Neurocognitive Deficits and Functional Outcome

in Schizophrenia: Are We Measuring the

"Right Stuff"?
by Michael Foster Qreen, Robert S. Kern, David L. Braff, and Jim Mint?:

Abstract all of the studies in the review were published in 1990 or

later, which demonstrates the recency of interest in the rela-
There has been a surge of interest in the functional tionships of neurocognition to functional outcome.
consequences of neurocognitive deficits in schizophre- Nonetheless, Heaton and colleagues anticipated this line of
nia. The published literature in this area has doubled investigation (Heaton and Pendleton 1981) and examined
in the last few years. In this paper, we will attempt to the functional consequences of neurocognitive deficits in
confirm the conclusions from a previous review that mixed psychiatric and neurological patients.
certain neurocognitive domains (secondary verbal Reflecting the surge of interest in this area of inquiry,
memory, immediate memory, executive functioning as the cumulative published literature on neurocognitive and
measured by card sorting, and vigilance) are associ- functional outcome in schizophrenia has doubled in the
ated with functional outcome. In addition to surveying few years since the Green (1996) review. In this paper, we
the number of replicated Findings and tallying box will address several issues. First, with a substantially
scores of results, we will approach the review of the larger data base, we will attempt to confirm the conclu-
studies in a more thorough and empirical manner by sions from the previous review. Second, we will approach
applying a meta-analysis. Lastly, we will discuss what the review of the studies in a more thorough and empirical
we see as a key limitation of this literature, specifically, manner by applying a meta-analysis. Third, we will evalu-
the relatively narrow selection of predictor measures. ate the literature critically and discuss what we see as its
This limitation has constrained identification of medi- key limitation, specifically, the restricted selection of
ating variables that may explain the mechanisms for appropriate predictor measures.
these relationships.
Keywords: Schizophrenia, neurocognition, func-
tional outcome, social cognition, learning potential. Literature Review
Schizophrenia Bulletin, 26(1):119-136, 2000.
The relevant literature on neurocognitive deficits and
functional outcome in schizophrenia is summarized in
In a previous review of the literature, we concluded that spe- table 1. In this table, we selected peer-reviewed studies
cific domains of neurocognition were significantly related to with well-defined neurocognitive and functional outcome
functional outcome (Green 1996). The neurocognitive measures, and a primary interest in patients with schizo-
domains most consistently related to functional outcome phrenia or schizoaffective disorder, or both. We included
included secondary verbal memory, immediate or working both cross-sectional and longitudinal studies.
verbal memory, executive functioning measured with card The table includes 37 studies that are divided into
sorting, and vigilance. Functional outcome was divided into three functional outcome domains, described below. Some
the distinct domains of (1) community outcome, (2) social of the studies (n = 4) examined the relationships between
problem solving, and (3) psychosocial skill acquisition. neurocognitive constructs and two of the functional out-
Only 16 studies were included in the review, and the studies come domains. When this occurred, we listed the study
were generally underpowered and exploratory. Hence, our
confidence in the conclusions was not strong, and the need Reprint requests should be sent to Dr. M.F. Green, UCLA
Neuropsychiatric Institute, 760 Westwood Plaza, C9-420, Los Angeles,
for future studies in this area of research was clear. Almost CA 90024-1759.

119
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
by guest
Co
Table 1. Neurocognition and functional outcome in schizophrenia <•)

a-

on 20 February 2018
A. Studies of community outcome and daily activities I
s-
s
Outcome S'
Study Sample1 (Power 2 ) Predictor/correlative measures measures Major findings Co

Addington 80 S outpatients (0.78) Neurocognitive Social Functioning Mainly negative findings with Social Functioning
and Addington Battery including measures of Scale, Quality of Scale; poor cognitive flexibility associated with
1999 verbal ability, verbal and visual Life Scale low scores on Quality of Life Scale. to
memory, executive functioning, p\
visual-spatial organization, vigilance Z
o
and early information processing
Symptom
PANSS

Addington 30 S outpatients (0.37) Battery including measures of Social Dysfunction No significant associations between cognitive
etal. 1998 verbal ability, memory, executive Index, Social measures and social functioning.
functioning, visual-spatial ability, Adjustment Scale II
and attention/vigilance

Bartels 129SorSA Neurocoqnitive Community living MMSE correlated with activities of daily living and
etal. 1997 geriatric, nursing Clinical Dementia Rating Scale, skills and activities community functioning more strongly than symptoms.
O
home, and MMSE of daily living from
community (0.93) Symptom the Specific Level of
BPRS Function Scale

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697

Bellack et al. 22 vs. 84 S DS CPT, Span, WAIS-R subtests, Good vs. poor Secondary memory (verbal and visual), Stroop, card
1999 outpatients (0.54) verbal fluency, WCST, secondary vocational outcome sorting, verbal fluency, and IQ subtests correlated with
memory (verbal and visual), vocational outcome; variables that best classified GVO
Trails A and B, Stroop differed from those that best classified PVO.

Breier 58 S or SA Neurocoqnitive Levels of Symptoms, particularly negative symptoms, were related

etal. 1991 patients followed WCST, Trails A and B, Functioning Scale, to functional outcome; WCST, Trails A and B, and verbal
from index verbal fluency GAS fluency related to poor social functioning; positive
hospitalization (0.64) Symptom symptoms at index hospitalization significantly predicted
BPRS, SANS social and work function at followup, while
negative symptoms predicted work only.

o
by guest
on 20 February 2018
Outcome re
c
Study Sample1 (Power2) Predictor/correlative measures measures Major findings

Brekke 40 S or SA Neurocognitive Strauss Carpenter, Stroop, digit symbol, verbal fluency, and BD correlated IB!
etal. 1997 outpatients (0.48) Stroop, verbal fluency, WAIS-R role functioning with independent living and work. o
subtests (DS, DSym, BD, arithmetic) 3

Symptom CL

BPRS

Buchanan 29 treatment- WMS-R, verbal fluency, BD, visual- Level of functioning Secondary verbal memory, visual memory, and verbal §
etal. 1994 resistant S spatial measures, WCST, Trails, quality of life fluency predicted quality of life; visual memory EL
O
patients (0.36) Stroop predicted level of functioning.

Dickerson 88 S or SA Neurocoanitive Social Functioning Visual motor, spatial organization, and aphasia I
etal. 1996 outpatients (0.82) WAIS-R subtests, logical memory, Scale correlated with social functioning. Composite
Rey-O, WCST, Trails, aphasia, neuropsych correlated with social functioning
verbal fluency more strongly than composite symptoms.
Symptom
PANSS

Goldman 19 S patients Neurocognitive Strauss Carpenter Secondary verbal memory predicted community
etal. 1993 (0.24) BD, Trails, vocabulary, selective Scale functioning; symptoms did not predict outcome.
reminding, DS
Symptom
Psychotic from BPRS, negative
from SANS

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697

Harvey et al. 208 elderly S MMSE, CERAD Cognitive Battery3 SAFE Composite measure of cognitive functioning correlated
1997 inpatients (0.99) significantly with all 17 items of the SAFE; regression
analysis using three SAFE factor scores indicated that
cognitive impairment was more strongly correlated with
skills deficits than behavioral undercontrol.

Harvey et al. Sample 1:97 Neurocognitive SAFE For each sample, composite cognitive score correlated a-
1998 chronic MMSE, CERAD Cognitive Battery3 with social-adaptive functioning more strongly than a
S'
inpatients (0.85) Symptom positive or negative symptoms.
Sample 2: 37 PANSS
TO
nursing home (0.45) S'
Sample 3: 31
acute inpatients (0.38)
All geriatrics

I
by guest
Table 1. Neurocognition and functional outcome in schizophrenia—Continued to
3-

on 20 February 2018
Outcome f
Study Sample1 (Power2) Predictor/correlative measures measures Major findings
I
Heslegrave 42 stable Neurocognitive Quality of life Span/CPT, iconic memory, positive, negative, and S'
etal. 1997 outpatients (0.50) COGLAB measured by the general symptoms correlated with the Sickness c
Symptom Sickness Impact Impact Profile.
PANSS Profile, two global Negative and general symptoms, but not neurocogni-
subjective items tion, correlated with the subjective items. £
Jaeger and 19 first episode WCST Social Adjustment Perseveration predicted social adjustment.
o
Douglas 1992 S patients (0.24) Scale
Johnstone 137 first episode Neurocognitive Occupational Neurocognitive measures did not predict outcome.
et al. 1990 S patients (0.95) PPVT, DSym functioning Psychotic behaviors did not predict occupational
Symptom outcome, but ratings of social withdrawal (negative
Behavioral ratings symptom) did.
Lysaker 89 S and SA WCST Occupational Card sorting predicted task orientation and
etal. 1995 patients (0.82) functioning, WPP social skills at work.
Meltzer Sample 1:48 Neurocognitive GAS, Quality of Life WCST correlated with GAS; mazes and VLL correlated
etal. 1996 treatment- WCST, FAS, verbal list learning, Scale, work status with both GAS and Quality of Life Scale. All measures
resistant patients mazes, DSym discriminated work outcome in separate sample
with S (0.55) Symptom (sample 2)—WCST was best.
Sample 2:82 BPRS
S patients (0.79)

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697

Mueser 28 S or SA Emotion and face recognition Social Behavior Facial emotion identification, discrimination, and
etal. 1996 inpatients (0.35) Scale recognition were associated with social competence
as measured by the Social Behavior Scale.
Penn et al. 27 S or SA Neurocognitive NOSIE-30 Card sorting and social cognition were related to
1996 chronic COGLAB, affect recognition, adaptive unit behavior. Relationships with social
inpatients (0.34) empathy, social scripts cognition (especially affect recognition) were
Symptom generally stronger than with COGLAB.
PANSS
Wykes et al. 28 S patients (0.35) Neurocognitive Independent living Complex RT predicted degree of independent
1990 Complex RT living; symptoms did not.
Symptom
Psychotic, negative items from
Present State Examination
by guest
on 20 February 2018
B. Studies of laboratory assessment of instrumental skills and social problem solving
I
o
Outcome
Study Sample1 (Power2) Predictor/correlative measures measures Major findings 90

Addington and 80 S outpatients (0.78) Battery including measures AIPSS Verbal ability, verbal memory, and poor cognitive flexibility I
Addington of verbal ability, verbal and were associated with AIPSS; visual-spatial ability and
1999 visual memory, executive vigilance were associated with the sending skills subscale.
functioning, visual-spatial
organization, vigilance, and
early information processing 9
o
Addington 30 S outpatients (0.37) Battery including measures AIPSS CPT, a measure of attention, was a significant predictor
etal. 1998 of verbal ability, memory, of processing and sending skills as assessed by the
I
executive functioning, visual- AIPSS.
spatial ability, and attention

Bellack 27 S inpatients (0.34) Neurocognitive Social Problem Verbal IQ, secondary memory, negative symptoms
etal. 1994 Verbal IQ, WMS-R Solving Assessment correlated with some outcome measures. Immediate
Symptom Battery memory, psychotic symptoms did not.
Psychotic from BPRS,
negative from SANS
to
Bowen 30 S inpatients (0.37) CPT, Span, DSDT AIPSS Vigilance correlated with overall problem solving;
etal. 1994 immediate verbal memory, early visual processing did not.

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697

Corrigan and 26 S or SA Neurocognitive AIPSS Social cue recognition and verbal memory were related
Toomey 1995 chronic CPT, RAVLT, WCST, DSDT, to AIPSS. Social cue recognition was more strongly
inpatients (0.33) social perception associated than most neurocognitive measures, and
Symptom this correlation held after controlling for positive
BPRS symptoms.

Corrigan 26 S inpatients (0.33) Neurocognitive Social Cue Immediate verbal memory, secondary verbal memory, a-
etal. 1994 CPT, Span, RAVLT, DSDT, WCST Recognition Test early visual processing, negative symptoms correlated
Symptom with social cue recognition summary score; vigilance,
Co
Psychotic and negative card sorting, psychotic symptoms did not.
from BPRS 5"
S'
Ikebuchi 20 S inpatients Neurocognitive Structured role-play Performance IQ was significantly correlated with receiving
etal. 1996 and outpatients (0.25) WAIS-R and processing skills and global score on role-play test. to
Symptom
BPRS
by guest
Table 1. Neurocognition and functional outcome in schizophrenia—Continued

on 20 February 2018
Outcome
Study Sample1 (Power2) Predictor/correlative measures measures Major findings
f
a
S'
Klapow 55 S older Neurocoqnitive DAFS MMSE was only significant predictor, even to
c
etal. 1997 outpatients (0.62) MMSE - simulated daily after controlling for symptom assessments
Symptom activities and demographics.
BPRS, SANS, SAPS

Mueser Male = 20 (0.25) WMS-R, visual-spatial, Standardized WMS-R (mainly visual) with omnibus test related
etal. 1995 Female = 18 (0.23) WAIS-R, FAS, language role-play to outcome for females, but not for males.
O
SorSA

Mueser 28 S or SA Emotion and face recognition Unstructured role- Facial emotion identification was associated with
etal. 1996 inpatients (0.35) play (Conversation nonverbal performance on the Conversation Probe.
Probe)

Penn et al. 31 Dutch S COGLAB4 Means-Ends Generally negative results

1993 inpatients (0.38) Problem-Solving
Test, alternative
solution generation
test

Penn et al. 38 S and SA COGLAB4 Role-playing Vigilance, RT correlated with role-playing.

1995 inpatients (0.46)

Penn et al. Male = 21 (0.27) COGLAB4 Unstructured role- RT, vigilance, masking, card sorting related to

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697

1996 Female = 17 (0.22) play outcome for females, not males. Results held
S or SA chronic after controlling for positive symptoms.
inpatients

Velligan Study 1:112(0.92) Neurocognitive Functional Needs Cognitive factor was the only one to enter the
etal. 1997 Study 2: 41 (0.49) Study 1: Mental status exam Assessment regression analysis. In path analysis, symptoms
chronic from Negative Symptoms were not needed. Same for study 2.
inpatients Assessment
Study 2: global from
neuropsych battery
Symptom
BPRS •

o
a
e.
by guest
C. Studies of psychosocial rehabilitation and skill acquisition

on 20 February 2018
I
Outcome
Study Sample1 (Power2) Predictor/correlative measures measures Major findings
II
Bowen et al. 30 S inpatients (0.37) CPT, Span, DSDT Single session skill Vigilance, early visual processing, and immediate
1994 acquisition verbal memory correlated with total skill acquisition.

Corrigan 30 S inpatients (0.37) Neurocoanitive Single session skill Immediate and secondary verbal memory, vigilance
§
etal. 1994 CPT, Span, RAVLT, DSDT, acquisition correlated with overall skill acquisition; card sorting,
WCST early visual processing, symptoms did not. O
Symptom
Psychotic and negative from
BPRS

Kern et al. 16 psychotic CPT, Span, RAVLT, WCST, Pre- and posttests, Secondary verbal memory correlated with pretest;
1992 inpatients, PPVT, Rey-O, Pin Test, on-task behaviors immediate and secondary verbal memory correlated
mainly S (0.21) Backward Masking during training with on-task behaviors; vigilance correlated with
change score.

Lysaker 53 S and SA Neurocognitive Work-related social Card sorting, proverb interpretation predicted
etal. 1995 inpatients and WCST, proverbs skills, WPP improvement in social skills; symptoms did not.
to
outpatients (0.60) Symptom
Psychotic and negative
from the PANSS

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697

McKee 19SorSA Neurocognitive Program attendanc Executive function (Stroop) was positively associated with
etal. 1997 inpatients (0.24) Symbol Digit Modalities Test, and overall level of level of participation; Symbol Digit Modalities Test,
RAVLT, WISC-III mazes, FAS, participation in WISC-III mazes, and negative symptoms were
Category Instance Generation Community Re- negatively correlated with program attendance.
Test, Stroop Color-Word Test Entry Program
Symptom
SAPS, SANS f
Mueser et al. 30 S and SA Neurocognitive Change scores from Immediate verbal, secondary verbal, and visual
1991 inpatients and WMS skills training memory predicted skills acquisition; symptoms to
outpatients (0.37) Symptom program did not.
Psychotic and negative from S'
BPRS
by guest
Table 1. Neurocognition and functional outcome in schizophrenia—Continued

on 20 February 2018
N
Predictor/correlative Outcome
1 2 a-
Study Sample (Power ) measures measures Major findings
a
5'
Silverstein 26 chronic S Perceptual organization, RT, MMLT, Verbal memory, verbal fluency, card sorting, and to
etal. 1998 inpatients (0.33) DS CPT, Sustained Attention observational inferential reasoning were correlated with the MMLT;
Test, RAVLT, FAS, Theory of ratings of 28 ward cognitive variables of RT and sustained attention were a'
Mind, WCST behaviors, hospital related to on-time behavior, and RT was related to
discharge keeping room clean; perceptual organization was
related to hospital discharge.

Weaver and 248 S inpatients Psychomotor tests Rehabilitation Psychomotor speed predicted patients selected o
Brooks 1964 and outpatients (including RT, dexterity, and potential for rehabilitation programs.
(0.99) motor learning)

Note.—AIPSS = Assessment of Problem Solving Skill; BD = block design; BPRS = Brief Psychiatric Rating Scale; CERAD = Consortium to Establish a Registry for Alzheimer's
Disease; CPT = Continuous Performance Test; DAFS = Direct Assessment of Functional Status Scale; DS = digit span; DSDT = Digit Span Distractibility Test; DSym = Digit Symbol
Substitution Test; FAS = a test of verbal fluency; GAS = Global Assessment Scale; GVO = good vocational outcome; MMLT = Micro Module Learning Test; MMSE = Mini Mental
State Examination; NOSIE-30 = Nurse's Observation Scale for Inpatient Evaluation; PANSS = Positive and Negative Syndrome Scale; POV = poor vocational outcome; PPVT =
Peabody Picture Vocabulary Test; RAVLT = Rey Auditory Verbal Learning Test; Rey-0 = Rey-Osterrieth Complex Figure Test; RT = reaction time; S = schizophrenia; SA = schizoaf-
fective disorder; SAFE = Social-Adaptive Functioning Evaluation; SAPS = Scale for the Assessment of Positive Symptoms; SANS = Scale for the Assessment of Negative
Symptoms; Span = Span of Apprehension; WAIS-R = Wechsler Adult Intelligence Scale-Revised; WCST = Wisconsin Card Sorting Test; WISC—III = Wechsler Intelligence Scale for
Children III; WMS = Wechsler Memory Scale; WMS-R = Wechsler Memory Scale-Revised; WPP = Work Personality Profile.
1
as Sample size of the analyses relevant to the current review.
Statistical power for a medium effect size of r= 0.30 with an alpha value of 0.05, two-tailed.
3
Consists of word list learning and delayed recall, praxic drawings, modified Boston Naming Test, and category fluency.
4

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697

A computer-based battery of cognitive tests, including vigilance, card sorting, masking, and RT.

£L
 Neurocognition and Functional Outcome Schizophrenia Bulletin, Vol. 26, No. 1,2000

twice, once in each outcome domain. The remaining stud- selected target stimuli and ignore all others (Davies and
ies considered only one area of functional outcome. Parasuraman 1982).
Hence, the table includes a total of 41 findings for the 37
separate studies. Executive Functioning/Card Sorting. The broader term
"executive functioning" refers to volition, planning, pur-
posive action, and self-monitoring of behavior. In this
Types of Neurocognitive Constructs group of studies, card sorting tests were the most common
way to assess executive functioning. Card sorting mea-
When starting to review this literature, one immediately
sures, such as the Wisconsin Card Sorting Test (WCST),
confronts the variety of both predictor and outcome mea-
were used in which the subject matches a stimulus card to
sures. There is considerable diversity in the selection of
one of four key cards according to certain matching prin-
neurocognitive measures. However, most of the measures
ciples (Heaton 1981). These tests assess the subject's abil-
can be clustered into a smaller number of key neurocogni-
ity to attain, maintain, and shift cognitive set.
tive constructs.
Verbal Fluency. These tests measure one's ability to gen-
Secondary Memory. There are many ways to subtype
erate words. Subjects may be asked to produce words that
memory and a large variety of assessment methods. Two
begin with a certain letter, or to produce words from a cer-
types of memory are especially relevant to the current
tain semantic category (e.g., animals). The term "fluency"
topic, namely, immediate memory and secondary mem-
is slightly misleading because it refers to the number of
ory. Secondary memory refers to the ability to acquire and
correct items generated, not whether the subject speaks
store information over a longer period of time (usually
fluently (Benton and Hamsher 1978).
lasting for several minutes and longer). This type of mem-
ory is usually assessed by asking individuals to attempt to
Early Visual Processing. Measures of early visual pro-
learn a list of words or recall passages of text (e.g., Delis
cessing evaluate the basic stages of visual processing,
et al. 1987). The amount of information to be remembered
such as the visual scanning of a display of stimuli and the
exceeds the immediate memory span, meaning that it con-
early detection and identification of visual stimuli.
tains too much information to be held "on-line" at any one
Assessments of these processes involve very brief, tachis-
time. This is distinct from remote memory because assess-
toscopic presentation of visual stimuli on a monitor, either
ments of secondary memory typically use delay periods of
alone (e.g., backward masking) or in the presence of com-
less than 1 hour, whereas remote memory typically refers
peting stimuli (Asamow et al. 1991).
to retention over days or years.
Psychomotor Skills. Assessments of psychomotor abili-
Immediate Memory. Immediate memory refers to the
ties are usually speeded tests and can be separated into
ability to hold a limited amount of information "on-line"
two types of skills: speed and dexterity. Motor speed is
for a brief period of time (usually a few seconds).
measured with rapid, repetitive finger movements.
Repeating a telephone number is an example of immedi-
Dexterity is assessed with tasks that involve fine manual
ate memory. Immediate memory (also called the phono-
manipulation (e.g., with pegs or pins). We have also
logical loop) is considered to be a component of "work-
included in this category assessments of reaction time in
ing" memory (Baddeley 1986). We use the term
which the subject responds as quickly as possible to a
"immediate memory" as opposed to working memory in
stimulus by pressing a button (Lezak 1995).
this article because almost all of the studies in this review
used measures requiring only brief storage of information,
not the ability to manipulate the information, which is Types of Functional Outcome
common in tests of working memory.
The outcome measures fit, more or less naturally, into
Vigilance. Also sometimes called "sustained attention," three general categories: (1) success in psychosocial reha-
vigilance refers to the ability to maintain a readiness to bilitation programs, (2) studies of laboratory assessment
respond to signal (i.e., target stimuli) and not respond to of social problem solving ability or analog measures of
noise (i.e., nontargets) over a period of time. In other instrumental skills, and (3) studies that have considered
words, it involves an ability, called "sensitivity," to distin- broader aspects of behavior in community outcome and
guish signal from noise. Vigilance is typically measured activities of daily living. These three outcome domains
with a Continuous Performance Test in which subjects are are similar, but not identical, to the outcome domains used
presented with a series of briefly presented stimuli on a in our previous review of the literature (Green 1996). The
computer screen and instructed to respond only to boundaries are not absolute. The first two areas are highly

127
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Schizophrenia Bulletin, Vol. 26, No. 1, 2000 M.F. Green et al.

interconnected because performance of skills obviously studies serve different functions. The first group provides
depends on acquiring those skills in the first place. We information about the overall magnitude of the relation-
also believe that the community outcome and activities of ships across measures, whereas the other group suggests
daily living are heavily dependent on skill acquisition. specific neurocognitive-functional connections.
The intention was to distinguish the intermediate cate-
gory, which includes acquisition and performance of iso-
lated skills, from measures that involve integration of Studies That Have Used
multiple skill areas. The categories also differ because the Global/Composite Measures of
intermediate cluster includes laboratory analogs of social Neurocognition
and work functioning, whereas the outcome in the com-
munity-based category is actual, not simulated. A brief As mentioned, some studies have used global or compos-
description of key outcome categories follows. ite measures of neurocognition. The advantage of these
measures is that they can provide an estimate of the total
Success in Psychosocial Skill Acquisition. Psychosocial amount of variance in functional outcome that can be
skills training is a widely used method of psychosocial treat- explained by neurocognition in general. The results of
ment for patients with schizophrenia. Psychosocial rehabili- these studies suggest that somewhere between 20 and 60
tation programs teach patients basic life skills (e.g., basic percent of the variance in outcome can be explained by
conversation skills, symptom and medication management, neurocognition. Consider these examples:
and leisure skills) and are designed to provide patients with 1. Using a sophisticated path analysis, Velligan and
greater functional independence. These training programs colleagues (1997) tested the pathways between positive
tend to be highly structured, and progress is closely moni- symptoms, negative symptoms, cognition, and activities
tored. It is possible to measure the amount of success in a of daily living in two separate samples. A global measure
skills training program, for example, by the degree of skill of cognition accounted for 48 percent and 42 percent of
acquisition (e.g., Mueser et al. 1991; Bowen et al. 1994). the variance in the activities of daily living for the first
and second samples, respectively. It is important to note
Laboratory Assessments of Instrumental Skills and that this study found that when the pathway from cogni-
Social Problem-Solving Ability. Studies in this category tive impairment to functional outcome was in the model,
examine performance of skills, mainly with laboratory direct pathways from psychotic and negative symptoms to
analog measures of social competence or social problem functional outcome were not needed. Their results provide
solving. In a typical assessment, subjects may watch rather strong support for the theory that cognitive impair-
videotaped vignettes that present an interpersonal prob- ment, rather than symptoms, influences functional out-
lem. Subjects may be asked to identify the problem, sug- come.
gest solutions to the problem, and demonstrate how they 2. Another study by Harvey et al. (1998) considered
would act out the solution through role-play (e.g., Bellack three separate groups of elderly schizophrenia patients
et al. 1994). that differed substantially in level of adaptive functioning.
One group from a nursing home had low levels of adap-
Community Outcome/Daily Activities. This is the most tive functioning, an acute group had relatively high levels,
varied category of functional outcome and includes such and a chronic group was intermediate. However, absolute
outcomes as occupational functioning, social attainment, levels of adaptive functioning did not affect the pattern of
and degree of independent living. This outcome area is correlations, which were the same for all groups. In each
based more on self-report than on demonstration but can group a composite measure of cognition correlated most
also be rated from hospital charts or caregivers' reports. strongly with adaptive functioning, explaining about 40 to
Assessments may include the activities of daily living, 50 percent of the variance. Slightly lower correlations
amount or level of work or school, and type and quality of were uncovered for negative symptoms, and positive
social support networks (e.g., Brekke et al. 1997). symptoms were essentially uncorrelated.
Table 1 includes two types of studies, and it will be
helpful to identify them up front. Some of the studies used
global measures of neurocognition, or composite mea- Studies That Examined Specific
sures in which performance across a battery is integrated Neurocognitive Constructs
into a single summary measure. In contrast, most of the
studies examined specific associations between particular Given the overall findings that global or composite indica-
neurocognitive constructs (e.g., immediate memory or tors of neurocognition are related to functional outcome,
vigilance) and functional outcome. These two types of the question shifts to whether we can identify specific

128
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Neurocognition and Functional Outcome Schizophrenia Bulletin, Vol. 26, No. 1, 2000

domains of neurocognition that account for the relation- Table 2. Box scores
ships. We can survey the literature three different ways:
(1) by the number of replicated findings, (2) through box Domain Positive Null Total
scores, and (3) through meta-analysis. Figure 1 depicts the Secondary verbal memory 13 5 18
replicated findings. In the figure, the neurocognitive con-
Immediate verbal memory 5 2 7
structs are placed on the left, and the three functional out-
come domains on the right. Associations are shown by Card sorting 11 11 22
two types of arrows that represent the number of replica- Vigilance 9 8 17
tions. A heavy arrow indicates that at least four separate
studies found a significant relationship between the neu-
Table 2 shows the box scores for some of the key
rocognitive construct and the outcome domain. The
neurocognitive constructs that were implicated in both the
smaller arrows indicate that two or three studies reported
previous review and the current one. The table shows the
a significant relationship.
number of positive findings compared with the number of
The figure reveals a large number of replicated find-
null findings. A complete box score should also show
ings that were reported in four or more studies. Secondary
"paradoxical" findings, that is, findings that are significant
verbal memory was reliably related to every outcome
and in the opposite direction. However, we did not
domain, and immediate memory was related to psychoso-
uncover any paradoxical relationships, an observation that
cial skill acquisition. Card sorting and verbal fluency were
is notable by itself because it reflects the rather consistent
both associated with community outcomes, and vigilance
direction of relationships.
was linked to skill performance.
The box scores are consistent with the replicated
Examining the replicated findings is partially, not
findings. For example, 13 out of 18 studies reported a sig-
fully, informative about the consistency of the findings.
nificant relationship between secondary verbal memory
The number of replicated findings tells us how many
and functional outcome, and 5 out of 7 reported signifi-
times a significant association was found, but it does not
cant associations between immediate memory and func-
tell us how many times an association was looked for. It is
tional outcome. The number of positive findings is
somewhat like knowing the number of hits that a baseball
impressive in light of the rather low statistical power of
player has without knowing his batting average. The
most of these studies (see table 1 for listing of power).
equivalent of the batting average for reviewing a literature
While the box scores are more informative than number
is the box score.
of replications alone, they have limitations. For one, in a
box score tally, all studies get one vote. However, there is
Figure 1. Neurocognitive prediction of functional
outcome wide variation in the number of subjects across the stud-
ies, and it would be helpful if the studies could be
Neurocognition Outcome
weighted accordingly. Another limitation is that box
Card Sorting Community / scores do not provide a clear sense of the magnitude of
1 jj Dally Activities
the effect. Knowing that secondary verbal memory is sig-
Secondary Verbal
Memory
nificant in 13 out of 18 studies does not tell us if the effect
Social Problem
Solving I
size is small, medium, or large. For these kinds of ques-
Immediate Verbal Instrumental Skills tions, a meta-analysis is valuable.
Memory
Results were summarized across studies for four of
PeychosocuH
Skill the key neurocognitive constructs (secondary verbal
Vigilance
Acquisition
memory, immediate memory, executive functioning/card
sorting, and vigilance) using standard meta-analytic pro-
Neurocognition Outcome
cedures for combining correlation coefficients (Hedges
Verbal Fluency
Community / and Olkin 1985). In almost all cases, study results were
Dally Activities

*. •*"
presented originally in the form of correlation coefficients
Psychomotnr Ability and
(Pearson r). In a very few, other statistics (e.g., t) were
Social PraMem
Reaction Time Solving ( converted into the equivalent value of r for meta-analysis.
Instrumental Skills
When values were not presented, we presumed them to be
nonsignificant and estimated them by a value halfway
Early Visual Processing Psychosocial \
sun I between 0.00 and the lowest possible significant correla-
Acquisition 1
tion based on the particular sample size. Results in the
f. 2-3stuofes | different outcome domains were treated and analyzed sep-
•^^•^ 4 or more studies 1
arately, but when more than one correlation coefficient

129
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Schizophrenia Bulletin, Vol. 26, No. 1, 2000 M.F. Green et al.

Table 3. Meta-analyses

Total sample Pooled Standard

Domain size estimated r1 error Effect size p value2
Secondary verbal memory 727 0.29 0.039 medium : 0.000001

Immediate verbal memory 188 0.40 0.077 medium-large < 0.000001

Card sorting 1002 0.23 0.033 small-medium < 0.000001

Vigilance 682 0.20 0.040 small-medium < 0.000001

1
Estimates weighted by sample size
2
Ratio of pooled estimate of r divided by its standard error referred to a normal distribution

was reported within one outcome domain in a single come and that the effect sizes (20%-60% for the composite
study, the separate values of r were first transformed using measures) are not trivial. Obviously, functional outcome is
Fisher's z and then averaged, yielding no more than one determined by a host of factors. The current literature offers
result for meta-analysis per domain per study. The meta- convincing support for the idea that neurocognitive abilities
analytic procedures followed several steps. The homo- constitute one key determinant. From this literature, we have
geneity of the various results was tested with the Q statis- learned about whether neurocognition is related to functional
tic. In addition, the pooled estimate of r (or, equivalently, outcome. However, we have learned very little about how
z) was obtained on the basis of weighted z values and neurocognition is related to functional outcome. Identifying
tested for significance by reference to the normal curve. mechanisms and mediators was not a goal for these studies.
An a priori hypothesis (discussed in a later section) about Instead, these studies were testing neurocognitive-functional
subsets within one domain was tested by separating the relationships to see if they existed, or to better understand
studies into groups and testing the difference using heterogeneity in functional outcome. Once relationships have
weighted linear regression. been demonstrated, it is reasonable to make a more con-
The results from four separate meta-analyses are dis- certed effort to identify mechanisms.
played in table 3. At a glance, one can see that the analy- The ways in which neurocognitive measures are
ses are adequately powered with sample sizes ranging selected has probably limited our ability to identify mecha-
from 188 to 1,002. All of the relationships between the nisms. The neurocognitive assessments for these studies
constructs and the outcome domains are highly significant typically come from two sources. For the most part, they are
(all p values < 0.000001). The estimated pooled r's for the selected from standard clinical neuropsychological mea-
relationships range from 0.20 to 0.40, and the effect sizes sures that would be found in neuropsychological clinics.
range from small-medium to medium-large. The Q statis- These measures, which are usually well standardized, were
tic revealed significant heterogeneity in only the group of developed for distinguishing impaired (often from focal or
studies on vigilance. The heterogeneity was largely due to diffuse cerebral injury) from normal performance. In addi-
two outliers: one study with a very high association tion, many laboratories add measures drawn from experi-
(Bowen et al. 1994) and one study with a very low associ- mental psychology, such as the Continuous Performance
ation (Bellack et al. 1999). Overall, the meta-analyses Test for vigilance (Nuechterlein 1991). Measures such as the
convincingly demonstrate what was suggested in the pre- Continuous Performance Test have shown considerable
vious review (Green 1996), that each of these four neu- promise as indicators of vulnerability to schizophrenia. In
rocognitive constructs have significant relationships with studying relationships with functional outcome, the tests are
functional outcome. being used for different purposes than the ones for which
they were developed.
While these neurocognitive tests perform reasonably
Are We Measuring the "Right Stuff"? well as predictors and correlates of functional outcome, we
might expect other tests designed and selected for assessing
From this review of the literature, one can conclude that neu-
related capacities to be more useful. To accomplish this
rocognitive variables are indeed related to functional out-
greater degree of "fine-tuning," we first need to clarify

130
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Neurocognition and Functional Outcome Schizophrenia Bulletin, Vol. 26, No. 1, 2000

exactly what we hope to measure. We previously suggested illiteracy, which was especially high among the national
that social cognition may be one promising mediator minorities of Central Asia (Wertsch 1985).
between basic neurocognition and social competence (Kee et Vigotsky and his students claimed that the tests avail-
al. 1998; Green and Nuechterlein 1999). Another key media- able at that time were not useful for their purposes because
tor would be the process(es) that underlie one's ability to the tests had been developed for other purposes. (We would
acquire and perform instrumental life skills. As mentioned argue that our review of the literature highlights the same
above, community functioning can be considered as the sum problem.) There were no tests that could be meaningfully
total of the acquisition and performance of key life skills. We
used with the national minorities, such as those in
can evaluate "success" in a rehabilitation program by using
Uzbekistan. The existing tests, even when translated into
gross measures of acquisition, which presumes a capacity for
local languages, almost always revealed a mental back-
learning. However, we may want to have a different mea-
sure, one that directly assesses learning potential. Learning wardness among the children of the national minorities.
potential involves a focus on latent capacity rather than on The flaw, according to Vigotsky, was not just in the specific
developed abilities (Grigorenko and Sternberg 1998). It is test; instead, an entirely new approach was needed. He
dependent on basic neurocognition and is related to psy- believed that new tests would have to assess a person's
chosocial skill acquisition, but it is not identical to either. capacity for learning, not only what he or she had previ-
The idea of "learning potential" requires a fundamen- ously learned. Vigotsky outlined the key components of a
tal shift in assessment: from what the individual currently theory of learning potential but was not able to formally
knows to what the individual is capable of learning. A test his theory because he died at an early age. Further, a
possible role for learning potential as a mediator between Stalinist decree banned his work soon after his death.
basic neurocognition and skill acquisition is illustrated in Hence, his insightful work was essentially unknown out-
figure 2. Assessments of learning potential fall under the side of the Soviet Union until several decades later.
general category of "dynamic assessment," which refers The concept of learning potential presents a substan-
to systematically eliciting and determining intra-individ- tial practical challenge regarding measurement. Even if
ual variability during the course of a test (Guthke et al. we agree that learning potential is a valuable idea, how do
1997). Before we discuss applications of these sorts of we turn that into a method for dynamic assessment? One
dynamic assessments to schizophrenia, we provide a brief
of the key psychometric influences for modern
discussion of the conceptual and psychometric origins of
approaches to dynamic assessment comes from Joseph
the concept of learning potential.
Zubin. His interests in this area appear to have been
entirely independent of the work of Vigotsky. In 1950,
Origins of Learning Potential: Vigotsky Zubin proposed four axioms about applying statistical
and Zubin methods to pathological conditions (Zubin 1950). Instead
of the more typical emphases on group means, his interest
Lev Vigotsky was a supremely gifted Soviet psychologist was on repeated testing and within-subject variability.
who was active in the years following the Russian revolu- Zubin wrote that the variability within a person is "char-
tion and civil war. He had profound influence on an acteristic of the individual and varies as much or more
impressive variety of content areas, and several volumes from person to person as does the level around which this
have been dedicated to his contributions (Kozulin 1990). variability occurs." To examine this within-subject vari-
Literacy became a major issue in the Soviet Union in the ability, repeated assessments are required. With repeated
1920s when Lenin mandated a massive drive to eliminate assessments, one can study the influences on perfor-
illiteracy (Sutton 1988). Vigotsky and his students dedi-
mance, such as training and intervention.
cated themselves to addressing this practical problem of
Zubin had anticipated the focus of dynamic assess-
Figure 2. Learning potential as a mediator for ment. As opposed to static assessment at one point in
functional outcome time, dynamic assessment requires repeated assessments,
often with feedback between the examiner and the test
taker. The focus is not how much someone knows to start
•Social with but how much someone can learn with intervention.
Networks
Skl« • Independent
This capacity for learning should be dependent on (but
Acquisition 1
Performance
Living not identical to) certain basic neurocognitive processes,
• Vocational
Outcome and it should be closely related to certain external out-
come variables relevant to schizophrenia, such as psy-
chosocial rehabilitation.

131
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Schizophrenia Bulletin, Vol. 26, No. 1, 2000 M.F. Green et al.

Applications of Dynamic Assessments brief psychosocial rehabilitation program. Recent data

indicate that the groups also differ significantly in vigi-
to Schizophrenia lance as measured by performance on the Continuous
Surprisingly few of the laboratories that have developed Performance Test (Wiedl et al., submitted).
dynamic assessment methods have applied them to the Returning to the review of the literature, the neu-
study of schizophrenia. One exception is work from rocognitive measures were largely static (based on single
Wiedl and colleagues at the University of Osnabruck assessment) as opposed to dynamic (based on within-sub-
(Wiedl and Schottke 1995). Their applications of dynamic ject change across assessments). However, some measures
assessment have involved a categorical approach in which of secondary memory were similar to dynamic assessment
they divide patients into subgroups based on performance measures. Secondary memory measures were of two
before and after a training intervention. The categorical types: (1) memory for lists of words and (2) memory for
approach was an attempt to address basic problems in passages and prose. The list learning measures are some-
analyzing data from dynamic assessment methods. what similar in format to dynamic assessment measures
Eventually, dynamic assessment methods may benefit because the former involve memory for a repeated list of
from dynamical, nonlinear models that have been used words and they assess within-session learning. The tests
with other types of tasks (Paulus et al. 1996). of passages and prose are also excellent measures of sec-
ondary memory, but they do not have a dynamic assess-
The measure of choice for Wiedl's work was the
ment component. If learning potential is truly relevant to
WCST, which has been the focus of numerous attempts at
functional outcome in schizophrenia as we propose, then
remediation with schizophrenic patients (Goldberg et al.
the list learning tests should be especially good predictors
1987; Kern et al. 1996). There is considerable between-
and correlates, even compared with the other tests of sec-
subject variability in the response to training on this mea-
ondary memory. We conducted two additional meta-
sure. In our initial publication on this topic (Green et al.
analyses in which we divided the 18 studies of secondary
1990), we suggested that training on the WCST appeared
verbal memory and functional outcome into two cate-
to reveal two groups of patients: learners and nonlearners.
gories: (1) studies that used list learning tests, and (2)
The idea of heterogeneity on WCST performance was
studies that used tests of passages and prose. The relation-
reinforced by our studies of the neuropsychology of schiz-
ships with functional outcome were significant in both
ophrenia (Braff et al. 1991). The idea of learning subtypes
groups of studies; however, the strengths of the associa-
has been thoughtfully developed within the context of
tions were significantly different. The estimated r for the
dynamic assessment by Wiedl and colleagues.
list learning studies was 0.42; for the other studies, it was
Based on performance on the WCST both before and 0.24. Hence, the only test in our review that had proper-
after training, three groups are identified (Schottke et al. ties of dynamic assessment seemed to be particularly
1993). Learners start out with poor performance and related to functional outcome, even compared with other
improve a requisite amount following instructions. memory measures that did not have those properties.
Nonlearners start out poorly and do not improve with
instruction. High scorers start out performing well and
continue to perform well after instruction. Group mem- Implications for Rehabilitation
bership cannot be determined by performance at the first
test (learners look like nonlearners) nor by change scores Can knowledge of a patient's neurocognitive strengths
(high scorers look like nonlearners). It should be empha- and weaknesses guide delivery of psychosocial rehabilita-
sized that the nonleamer status is not general—it is spe- tion? One approach might involve maintaining the same
cific to this particular task and this particular training type of intervention but changing its intensity, based on
method. Different types of instruction, such as more the patient's neurocognitive abilities. For example, assess-
detailed, errorless learning techniques (Kern et al. 1996), ment of neurocognitive abilities might be used to place
would be well suited for patients who are nonlearners on patients in different rehabilitation tracks. If patients have
the briefer instructional method. deficits in verbal memory, the instructor may decide to
Validity of learner status for schizophrenia patients is use more repetition of the instructional material or, alter-
starting to be demonstrated (Wiedl and Weinhobst 1999). natively, present the material at a slower pace. Note that
The groups have been shown to differ in length of hospi- this type of tailoring the instruction to the neurocognitive
talization and in the level of demand of their current reha- needs of the patients does not require substantial modifi-
bilitation program (i.e., nonlearners were placed in less cation of the content of the training program.
demanding types of rehabilitation programs than learners Another approach would be to use a neurocognitive
and high scorers). In terms of predictive validity, learner assessment to assign patients to a separate training
status at baseline was related to degree of success on a method altogether. For example, subjects who are con-

132
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Neurocognition and Functional Outcome Schizophrenia Bulletin, Vol. 26, No. 1,2000

sidered nonlearners based on the dynamic assessment neurocognitive-functional relationships are by no means
mentioned above are still likely to be good candidates unique to schizophrenia. The impact of neurocognitive
for alternative types of instruction, such as those based deficits in other disorders such as multiple sclerosis (Rao
on errorless learning principles. Errorless learning is a et al. 1991) and AIDS (Heaton et al. 1994) are well docu-
method of training that minimizes errors during skill mented. Indeed, even in a nonclinical elderly sample, cog-
acquisition. Training begins at the simplest, most basic nitive deficits assessed by a mental status examination
skill level in which there is a high probability for suc- predicted activities of daily living (Moritz et al. 1995).
cess and progresses stepwise through a series of increas- Hence, we should not be surprised that neurocognition is
ingly complex training stages. It has been used effec- related to functional outcome in schizophrenia. On the
tively with the developmentally disabled and more contrary, if the findings did not show such a relationship,
recently in studies with schizophrenia patients (Kern et we would be confronted with a much more puzzling mys-
al. 1996). Errorless learning and similar teaching tery to explain.
approaches deemphasize the acquisition of skill through
In conclusion, we can state with considerable confi-
didactic instruction and instead emphasize procedural
dence, based on the updated review of the rapidly
aspects of learning, an area that is putatively more intact
expanding literature, that certain aspects of neurocogni-
in persons with schizophrenia. Alternative rehabilitation
tion (e.g., secondary verbal memory, immediate mem-
instructional modes, while promising, are largely experi-
ory, vigilance, and executive functioning/card sorting)
mental and have not yet been developed for use on a
are related to functional outcome in schizophrenia.
large scale.
These relationships are highly significant with medium
to large effect sizes. In addition, when studies examine
Conclusions the effects of composite neurocognitive measures, the
percentage of variance explained in functional outcome
In our review of the literature, we uncovered significant is not small (generally 20%-60%). The major limitation
associations between neurocognition and functional out- of this line of investigation is that we still know rather
comes though a survey of replicated findings, box scores, little about the underlying mechanisms through which
and meta-analyses. The meta-analyses revealed associa- these effects operate. One possibility, admittedly specu-
tions between specific neurocognitive constructs and lative at this time, is that the neurocognitive effects on
functional outcome. The relationships are larger when we functional outcome operate through the construct of
consider global or composite measures of neurocognition. learning potential.
We propose that a concept like learning potential can be
viewed as a mediator—one mechanism through which
basic neurocognitive processes are related to skill acquisi- References
tion and functional outcome in schizophrenia (see figure
2). It is not the only one—we previously suggested that Addington, J., and Addington, D. Neurocognitive and
social cognition was another promising mediator (Kee et social functioning in schizophrenia. Schizophrenia
al. 1998; Green and Nuechterlein 1999). The search for Bulletin, 25(1): 173-182,1999.
mediators between neurocognition and functional out- Addington, J.; McCleary, L.; and Munroe-Blum, H.
come is important for several reasons. First, this literature Relationship between cognitive and social dysfunction
has been largely atheoretical, and identifying mediating in schizophrenia. Schizophrenia Research, 34:59-66,
variables will help to provide a theoretical framework for 1998.
understanding the relationships between neurocognition
Asarnow, R.F.; Granholm, E.; and Sherman, T. Span of
and outcome. Second, the mediators themselves might
apprehension in schizophrenia. In: Steinhauer, S.R.;
reasonably become the target of interventions. For exam-
Gruzelier, J.H.: and Zubin, J., eds. Handbook of
ple, interventions to improve perception of emotion are
Schizophrenia: Neuropsychology, Psychophysiology, and
being developed. Even learning potential may be a char-
Information Processing. Amsterdam, The Netherlands:
acteristic of a person who is available for intervention.
Elsevier Science Publishers, 1991. pp. 335-370.
Third, assessments of potential mediators such as learning
potential and social cognition might reasonably be added Baddeley, A.D. Working Memory. New York, NY: Oxford
to existing batteries in an effort to assess the "right stuff" University Press, 1986.
for adaptive functioning in schizophrenia. Bartels, S.J.; Mueser, K.T.; and Miles, K.M. A compara-
Future studies will explore the mechanisms and tive study of elderly patients with schizophrenia and bipo-
mediators underlying relationships between neurocogni- lar disorder in nursing homes and the community.
tion and functional outcome. It should be noted that these Schizophrenia Research, 27:181-190, 1997.

133
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Schizophrenia Bulletin, Vol. 26, No. 1, 2000 M.F. Green et al.

Bellack, A.S.; Gold, J.M.; and Buchanan, R.W. Goldberg, T.E.; Weinberger, D.R.; Berman, K.F.; Pliskin,
Cognitive rehabilitation for schizophrenia: Problems, N.H.; and Podd, M.H. Further evidence for dementia of
prospects, and strategies. Schizophrenia Bulletin, the prefrontal type in schizophrenia? A controlled study of
25(2):257-274, 1999. teaching the Wisconsin Card Sorting Test. Archives of
Bellack, A.S.; Sayers, M.; Mueser, K.; and Bennett, M. General Psychiatry, 44:1008-1014, 1987.
Evaluation of social problem solving in schizophrenia. Goldman, R.S.; Axelrod, B.N.; Tandon, R.; Ribeiro,
Journal of Abnormal Psychology, 103:371-378, 1994. S.C.M.; Craig, K.; and Berent, S. Neuropsychological
Benton, A.L., and Hamsher, K. Multilingual Aphasia prediction of treatment efficacy and one-year outcome in
Examination Manual (Revised). Iowa City, IA: University schizophrenia. Psychopathology, 126:122-126, 1993.
of Iowa, 1978. Green, M.F. What are the functional consequences of neu-
rocognitive deficits in schizophrenia? American Journal
Bowen, L.; Wallace, C.J.; Glynn, S.M.; Nuechterlein,
of Psychiatry, 153:321-330, 1996.
K.H.; Lutzker, J.R.; and Kuehnel, T.G. Schizophrenic
individuals' cognitive functioning and performance in Green, M.F.; Ganzell, S.; Satz, P.; and Vaclav, J.F.
interpersonal interactions and skills training procedures. Teaching the Wisconsin Card Sort to schizophrenic
Journal of Psychiatric Research, 28:289-301,1994. patients. Archives of General Psychiatry, 47:91-92, 1990.
Braff, D.L.; Heaton, R.K.; Kuck, J.; and Cullum, M. The Green, M.F., and Nuechterlein, K.H. (1999). Should
generalized pattern of neuropsychological deficits in out- schizophrenia be treated as a neurocognitive disorder?
patients with chronic schizophrenia with heterogeneous Schizophrenia Bulletin, 25(2):309-319, 1999.
Wisconsin Card Sorting Test results. Archives of General Grigorenko, E.L., and Steinberg, R.J. Dynamic testing.
Psychiatry, 48:891-898, 1991. Psychological Bulletin, 124:75-111, 1998.
Breier, A.; Schreiber, J.L.; Dyer, J.; and Pickar, D. Guthke, J.; Beckmann, J.F.; and Dobat, H. Dynamic test-
National Institutes of Mental Health longitudinal study of ing—problems, uses, trends, and evidence of validity.
chronic schizophrenia: Prognosis and predictors of out- Educational and Child Psychology, 14:17-32, 1997.
come. Archives of General Psychiatry, 48:239-246, 1991. Harvey, P.D.; Howanitz, E.; Parrella, M.; White, L.;
Brekke, J.S.; Raine, A.; Ansel, M.; and Lencz, T. Davidson, M.; Mohs, R.C.; Hoblyn, J.; and Davis, K.L.
Neuropsychological and psychophysiological correlates Symptoms, cognitive functioning, and adaptive skills in
of psychosocial functioning in schizophrenia. geriatric patients with lifelong schizophrenia: A compari-
Schizophrenia Bulletin, 23(1): 19-28, 1997. son across treatment sites. American Journal of
Psychiatry, 155:1080-1086, 1998.
Buchanan, R.W.; Holstein, C ; and Breier, A. The compar-
ative efficacy and long-term effect of clozapine treatment Harvey, P.D.; Sukhodolsky, D.; Parrella, M.; White, L.;
on neuropsychological test performance. Biological and Davidson, M. The association between adaptive and
Psychiatry, 36:717-725, 1994. cognitive deficits in geriatric chronic schizophrenic
patients. Schizophrenia Research, 27:211-218, 1997.
Corrigan, P.W.; Green, M.F.; and Toomey, R. Cognitive
correlates to social cue perception in schizophrenia. Heaton, R.K. Wisconsin Card Sorting Test Manual.
Psychiatry Research, 53:141-151, 1994. Odessa, FL: Psychological Assessment Resources, 1981.
Heaton, R.K., and Pendleton, M.G. Use of neuropsycho-
Corrigan, P.W., and Toomey, R. Interpersonal problem
solving and information processing in schizophrenia. logical tests to predict adult patients' everyday function-
Schizophrenia Bulletin, 21(3):395-^O3, 1995. ing. Journal of Consulting and Clinical Psychology,
49:807-821, 1981.
Corrigan, P.W.; Wallace, C.J.; Schade, M.L.; and Green,
Heaton, R.K.; Velin, R.V.; McCutchan, A.; Gulevich, S.J.;
M.F. Cognitive dysfunctions and psychosocial skill learn-
Atkinson, J.H.; Wallace, M.R.; Godfrey, H.P.D.; Kirson,
ing in schizophrenia. Behavior Therapy, 25:5-15, 1994.
D.A.; and Grant, I. Neuropsychological impairment in
Davies, D.R., and Parasuraman, R. The psychology of vig- human immunodeficiency virus-infection: Implications for
ilance. London, England: Academic Press, 1982. employment. Psychosomatic Medicine, 56(1):8—17, 1994.
Delis, D.C.; Kramer, J.H.; Kaplan, E.; and Ober, B.A. The Hedges, L.V., and Olkin, I. Statistical Methods for Meta-
California Verbal Learning Test (Manual). San Antonio, analysis. Orlando, FL: Academic Press, 1985.
TX: Psychological Corp, 1987. Heslegrave, R.J.; Awad, G.; and Voruganti, L.N.P. The
Dickerson, F.; Boronow, J.J.; Ringel, N.; and Parente, F. influence of neurocognitive deficits and symptoms on
Neurocognitive deficits and social functioning in outpatients quality of life in schizophrenia. Journal of Psychiatry and
with schizophrenia. Schizophrenia Research, 21:75-83, 1996. Neuroscience, 22:235-243, 1997.

134
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Neurocognition and Functional Outcome Schizophrenia Bulletin, Vol. 26, No. 1, 2000

Dcebuchi, E.; Nakagome, K.; Tugawa, R.; Asada, Y.; Mori, Mueser, K.T.; Bellack, A.S.; Douglas, M.S.; and Wade,
K.; Takahashi, N.; Takazawa, S.; Ichikawa, I.; and Akaho, J.H. Prediction of social skill acquisition in schizophrenic
R. What influences social skills in patients with schizo- and major affective disorder patients from memory and
phrenia? Preliminary study using the role play test, symptomatology. Psychiatry Research, 37:281-296, 1991.
WAIS-R and event-related potential. Schizophrenia Mueser, K.T.; Blanchard, J.J.; and Bellack, A.S. Memory
Research, 22:143-150, 1996. and social skill in schizophrenia: The role of gender.
Jaeger, J., and Douglas, E. Neuropsychiatric rehabilitation Psychiatry Research, 57:141-153, 1995.
for persistent mental illness. Psychiatric Quarterly, Mueser, K.T.; Doonan, B.; Penn, D.L.; Blanchard, J.J.;
63:71-94,1992. Bellack, A.S.; Nishith, P.; and DeLeon, J. Emotion recog-
Johnstone, E.C.; Macmillan, J.F.; Frith, CD.; Benn, D.K.; nition and social competence in chronic schizophrenia.
and Crow, T.J. Further investigation of the predictors of Journal of Abnormal Psychology, 105:271-275, 1996.
outcome following first schizophrenic episodes. British Nuechterlein, K.H. Vigilance in schizophrenia and related
Journal of Psychiatry, 157:182-189, 1990. disorders. In: Steinhauer, S.R.; Gruzelier, J.H.; and Zubin,
Kee, K.S.; Kern, R.S.; and Green, M.F. Perception of J., eds. Handbook of Schizophrenia. Amsterdam, The
emotion and neurocognitive functioning in schizophrenia: Netherlands: Elsevier, 1991. pp. 397^33.
What's the link? Psychiatry Research, 81, 57-65,1998.
Paulus, M.P.; Geyer, M.A.; and Braff, D.L. Use of methods
Kern, R.S.; Green, M.F.; and Satz, P. Neuropsychological from chaos theory to quantify a fundamental dysfunction in
predictors of skills training for chronic psychiatric the behavioral organization of schizophrenic patients.
patients. Psychiatry Research, 43:223-230, 1992. American Journal of Psychiatry, 153:714-717, 1996.
Kern, R.S.; Wallace, C.J.; Hellman, S.G.; Womack, L.M.; Penn, D.L.; Mueser, K.T.; and Spaulding, W. Information
and Green, M.F. A training procedure for remediating processing, social skill, and gender in schizophrenia.
WCST deficits in chronic psychotic patients: An adapta- Psychiatry Research, 59:213-220, 1996.
tion of errorless learning principles. Journal of
Penn, D.L.; Mueser, K.T.; Spaulding, W.; Hope, D.A.; and
Psychiatric Research, 30:283-294, 1996.
Reed, D. Information processing and social competence
Klapow, J.C.; Evans, J.; Patterson, T.L.; Heaton, R.K.; in chronic schizophrenia. Schizophrenia Bulletin,
Koch, W.L.; and Jeste, D.V. Direct assessment of func- 21(2):269-281, 1995.
tional status in older patients with schizophrenia.
Penn, D.L.; Spaulding, W.D.; Reed, D.; and Sullivan, M.
American Journal of Psychiatry, 154:1022-1024, 1997.
The relationship of social cognition to ward behavior in
Kozulin, A. Vygotsky's Psychology. Cambridge, MA: chronic schizophrenia. Schizophrenia Research,
Harvard University Press, 1990. 20:327-335, 1996.
Lezak, M.D. Neuropsychological Assessment. 3rd ed. Penn, D.L.; Van Der Does, W.; Spaulding, W.; Garbin,
New York, NY: Oxford University Press, 1995. C.P.; Linszen, D.; and Dingemans, P. Information process-
Lysaker, P.; Bell, M.; and Beam-Goulet, J. Wisconsin ing and social cognitive problem solving in schizophrenia.
Card Sorting Test and work performance in schizophre- Journal of Nervous and Mental Disease, 181:13-20,
nia. Psychiatry Research, 56:45-51, 1995. 1993.
Lysaker, PH.; Bell, M.D.; Zito, W.S.; and Bioty, S.M. Rao, S.M.; Leo, G.J.; Ellington, L.; and Nauertz, T.
Social skills at work: Deficits and predictors of improve- Cognitive dysfunction in multiple sclerosis: II. Impact on
ment in schizophrenia. Journal of Nervous and Mental employment and social functioning. Neurology,
Disease, 183:688-692, 1995. 41:692-696, 1991.
McKee, M.; Hull, J.W.; and Smith, T.E. Cognitive and Schottke, H.; Bartram, M.; and Wiedl, K.H. Psychometric
symptom correlates of participation in social skills train- implications of learning potential assessment: A typologi-
ing groups. Schizophrenia Research, 23:223-229, 1997. cal approach. In: Hamers, J.H.M.; Sijtsma, K.; and
Meltzer, H.Y.; Thompson, P.A.; Lee, M.A.; and Ranjan, Ruijssenaars, A.J.J.M., eds. Learning Potential
R. Neuropsychologic deficits in schizophrenia: Relation Assessment: Theoretical, Methodological, and Practical
to social function and effect of antipsychotic drug treat- Issues. Amsterdam, The Netherlands: Swets and
ment. Neuropsychopharmacology, 14(3S):27-33, 1996. Zeitlinger, 1993. pp. 153-173.
Moritz, D.J.; Kasl, S.V.; and Berkman, L.F. Cognitive Silverstein, S.M.; Schenkel, L.S.; Valone, C ; and
functioning and the incidence of limitations in activities Nuernberger, S.W. Cognitive deficits and psychiatric reha-
of daily living in an elderly community sample. American bilitation outcomes in schizophrenia. Psychiatric
Journal of Epidemiology, 141:41-49, 1995. Quarterly, 69:169-191, 1998.

135
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018
 Schizophrenia Bulletin, Vol. 26, No. 1, 2000 M.F. Green et al.

Sutton, A.L.S. Vygotskii: The cultural-historical theory, phrenic patients differing in learning proficiency in the
national minorities and the zone of next development. Wisconsin Card Sorting Test. Submitted.
In: Gupta, R., and Coxhead, P., eds. Cultural Diversity
Wykes, T.; Sturt, E.; and Katz, R. The prediction of reha-
and Learning Efficiency: Recent Developments in
bilitative success after three years: The use of social,
Testing. New York, NY: St. Martin's, 1988. pp. 89-117.
symptom and cognitive variables. British Journal of
Velligan, D.I.; Mahurin, R.K.; Diamond, P.L.; Hazelton, Psychiatry, 157:865-870, 1990.
B.C.; Eckert, S.L.; and Miller, A.L. The functional signifi-
cance of symptomatology and cognitive function in schiz- Zubin, J. Symposium on statistics for the clinician.
ophrenia. Schizophrenia Research, 25:21-31, 1997. Journal of Clinical Psychology, 6:1-6, 1950.

Weaver, L.A., and Brooks, G.W. The use of psychomet-

ric tests in predicting the potential of chronic schizo- Acknowledgments
phrenics. Journal of Neuropsychiatry, 5:170-180, 1964.
The authors thank Mary Jane Robertson, M.S., for her
Wertsch, J.V. Vigotsky and the Social Formation of
help in the preparation of the manuscript and figures.
Mind. Cambridge, MA: Harvard University Press, 1985.
Kimmy S. Kee, Ph.D., and Mark J. Sergi, Ph.D., con-
Wiedl, K.H., and Schottke, H. Dynamic assessment of tributed to the literature review.
selective attention in schizophrenic subjects: The analysis
of intraindividual variability of performance. In: Carlson,
J.S., ed. Advances in Cognition and Educational Practice. The Authors
Greenwich, CT: JAI Press, 1995. pp. 185-208.
Michael Foster Green, Ph.D., is Professor in Residence,
Wiedl, K.H., and Weinhobst, J. Intraindividual differ- and Robert S. Kern, Ph.D., is Assistant Research
ences in cognitive remediation research with schizo- Psychologist, Department of Psychiatry and Behavioral
phrenic patients—indicators of rehabilitation potential? Sciences, University of California, Los Angeles (UCLA);
International Journal of Rehabilitation Research, David L. Braff, M.D., is Professor, Department of
22:1-5, 1999. Psychiatry, University of California, San Diego; Jim
Wiedl, K.H.;Weinobst, J.; Schottke, H.H.; Green, M.F; Mintz, Ph.D., is Professor in Residence, Department of
Nuechterlein, K.H. Attentional characteristics of schizo- Psychiatry and Behavioral Sciences, UCLA.

136
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/26/1/119/1927697
by guest
on 20 February 2018