Negative v Positive Schizophrenia

Definition and Validation

Nancy C. Andreasen, MD, PhD, Scott Olsen, PhD

\s=b\ We developed criteria for dividing the schizophrenic syn- none of which has emerged as preeminent. One
Systems,
drome into three subtypes: positive, negative, and mixed schizo- set of approaches has emphasized cross-sectional phenom-
phrenia. Positive schizophrenia is characterized by prominent enology, such as the traditional Kraepelinian-Bleulerian
delusions, hallucinations, positive formal thought disorder, and division into hebephrenic, catatonic, paranoid, and simple,
persistently bizarre behavior; negative schizophrenia, by affec- which still forms the basis for the DSM-III."
tive flattening, alogia, avolition, anhedonia, and attentional Other cross-sectional subdivisions have included group-
impairment. In mixed schizophrenia either both negative and ings such as schizoaffective v nonaffective5 and paranoid v
positive symptoms are prominent, or neither is prominent. We nonparanoid.6 Another set of subtypes has emphasized
explored the validity of these criteria in a variety of ways. longitudinal course, leading to subtypes such as acute v
Significant differences between the three types were noted chronic,7 process v reactive,8 or good v poor prognosis.9
using external validators such as premorbid adjustment, indices Surprisingly few recent attempts have been made to relate
of cognitive dysfunction, ventricular brain ratio, and course in classification either to functional brain systems, such as
hospital. The correlational structure of the symptom complexes language or auditory perception, or to other possible
also provided further support for our approach to subtyping. etiologic constructs. None of the classification systems
(Arch Gen Psychiatry 1982;39:789-794) currently available enjoys widespread acceptance, because
none has well-documented predictive power for estimating
outcome or facilitating the search for causes.

Most clinicians and investigators agree that the group

of disorders called schizophrenia is heterogeneous.
Although Kraepelin is sometimes said to represent a
During recent years, however, a consensus has emerged
that investigators should limit the concept of schizophre-
nia to relatively chronic forms of the disorder and to forms
"unitary position," in Dementia Praecox and Paraphrenia that lack prominent affective symptoms. That is, acute,
he indicated quite clearly his belief that the disorder good-prognosis, latent, and schizoaffective schizophrenia
should be divided into subtypes and that the various are no longer pooled with other types in most research
subtypes might reflect different cerebral localizations in studies, and they are also seen as requiring different
areas such as the frontal or temporal lobes.1 Bleuler clinical management. Although this narrowing of the
emphasized the importance of subtypes by subtitling his concept of schizophrenia is still somewhat controversial, it
book "the group of schizophrenias."2 The Kleist-Leonhard will undoubtedly facilitate research on biologic correlates
school has proposed an elaborate system for subtyping by reducing the variability within samples studied.10 Nev-
schizophrenia that is also based on different cerebral ertheless, even within the chronic schizophrenias, consid-
localizations.3 erable variability clearly remains, together with a sense
American psychiatry has also emphasized the impor- that chronic schizophrenia (as defined by a duration of
tance of subtyping, although there has been little consen- longer than six months) is still a heterogeneous group of
sus about the best system. As a consequence, the history of disorders. Much like collagen-vascular disease, it is proba-
the nosology of schizophrenia is one of competing nosologie bly a group of related disorders that vary in their manifes-
tations depending on the neurochemical or functional
brain system being affected. As long as these different
disorders are pooled in research studies, the search for
Accepted for publication March 3, 1982.
From the Department of Psychiatry, University of Iowa, Iowa City. biologic correlates, markers, or etiologic factors is likely to
Reprint requests to Department of Psychiatry, University of Iowa, Iowa remain inconclusive.
City, IA 52242 (Dr Andreasen). Clearly a fresh approach to the subtyping of schizophre-

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Auckland User on 06/04/2015

 nia is needed. One such approach that warrants further the inclusion of additional cases of "negative schizophre-
investigation is a division based on whether the symptoms nia" in the category.
of the disorder are predominantly positive (or florid) or
whether they are predominantly negative (or defect). This Positive Schizophrenia
distinction, originally proposed by the neurologist Hugh- 1. At least one of the following is a prominent part of
lings-Jackson," has recently been revived by Strauss et the illness.
al,12 Crow,13 Andreasen et al,1415 and Angrist et al.16 The a. Severe hallucinations that dominate the clinical
distinction has led to the hypothesis that patients with picture (auditory, haptic, or olfactory) (The judgment
prominent positive symptoms (delusions, hallucinations, of severity should be based on various factors such as
positive formal thought disorder, or bizarre behavior) are persistence, frequency, and effect on lifestyle.)
likely to differ in a variety of important ways from b. Severe delusions (may be persecutory, jealous, somat-
patients who have prominent negative symptoms or a ic, religious, grandiose, or fantastic) (The judgment of
defect state (alogia, affective flattening, avolition, anhe- frequency should be made as described for severity.)
donia-asociality, and attentional impairment). c. Marked positive formal thought disorder (manifested
This subdivision, which is reminiscent of Bleuler's dis- by marked incoherence, derailment, tangentiality, or
tinction between fundamental and accessory symptoms, is illogicality)
one that clinicians who treat large numbers of schizo- d. Repeated instances of bizarre or disorganized behav-
phrenic patients have recognized for some time. Although ior
positive symptoms tend to improve with aggressive treat- 2. None of the following is present to a marked
ment, negative symptoms tend to be more refractory and degree.
ultimately more crippling. a. Alogia
Recently, investigators have become increasingly inter- b. Affective flattening
ested in studying this distinction in an empirical and c. Avolition-apathy
data-based manner. It has been hypothesized that negative d. Anhedonia-asociality
symptoms define one end of a continuum of disorders and e. Attentional impairment
are to be correlated with poor premorbid adjust-
likely
ment, poor response to neuroleptic therapy, a chronic Negative Schizophrenia
course and poor outcome, cognitive impairment, and a 1. At least two of the following are present to a marked
different underlying pathologic process, such as atrophie degree.
changes in the brain." On the other hand, positive symp- a. Alogia (eg, marked poverty of speech, poverty of
toms are likely to be correlated with a better premorbid content of speech)
adjustment, better response to neuroleptic therapy, a less b. Affective flattening
malignant course, a normal sensorium, and an underlying c. Anhedonia-asociality (eg, inability to experience plea-
pathologic process that is predominantly neurochemical. sure or to feel intimacy, few social contacts)
This distinction remains largely untested. Most of its d. Avolition-apathy (eg, anergia, impersistence at work
advocates are committed to it primarily because of its or school)
heuristic and hypothesis-generating value. Exploration of e. Attentional impairment
the distinction has proceeded slowly and fitfully in spite of 2. None of the following dominates the clinical picture
considerable interest in it, primarily because adequate or is present to a marked degree.
methods of phénoménologie description and nosologie cat- a. Hallucinations
egorization have not been available. In a previous report, b. Delusions
we described a Scale for the Assessment of Negative c. Positive formal thought disorder
Symptoms (SANS),17 which will permit investigators to d. Bizarre behavior
assess negative symptoms reliably. We report herein on a
set of diagnostic criteria that may be used to subdivide Mixed Schizophrenia
schizophrenic patients into three groups based on the This category includes those patients that do not meet
nature of their current symptoms: positive (or florid), criteria for either positive or negative schizophrenia, or
negative (or defect), and mixed schizophrenia. We have meet criteria for both.
examined the predictive and mathematical validity of
these criteria and report some interesting findings that Basis for Criteria
suggest that they may be useful for other investigators. Our criteria are based on the SANS,1718 the Scale for the
DIAGNOSTIC CRITERIA
Assessment of Thought, Language, and Communication
(TLC),19 to evaluate positive formal thought disorder, and a
The following diagnostic criteria provide a simple, logi- modified version of the Schedule for Affective Disorders
cal, and coherent method for classifying patients cross- and Schizophrenia (SADS),20 to develop a global rating of
sectionally as positive, negative, or mixed schizophrenics. hallucinations, delusions, and bizarre behavior. These
These criteria were written to characterize the nature of symptoms were considered to be present to a prominent or
the illness during the index evaluation or admission. marked degree when a rating of at least 4 was given on a
Because these are criteria for subtyping schizophrenia, all scale of 0 to 5.
patients must first meet DSM-III criteria for schizophre- Several aspects of these criteria should be noted. First,
nia. Based on our own research experience, however, we they are designed to evaluate patients cross-sectionally
suggest that DSM-III criterion A-6 be slightly modified to during the index episodes. In our own use of the criteria,
include marked poverty of speech in addition to the other we evaluated the severity of symptoms during the preced-
manifestations of disorganized language and cognition, ing month. The most prominent symptoms may fluctuate
such as incoherence or poverty of content of speech. This in some patients over the course of months or years. These
modification recognizes an important linguistic-cognitive criteria do not take such fluctuations into account and
symptom, as well as a negative symptom, and may permit instead attempt to characterize only current symptoms.

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Auckland User on 06/04/2015

 Table 1. Sociodemographic Characteristics
—

Type of Schizophrenia
,-A-, Duncan's
Negative (N) Mixed (M) Positive (P) Follow-up Test
_Variable_(n 16)_(n 18)_(n 18)_P> = = =
F_(a .05) =

Age, yr (mean ± SD)_34.37 ± 14.49_27.28 ± 7.73_28.72 ± 8.09_.1241_

Education, yr (mean ± SD)_11.06 ± 1.48_12.05 ± 2.44_13.55 ± 2.04_.0032_P > M, N
Premorbid adjustment, Phillips
SD)_8.40 ± 3.58_5.50 ± 3.20_5.05 ± 1.95_.0047_N > M, P
scale (mean ±
Female, %_50_44_50_.9336'_
Single. %_63 _67_78_.6175'_-_^^_^
Employed, % 6 41 55 .0066' P, M > N
*
F statistics result from a normal approximation to the binomial distribution.

Table 2. Previous Course of Illness (Mean ± SD)

—

Type of Schizophrenia
,-~-. Duncan's
Negative Mixed Positive Follow-up Test
_Variable'_(n =
16)_(n 18)_(n 18)_P> = =
F_(a .05) =

No. of hospitalizations_6.75 ± 6.14_3.94 ± 3.09_4.50 ± 4.40_.1922_._J_._

Duration ot hospitalizations, mo_22.73 ± 37.50_6.75 ± 6.12_8.50 ± 8.38_.0994_._J_1_
Age at onset, yr_23.00 ±
8.02_21.22 ± 4.95_21.88 ± 5.90_.7219_i^^_^
Months since onset 72.00 ± 30.93 53.05 ± 35.42 54.83 ± 36.74 .2450

'Some observations are missing for some of the variables.

We working on an interview and criteria that will

are
Table 3. Previous Treatment and Substance Abuse
characterize patients longitudinally as well. —

Second, our application of the criteria is based on the use Type of Schizophrenia Duncan's
of a systematic structured interview that evaluates the ,-. Follow-up
various positive and negative symptoms in terms of fre- Negative (N) Mixed Positive (P) Test
Variable (n 16) =
(n 18) (n 18) P> F' (a .05)
= = =

quency, persistence, severity, and impact upon the ECT.t %_56_27_5 .0037 N > P
patient's lifestyle. To work well, these criteria should be Neuroleptic
based on a thorough interview and a clear sense of what therapy,
constitutes sufficient severity for the symptom to be %_93_83_72 .2953
considered present at the criterion level. Most schizo- Drug abuse, _

phrenic patients have many of these symptoms to a mild %_6_33_33 .1186

degree. The distinction between the positive and negative Alcohol _

abuse, % .6444
syndrome turns on the severity of individual symptoms. 0 5 5

*F statistics result from a normal approximation to the binomial distribu-

CLINICAL VALIDATION OF THE CRITERIA tion.
We evaluated a consecutive sample of 52 patients who were tECT indicates electroconvulsive therapy.
admitted to Iowa Psychiatric Hospital (Iowa City) and who met
DSM-III criteria for schizophrenia. The total data base on these
patients consisted of a modified SADS interview, a preliminary
version of the SANS, the Phillips Premorbid Adjustment Scale,21 found that 16 met criteria for negative schizophrenia, 18 for mixed
the Personal History for Demographic Data,22 the TLC scale, and schizophrenia, and 18 for positive schizophrenia. We hypothesized
additional historic data describing past treatment and course of that if the criteria had clinical validity, these three groups would
illness. Cognitive function was assessed according to the Mini differ on a number of dependent variables usually considered to be
Mental Status scores.23 Computerized tomographic (CT) scans of external validators. Patients with negative schizophrenia were
the head were obtained using a head scanner (EMI model 1005, hypothesized to have more frequent family histories of schizo-
Mark I). Methods for measuring ventricular-brain ratio (VBR) phrenia, poorer premorbid adjustment, larger VBRs, impairment
have been described previously.24 Family history was assessed of the sensoria as assessed by Mini Mental Status scores, and
using the Family History Research Diagnostic Criteria.25 relatively poor responses to treatment.
Our patients were relatively young, with a mean (±SD) age of The sociodemographic characteristics of the three groups are
29.96 ± 10.61 years. The average age at onset of illness was given in Table 1. Although the patients with negative schizophre-
21.96 ± 6.22 years, and they had been hospitalized a mean of nia were somewhat older, the difference was not statistically
5 ± 4.71 times. The mean duration of hospitalization was 12.44 ± significant (a .05). There were also no differences between the
=

22.68 months. Alcohol abuse was quite uncommon, occurring in three groups in sex ratio or in marital status.
only 4% of the sample. A larger number had histories of drug On the other hand, the three groups differed in important ways
abuse (25% ), but the abuse had not been sufficiently severe to that are consistent with the nature of the three different syn-
account for the diagnosis of schizophrenia. Eighty-two percent dromes. Patients with negative schizophrenia had the least educa-
had been treated with neuroleptics at some time in the past, and tion (11.06 ± 1.48 years), which suggests that the typical negative
29% with electroconvulsive therapy (ECT). schizophrenic is unable to complete high school. On the other
Using the criteria described to classify these 52 patients, we hand, the positive schizophrenics had a mean educational level of

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Auckland User on 06/04/2015

 Table 4.—Indices of Cognitive Dysfunction and Outcome
Type of Schizophrenia
,- -, Duncan's
Negative Mixed Positive Follow-up Test
_Variable'_(n 16)_(n 18)_(n 18)_P> = = =
F_(a = .05)
VBR, mean ± SD_8.48 ± 4.55_5.20 ± 2.93_4.59 ± 3.24_.0063_N > M, P
Mini Mental Status, mean ± SD_20.60 ±
8.96_26.47 ± 3.00_28.67 ± 1.88_.0011_P, M > N
GAS (admission), mean ± SD_20.31 ±
6.04_27.25 ± 6.04_26.61 ± 5.68_.0026_P, M > N
GAS (discharge), mean ± SD_28.00 ±
10.62_37.50 ± 12.30_33.33 ± 7.30_.0381_M > N
Right-handedness 67 83 100 .0403t P > N
*Some observations are missing for some of the variables. VBR indicates ventricular-brain ratio; GAS, global assessment scale.
fF statistics result from a normal approximation to the binomial distribution.

Table 5.—Correlations Between Positive and Negative Symptoms

Positive
Formal Catatonic
Affective Attentional Hallu- Thought Bizarre Motor
_Flattening Alogia Avolition Anhedonia Impairment cinations Delusions Disorder Behavior Behavior
Affective flattening 1.000
Alogia_.632_1.000 _„_._._^____„_._„_._._^_
Avolition_.495_.673 1.000_._L._._^__._^__._l_._„j_.___
Anhedonia_.442_.654_.836_1300_„_._._^___._._^_._L1_u__
Attentional .146 .560 .560 .532 1.000
impairment_
Hallucinations_-.311_-.410 -.469_-.286_-.363_1.000_._^_._^_._^_
Delusions_-.299 -.688 -.502_-.418_-^562 .470_1.000__J_._l^__
Positive formal -.080 -.323 -.027 -.124 -.014 .156 .200 1.000
thought
disorder_
Bizarre behavior .167 .049 .176 .107 -.027 .086 .045 .188 1.000
Catatonic motor .195 .446 .211 .261 .562 -.083 -.256 -.155 -.052 1.000
behavior

13.55 ± 2.04 years, which is equivalent to a year and a half of Table 4, which summarizes various indices of cognitive dysfunc-
college. tion and outcome. The three patient groups differed significantly
The three groups also differed on premorbid adjustment as on all these variables. The patients with negative schizophrenia
assessed by the Phillips scale. On this scale, a score of 12 had significantly larger VBRs than did the patients with mixed or
represents the worst possible premorbid adjustment, and a score florid schizophrenia. They also had a significantly lower score on
of 0 the best. The negative schizophrenics had significantly poorer Mini Mental Status. Taken together, these findings suggest that
premorbid adjustment than did the patients with mixed or patients with negative schizophrenia may have an underlying
positive schizophrenia. pathologic process involving cerebral atrophy that is reflected by
The three groups also differed significantly in employment rate. such measures of the sensorium as orientation, ability to calcu-
Only 6% of the patients with negative schizophrenia were late, and memory functions. Possibly related is the fact that the
employed, compared with 41% of the patients with mixed schizo- three groups also differed significantly in the rate of right-
phrenia and 55% of those with positive schizophrenia. handedness. All the patients with positive schizophrenia were
The past course of illness was compared in all three groups, and right-handed, compared with only 69% of those with negative
no significant differences were noted. These data are summarized schizophrenia. As some research suggests that left-handedness in
in Table 2. The patients with negative schizophrenia tended to some patients results from head injury occurring early in life,26
have a somewhat longer total duration of past hospitalization and this finding is also consistent with the larger VBRs and lower
a somewhat larger number of past hospitalizations. No differences Mini Mental Status scores in the patients with negative schizo-
were noted among the three groups in age at onset. phrenia.
Data concerning prior treatment and previous substance abuse The data concerning scores on the global assessment scale
are summarized in Table 3. No significant differences were noted (GAS) at admission and discharge should be evaluated cautiously.
in rate of substance abuse, but there was a tendency for more The GAS is a 100-point scale on which high scores indicate good
abuse to occur in the patients with mixed and florid schizophrenia health. The GAS rating on admission reflects the overall severity
than in those with negative schizophrenia. There was also a of illness at the time of initial index evaluation. The patients with
tendency for the patients with negative schizophrenia to have negative schizophrenia had significantly lower GAS ratings on
received more treatments of all types, but the only significant admission than did the patients with positive schizophrenia. The
difference was in the rate of ECT. Fifty-six percent of the patients rating at discharge might be considered to represent the level of
with negative schizophrenia had received ECT v 27% of those with improvement achieved, an indirect index of response to treatment.
mixed schizophrenia and only 5% of those with positive schizo- As Table 4 indicates, the patients with negative schizophrenia had
phrenia (P < .004). This frequent use of ECT in the patients with significantly lower GAS scores at the time of discharge than did
negative schizophrenia probably represents an aggressive but the patients with mixed schizophrenia, and their scores continued
obviously unsuccessful attempt to eradicate their negative symp- to be lower than those of both the mixed and the positive
toms. schizophrenics. However, because no control was exerted over
Data concerning major external validators are presented in treatment, our results indicate response to treatment only very

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Auckland User on 06/04/2015

 Table 6.—Factor Analysis of Positive and Negative Symptoms
opposite dimensions of specific forms of behavioral activation.
Factor 3 was weighted positively on hallucinations and negatively
Factor on positive formal thought disorder, and factor 4 had its heaviest
loading on catatonic motor behavior. The presence of the large
_1_2_3_4 bipolar general factor, which accounts for 42% of the variance,
Affective provides additional support for the independence of the positive
flattening_.60898 .27128 .37747 -.25254 and negative symptom syndromes.
Alogia_.90155 -.11166 .19401 -.09262
COMMENT
Avolition_.84937 .28705 .03470 -.02729
Anhedonia_.80090 .15119 .16289 .04975 Our findings provide support from two different per-
Attentional spectives for the subtyping of schizophrenia into positive
impairment_.73903 -.18304 -.33719 .40279 and negative. The first approach involved the a priori
Hallucinations_-.55927 -.21871 .53879 .37766 definition of criteria for positive, negative, and mixed
Delusions_-.74211 .16852 .17909 .18587
schizophrenia. These three groups differed from one
Positive formal another on a number of dependent variables that serve as
thought external validators. As hypothesized, patients with nega-
disorder_-.18463 .66728 -.56326 .26039
tive schizophrenia had poor premorbid adjustment, a
Bizarre
behavior_.06790 .66301 .41414 .40609
Catatonic motor
lower overall level of functioning as measured by the GAS,
behavior_.49033 -.43488 -.02734 .59675 impaired cognitive function, and indications of previous
Variance brain injury and cerebral atrophy. On the other hand,
explained,* 42.3 13.7 11.3 10.0 patients with positive schizophrenia had better premorbid
adjustment, better overall levels of functioning, normal
sensoria, and no evidence of cerebral atrophy. The mixed
group consistently occupied a middle ground between these
indirectly. Nevertheless, they do seem to indicate that, whatever two extremes. Thus the distinction does appear to have
its cause, negative schizophrenia leads to a more severe syndrome
than does positive schizophrenia, both at the time of admission some predictive validity.
and at the time of discharge from the hospital. Examination of the correlational structure of the symp-
There were no significant differences among the three groups in tom complexes provided additional support for the value of
familial rate of schizophrenia. Only eight patients in the sample the positive-negative distinction, suggesting at the very
had a family history of schizophrenia. least that the positive and negative syndromes are at
INTERNAL CONSISTENCY AND FACTOR ANALYSIS opposite ends of a continuum. Negative symptoms were
highly correlated with one another, as were positive symp-
Although our data suggest that the concepts of positive and toms to a lesser extent. On the other hand, the correlations
negative schizophrenia may have some predictive validity, we also between positive and negative symptoms were negative.
considered it useful to explore the correlational relationships These correlations emerged clearly when the data were
between the positive and negative, symptom complexes as an
internal validator. In these analyses we also included a tenth subjected to principal components analysis, which yielded
a large bipolar general factor with opposite loadings for
symptom, catatonic motor behavior, in addition to the four
positive and five negative symptoms described above. This symp- positive and negative symptoms.
tom was evaluated by using SADS interview items plus a global Although the diagnostic criteria proposed in this inves-
rating of severity. tigation thus appear to be both useful and internally
The correlations among these ten symptoms are shown in Table consistent, they must be viewed as preliminary for several
5. On the whole, the positive symptoms tended to be positively reasons. First, they provide only a cross-sectional index.
correlated with one another, as did the negative symptoms. Clinical observation suggests that patients whose symp-
Cronbach's a, an index of internal consistency, was calculated for toms are initially negative tend to remain negative when
the four positive symptoms and the five negative symptoms. The a followed longitudinally, whereas in many patients with
for positive symptoms was .397, and the a for negative symptoms initial positive symptoms, negative symptoms eventually
was .849. The a for all nine symptoms was .302. These results
suggest that the negative symptoms are measuring a unitary develop. For some patients, the elimination of delusions,
dimension. This measure of internal consistency for negative hallucinations, or positive thought disorder through the
symptoms is quite high, in spite of the fact that it is based on a use of neuroleptics leaves an underlying residuum of
relatively small item set for the calculation of internal consisten- relatively severe negative symptoms. Subsequent investi-
cy. On the other hand, the a for positive symptoms was relatively gations should explore the evolution of positive and nega-
low, which implies that the group of positive symptoms may tive symptoms through time, in relation both to treatment
represent more than one type of symptom complex. Because received and to the severity of the course of the illness.
internal consistency ordinarily increases as the number of items The relationship of mixed schizophrenia to the positive
increases, the further decrease in internal consistency when all and negative syndromes also warrants further investiga-
nine items are pooled also suggests that these symptom complexes
represent several different dimensions rather than a single
tion. The mixed group may represent a bridge between two
dimension. ends of a continuum, it may be a distinct subtype, or it may
A principal components analysis was also performed, and its be a group of patients with positive schizophrenia pro-
results appear in Table 6. This analysis yielded four components. gressing toward negative schizophrenia. Furthermore, the
The first was a large general factor that is bipolar and accounts mixed group may contain several different subgroups, as
for 42% of the variance. This factor had very large positive some patients in it meet criteria for both positive and
loadings on all five negative symptoms and large negative load- negative schizophrenia, and others meet criteria for nei-
ings on delusions and hallucinations. Catatonic motor behavior ther.
also had a relatively high positive loading, suggesting that it may Yet another area requiring investigation is the relation-
also be a negative symptom. Factor 2, which explains 14% of the
variance, had large positive loadings on positive formal thought ship between negative schizophrenia and simple schizo-
disorder and bizarre behavior and a negative loading on catatonic phrenia or schizotypal personality. None of our patients
motor behavior. Whereas factor 1 appears to tap the general could be considered simple or schizotypal, as all met
positive-negative symptom dimensions, factor 2 appears to tap DSM-III criteria for schizophrenia, which require the

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Auckland User on 06/04/2015

 presence of positive symptoms at some time during the subtypes. That is, if one assumes that diagnostic subtypes
illness. The requirement that positive symptoms be reflect differing underlying causes, phénoménologie char-
present was introduced by the authors of DSM-III to acteristics may be a poor way to identify etiologic sub-
narrow the concept of schizophrenia and to exclude non- types. Syphilis is of course the classic case of an illness
psychotic or latent forms. Although it was recognized that with many different clinical symptoms but a single patho-
some negative symptoms, such as affective flattening, are genesis. In fact, the many manifestations were often not
important indices of schizophrenia, negative symptoms recognized as a single disease until the spirochete was
were de-emphasized in the criteria because of a concern identified as the cause. Once we know the cause (or causes)
about poor reliability. Nevertheless, the DSM-III criteria of schizophrenia and have a laboratory test for making the
may give an excessive prominence to positive symptoms, diagnosis, the use of cross-sectional phenomenology to
for negative symptoms may be more important as prognos- define subtypes may well appear meaningless. In the
tic indicators. Although it represents a more severe syn- meantime, however, clinical diagnosis and phenomenology
drome, negative schizophrenia does not differ markedly aids in identifying biologic correlates and establishing
from schizotypal personality, and it is similar to the causes. Indefinite as it is, it is the most definite thing we
concept of simple schizophrenia that appeared in DSM-I have.
and -II but was dropped in DSM-III.
This research was supported in part by National Institute of Mental
Finally, in spite of the various significant findings in our Health grant MH 31593.
investigation, it may be that cross-sectional phenomenolo- Lydia Jeffries and Kelly Rowe, MA, assisted in interviewing the
gy is not the best method for identifying diagnostic patients.
References

1. Kraepelin E: Dementia Praecox and Paraphrenia, Zinkin J (trans). Psychiutry 1982;139:297-302.

New York, International Universities Press Inc, 1950. 15. Andreasen NC: Thought, language, and communication disorders: II.
2. Bleuler E: Dementia Praecox or the Group of Schizophrenias, Zinkin J Diagnostic significance. Arch Gen Psychiatry 1979;36:1325-1330.
(trans). New York, International Universities Press Inc, 1950. 16. Angrist B, Rotrosen J, Gershon S: Differential effects of neuroleptics
3. Kleist K: Schizophrenic symptoms and cerebral pathology. J Ment Sci on negative versus positive symptoms in schizophrenia. Psychopharmacol-
1970;106:246-255. ogy 1980;72:17-19.
4. American Psychiatric Association, Committee on Nomenclature and 17. Andreasen NC: Scale for the Assessment of Negative Symptoms
Statistics: Diagnostic and Statistical Manual of Mental Disorders, ed 3. (SANS). Iowa City, University of Iowa, 1981.
Washington, DC, American Psychiatric Association, 1980. 18. Andreasen NC: Negative symptoms in schizophrenia: Definition and
5. Kasanin J: The acute schizoaffective psychoses. Am J Psychiatry reliability. Arch Gen Psychiatry, in press.
1933;90:97-126. 19. Andreasen NC: Thought, language, and communication disorders: I.
6. Tsuang MT, Winokur G: Criteria for subtyping schizophrenia: Clinical Clinical assessment, definition of terms, and evaluation of their reliability.
differentiation of hebephrenic and paranoid schizophrenia. Arch Gen Arch Gen Psychiatry 1979;36:1315-1321.
Psychiatry 1974;31:43-47. 20. Endicott J, Spitzer RL: A diagnostic interview: The schedule for
7. American Psychiatric Association, Committee on Nomenclature and affective disorders and schizophrenia. Arch Gen Psychiutry 1978;35:837-
Statistics: Diagnostic and Statistical Manual of Mental Disorders, ed 2. 844.
Washington, DC, American Psychiatric Association, 1968. 21. Harris JG: An abbreviated form of the Phillips rating scale of
8. Stephens JH, Astrup C, Mangrum JC: Prognostic factors in recovered premorbid adjustment in schizophrenia. J Abnorm Psychol 1975;84:129-
and deteriorated schizophrenics. Am J Psychiatry 1966;122:1116-1121. 137.
9. Vaillant GE: Prospective prediction of schizophrenic remission. Arch 22. Personell History for Demographic Duta, NIMH Collaborative Study
Gen Psychiatry 1964;11:509-518. of the Psychobiology of Depression. National Institute of Mental Health,
10. Kety SS: Biochemical theories of schizophrenia: I. Science 1978.
1959;129:1528-1532. 23. Folstein MF, McHugh PR: Mini mental state. J Psychiatr Res
11. Hughlings-Jackson J: Selected Writings, Taylor J (ed). London, Hod- 1975;12:189-198.
der & Stoughton Ltd, 1931. 24. Andreasen NC, Smith MR, Jacoby CG, et al: Ventricular enlargement
12. Strauss JS, Carpenter WT, Bartko JJ: The diagnosis and understand- in schizophrenia: Definition and prevalence. Am J Psychiatry 1982;139:292-
ing of schizophrenia: II. Speculations on the processes that underlie 296.
schizophrenic symptoms and signs. Schizophr Bull 1974;11:61-76. 25. Andreasen NC, Endicott J, Spitzer RL, et al: The family history
13. Crow TJ: Molecular pathology of schizophrenia: More than one method using diagnostic criteria: Reliability and validity. Arch Gen Psychi-
disease process? Br Med J 1980;280:1-9. atry 1977;34:1229-1235.
14. Andreasen NC, Olsen SA, DennertJW, et al: Ventricular enlargement 26. Satz P: Pathological left handedness: An explanation. Cortex
in schizophrenia: Relationship to positive and negative symptoms. Am J 1972;8:121-125.

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Auckland User on 06/04/2015