to Cognitive Deficits in Schizophrenia€¦ · schizophrenia"; Crow 1985) is a variant of this approach. Most re-cently, interest emerged in the cognitive psychology of specific symptoms

VOL 19, NO. 2, 1993 Relations of Symptomsto Cognitive Deficits inSchizophrenia

215

by Milton E. Strauss Abstract

Schizophrenia is characterized bya variety of cognitive dysfunc-tions. Information-processing dys-functions differ between clinicalsubtypes such that nonparanoidschizophrenia patients attend lessthan paranoid schizophrenia pa-tients to connotative or con-textual aspects of stimuli. Thepositive and negative symptomdimensions are also associatedwith distinct cognitive deficits.In general, positive symptomsare related to auditory-processingdeficits and negative symptomsto visual/motor dysfunctions. Theinteraction of frontal and sep-tohippocampal brain systems,and failures of information-processing automaticity and self-monitoring, have been proposedas the bases of positive symp-toms. Negative symptoms arethought to arise from abnor-malities in the complex interac-tions of frontal and striatal sys-tems. Recent theoretical analyseshave recommended a focus onthe cognitive and neuropsycho-logical analysis of specific symp-toms (e.g., hallucinations and de-lusions) instead of on the moreheterogeneous symptom clustersor dimensions. Studies of spe-cific symptoms indicate that pa-tients with hallucinations havedeficits in discriminating thesource of information. Delusionshave been related to abnormalinference processes as well asabnormal perceptual experiences.Studies should now examine thelinks between information-processing abnormalities andsymptoms over time, as the latterchange, within the framework ofexplicit, disconfirmable theoreti-cal models.

People with schizophrenia gener-ally perform inefficiently oninformation-processing tasks. Thecognitive dysfunction of schizo-phrenia, or "psychological deficit"(Hunt and Cofer 1944), has beenthe subject of extensive empiricaland theoretical analysis over thelast 50 years (Hunt and Cofer1944; Buss and Lang 1965; Langand Buss 1965; Chapman andChapman 1973; Cromwell 1975).Advances in cognitive psychology,neuropsychology, and neurosciencenow allow us to describe mecha-nisms of the cognitive dysfunctionstypically observed in schizo-phrenia, as well as possible neuro-biological substrates (Steinhauer etal. 1991).

Schizophrenia is an experientialand behavioral disorder that re-quires analysis at biological, psy-chological, and social levels. Thestudy of cognitive processes hasbeen an important avenue for link-ing the biological and social as-pects of the illness (Hemsley 1991).Three principal approaches havebeen used to join the phenomenol-ogy and cognitive psychology ofschizophrenia. The first was thecomparison of Kraepelinian sub-types to determine whether therewere differences in cognitive defi-cits. In the second, positive andnegative symptom dimensionswere studied, following on thesuggestion by Strauss et al. (1974)that these independent dimensionsmight reflect disturbances in dif-ferent underlying processes ormechanisms. The comparison of

Reprint requests should be sent toDr. M.E. Strauss, Dept. of Psychology,Case Western Reserve University,10900 Euclid Ave., Cleveland, OH44106-7123.

by guest on October 5, 2016

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

http://schizophreniabulletin.oxfordjournals.org/

216 SCHIZOPHRENIA BULLETIN

cognitive deficits in patients classi-fied as Type I ("positive schizo-phrenia") and Type II ("negativeschizophrenia"; Crow 1985) is avariant of this approach. Most re-cently, interest emerged in thecognitive psychology of specificsymptoms themselves. This articlereviews these approaches to thestudy of symptom-cognition rela-tions in schizophrenia and de-scribes principal findings, meth-odological and conceptual issues,and promising trends.

Psychological Deficits InSchizophrenia Subtypes

The psychological deficits ofschizophrenia are pervasive; pa-tients perform less well than othergroups on most cognitive or atten-tional tasks (Hemsley 1991). And,as was evident in the first reviewof this field (Hunt and Cofer1944), these deficits are also quitevariable in groups of patients. In-deed, the variability of perform-ance in a group of patients withschizophrenia may exceed theaverage difference between themand controls (Cromwell 1975).

The great variability of schizo-phrenia patients' performance onpsychological tasks was demon-strated in Shakow and colleagues'early work on the mechanisms ofpsychological deficit in schizo-phrenia (Shakow and Huston 1936;Huston et al. 1937; Rodnick andShakow 1940; Shakow 1962). In thecourse of their research, Shakow etal. discovered substantial correla-tions (r = 0.4-0.6) between the pa-tients' performance and theirmotivation, cooperativeness, andinterest.

While reduced motivation con-tributes to the poor performanceof schizophrenia patients on psy-

chological tasks, it is by no meansthe only cause. Some recent cogni-tive studies have demonstratedthat the performance of schizo-phrenia patients on cognitive taskscan be better than that of controls.For example, schizophrenia subjectsperform better than controls onvisual estimation of numbers be-cause they have a deficit in theirability to group them (Schwartz-Place and Gilmore 1980). Patientswith schizophrenia also may learnan association more rapidly thancontrols because of dysfunctions ininhibitory or selective attentionprocesses, as in the latent inhibi-tion studies described below.

Reduced motivation and lack ofinterest are certainly importantcontributors to the variability ofperformance seen within groups ofschizophrenia subjects. Other fac-tors more specific to schizophreniahave also been shown to contrib-ute to variability in schizophrenicpsychological deficit. The threethat have been investigated mostextensively are symptom pattern,stage of illness (chronicity), andpremorbid adjustment (Cromwell1975).

Like cognitive deficits, symptompatterns are variable in schizo-phrenia, as the disorder is not de-fined by a conjunctive symptomset. In both Kraepelinian andBleulerian formulations, the com-mon feature(s) among patients wassomething other than symptoms.For Kraepelin (1919/1971), thecommon features were onset(early) and course (progressive de-terioration), while for Bleuler(1911/1950) the commonality wasin underlying mechanism (breakingof associative threads; Neale 1987).Some symptomatic homogeneitywas achieved through classifyingpatients into clinical subtypes (e.g.,paranoid, catatonic, hebephrenic).

The investigation of relations be-tween these subtypes and others(chronic vs. acute, process vs. reac-tive) and both pattern and severityof psychological deficit was thefirst approach taken to relatesymptoms to cognitive disorder inschizophrenia.

In seminal papers, Silverman(1964) and Venables (1964) re-viewed the already large literatureon relations of psychological deficitto Kraepelinian subtype (particu-larly paranoid vs. nonparanoid),premorbid adjustment, and chron-icity (Schooler and Feldman 1964).Both Silverman (1964) and Ven-ables (1964) noted that subclassi-fying patients conjointly on thethree dimensions of subtype, pre-morbid adjustment, and chronicityreduced the variability of perform-ance in schizophrenia subgroups.Furthermore, this tripartite sub-classification identified distinct pat-terns of cognitive and psycho-physiologic deficits. For instance,nonparanoid, poor premorbid, andchronic patients were found toperform more poorly than controls.Paranoid patients, especially thosewho had made adequate psychoso-cial adjustments before the onsetof the illness and who had beenill for only a short time, often didnot differ from normals and some-times performed better than con-trol groups on cognitive/attentionaltasks (Cromwell 1975).

The study of relationships be-tween cognitive dysfunctions andclinical characteristics appeared tobe an avenue for rationalizing thecognitive variability in schizo-phrenia. It permitted the identi-fication of subgroups whoseinformation-processing characteris-tics were different from each otheras well as from controls. Three re-lated principles emerged from thisresearch to organize our under-



ownloaded from


VOL 19, NO. 2, 1993 217

standing of cognition in schizo-phrenia: (1) fundamentally differentpatterns of information processingor cognitive style are associatedwith paranoid and nonparanoidschizophrenia; (2) the pattern ofcognitive characteristics observedin schizophrenia is tempered bythe premorbid history of the pa-tients; and (3) these cognitivestyles appear to change over thecourse of the illness, possibly asan adaptive process (Venables1964; Silverman 1972).

The study of subgroup dif-ferences was the basis of a num-ber of hypotheses about themechanisms of schizophrenia. Sil-verman's (1964, 1967, 1972) theoryof relations between cognition andclinical features of schizophreniafocused on three aspects of atten-tion: the modulation of intensity ofinput ("stimulus intensity control"),breadth of attention ("scanningcontrol"), and relative focus onsensory rather than conceptual at-tributes of stimuli ("sensory inputprocessing/ideational gating"). Sil-verman (1967) saw cognitivecharacteristics changing over thecourse of the disorder, as did Ven-ables (1964), who suggested thatchronic patients developed a cogni-tive style of narrowing attention inresponse to states of hyperarousal.These theories were based oncross-sectional data rather thanlongitudinal studies of patients,and there was little direct supportfor their hypotheses of cognitivechanges over the course of schizo-phrenia (Strauss 1973).

Silverman (1967) also suggestedthat nonparanoid and poor pre-morbid schizophrenia patients at-tended less to connotative or con-textual aspects of stimuli and moreto sensory/perceptual propertiesthan normal subjects. This ideawas elaborated by Magaro (1980)

in a comprehensive theory dif-ferentiating the cognitive style ofparanoid from other schizophreniasubjects. Magaro (1980) proposedthat the paranoid schizophreniapatient is biased toward the use ofconceptual rather than perceptualprocessing strategies; that is, his orher information processing is moregoverned by schemata and expec-tations than by perceptual analysisand inference.

Interest in subtype, premorbid,and chronicity constructs waned inthe late 1970s, when the schizo-phrenia concept was being refor-mulated along more Kraepelinianlines. The narrower construct ofschizophrenia as specified byFeighner et al. (1972), Research Di-agnostic Criteria (RDC; Spitzer etal. 1978), and DSM-HI criteria(American Psychiatric Association1980), which include chronicity(poor outcome/duration) as part ofthe diagnostic criteria, led to re-duction in the variability of pre-morbid adjustment and chronicity(Westermeyer and Harrow 1984).In the 1980s, symptom patterns ofschizophrenia and their relation toneurocognitive dysfunctions cameto be studied in terms of the posi-tive and negative symptomdimensions.

Psychological Deficit inRelation to Positive andNegative Symptoms

Crow (1985) proposed two distinctforms of schizophrenia. Type Iwas characterized by positive orproductive symptoms caused byreversible neurotransmitter abnor-malities, while Type II wasmarked by the presence of cogni-tive deficits, negative symptoms,and structural brain abnormalities.Crow's hypothesis, the develop-

ment of frameworks (e.g., Strausset al. 1974), and methods for thereliable assessment of these symp-toms (e.g., Schedule for Assess-ment of Negative Symptoms[SANS; Andreasen 1984o] andSchedule for Assessment of Posi-tive Symptoms, [SAPS; Andreasen1984b]) stimulated research on thepositive and negative symptom ty-pology (Walker and Lewine 1988).

Crow (1985) hypothesized a rela-tion between cognitive dysfunctionand negative symptoms, seeingboth as mediated by structuralbrain impairment. The associationbetween cognitive dysfunction andnegative symptoms was confirmedin a number of studies that usedclinical assessments of mental sta-tus to assess cognition. However,when more detailed assessments ofcognitive processes were made, therelationships between positive andnegative symptoms and cognitionwere equivocal (Walker andLewine 1988). The measurementand conceptualization of positiveand negative symptoms alsoproved more problematic than an-ticipated (Andreasen and Grove1986; Carpenter et al. 1991).

Negative symptoms form a morecoherent construct than do positivesymptoms (Andreasen and Grove1986). The internal consistency re-liability (a) for the SANS (0.85,Andreasen and Grove 1986) issubstantially higher than for theSAPS (0.48, Andreasen and Grove1986; see also Andreasen andOlsen 1982). Internal consistencyreliability (a) is evidence of theconstruct validity of a dimensionand has Important consequencesfor empirical research. Reliabilitysets an upper bound on the cor-relation of a measure with othervariables and so affects what canbe found in a study. Scale re-liability coefficients, or interitem



ownloaded from



correlations among positive andnegative symptom items, are rarelyreported in studies on cognitivecorrelates of symptom patterns. Al-though interrater reliability is typ-ically reported, it is not a sub-stitute for scale reliability.

The importance of examiningscale reliability is illustrated byAddington and Addington's (1991;Addington et al. 1991) recent lon-gitudinal study. Moderate interitemcorrelations were found amongnegative symptoms (0.45 =£ r =£0.57), but interitem associations forpositive symptoms were so low(-0.01 =£ rs « 0.35) that the inves-tigators felt it inappropriate tosum the SAPS item scores into asingle scale. These investigatorsfound a number of significant rela-tions between negative symptomtotal and neuropsychological meas-ures, but virtually none with thepositive symptom items, each ofwhich was considered separately.Single item "scales" are generallyunreliable, so the absence of cor-relations between each positivesymptom and neuropsychologicalmeasures is not surprising. This isan extreme case, but it demon-strates the importance of the psy-chometric properties of symptommeasures in studies of theircorrelates.

Green and Walker (1986) ex-amined the relation between theseverity of positive and negativesymptoms and a wide range ofneuropsychological functions in aninpatient sample. (Unless otherwiseindicated, the studies reviewedevaluated positive and negativesymptoms with the SAPS and theSANS, respectively.) The data wereevaluated by (1) comparing theperformance of patients classifiedinto positive type, negative type,and mixed type, depending on thepreponderance of symptoms and

(2) multiple regression models. Inthe latter, positive and negativesymptom scores were treated ascontinuous variables so that varia-tion in cognitive processing couldbe related to individual differenceson both dimensions as well as ontheir interaction. Green and Walker(1986) found few consistent trendsin the group comparisons, but aninteresting pattern emerged in thecorrelations. Higher negative symp-tom levels were associated withpoorer performance on visualmemory (Benton Visual RetentionTest; Lezak 1983) and motor speedand dexterity tasks (Perdue Peg-board; Lezak 1983). On the otherhand, higher positive symptomlevels were associated with deficitson verbal memory (Buschke's se-lective reminding procedure[Buschke and Fuld 1974]) and lan-guage comprehension (Token Test)deficits.

On the basis of the Green andWalker (1986) study and others,Walker and Lewine (1988) offeredthe generalization that visual-processing and motor deficits areassociated with negative symp-toms, while auditory-processingdeficits are linked with positivesymptoms. Nuechterlein et al.(1986) found negative symptoms(assessed with the Brief PsychiatricRating Scale [BPRS; Overall andGorham 1962]) linked with poorperformance on tasks that makeheavy demands on visual and mo-tor processes, visual vigilance onthe Continuous Performance Test(CPT; Rosvold et al. 1956), andtarget discrimination on the Spanof Apprehension (Neale 1971)tasks; this has been partially repli-cated by Strauss et al. (in press),who found a correlation betweenBPRS negative symptoms andSpan of Apprehension, but notCPT. Studies of backward masking

also indicate that negative symp-toms are associated with deficitson tasks involving rapid process-ing of visual information (seeGreen and Walker 1984; Braff1989; and Merriam et al. 1990,who used the Positive and Nega-tive Syndrome Scale [PANSS; Kayet al. 1989] to assess symptoms).

Difficulty in processing auditoryverbal information, especially ontests of distractibility, is a robustcorrelate of positive symptoms. Ina study that assessed selective at-tention by measuring the decreasein digit recall caused by distrac-tion, Oltmanns and Neale (1975)demonstrated that auditory dis-tractibility was a specific, differen-tial deficit. Using this task withgroups of positive symptom, nega-tive symptom, and mixed symp-tom schizophrenia subjects, Walkerand her colleagues (Green andWalker 1986; Walker and Harvey1986) found significant selective at-tention deficits only for the posi-tive symptom group. Unfor-tunately, correlational analyseswere not reported in these studies.Green and Walker (1986) alsostudied dichotic shadowing alonewith digit-span distraction. Therewere no differences betweensymptoms-type groups on theshadowing task, although therewere on auditory distraction. So,the association between positivesymptoms and auditory processingdeficits is not seen consistently.

The effects of distractions on anauditory task were also studied byComblatt et al. (1985). They inves-tigated the effects of two kinds ofdistraction on the performance ofa task in which auditorily pre-sented names had to be matchedto pictures. One distraction condi-tion consisted of irrelevant noisesthat partially overlapped with thecritical auditory stimuli. In the sec-



ownloaded from


VOL. 19, NO. 2, 1993 219

ond distraction condition a storywas recited by a male voice, whilethe target words were spoken bya female voice. Negative symptomswere not correlated with perform-ance in any of the conditions.Positive symptoms were unrelatedto scores in the baseline no-distraction condition but were cor-related with increased errors withboth kinds of distractions.

The partially masking distractioncondition was a selective attentiontask because the sounds were ir-relevant information. The narrativetext condition, however, was a di-vided attention task. The auditorilypresented text was not irrelevantfor the subjects: they were warnedthat a memory test for the narra-tive material would follow theword-picture task. Although posi-tive symptoms were related toperformance on the name-picturematch task in both selective- anddivided-attention conditions, theywere not related to memory forthe narrative material. Instead, per-formance on the multiple choicememory test was related to nega-tive symptoms. The "overload con-dition" memory task, as Comblattet al. (1985) termed the narrativedistraction condition, is psychologi-cally complex, so it is difficult toattribute the symptom associationto any specific psychologicalprocess.

It is noteworthy that positivesymptoms predict performance ontasks in which language stimuliare presented auditorily becausehallucinations in schizophrenia arepredominantly auditory and verbal.It would be interesting to evaluatethe differential relationship of au-ditory and visual verbal learningtasks and of tone and shape dis-crimination tasks to positive andnegative symptoms. This wouldclarify whether positive symptoms

are related to processing in theauditory modality or to languagestimuli.

The hypothesis that stimulusmodality accounts for relations ofcognitive tasks with positive andnegative symptoms may be correct,but it does not address the proc-esses that connect performancedeficits and symptom manifesta-tions. Identifying the mechanismsin single tasks is itself problematicsince there is no isomorphism be-tween a performance deficit and asingle processing impairment. Thedeficits that schizophrenia subjectshave on even "simple" tasks suchas backward masking and visualvigilance may indicate dysfunctionsat sensory/input processing or la-ter cognitive levels.

Isolating deficient neurocognitiveprocesses requires an integratedseries of studies that use distinctlydifferent, theoretically linked tasksor measures. Knight (1987) arguedthat such focused, analytic experi-mental designs are also necessaryfor the study of cognitive dif-ferences between schizophreniasubgroups or of relations betweencognitive dysfunctions and symp-toms. In his view, theoretical mod-els should guide the design of aset of information-processing para-digms to measure the cognitivefunction of interest. The test of ahypothesized relationship betweena symptom, or symptom cluster,and information processing then isin the pattern of performance defi-cits across a series of measures.Knight (1987) suggests that thisapproach is conceptually morepowerful than the psychometrictask-matching strategy proposed byChapman and Chapman (1978).His approach to testing theoriesdepends on the availability of cog-nitive models that are preciseenough to permit the design of a

set of converging experimentalparadigms. A statistical approachto more explicit specification ofhypothesized mediating processesis offered by latent trait modeling(Bentler 1980). This technique hasbeen useful in some areas of psy-chological research but may haveonly limited applicability in psy-chopathology research. The designrequirements of latent trait studies,particularly multiple measures ofeach construct, large samples, andcross-validation, are difficult to im-plement in clinical studies ofschizophrenia (see Breckler [1990]for a review of uses and misusesof this approach). The cognitivescience strategy of deriving experi-mental manipulations from con-nectionist models may be a morefeasible approach to theory de-velopment and testing. This is il-lustrated in Cohen and Servan-Schreiber's work (1991, 1992),which is discussed below.

The interpretation of relationsbetween cognitive and symptommeasures is plagued by a numberof issues. As with symptom meas-ures, insufficient attention has beenpaid to the psychometric proper-ties of test batteries. It is difficultto isolate constructs with tests thatare factorially or componentiallycomplex, as are most neuropsycho-logical measures (Lezak 1983). Fur-ther, relations between clinicalmeasures and cognitive tasks canbe psychometric artifacts caused bythe differences among tasks in dis-criminating power in the specificsamples studied (Chapman andChapman 1978; Strauss and AHred1987). Studying symptom-performance associations using testbatteries may not permit the iden-tification of links to specific defi-cits. The correlational patterns mayreflect general performance defi-ciencies, especially if the tasks



ownloaded from



have differential sensitivity(Chapman and Chapman 1978;Knight 1987).

Measures that are independentin some populations may be corre-lated in patients with schizo-phrenia, raising the possibility thatin schizophrenia the cognitive taskmeasures a generalized deficitrather than a specific ability orprocess (Chapman and Chapman1978). An example of this may befound in Addington et al.'s (1991)study of the relations betweenneuropsychological deficits andpositive and negative symptoms.They found significant relations be-tween negative symptoms andboth general intellectual ability(IQ) and executive functions (asassessed by the Wisconsin CardSorting Test [WCST; Heaton 1981]and Category Instances FluencyTest [Lezak 1983]) both at indexhospitalization and at a 6-monthfollowup. Factor analysis showedthat the executive function meas-ures and IQ loaded on the samedimension. Executive functionmeasures, such as the WCST andCategory Fluency, and general IQtend to be independent in othergroups (Stuss and Benson 1986).This suggests that all three testsmay be measures of general intel-lectual inefficiency in schizo-phrenia. Wagman et al. (1987),using a test battery selected to tapindependent functions, also founda general cognitive performancefactor in their factor analysis of aneuropsychological battery in asample of outpatients withschizophrenia.

As noted earlier, the measure-ment of positive and negativesymptoms has also been problem-atic. The constructs themselves arenot without ambiguity. There areseveral overlapping but far fromisomorphic conceptualizations of

positive and negative symptomsand measures thereof (McGlashanand Fenton 1992). Initially, positiveand negative symptoms were con-strued as unidimensional (An-dreasen and Olsen 1982), givingrise to the classification of patientsinto positive, mixed, or negativesymptom types. Other investigatorsreported that the two dimensionswere independent, conforming toStrauss et al.'s (1974) earlier for-mulation (Walker and Lewine1988).

Neither a unidimensional nor atwo-factor model appears to ac-count for the correlation patternstypically observed between SANSand SAPS component scales. Amdtet al. (1991) recently confirmed thefindings of Bilder et al. (1985) andLiddle (1987) that the SANS andSAPS subscales are best sum-marized by a three-factor model.The SANS components all load onone factor. The SAPS scales divideinto two independent dimensions,one defined by delusions and hal-lucinations and the other by posi-tive thought disorder and bizarrebehavior. Amdt and associates'(1991) study applied multiple ap-proaches to factor analysis to threereasonably large samples (n's =55-93). The findings of this study,combined with the findings fromsamples in New York (Bilder et al.1985) and England (Liddle 1987),indicate that the three-componentstructure of negative symptoms,delusions/hallucinations, andthought disorder is robust.Addington and Addington (1991)reported this same structure in asample of clinically stable patients,although the structure was not asclear-cut when the symptom rat-ings were obtained while the pa-tients were acutely psychotic.

The studies reviewed have as-sessed negative symptoms cross-

sectionally and without regard toorigins, an approach criticized byCarpenter et al. (1991), who arguethat transitory negative symptomsmust be differentiated from per-sistent (present longitudinally), pri-mary (independent of episode, de-pression, or medication) negativesymptoms, which constitute theirdeficit state of schizophrenia. Thelatter should be the focus of studyof neurocognitive correlates ofnegative symptoms.

As noted above, two approacheshave been used to analyze the re-lations between positive/negativesymptoms and cognitive measures:(1) comparing patients grouped aspositive, negative, or mixed, and(2) correlating symptom scores andcognitive scores. Which one ispreferable is partly a psychometricissue. More information is used inthe correlational analyses (Greenand Walker 1986), and any groupcomparison analysis of variancecan be modeled with regressionusing the full range of scores (Co-hen and Cohen 1983). Furthermore,grouping patients into positive,negative, and mixed types is ap-propriate only if positive andnegative symptom clusters arepoles of a single dimension, asAndreasen and Olsen (1982) andCrow (1985) initially proposed.Unidimensionality is indicated by anegative correlation between posi-tive and negative symptom scores,which is not commonly found. Thecorrelation between positive andnegative symptoms varies with thepopulation studied (Walker andLewine 1988).

In a cross-sectional study, Bilderand associates (1985) found thatthe thought disorder/bizarre be-havior factor and, to a lesser ex-tent, the negative symptom factorwere correlated with neuropsycho-logical deficits. Hallucinations and



ownloaded from


VOL 19, NO. 2, 1993 221

delusions did not correlate withany neuropsychologial measures.Some of the cognitive deficits inschizophrenia appear to be statemarkers that wax and wane withepisodes of psychosis, while othersare more enduring trait-like fea-tures (Zubin and Spring 1977).Negative symptoms seem to bemore trait-like than hallucinationsand delusions and are less respon-sive to neuroleptic treatment (e.g.,Lewine 1990). Consequently, itwould be interesting to determinewhether the cognitive correlates ofnegative symptoms are more per-sistent than those of the psychoticsymptoms. Longitudinal studies ofcognitive deficits in schizophreniain relation to clinical state areneeded to answer this question.

In a rare longitudinal study, Ad-dington et al. (1991) examined 38patients while they were acutely illand again at 6-month followup.Negative symptoms were relatedto poor performance on measuresof verbal reasoning both duringhospitalization and at followup.Poor verbal fluency was associatedwith high levels of negative symp-toms and low ratings of thoughtdisorder on both occasions. At ini-tial testing, better verbal reasoningwas associated with presence ofdelusions. Bizarre behavior was re-lated to poor verbal reasoningduring hospitalization, as Bilder etal. (1985) also found. The associa-tion between delusions and cogni-tive performance was not seen 6months later, but the relation ofcognitive performance with bizarrebehavior continued. Substantialchange in positive symptoms oc-curred in the 6-month period. Im-provement in these symptoms wasrelated to improvements in cogni-tive functions. Conversely, negativesymptoms changed little, and asexpected this small change was

not associated with change inneuropsychological performance.Additional studies of the relation-ship of symptoms and cognitivedeficits across time would beuseful.

Neuropsychological Analysisof Positive and NegativeSymptoms

The current status of the positiveand negative symptom constructsfrom descriptive, genetic, andtreatment approaches was the sub-ject of an international conference(Marneros et al. 1991). In addition,two novel, provocative, conceptualanalyses of the neuropsychologicalsubstrates of positive and negativesymptoms have also been ad-vanced. The first is by Frith, who,in two essays (Frith 1987; Frithand Done 1988), proposed that afailure to monitor actions internallyresults in the positive symptoms ofhallucinations and delusions, whiledefects in the initiation of spon-taneous action underlie negativesymptoms. Both deficits are meta-cognitive processes, that is, moni-toring or executive systems, ratherthan processing mechanisms. In-deed, Frith explicitly rejected hisearlier theory, which emphasizedselective attention deficits in thegenesis of positive symptoms(Frith 1987; Frith and Done 1988).He links his current model toneuropsychological evidence fortwo separate routes to action, onethat is stimulus-driven by environ-mental events and a second that isself-generated ("willed intention").More specifically, this model pro-poses that information about self-generated acts fails to reach amonitor system, resulting in theexperience of "alien forces" as indelusions of control, unintended

thoughts, the experience of pas-sivity, and other delusions andhallucinations.

In a shady of performance ontwo arcade-like shooting tasks,Frith and Done (1989) tested thehypothesis that experiences ofalien control reflect a deficit in aself-monitoring system. In bothtasks the subject had to "shootdown" a bird that appeared oneither the left or right of thescreen by firing the "gun" on theopposite side of the screen. Sub-jects could correct errors at anypoint in a trial. In one task, therewas immediate visual feedback:the "bullet" emerged from thegun. In the second, a wall hid thetrajectory of the "bullet" for thefirst 2,000 msec after the response.Four groups were compared:schizophrenia patients with delu-sions of alien control, schizo-phrenia patients without thissymptom, affective psychotic pa-tients, and normal controls. Thegroups did not differ in errorrates when there was immediatevisual feedback, and virtually allerrors were corrected within 2,000msec. However, when the trajec-tory of the "bullet" was hiddenfrom view, patients with ex-periences of passivity and aliencontrol were significantly less ableto correct their errors. They caughtup once the "bullet" became vis-ible in the last 800 msec of a trial,so the deficit in error correctionwas restricted to the period inwhich error correction dependedon internal monitoring of anaction.

The authors suggest that a simi-lar deficit in monitoring one's ownresponses can account for anoma-lies of auditory event-related po-tentials (ERPs) in schizophrenia. Innormal controls, the P300 compo-nent of the auditory ERP is



ownloaded from



smaller when subjects trigger stim-uli than when the tones are exter-nally controlled. This expectationeffect is attenuated in schizo-phrenia (Braff et al. 1977). Inter-estingly, problems in bothmonitoring and initiation of be-havior have been demonstrated inanimals following lesions of thefrontal cortex (Fuster 1989). Frithand Done have proposed thatfrontal damage is the basis of bothpositive and negative symptoms(Frith 1987; Frith and Done 1988).In their theory, positive symptomsand associated cognitive deficits re-flect prefrontal cortical impairmentsin interaction with the septohip-pocampal system. Negative symp-toms and associated cognitive defi-cits are due to disruptions infrontal-striatal connections.

The second new neuropsycho-logical analysis of symptoms ofschizophrenia comes from Gray etal. (1991). Their detailed, admit-tedly somewhat speculative, modelfocuses on positive symptoms ofschizophrenia, which, followingCrow (1985), they refer to asType I schizophrenia. The keyconcept in their model is that au-tomaticity in information process-ing fails to develop fully in Type Ischizophrenia. Consequently thebehavior of patients with positivesymptoms is less influenced byregularities of past experience thanis the behavior of normal controls.

One way that immediate pastexperience influences learning andmemory in humans and othermammals is shown in the phe-nomenon of latent inhibition (LI)(Lubow 1989). LI refers to the re-tardation of association learning bypreexposure to the stimulus that isto be conditioned. During pre-exposure, the stimulus is repeat-edly presented but has no signalvalue. Normally, animals habituate

to repetitive, uninformative signals.Consequently, when contingencieschange and the stimulus becomesinformative, it takes longer forpreexposed subjects to leam thecontingency than subjects not pre-viously exposed to the stimulus.There is a dopaminergic basis forLI, and preliminary evidence indi-cates that preexposure does notproduce U in symptomatic Type Ipatients (Gray et al. 1991), whoare presumably in a hyper-dopaminergic state. Acutely ill pa-tients do not show the pre-exposure effect in a number ofexperimental paradigms, whilechronic patients (i.e., those withfewer positive symptoms) and pa-tients with short histories of illnesswho are effectively treated withneuroleptics do show LI (Gray etal. 1991).

Prepulse inhibition of eye-blinkstartle (Braff et al. 1991) and sen-sory gating evoked-potential para-digms (Freedman and Mirsky1991) also reveal abnormalities inthe responses of schizophreniasubjects because an immediatelypreceding stimulus fails to controlinformation processing. Schizo-phrenia patients show less inhibi-tion of a startle response precededby a weak stimulus trial and showless attenuation of the ERP P50 tothe second of a pair of auditorysignals. The failure of selective at-tention produces the typical out-come in these studies of schizo-phrenia: the performance ofpatients is poorer than that of con-trols. In LI paradigms, on theother hand, the failure of selectiveattention in positive symptom pa-tients results in better learningthan in controls.

Gray et al. (1991) submit thatthe genesis of positive symptomsand cognitive abnormalities inType I patients lies in the interac-

tions of corticolimbicstriatal brainsystems and the septohippocampalcomplex, which serves as a moni-tor or comparator of current stim-uli, memory, plans, and motorprograms. As previously discussed,the septohippocampal system isalso central to Frith and Done's(1988) account of self-monitoringdisturbances in positive symptomschizophrenia.

There are differences as well assimilarities in these two recent the-ories. Frith's theory focuses on in-ternal monitoring and planning ofaction (Frith 1987; Frith and Done1988). In contrast, the Gray et al.(1991) theory emphasizes inputdisturbances; that is, disruptions inthe attentional mechanisms that fa-cilitate the development of auto-matic processing (see also Hemsley1991, in press). Both theories areexciting integrations of psychologi-cal and neuroscience perspectiveson schizophrenia. They offer ex-plicit, testable models of the mech-anisms of symptoms and cognitivedeficits. Both the Frith-Done andGray-Hemsley theories providemore detail on positive than onnegative symptoms, although eachoffers hypotheses about the neuro-cognitive bases of the latter aswell. The Gray-Hemsley model hastreated positive symptoms as amore global construct than has theFrith and Done model, whichfocuses more on discrete symp-toms and specific associated cogni-tive abnormalities. As discussedearlier, there is substantial psycho-metric evidence of the independ-ence of delusions and hallucina-tions from thought disorder.Consequently, a focus on specificsymptoms rather than on the posi-tive symptom category would bebeneficial in the further develop-ment of neuropsychologicaltheories.



ownloaded from


VOL. 19, NO. 2, 1993 223

Relations of CognitiveFunctions to SpecificSymptoms

Persons (1986) and, more recently,Costello (1992) have discussed ad-vantages of focusing psychological(or biological) analysis on the ex-planation of discrete symptoms in-stead of syndromes. Both of theseauthors noted the heterogeneity inpatient groups that is inevitablewhen disjunctive criteria are usedto define a diagnostic construct.Persons (1986) also posited that afocus on the cognitive analysis ofspecific symptoms rather thansymptom clusters (syndromes) canbenefit research. Focusing onsymptoms underscores the con-tinuity between normal and clinicalphenomena and encourages morecareful definition and analysis ofsymptoms themselves. Much of theresearch on symptom-cognition re-lations is based on plausibilityrather than on firmly groundedhypotheses (Neale et al. 1985).This makes inconsistent findingsdifficult to reconcile.

Certainly, a focus on discretesymptoms can facilitate theoreticaldevelopment by fostering moreelaborate, tighter explanatory linksbetween proposed mechanisms andthe clinical phenomena (Persons1986; Costello 1992). Knight (1987),however, noted a potential pitfallin this approach. If symptoms arestudied independently of diagnosis,one must assume that the mecha-nisms of symptoms are the sameacross syndromes. Although theremay be a final pathway that iscommon to delusions in schizo-phrenia and, for example, mania,there may be different processesinvolved as well (Neale 1988).Knight (1987) refers to the pos-sibility of different symptom-psychological mechanism relations

across disorders as symptom-process equivocality.

As discussed above, subtypingschizophrenia into paranoid-nonparanoid, premorbid adjust-ment, and chronicity subgroups re-duced within-group variability andisolated some distinctive patternsof cognitive deficits. The subgroup-ing variables thus were "markers"for cognitive differences withinschizophrenia. Theoretical analysisin this literature, as that of TypeI/Type II group differences,tended to deal with the mecha-nisms of cognitive deficits ratherthan the mechanisms of the symp-toms or the processes that mightmediate both (Silverman 1964,1972; Venables 1964; Cromwell1975). Persons' (1986) and Cos-tello's (1992) articles emphasize adifferent tack in the study of cog-nitive processes in conjunctionwith symptomatic characteristics ofschizophrenia: the psychologicalanalysis of the symptoms them-selves. The Frith-Done and Gray-Hemsley theories are examples ofhighly generalized studies of thissort of analysis. Two specificsymptoms, hallucinations and de-lusions, have been the subject ofrecent empirical study and theo-retical analysis (Persons 1986;Oltmanns and Maher 1988; Sladeand Bentall 1988; Bentall 1990).

Hallucinations. A number ofcognitive mechanisms of hallucina-tions have been proposed, andfour principal theoretical ap-proaches were reviewed by Bentall(1990): (1) classical conditioning ofhallucination-like experiences (e.g.,visual afterimages); (2) "seepage"of preconscious mental processesor contents, either through over-arousal or dysfunction in executivecontrols over access of pre-

conscious processing to conscious-ness (Frith 1979); (3) abnormallyvivid mental imagery in conjunc-tion with defective reality testing,or deficits in imagery with mis-attribution of those images that areexperienced; and (4) subvocaliza-tion theories. This last group oftheories posits that hallucinationsare caused by dysfunctions in theneural mechanisms of the sub-vocalizations that typically accom-pany thought. Bentall (1990) dis-cussed a number of limitations ofeach of these approaches andelaborated the view that hallucina-tions represent a breakdown in themetacognitive processes of realitydiscrimination. Reality discrimina-tion refers to the idea "that hallu-cinators mistake their own internal,mental, or private events for exter-nal, publicly observable events"(Bentall 1990, p. 88). That is, hallu-cinations result from deficits in theability to discriminate between realand imaginary events.

Failures in reality discriminationhave been demonstrated byHeUbrun and his associates instudies of patients with histories ofauditory hallucinations. In the firststudy, Heilbrun (1980) contrastedthe performance of hallucinatingand nonhallucinating inpatients ontasks that required recognizing ex-pressed thoughts as one's own.Patients were first interviewed,and their responses to five generalquestions were recorded. Oneweek later, subjects were tested fortheir ability to recognize their ownexpressed thoughts using multiplechoice tests. Patients with a historyof hallucinations were less able torecognize their own thoughts, al-though their memory, communica-tion skill, or stability of opinionsdid not differ. In a subsequentstudy, Heilbrun et al. (1983) foundthat patients with a history of au-



ownloaded from



ditory hallucinations and poor pre-morbid adjustment were less accu-rate in localizing the direction ofthe experimenter's voice than non-hallucinating patients or hallucinat-ing patients with good premorbidadjustment. The patients studiedincluded those with depressiveand personality disorders as wellas schizophrenia. The authorsstated that comparisons of onlythe schizophrenia patients wouldhave produced even strongereffects.

The reality-discriminating abilitiesof currently hallucinating schizo-phrenia patients, delusional psy-chotic patients without halluci-nations (mainly schizophreniasubjects), and normals were stud-ied by Bentall et al. (1991a). Inthis study, subjects were first re-quired to provide responses to aseries of cues (e.g., name a kindof dwelling that begins with H) orwere provided with responses tocues (a type of footwear is ashoe). One week later, withoutforewarning, the subjects were re-quired to recall which responseswere self-generated and which hadbeen provided. Hallucinators wereless able to discriminate betweenmemories of their own thoughtsand memories of information givento them by the experimenter.

Hoffman (1986) also proposedthat misattribution processes ac-counted for auditory hallucinationsas part of a more comprehensivetheory of language productionfailures in schizophrenia. In brief,Hoffman's (1986) hypothesis is thathallucinators experience alienvoices because of disorders of dis-course planning that lead to theexperience of unintended verbalimages. This model brings togetherspeech disorganization, languageprocessing, and hallucinations, thusdirectly linking specific cognitive

deficits to specific symptoms. Ascommentators on Hoffman's (1986)article noted, there are no directtests of this idea, but it can leadto interesting experimental studies.

Schizophrenia is characterized byhallucinations in modalities otherthan audition, so a focus on meta-cognitive processes rather thanspecific cognitive mechanisms suchas audition or language processingmay be a useful strategy. A greatdeal is known about the similari-ties, differences, and interferencesbetween imagination and percep-tion in normal cognition. Johnson(1988) recently reviewed the cogni-tive psychology of confusion be-tween perception and imaginationin both experiencing and remem-bering. She summarized evidencefrom disorders other than schizo-phrenia that suggested that bothvisual and auditory hallucinationsmay be the results of improperinterpretation of fragmentarysensory/perceptual experiences.

Delusions. As noted earlier, fac-tor analyses of positive symptomscales indicate that hallucinationsand delusions both load on onefactor and are independent ofthought disorder. Thus, it mightbe expected that some of the cog-nitive processing characteristics ofhallucinating schizophrenia patientswould also be found in delusionalpatients. Apparently, however,there have been no direct com-parisons of schizophrenia patientswith delusions and those withonly hallucinations.

There is much more theory thanresearch on the psychology of de-lusions (Winters and Neale 1983;Butler and Braff 1991). Oltmannsand Maher (1988) edited a volumethat presents a number of theoreti-cal perspectives on delusions de-rived from cognitive psychology,

hypnosis, and social psychology, aswell as Maher's views, which area jumping-off point for several ofthe other chapters.

A theoretical link between delu-sions and hallucinations was madeby Johnson (1988) in terms of real-ity monitoring, a perspective thatbuilds on the theoretical views ofMaher (1988). Maher proposed thatdelusions are the product of nor-mal reasoning processes applied toaberrant experience. He docu-mented both the range of evidencefor anomalous experience in thegenesis of delusions and the ab-sence of evidence of abnormal log-ical thinking in studies that com-pare paranoid patients withcontrols.

Chapman and Chapman (1988)presented findings from their high-risk group study of college stu-dents that are germane to Maher's(1988) theory. They studied collegestudents who scored very high(> 2 standard deviations [SDs]above the mean) on their Percep-tual Aberration and/or MagicalIdeation scales (Chapman andChapman 1988). These subjects, aswell as controls, were interviewedtwice over a 2-year period with anexpanded version of the Schedulefor Affective Disorders and Schizo-phrenia (SADS-L; Spitzer andEndicott 1977) which permitted rat-ings of 80 types of deviant ex-periences. At the first interview,about half the psychometric high-risk group met DSM-III criteriafor schizotypal personality disor-der. In the small number of high-risk subjects who developed psy-chosis over the followup period(n = 3), there was similarity be-tween the aberrant beliefs ex-pressed in the first interview andthe delusions exhibited during thepsychosis. Isolated delusions werereported on followup by nine ad-



ownloaded from


VOL. 19, NO. 2, 1993 225

ditional high-risk subjects; thesesymptoms were more extreme thanthose manifested on the first inter-view. These findings suggest acontinuity of aberrant beliefs anddelusions in schizophrenia. Suchcontinuity was also seen byHarrow et aJ. (1988) in a retro-spective study of patients withschizophrenia.

Chapman and Chapman (1988)also noted that not all subjectswith similar anomalous experiencesdeveloped aberrant beliefs. Like-wise, some subjects developed de-lusional interpretations of ex-periences generally not consideredanomalous (e.g., at a party afriend "reads your mind" andsays, "Let's go"). The Chapmanssuggested that the genesis of delu-sions lies in the interplay ofanomalous experiences and twocognitive abnormalities of schizo-phrenia. These two abnormalitiesare cognitive slippage, which be-came apparent in their subjectswhen they talked about psychoticand psychotic-like experiences, andthe tendency to process informa-tion in a biased manner. TheChapmans suggest that peoplewho develop delusions assign ex-cessive significance to certain as-pects of their experiences andshow constriction in the range ofinformation they use.

Magaro (1980) also commentedon the tendency of paranoid sub-jects to attribute meaning accord-ing to rigid conceptual expecta-tions. Paranoid patients show arush to closure (Neufeld 1990) andtend to discover relationships be-tween stimuli that do not exist(Brennan and Hemsley 1984). Sincethe "discovery" of illusory correla-tions is a normal cognitive phe-nomenon (Chapman and Chapman1969), it may be that paranoidpatients reason like the rest of us,

but to a greater extent.In a theoretical review of the

formation and maintenance of de-lusions, Hemsley and Garety (1986)proposed that deficits in the abilityto weigh evidence may underliedelusions. They suggested thatBayesian inference could serve asa normative model against whichto evaluate the inference processesof deluded and nondeluded indi-viduals. Within a Bayesian frame-work, cognitive biases can occur ata number of points in the genera-tion of hypotheses (plausible in-ferences), estimation of probabili-ties, and assessment of likelihoodratios. Garety et al. (1991) subse-quently studied the reasoning ofdelusional patients (paranoia/delusional disorder), deludedschizophrenia patients, and anxiouspatients, as well as normal con-trols, on probabilistic inferencetasks. Since delusion-relevant con-tent can perturb cognition, twocontent-neutral tasks were used.As expected, they found that bothdelusional groups requested lessinformation before making a deci-sion. However, the hypothesis thatdelusional groups would be moreoverconfident in their probabilityjudgments was not supported. InBentall and associates' (1991&) astudy of social attribution, delu-sional patients were more certainabout their choices than normalcontrols, but they did differ fromdepressive patients.

Garety et al. (1991) reported thatdelusional patients were morelikely to be extreme respondersthan were members of the othergroups. In post hoc comparisonsof extreme and nonextreme delu-sional responders, they found ex-treme responders had higher de-grees of conviction in their ownbeliefs and were more likely tohave had anomalous experiences

(e.g., auditory hallucinations,strange coincidences). Thus, for atleast a subgroup of patients, ab-normal inferences and abnormalperceptions went hand in hand.Other delusional patients showedneither characteristic (Garety et al.1991). Disregarding the possibilitythat the measurement of anoma-lous experience and inference proc-esses may not have been suffi-ciently sensitive, Garety et al.'s(1991) findings suggest possibleheterogeneity of the mechanisms ofdelusions. On the other hand,there were no differences betweenthe schizophrenia patients with de-lusions and the nonschizophreniagroups.

Bentall and associates (Kaneyand Bentall 1989; Bentall et al.1991b) also studied the notion thatthe differences between the reason-ing proclivities of delusional andnondelusional patients are mainlyquantitative. Social attribution the-ory was the framework used tostudy reasoning in delusional, de-pressed, and normal subjects. Inthe first of two reports (Kaneyand Bentall 1989), delusional pa-tients, like depressed patients, ex-pressed strong beliefs in chanceevents and made excessively stableand global attributions in com-parison with normal subjects. Thedelusional patients, however, madeexternal attributions for negativeevents, while the depressed pa-tients made internal attributionsfor negative events and externalattributions for positive events. Intheir second study (Bentall et al.1991b), subjects selected one ofthree explanations for the actionsof one person toward another. Thechoices were attributing the causeto a person, to a stimulus, or tothe circumstances. The paranoidpatients made more person attribu-tions, especially for negatively val-



ownloaded from



ued actions, although there weremany similarities in performanceamong the groups.

The content of thought, as wellas form, were considered byHemsley and Garety (1986) intheir theoretical analysis of delu-sions. An earlier study from theirlaboratory (Brennan and Hemsley1984) had shown that there werestrong perceptions of correlationbetween unrelated events ("illusorycorrelation") for stimuli related topatients' delusions. Chapman andChapman (1988) also noted greatercognitive slippage among theirhigh-risk subjects when they de-scribed delusional ideas.

Further evidence of the disrup-tive effects of personally relevantstimuli on the information process-ing of delusional patients has beenprovided by Bentall and Kaney(1989). These investigators admin-istered a Stroop color-word inter-ference test (Lezak 1983), a meas-ure of selective attention, to agroup of patients with persecutorydelusions, to a depressed group,and to normal controls. The stim-ulus words were either content-neutral or had depressive or para-noid content. The subject's taskwas to name the color in whichthe word was printed. Usingcolor-naming speed for neutralwords as a referent, the re-searchers found that depressed pa-tients showed greater interferencewith depressive words and de-luded patients were slow onwords with paranoid themes. De-pressed and paranoid patients alsoshowed memory biases in recall ofthematic material. Depressedsubjects more easily recalleddepression-related stimuli, whiledelusional patients recalled threat-related material more easily(Kaney et al. 1991). Both symptom-specific and more general

information-processing abnor-malities are found among delu-sional patients.

Overview and Future Needs

Systematic relationships exist be-tween clinical aspects of schizo-phrenia and cognitive or neuropsy-chological parameters whether theyare studied at the level of sub-types, symptom dimensions, orspecific symptoms. Individual dif-ferences among patients in the se-verity of negative symptoms arerelated to some classes of informa-tion processing, while variability inpositive symptoms is related todifferent cognitive measures. Spe-cifically, deficits in memory,perceptual-motor integration, andvisual information processing ap-pear to be associated with nega-tive symptoms, while auditory in-formation processing, especially oflanguage stimuli, is a principalcorrelate of positive symptoms.

The theoretical basis of these re-lations is relatively undeveloped.The first wave of interest in thistopic occurred at a time (the1960s) when knowledge of neuralmechanisms of cognition and cog-nitive theory itself were justemerging, and the boundaries ofthe schizophrenia construct werenot well conceptualized. The sec-ond wave of research, that onpositive and negative symptom di-mensions, was stimulated by inter-est in evaluating Crow's (1985) as-sertions about relations betweennegative symptoms, brain abnor-malities, and cognitive deficits.This research was driven by inter-est in the concept of two distinc-tive syndromes of schizophrenia,Type I and Type II. Research hasbeen mainly descriptive, with fewstudies attempting to test directly

hypotheses about mechanisms. Therelationship between stimulusmodality (auditory/visual) andsymptom cluster (positive/negative) is open to a number ofinterpretations.

Two trends are apparent in themost recent work. First, specificpositive symptoms rather than thepositive symptom construct arebeing studied. Since it is nowclear that positive symptoms arenot a single construct, this is avaluable change. Conversely,greater coherence exists amongnegative symptoms, so they mayusefully be treated as a construct.However, the etiology and neuro-behavioral mechanisms of thisnegative construct remain unclear(Carpenter et al. 1991; Marneros etal. 1991). The second trend is thedevelopment of explicit neuropsy-chological theories of schizo-phrenia, such as Frith and Done's(1988), Gray et al.'s, (1991), andHoffman's (1986). Although thespecifics of each these theories aresubject to criticism (see publishedcommentaries in articles by Grayet al. [1991] and Hoffman [1986]),each has enough detail to allowclear hypotheses and tests of thesehypotheses.

A common ground is emergingfor the explanation of symptomsand cognition in schizophrenia inthe interaction of discrete, thoughdistributed neural systems. Cohenand Servan-Schreiber (1992) re-cently proposed a connectionistmodel of attention and languageprocessing based on the psycho-biology of the prefrontal cortexand the mesolimbic dopamine sys-tem. Their computational modelsuccessfully simulated the perform-ance of schizophrenia patients onStroop color-word interference, vis-ual vigilance deficits, and lexicalambiguity tasks. Preliminary tests



ownloaded from


VOL 19, NO. 2, 1993 227

of predictions of the model inother information-processing tasksare encouraging (Cohen andServan-Scheiber 1991). As the phe-nomenology and cognitive psychol-ogy of schizophrenia come to bebetter understood in terms ofneurobiology and neuropsychology,the mechanisms linking these twodomains of abnormality will be-come clearer.

Important questions remainabout the relationship betweencognition and symptoms. Cognitivevulnerability markers should bepresent before clinical symptoms.This suggests that the neuropsy-chological processes of cognitivedeficits affect the later develop-ment of symptoms. But how tight-ly linked are cognitive disordersand symptom patterns develop-mentally? Are they driven by com-mon mechanism or is the linkindirect?

Clinically, schizophrenia is a dy-namic disorder marked by a vari-ety of courses in symptom expres-sion (Strauss 1987; Mameros et al.1991). Delusions and hallucinationsappear to dissipate over time(Depue and Woodburn 1975; Pfohland Winokur 1982). But paranoidschizophrenia patients are also lesslikely to be rehospitalized, so theyare less available for study insamples of patients with lengthyillnesses (Strauss 1973). Whensymptoms change, do neuropsy-chological or information-processing measures also change?There is scant direct evidence ofchange in cognitive characteristicsover the course of schizophrenia.Most of our knowledge is cross-sectional, measuring differences be-tween chronic patients and thosewho have not been ill for a longperiod at one point in time. Suchdifferences may reflect samplebiases rather than change (Strauss

1973). To learn whether changes insymptom expression are paralleledby changes in putative cognitivemechanisms, both longitudinalstudies and studies of pharmaco-logic interventions are necessary.

References

Addington, J., and Addington, D.Positive and negative symptoms ofschizophrenia. Schizophrenia Re-search, 5:51-59, 1991.Addington, J.; Addington, D.; andMaticka-Tyndale, E. Cognitivefunctioning and positive and nega-tive symptoms in schizophrenia.Schizophrenia Research, 5:123-134,1991.American Psychiatric Association.DSM-III: Diagnostic and StatisticalManual of Mental Disorders. 3rded. Washington, DC: The Associa-tion, 1980.Andreasen, N.C. The Scale for theAssessment of Negative Symptoms(SANS). Iowa City, IA: The Uni-versity of Iowa, 1984fl.Andreasen, N.C. The Scale for theAssessment of Positive Symptoms(SAPS). Iowa City, IA: The Uni-versity of Iowa, 1984b.

Andreasen, N.C, and Grove, M.D.Evaluation of positive and negativesymptoms in schizophrenia. Psychi-atrie & Psychobiologie, 1:108-121,1986.

Andreasen, N.C, and Olsen, S.A.Negative v positive schizophrenia:Definition and validation. Archivesof General Psychiatry, 39:789-794,1982.Amdt, S.; Alliger, R.J.; and An-dreasen, N.C. The distinction ofpositive and negative symptoms.British journal of Psychiatry,158:317-322, 1991.

Bentall, R.P. The illusion of reality:A review and integration of psy-

chological research on hallucina-tions. Psychological Bulletin, 107:82-95, 1990.

Bentall, R.P.; BaJcer, G.A.; andHavers, S. Reality monitoring andpsychotic hallucinations. BritishJournal of Clinical Psychology,30:213-222, 1991fl.Bentall, R.P., and Kaney, S. Con-tent specific information processingand persecutory delusions: An in-vestigation using the emotionalStroop test. British Journal of Medi-cal Psychology, 62:355-364, 1989.Bentall, R.P.; Kaney, S.; andDewey, M.E. Paranoia and socialreasoning: An attribution theoryanalysis. British Journal of ClinicalPsychology, 30:13-23, 1991£>.

Bentler, P.M. Multivariate analysiswith latent variables: Causal mod-eling. Annual Review of Psychology,31:419-456, 1980.

Bilder, R.M.; Mukherjee, S.; Rieder,R.O.; and Pandurangi, A.K. Symp-tomatic and neuropsychologicalcomponents of defect states.Schizophrenia Bulletin, 11:409-419,1985.

Bleuler, E. Dementia Praecox or theGroup of Schizophrenias. (1911)Translated by J. Zinkin. New York,NY: International UniversitiesPress, 1950.

Braff, D.L. Sensory input deficitsand negative symptoms in schizo-phrenic patients. American Journalof Psychiatry, 146:1006-1011, 1989.

Braff, D.L.; Callaway, E.; andNaylor, H. Very short-term mem-ory dysfunction in schizophrenia:Defective short-term constant infor-mation processing in schizophrenia.Archives of General Psychiatry,34:25-30, 1977.

Braff, D.L.; Saccuzzo, D.P.; andGeyer, M.A. Information processingdysfunctions in schizophrenia:Studies of visual backward mask-



ownloaded from



ing, sensorimotor gating, and ha-bituation. Handbook of Schizo-phrenia, 5:303-334, 1991.

Breckler, SJ. Applications ofcovariance structure modeling inpsychology: Cause for concern?Psychological Bulletin, 107:260-273,1990.

Brennan, J.H., and Hemsley, D.R.Illusory correlations in paranoidand non-paranoid schizophrenia.British Journal of Clinical Psychol-ogy, 23:225-226, 1984.

Buschke, H., and Fuld, P.A.Evaluating storage, retention andretrieval in disordered memoryand learning. Neurology, 11:1019-1025, 1974.

Buss, A.H., and Lang, P.J. Psycho-logical deficit in schizophrenia: I.Affect, reinforcement, and conceptattainment. Journal of AbnormalPsychology, 1:2-24, 1965.

Butler, R.W., and Braff, D.L. Delu-sions: A review and integration.Schizophrenia Bulletin, 17:633-649,1991.

Carpenter, W.T., Jr.; Buchannan,R.W.; Kirkpatrick, B.; Thaker, G.;and Tamminga, C. Negative symp-toms: A critique of current ap-proaches. In: Mameros, A.; An-dreasen, N.C.; and Tsuang, M.T.,eds. Negative Versus Positive Symp-toms. Berlin, Germany: Springer-Verlag, 1991. pp. 126-133.Chapman, L.J., and Chapman, J.P.Illusory correlation as an obstacleto the use of valid psychodiagnos-tic signs. Journal of Abnormal Psy-chology, 74:271-280, 1969.Chapman, L.J., and Chapman, J.P.Disordered Thoughts in Schizo-phrenia. Englewood Cliffs, NJ:Prentice-Hall, 1973.Chapman, L.J., and Chapman, J.P.The measurement of differentialdeficit. Journal of Psychiatric Re-search, 14:303-311, 1978.

Chapman, L.J., and Chapman, J.P.The genesis of delusions. In:Oltmanns, T.F., and Maher, B.A.,eds. Delusional Beliefs. New York,NY: John Wiley & Sons, 1988.pp. 167-183.Cohen, J.D., and Cohen, P. AppliedMultiple Regression/Correlation Anal-ysis for the Behavioral Sciences.Hillsdale, NJ: Erlbaum, 1983.

Cohen, J.D., and Servan-Schreiber,D. "Schizophrenic Performance ina Variant of the CPT-AX: A Testof Theoretical Predictions Concern-ing the Processing of Context."Presented at the Sixth AnnualMeeting of the Society for Re-search in Psychopathology,Cambridge, MA, November 1991.

Cohen, J.D., and Servan-Schreiber,D. Context, cortex, and dopamine:A connectionist approach to be-havior and biology in schizo-phrenia. Psychological Review,99:45-77, 1992.

Comblatt, B.A.; Lenzenweger, M.F.;Dworkin, R.H.; and Erlenmeyer-Kimling, L. Positive and negativeschizophrenic symptoms, attention,and information processing. Schizo-phrenia Bulletin, 11:397-408, 1985.

Costello, C.G. Research symptomsversus research on syndromes: Ar-guments in favour of allocatingmore time to the study of symp-toms. British Journal of Psychiatry,160:304-308, 1992.

Cromwell, R.L. Assessment ofschizophrenia. Annual Review ofPsychology, 26:593-619, 1975.

Crow, T.J. Positive and negativesymptoms and the role ofdopamine. British Journal of Psychi-atry, 137:383-386, 1985.

Depue, R.A., and Woodbum, L.Disappearance of paranoid symp-toms with chronicity. Journal ofAbnormal Psychology, 84:84-86,1975.

Feighner, J.P.; Robins, E.; Guze,S.B.; Woodruff, R.A.; Winokur, G.;and Munoz, R. Diagnostic criteriafor use in psychiatric research. Ar-chives of General Psychiatry, 26:57-63, 1972.

Freedman, R., and Mirsky, A.F.Event-related potentials: Exogenouscomponents. Handbook of Schizo-phrenia, 5:71-90, 1991.

Frith, CD. Consciousness, informa-tion processing and schizophrenia.British Journal of Psychiatry,134:225-235, 1979.Frith, CD. The positive and nega-tive symptoms of schizophrenia re-flect impairments in the perceptionand initiation of action. Psychologi-cal Medicine, 17:631-648, 1987.Frith, CD., and Done, J.D. To-wards a neuropsychology ofschizophrenia. British Journal ofPsychiatry, 153:437^143, 1988.Frith, CD., and Done, J.D. Ex-periences of alien control inschizophrenia reflect a disorder inthe central monitoring of action.Psychological Medicine, 19:359-363,1989.

Fuster, J.M. The Prefrontal Cortex:Anatomy, Physiology, and Neuropsy-chology of the Frontal Lobe. 2nd ed.New York, NY: Raven Press, 1989.

Garety, P.A.; Hemsley, D.R.; andWessely, S. Reasoning in deludedschizophrenic and paranoid pa-tients: Biases in performance on aprobalistic inference task. Journal ofNervous and Mental Disease,179:194-201, 1991.Gray, J.A.; Feldon, J.; Rawlins,J.N.P., Hemsley, D.R.; and Smith,A.D. The neuropsychology ofschizophrenia. Behavioral and BrainSciences, 14:1-84, 1991.Green, M., and Walker, E. Suscep-tibility to backward masking inschizophrenic patients with positiveor negative symptoms. American



ownloaded from


VOL. 19, NO. 2, 1993 229

Journal of Psychiatry, 141:1273-1275,1984.

Green, M, and Walker, E. Atten-tional performance in positive- andnegative-symptom schizophrenia.Journal of Nervous and Mental Dis-ease, 174:208-213, 1986.

Harrow, M.; Rattenbury, F.; andStoll, F. Schizophrenic delusions:An analysis of their persistence, ofrelated premorbid ideas, and ofthree major dimensions. In:Oltmanns, T.F., and Maher, B.A.,eds. Delusional Beliefs. New York,NY: John Wiley & Sons, 1988.pp. 184-211.

Heaton, R.K. Wisconsin Card Sort-ing Test Manual. Odessa, FL: Psy-chological Assessment Resources,1981.Heilbrun, A.B. Impaired recogni-tion of self-expressed thought inpatients with auditory hallucina-tions. Journal of Abnormal Psychol-ogy, 89:728-736, 1980.

Heilbrun, A.B.; Blum, N.; andHaas, M. Cognitive vulnerability toauditory hallucination: Preferredimagery mode and spatial locationof sounds. British Journal of Psychi-atry, 143:294-299, 1983.

Hemsley, D.R. What have cogni-tive deficits to do with schizo-phrenia? In: Huber, G., ed. Idio-pathische Psychosen. Stuttgart,Germany: F.K. Schattauer, 1991.pp. 111-125.Hemsley, D.R. Perception and cog-nition in schizophrenia. In: Crom-well, R.L., and Snyder, C.R., eds.Schizophrenia: Origins, Processes,Treatment, and Outcome. NewYork, NY: Oxford University Press,in press.

Hemsley, D.R., and Garety, P.A.The formation of maintenance ofdelusions: A Bayesian analysis.British Journal of Psychiatry, 149:51-56, 1986.

Hoffman, R.E. Verbal hallucinationsand language production processesin schizophrenia. Behavioral andBrain Sciences, 9:503-548, 1986.

Hunt, J.M., and Cofer, C.N. Psy-chological deficit. In: Hunt, J.M.,ed. Personality and the BehaviorDisorder. Vol. 2. New York, NY:Ronald Press, 1944. pp. 971-1032.

Huston, P.E.; Shakow, D.; andRiggs, L.A. Studies of motor func-tion in schizophrenia: II. Speed oftapping. Journal of General Psychol-ogy, 15:39-82, 1937.

Johnson, M.K. Discriminatingthe origin of information. In:Oltmanns, T.F., and Maher, B.A.,eds. Delusional Beliefs. New York,NY: John Wiley & Sons, 1988.pp. 34-65.

Kaney, S., and Bentall, R.P. Per-secutory delusions and attributionalstyle. British Journal of MedicalPsychology, 62:191-198, 1989.

Kaney, S.; Wolfenden, M.; Dewey,M.E.; and Bentall, R.P. Persecutorydelusions and recall of threateningpropositions. British Journal ofClinical Psychology, 31:85-87, 1991.

Kay, S.R.; Opler, L.A.; and Iinden-mayer, J.P. The Positive and Nega-tive Syndrome Scale (PANSS): Ra-tionale and standardization. BritishJournal of Psychiatry, 155(Suppl.7):59-65, 1989.

Knight, R.A. Relating cognitiveprocesses to symptoms: A strategyto counter methodological diffi-culties. In: Harvey, P.D., andWalker, E.E., eds. Positive andNegative Symptoms of Psychosis.Hillsdale, NJ: Erlbaum, 1987.pp. 1-29.

Kraepelin, E. Dementia Praecox andParaphrenia. (1919) Translated byR.M. Barclay. New York, NY:Robert E. Krieger Publishing Com-pany, Inc., 1971. pp. 282-329.

Lang, P.J., and Buss, A.H. Psycho-logical deficit in schizophrenia: HInterference and activation. Journalof Abnormal Psychology, 2:77-106,1965.

Lewine, R.J. A discriminant valid-ity study of negative symptomswith a special focus on depressionand antipsychotic medication.American Journal of Psychiatry,147:1463-1466, 1990.

Lezak, M.D. Neuropsychological As-sessment. 2nd ed. New York, NY:Oxford University Press, 1983.

Liddle, P.F. The symptoms ofchronic schizophrenia: A re-examination of the positive andnegative dichotomy. British Journalof Psychiatry, 151:145-151, 1987.

Lubow, R.E. Latent Inhibitions andConditioned Attention Theory. NewYork, NY: Cambridge UniversityPress, 1989.

Magaro, P.A. Cognition in Schizo-phrenia and Paranoia: The Integra-tion of Cognitive Processes. Hills-dale, NJ: Erlbaum, 1980.

Maher, B.A. Anomalous experienceand delusional thinking: The logicof explanations. In: Oltmanns, T.F.,and Maher, B.A., eds. DelusionalBeliefs. New York, NY: John Wiley& Sons, 1988. pp. 15-33.

Marneros, A.; Andreasen, N.C.;and Tsuang, M.T., eds. NegativeVersus Positive Symptoms. Berlin,Germany: Springer-Verlag, 1991.

McGlashan, T.H., and Fenton, W.S.The positive-negative distinction inschizophrenia: Review of naturalhistory validators. Archives of Gen-eral Psychiatry, 49:63-72, 1992.

Merriam, A.E.; Kay, S.R.; Weiner,R.U.; and Opler, L.A. Negativeschizophrenic symptoms are associ-ated with backward masking vul-nerability. Schizophrenia Research,3:355-356, 1990.



ownloaded from



Neale, J.M. Perceptual span inschizophrenia. Journal of AbnormalPsychology, 77:196-204, 1971.

Neale, J.M. Afterword. In: Harvey,P.D., and Walker, E.E., eds. Posi-tive and Negative Symptoms of Psy-chosis. HUlsdale, NJ: Erlbaum, 1987.pp. 289-293.

Neale, J.M. Defensive functions ofmanic episodes. In: Oltmanns, T.F.,and Maher, B.A., eds. DelusionalBeliefs. New York, NY: John Wiley& Sons, 1988. pp. 138-156.

Neale, J.M.; Oltmanns, T.F.; andHarvey, P.D. The need to relatecognitive deficits to specificbehavioral referents of schizo-phrenia. Schizophrenia Bulletin,11:286-291, 1985.Neufeld, R.W.J. Memory in para-noid schizophrenia. In: Magaro,P.A., ed. Cognitive Bases of MentalDisorders: The Annual Review ofPsychopathology. Vol. 1. NewburyPark, CA: Sage Publications, 1990.pp. 231-261.Nuechterlein, K.H.; Edell, W.S.;Norris, M.; and Dawson, M.E. At-tentional vulnerability indicators,thought disorder, and negativesymptoms. Schizophrenia Bulletin,12:408-426, 1986.Oltmanns, T.F., and Maher, B.A.,eds. Delusional Beliefs. New York,NY: John Wiley & Sons, 1988.Oltmanns, T.F., and Neale, J.M.Abstract conceptualization inschizophrenia: A fundamental psy-chological deficit? Perceptual andMotor Skills, 41:807-811, 1975.Overall, J.E., and Gorham, D.R.The Brief Psychiatric Rating Scale.Psychological Reports, 10:799-812,1962.

Persons, J. The advantages ofstudying psychological phenomenarather than psychiatric diagnosis.American Psychologist, 41:1252-1260,1986.

Pfohl, B., and Winokur, G. Theevolution of symptoms in institu-tionalized hebephrenic/catatonicschizophrenics. British Journal ofPsychiatry, 141:567-572, 1982.

Rodnick, E.H., and Shakow, D. Setin the schizophrenic as measuredby composite reaction time index.American Journal of Psychiatry,97:214-225, 1940.

Rosvold, H.E.; Mirsky, A.F.;Sarason, I.; Bransome, E.D., Jr.;and Beck, L.H. A continuous per-formance test of brain damage.Journal of Consulting Psychology,20:343-350, 1956.

Schooler, C, and Feldman, S.E.,eds. Experimental Studies of Schizo-phrenia. Goleta, CA: PsychonomicPress, 1964.

Schwartz-Place, E.J., and Gilmore,G.C. Perceptual organization inschizophrenia. Journal of AbnormalPsychology, 89:409-418, 1980.

Shakow, D. Segmental set: A the-ory of the formal psychologicaldeficit in schizophrenia. Archives ofGeneral Psychiatry, 6:1-17, 1962.

Shakow, D., and Huston, P.E.Studies of motor function inschizophrenia: I. Speed of tapping.Journal of General Psychology,15:63-108, 1936.

Silverman, J. The problem of atten-tion in research and theory inschizophrenia. Psychological Review,71:352-379, 1964.

Silverman, J. Variations in cogni-tive control and psychophysiologi-cal defense in the schizophrenias.Psychosomatic Medicine, 29:225-251,1967.

Silverman, J. Stimulus intensitymodulation and psychological dis-ease. Psychopharmacologia, 24:42-80,1972.

Slade, P.D., and Bentall, R.P. Sen-sory Deception: Towards a Scientific

Analysis of Hallucinations. London,England: Croom Helm, 1988.

Spitzer, R.L., and Endicott, J.Schedule for Affective Disorders andSchizophrenia—Lifetime Version(SADS-L). New York, NY: NewYork State Psychiatric Institute,1977.

Spitzer, R.L.; Endicott, J.; andRobins, E. Research DiagnosticCriteria: Rationale and reliability.Archives of General Psychiatry,35:773-782, 1978.

Steinhauer, S.R.; Gruzelier, J.H.;and Zubin, J., eds. Handbook ofSchizophrenia. Vol. 5. New York,NY: Elsevier, 1991.

Strauss, J.S. Processes of healingand chronicity in schizophrenia. In:Hafner, H.; Gattaz, W.F.; and Jan-zarik, W., eds. Search for theCauses of Schizophrenia. New York,NY: Springer-Verlag, 1987. pp. 75-87.

Strauss, J.S.; Carpenter, W.T., Jr.;and Bartko, J.J. An approach tothe diagnosis and understanding ofschizophrenia. Part HI: Speculationson the processes that underlieschizophrenic-symptoms and signs.Schizophrenia Bulletin, ^Experimen-tal Issue No. ll):61-69, 1974.

Strauss, M.E. Behavioral differencesbetween acute and chronic schizo-phrenics: Course of psychosis,effects of institutionalization orsampling biases? Psychological Bul-letin, 79:271-279, 1973.

Strauss, M.E., and Allred, L.J.Measurement of differential cogni-tive deficits after head injury. In:Levin, H.S; Eisenberg, H.M.; andGrafman, J., eds. NeurobehavioralRecovery From Head Injury. NewYork, NY: Oxford University Press,1987. pp. 88-107.

Strauss, M.E.; Buchanan, R.W.; andHale, J. Relations among attentiondeficits and clinical symptoms in



ownloaded from


VOL 19, NO. 2, 1993 231

schizophrenic outpatients. Psychia-try Research, in press.Stuss, D.T., and Benson, D.F. TheFrontal Lobes. New York, NY:Raven Press, 1986.Venables, P.H. Input dysfunctionin schizophrenia. In: Maher, B.A.,ed. Progress in Experimental Person-ality Research. New York, NY: Aca-demic Press, 1964. pp. 1-42.Wagman, A.M.I.; Heinrichs, D.W.;and Carpenter, W.T., Jr. Deficitand nondeficit forms of schizo-phrenia: Neuropsychologicalevaluation. Psychiatry Research,33:319-330, 1987.

Walker, E., and Harvey, P. Positiveand negative symptoms in schizo-phrenia: Attentional performancecorrelates. Psychopathology, 19:294-302, 1986.

Walker, E., and Lewine, R.J. Nega-tive symptom distinction in schizo-phrenia: Validity and etiologicalrelevance. Schizophrenia Research,1:315-328, 1988.

Westermeyer, J.F., and Harrow, M.Prognosis and outcome usingbroad (DMS-II) and narrow(DSM-IH) concepts of schizo-phrenia. Schizophrenia Bulletin,10:624-637, 1984.

Winters, K.C., and Neale, J.M. De-lusions and delusional thinking inpsychotics: A review of the litera-ture. Clinical Psychology Review,3:227-253, 1983.

Zubin, J., and Spring, B.Vulnerability—A new view ofschizophrenia. Journal of AbnormalPsychology, 86:103-126, 1977.

Acknowledgments

The author is grateful for the com-ments and suggestions of Drs. Al-thea M.I. Wagman and Rue Crom-well on an early draft, and alsothose of Dr. David Braff and Mr.Brian Carpenter.

The Author

Milton E. Strauss, Ph.D., is Pro-fessor, Departments of Psychology,Psychiatry, and Neurology, andDirector, Clinical Psychology Train-ing, Department of Psychology,Case Western Reserve University,Cleveland, OH.

Announcement The Kurt-Schneider Award is spon-sored by JANSSEN, GmbH, for ex-ceptional scientific achievements.The aim of the award is to en-courage psychiatric research,especially in the field of schizo-phrenia, including basic research(clinical psychopathology, bio-chemistry, physiology, pharmacol-ogy, genetics, epidemiology), diag-nostics, therapy, and rehabilitation.

Articles will be accepted ineither German or English andmust be received by September 15,1993. All entries (6 copies with a1-page abstract) are to be sent toProfessor Gerd Huber, Departmentof Nervous Diseases, UniversitatsNervenklinik und Poliklinik Psychi-atric D-5300 Bonn 1, Venusberg,Germany.



ownloaded from



Call for Papers Innovations & Research addressesa broad range of topics related toserious and persistent mental ill-ness for families, consumers, andprofessionals. The journal publishesinformation on medical and psy-chosocial rehabilitation research, in-novative service programs, medi-cation approaches, policy issues,professional training, self-help, self-report, and advocacy.

Innovations & Research is so-liciting manuscripts for upcomingissues from families, consumers,and professionals. Manuscriptsshould be submitted according tothe following guidelines:

• Send original and two copiesto the Executive Publisher.

• Include an abstract of a max-imum of 300 words.

• Place all references alpha-betically at the end of the article.

• Authors should provide abrief 2- to 3-line vita on a sepa-rate page, including name, degree,title, and position.

• Major articles should not ex-ceed 30 double-spaced typedpages.

• Follow American PsychologicalAssociation guidelines.

In addition to the above generalguidelines for authors, individualswho submit articles on innovativeprograms should describe the pro-grams in a way that will allowfamilies, consumers, or profes-sionals to implement the innova-tive program or some aspect of it.Innovation program articles shouldinclude discussion of administra-tive issues, policy concerns,budget/ financing, staffing andtraining, scope of the program,linkage to other programs, androle of families and consumers inprogram planning, implementation,and evaluation.

Send the manuscript copies andcover letter to:

LeRoy Spaniol, Ph.D.Executive PublisherInnovations and Research730 Commonwealth Ave.Boston, MA 02215Telephone: 617-353-3549



ownloaded from


to Cognitive Deficits in Schizophrenia€¦ · schizophrenia"; Crow 1985) is a variant of this approach. Most re-cently, interest emerged in the cognitive psychology of specific symptoms

Documents