findings, these models must assume a modest genetic contribution to schizophrenia, leaving a large environmental contribution. A similar situation obtains with mixed genetic models (a single gene combined with polygenes), which can be made consistent with either (a) a rare (e.g., 10% of schizophrenic cases) single gene with high penetrance (e.g., .60) or (b) a much more common single gene that has low penetrance (Gottesman & McGue, 1991; Kendler & Diehl, 1993). These latter two models probably are more interesting than the limited-loci-polygenic models because they are consistent with relatively high heritability overall (single gene plus polygenes) and potentially offer an explanation for some of the heterogeneity in the schizophrenia phenotype. The common single gene (or major gene) in particular is of interest because it might be associated with a more chronic course (Iacono, 1998). Nevertheless, the MFP model is a major component of any model consistent with genetic data on schizophrenia—even if there are one or more major genes with larger effects. Additionally, a broad review of the behavior genetics literature suggests that the kind of behavior represented by the construct of schizophrenia is likely to involve polygenic influences (Plomin, 1990; Plomin, DeFries, McLearn, & Rutter, 1997).

The preceding reasoning with respect to the MFP model applies to the majority of schizophrenia cases and does not exclude rare cases of schizophrenia or schizophrenia-like psychoses attributable to another etiology. For example, Tsuang and Faraone (1995) concluded that in some rare cases gross chromosomal abnormalities cause schizophrenia, and Gottesman (1991) listed numerous drugs, somatic disorders, and genetic and chromosomal factors that produce phenocopies of schizophrenia (i.e., schizophrenia-like psychoses that do not share the etiology of typical schizophrenia).

A small deletion of genetic material on the long arm of Chromosome 22 (22q) has been associated with a form of schizophrenia that develops in roughly 25% of individuals with the deletion and accounts for up to 2% of individuals with schizophrenia (Bassett et al., 2001). Although the 22q deletion is inherited in an autosomal dominant pattern, due to reproductive disadvantage such transmission is greatly reduced. Consequently, over 90% of the cases are due to de novo mutations, which in turn are more likely with increasing paternal age. More broadly, Gottesman (1991), when comparing schizophrenia to some extent with mental retardation, speculated that a mixed model will prove most appropriate for schizophrenia. In this model, some rare cases are attributed to single genes and others to primarily environmental factors, but the vast majority are attributed to one of two variations of the MFP. Of this MFP group, up to 10% might reflect the effect of a specific major gene combined with polygenic and environmental influences (the mixed genetic model with a more common single gene described previously), whereas the remaining 90% reflect the standard MFP model of polygenic and environmental influences. Although a specific major gene operating in a polygenic context has not yet been identified, identification of such a gene would constitute a major breakthrough.

Specific and Nonspecific Liability

As noted previously, the MFP model assumes that polygenic and environmental influences combine to produce total liability with a threshold for the appearance of schizophrenia. The liability is further assumed to be distributed normally in the population as a whole and to consist of specific genetic, nonspecific genetic, and nonspecific environmental liability, as well as genetic and environmental assets that reduce liability (Gottesman, 1991). The modifier specific means that this aspect of the genetic contribution is specific to schizophrenia and not to any other disorders, whereas nonspecific factors affect liability but are not themselves specific to schizophrenia.

Relatively little has been said about the nature of nonspecific liability. As conceptualized within a diathesis-stress framework, environmental stress (physical or psychosocial) has nonspecific effects that influence but are not limited to the risk for schizophrenia. Similarly, it is reasonable to suggest that among the nonspecific genetic liability will be characteristics that influence the amount of stress the person experiences (Fowles, 1992b; also see Meehl’s classic 1962 paper for discussion of nonspecific factors involving vulnerability to stress). Two examples can serve to illustrate this point. First, all other things being equal, individuals with an anxiety-prone or stress-reactive temperament will experience more stress than would less anxious individuals, thereby contributing to nonspecific liability. Second, individuals whose genetic endowment pushes toward lower intelligence are more likely to experience academic, occupational, economic, and even interpersonal aversive experiences that make the world objectively more stressful. These suggestions are consistent with Heston and Denney’s (1968) report that among the adopted-away offspring of mothers with schizophrenia, there was a trend toward an increased rate of mental retardation and anxiety disorders. In selecting for mothers who clearly had schizophrenia, the authors may have selected for nonspecific genetic liability as well as specific genetic liability. On the other hand, inasmuch as presumably there are a large number of nonspecific effects, selection for any specific characteristic (e.g., low intelligence) would be weak and difficult to demonstrate. Furthermore, this nonspecific liability would segregate independently of specific liability, contributing to a varied collection of vulnerabilities and deficits among the relatives of schizophrenia probands— consistent with the impression of substantial rates of psychiatric abnormalities among relatives (Gottesman, 1991; Gottesman & Shields, 1976)—that add to the difficulty of identifying the true schizophrenic spectrum.

Environmental Liability and Episodic Course

An important feature of the MFP model needs to be underscored. With a fixed threshold and additive genetic and environmental liability, a trade-off exists between the genetic and environmental liability. As genetic liability increases, less environmental liability is needed to reach threshold (Fowles, 1992b, 1994; Gottesman, 1991; Gottesman & Shields, 1982; Siris & Lavin, 1995; Zubin & Spring, 1977). We do not know how many genes are involved or how high the maximum liability can be, but it is generally assumed that there can be enough genetic liability so that relatively little environmental stress is needed to reach threshold. Such individuals are likely to run a chronic course because their total liability will be above threshold most or all of the time. As the genetic contribution to liability diminishes, it must be offset with greater contributions from the environment. In those cases, the course is likely to be episodic because most of the time the individual will not encounter that much stress. Of course, stress can be more or less chronic, in which case even substantial contributions to liability from stress can be associated with chronicity (Wing, 1978). Finally, as the genetic liability diminishes even further, a point can be reached at which normal amounts of environmental stress will not be sufficient to bring total liability to threshold, making such fortunate individuals not schizophrenic for genetic reasons. In this conceptualization of the underlying etiology, episodic schizophrenia differs from chronic schizophrenia quantitatively but not qualitatively. It may be practical to distinguish between cases that—on average at least—are more genetic versus less genetic, but the etiologic model remains the same for both.

Summary

It can be seen from this brief review that the MFP model accounts well for genetic data, but one cannot exclude contributions from single rare genes with relatively high penetrance, purely environmental etiologies, and more common single genes with low penetrance. It is important to note that all attempts to fit genetic models to extant data conclude that environmental factors cannot be neglected. Ironically, in this sense the genetic data provide the strongest evidence for environmental contributions to schizophrenia. This evidence is neutral with respect to whether the environmental contribution involves psychosocial factors or physical environmental factors. Two major lines of research provide support for the hypothesis that psychosocial stress does contribute to the onset of schizophrenia—life events and aversive family interactions, both of which are reviewed later in this research paper— but evidence also strongly supports physical environmental contributions.

Environmental Influences

Variability in Outcome

A contribution of life events to the onset of schizophrenia became more likely after Bleuler broadened the concept of schizophrenia to include those who would show improvement. Additionally, with this outcome heterogeneity, it was inevitable that investigators would try to find predictors of good versus poor outcome. One approach distinguished between process and reactive schizophrenia (Chapman & Chapman, 1973; Neale & Oltmanns, 1980). Reactive schizophrenia has a rapid onset associated with a life event, a normal premorbid adjustment, and a good prognosis. Process schizophrenia has an insidious onset, poor premorbid social adjustment, affective flattening (a classic symptom of chronic schizophrenia), and a poor prognosis. Premorbid adjustment is the most powerful correlate of prognosis. In another tradition, the term schizophreniform disorder originally referred to psychoses with good prognoses in order to distinguish them from so-called genuine schizophrenia (Siris & Lavin, 1995)—that is, Kraepelinian chronic, deteriorating schizophrenia. The predictors of good outcome are now found in the good prognosis subtype of schizophreniform disorder in DSM-IV (American Psychiatric Association, 1994, 2000; see also Siris & Lavin, 1995) with confusion or perplexity during the episode and absence of blunted or flat affect being added to good premorbid adjustment and rapid onset as predictors of good outcome. The presence of affective symptoms also indicate a better prognosis (Siris & Lavin, 1995).

Life Events

Interpretation of Outcome Heterogeneity

Aplausible interpretation of the process-reactive or poor versus good prognosis distinction is that the process or poor prognosis cases have greater (possibly largely genetic) vulnerability that adversely affects the premorbid personality and develops into schizophrenia without the additional liability of a noteworthy life event. In contrast, the reactive or good prognosis cases are somewhat less vulnerable to schizophrenia and develop normally until a stressful life event adds to liability and raises them above threshold for schizophrenia. In many cases, the stress response to negative life events does not last forever, the stress-based liability diminishes, and the person recovers from schizophrenic symptoms. This interpretation of the nature of the process-reactive distinction implies that life events and the associated stress response can contribute to the onset of schizophrenia but are likely to do so most obviously for individuals with a sudden onset of schizophrenic symptoms.

Life Event Studies

In a classic study consistent with the aforementioned reasoning, Birley and Brown (1970) assessed life events during the 12 weeks prior to the onset of symptoms among patients whose symptom onset could be dated within a week. This datable-onset requirement necessarily precludes patients with an insidious onset and samples strongly in favor of reactive or good prognosis schizophrenia (approximately 50% of the schizophrenic patients were excluded by this criterion). Nonpatient controls from the community were asked about life events in the 12 weeks before the interview. The schizophrenic patients showed an increased rate of life events (60% of patients) in the 3 weeks prior to onset relative to earlier 3-week periods and relative to all 3-week periods for the controls (averaging about 21%).

For life events, there is a concern about the direction of effect—whether incipient psychotic symptoms cause the life event rather than the reverse. The authors rated the life events for their independence of the individual’s behavior (e.g., losing one’s job because a factory closed clearly would be independent). The results were similar to those previously mentioned for the restricted group of events classified as independent: 46% for the most recent 3 weeks for schizophrenic patients compared with 12–14% for other 3-week periods for patients and all 3-week periods for controls. Thus, this study found that life events were associated with onset of symptoms among the roughly 50% of patients with a datable onset and that the time frame for life event to symptoms was relatively short—about 3 weeks.

It was 17 years before a replication of this important finding was published. The World Health Organization reported data from eight cities around the world (Day et al., 1987), each site essentially constituting an attempt at independent replication. Because control subjects were not available, this study focused on the percentage of patients with life events in the 3 weeks before a datable onset relative to the three earlier 3-week periods. The predicted elevations were statistically significant at six of the eight sites for total life events and five of the eight sites for independent life events with identical trends at the remaining sites (not significant because of smaller numbers of patients).

Ventura, Nuechterlein, Lukoff, and Hardesty (1989) noted that the previous two studies involved retrospective memories of life events after the onset of schizophrenic symptoms (possibly permitting memory bias for patients who want to attribute the onset of schizophrenia to stress) and argued that it was important to replicate the results in a prospective study. They tracked 30 patients with schizophrenia following discharge from the hospital with assessment of schizophrenic symptoms biweekly (to detect relapse) and of life events monthly. Using data from the last life event interview before relapse-exacerbation, they found an increased number of life events (both total and independent) relative to other time periods for the 11-patient relapse group and relative to comparable months for the nonrelapse group. Although these patients were not selected for a datable onset, they had shown sufficient recovery that they could be monitored for relapse. At least in that sense they were selected not to include the most chronic, process cases of schizophrenia that tend to be less responsive to treatment. Thus, similar results were obtained in a prospective study, thereby eliminating the possibility of memory bias present in the Birley and Brown and the Day et al. studies.

Extreme Life Events

The aforementioned studies were conducted with patients suffering from typical schizophrenia. Dohrenwend and Egri (1981) took the argument a step further and applied it to battlefield psychoses that—at a symptom level—are indistinguishable from schizophrenia, although they are likely to remit quickly. Finding that the individuals involved had shown no other evidence of schizophrenia, they proposed that very severe stress can contribute to a schizophrenic reaction in those without any obvious vulnerability. Although most would argue that these battlefield psychoses are not cases of true schizophrenia, a parsimonious model would suggest that these individuals may have some genetic liability to schizophrenia but only at a level that requires extreme stress to bring them to threshold. As soon as that extreme stress passes, they should recover and (hopefully) never again experience psychotic symptoms. Inasmuch as the vast majority of battlefield casualties do not involve symptoms of schizophrenia (Gottesman, 1991), the assumption that some degree of genetic liability combines with the extreme stress to produce a schizophrenic syndrome offers a possible explanation for why only certain individuals are afflicted. This perspective perhaps parallels that for amphetamine psychosis—in which massive increases in dopamine produced by stimulant drugs produce a schizophrenia-like condition in normal individuals—that routinely is cited as evidence of the involvement of dopaminergic activity in schizophrenia.

Life Events Act on a Small Percent of the Population

The implications of this literature on life events—in the context of the MFP model—are that noteworthy stressful life events are not necessary for the small number of individuals with a heavy genetic loading for schizophrenia and are not sufficient for the large number (perhaps over 95% of the population) with a modest or no genetic loading (Gottesman, 1991). Between these extremes, a range of genetic liability for schizophrenia can be combined with stressful life events to bring the individual to the threshold for manifestation of schizophrenic symptoms. In a quantitative sense, the contribution of stress need not be large for the population as a whole because it centers on a small percentage of the population with enough genetic liability to be vulnerable to environmental stress (Fowles, 1992b, 1994; Gottesman & Shields, 1976; Siris & Lavin, 1995).

Aversive Family Interactions

Not all stressful environments involve obvious life events. Some can be more subtle, as in the case of aversive family interactions. In a study that initiated an important line of research on family interactions and schizophrenia, Brown, Birley, and Wing (1972) monitored 9-month relapse among 101 patients with schizophrenia discharged from the hospital to live with family. At the patient’s admission to the hospital, the authors rated an extensive interview with key relatives for number of critical comments and unusual emotional overinvolvement, these ratings being combined into a single index of family expressed emotion (EE). High EE was associated with more frequent relapse and interacted with both hours of contact with high-EE families and compliance with medication. This study was replicated by Vaughn and Leff (1976) and the pooled results summarized by Leff (1976) and Leff and Vaughn (1985). Based on these pooled data, 51% of patients returning to high-EE families relapsed versus 13% for those returning to low-EE families. Of those in high-EE families, 69% with more than 35 hours/week contact (in the same room) with their relatives relapsed compared with only 28% with less than 35 hours/week contact.Additionally, medication reduced relapse in high-EE families, with relapse ranging from 15% for those on medication and having fewer than 35 hours/week contact to 92% for the combination of high contact and no medication. Hours of contact and medication had no effect on relapse among patients with low-EE families.

The basic finding that aversive family interactions in the form of high EE predicts a higher rate of relapse has been replicated many times (Hooley & Hiller, 1998). Debate has centered on the direction of the effect: Does high EE cause relapse, or could both relapse and family reactions to the patient reflect the effects of a third variable, such as severity of the symptoms of schizophrenia? Although this question has not been fully resolved and possibly there are bidirectional effects, evidence suggests strongly that high EE among relatives does contribute in important ways to relapse of schizophrenic symptoms (Hooley & Hiller, 1998; Linszen et al., 1997; Nuechterlein, Snyder, & Mintz, 1992).

Institutional Environments

In the famous three-hospital study, Wing and Brown (1970) found that negative symptoms in schizophrenia (slowness, underactivity, blunting of affect, and poverty of speech), called the clinical poverty syndrome, varied across time and hospitals as a function of the institutional environment—a phenomenon termed institutionalism. Clinical poverty was high on wards that placed many restrictions on patients and provided few opportunities for activities; in contrast, it was minimized on wards that placed fewer restrictions and required more socially positive activities. Wing (1978) attributed clinical poverty to “a protective reaction against the painful effects of social interaction when one has inadequate equipment for communication” (p. 606)—that is, some negative symptoms can be seen as a form of passive or active avoidance. In agreement, Carpenter, Heinrichs, and Wagman (1985) noted that in some cases, social withdrawal may reflect a combination of diminished social drive and a reaction to the development of positive symptoms and to aversive environments. In a similar observation, Manfred Bleuler (1974) described the elimination of what he called catastrophic schizophrenia by improved hospital care prior to the advent of antipsychotic medications. Catastrophic schizophrenia, which was the prototype for Kraepelin’s dementia praecox, was characterized by an acute onset of a severe psychosis, followed with little improvement by a severe chronic psychosis lasting until death.

These phenomena reflect an interaction between a vulnerability among schizophrenia patients and the hospital environment. The fundamental point is that negative symptoms are not hard-wired manifestations of a genetic disease; rather, they are secondary to the schizophrenic process interacting with the environment. Thus, the environment contributes to some degree even to a classic Kraepelinian clinical picture.

Cortisol as an Index of Stress During Episodes

If stress contributes to liability in schizophrenia, at least a portion of patients should be experiencing stress during episodes. Cortisol in blood or saliva in humans serves as an index of activation of the hypothalamic-pituitary-adrenal (HPA) axis, in turn a primary manifestation of the stress response in humans. Additionally, failure to suppress cortisol secretion in response to administration of the synthetic glucocorticoid dexamethasone reflects poor feedback HPAregulation. In studies reviewed by Walker and Diforio (1997), increased baseline cortisol levels were associated with positive psychotic symptoms among patients with schizophrenia in cross-sectional and longitudinal studies and were high immediately prior to psychotic episodes, consistent with precipitation of symptom exacerbation by increased cortisol levels. Failure to suppress cortisol in the dexamethasone suppression test (DST) was associated more strongly with negative than with positive symptoms, although the difference might have been due to greater reduction of positive symptoms by medication. Similarly, DST failure was associated with several indicators of poor prognosis—poor premorbid adjustment, enlarged ventricals, worsening when medication is suspended, and poorer outcome over time. These findings strengthen the inference that psychosocial stress contributes to nonspecific liability in schizophrenia.

Physical Environmental Influences

Physical Environment as a Primary Cause of Schizophrenia

The possibility that schizophrenia is attributable to nongenetic biological factors has been attractive to many investigators. One favorite has been the hypothesis that an unspecified virus causes brain damage that presents—perhaps years later—as schizophrenia. However, one famous British advocate of that position later came to the conclusion that there was no evidence to support it (Crow in Liddle, Carpenter, & Crow, 1993). Invoking another important candidate, early trauma such as birth injury was proposed by Murray, Lewis, and Reveley (1985) to produce nongenetic forms of schizophrenia. They suggested that such trauma might account for the afflicted twin in discordant MZ twin pairs.

Gottesman and Bertelsen (1989) rebutted this argument by examining the risk of schizophrenia spectrum disorders among the offspring of both twins in discordant MZ twin pairs. They found an elevated risk of spectrum disorders that was equally high for the offspring of the healthy and afflicted cotwins and comparable to that for offspring of schizophrenia probands in general, exactly what would be predicted by a genetic hypothesis because of the identical genotype for the healthy cotwin. According to the Murray et al. hypothesis, there is no reason for increased risk for the offspring of the healthy twin (or the afflicted twin, for that matter; see Gottesman et al., 1987). The Gottesman and Bertelsen results demonstrate that the genetic hypothesis applies to a large majority of discordant MZ twins (and, by implication, to a large majority of all other cases of schizophrenia). Such evidence does not preclude the possibility that there are some cases of schizophrenia attributable to viral infection, obstetrical complications, or head injuries, but the number of such cases must be so few that they did not affect the results in the Gottesman and Bertelsen study, and there is little empirical support for environmental factors as a primary etiology of schizophrenia (Bassett et al., 2001).

Physical Environment as Nonspecific Liability

In the context of the expectation of environmental contributions to schizophrenia in the MFP model, physical insults to the brain may constitute part of the nonspecific contribution of the environment, acting as a stressor to increase risk of schizophrenia in individuals with a genetic vulnerability (Walker & Diforio, 1997). As discussed previously in connection with the process-reactive distinction, patients whose onset of schizophrenia is associated with a life event should—on average—have less genetic liability than do patients without a life event. On that assumption, the relatives of persons with reactive schizophrenia should be at lower risk for schizophrenia than are the relatives of persons with a process schizophrenia. Gottesman (1991) cited an older study in Germany that included head injuries as somatic life events associated with the onset of schizophrenia. As expected from this reasoning, the siblings of patients with somatic stressors had a lower risk of schizophrenia (4.8%) than did siblings of patients with no stressors of any kind (10%). Along with other, more anecdotal (but nevertheless informative) evidence, these findings led Gottesman (1991) to include head injuries among the list of stressors that increase the risk of episodes among genetically vulnerable individuals.

Although the evidence is mixed, pregnancy and birth complications appear to be associated with an increased risk of schizophrenia (Johnstone, 1999b)—especially in the second trimester (Bunney & Bunney, 1999). Consistent with this finding, Gottesman (1991) cited a review of twin studies by McNeil and Kaij suggesting that obstetrical complications (pregnancy or birth complications and problems within 4 weeks of birth) may be stressors that combine with genetic vulnerability to produce schizophrenia. Similarly, Andreasen (2001) summarized findings that viral infections during pregnancy and other birth complications contribute to the development of schizophrenia, but they apparently do so in combination with genetic factors. Tsuang and Faraone (1995) found support for both obstetric complications and viral infections and proposed that their effects could best be conceptualized in the context of the MFP model.

The Genain Quadruplets

The contribution of other than genetic factors to heterogeneity is underscored by a famous study of the Genain (identical) quadruplets (Mirsky et al., 2000; Rosenthal, 1963), all of whom developed schizophrenia. Although they were genetically identical, these four sisters showed very different clinical pictures, age at onset, course, and outcome. This varied manifestation of schizophrenia in spite of genetic identity underscores the importance of environmental factors—both physical (e.g., brain injury at birth) and psychosocial (e.g., differential treatment by parents)—as contributors to the heterogeneity of schizophrenia (Mirsky et al., 2000).

Summary

To summarize, all attempts to fit genetic models to the data on family, twin, and adoption studies conclude that the environment makes some contribution to the etiology of schizophrenia; this conclusion is consistent with the diathesis-stress hypothesis. More direct evidence supports the inclusion of psychological stressors such as life events, aversive interpersonal interactions, and battlefield conditions as contributors to the onset of episodes of schizophrenia. Similarly, institutional environments influence negative symptoms in schizophrenia. Additionally, physical insults such as head injuries, obstetrical complications, and possibly viral infections contribute to the development of schizophrenia in genetically vulnerable individuals, although it may be that in relatively rare cases, physical insults alone produce a schizophrenialike clinical condition. On average, the greater the contribution of psychological or physical stress to overall liability, the less the genetic contribution needs to be—with subsequent reduced risk of schizophrenia among the relatives.

The Concept of Schizophrenia from a Genetic Perspective

In addition to pointing toward environmental contributions, the genetics literature provides an important perspective in a number of ways on the conceptualization of phenotypic heterogeneity. First, variation in genetic liability in the MFP model has implications for conceptualizing phenotypic heterogeneity. Second, it is possible to compare the validity of the diagnosis of schizophrenia for different diagnostic approaches from a genetic perspective. Assuming that invalid diagnoses will add error variance and undermine the magnitude of findings, diagnostic approaches that yield weaker results are less valid. Third, the genetic perspective can be used to evaluate different subtypes for inclusion in the schizophrenia spectrum.

Severity of Genetic Loading

As already implied, one form of genetic heterogeneity with phenotypic consequences is to be expected from the MFP model: a dimension of severity of genetic liability (Gottesman, 1991; Gottesman et al., 1987) with consequences for family risk and course. This assumption has been documented most clearly with family risk data, in which family members of more severe phenotypes (e.g., nuclear, Kraepelin’s hebephrenic and catatonic, process, chronic, negative symptom, Crow’s Type II) have a greater risk of schizophrenia than do relatives of those with other, less severe phenotypes (e.g., nonnuclear, Kraepelin’s paranoid and simple, reactive, nonchronic, positive symptom, Crow’s Type I). For example, children of those with hebephrenic and catatonic subtypes—sometimes called nuclear schizophrenia because of their chronic, severe course—have about a 20% risk of schizophrenia compared with only 10% for the children of the milder paranoid and simple schizophrenic subtypes (Gottesman, 1991). Additionally, differences in the severity of genetic liability will on average be related to differences in course—high genetic liability is associated with a more chronic course and less high genetic liability with a more episodic course combined with life events as precipitants of episodes (Gottesman, 1991; Siris & Lavin, 1995). In examining this question, Gottesman et al. (1987) found no results that could not be attributed to a continuum of severity within the concept of schizophrenia.

If these variations in family risk, course, and outcome are reflections of severity of liability in the MFP model and are not attributable to qualitatively different genetic etiologies, then subtypes should not breed true, so to speak, in family studies—that is, all subtypes or forms of schizophrenia should be found among relatives, regardless of the subtype of the index case. The results have been clear: Although there is some tendency for subtypes to breed true, all types of schizophrenia are found in family pedigrees, contradicting the hypothesis that schizophrenia can be subdivided on the basis of qualitatively different genetic effects (Gottesman, 1991; Gottesman et al., 1987; Gottesman & Shields, 1982). The polygenic model, then, can account for some of the heterogeneity in schizophrenia without recourse to hypothesizing qualitatively different underlying genetic etiologies. Of course, etiologic heterogeneity remains a possibility—as indicated previously—in the form of major genes in a polygenic context, rare single genes with high penetrance, and purely environmental effects (see also Kendler & Diehl, 1993).

Breadth of the Concept of Schizophrenia

Twin Studies

Twin studies have been used to advantage to examine the optimal breadth of the concept of schizophrenia. Gottesman (1991) had eight experts diagnose 120 case histories compiled in the course of the Gottesman and Shields (1966) Maudsley twin study. The breadth of the concept of schizophrenia varied across diagnosticians, ranging from a narrow approach that identified only17 cases out of the 120 to a broad concept that diagnosed 79 cases. Each expert’s diagnoses were evaluated on the basis of the largest difference in MZ versus DZ concordance. A middle-of-the-road breadth of concept was found to be superior both to very broad and very narrow approaches.

Gottesman et al. (1987; Gottesman, 1991) used the Gottesman and Shields twin sample to evaluate DSM-III, RDC, and Feighner criteria. DSM-III yielded good results: 47.5% MZ, 9.5% DZ concordance. RDC and Feighner criteria yielded comparable results, but only if these stringent diagnostic systems were broadened to include probable schizophrenia. Use of Schneider first rank symptoms alone (presence of any one symptom justifies a diagnosis of schizophrenia) yielded very poor results: Only a small number of diagnoses of schizophrenia and the anomalous finding that DZ concordance was higher than MZ concordance. Use of Crow’s Type II criteria (discussed later in this research paper) yielded only three cases, necessitating substitution of a mixed type (Type I plus Type II) for comparison with a positivesymptom-only type (cf. Fenton & McGlashan, 1992, for similar findings). It is important to underscore that this test of the diagnostic approaches was based on longitudinal research in which a great deal of information about the course of the disorder over a period of time had been collected. Diagnostic approaches requiring a chronic course do much better with such comprehensive data than they do when applied to crosssectional studies involving a single assessment at one point in time. Gottesman et al. (1987), for example, cite a study in which the Feighner criteria failed to diagnose 32% of cases based on cross-sectional data when compared with later diagnoses based on longitudinal data.

IntheoriginalreportusingtheMaudsleytwinstudysample with DSM-III diagnoses, Farmer, McGuffin, and Gottesman (1987) used the ratio of MZ concordance to DZ concordance to evaluate whether broadening the spectrum to include one additional diagnosis along with schizophrenia improved results.The ratio for schizophrenia alone was 5.01 (using data from the preceding paragraph). Improvements were found with adding schizotypal personality disorder (6.01), adding affective disorder with mood-incongruent delusions (6.31), and atypical psychosis (5.23), whereas there was no effect for adding schizophreniform disorder (5.00).

Adoption and Family Studies

Adoption and family studies similarly can be used to provide clues as to which variations on the schizophrenia theme should be included in the schizophrenia spectrum. In the Danish adoption study, Kety et al. (1968) reported results based on hospital record diagnoses from the Copenhagen sample. Over time, these investigators enlarged the sample to include the rest of Denmark and conducted diagnostic interviews with all participants. Kendler et al. (1994) reported results based on DSM-III diagnoses for this national sample, which serves as an excellent illustration because of the relatively large sample size (for an adoption study) and high quality of the data. As in the original study, a broader concept of schizophrenia spectrum disorders was used, which included schizophrenia; schizoaffective disorder, mainly schizophrenia (SADMS); schizotypal personality disorder (SPD); and paranoid personality disorder (PPD). The risk of these spectrum disorders was elevated among the relatives of schizophrenic (13.0%), SAD-MS (12.8%), and SPD (19.1%) index case groups compared with a risk of only 3.0% among relatives of the control adoptees. The specific diagnoses were tallied among the relatives of the index cases and controls, and comparisons were made to see whether the particular diagnosis was elevated. When specific diagnoses were tallied, the diagnosis of SPD was elevated among the relatives of all three index case groups—schizophrenia (7.3%), SAD-MS (7.7%), and SPD (14.9%)—compared with relatives of controls (2.3%). The diagnosis of schizophrenia was elevated among the relatives of schizophrenic (3.3%) and SAD-MS (5.1%) but not SPD (0%) index groups compared with controls (0.3%). SAD-MS and PPD were not significantly elevated among the relatives of any index groups. Note that SPD was elevated among relatives of schizophrenia probands, but schizophrenia was not found among the relatives of SPD probands. In an opposite pattern, schizophrenia was found among relatives of SAD-MS probands, but SAD-MS was not found among the relatives of schizophrenia probands.

These analyses strongly support the inclusion of SPD in the spectrum, based on the elevated risk of SPD among relatives of schizophrenia probands. This conclusion consistently has been supported in reviews of the family, twin, and adoption literature (e.g., Asarnow et al., 2001; Battaglia & Torgersen, 1996; Kendler & Diehl, 1995). In a polygenic model in which SPD is a milder form, the risk of schizophrenia among relatives of SPD probands would be expected to be relatively low and difficult to detect with small samples (Battaglia & Torgersen, 1996), such as was the case in the Kendler et al. (1994) study when the data were reported separately for each spectrum diagnosis. Nevertheless, enough studies have found elevated risk of schizophrenia among relatives of SPD probands that Battaglia and Torgersen (1996) conclude the evidence is convincing.

Schizoaffective disorder clearly falls within the schizophrenic spectrum, as long as it is restricted to mainly schizophrenic subtypes. The unidirectional pattern of results for SAD-MS in the Kendler et al. (1994) study is consistent with the notion (discussed later in this research paper) that SAD-MS involves a genetic contribution from schizophrenia with perhaps a synergistic but independent contribution of affective symptoms—such that the combination will emerge from the population as a whole when both happen to be present in moderate degree. However, when starting with schizophrenia in the proband, because of the independence of the two components, affective symptoms would not necessarily combine with schizophrenia symptoms among the relatives in high numbers. Consistent with this model, Kendler and Diehl (1995) report that in two large-scale studies, although there was no increase in risk of mood disorders among the relatives of schizophrenia probands (as expected), the risk of psychotic features was more than twice as likely among the affectively ill relatives of schizophrenia probands compared with relatives of controls. This finding supports the notion that genetic liability to schizophrenia combines with liability to affective disorder to produce psychotic reactions.

The results for PPD were equivocal in the Danish study. Similarly, some reviewers include it (e.g., Kendler & Diehl, 1995), whereas others find the evidence equivocal (Asarnow et al., 2001). PPD appears to be less strongly related to schizophrenia than is SPD but may well be included in the spectrum.

Schizophreniform Disorder

Overall, the evidence for inclusion of schizophreniform disorder has been somewhat inconsistent, possibly as a function of varying conceptualizations of this disorder. As noted earlier, the diagnosis originally referred to psychotic patients with good prognosis in order to distinguish them from Kraepelinian chronic schizophrenia (Siris & Lavin, 1995). Since DSM-III (in the United States), the term has been narrowed to refer to a condition identical to schizophrenia except that it lasts less than 6 months and does not require deterioration in social and occupational functioning. In DSM-IV, a minimum of 1-month duration is required to distinguish schizophreniform disorder from brief psychotic disorder (APA, 1994, 2000). The latter diagnosis to some extent reflects a similar traditional Scandinavian concept of brief reactive psychosis that emphasized a severe stressor shortly before the onset of the psychosis, the understandability of the presenting symptoms as a result of the stressor, and full recovery to normal function, although it should be noted that the precise definition has varied over time and among authors (Siris & Lavin, 1995). Also, DSM-IV does not require a stressful event. Thus, the most salient aspect of these distinctions has to do with the duration of one or more episodes, combined with at least enough recovery to terminate an episode and an absence of chronic negative symptoms that would meet the 6-month duration requirement for schizophrenia in DSM-IV.

An episodic course per se is not an issue for inclusion in the concept of schizophrenia: Even with narrow diagnostic approaches, an episodic course is not precluded as long as the 6month criterion is met at some point. Similarly, an initial presentation formally meeting the criteria for schizophreniform disorder is not incompatible with a later diagnosis of schizophrenia. In many studies, the diagnosis is changed to schizophrenia at follow-up, leaving a smaller (albeit still significant) number who show a remitting nonaffective psychosis (Siris & Lavin, 1995). Thus, it is only schizophreniform disorder based on a longitudinal perspective for which there is a question of relatedness to schizophrenia.

In the Danish adoption study (Kendler et al., 1994), schizophreniform disorder was included in a psychotic nonspectrum group (i.e., was not seen as part of the schizophrenia spectrum), but other evidence has been more supportive. In a DSM-III (lifetime diagnosis) reanalysis of the Iowa Family Study, Kendler, Gruenberg, and Tsuang (1986) reported as many schizophrenia cases among the relatives of schizophreniform probands (3.6%) as among the relatives of schizophrenia probands (3.7%), finding that the two diagnoses were indistinguishable from a familial perspective. Kendler and Diehl (1995) cited two large family studies that reported an increased risk of remitting or atypical psychoses (nonaffective psychoses that do not meet criteria for schizophrenia and thus would include schizophreniform disorder) among the relatives of probands with schizophrenia and suggested that family liability to schizophrenia increases the risk for several nonschizophrenic psychotic disorders.

From the perspective of a polygenic model, the shorter the episodes and the more complete the recovery, the lower is the presumed genetic liability—therefore, the more difficult it is to demonstrate genetic relatedness in family and adoption studies. Also, lower genetic liability (schizophreniform disorder) should be more easily seen among the relatives of probands with higher genetic liability (schizophrenia) than it is in the reverse direction. The evidence seems reasonably supportive of this expectation. As just noted, evidence from family studies for inclusion of schizophreniform disorder in the schizophrenia spectrum is reasonably strong. Furthermore, based on his reading of the genetics literature, Gottesman (1991) included many cases of schizophreniform disorder as part of schizophrenia: Although he accepted that there may be some psychogenic psychoses, he argued that schizophrenia can involve only one or two episodes, the duration criterion is arbitrary and can lead to underdiagnosis, and low-risk genotypes could develop mild and remitting schizophrenia, reflecting the unity of schizophrenia with manifestations along a broad continuum.

Brief Psychotic Disorder

Little or no evidence supports the inclusion of brief psychotic disorder or brief reactive psychosis as part of the schizophrenia spectrum (Siris & Lavin, 1995). Several family studies (unfortunately often methodologically flawed) found that brief reactive psychosis tends to run in families and also may be related to mood disorder but is not associated with schizophrenia. Of even greater interest is that a Danish study of matings between persons with schizophrenia and reactive psychoses did not increase the risk of schizophrenia in offspring over that expected for children with one parent with schizophrenia and one with no psychopathology (Gottesman, 1991; Gottesman et al., 1987). Although the small sample size precludes firm conclusions, the result suggests no genetic or environmental contribution to liability for schizophrenia from brief reactive psychosis. Although it would be parsimonious to speculate that a very modest specific genetic liability for schizophrenia is required for a brief psychotic reaction and that this small genetic effect is undetectable without very large samples, it is also possible that processes having little or nothing to do with schizophrenia cause a portion of brief reactive psychoses.

Blurred Boundaries in the Polygenic Model

It should be appreciated that—as emphasized previously—in a polygenic model the phenotypic manifestation of schizophrenia may be influenced by contributions from different types of nonspecific liability in different subgroups of patients. Due to the varying importance of these contributions, it will be difficult to demonstrate them with the sample sizes normally available. Stating this point another way, in such a model the boundaries of the spectrum are not distinct but rather fall off gradually with quantitative variations in specific and nonspecific genetic liability. Consequently, the boundaries of the schizophrenic spectrum may be inherently blurred; only the strongest and most common manifestations of schizophrenia may be reliably demonstrable.

Summary

From the preceding examples, it can be seen that the genetics literature provides a valuable tool for evaluating the breadth of the concept of schizophrenia and the validity of specific diagnoses for inclusion in the schizophrenic spectrum (diagnoses with similarities to schizophrenia that are genetically related to schizophrenia). Based on the data reviewed so far, there has been no support for a subdivision of schizophrenia into two or more distinct genetic disorders, in spite of the heterogeneity of clinical features. A middle-of-the-road diagnosis is supported, comparable to DSM-III and to RDC and Feighner criteria diagnoses that were broadened to include probably schizophrenia—at least when these diagnoses are applied to longitudinal data. Additionally, the concept of schizophrenia needs to be broadened to include schizotypal personality disorder and—in some sense—schizoaffective disorder, mainly schizophrenic. Some phenotypic heterogeneity can be attributed to differences in the degree of genetic vulnerability with a higher genetic loading implicated for chronic, severe schizophrenia and the hebephrenic and catatonic subtypes. On the other hand, the life event, expressed emotion, and physical environmental studies and the example of the Genain quadruplets indicate that environmental factors affect the clinical picture.

Heterogeneity: A Challenge to the Conceptualization of Schizophrenia

Although differences in chronicity can be explained in terms of a dimension of severity of genetic loading, other aspects of the heterogeneity of schizophrenia continue to challenge the conceptualization of schizophrenia. In particular, the positive versus negative symptom distinction and the high frequency of a schizoaffective clinical picture represent major phenomena in need of explanation—that is, how can any unitary concept of schizophrenia account for these varied clinical pictures, or (alternatively) what other sources of variance might account for them? One neglected possibility is that nonspecific liability may account for some heterogeneity.

Positive Versus Negative Symptoms

The Two-Dimensional Model

As noted previously, debate has centered on whether positive or negative symptoms constitute the core symptoms of schizophrenia. Rather than trying to subordinate one to the other, Strauss, Carpenter, and Bartko (1974) viewed positive and negative symptoms as representing semi-independent processes. In a well-known proposal consistent with semiindependence, Crow (1980, 1985) distinguished between Type I and Type II schizophrenia, which he held are two independent dimensions or pathological processes that underlie schizophrenic symptomatology. In his formulation, Type I schizophrenia is characterized by positive symptoms (e.g., delusions, hallucinations), acute onset, an episodic course, good premorbid adjustment, and good response to antipsychotic medication, whereas Type II schizophrenia is characterized by negative symptoms, insidious onset, intellectual deterioration, poor premorbid functioning, a chronic course, and a poorer response to antipsychotic medication. Crow further attributed Type I schizophrenia toa neurochemical disturbance involving the neurotransmitter dopamine and Type II schizophrenia to structural brain changes (e.g., enlarged cerebral ventricles as assessed by computed tomographic studies—discussed later in this research paper). Given the independence of these processes, patients may present with only Type I or only Type II symptoms or a combination of the two. However, patients with exclusively negative symptoms are rare (Andreasen, Flaum, Swayze, Tyrell, & Arndt, 1990; Gottesman et al., 1987).

Although Crow’s hypothesis of an association between ventricular enlargement and negative symptoms has not been supported (Andreasen et al., 1990; Gottesman & Bertelsen, 1989; Liddle et al., 1993), involvement of dopaminergic activity in schizophrenia is indicated by three lines of evidence. First, the potency of typical antipsychotic medications correlates strongly with their ability to block dopamine receptors, especially D2 receptors (e.g., Byne, Kemether, Jones, Haroutunian, & Davis, 1999; Johnstone, 1999a). Second, dopamine agonists (e.g., amphetamine) in large doses can produce a clinical syndrome indistinguishable from paranoid schizophrenia in nonschizophrenic individuals (Krystal, Abi-Dargham, Laruelle, & Moghaddam, 1999). Third, the same drugs in small doses exacerbate symptoms in schizophrenic patients (Andreasen, 1985; Krystal et al., 1999) or cause transient symptoms in schizophrenic patients in remission (Losonczy, Davidson, & Davis, 1987). Additionally, as proposed by Crow, a good response to typical antipsychotic medication is associated with positive symptoms and a reactive schizophrenic pattern (e.g., Andreasen, 1985; Andreasen et al., 1990; Losonczy et al., 1987; Reynolds, 1989). On the other hand, after many years of research on this point it appears that although the dopamine hypothesis of the mechanism of traditional antipsychotic drugs is strongly supported, there is no positive support for a primary excess of dopamine (or dopamine receptors) in schizophrenia (Byne et al., 1999; Crow in Liddle et al., 1993; Weinberger & Lipska, 1995). Consistent with seeing dopamine as not a primary cause of schizophrenia, the newer, atypical antipsychotics (clozapine and risperidone) have pointed to the importance of inhibiting serotonergic activity, especially for treating negative symptoms (Kapur & Remington, 1996). Undoubtedly, dopamine plays an important role in schizophrenia, but it probably combines with other factors (Andreasen, 2001), possibly interacting with some other primary deficit in schizophrenia, to facilitate development of positive symptoms.

Like Crow, Carpenter (1992) addressed the theoretical implications of the divergence of positive and negative symptoms. He considered and rejected two traditional explanations for heterogeneity. One would reduce heterogeneity by dividing schizophrenia into several disease entities. The second would consider schizophrenia as a single disease entity and attribute the heterogeneity to the interaction between a primary pathology and other characteristics of the individual and the environment (cf. the MFP model discussed previously). Carpenter proposed instead that several distinct pathophysiological processes (with different neural circuits, pathophysiology, and etiology) combine differently in different individuals to produce the syndrome of schizophrenia. Rather than studying schizophrenia (comparing schizophrenic individuals with nonschizophrenic individuals) one should study the distinct pathophysiological processes (compare negative-symptom schizophrenic individuals with nonnegative-symptom schizophrenic individuals).

The Three-Dimensional Model

With time, the positive and negative symptom two-dimensional model has been replaced by a three-dimensional model of schizophrenic symptoms in which the positive psychotic symptoms have split into two dimensions (Kirkpatrick, Buchanan, Ross, & Carpenter, 2001). Using factor analysis, Liddle (1987) found dimensions of psychomotor poverty (negative symptoms of poverty of speech, blunted affect, and decreased movement), reality distortion (positive symptoms of various delusions and hallucinations), and disorganization (formal thought disorder, inappropriate affect, and poverty of content of speech). Replication of these dimensions in later research and widespread acceptance of the findings (Andreasen, Arndt, Alliger, Miller, & Flaum, 1995; Cuesta, Peralta, & Caro, 1999; Johnstone, 1999a; Kirkpatrick et al., 2001; Liddle et al., 1993) make it clear that multiple dimensions of symptomatology must be incorporated into any fully adequate theory of schizophrenia; however, the implications of the addition of a third major symptom dimension have not been fully developed, and this research paper focuses on the twodimensional model.

Genetic Versus Pharmacological Approaches

The conceptual challenge of the independence of important symptom dimensions can be seen from another perspective. As noted earlier, more chronic cases—including negativesymptom cases (Gottesman et al., 1987)—seem to carry a greater genetic liability. This association of a strong genetic contribution to etiology with negative symptoms in contrast to the association of the pharmacological dopamine hypothesis with positive symptoms underscores the importance of understanding both negative and positive symptoms. Given that response to biological treatment and genetic contributions to etiology are the two cornerstones of biological theories of schizophrenia and yet are associated with semiindependent dimensions of schizophrenia, it is obvious that something more is needed for an adequate conceptualization of schizophrenia.

Activity-Withdrawal

Although it is less frequently recognized as defining heterogeneity in schizophrenia, the activity-withdrawal dimension nevertheless is of some importance. Assessed by ward ratings on 10 items (Venables, 1957), patients with schizophrenia at the active end are described as restless, loud, overtalkative, overactive, and having many friends and interests (actually reminiscent of mania), whereas withdrawn cases are described by an absence of these features. In spite of being behaviorally inactive, withdrawn patients with schizophrenia were found to be more highly aroused on the basis of two perceptual measures of cortical reactivity and one autonomic measure (Venables, 1963a, 1963b, 1967; Venables & Wing, 1962), whereas active patients were low on arousal. Thus, the active schizophrenic patients showed a combination of behavioral activation, reward seeking, and low cortical arousal, whereas the withdrawn schizophrenic patients showed a pattern of anhedonia, behavioral inactivity, and high arousal (presumably aversive arousal or anxiety, discussed later in this research paper). Depue (1976) demonstrated a close congruence between activity-withdrawal and the good-poor premorbid adjustment distinction (withdrawn individuals showing poor premorbid adjustment). Wing’s work on withdrawal in schizophrenia provided the background for the negative symptom component of Crowe’s Type II symptom pattern (Crow, 1985), indicating the relevance of these findings to negative symptoms. The active cases have received little attention in the recent literature.

Schizoaffective Symptoms

The traditional disease model implies a discrete, categorical distinction for different psychiatric disorders. Gradations of severity have long challenged that view, making it difficult to draw a sharp distinction between health and illness. Schizoaffective symptoms challenge the categorical approach in a different way: Even the boundaries between hypothesized disorders are blurred. Although it is not the concern of the present review, the past two decades or so have brought increasing awareness of the generality of comorbidity across many forms of psychopathology and the subsequent blurring of boundaries (Mineka, Watson, & Clark, 1998). Schizoaffective symptoms illustrate this phenomenon well.

As already noted, when patients are arranged along a continuum from pure schizophrenia at one end to pure affective disorder at the other end, the distribution is unimodal; patients with schizoaffective symptoms outnumber those with purely schizophrenic or purely affective symptoms (Kendell, 1982). The prevalence of schizoaffective disorder is far too common to be attributed to the chance occurrence of two relatively rare disorders (Procci, 1989; Siris & Lavin, 1995). Among hospital admissions for functional psychosis, depending on diagnostic criteria, 10–30% manifest schizoaffective disorder (Siris & Lavin, 1995). Even with narrow diagnostic systems that preclude a diagnosis of schizophrenia when prominent depression is present, significant depression is seen during the longitudinal course of 25–50% of schizophrenia patients (Sands & Harrow, 1999). The theoretical challenge of the large number of schizoaffective disorders is well recognized (Baron & Gruen, 1991; Crow, 1986, 1991; Grossman, Harrow, Goldberg, & Fichtner, 1991; Kendell, 1982; Maier et al., 1993; Meltzer, 1984; Sands & Harrow, 1999; Siris & Lavin, 1995; Taylor, 1992; Taylor, Berenbaum, Jampala, & Cloninger, 1993). Attempts to dichotomize the schizoaffective continuum on the basis of underlying etiology has failed, leading Crow (1998) to conclude that no objective genetic boundaries can be drawn between predominantly schizophrenic and predominantly affective patients.

Although for a while it was popular to argue that schizoaffective disorders are a variant of affective disorders (Meltzer, 1984), the evidence has not been consistent (Procci, 1989), and in any case, supportive studies tend to select patients for more prominent affective than schizophrenic symptomatology (Siris & Lavin, 1995; Williams & McGlashan, 1987). As noted earlier, when patients are selected to have schizoaffective symptoms with more prominent schizophrenic symptoms, they have more familial schizophrenia and less familial affective disorder (Kendler et al., 1994; Levinson & Levitt, 1987). Furthermore, when both the full schizophrenic syndrome and the full affective disorder syndrome are present, follow-up data point to a closer relationship to schizophrenia (Williams & McGlashan, 1987).

Such results are consistent with a continuum in which the vulnerability processes underlying affective disorders combine additively with vulnerability to schizophrenia to reach threshold for development of a psychotic episode (Braden, 1984; Fowles, 1992b; Siris & Lavin, 1995). In this model, the prominence of schizophrenic versus affective symptoms varies with the magnitude of the underlying contribution from schizophrenia or affective disorder. If vulnerability to schizophrenia is the major contributor and is combined with a small contribution from affective processes, then the clinical picture will be largely schizophrenic or schizoaffective, mainly schizophrenic. As the vulnerability to schizophrenia decreases, more liability from the processes underlying affective disorders is required to reach threshold for schizophrenia, and the associated affective symptoms become more prominent. This model can account for the continuum from schizophrenia to affective disorders, the temporal coupling between schizophrenic and affective symptomatology seen in many patients, the more favorable prognosis when affective symptoms are prominent, the higher number of schizoaffective diagnoses that is predicted on the basis of the chance occurrence of two relatively infrequent disorders, and the family and genetic findings (Siris & Lavin, 1995).

The model just presented can be strengthened by evidence that the processes underlying affective disorders should contribute to liability for schizophrenia—that is, some linkage between the two liabilities that makes affective liability relevant to the MFP model of schizophrenia. In order to make this argument effectively and to address contributors to positive and negative symptoms, it is necessary to consider the implications of the MFP model for the importance of affective processes and the associated underlying systems.

Affective Neurobehavioral Systems as a Nonspecific Liability

In considering the problem of heterogeneity in schizophrenia, the diathesis-stress and MFP models point to the processes associated with stress as the origin of an important component of nonspecific liability for schizophrenia—consistent with recent conclusions that genetic factors act via various dimensional risk factors involved in the multifactorial origins of psychopathology rather than producing discrete categorical diseases (Rutter, 1997). Neurobehavioral systems associated with affective-emotional responses are obviously relevant— both as affective responses to stressful environments and as dimensions of temperament-based individual differences in stress reactivity. Activity in these systems might be expected to constitute nonspecific liability and to influence the clinical presentation of schizophrenia. In this context, Fowles (1992b, 1994) cited Gray’s and Depue’s neurobehavioral theories as particularly useful. For Gray (1982, 1987; Gray & McNaughton, 1996, 2000) the motivational-affective systems in question are the behavioral approach or behavioral activation system (BAS), the behavioral inhibition system (BIS; Gray & McNaughton, 2000), and the fight-flight system (Gray & McNaughton, 2000), all derived from the literature on animal learning and motivation but integrated with findings in behavioral neuroscience. Depue (Depue & Collins, 1999; Depue, Collins, & Luciana, 1996; Depue & Iacono, 1989; Depue & Lenzenweger, 2001) uses the term behavioral facilitation system (BFS) to describe a system almost identical to the BAS, and he has written extensively on the relevance to personality and affective disorders. Depue’s and Gray’s primary contributions have been to the BFS and BIS, respectively; thus, those terms are used here.

The BFS: Reward Seeking, Coping With Stress, Positive Affect, and Dopamine

The BFS facilitates goal-directed or reward-seeking behavior and increases nonspecific arousal in response to conditioned stimuli (CS) for reward. It is important to note that the BFS also facilitates punishment-avoiding behavior in response to cues for relieving nonpunishment in active avoidance situations (in which some instrumental response can avoid a threatened punishment). The positive affective states accompanying the behavioral activation are hope and relief, respectively. The substrate for the BAS-BFS centers on the ventral tegmental area (VTA) dopamine projection system—the VTA dopamine projection to the nucleus accumbens (the mesolimbic dopaminergic pathway) and other structures (see Depue & Collins, 1999, for additional dopamine projections in this system).

Several aspects of the BFS are especially important in the present context. First, its role in active avoidance means that it will be activated during any stressful situation in which the animal (or person) expects that some coping response may be effective in dealing with the potentially negative outcome. Second, the neurotransmitter dopamine is centrally involved in the substrate for the BAS (Gray, 1987) or the BFS (Depue & Collins, 1999; Depue & Iacono, 1989). The same dopaminergic pathways mediate the rewarding effects of many addictive drugs such as amphetamines, cocaine, and heroin (Leshner, 1997; Wise & Bozarth, 1987; Wise & Rompre, 1989). There is agreement, therefore, that dopamine pathways are involved in activating behavior in response to cues for reward, and both Gray and Depue describe a behavioral system that would be involved in coping with stress. Consistent with that view, it is well established that exposure to stress increases dopamine release (Walker & Diforio, 1997; Weinberger, 1987). Third, the BFS has been implicated in mania and depression, making it relevant to schizoaffective disorders. Fourth, the BFS is likely to influence the degree of behavioral activation seen clinically in different subtypes of schizophrenia.

The BIS: Passive Avoidance, Extinction, Anxiety, and Anxiolytic Drugs

The BIS inhibits behavior, increases nonspecific arousal, and facilitates attention to the environment in two important situations involving goal conflict: (a) approach-avoidance conflict (also called passive avoidance), in which an animal may receive punishment for making a rewarded response, such as crossing an electrified grid in order to reach food, or—in a more naturalistic setting—exploring for food when a predator might be present; and (b) extinction, in which the absence of an expected reward produces frustration. Both situations create a conflict between the desire to approach the reward and the desire to avoid the punishment or frustration by not approaching. All anxiolytic drugs—alcohol, barbiturates, benzodiazepines, the novel anxiolytics (e.g., buspirone), and the anxiolytic antidepressant imipramine—produce behavioral and neurobiological effects that can be conceptualized as weakening the BIS (Gray & McNaughton, 2000). These findings contribute to Gray’s conclusion that the BIS is the anxiety system, and the BIS is seen as a substrate for Barlow’s concept of anxiety that is common to all of the anxiety disorders (Barlow, Chorpita, & Turovsky, 1996; Fowles, 1992a). The BIS, therefore, processes stressful (i.e., anxietyproducing) stimuli, increases nonspecific arousal, and produces a clinical picture of reduced approach behavior. The substrate for the BIS centers on the septo-hippocampal system, which includes the hippocampus, the dentate gyrus, the subicular area, the entorhinal cortex, and the posterior cingulate cortex (Gray & McNaughton, 2000).

The Fight-Flight System: Dealing with Imminent Threats

Gray has described a third system, the well-known fight-flight system, that prepares the organism for vigorous activity in the form of flight or flight. In his earlier work Gray (1987) said the fight-flight system is activated by unconditioned punishment stimuli, but recently Gray and McNaughton (2000) described the fight-flight system as activated when there is an actual threat (as opposed to a potential threat), such as when a predator is present. The term fear is commonly used (by Gray and others) to refer to activation of the fight-flight system.The fight-flight system, then, is a third stress-relevant system, but it is one that responds to imminent danger. The neurobiological substrate of the fight-flight system centers on the dorsal periaqueductal gray, but Gray and McNaughton (2000) suggested viewing it as part of a distributed system that also includes the hypothalamus, amygdala, anterior and posterior cingulate, and hippocampus.

The BIS-BFS and Heterogeneity in Schizophrenia

Fowles (1992b, 1994) suggested that the concepts of a BFS and a BIS offer potential explanations for some aspects of heterogeneity in schizophrenia. The major considerations are that the BFS is associated with behavioral activation (active coping with stress, reward-seeking behaviors), positive affect, and dopaminergic activity, whereas BIS activation is associated with behavioral inhibition in conflict situations and with increased aversive arousal and negative affect in the form of anxiety. Although both systems are implicated in responding to threatening stimuli, both the behavioral effects and the underlying systems are quite different and can be expected to affect those vulnerable to schizophrenia differently.

The BFS is directly relevant to the manic form of schizoaffective disorder. Depue (Depue & Iacono, 1989; Depue, Krauss, & Spoont, 1987; Depue & Zald, 1993) argues that the features of manic episodes can be understood as resulting from uncontrolled activation of the BFS (with resultant high levels of dopaminergic activity) as a consequence of a breakdown of regulation. For example, the manic features of increased activity, positive affect, and uncritical optimism while engaging in risky activities are to be expected from uncontrolled activation of a moderately strong BFS. The resultant high levels of dopaminergic activity would combine with specific vulnerability to schizophrenia, as suggested previously, to produce a continuum of schizoaffective symptomatology involving positive symptoms (see Braden, 1984, for a similar proposal, albeit without notions of the BFS, in which he suggests that the dopamine-based behavioral activation seen in mania combines with vulnerability to schizophrenia to produce schizoaffectivemanic disorder).

The activity-withdrawal dimension should reflect the balance between the BFS and the BIS. Atemperamentally strong BFS would be expected to produce positive schizophrenic symptoms in connection with a behaviorally active, rewardseeking individual, as described previously for cases of active schizophrenia. Additionally, stressful situations that involve active coping by the individual would be expected to produce an increase in dopaminergic activity, triggering positive symptom episodes of schizophrenia in vulnerable individuals. In contrast, a temperamentally weak BFS would result in weak reward-seeking and active avoidance with little positive affect. Additionally, the finding of high arousal in cases of withdrawn schizophrenia suggests that these negative-symptom patients suffer from high aversive arousal combined with behavioral inhibition (passive avoidance and extinction) due to activation of the BIS. The consequences of such temperament factors would be increased by an environment in which few responses were rewarded and aversive control was prominent and would also be increased by depression.

The first consideration in the application to schizoaffective disorder with depression is that depression involves low positive affect (e.g., Mineka et al., 1998; Tellegen, 1985), which can be seen as a result of greatly diminished activation of the BFS (e.g., Fowles, 1994). All behavioral theories of depression emphasize a blocking of reward-seeking behavior, punishment-avoiding behavior, or both as core etiologic factors (Eastman, 1976). Cognitive theories emphasize hopelessness, by which they mean a lack of hope that behavior will be effective in achieving desired goals (obtaining rewards or avoiding punishments). Furthermore, Depue (Depue & Iacono, 1989; Depue et al., 1987; Depue & Zald, 1993) concluded that depression in the context of bipolar disorder can be understood as a loss of BFS functioning—that is, the BFS no longer responds to the usual reward incentive cues. Thus, whether from a behavioral, cognitive, or biological perspective, depression is seen as involving a lack of the activities attributed to the BFS.

At the same time, depression, which shares a genetic diathesis with generalized anxiety disorder in the form of a predisposition to general distress (Kendler, Neale, Kessler, Heath, & Eaves, 1992), is associated with stress-related HPA activation (Walker & Diforio, 1997). This high level of negative affectanxiety-distress should contribute nonspecifically to liability for schizophrenia, combining with specific liability to produce a continuum of schizoaffective-depressed symptomatology—a proposal also articulated by others. For example, Kendler and Diehl (1993) commented that “affective illness represents a stress that might precipitate a psychosis” (p. 269). More formally, Siris and Lavin (1995) articulated a shared diathesis in which “an episode of a major mood disorder may constitute a sufficient stressor” to elicit schizoaffective symptoms in individuals with a moderate schizophrenia diathesis (p. 1021). Presumably, it must be the high negative affect component of depression that functions as a stressor.

According to this analysis, the combination of minimal BFS functioning and high negative affect is one etiologic pathway to the withdrawn clinical picture, characterized by aversive arousal and lack of behavioral activation (prominent negative symptoms). The well-established evidence that depression can produce negative symptoms (e.g., Sommers, 1985) is consistent with this suggestion. Furthermore, the well-known negative symptom phenomenon of institutionalism, seen in schizophrenia patients exposed to nonstimulating and unrewarding psychiatric hospital environments (e.g., Sommers, 1985; Wing, 1978), can be understood as an extinction phenomenon that disrupts BFS activation and produces depression.

Thus, the BFS and the BIS are likely contributors to nonspecific liability in schizophrenia. Because of their different features, they offer explanations for some of the heterogeneity in schizophrenia. Although the precise conceptualization of these stress-relevant affective-motivational systems will evolve with additional research, the genetics literature in the form of the MFP model and its nonspecific contributors to liability suggests that they are relevant to understanding the etiology and heterogeneity in schizophrenia. The contribution of the fight-flight system and panic attacks is less clear, although it seems likely that panic attacks constitute biological stressors and that they promote withdrawal and avoidance of situations in which they occur.

The Deficit Syndrome

Carpenter and his colleagues have focused on a deficit syndrome, reflecting Kraepelin’s avolitional psychopathology (Carpenter, 1994; Carpenter, Heinrichs, & Wagman, 1988). A key to defining the deficit syndrome is the long-standing recognition that negative symptoms arise from a number of sources other than from schizophrenia per se (Sommers, 1985). To the institutionalism and postpsychotic depression mentioned previously can be added akinesia secondary to pharmacological treatment with dopamine antagonists. Narrowly defined, akinesia refers to extrapyramidal motor symptoms (diminished arm swing, shortened stride, and rigid posture); broadly defined, however, it includes a broad range of negative features (e.g., lack of emotional reactivity, lack of goal-directedness, lack of or retarded spontaneous speech, and masklike facial expressions very similar to flat affect), quite possibly due to impairment of a dopamine-based neurobehavioral system (Harrow, Yonan, Sands, & Marengo, 1994) conceptualized here as the BFS. Carpenter, Buchanan, Kirkpatrick, Thaker, and Tamminga (1991) developed diagnostic criteria that attempted to exclude secondary negative symptoms and to define a primary deficit syndrome. These criteria required the enduring presence of any two of the following: restricted affect, diminished emotional range, poverty of speech, curbing of interests, diminished sense of purpose, and diminished social drive. This definition is narrower than those of other commonly used scales for rating negative symptoms, which do not exclude secondary negative symptoms (Kirkpatrick et al., 2001).

Recently, Carpenter and his colleagues (Kirkpatrick et al., 2001) proposed that deficit psychopathology or a deficit schizophrenia subtype is a disease distinct from schizophrenia without deficit features, at least based on preliminary evidence to date. They estimated that the deficit schizophrenia is seen in about 15% of first-episode schizophrenia cases. In support of the view that deficit schizophrenia is a distinct disease, Kirkpatrick et al. (2001) cite a range of evidence comparing deficit and nondeficit schizophrenia, among which were (a) a worse premorbid adjustment with insidious onset; (b) poorer occupational and social functioning; (c) for relatives of deficit schizophrenia probands, an increased risk of schizophrenia, more severe social withdrawal, and a threefold increase in the risk of deficit rather than nondeficit schizophrenia; (d) a small increase in the prevalence of antibodies for the Borna disease virus (and these antibodies are associated with greater severity of negative but not positive symptoms); (e) an increase in summer rather than winter births (suggesting that winter births are associated with a more benign course of nondeficit schizophrenia); (f) impairments on neurocognitive measures sensitive to frontal and parietal lobe dysfunction; (g) greater oculomotor (eye-tracking) dysfunction; and (h) decreased activity in the dorsolateral prefrontal cortex and the inferior parietal cortex that they suggest is consistent with dysfunction of the dorsolateral prefrontal basal ganglia-thalamocortical circuit (DLPFC).

If the deficit syndrome simply were a more severe version than the nondeficit syndrome of schizophrenia, then (the authors argued) the deficit syndrome should appear to be more extreme on a variety of problems characterizing schizophrenia in general. Contrary to this expectation, deficit schizophrenia patients have been found not to be more severe in terms of overall psychotic symptoms and have been found to be less severe in terms of (a) reduced frontal lobe white matter, (b) winter births, (c) and some measures of positive symptoms. The authors also cite less severe depression and other dysphoric affect even on follow-up, but it is not clear to what extent this correlate might be contaminated by excluding patients with depression in the initial diagnosis of deficit schizophrenia. Additionally, among the nonpsychotic relatives of deficit schizophrenia probands, dysphoria and psychotic-like symptoms are less severe than they are among the nonpsychotic relatives of nondeficit schizophrenia probands. Finally, among deficit schizophrenia patients, the disorganization symptom dimension is uncorrelated with impairment in the sequencing of complex motor acts, whereas among schizophrenia patients in general, there is a positive correlation.

These findings underscore the importance of the positive versus negative symptom distinction, and the authors have argued against a single dimension in which the negative symptom or deficit schizophrenia group is more severe on all dimensions. On the other hand, it does not appear to be the case that deficit schizophrenia breeds true, even with the more restrictive diagnostic approach developed by Carpenter and his colleagues (i.e., the risk of deficit schizophrenia among relatives of deficit probands is increased quantitatively, but nondeficit cases still appear).

An alternative interpretation would attribute their findings to greater severity of genetic loading combined with contributions from nonspecific liability. As noted previously, when examining negative symptom patients, Gottesman et al. (1987) found no results that could not be attributed to a continuum of severity. With an underlying continuum of severity, threshold effects are possible that create apparent discontinuous phenotypic variation (Gottesman & Shields, 1972). The difference in winter births and reduced frontal lobe white matter could reflect a need for additional stressors (in this case, physical environmental) for nondeficit schizophrenia that involves a lesser genetic loading, as suggested previously.

The implications of the association between negative symptoms and the presence of antibodies for the Borna disease virus are not clear. Borna disease is an immune-mediated meningoencephalitis known to infect a number of animal species and recently believed to infect humans as well. It is believed to cause damage to the hippocampus that (at least in theory), in turn, results in diminished activity in the frontal cortex (hypofrontality as discussed later) and a hypodopaminergic state (Waltrip et al., 1997). Thus, Borna disease might constitute a contributor to liability that is nonspecificwithrespecttoschizophreniabutspecificwithrespect to negative symptoms due to the hypodopaminergic state. Needless to say, more research is needed on the contribution of Borna disease.

However, the position adopted here is that some heterogeneity cannot be accounted for simply in terms of severity; rather, it reflects multidimensional components of nonspecific liability. The negative syndrome may reflect a weak BFS and a strong BIS in combination with an unrewarding and/or aversive environment, producing a failure to initiate approach and active avoidance behavior and a lack of positive affect—components of the avolitional syndrome. The increased BIS-related aversive arousal may constitute an important source of nonspecific liability in these patients. The reduced dopaminergic activity associated with reduced BFS activity would reduce positive symptomatology, making deficit syndrome patients appear less severe on positive symptoms. In addition to temperament variables, it is possible that more severe schizophrenia is associated with a central nervous system (CNS) dysfunction that directly or indirectly disrupts functioning of the BFS, adding to the effects of temperament variables. Such a CNS dysfunction could possibly reflect the effects of a single major gene, acting in a polygenic context in a minority of schizophrenic patients and producing a more severe form of schizophrenia. To date, no such gene has been identified, but as noted previously, it is not incompatible with existing data.

Brain Dysfunction Hypotheses

The features of schizophrenia are so extreme that it is difficult to imagine that they could be produced by psychosocial experiences alone as a normal part of learning. Consequently, a great many researchers believe that some type of brain dysfunction plays an important role in schizophrenia. Research on this hypothesis has been greatly facilitated by the development of modern technology that allows imaging of both brain structure and function. Improvements in resolution of these techniques have facilitated this effort. Two major findings have been enlarged ventricles and hypofrontality (diminished activation of the frontal cortex) in cases of schizophrenia.

Enlarged Ventricles and Hypofrontality

Ventricles (pools of cerebrospinal fluid in the brain) can be measured with structural techniques such as magnetic resonance imagery. Enlarged ventricles (not attributable to medication) repeatedly have been found in a portion of cases, and it is also interesting to note that increased ventricular size is associated with the schizophrenic twin in discordant MZ twins (Berman & Weinberger, 1999; Bunney & Bunney, 1999). The enlarged ventricles imply smaller brain size but do not indicate the cause or locus. Although some evidence has supported reduced neural tissue in a number of areas, no clear picture has emerged at present (Andreasen, 2001; Berman & Weinberger, 1999; Bunney & Bunney, 1999), and it is quite likely that no single area accounts for the enlarged ventricles. Nevertheless, these findings strongly encourage theories of brain dysfunction. As discussed later in this research paper, those theories have suggested a problem in neural development rather than a degeneration.

Another well-replicated finding—relying on functional assessments of brain activity (changes in regional blood flow secondary to neural activity)—is that schizophrenia involves less activation of the prefrontal cortex during cognitive activation tasks, especially those involving working memory (Berman & Weinberger, 1999). This presumed deficit in prefrontal activity, which is not attributable to medication, is strongly supported by performance deficits on a wide range of neuropsychological tests tapping prefrontal cortical function, and both hypofrontality and neuropsychological deficits appear to be more characteristic of individuals with negativesymptom schizophrenia. Hypoactivation or hyperactivation have been reported in other regions of the brain, but the results have not been as consistent as those for the prefrontal cortex.

The findings of hypofrontality and its association with negative symptoms have served as the basis of speculations concerning a complex role of dopamine. In a classic paper, Weinberger (1987) suggested that a pathological process of unknown origin early in life interferes with the functioning of parts of the prefrontal cortex (especially the DLPFC) and limbic system, functionally compromising activity in the mesocortical dopamine pathways (i.e., producing underactivity of this dopamine pathway). The hypoactivity in the prefrontal cortex releases the mesolimbic dopamine pathways from feedback control, producing hyperactivity in this dopamine pathway. In this theory, negative symptoms are related to underactivity of the mesocortical dopamine pathways, whereas positive psychotic symptoms are related to overactivity of the mesolimbic pathways.

More recently, Byne et al. (1999) found support for an association between hypofrontality and decreased activity of mesocortical dopamine pathways, for a causal connection between hypofrontality and negative symptoms, and for an inverse relationship between prefrontal cortex dopamine activity and subcortical dopamine activity. However, evidence of increased subcortical D₂ dopamine receptors in schizophrenia is suggestive of underactivity of subcortical dopamine.An attempt to resolve this contradiction proposed that prefrontal cortical afferents fail to produce enough tonic mesolimbic dopamine release that results in up-regulation (i.e., an increase) of D₂ receptors. The increased D₂ receptors then respond more strongly to dopamine release caused by neural firing, producing a functional hyperactivity of this pathway in response to stimulation. This revised model permits the same predictions as the Weinberger (1987) model by virtue of modifications that take into account findings of up-regulation of D₂ receptors in mesolimbic pathways. The authors emphasize, however, that the model is speculation that has yet to be tested. Nevertheless, it serves to illustrate the complexity of current versions of the dopamine hypothesis.

Neurodevelopmental Hypotheses: Distributed Networks and Poor Connectivity

Recent theorizing has shifted from searching for a deficit in a single area of the brain to an emphasis on interactions among many different regions, leading to hypotheses of abnormal interconnectedness of different areas as a result of neurodevelopmental failures (Andreasen, 2001; Berman & Weinberger, 1999; Bunney & Bunney, 1999; Weinberger, 1987; Weinberger & Lipska, 1995). Weinberger and Lipska (1995) noted that during the second trimester of pregnancy, young neurons migrate and settle into their appropriate target sites; they suggested that schizophrenia involves a failure of this process in which neurons may not only fail to make some connections, but may also even make incorrect connections, resulting in inefficient, noisy processing (rather than no processing) between areas of the brain—specifically between the prefrontal and temporal and limbic areas, although other areas may be involved (see Berman & Weinberger, 1999). These neural systems are implicated not only in complex cognitive and psychological behaviors, but also in the regulation of subcortical dopamine systems during periods of stress. The authors suggested that this deficit is almost universally present in schizophrenia but falls on a continuum of severity. Animal data indicate that problems in dopaminergic regulation due to malfunction of the prefrontal-temporolimbic cortical neural systems do not appear until postpuberty, consistent with the age of onset of schizophrenia.

Bunney and Bunney (1999) suggested a similar neurodevelopmental model, also emphasizing abnormal functional connectivity between brain regions that arises especially during the second trimester of pregnancy. They focused on the cortical subplate, a transitory structure critical to formation of neural connections in the cortex during early brain development, a disturbance of which is suggested to lead to abnormal connections, particularly in the frontal and limbic regions and in thalamocortical connections. Disruption of this process during the second trimester by obstetrical complications and viral infections is seen as a second hit that adds to genetic vulnerability in a minority of schizophrenic patients. Although the authors did not explicitly suggest the idea, it is easy to imagine (as suggested previously) that these second hits play a critical role in individuals with a moderate degree of genetic liability for schizophrenia but are not required for individuals with greater genetic liability.

Andreasen (2001) also mentioned pregnancy and birth complications, but she proposed that negative influences on brain development probably occur at multiple times, acting in an additive manner and potentially extending from the intrauterine period to late adolescence or young adulthood (during which important neural connections are being made). The range of environmental influences that affect brain development include “head or birth injuries, viral infections, exposure to toxins and drugs of abuse, hormonal changes, and other factors” (p. 206). Although she cites evidence of decreased size of the prefrontal cortex and the hippocampus (a part of the limbic system), she reported that schizophrenic patients show abnormal patterns of neural activity (inferred from blood flow) in the thalamus and the cerebellum in addition to the prefrontal cortex to a wide variety of tasks. However, Andreasen concluded that no specific regional abnormalities have been identified in schizophrenia. She suggested that the fundamental problem is a dysfunction of functional connectivity between distributed neural circuits (different areas of the brain)—a misconnection syndrome. It is interesting to note that she encouraged further work on cognitive relearning treatment programs as an adjunct to medication for schizophrenia patients, in the hope that such extensive retraining may gradually form new neural connections.

Thus, there appears to be a convergence among different investigators that the brain dysfunction in schizophrenia is more subtle and complex than a lesion in a specific area and that it is more than likely that the problem arises as a failure of development. Additionally, the notions of faulty neural wiring or misconnections suggest a more diffuse deficit rather than the loss of a specific single psychological process. These theoretical models, then, are consistent with the complexity of deficits seen in schizophrenia and perhaps explain why it has been so difficult to identify a specific psychological deficit in schizophrenia (Chapman & Chapman, 1973; Strauss, 2001). These models also suggest that pharmacological manipulations can only partially ameliorate the problem, inasmuch as drugs cannot restore appropriate neural connections. Although this neurodevelopmental perspective has many positive features, it is an approach to looking for the neurobiological substrate of schizophrenia rather than providing support for any specific theory. Indeed, there is no consensus on the specific developmental insult that contributes to the etiology of schizophrenia (Meinecke, 2001).

Summary

The emotional, cognitive, and behavioral problems subsumed under the terms schizophrenia and schizophrenia spectrum constitute a complex and heterogeneous phenomenon that is only approximated by our current conceptualizations and diagnostic approaches. Opinions differ with respect to the importanceof—andwaystoconceptualize—chronicity,negative versus positive symptoms, and affective symptoms, and decisions on these issues exert a large effect on attempts to diagnose schizophrenia. In spite of these uncertainties, a large literature on family, twin, and adoption studies have documented a very large genetic contribution to schizophrenia. These same studies indicate that schizotypal personality disorder and schizoaffective disorder, mainly schizophrenic belong in the schizophrenia spectrum and suggest that schizophreniform disorder and paranoid personality disorder may well be included. There is little or no positive support for inclusion of brief reactive psychosis.

The extant genetic data are most consistent with the MFP model, in which a large number of genes, each often assumed to be of small effect, and environmental influences contribute in an additive fashion to the overall liability for schizophrenia. Sources of liability include specific genetic, nonspecific genetic, and nonspecific environment. Schizophrenia develops when total liability exceeds a threshold. Although a large majority of cases of schizophrenia are attributed to the MFP model, it is quite possible that some rare cases eventually will be attributed primarily to single genes and others primarily to environmental factors. Furthermore, within the MFP majority, it is possible that a minority eventually will be attributable in part to a major gene (combined with polygenes and the environment) with low penetrance.

The phenotypic heterogeneity of schizophrenia challenges all theoretical approaches. This research paper attempted to conceptualize heterogeneity in terms of the MFP model, including a consideration of affective systems relevant to nonspecific liability. From the perspective of the MFP model, a greater genetic loading will on average be associated with a more chronic course, accounting for one aspect of heterogeneity. The frequent pairing of affective and schizophrenic symptoms suggests that the affective and schizophrenia vulnerabilities combine additively to produce a continuum from relatively pure schizophrenia to relatively pure affective disorder with many cases of schizoaffective symptomatology. Depue’s work indicates that mania involves uncontrolled activation of the BFS, and it was suggested here that the dopaminergic activity central to this system contributes nonspecifically to schizophrenia liability. Depression involves suppression of this same system, combined with high negative affect–anxiety. It was proposed that the negative affect contributes nonspecifically to liability and that the failure to activate the BFS results in negative symptoms in the form of social withdrawal. Negative symptoms also were attributed to a temperamentally weak BFS combined with a temperamentally strong BIS, along with a social environment that is unrewarding and aversive. The extreme negative symptoms seen in the deficit schizophrenia subtype described by Carpenter and his colleagues are likely to reflect poor functioning of the BFS, possibly due to a CNS dysfunction that directly or indirectly disrupts the BFS or due to extremes of temperament combined with unrewarding-aversive environments.

Given the difficulty of accounting for schizophrenia symptoms in terms of psychosocial environmental events alone, intense interest naturally has focused on theories of brain dysfunction. These efforts, however, have been hindered by the crudeness of techniques for assessing CNS functioning and by the complexity of schizophrenia. Two major findings have been enlarged ventricles (a structural measure indirectly indicating smaller brain size) and hypofrontality (a functional measure showing diminished activation of the frontal cortex). Neither finding has resulted in clear conclusions concerning the neurobiological substrate for schizophrenia. More recently, the earlier search for a deficit in a single area of the brain has been replaced by an emphasis on interactions among many different regions, with hypotheses centering on abnormal interconnectedness of different areas as a result of neurodevelopmental failures. As interesting as these hypotheses are, again no firm conclusions have been reached.

The conceptualization of schizophrenia in the present research paper attempted to account for heterogeneity while retaining the parsimonious assumption of an underlying schizophrenic process varying in severity, to a substantial degree as a function of genetic liability. In that sense, it embraced Carpenter’s (1992) second model, seeing schizophrenia as a single disease entity with heterogeneity attributed to the interaction between this primary pathology and other characteristics of the individual and the environment. By hypothesizing two affective neurobehavioral systems that contribute different types of nonspecific liability and have different effects on behavior, it was possible to account for important aspects of heterogeneity without giving up the concept of schizophrenia as a distinct phenomenon. While this approach, like all biological theories of schizophrenia, must be considered a work in progress, it is argued here that such important functional neurobehavioral systems cannot be ignored in any adequate approach to developing psychobiological theories of psychopathology.

Bibliography:

1. (1980).Diagnosticandstatistical manual of mental disorders (3rd ed.). Washington, DC: Author.
2. American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed. Rev.). Washington, DC: Author.
3. (1994).Diagnosticandstatistical manual of mental disorders (4th ed.). Washington, DC: Author.
4. (2000).Diagnosticandstatistical manual of mental disorders (4th ed.,Text Revision).Washington, DC:Author.
5. Andreasen, N. C. (1985). Positive vs. negative schizophrenia: A critical evaluation. Schizophrenia Bulletin, 11, 380–389.
6. Andreasen, N. C. (2001). Brave new brain. New York: Oxford University Press.
7. Andreasen, N. C., Arndt, S., Alliger, R., Miller, D., & Flaum, M. (1995). Symptoms of schizophrenia: Methods, meanings, and mechanisms. Archives of General Psychiatry, 52, 341–351.
8. Andreasen, N. C., & Carpenter, W. T., Jr. (1993). Diagnosis and classification of schizophrenia. Schizophrenia Bulletin, 19, 199–214.
9. Andreasen, N. C., Flaum, M., Swayze, V. W., II, Tyrell, G., &Arndt, S. (1990). Positive and negative symptoms in schizophrenia: A critical reappraisal. Archives of General Psychiatry, 47, 615–621.
10. Asarnow, R. F., Nuechterlein, K. H., Fogelson, D., Subotnik, K. L., Payne, D. A., Russell, A. T., Asamen, J., Kuppinger, H., & Kendler,K.S.(2001).Schizophreniaandschizophrenia-spectrum personality disorders in the first-degree relatives of children with schizophrenia. Archives of General Psychiatry, 58, 581–588.
11. Astrachan, B. M., Harrow, M., Adler, D., Brauer, L., Schwartz, A., Schwartz, C., & Tucker, G. (1972). A checklist for the diagnosis of schizophrenia. British Journal of Psychiatry, 121, 529–539.
12. Barlow, D. H., Chorpita, B. F., & Turovsky, J. (1996). Fear, panic, anxiety, and disorders of emotion. In D. A. Hope (Ed.), Nebraska Symposium on Motivation (Vol. 43, pp. 251–328). Lincoln: University of Nebraska Press.
13. Baron, M., & Gruen, R. S. (1991). Schizophrenia and affective disorder: Are they genetically linked? British Journal of Psychiatry, 159, 267–270.
14. Bassett, A. S., Chow, E. W. C., O’Neill, S., & Brzustowicz, L. M. (2001). Genetic insights into the neurodevelopmental hypothesis of schizophrenia. Schizophrenia Bulletin, 27, 417–430.
15. Battaglia, M., & Torgersen, S. (1996). Schizotypal disorder: At the crossroads of genetics and nosology. Acta Psychiatrica Scandinavica, 94, 303–310.
16. Berman, K. F., & Weinberger, D. R. (1999). Neuroimaging studies of schizophrenia. In D. S. Charney, E. J. Nestler, & B. S. Bunney (Eds.), Neurobiology of mental illness (pp. 246–257). Oxford, England: Oxford University Press.
17. Birley, J. L. T., & Brown, G. W. (1970). Crises and life changes preceding the onset or relapse of acute schizophrenia: Clinical aspects. British Journal of Psychiatry, 116, 327–333.
18. Bleuler, E. (1950). Dementia praecox or the group of schizophrenias. New York: International Universities Press. (Original work published 1911)
19. Bleuler, M. (1974). The long-term course of schizophrenic psychoses. Psychological Medicine, 4, 244–254.
20. Braden, W. (1984). Vulnerability and schizoaffective psychosis: A two-factor model. Schizophrenia Bulletin, 10, 71–86.
21. Brown, G., Birley, J. L. T., & Wing, J. K. (1972). Influence of family life on the course of schizophrenic disorder: A replication. British Journal of Psychiatry, 121, 241–258.
22. Bunney,W. E.,Jr.,&Bunney,B. G.(1999). Neurodevelopmental hypothesis of schizophrenia. In D. S. Charney, E. J. Nestler, & B. S.
23. Bunney (Eds.), Neurobiology of mental illness (pp. 225–235). Oxford, England: Oxford University Press.
24. Byne, W., Kemether, E., Jones, L., Haroutunian, V., & Davis, K. (1999). The neurochemistry of schizophrenia. In D. S. Charney, E. J. Nestler, & B. S. Bunney (Eds.), Neurobiology of mental illness (pp. 236–245). Oxford, England: Oxford University
26. Cardno, A. G., & Gottesman, I. I. (2000). Twin studies of schizophrenia: From bow-and-arrow concordances to Star Wars Mx and functional genomics. American Journal of Medical Genetics, 97, 12–17.
27. Cardno, A. G., Marshall, E. J., Coid, B., Macdonald, A. M., Ribchester, T. R., Davies, N. J., Venturi, P., Jones, L. A., Lewis, S. W., Sham, P. C., Gottesman, I. I., Farmer, A. E., McGuffin, P., Reveley, A. M., & Murray, R. M. (1999). Heritability estimates for psychotic disorders: The Maudsley twin psychosis series. Archives of General Psychiatry, 56, 162–168.
28. Carpenter, W. T., Jr. (1992). The negative symptom challenge. Archives of General Psychiatry, 49, 236–237.
29. Carpenter, W. T., Jr. (1994). The deficit syndrome. American Journal of Psychiatry, 151, 327–329.
30. Carpenter, W. T., Jr., Buchanan, R. W., Kirkpatrick, B., Thaker, G., & Tamminga, C. (1991). Negative symptoms: A critique of current approaches. In A. Marneros, N. C. Andreasen, & M. T. Tsuang (Eds.), Negative versus positive schizophrenia (pp. 126–133). New York: Springer-Verlag.
31. Carpenter, W. T., Jr., Heinrichs, D. W., & Wagman, A. M. I. (1985). On the heterogeneity of schizophrenia. In M. Alpert (Ed.), Controversies in schizophrenia: Changes and constancies (pp. 38–47). New York: Guilford Press.
32. Carpenter, W. T., Jr., Heinrichs, D. W., & Wagman, A. M. I. (1988). Deficit and nondeficit forms of schizophrenia: The concept. American Journal of Psychiatry, 145, 578–583.
33. Carpenter, W. T., Jr., Strauss, J. S., & Bartko, J. J. (1973). Flexible system for the diagnosis of schizophrenia: Report from the WHO International Pilot Study of Schizophrenia. Science, 182, 1275– 1278.
34. Chapman, L. J., & Chapman, J. P. (1973). Disordered thought in schizophrenia. Englewood Cliffs, NJ: Prentice-Hall.
35. Crow, T. J. (1980). Molecular pathology of schizophrenia: More than one disease process? British Medical Journal, 180, 66–68.
36. Crow, T. J. (1985). The two-syndrome concept: Origins and current status. Schizophrenia Bulletin, 11, 471–486.
37. Crow, T. J. (1986). The continuum of psychosis and its implications for the structure of the gene. British Journal of Psychiatry, 149, 419–429.
38. Crow, T. J. (1991). The search for the psychosis gene. British Journal of Psychiatry, 158, 611–614.
39. Crow, T. J. (1998). From Kraepelin to Kretschmer leavened by Schneider. Archives of General Psychiatry, 55, 502–504.
40. Cuesta, M. J., Peralta, V., & Caro, F. (1999). Premorbid personality in psychoses. Schizophrenia Bulletin, 25, 801–811.
41. Day, R., Nielsen, J. , Korten, A., Ernberg, G., Dube, K. C., Gebhart, J., Jablensky, A., Leon, C., Marsella, A., Olatawura, M., Sartorius, N., Stromgren, E., Takahashi, R., Wig, N., & Wynne, L. C. (1987). Stressful life events preceding the acute onset of schizophrenia: A cross-national study from the World Health Organization. Culture, Medicine, and Psychiatry, 11, 123–205.
42. Depue, R. A. (1976). An activity-withdrawal distinction in schizophrenia: Behavioral, clinical brain damage, and neurophysiological correlates. Journal of Abnormal Psychology, 85, 174– 185.
43. Depue, R. A., & Collins, P. F. (1999). Neurobiology of the structure of personality: Dopamine, facilitation of incentive motivation, and extraversion. Behavioral and Brain Sciences, 22, 491–569.
44. Depue, R. A., Collins, P. F., & Luciana, M. (1996). A model of neurobiology-environment interaction in developmental psychopathology. In M. F. Lenzenweger & J. J. Haugaard (Eds.), Frontiers of developmental psychopathology (pp. 44–77). New York: Oxford University Press.
45. Depue, R. A., & Iacono, W. G. (1989). Neurobehavioral aspects of affective disorders. Annual Review of Psychology, 40, 457–492.
46. Depue, R. A., Krauss, S. P., & Spoont, M. R. (1987). A twodimensional threshold model of seasonal bipolar affective disorder. In D. Magnusson & A. Ohman (Eds.), Psychopathology: An interactionist perspective (pp. 95–123). New York: Academic Press.
47. Depue, R. A., & Lenzenweger, M. F. (2001). A neurobehavioral dimensional model. In W. J. Livesley (Ed.), Handbook of personality disorders: Theory, research, and treatment (pp. 136–176). New York: Guilford Press.
48. Depue, R. A., & Zald, D. (1993). Biological and environmental processes in nonpsychotic psychopathology: A neurobehavioral perspective. In C. G. Costello (Ed.), Basic issues in psychopathology (pp. 127–237). New York: Guilford Press.
49. Dohrenwend, B. P., & Egri, G. (1981). Recent stressful life events and episodes of schizophrenia. Schizophrenia Bulletin, 7, 12–23.
50. Eastman, C. (1976). Behavioral formulations of depression. Psychological Review, 83, 277–291.
51. Endicott, J., Nee, J., Fleiss, J., Cohen, J., Williams, J. B. W., & Simon, R. (1982). Diagnostic criteria for schizophrenia. Archives of General Psychiatry, 39, 884–889.
52. Faraone, S. V., Green, A. I., Seidman, L. J., & Tsuang, M. T. (2001). “Schizotaxia”: Clinical implications and new directions for research. Schizophrenia Bulletin, 27, 1–18.
53. Faraone, S. V., & Tsuang, M. T. (1985). Quantitative models of the genetic transmission of schizophrenia. Psychological Bulletin, 98, 41–66.
54. Farmer, A. E., McGuffin, P., & Gottesman, I. I. (1987). Twin concordance for DSM-III schizophrenia. Archives of General Psychiatry, 44, 634–641.
55. Feighner, J. P., Robins, E., Guze, S. B., Woodruff, R. A., Winokur, G., & Munoz, R. (1972). Diagnostic criteria for use in psychiatric research. Archives of General Psychiatry, 26, 57–63.
56. Fenton, W. S., & McGlashan, T. H. (1992). Testing systems for assessment of negative symptoms in schizophrenia. Archives of General Psychiatry, 49, 179–184.
57. Fowles, D. (1992a). Motivational approach to anxiety disorders. In D. G. Forgays, T. Sosnowski, & K. Wrzesniewski (Eds.), Anxiety: Recent developments in cognitive, psychophysiological, and health research (pp. 181–192). Washington: Hemisphere.
58. Fowles, D. (1992b). Schizophrenia: Diathesis-stress revisited. Annual Review of Psychology, 43, 303–336.
59. Fowles, D. (1994). A motivational theory of psychopathology. In W. Spaulding (Ed.), Nebraska Symposium on Motivation: Integrated views of motivation and emotion (Vol. 41, pp. 181–238). Lincoln: University of Nebraska Press.
60. Gottesman, I. I. (1991). Schizophrenia genesis. New York: W. H. Freeman.
61. Gottesman, I. I., & Bertelsen, A. (1989). Confirming unexpressed genotypes for schizophrenia. Archives of General Psychiatry, 46, 867–872.
62. Gottesman, I. I., & McGue, M. (1991). Mixed and mixed-up models for the transmission of schizophrenia. In D. Cicchetti & W. M. Grove (Eds.), Thinking clearly about psychology: Vol. 2. Personality and psychopathology (Essays in honor of Paul E. Meehl) (pp. 295–312). Minneapolis: University of Minnesota Press.
63. Gottesman, I. I., McGuffin, P., & Farmer, A. E. (1987). Clinical genetics as clues to the “real” genetics of schizophrenia (A decade of modest gains while playing for time). Schizophrenia Bulletin, 13, 23–47.
64. Gottesman, I. I., & Moldin, S. O. (1998). Genotypes, genes, genesis, and pathogenesis in schizophrenia. In M. F. Lenzenweger & R. H. Dworkin (Eds.), Origins and development of schizophrenia (pp.5–26).Washington,DC:AmericanPsychologicalAssociation.
65. Gottesman, I. I., & Shields, J. A. (1966). Schizophrenia in twins: Sixteen years’ consecutive admissions to a psychiatric clinic. British Journal of Psychiatry, 112, 809–818.
66. Gottesman, I. I., & Shields, J.A. (1967).Apolygenic theory of schizophrenia. Proceedings of the National Academy of Sciences, 58, 199–205.
67. Gottesman, I. I., & Shields, J. A. (1972). Schizophrenia and genetics: A twin study vantage point. New York: Academic Press.
68. Gottesman, I. I., & Shields, J. A. (1976). A critical review of recent adoption, twin, and family studies of schizophrenia: Behavioral genetics perspective. Schizophrenia Bulletin, 2, 360–401.
69. Gottesman, I. I., & Shields, J. A. (1982). Schizophrenia: The epigenetic puzzle. New York: Cambridge University Press.
70. Gray, J. A. (1982). The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system. Oxford, England: Oxford University Press.
71. Gray, J. A. (1987). The psychology of fear and stress (2nd ed.). Cambridge, England: Cambridge University Press.
72. Gray, J. A., & McNaughton, N. (1996). The neuropsychology of anxiety: Reprise. In D. A. Hope (Ed.), Nebraska Symposium on Motivation (Vol. 43, pp. 61–134). Lincoln: University of Nebraska Press.
73. Gray, J. A., & McNaughton, N. (2000). The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system (2nd ed.). Oxford, England: Oxford University Press.
74. Grossman, L. S., Harrow, M., Goldberg, J. F., & Fichtner, C. G. (1991). Outcome of schizoaffective disorder at two long-term follow-ups:Comparisonswithoutcomeofschizophreniaandaffective disorders. American Journal of Psychiatry, 148, 1359–1365.
75. Harrow, M., Yonan, C. A., Sands, J. R., & Marengo, J. (1994). Depression in schizophrenia: Are neuroleptics, akinesia, or anhedonia involved? Schizophrenia Bulletin, 20, 327–338.
76. Heston, L. L., & Denney, D. (1968). Interactions between early life experience and biological factors in schizophrenia. In D. Rosenthal & S. S. Kety (Eds.), The transmission of schizophrenia. New York: Pergamon Press.
77. Hooley, J. M., & Hiller, J. B. (1998). Expressed emotion and the pathogenesis of relapse in schizophrenia. In M. F. Lenzenweger & R. H. Dworkin (Eds.), Origins and development of schizophrenia: Advances in experimental psychopathology (pp. 447–468). Washington, DC:American Psychological Association.
78. Iacono, W. G. (1998). Identifying psychophysiological risk for psychopathology: Examples from substance abuse and schizophrenia research. Psychophysiology, 35, 621–637.
79. Johnstone, E. C. (1999a). Diagnostic issues: Concepts of the disorder. In E. Johnstone, M. Humphreys, F. Lang, S. Lawrie, & R. Sandler (Eds.), Schizophrenia: Concepts and clinical management (pp. 20–43). Cambridge, England: Cambridge University Press.
80. Johnstone, E. C. (1999b). Epidemiology and genetics. In E. Johnstone, M. Humphreys, F. Lang, S. Lawrie, & R. Sandler (Eds.), Schizophrenia: Concepts and clinical management (pp. 145– 160). Cambridge, England: Cambridge University Press.
81. Johnstone, E. C. (1999c). Introduction. In E. Johnstone, M. Humphreys, F. Lang, S. Lawrie, & R. Sandler (Eds.), Schizophrenia: Concepts and clinical management (pp. 1–19). Cambridge, England: Cambridge University Press.
82. Kapur, S., & Remington, G. (1996). Serotonin-dopamine interaction and its relevance to schizophrenia. American Journal of Psychiatry, 153, 466–476.
83. Kendell, R. E. (1982).The choice of diagnostic criteria for biological research. Archives of General Psychiatry, 39, 1334–1339.
84. Kendler, K. S. (1999). Molecular genetics of schizophrenia. In D. S. Charney, E. J. Nestler, & B. S. Bunney (Eds.), Neurobiology of mental illness (pp. 203–213). Oxford, England: Oxford University Press.
85. Kendler, K. S., & Diehl, S. R. (1993). The genetics of schizophrenia. Schizophrenia Bulletin, 19, 261–285.
86. Kendler, K. S., & Diehl, S. R. (1995). Schizophrenia: Genetics. In H. I. Kaplan & B. Sadock (Eds.), Comprehensive textbook of psychiatry (6th ed., pp. 942–957). Baltimore: Williams & Wilkins.
87. Kendler, K. S., Gruenberg, A. M., & Kinney, D., K. (1994). Independent diagnoses of adoptees and relatives as defined by DSMIII in the Provincial and National samples of the Danish Adoption Study of Schizophrenia. Archives of General Psychiatry, 51, 456–468.
88. Kendler,K.S.,Gruenberg,A.M.,&Tsuang,M.T.(1986).ADSM-III family study of the non-schizophrenic psychotic disorders. American Journal of Psychiatry, 143, 1098–1105.
89. Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. (1992). Major depression and generalized anxiety disorder: Same genes, (partly) different environments? Archives of General Psychiatry, 49, 716–722.
90. Kety, S. S., Rosenthal, D., Wender, P. H., & Schulsinger, F. (1968). The types and prevalence of mental illness in the biological and adoptive families of adopted schizophrenics. In D. Rosenthal & S. S. Kety (Eds.), The transmission of schizophrenia (pp. 345–362). Oxford, England: Pergamon.
91. Kety, S., Rosenthal, D., Wender, P. H., Schulsinger, F., & Jacobsen, B. (1978). The biological and adoptive families of adopted individual who became schizophrenic: Prevalence of mental illness and other characteristics. In L. C. Wynne, R. L. Cromwell, & S. Matthysse (Eds.), The nature of schizophrenia (pp. 25–37). New York:Wiley.
92. Kety, S. S., Wender, P. H., Jacobsen, B., Ingraham, L. J., Jansson, L., Faber, B., & Kinney, D. K. (1994). Mental illness in the biological and adoptive relatives of schizophrenic adoptees: Replication of the Copenhagen study in the rest of Denmark. Archives of General Psychiatry, 51, 442–455.
93. Kirkpatrick, B., Buchanan, R. W., Ross, D. E., & Carpenter, W. T., Jr. (2001). A separate disease within the syndrome of schizophrenia. Archives of General Psychiatry, 58, 165–171.
94. Krystal, J. H., Abi-Dargham, A., Laruelle, M., & Moghaddam, B. (1999). Pharmacologic models of psychoses. In D. S. Charney, E. J. Nestler, & B. S. Bunney (Eds.), Neurobiology of mental illness (pp. 214–224). Oxford, England: Oxford University
96. Leff, J. P. (1976). Schizophrenia and sensitivity to the family environment. Schizophrenia Bulletin, 2, 566–574.
97. Leff, J. P., &Vaughn, C. E. (1985). Expressed emotion in families: Its significance for mental illness. NewYork: Guilford Press.
98. Leshner, A. I. (1997). Addiction is a brain disease, and it matters. Science, 278, 45–47.
99. Levinson,D.F.,&Levitt,M.E.M.(1987).Schizoaffectivemaniareconsidered. The American Journal of Psychiatry, 144, 415–425.
100. Lewine, R. J. (1985). Negative symptoms in schizophrenia: Editor’s introduction. Schizophrenia Bulletin, 11, 361–363.
101. Liddle, P. F. (1987). The symptoms of chronic schizophrenia: A reexamination of the positive-negative dichotomy. British Journal of Psychiatry, 151, 145–151.
102. Liddle, P., Carpenter, W. T., & Crow, T. (1993). Syndromes of schizophrenia: Classic literature [Editorial]. British Journal of Psychiatry, 165, 721–727.
103. Linszen, D. H., Dingemans, P. M., Nugter, M. A., Van der Does, A. J. W., Scholte, W. F., & Lenior, M. A. (1997). Patient attributes and expressed emotion as risk factors for psychotic relapse. Schizophrenia Bulletin, 23, 119–130.
104. Losonczy, M. F., Davidson, M., & Davis, K. L. (1987). The dopamine hypothesis of schizophrenia. In H. Y. Meltzer (Ed.), Psychopharmacology: The third generation of progress (pp. 715–726). New York: Raven Press.
105. Maier, W., Lichtermann, D., Minges, J., Hallmayer, J., Heun, R., Benkert, O., & Levinson, D. F. (1993). Continuity and discontinuity of affective disorders and schizophrenia: Results of a controlled family study. Archives of General Psychiatry, 50, 871–883.
106. McGue, M., & Gottesman, I. I. (1989). Genetic linkage in schizophrenia: Perspectives from genetic epidemiology. Schizophrenia Bulletin, 15, 453–464.
107. McGuffin, P., Farmer, A. E., Gottesman, I. I., Murray, R. M., & Reveley, A. M. (1984). The concordance of operationally defined schizophrenia: Confirmation of familiality and heritability. Archives of General Psychiatry, 41, 541–545.
108. McNaughton, N., & Gray, J. A. (2000). Anxiolytic action on the behaviouralinhibitionsystemimpliesmultipletypesofarousalcontribute to anxiety. Journal of Affective Disorders, 61, 161–176.
109. Meehl, P. E. (1962). Schizotaxia, schizotypy, schizophrenia. American Psychologist, 17, 827–838.
110. Meinecke, D. L. (2001). The developmental etiology of schizophrenia hypothesis: What is the evidence? Editor’s introduction. Schizophrenia Bulletin, 27, 335–336.
111. Meltzer, H. Y. (1984). Schizoaffective disorder; Editor’s introduction. Schizophrenia Bulletin, 10, 11–13.
112. Mineka, S., Watson, D., & Clark, L. A. (1998). Comorbidity of anxiety and unipolar mood disorders. Annual Review of Psychology, 49, 377–412.
113. Mirsky, A. F., Bieliauskas, L. A., French, L. M., Van Kammen, D. P., Jönsson, E., & Sedvall, G. (2000). A 39-year follow-up of the Genain quadruplets. Schizophrenia Bulletin, 26, 699–708.
114. Moldin, S. O., & Gottesman, I. I. (1997). At issue: Genes, experience, and chance in schizophrenia—positioning for the 21st century. Schizophrenia Bulletin, 23, 547–561.
115. Morey, L. C. (1991). Classification of mental disorder as a collection of hypothetical constructs. Journal of Abnormal Psychology, 100, 289–293.
116. Murray, R. M., Lewis, S., & Reveley, A. M. (1985). Towards an aetiological classification of schizophrenia. Lancet, 1, 1023–1026.
117. Neale, J. M., & Oltmanns, T. F. (1980). Schizophrenia. New York: Wiley.
118. Nuechterlein, K. H., Snyder, K. S., & Mintz, J. (1992). Paths to relapse: Possible transactional processes connecting patient illness onset, expressed emotion, and psychotic relapse. British Journal of Psychiatry, 161, 88–96.
119. Plomin, R. (1990). The role of inheritance in behavior. Science, 248, 183–188.
120. Plomin, R., DeFries, J. C., McLearn, G. E., & Rutter, M. (1997). Behavioral genetics (3rd ed). New York: W. H. Freeman.
121. Procci, W. R. (1989). Psychotic disorders not elsewhere classified. InH.Kaplan&B.Sadock(Eds.),Comprehensivetextbookofpsychiatry, (Vol. 5, pp. 830–842). Baltimore: Williams & Williams.
122. Reynolds, G. P. (1989). Beyond the dopamine hypothesis: The neurochemical pathology of schizophrenia. British Journal of Psychiatry, 155, 305–316.
123. Rosenthal, D. (Ed.). (1963). The Genain quadruplets: A case study and theoretical analysis of heredity and environment in schizophrenia. New York: Basic Books.
124. Rosenthal, D. (1970). Genetic theory and abnormal behavior. New York: McGraw-Hill.
125. Rutter, M. L. (1997). Nature-nurture integration. The example of antisocial behavior. American Psychologist, 52, 390–398.
126. Sands, J. R., & Harrow, M. (1999). Depression during the longitudinalcourseofschizophrenia.SchizophreniaBulletin,25,157–171.
127. Siris, S. G., & Lavin, M. R. (1995). Schizoaffective disorder, schizophreniform disorder, and brief reactive psychotic disorder. In H. I. Kaplan & B. Sadock (Eds.), Comprehensive textbook of psychiatry (6th ed., pp. 1019–1031). Baltimore:Williams &Wilkins.
128. Snyder, S. H. (1978). Dopamine and schizophrenia. In L. C. Wynne, R. L. Cromwell, & S. Matthysse (Eds.), The nature of schizophrenia (pp. 87–94). New York: Wiley.
129. Sommers, A. A. (1985). “Negative symptoms”: Conceptual and methodological problems. Schizophrenia Bulletin, 11, 364–379.
130. Spitzer, R. L., Endicott, J., & Robins, E. (1978). Research diagnostic criteria: Rationale and reliability. Archives of General Psychiatry, 35, 773–789.
131. Strauss, J., Carpenter, W. T., Jr., & Bartko, J. (1974). The diagnosis and understanding of schizophrenia: Part III. Speculations on the processes that underlie schizophrenic symptoms and signs [Experimental issue]. Schizophrenia Bulletin, 1, 61–69.
132. Strauss,J.,&Gift,T.E.(1977).Choosinganapproachfordiagnosing schizophrenia. Archives of General Psychology, 34, 1248–1253.
133. Strauss, M. (2001). Demonstrating specific cognitive deficits:Apsychometric perspective. Journal of Abnormal Psychology, 110, 6–14.
134. Taylor, M. A. (1992). Are schizophrenia and affective disorder related? Aselective review of the literature. American Journal of Psychiatry, 149, 22–32.
135. Taylor, M. A., Berenbaum, S. A., Jampala, V. C., & Cloninger, C. R. (1993). Are schizophrenia and affective disorder related? Preliminary data from a family study. American Journal of Psychiatry, 150, 278–285.
136. Tellegen, A. (1985). Structures of mood and personality and their relevance to assessing anxiety, with an emphasis on self-report. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 681–706). Hillsdale, NJ: Erlbaum.
137. Tienari, P. (1991). Interaction between genetic vulnerability and family environment: The Finnish adoptive family study of schizophrenia. Acta Psychiatrica Scandinavica, 84, 460–465.
138. Tsuang, M. T., & Faraone, S. V. (1995). The case for heterogeneity in the etiology of schizophrenia. Schizophrenia Research, 17,
139. 161–175.
140. Tsuang, M. T., & Faraone, S. V. (1997). Schizophrenia: The facts (2nd ed.). New York: Oxford University Press.
141. Vaughn, C., & Leff, J. (1976). The influence of family and social factors on the course of psychiatric illness. British Journal of Psychiatry, 129, 125–137.
142. Venables,P.H.(1957).Ashortscaleforrating“Activity-Withdrawal” in schizophrenics. Journal of Mental Science, 103, 197–199.
143. Venables, P. H. (1963a). The relationship between level of skin potential and fusion of paired light flashes in schizophrenic and normal subjects. Journal of Psychiatric Research, 1, 279–287.
144. Venables, P. H. (1963b). Selectivity of attention, withdrawal, and cortical activation. Archives of General Psychiatry, 9, 74–78.
145. Venables,P.H.(1967).Therelationoftwoflashandtwoclickthresholds to withdrawal in paranoid and non-paranoid schizophrenics. British Journal of Social and Clinical Psychology, 6, 60–62.
146. Venables, P. H., & Wing, J. K. (1962). Level of arousal and the subclassification of schizophrenia. Archives of General Psychiatry, 7, 114–119.
147. Ventura, J., Nuechterlein, K. H., Lukoff, D., & Hardesty, J. P. (1989). A prospective study of stressful life events and schizophrenic relapse. Journal of Abnormal Psychology, 98, 407–411.
148. Walker, E., & Diforio, D. (1997). Schizophrenia: Aneural diathesisstress model. Psychological Review, 104, 667–685.
149. Waltrip, R. W., Buchanan, R. W., Carpenter, W. T., Kirkpatrick, B., Summerfelt, A., Breier, A., Rubin, S. A., & Carbone, K. M. (1997). Borna disease virus antibodies and the deficit syndrome of schizophrenia. Schizophrenia Research, 23, 253–257.
150. Weinberger, D. R. (1987). Implications of normal brain development for the pathogenesis of schizophrenia. Archives of General Psychiatry, 44, 660–669.
151. Weinberger, D. R., & Lipska, B. K. (1995). Cortical maldevelopment, anti-psychotic drugs, and schizophrenia: Asearch for common ground. Schizophrenia Research, 16, 87–110.
152. Williams, P. V., & McGlashan, T. H. (1987). Schizoaffective psychosis. Archives of General Psychiatry, 44, 130–137.
153. Wing, J. K. (1978). Social influences on the course of schizophrenia. In L. C. Wynne, R. L. Cromwell, & S. Matthysse (Eds.), The nature of schizophrenia (pp. 599–616). New York: Wiley.
154. Wing, J. K., & Brown, G. W. (1970). Institutionalism and schizophrenia. London: Cambridge University Press.
155. Wise, R. A., & Bozarth, M. A. (1987). Apsychomotor stimulant theory of addiction. Psychological Review, 94, 469–492.
156. Wise, R. A., & Rompre, P.-P. (1989). Brain dopamine and reward. Annual Review of Psychology, 40, 191–225.
157. Zubin, J., & Spring, B. (1977). Vulnerability—Anew view of schizophrenia. Journal of Abnormal Psychology, 86, 103–126.