Neuropsychological Impairmentsin Deficit vs
Nondeficit Forms ofSchizophrenia
Robert W. Buchanan, MD; Milton E.Strauss, PhD; BrianKirkpatrick, MD;
Constanze Holstein, MS;Alan Breier, MD; William T.Carpenter,Jr,MD
Background:Previous studies havesuggestedthat
functionalimpairmentsof the frontal andparietal
lobes are related to the deficitsymptomsof schizo-
phrenia.Thepurposeof the currentstudywas to
examine whetherneuropsychologicalmeasures of
frontal andparietallobe function differentiated deficit
from nondeficitpatients. Neuropsychologicalmea-
sures oftemporallobe function were used as contrast
Methods: Theperformanceof 18 deficit and 21 non$-
deficitschizophrenic patientswas examined on neuro-
psychologicalmeasures of executive, visuospatial,and
memory functions, selected on the basis of their asso-
ciation with lesions of either the frontal, parietal,
ortemporallobes. The results from the schizo-
phrenic subgroupswerecomparedwith the results
on the same measures obtained from 30 normal
deficitpatientson two frontal lobemeasures,theStroop
Color-Word InterferenceandTrailsMakingBtests, andone
parietallobemeasure,theMooneyFaces Closure Test. There
were no differences inperformanceon thetemporallobe
measuresbetween the twogroups.Bothgroupsperformed
morepoorlyon the tests than the normal controls.
Conclusion: The resultssuggestthatdeficitpatients may
logicalmeasures associated with frontal andparietalneu-
studies have demonstrated thetemporal
ityofthe deficit-nondeficit distinction6 and
shown deficitpatientsto havepoorer
premorbid adjustment,1increased social
tures,3and a diminishedresponseto meth-
ylphenidate hydrochloride(D. Mayerhoff,
MD,S.Alvir, MD,J. Lieberman, MD;un¬
also been shown to be moreneurologi-
cally impaired, especiallywithrespectto
neurological signssensitive toparietallobe
function1; exhibit increased volitional sac-
sensitive to frontal and/orparietallobe im-
validity ofusing pri¬
phreniahas been exam¬
ined in a number of studies.1"6 These
pairment7;and have reducedglucoseuti¬
lization as assessed withpositronemis¬
siontomographyin the frontal andparietal
cortices and thalamus.4
Theneurological signs, eye-
tracking,and results ofpositronemis¬
siontomographystudies have been con¬
sistent in theirimplicationof the frontal
andparietallobes in the neuroanatomi-
cal substrate of deficitsymptoms.The
possibleinvolvement of the frontal and
parietallobes in the neural substrate un¬
toms is furthersupported bythe results
have found arelationshipbetweennega-
From theDepartment of
University ofMaryland School
(Drs Buchanan, Kirkpatrick,
of Psychology,Case Western
Ohio (DrStrauss); and the
Branch, National Instituteof
Mental Health, Rockville, Md
Thirty-ninepatientsmeetingD5M-HI-R24 criteria for schizo¬
phreniawere selected from theMaryland PsychiatricRe¬
search CenterOutpatientClinic, Baltimore,forentryinto
thestudy.All available deficitpatients (N=18) were en¬
tered into thestudy.The nondeficitpatients (N=21) were
selected from the clinicpopulationon the basis oftheir simi¬
larityto the deficitpatientson thefollowing demographic
variables:age,race,and sex.Diagnoseswere madeusinga
best estimatediagnostic approachthat used all available in¬
formation from directassessment, familyinformants, and
pastmedical records. Patients with anorganicbrain dis¬
order,mentalretardation, ahistoryof severe headtrauma,
or ahistoryofdrugabuse/dependencewere excluded from
The 30 normal controls were selected from thegen¬
eralpopulation through neighborhood newspapernotices
andflyers.The controls did not have apastorcurrentDSM-
I1I-RAxis I or IIdisorder, as determinedbyresults of the
Structured Clinical Interview for DSM-III-R-Patient Ver¬
sion,25 or ahistoryoforganicbraindisorder, mental retar¬
dation, or severe head trauma.
Thepatientsand normal controlsgaveinformed con¬
sentpriortoparticipatingin thestudy.The normal con¬
trols were reimbursed for theirparticipationin thestudy.
Thepatientswerecategorizedinto deficit and nondeficit
subgroupson the basis of the consensusdiagnosisof two
of us (R.W.B. andB.K.), usingthe Schedule for the Defi¬
citSyndrome(SDS), a semistructured interview of docu¬
mentedreliabilityin thispopulation.26The SDSprovides
specific criteria forassessingthepresenceofnegative
symptoms,the duration of thesymptoms,and whether
interviewed whiletheywere betweenepisodes, during
periodsof clinicalstability.Additional information was
obtained from clinicians withlong-standingcontact with
thepatientsand fromfamilymembers. The BriefPsychi¬
atricRatingScale (BPRS) ThoughtDisorder Factor was
used to measure positive symptoms. Anticholinergic
doses wereexpressedinbenztropine equivalents,and
meanneurolepticdoses for eachgroupwere determined
usingthe conversionsystem developed bySchooler.27
Table 1,28"38 The tests are measures of executive, visuo-
spatial,andmemoryfunctions and were selected on the ba¬
sis oftheir association with lesions ofeither thefrontal,pa¬
rietal, ortemporallobes.20-3841Although performanceon
the tests selected for thisstudyisadverselyaffectedbyle-
sions of the stated brainregions, neuropsychologicaltest
performanceisdependenton the functionalintegrityofmul¬
tiplebrainstructures,andperformanceon these testsmay
be affectedbylesions of other brainregions.38'42There¬
fore,to increase the likelihood ofdetectingfunctional dis¬
turbances in aspecificcortical brainregionorsystemof
which it is apart, multipletests are used toprovidea com¬
prehensiveassessment of functions associated with each
of the cortical brainregions.
Theneuropsychologicaltests were administered ac¬
cordingto standardizedtesting procedures,with minimal
reinforcement and extraneous interaction to maximize the
similarityintestingsituations for eachsubject.The test bat¬
terywas administered to theschizophrenic patientswhile
theywereclinicallystable; patientshad beenjudgedto be
stable for a median of 18.0 weeks(range,1 to 121weeks).
Allpatientswerereceiving neurolepticsat the time of test¬
ing.The testbatterywas administered to the normal con¬
trols oncompletionoftheirdiagnosticevaluation. The tests
were administered in a fixed rather than random order to
ensure that more difficult tasks did not occur consecu¬
tively,and short tasks wereinterspersedbetweenlonger
ones.38Frequentbreaks wereallowed, and,ifnecessary,the
tests were administered over two test sessions. The total
time for the administration of the testbatterywas 3 hours.
Thesubject'sattentiveness, effort, and motivation for each
of the tests was assessed and coded asadequateor inad¬
equate bythe examiner. The assessment of theadequacy
of thesubjects' performanceson the tests was based on the
subjects'behavior and commentsduringthetestingses¬
sion and indicates the examiner'sjudgmentabout the de¬
greeto which thesubjects' performancesreflect their abili¬
ties. Test scores for whichperformancewas characterized
by inadequateattention, effort, and motivation, as deter¬
minedbythe examinerpriortoanalysesand blind to the
study hypothesesand SDScategorizations,were excluded
from theanalyses.Foursubjectshad testscores excluded:
two deficitpatients(fromTrailsMaking and BentonJudg¬
ment of Linestests)and two nondeficitpatients (Wiscon¬
sin CardSortingTest [WCST]and BentonJudgmentofLines
The variables used to assess testperformancearepre¬
sented in Table 1. The distribution of each variable was ex¬
amined to determineifitwasnormallydistributed. Theonly
variables that were notnormallydistributed were the total
timesrequiredtocompleteTrailsMakingA and tests.Log
transformation was used to normalize these scores.
Derived scores weredevelopedfor the Trails Mak¬
ingandStroopColor-Word Interference tests to have
measuresspecificallyfocused on the executive (frontal)
function construct we wished to measure.38 For the
TrailsMaking tests,the measure of interest is the influ¬
ence of set maintenance and set shiftrequirements
Continued on nextpage
(TrailsMaking Test)independentofgeneral psychomo-
Test). To evaluateimpairmentsin setfunctioning,we mea¬
sured the difference between eachsubject'sactual Trails
Making score and the TrailsMaking score estimated
time onlog-transformedTrailsMaking time. Theregres¬
sionequationwascomputed usingthe data from the nor¬
mal controls.Accordingly,the meandiscrepancyscore for
the normal controlgroupwas zero. Thehigherthe aver¬
age discrepancyscore was for the twoschizophrenicsub¬
groups,the slower the TrailsMaking time was than ex¬
pectedfrom the correlation ofTrailsMakingA and times
in normal controls. A similarapproachwas used to esti¬
mate the residual score for theStroopColor-Word Inter¬
ference Test measure. The residual score was the differ¬
ence between the observed Color-Word score and the score
predicted bythe Color-Word score with Color score re¬
gressionin normal controls.
Thefollowingmodification was made in theproce¬
dure used to assessperformanceon the BentonJudgment
of Lines Test. This test isusuallyscoredby countingthe
number ofcorrectlyidentified linepairs.To measure the
magnitudeof errors injudgment,as well as their fre¬
quency,we calculated the number ofpositions bywhich
thesubjectmisidentified the correct location of the lines.
This score canrangefrom 0 to 355.Weightedraw scores
were used for the individual WechslerMemoryScale-
Revised subscales.43 Scores on theFigurai Memoryand
Visual Pairs subscale were summed, on the basis of the
similarityof thememoryfunction assessedbyboth
scales, to maximize thesimilarityin the number of vari¬
ables used toassess frontal,parietal,andtemporallobe func¬
tion. The standard measures were used on each of the re¬
formed tocompare neuropsychologicaltestperformance
amongdeficitpatients,nondeficitpatients,and normal con-
trois. Allsubjectswithcompletedata sets were included
in the MANOVA.Analysesofvariance(ANOVAs) wereper¬
formed to assessgroupdifferences on the individual neu¬
ropsychologicaltests. Allsubjectswith valid scores for a
particulartest were included in the ANOVAs. Thesignifi¬
cance level was set at a=.05.
The ANOVAs wereperformedto determine if deficit
and nondeficitpatientsdiffered on medication (ie, anti-
cholinergicandneuroleptic dosages)and clinical vari¬
ables (ie,level ofpsychotic symptomsand duration of ill¬
ness) that couldpotentiallyconfound deficit-nondeficit
When differences in control variables wereobserved,analy¬
ses ofcovariance(ANCOVAs) wereperformedtoassess the
effect of each variable ongroupdifferences inneuropsy¬
Differences inintelligenceorgeneral ability may
also be associated with differences ingroup performance
onneuropsychologicalmeasures. Therefore, the Wech¬
sler AdultIntelligenceScale-Revised (WAIS-R) Vocabu¬
laryand PictureCompletionsubscale scores were used
to estimate GeneralAbility (WAIS-R GeneralAbility).35
These two measures were selected becausetheyare the
two WAIS-R verbal andperformance measures, respec¬
tively,with thehighestcorrelation withgeneral ability.38
An ANOVA wasperformedto assess if deficit and non-
deficitpatientsdiffered ingeneral ability, and, if differ¬
ences weresignificant, ANCOVAs wereperformedto
assess the effect ofgroupdifferences inintelligenceon
deficit-nondeficit differences inneuropsychologicaltest
performance.The summedage-corrected Vocabulary
and PictureCompletionsubscale scaled scores from the
WAIS-R were used as the covariate.
Prior toconducting anyANCOVAanalysis,the
relationshipbetween thedependentand control vari¬
ables was assessed in eachschizophrenic subgroupto
determine if theassumptionofhomogeneityofregres¬
sion was met.48
tivesymptomsand decreasedglucoseutilization8·9 or de¬
creasedregionalcerebralbloodflow (rCBF)lul1in the fron¬
tal lobes and betweennegative symptomsand decreased
rCBF in the inferiorparietalcortex.10 The frontal andpa¬
rietal cortices have extensivereciprocalinterconnec¬
tions1213 and extensive connections with the thalamus,
the other brainregion implicatedin thestudyofglucose
utilization.1415 The frontal andparietalcortices are also
connectedthroughtheir involvement in the basalganglia-
thalamocortical circuits describedbyAlexander16 and Al¬
exander and Crutcher.17 Each of theseparallel segre¬
gatedcircuits ishypothesizedto subserve a discreterange
of functions. We havehypothesizedthe involvement of
the dorsolateralprefrontalcircuit in theproductionof
deficitsymptoms,as it iscomposedof the brainregions,
ie, the dorsolateralprefrontaland inferiorparietalcor-
tices and the thalamus, implicatedin thepathophysi-
ologicalfeatures of the deficitsyndrome.1,3,4
In thepresent study,the functionalintegrityof the
frontal,parietal,andtemporallobes was assessed in defi¬
cit and nondeficitpatientswithneuropsychologicalmea¬
sures selected on the basis of theirsensitivityto lesions
of these areas. Toprovidea framework forinterpreting
the deficit-nondeficitcomparisons,the results from the
twoschizophrenic subgroupswerecomparedwith those
obtained from normal controls. On the basis of ourhy¬
pothesisofdorsolateralprefrontalcircuit involvement in
theproductionof deficitsymptoms,the deficitpatients
werehypothesizedtoperformmorepoorlyon frontal and
parietalmeasures than the nondeficitpatients.The neu¬
ropsychologicalmeasures sensitive totemporallobe le¬
sions were used as control tasks. The twogroupswere
*See the "Methods" section of the text for a more detaileddescription
*Determined bythe Hollingshead-Redlich rating of one to five, in which
1 is thehighestand 5 the lowest. Socioeconomic status could not be
determined for one deficitpatient (N=17)and was not available for three
heads of households(N=15)of deficitpatients.
1Based on a scale of zero to four, in which 0 indicates nodrug; 1,low
dosage;and4, high dosage.
hypothesizednot to differ in theirperformanceon these
tasks,becausetemporallobe abnormalities have been im¬
plicatedin theproductionofpositive symptoms,18and
bothpatient groupswere characterizedbythepresence
ofthesesymptoms.However, bothschizophrenic groups
wereexpectedtoperformless well on these measures than
the normal controls.19"21 Twodesignfeatures distin¬
guishedthe currentstudyfrompreviousstudies of the
neuropsychologyofnegative symptoms. First, the dif¬
ferentiation ofdeficitsymptomsfrom othertypesofnega¬
tivesymptomsrestricts thescopeofinvestigationto those
negative symptomsthat were mostlikelytorepresentin¬
trinsic trait features of theschizophrenicillness.22 Sec¬
ond,subjectswereprospectivelyselected on the basis of
their exhibition of thepsychopathologicalfeatures of in¬
terest, ie, deficitsymptoms,and theirsimilarityin de¬
mographicand clinical variables. These featuresopti¬
mized theabilitytointerpret studyresults in the context
Thedemographicandclinical characteristics of thestudy
samplearepresentedin Table 2. There were nosignifi¬
cant differences between the twoschizophrenicsub-
groupsandnormal controls inage,sex, or race distribu¬
tion or handedness. Deficitpatients, comparedwith
nomic status(SES) (F=8.12;d/=l,36;P<.01). However,
there was nosignificantdifference between the deficit and
nondeficitpatientswithrespectto head ofhousehold SES
in thefamilyoforigin (F=1.64; d/=l,34; P=.21).The defi¬
cit and nondeficitpatientshad similar numbers ofyears
ofeducation. The normal controls hadsignificantlymore
yearsof education andhigherSES than both the deficit
and nondeficitpatients,but there were nosignificantdif¬
ferences between head ofhousehold SESamongthe nor¬
mal controls and theschizophrenic subgroups.
Anticholinergicandneuroleptic dosages, lengthof
illness, andageat onset were similar between the two
schizophrenic subgroups.The deficitpatientshad less
severepsychotic symptomsthan the nondeficitpatients
at the time of theneuropsychological testing (F=8.88;
d/=l,37;P<.01).The test forhomogeneityofregression
revealed asignificant group byBPRS-Factor 1 interac¬
tion formultiple neuropsychologicalmeasures. The fail¬
ure to meet theassumptionofhomogeneityofregres¬
sion for all theneuropsychologicalmeasuresprecluded
The overall MANOVA for theneuropsychological
*P<.05, for deficit patients vs normal controls.
^Scoresof one nondeficitpatient were excluded.
<.05,for deficit vs nondeficitpatients.
WScoresof one deficitpatient were excluded.
^Subscales are from the WechslerMemoryScale-Revised.
testperformancedemonstrated asignificant groupbytest
interaction(n=65; F=3.07;ci/=22,104; P<.0001).The re¬
sult indicates that thepatternofdeficit, nondeficit, and
normal controlgroupdifferences varied across the tests.
Univariate tests wereperformedto assessspecific group
differences for each test.
Deficit vs Nondeficit
The mean±SD scores for the deficit and nondeficitpa¬
tients arepresentedinTable 3. The deficitpatients per¬
formedsignificantlymorepoorlyon two of the four ex¬
Interference(F=4.69; d/=l,37; P=.04)and the Trails Mak¬
ing tests (F=5.92; cij=l,36; P=.02)—andon one of the
Closure Test(F=7.14;df=l,37;P=.01).There were nosig¬
nificant differences between the twoschizophrenicsub¬
The deficitpatients, comparedwith the nondeficit
patients,hadsignificantlylower WAIS-R General Abil¬
ityscores(F=8.35; d/=l,36; P<.01).Because theregres¬
sions between theWAIS-R GeneralAbilityscores and neu¬
ropsychologicaltestperformancein bothgroupsdid not
differ, the GeneralAbilitymeasure was used as a covar-
iate to determine its influence on thegroupdifferences
on theneuropsychologicaltest measures. Thegeneral pat¬
tern ofdeficit-nondeficitgroupdifferences did notchange,
but themagnitudesofthe differences were reduced. The
differences on theStroopColor-Word Interference
(F=3.03;cJ/=l,35;P=.09)andMooneyFaces Closure tests
(F=3.80; d/=l,35; P=.06)were reduced to trend levels,
and the differences on the TrailsMaking Test(F=0.83;
d/=l,34; P=.36)were nolonger significant.
Schizophrenic Subgroupsvs Normal Controls
The mean±SD scores for the normal controls arepre¬
sented in Table 3. The deficitpatients performed signifi¬
cantlyworse than the normal controls on all the neuro¬
psychologicalmeasures. The samepattern appearswith
the nondeficitpatients, exceptthat there were nosig¬
nificant differences between nondeficitpatientsand nor¬
mal controls on the three tests (ie,StroopColor-Word
Interference, TrailsMakingB, andMooneyFaces Clo¬
sure) thatsignificantlydifferentiated deficit from non-
The overall MANOVA demonstrated asignificantinter¬
action betweengroupandneuropsychologicaltest, in¬
dicatingthat there were nonuniformgroup perfor¬
mance differences across the tests. Thefollow-up
univariateanalysesrevealed that deficit and nondeficit
patientsdiffered inperformanceson selected frontal and
parietalmeasures. Deficitpatients performed signifi¬
cantlymorepoorlyon two of the four frontal tests, the
StroopColor-Word Interference and TrailsMaking tests,
and on one of threeparietaltests, theMooneyFaces Clo¬
sure Test. There were nosignificant performancediffer¬
ences between the twoschizophrenic subgroupson the
other frontal orparietalmeasures. The deficitpatients
performedsignificantlymorepoorlythan the normal con¬
trols on all the frontal andparietallobe measures. The
differences between the nondeficitpatientsand normal
controls were not asnumerous;the twogroupsdid not
differ on the three frontal andparietal measures, ie, the
StroopColor-Word Interference, TrailsMakingB, and
MooneyFaces Closure tests, thatsignificantlydifferen¬
tiated the deficit from the nondeficitpatients.The re¬
sults are consistent with the apriori hypothesisthat defi¬
citpatients, comparedwith nondeficitpatients, are
characterizedby impairmentsin frontal andparietallobe
functions. Thesimilaritybetween the twogroupson mul¬
tipleclinical anddemographicvariables enhances the like¬
lihood that the observed deficit-nondeficitneuropsycho¬
logicaltestperformancedifferences are due to deficit
symptoms.The correlation of testperformancewith the
results of either structural or functionalimagingmea-
sures of the frontal andparietallobes arerequiredto
confirm the involvement of these brainregionsin the
chologicaltests sensitive to frontal lobedysfunction.Ex¬
amination ofthe studies that used some ofthe same frontal
measures used in the currentstudyreveals afairlycon¬
sistentpatternof results. In the currentstudy,we failed
to find aunique relationshipbetween the WCST or Ver¬
deficit and nondeficitpatientsbothperformedpoorlyon
the tasks. In three studies that have examined the con¬
negative symptoms,onestudyfound WCSTpersevera-
tive errors to besignificantlycorrelated with bothposi¬
tive andnegative symptoms,50and two studies were un¬
able to document asignificantrelationshipbetweenWCST
perseverativeerrors andnegative symptoms.56,57Simi¬
larly,Liddle and Morris56 found VerbalFluency perfor¬
mance to be associatedwithbothnegative symptomsand
thoughtdisorder,and Morrison-Stewart et al57 were un¬
able to demonstrate arelationshipbetween Verbal Flu¬
encyTestperformanceandnegative symptoms.In con¬
trast, Stolar et al34 observed VerbalFluencyTest
performanceto besignificantlyandinverselyrelated to
alogiaand affectiveflatteningand not related topositive
symptoms.The contrastwith thefindingsof the current
study maybe related to theiremphasisonalogiain de¬
finingtheirnegative symptomconstruct.OnlyBreier et
al53 have found auniquerelationshipbetweenWCSTper¬
formance andnegative symptoms. Theyalso found Ver¬
balFluencyTestperformanceto beuniquelyrelated to
negative symptoms.In thatstudy,thereportedrelation¬
shipswere between indexneuropsychologicaltestper¬
formance andfollow-up negative symptom ratings.Fi¬
nally,Liddle and Morris56 did not find an association
betweennegative symptomsand measures ofperfor¬
mance on theStroopColor-Word Interference and Trails
Making tests, similar measures to those used in the cur¬
major methodologicaldifferences,are consistent in their
failure to find aunique relationshipbetween WCST and
toms. There are less data withrespectto theperfor¬
mance on theStroopColor-Word Interference and Trails
Making tests, butonlythe currentstudywasdesigned
tospecificallyisolateprimaryandenduring negative symp¬
association of frontal lobe func¬
is consistent with thefindingsofprevi¬
tivesymptoms. Most,50"54 but not all,55"57
studies have demonstrated arelationshipbetweennega¬
There have beenrelativelyfew studiesexamining
ofparietallobe function andnegative symptoms,al¬
thoughGreen and Walker58 did findperformanceon the
PurduePegboardand BlockDesigntests related tonega¬
tivesymptoms.Inaddition, Liddle et al10reporteda re¬
lationshipbetween decreased rCBF in the inferiorpari¬
etal cortex andnegative symptoms.
There were nosignificantdeficit-nondeficit differ¬
ences onanyofthetemporallobe/memorymeasures. Both
schizophrenic subgroups performed,aspredicted,less
well than the normal controls on thetemporallobe tasks.
The observation ofimpaired memoryfunction in the two
schizophrenic subgroupsis consistent withprevious
studiesdocumenting memory impairmentsin schizo¬
phrenic patients.20,21,55Moreover, the sharedimpair¬
ments inmemoryfunctionsupport hypothesesof tem¬
porallobe involvement in theproductionofpsychotic
There are severalmethodologicalissues that re¬
quirecomment. First,althoughthe deficit and nondefi¬
citpatientswere similar onmultipleclinical and demo¬
graphic variables, the nondeficitpatientsdid have a
significantly higherlevel ofpsychotic symptomsat the
time ofneuropsychological testingthan the deficitpa¬
tients. The failure to meet theassumptionofhomoge¬
neityofregression precludedthe use of BPRS-Factor 1
as a covariate tostatisticallyassess the effect of the level
ofpsychotic symptomson deficit-nondeficit differences
differences in the level ofpsychotic symptomsunderlie
the deficit-nondeficit differences on theStroopColor-
Word Interference, TrailsMakingB,andMooneyFaces
Closure tests,aspsychotic symptomswouldbe morelikely
to interfere with testperformance. However, thepossi¬
bilityexists that thehigherlevel ofpsychotic symptoms
in the nondeficitpatients differentiallyinterfered with
theirperformanceon theneuropsychologicaltest mea¬
sures, leadingto the failure to find deficit-nondeficitper¬
formance differences on the otherneuropsychologicaltest
measures. The selection of deficit and nondeficitpa¬
tients withcomparablelevels ofpsychotic symptomsis
requiredto examine thispossibility.
Second,the addition oftheintelligence/generalabil¬
itymeasure as a covariate altered themagnitudeofdeficit-
nondeficitgroupdifferences on theStroopColor-Word
Interference, MooneyFacesClosure, and TrailsMaking
tests. An examination of the 2value,an estimation of
the effect size due to theexperimental manipulation (ie,
the deficit-nondeficitcategorization62), suggeststhat the
deficit-nondeficitgroupdifferences on theStroopColor-
Word Interference andMooneyFaces Closure tests, al¬
thoughreduced to a trendlevel,are nontrivial and remain
potentiallymeaningful.The effect sizes for the three tests,
ity,were. 14fortheStroopColor-Word InterferenceTest;
.09 fortheMooneyFaces ClosureTest;and. 12 fortheTrails
Making Test. After the covarianceadjustmentfor
intelligence/generalability,the effect sizes for theStroop
Color-Word Interference andMooneyFaces Closure tests
were .09 and.05,respectively.The effectsize for theTrails
Making Test was zero.Parenthetically,the effect sizes
for the threetemporallobe measures thatweregreaterthan
Pairssubscales,equaledzero after covarianceadjustment
forintelligence. By wayofcomparison,the 2 value for
the effect ofcigarette smokingon heart disease in the
Framingham studywas .01.63 Thiswas,ofcourse,statis¬
ticallysignificantbecause of thelarge samplesize in that
epidemiologicalstudy,whereas thepowerin thepresent
studywasmuch less becauseofthe smallsamplesizes,and
so ourfindingsaresuggestiverather than conclusive.
Understandingthe role ofintelligence/generalabil¬
ityin deficit-nondeficitperformancedifferences is an im¬
portantissue ininterpretingthepresent studyresults.
When thisstudywasdesigned, intelligence/generalabil¬
itywas construed as a "nuisance" variable thatrequired
control, ie, theintelligence/general abilitymeasure was
to control forgroup performancedifferences unrelated
to the deficit/nondeficitcategorization.Thismayhavebeen
an incorrectperspective.A number ofinvestigationshave
reported significant relationshipsbetweenintelligence
andnegative symptoms.50,34,64,65In addition, intelli¬
genceinschizophrenic patientshas also been related to
other features that differentiate deficit and nondeficitpa¬
tients, specifically neurological signsandpoor premor-
bidadjustment.66"68Further,SES has also been shown to
be correlatedwithintelligence,and,in the currentstudy,
deficitpatientshad a lower SES than nondeficitpa¬
tients.Thus,it is not clear thatintelligence/generalabil¬
ityshouldbeconceptualizedas a nuisance variable. The
ityand the deficitsyndrome suggeststhat itmaynot be
possibletodisentanglethe effects ofintelligencefrom more
specific impairmentson standardizedneuropsychologi¬
cal testsusingstatistical means.
Third, thepossibilityexists that the deficit-
nondeficit differences on the selected frontal andpari¬
etal measures are due to the differentialdiscriminating
powerof theneuropsychologicalmeasures and notspe¬
cific behavioral abnormalities.69Alternatively,discrimi¬
nating powerdifferences could also account for the fail¬
ure of all of the frontal andparietalmeasures to
differentiate deficit from nondeficitpatients.In the fu¬
ture, itmaybenecessarytodevelop psychometrically
matched tasksdesignedtotap specific neuropsychologi¬
cal functions of research interest toeffectivelydeal with
theproblemof differential deficit in studies concerned
with the identification ofneurocognitivefunctional dif¬
ferences withinschizophrenic subgroups.69
Inconclusion, the results of the currentstudy pro¬
toms areassociatedwith frontal andparietal neuropsycho¬
logical impairments.If theresults areconfirmed, impor¬
are related to theproductionofdeficitsymptoms.Future
studieswillattemptto elucidate thespecific cognitiveim¬
pairments occurringindeficitpatientsand combine theo¬
to examine morepreciselythe involvement ofthe frontal
andparietalcortices and othercomponentsofthe dorso¬
Acceptedfor publication June 23,1994.
Supported by grantsMH40279, MH45074, MH48225,
and MH00925fromthe National InstituteofMental Health,
Rockville, Md, andbythe National AllianceforResearch
inSchizophreniaandDepression, Chicago, III, (Drs
Reprint requeststoPsychiatricResearch Center, PO
Box21247, Baltimore,MD 21228(Dr Buchanan).
1. Buchanan RW, Kirkpatrick B, HeinrichsDW, CarpenterWT. Clinical correlates
of the deficitsyndromeofschizophrenia.Am JPsychiatry. 1990;147:290-294.
2. Kirkpatrick B, Buchanan RW. Anhedonia and the deficit syndrome of schizo-
phrenia. Psychiatry Res. 1989;31:25-30.
3. Thaker G, Kirkpatrick B, Buchanan RW, Ellsberry R,LahtiA, TammingaCA.
Oculomotor abnormalities and their clinical correlates in schizophrenia. Psy-
4. Tamminga CA, Thaker GK, Buchanan R, Kirkpatrick B, Alphs LD, Chase TN,
CarpenterWT. Limbicsystemabnormalities identified in schizophrenia using
positronemission tomographywith fluorodeoxyglucoseand neocortical alter-
ations with deficitsyndrome.Arch GenPsychiatry. 1992;49:522-530.
5. GurRE, Mozley PD,ResnickSM, LevickS,ErwinR, Saykin AJ, Gur RC. Re-
lations amongclinical scales in schizophrenia.Am J Psychiatry. 1991;148:
6. Fenton WS, McGlashan TH.Testing systemsfor assessment ofnegative symp-
toms in schizophrenia. Arch GenPsychiatry. 1992;49:179-184.
7. Pierrot-Deseilligny CH, Rivaud S, Gaymard G, Agid Y. Cortical control of re-
flexivevisually-guidedsaccades. Brain. 1991;114:1473-1485.
8. Volkow ND,WolfAP,Van GelderP, BrodieJD,OverallJE,CancroR,Gomez-
Mont F. Phenomenologicalcorrelates of metabolic activityin 18 patientswith
chronic schizophrenia. Am JPsychiatry. 1987;144:151-158.
9. WorlkinA, Sanfilipo M, Wolf AP, Angrist B, BrodieJD, Rotrosen J. Negative
symptomsand hypofrontality in chronic schizophrenia.Arch GenPsychiatry.
10. Liddle PF,FristonKJ,FrithCD,HirschSR,JonesT,Frackowiak RSJ. Patterns
of cerebral blood flow inschizophrenia. Br J Psychiatry. 1992;160:179-186.
11. Andreasen NC,RezaiK, Alliger R, SwayzeVW III, Flaum M, KirchnerP,Cohen
G, O'LearyDS.Hypofrontalityin neuroleptic-naive patientsand inpatientswith
chronic schizophrenia: assessment with Xenon 133 single-photon emission
computed tomographyand the Tower of London. Arch Gen Psychiatry. 1992;
12. Cavada C, Goldman-Rakic PS. Posterial parietal cortex in Rhesus monkey, I:
parcellation of areas based on distinctive limbic andsensorycorticocortical
connections. J CompNeurol. 1989;287:393-421.
13. CavadaC,Goldman-Rakic PS. Posteriorparietal cortex in Rhesusmonkey,II:
evidence for segregated corticocortical networks linking sensoryand limbic
areas with the frontal lobe. J CompNeurol. 1989;287:422-445.
14. Mesulam M-M. Principles of BehavioralNeurology. Philadelphia, Pa: FA Davis Download full-text
15. Fuster JM. The Prefrontal Cortex. NewYork, NY: Raven Press; 1989.
16. Alexander GE. Parallelorganization offunctionally segregated circuits linking
basal gangliaand cortex. Annu Rev Neurosci. 1986;9:357-381.
17. AlexanderGE, Crutcher MD. Functional architecture of basal gangliacircuits:
neural substrates ofparallel processing. Trends Neurosci. 1990;13:266-270.
18. Buchanan RW, Carpenter WT. Domains ofpsychopathology: conceptand va-
lidity. J Nerv Ment Dis. 1994;182:193-204.
19. Kolb B, Whishaw IQ. Performance ofschizophrenic patients on tests sensitive
to left or right frontal, temporal, or parietalfunction inneurological patients.
J Nerv Ment Dis. 1983;171:435-443.
20. GruzelierJ, Seymour K,Wilson L, Jolley A, Hirsch S. Impairments on neuro-
psychologictests oftemporohippocampalandfrontohippocampalfunctions and
wordfluencyinremitting schizophreniaand affective disorders. Arch GenPsy-
21. Saykin AJ, Gur RC, GurRE, Mozley D, Mozley LH, Resnick SM, Kester DB,
Stafiniak P. Neuropsychological function in schizophrenia: selective impair-
ment inmemoryand learning. Arch GenPsychiatry. 1991;48:618-624.
22. CarpenterWTJr,HeinrichsDW, WagmanAMI. Deficit and nondeficit forms of
schizophrenia:theconcept.Am J Psychiatry. 1988;145:578-583.
23. Carpenter WTJr, BuchananRW, Kirkpatrick B, Tamminga C,Wood F. Strong
inference, theory testing, and theneuroanatomyofschizophrenia. Arch Gen
24. AmericanPsychiatric Association,Committee on Nomenclature and Statistics.
Diagnosticand Statistical Manual of MentalDisorders, Revised Third Edition.
Washington, DC: American Psychiatric Association; 1987.
25. Spitzer RL, WilliamsJBW, Gibbon M, First MB. Structured Clinical Interview
for DSM-III-R\p=m-\PatientVersion(SCID-P, 5/1/89).NewYork,NY: Biometrics Re-
search Dept, New York State Psychiatric Institute; 1989.
26. Kirkpatrick B, Buchanan RW, McKenney PD, Alphs LD, Carpenter WT Jr. The
Schedule for the Deficit Syndrome: an instrument for research inschizophre-
nia. Psychiatry Res. 1989;30:119-123.
27. Schooler N. Antipsychotic medications in schizophrenia: their effect in acute
and maintenance treatment. In: CromwellRL,ed.Schizophrenia: Origins, Pro-
cesses, Treatment,and Outcome. NewYork,NY: OxfordUniversity Press;1993.
28. Heaton RK. Wisconsin CardSorting Test, Manual. Odessa, Fla: Psychological
Assessment Resources; 1981.
29. Taylor LB. Psychologicalassessment ofneurosurgical patients. In: Rasmus-
sen T, Marino R, eds. FunctionalNeurosurgery. NewYork, NY: Raven Press;
30. Milner B. Some effects of frontal lobectomyin man. In: WarrenJM,AkertK,
eds. The Frontal Granular Cortex and Behavior. New York, NY: McGraw-Hill
International Book Co; 1964:313-334.
31. Golden CJ.Stroop Color and Word Test: A Manual for Clinical andExperimen-
tal Uses. Chicago, III: Stoelting Co; 1978.
32. Reitan RM,Wolfson D. The Halstead-ReitanNeuropsychologicalTestBattery:
Theoryand ClinicalInterpretation. Tucson,Ariz:Neuropsychology Press;1985.
33. MooneyCM.Agein thedevelopmentof closureabilityin children. Can JPsy-
34. BentonAL, Hannay HJ, VarneyNR. Visualperception of line direction inpa-
tients with unilateral disease. Neurology. 1975;25:907-910.
35. Wechsler D. WechslerAdultIntelligence Scale\p=m-\Revised, Manual.Cleveland,Ohio:
ThePsychological Corp; 1981.
36. Wechsler D. A standardizedmemoryscale for clinical use. JPsychol. 1945;
37. Russell EW. Amultiple scoring method for assessment ofcomplex memory
functions. J Consult ClinPsychol. 1975;43:800-809.
38. Lezak MD. Neuropsychological Assessment. 2nd ed. New York, NY: Oxford
University Press; 1983.
39. Perrett E. The left frontal lobe of man and the suppression of habitual re-
sponsesin verbalcategoricalbehaviour. Neuropsychologia. 1974;12:323-330.
40. Fried I, MateerC, Ojemann G,WohnsR, Fedio P.Organization ofvisuospatial
functions in human cortex. Brain. 1982;105:349-371.
41. Goldberg TE, KelsoeJR, Weinberger DR, Pliskin NH, Kirwin PD, Berman KF.
Performance ofschizophrenic patientsonputative neuropsychological tests of
frontal lobe function. Int J Neurosci.1988;42:51-58.
42. Anderson SW, Damasio H,Jones RD,Tranel D. Wisconsin Card Sorting Test
performance as a measure of frontal lobedamage. J ClinExp Neuropsychol.
43. Wechsler D. WMS-R, WechslerMemory Scale\p=m-\Revised Manual. NewYork,
NY: The Psychological Corporation, Harcourt BraceJovanovich, Inc; 1987.
44. PerlickD, Stastny P, KatzI, Mayer M, Mattis S.Memorydeficits and anticho-
linergic levels in chronic schizophrenia. Am J Psychol. 1986;143:230-232.
45. StraussME, Reynolds KA, Jayaram G,Tune LE. Effects ofanticholinergicmedi-
cation onmemoryinschizophrenia. SchizophrRes. Inpress.
46. MedaliaA,GoldJ,Merriam A. The effects ofneuroleptics on neuropsychological
test results ofschizophrenics.Arch ClinNeuropsychol. 1988;3:249-271.
47. CassensG, Inglis AK, Appelbaum PS,Gutheil TG.Neuroleptics:effects on neu-
ropsychologicalfunction in chronicschizophrenic patients. SchizophrBull.1990;
48. Stevens J.AppliedMultivariate Statistics for the Social Sciences.Hillsdale,NJ:
Lawrence Erlbaum Associates; 1986.
49. Goldstein G.Comprehensive neuropsychological test batteries and research in
schizophrenia. In: SteinhauerSR, GruzelierJH,ZubinZ,eds. Handbook ofSchizo-
phrenia: Neuropsychology, Psychophysiologyand InformationProcessing.New
York, NY: Elsevier SciencePublishers; 1991;5:525-551.
50. Braff DL, Heaton R, KuckJ, Cullum M, MoranvilleJ, GrantI, Zisook S. The
generalized patternofneuropsychologicaldeficits in outpatientswith chronic
schizophreniawith heterogeneousWisconsin Card Sorting Test results. Arch
51. Guy JD, Liaboe GP, Wallace CJ. Premorbid adjustment in adult male schizo-
phrenics,as related toprocessvsreactive,chronic vs acute,ageofonset,and
neurologically impaired vs non-impaired. J ClinPsychol. 1986;42:62-67.
52. KemaliD, Maj M,GalderisiS, Monteleone P, Mucci A. Conditional associative
learning in drug-free schizophrenic patients. Neuropsychobiology. 1987;17:
53. BreierA, SchreiberJL, Dyer J, Pickar D. National Institute of Mental Health
longitudinal studyof chronicschizophrenia: prognosisand predictorsof out$-
come. Arch GenPsychiatry. 1990;48:239-246.
54. StolarN, Berenbaum H, Banich MT, Barch D. Neuropsychological correlates of
alogiaand affectiveflatteninginschizophrenia.BiolPsychiatry. 1994;35:164-172.
55. Goldberg TE, Ragland JD, Fuller-Torrey E, GoldJM, Bigelow LB, Weinberger
DR.Neuropsychological assessment ofmonozygotictwins discordant for schizo-
phrenia. Arch GenPsychiatry. 1990;47:1066-1072.
56. LiddlePF,Morris DL. Schizophrenic syndromes and frontal lobeperformance.
Br JPsychiatry. 1991;158:340-345.
57. Morrison-StewartSL,WilliamsonPC, Corning WC,KutcherSP,SnowWG,Mer-
skeyH. Frontal and non-frontal lobeneuropsychological test performanceand
clinicalsymptomatology in schizophrenia. PsycholMed. 1992;22:353-359.
58. Green M, Walker E. Neuropsychological performanceand positiveandnega-
tivesymptomsin schizophrenia. J AbnormPsychol. 1985;94:460-469.
59. WeinbergerDR. Implications of normal brain developmentfor the pathogen-
esis ofschizophrenia. Arch GenPsychiatry. 1987;44:660-669.
60. Barta PE, Pearlson GD, Powers RE, Richards SS, Tune LE. Auditory halluci-
nations and smallersuperior temporal gyral volume in schizophrenia. Am J
61. Cleghorn JM, Franco S, SzechtmanB, Kaplan RD,SzechtmanH, Brown GM,
NahmiasC, Garnett ES. Toward a brainmapofauditoryhallucinations. Am J
62. Keppel G. DesignandAnalysis:A Researcher's Handbook. Englewood Cliffs,
NJ: Prentice-Hall International Inc; 1973.
63. RosenthalR, Rosenow RL. Essentials of Behavioural Research: Methods and
DataAnalysis. NewYork, NY: McGraw Hill International BookCo; 1991.
64. Crow TJ. Thetwo-syndrome concept: origins and current status. Schizophr
65. Addington J, Addington D, Maticka-TyndaleE.Cognitive functioningandposi-
tive and negative symptomsin schizophrenia. SchizophrRes. 1991;5:123\x=req-\
66. MosherLR, PollinW,Stabenau JR. Identical twins discordant forschizophre-
nia. Arch GenPsychiatry. 1971;24:422-430.
67. ManschreckTC,Ames D.Neurologic features and psychopathology. BiolPsy-
68. Garmezy N. Process and reactive schizophrenia:someconceptions and is-
sues. In: Katz MM, Cole JO, Barton WE, eds. The Role andMethodologyof
Classification in Psychiatry and Psychopathology. Washington, DC: US
GovernmentPrinting Office; 1968. USDeptofHealth, Education,and Welfare
69. Chapman LJ, Chapman JP. The measurement of differential deficits. JPsy-
chiatr Res. 1978;14:303-311.