Cognitive Functioning in Schizophrenia:
Implications for Psychiatric Rehabilitation
by William D. Spaulding, Shelley K. Fleming, Dorie Reed,
Mary Sullivan, Daniel Storzbach, and Mona Lam
Abstract
Research in psychopathology and the cognitive neuro-
sciences suggests new applications in psychiatric reha-
bilitation. Analysis of performance deficits on labora-
tory tasks can contribute to treatment planning,
individual and family counseling, and staff consulta-
tion, much like it does in cases of brain injury and
other types of central nervous system neuropathology.
Recognition of the nature of cognitive impairments in
schizophrenia can inform design of psychosocial tech-
niques such as social and living skills training.
Cognitive impairments are increasingly seen as poten-
tial targets for pharmacological and psychosocial
treatment and rehabilitation. In this article, three key
issues for application of cognitive technology in psychi-
atric rehabilitation of schizophrenia and related disor-
ders are formulated as straightforward, clinically rele-
vant questions: (1) What is the prognostic significance
of cognitive impairment in acute psychosis? (2) Can
cognitive functioning improve in the chronic, residual
course? (3) How does cognitive improvement benefit
other aspects of recovery and rehabilitation? These
questions are addressed through review of previous
findings and new multivariate analyses of cognitive
functioning in the acute, post-acute, and chronic resid-
ual phases of schizophrenia.
Key words: Cognitive rehabilitation.
Schizophrenia Bulletin, 25(2):275-289,1999.
Historically, the preponderance of scientific interest in the
cognitive aspects of schizophrenia has been directed
toward illuminating the disorder's etiology, with only
indirect consideration of clinical applications. Never-
theless, cognitive measures and models are increasingly
proposed as potentially useful prognostic indicators
(Wykes et al. 1990; Wykes and Dunn 1992; Bowen et al.
1994), as limiting factors in social functioning (Penn
1991; Kern and Green 1994; Penn et al. 1995, 1996;
275
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
Green 1996; Meltzer et al. 1996), and as targets for direct
treatment (Brenner et al. 1983, 1992; Spaulding et al.
1986, 1998; Corrigan and Storzbach 1993; Lee et al.
1994). These proposals reflect a convergence, on the one
hand, of research in the neurosciences, experimental psy-
chopathology, neuropsychology, and cognitive-behavioral
psychotherapy and, on the other hand, a perceived need to
translate the methods, principles, and insights of cognitive
psychopathology into practice.
New roles for cognitive psychopathology suggest
new applications for a number of familiar clinical tech-
nologies in the domain of psychological and neuropsy-
chological assessment. For example, analysis of perfor-
mance deficits on laboratory tasks can contribute to
treatment planning, individual and family counseling, and
staff consultation, much like it does in cases of brain
injury and other types of central nervous system neu-
ropathology (Erickson and Binder 1986; Erickson 1988,
1994). Similarly, recognition of the nature of schizo-
phrenic cognitive impairments can inform design of psy-
chosocial techniques such as social and living skills train-
ing (Liberman et al. 1982; Kern and Green 1994).
The issues for application of cognitive technology
are not necessarily the same as those that drive develop-
ment of etiological models. Also, the issues raised by
cognitive applications in classical neuropsychology do
not overlap completely with issues raised in assessment
and treatment of schizophrenic impairments. For exam-
ple, some parameters that give necessary context to tradi-
tional neuropsychological assessment, such as time since
injury, are poorly understood and possibly irrelevant in
schizophrenia.
In the following sections, three key issues applicable
to psychiatric rehabilitation of schizophrenia and related
disorders are formulated as straightforward, clinically rel-
evant questions:
Reprint requests should be sent to Dr. W.D. Spaulding, Dept. of
Psychology, University of Nebraska-Lincoln, Lincoln, NE 68588-0308.
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
1. What is the prognostic significance of cognitive
impairment in acute psychosis?
2. Can cognitive functioning improve in the chronic,
residual course?
3. How does cognitive improvement benefit other aspects
of recovery and rehabilitation?
Recent research findings from our laboratory and others
are marshaled to help clarify the nature of the questions
and in some cases provide preliminary answers. In some
instances, contemporary theoretical frameworks are help-
ful or even essential for informing application strategies.
Neurodevelopmental models of schizophrenia (e.g.,
Walker 1994; Weinberger and Lipska 1995; Weinberger
1996) are especially important in this regard. In other
instances, clinical considerations, observations, or impera-
tives suggest a need for further development of theoretical
models. In the final section we propose theoretical mod-
els to account for the important new finding that cognitive
impairments respond to psychosocial interventions in the
chronic course of schizophrenia.
What Is the Prognostic Significance of
Cognitive Impairment in Acute
Psychosis?
Gross disruption of cognitive functioning is usually part
of the clinical picture of acute psychosis, and reduction of
cognitive impairment usually accompanies resolution of
the acute episode. Experimental studies of the relation-
ship between symptomatology and cognitive functioning
in acute psychosis may provide clues about the neuro-
physiological and neuropsychological mechanisms that
produce them (e.g., Gilbertson et al. 1994). However,
what is known about the long-term prognostic signifi-
cance of cognitive impairment is generally limited to the
severity of residual impairment after resolution of acute
psychosis (for a review, see Meltzer et al. 1996).
Cognitive functioning at the point of hospitalization has
been found to predict some important consequences, for
example, emotional blunting after remission of the acute
psychosis (Silverstein et al. 1994), but little more is
known about the portent of cognitive functioning during
exacerbations. Hospitalization during acute psychosis is
often the starting point for long-term treatment and reha-
bilitation. A better understanding of the prognostic impli-
cations of cognitive functioning at this point could
enhance the planning process.
To determine whether cognitive measures during an
acute episode have prognostic value over and above diag-
nostic, symptomatic, and behavioral indicators, we reana-
lyzed data collected in a recent study of patients admitted
to an acute short-term unit in a State hospital (Lam 1997).
276
W.D. Spaulding et al.
During the study, 31 patients were admitted, and subse-
quently discharged or transferred to a long-term inpatient
unit; 19 (61%) had a diagnosis of schizophrenia, 10 (32%)
had a diagnosis of major depression or bipolar disorder, 1
had an eating disorder, and 1 had substance abuse-related
psychosis. Each was assessed with a structured interview,
the Brief Psychiatric Rating Scale (BPRS; Ventura et al.
1993); a self-report symptom and problem inventory,
Symptom Checklist-90R (SCL-90R; Peveler and Fairburn
1990); and a neurocognitive test battery (COGLAB;
Spaulding et al. 1989). Assessment was performed twice:
within 14 days of hospitalization and 4 weeks later. Two
assessments were included to determine whether cogni-
tive functioning at the start of treatment has different
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
prognostic significance than later. Two indices of short-
term outcome were analyzed: length of stay in the acute
unit and discharge destination. The latter is the setting to
which the patient went upon discharge, coded as a scale
reflecting graduated levels of restrictiveness from inde-
pendent living to locked, long-term inpatient care.
With data available at the first assessment, including
demographic characteristics, it was possible to construct a
multiple regression equation that predicted about 65 per-
cent of the variance in length of stay (adjusted R2 = 0.647,
F = 11.28, p < 0.00005). The regression statistics for this
formula are shown in table 1. Interestingly, diagnosis is
not significantly associated with outcome in this context
The patient's age, age at onset, and marital status are
powerful predictors of length of stay; a BPRS and an
SCL-90R subscale also contribute to a statistically signifi-
cant degree. Even after those factors are taken into
account, the perseverative error score on the COGLAB
card sorting task contributes uniquely to the prediction.
That is the only score of the 11 derived from the
COGLAB battery that was associated with outcome, but
the size of its beta weight is sufficient to allow confidence
that it is not an artifact of multiple correlational tests.
Table 1. Multiple regression statistics for
predicting length of stay from demographic,
clinical, and cognitive characteristics at
admission
Variable Beta Significance1
Age 0.34 0.0248
Age at onset 0.39 0.0057
Marital status 0.56 0.0002
BPRS paranoia factor 0.31 0.0126
COGLAB Card Sort
perseverative errors 0.52 0.0002
Note.BPRS = Brief Psychiatric Rating Scale.
1
Significance of T, reflecting the probability that the variable's
unique contribution to the overall prediction of the dependent vari-
able is zero.
Cognitive Functioning
The demographic variables were not so effective for
predicting discharge destination, but card sorting perfor-
mance was (see table 2). The equation still accounted for
37 percent of the variance in discharge destination
(adjusted R2 = 0.365, F = 4.34, p < 0.006). These analy-
ses indicate that patients with more impaired card sorting
performance shortly after acute hospitalization stayed in
the acute ward longer and were more likely to be dis-
charged to a residential or long-term hospital unit than to
supported or independent community living.
Table 2. Multiple regression statistics for
predicting discharge destination from
demographic, clinical, and cognitive
characteristics at admission
Variable Beta Significance1
Age 0.11 0.5434
Age at onset 0.03 0.8519
Marital status 0.03 0.8687
BPRS paranoia factor 0.51 0.0025
COGLAB Card Sort
perseverative errors 0.43 0.0100
NoteBPRS = Brief Psychiatric Rating Scale.
Significance of T, reflecting the probability that the variable's
unique contribution to the overall prediction of the dependent vari-
able is zero.
Data from the second assessment were then analyzed
for additional predictive value. The contribution of age
was eclipsed by other predictors at the second assessment,
so it was excluded from the prediction equation. Card
sorting performance at the second assessment no longer
contributed either, but another COGLAB measureper-
formance on a backward visual masking taskdid con-
tribute. The best equation including these variables, and
the remaining variables from the first assessment,
accounted for almost 75 percent of the variance in length
of stay (adjusted R2 = 0.744, F = 17.28, p < 0.00005). As
table 3 shows, card sorting performance at the first assess-
ment and masking performance at the second assessment
contributed uniquely to this prediction.
It is interesting that both paranoid behavior and card
sorting perseverative errors contributed to outcome in this
situation. Perseverative errors have long been associated
with paranoia (Spaulding 1978), presumably because the
rigidity of thinking that produces this type of error also
produces the inflexible beliefs and attributions of para-
noia. (The perseverative error score here uses the original
criteria of the Wisconsin Card Sorting Task (WCST),
which are different from the criteria used by WCST ver-
sions in most common use today; see Wagman and
Wagman (1992). Perseverative errors as scored by the
current criteria may be differentially associated with
277
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
Table 3. Multiple regression statistics for
predicting length of stay from demographic,
clinical, and cognitive data available after 6
weeks of hospitalization
Variable Beta Significance1
Age at onset 0.24 0.0234
Marital status 0.45 0.0001
BPRS paranoia factor at
admission 0.28 0.0104
COGLAB Card Sort
perseverative errors at
admission 0.49 0.0001
COGLAB Backward Masking
task at second assessment 0.41 0.0005
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
Note.BPRS = Brief Psychiatric Rating Scale.
1
Significance of T, reflecting the probability that the variable's
unique contribution to the overall prediction of the dependent vari-
able is zero.
schizophrenia, but they are not necessarily associated with
paranoia.) It would make sense that a syndrome of rigid
thinking and hostile, suspicious, and uncooperative social
behavior would inhibit efficacious treatment and disposi-
tion in a psychiatric treatment setting, but the current
analyses indicate that the behavioral and the cognitive
aspects of this syndrome contribute independently to out-
come. The role of the masking measure at the second
assessment is also interesting. Masking measures an early
stage of visual information processing, but it is known to
be responsive to antipsychotic drug interventions (Braff
and Sacuzzo 1982). In this analysis, the masking measure
may be detecting a subgroup of patients who respond less
well to antipsychotic medication, who remain more
actively psychotic after several weeks of treatment, and
who therefore require a longer hospitalization.
These analyses provide some encouragement that
cognitive assessment in the acute and early post-acute
phase can be made useful for predicting post-acute and
chronic baseline functioning, at least to the degree that
length of hospital stay and discharge destination reflect
baseline functioning. This prediction is likely to be easier
in a mixed sample that includes patients with both high
and low levels of baseline functioning. Within a group of
patients with more impaired baseline functioning, higher
levels of symptoms between episodes, and suboptimal
medication response, it may be extremely difficult to pre-
dict baseline and residual levels of cognitive impairment.
It may be difficult even to determine when the patient has
reached stabilization of the acute state. In such cases, sys-
tematic longitudinal assessment may be necessary to
determine the significance of cognitive impairments for
longer-term treatment planning.
Schizophrenia Bulletin, Vol. 25, No. 2, 1999 W.D. Spaulding et al.

Can Cognitive Functioning Improve in
the Chronic, Residual Course?
Apart from those examining treatment of residual depres-
sion or negative symptoms (e.g., Plasky 1991; Silver and
Nassar 1992; Siris 1994), there are relatively few system-
atic studies showing cognitive recovery independent of
resolution of acute psychosis ("recovery" in this context
means regaining some degree of premorbid functioning,
not necessarily full return to premorbid levels). Studies of
long-term outcome of schizophrenia and related disorders
(Harding et al. 1992) have found surprisingly high levels
of recovery of personal and social functioning, indirectly
suggesting that some degree of cognitive recovery under-
help recover more impaired ones? The answers to these
questions would provide important clues about how cog-
nitive recovery can come about.
For the present article, in order to further articulate
the nature of the cognitive changes observed in the
Spaulding et al. (1998) study, we performed a principal
components analysis of the data from that study. The
analysis included change scores for the nine measures that
showed a change over the 6-month study period. The
change scores are simply the arithmetic difference
between pre and posttreatment scores adjusted for initial
values effects by residualization in linear regression (a
statistical procedure to control for the problem that
change scores may have different meanings at different

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016

lies this. On the other hand, long-term behavioral recov- levels of a measure). The results are summarized in
ery could also reflect the cumulative effects of socializa- table 4.
tion uninterrupted by acute episodes, without any changes Using conventional extraction and rotation parame-
in cognition per se. The atypical antipsychotics, espe- ters, the analysis identifies three separate sources of vari-
cially risperidone, may produce somewhat better cogni- ance that collectively account for about half of the total
tive functioning than typicals in stabilized patients variance in cognitive change. Although there is some
(Green 1998). This finding suggests the possibility of overlap, the three factors clearly differentiate changes in
some remediational process independent of episode reso- different domains of functioning. The first factor is pre-
lution, and the search is on for pharmacotherapeutic dominantly a verbal processing factor that spans elemen-
strategies that minimize residual impairment while maxi- tal and complex levels of cognition. (The backward mask-
mizing stabilization and symptom control. Meanwhile, ing task requires subjects to identify alphanumeric
the search also continues for psychosocial strategies to characters presented in a preattentional time frame, while
counteract residual cognitive impairment. verbal memory tasks require integrated storage and
A previous analysis in our laboratory (Spaulding
1993) of 110 severely disabled but stabilized patients in a
State hospital long-term unit showed that cognitive func- Table 4. Structure matrix of neurocognltlve
tioning undergoes little or no change over a 6-month change scores, oblique rotation of principal
period. A more recent longitudinal study, also completed component extraction
in our lab (Spaulding et al. 1998), assesses cognitive Factor Factor Factor
changes in equally disabled and stabilized chronic Measure 1 2 3
patients over the same time period but in an enriched psy- Rey Verbal Learning
chosocial treatment environment. The second study Test -Visual 1 0.76 0.12 0.01
showed improvements in 9 of 12 measures of cognitive COG LAB Backward
functioning in chronic schizophrenia patients with espe- Masking 0.75 0.00 0.09
cially severe residual impairments. Measures of memory Denman Verbal Memory
and executive functioning showed improvement, while Quotient 0.55 0.41 0.25
Rey Verbal Learning
measures of reaction time and continuous attention did
Test-Auditory1 0.46 0.42 0.01
not. One measure of preattentional visual processing
Halstead Reitan
backward maskingchanged, while another measure of Trailmaking B 0.09 0.79 0.25
preattentional processinga span of apprehension task Halstead Reitan Tactile
did not. Thus, there is evidence that at least under certain Performance Test 0.07 0.71 0.17
conditions, some aspects of impaired schizophrenic cog- Denman Nonverbal Memory
nition are subject to improvement in the chronic course. Quotient 0.49 0.60 -0.33
What is the qualitative nature of the observed cogni- COGLAB Card Sorting Task 0.00 0.27 0.77
tive change? Are some aspects of cognitive functioning Halstead Reitan Categories 0.37 0.15 0.76
1
more amenable to recovery than others? Does the The Rey Verbal Learning Test in its standard form uses auditory
presentation of the stimulus words. The nonstandard visual form
improvement simply reflect a general, overall improve- uses presentation of printed words. Change scores are residual-
ment in neurophysiological, cognitive, and behavioral ized for initial values (i.e., the effect of the initial overall severity of
functioning? Do patients use less impaired abilities to the impairment has been statistically removed).

278
Cognitive Functioning
retrieval of verbal information in a much longer time
frame, in both visual and auditory modalities.) The sec-
ond factor is a spatial information processing factor (all
the tasks require processing of spatial relations, whether
or not memory is also involved), and the third is a concept
formation and manipulation factor. These results indicate
that, in addition to the predominance of more global
changes, cognitive change in the chronic residual phase is
somewhat segregated into verbal, spatial, and conceptual
domains.
Segregation of verbal and spatial functions suggests
lateralized hemispheric differences. The findings on later-
alized neuropsychological deficits in schizophrenia
(Gruzelier 1979, 1991; see also Gold et al. 1994) suggest
that both right and left lateralization can occur in a hetero-
geneous schizophrenia population. The segregation of
verbal and nonverbal performance into separate factors, as
in the table 1 factor structure, would be consistent with a
mixed population of individuals with varying degrees of
impairment lateralized to one or another cerebral hemi-
sphere. However, the segregation of cognitive processes
into factors could also reflect purely functional relation-
ships between various aspects of behavioral performance,
independent of hemispheric or any other kind of anatomi-
cal localization. They could also reflect relationships
between specific cognitive abilities and anatomically dis-
persed neurophysiological mechanisms (e.g., neurotrans-
mitter or neuroendocrine systems). Changes in these cog-
nitive abilities could reflect slowly occurring changes in
lateralized neurophysiological processes, or functional
organization of cognitive processes, or combinations of
the two.
Interestingly, Factors 1 and 2 are inversely correlated
(r = -0.23), meaning improvement on one is associated
with a lack of improvement, or deterioration, on the other.
This correlation could occur because in patients with lat-
eralized impairments, recovery from episode-linked
deficits in the more impaired hemisphere is quantitatively
greater, and therefore more apparent on neuropsychologi-
cal tasks, or because vulnerability-linked deficits prevent
recovery in the more impaired hemisphere. Also, there
may be "trade-offs" involved in recovery processes. In
Factor 3, there appears to be such a trade-off between
nonverbal memory and concept manipulation perfor-
mance, evidenced by the negative factor loading of the
nonverbal memory measure. Such a trade-off could be
the result of reallocating limited processing capacity from
one functional domain to another (Knight and Russell
1978; Gjerde 1983). In that sense, cognitive recovery
may in part represent more efficient allocation of cogni-
tive capacity in response to environmental demands. In
cases where there are significant enduring deficits, reallo-
cation of capacity would be further improved by taking
279
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
one's own strengths and weaknesses into account. If
enduring deficits were lateralized, better executive reallo-
cation would produce inverse relationships, as with
Factors 1 and 2.
To further explore the implications of lateralization,
we identified left- and right-lateralized deficits in this
sample with a crude but clinically useful convention, a
difference between Wechsler Adult Intelligence Scale
1981 (WAIS) Vocabulary and Picture Arrangement of
three standardized subscale points in either direction.
With this criterion, 34 percent of the sample showed later-
alization to the right hemisphere, and 12 percent showed
lateralization to the left. The patients with left lateraliza-
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
tion had a strikingly poor outcome. Whereas the entire
sample showed significant gains in social competence as
measured by the Assessment of Interpersonal Problem
Solving Skills (AIPSS; Donohoe et al. 1990) over the
course of the study (discussed further below), the sub-
group of left-lateralized patients showed no change.
Explicit targeting of cognitive impairment in psy-
chosocial rehabilitation enhances cognitive improvement.
The Spaulding et al. (1998) outcome study included a ran-
domized controlled comparison of comprehensive biopsy-
chosocial rehabilitation with and without a group treat-
ment modality designed to exercise cognitive abilities
believed to be prerequisite to sociobehavioral functioning.
The modality was constructed from the cognitive subpro-
grams of Integrated Psychological Therapy (IPT; Brenner
et al. 1994), which combine social skills training with
other psychosocial techniques. Control subjects received
supportive group therapy in identical doses, and all sub-
jects received a standard regimen of intensive pharma-
cotherapy, social and living skills training, contingency
management, and other rehabilitation modalities. Over
the course of 6 months of rehabilitation, patients who
received the cognitive treatment showed improvement in
some domains of cognition, but the subjects receiving
intensive rehabilitation without the cognitive component
did not. There was strong evidence for differential
improvement in span of apprehension, a preattentional
level of cognitive processing. (The overall change on this
measure was not significant because of the complete lack
of improvement in the control subjects, and. so this mea-
sure was not included in the factor analysis of changes
described above.) This particular domain of cognition is
known to be sensitive to "top-down" information process-
ing factors (Silverstein et al. 1996), so it is possible that
the differential improvement reflects better executive-
level modulation of preattentional processing routines, as
opposed to improvement in the rate or efficiency of the
routines themselves. There was weaker evidence of dif-
ferential change in more global cognitive functions, as
measured by the COGLAB card sorting task. In the con-
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
text of the large improvements in global functions in both
groups, further enhancement by cognitive treatment in
this domain may be less significant, or simply more diffi-
cult to detect. Taken together, these findings suggest that
part of the effect of the cognitive treatment was to con-
tribute a unique effect on the ability of higher-level func-
tions to optimally activate specialized elemental processes
in response to particular environmental demands.
In summary, cognitive recovery can occur in stabi-
lized patients, and it appears to involve a spectrum of spe-
cific processes. Generally, the changes appear more
robustly in higher levels of cognition. For both post-acute
and chronic course recovery, the predominance of
changes in measures of more complex cognition in the
available research data is consistent with the hypothesis
that recovery primarily involves improvements in execu-
tive-level organization, modulation, and integration of
elemental cognitive processes, as opposed to better opera-
tion of the elemental processes themselves. (Medication-
related improvement in backward masking may be a sig-
nificant exception to this rule.) Differential recovery of
different brain functions, related to each other through
anatomical, neurophysiological, or functional organiza-
tion, may produce individual differences in the recovery
process as well. Lateralization of impairments may be
especially significant in this regard. Recovery is subject
to enhancement by environmental and psychosocial fac-
tors, which may range from "therapeutic milieu" condi-
tions to therapeutic procedures that explicitly target cogni-
tive functioning.
How Does Cognitive Improvement
Benefit Other Aspects of Recovery and
Rehabilitation?
Whether or not cognitive change can be enhanced by
treatment and rehabilitation, it is important to know
whether recovery of specific functions has meaningful
benefits for the patient. Such benefits could logically lie
in either objective or subjective domains. Subjectively, it
would be logical to expect that better cognitive function-
ing would be associated with reductions in distress or
greater self-efficacy. However, it would be difficult to
demonstrate this empirically because subjective expres-
sions could be influenced by many factors, including
symptom reduction and improvement in behavioral func-
tioning unrelated to cognitive changes. The systematic
studies needed to determine the true subjective value of
cognitive recovery in schizophrenia have not been per-
formed.
Many experimental and case studies show perfor-
mance improvements in cognitive domains in response to
280
W.D. Spaulding et al.
practice, special training, or other treatment-like proce-
dures (reviewed by Reed et al. 1992; Corrigan and
Storzbach 1993). There can be little doubt that people
with stabilized chronic schizophrenia can achieve perfor-
mance improvements on a variety of cognitively challeng-
ing tasks. However, the improvements observed in previ-
ous studies may not be any different from changes in
normal subjects who practice an unfamiliar task, and there
are no previous indications that such changes confer eco-
logically significant benefits in personal or social func-
tioning. The Spaulding et al. (1998) study is the first
large-scale controlled treatment trial to show that treat-
ment-induced cognitive changes in the chronic phase have
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
important objective benefits.
The primary outcome measure in that study, the
AIPSS, is an elaborate assessment of social competence.
In the AIPSS, the subject is shown a series of videotaped
vignettes of people engaged in some interpersonal prob-
lem or conflict. After each vignette the subjects are inter-
viewed to ascertain their understanding of the portrayed
problem and their ability to resolve it. Finally, they are
asked to role-play a solution. The entire assessment is
itself videotaped and later scored by trained observers
using a structured protocol. Assessments like the AIPSS
detect improvements in social competence that are known
to have implications for long-term outcome, including
relapse and hospital recidivism (Wallace and Liberman
1985; Hogarty et al. 1987). The size of the treatment
effect of the cognitive intervention, as measured by the
AIPSS, is comparable to the effect size of atypical
antipsychotics over typicals in suppressing symptoms and
enhancing behavioral functioning. Longer-term appraisal
of overall rehabilitation outcome (Weilage 1997) confirms
that the Spaulding et al. (1998) cognitive treatment group
showed higher levels of rehabilitation success, as mea-
sured by judgment of staff who were blind to treatment
group assignment.
There are a number of possible causal relationships
between the domains of functioning measured in the
Spaulding et al. (1998) study. Cognitive changes may be
links in a cascade of changes, from stabilization of the
neurophysiological level of acute psychosis to social
behavior, or they may be incidental "markers" of neuro-
physiological changes, unrelated to behavioral conse-
quences. In addition to having different roles in the
causal sequence of recovery, beneficial effects of cogni-
tive recovery may logically be expected to be direct or
indirect. The latter would require interaction with some
other condition. For example, improvement in specific
attentional or memory capabilities may have no direct
effects on behavioral functioning, but may nevertheless
enhance a patient's response to social skills training or
related rehabilitation interventions. This could be because
Cognitive Functioning
the cognitive abilities in question have limited importance
to ongoing performance of social skills but are critical to
acquisition or reacquisition of those skills.
For the purposes of this article, we conducted a series
of analyses to explore in greater detail the mechanisms of
the cognitive treatment effect reported by Spaulding et al.
(1998). The first step in these analyses was to identify all
statistically significant bivariate correlations between the
degree of improvement in social competence and mea-
sures of cognition, behavioral functioning, and symptoms.
The measure of social competence change is the change
score on the ATPSS overall performance scale, residual-
ized by linear regression to remove variance attributable
to the initial value of the score. Five measures of clinical
status (the five factors of the BPRS) and 13 neurocogni-
tive measures were analyzed as predictor variables. The
initial value and the residualized change score for each
brought the total in the predictor pool to 36. Five of the
pretreatment measures and two residualized change scores
showed significant correlations with the AIPSS change
score (p < 0.05, uncorrected for multiple correlations). In
addition, as expected from the significant cognitive treat-
ment effect, the amount of cognitive treatment provided
(expressed as a dose unit of 0 or 1, reflecting whether or
not the patient received 6 months of cognitive treatment)
was correlated with AIPSS change. Together, these eight
measures account for about 21 percent of the total change
in AIPSS performance (adjusted R2 = 0.206, p < 0.0036).
The measures fall in four separate domains: four mea-
sures of cognitive functioning at the start of treatment,
one measure of behavioral functioning before and during
treatment (the Nurses Observation Scale for Inpatient
Evaluation (NOSIE-30) Total Assets Scale; Honigfeld et
al. 1966), two measures of change in cognitive function-
ing, and the "dose" of cognitive treatment.
The sequence and direction of possible causal rela-
tionships among measures in the different domains can be
logically assumed; for example, the sequence of pretreat-
ment functioning, administration of treatment, and
changes during treatment is known. This allows construc-
tion of the path model shown in figure 1. A path model is
a sequentially ordered set of multiple regression formulae
constructed to represent causal, as opposed to merely cor-
relational, relationships among variables. The strengths
of such relationships are expressed as beta weights, as in
conventional multiple regression, and they represent the
unique contribution of each independent variable to the
dependent variables. When causal sequences are
unknown, alternative path models can be constructed and
evaluated to determine which model best fits the available
correlational data. When the possible causal pathways are
constrained by a known sequence of events in real time,
as in the present case, the path model that best fits the
281
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
available data can provide information about how specific
sets of variables influence an outcome.
The path model in figure 1 shows that the amount of
cognitive treatment (i.e., the "dose" of cognitive therapy,
expressed as a dosage unit of 1 or 0) directly influences
AIPSS performance, without mediation by any of the cog-
nitive measures in the protocol. This means that whatever
the cognitive mechanism of the treatment effect may be, it
is not identified in this model. Changes in COGLAB card
sorting random errors and performance on the Halstead
Reitan Trailmaking-Test B (Trails B) (Goldstein 1991)
performance are also associated with AIPSS change, but
this happens independently of the cognitive treatment.
Trails B change is associated in turn with pretreatment
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
levels of the visual form of the Rey Verbal Learning Test
(VLT; Kern and Green 1994) and the false alarm rate on
the COGLAB continuous performance test/Span of
Apprehension Test. The sizes of the correlations do not
permit much confidence in the relative contributions of
the pretreatment variables or the Trails B change. The
contributions of the card sorting change score and treat-
ment type, however, do permit a confident conclusion that
these two variables are independently associated with
change in AIPSS performance. This would be consistent
with (but does not prove) the hypothesis that improve-
ments in the cognitive functions measured by the card
sorting task enhance acquisition of social competence in
the course of rehabilitation. However, what is most sig-
nificant about this analysis is what it does not show: The
cognitive treatment effect does not operate through any of
the measured cognitive functions, and so its cognitive
mechanisms remain unknown.
The final step in these analyses was to employ inter-
action terms (Downs and Robertson 1991; Cortina 1993)
as independent (predictor) variables. An interaction term
in this context is the arithmetic product of a quantitative
expression of the treatment condition (the quantitative
"dose" of treatment) and any organismic variable (e.g., a
measure of cognitive functioning). The organismic vari-
able can be analyzed as it existed at the outset of treat-
ment, or it can be converted to a change score, that is, the
difference between pre and posttreatment scores. When
included as an independent (or "predictor") variable in a
multiple regression equation whose dependent (or "tar-
get") variable is also a change score, the interaction term
shows how the two different factors (the treatment and the
organismic variable) work together to produce change.
Interaction terms allow the possibility that a variable may
behave differently in the presence of another variable. For
example, cognitive functioning may affect acquisition of
social competence only in the presence of cognitive treat-
ment. This contrasts with conventional multiple regres-
sion, which can show only the unique contributions of the
Schizophrenia Bulletin, Vol. 25, No. 2, 1999 W.D. Spaulding et al.

Figure 1. Path model constructed from bivarlate correlations between change on the AlPSS total
score, other measures in the assessment protocol, and the "dose" (1 or 0 dose units) of cognitive
therapy

Pre-treatment status Changes during treatment Outcome

Change in COGLRB
Card Sorting

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016

Change
Rey ULT Trails B
uisual form

/ False alarms on \ Change In

\COGLRB CPT/Span / ~ HIPSS Total

[ N0SIE-3B thru
v. treatment period

WfllS arithmetic

Dose of cognitive
treatment
LURIS Digit Span

Note.Numbers on paths indicate beta weights. AlPSS = Assessment of Interpersonal Problem Solving Skills; CPT =. continuous perfor-
mance test; NOSIE = Nurses Observation Scale for Inpatient Evaluation; Trails B = Halstead Reitan Trailmaking Test B; VLT = Verbal
Learning Test; WAIS = Wechsler Adult Intelligence Scale.
*p < 0.05.
"p < 0.02.
*"p<0.01.

individual variables. Interactions are assumed to operate
over the course of the treatment period, so regression val-
ues cannot be ordered or sequenced, as in a path model.
However, a single regression formula can be constructed
that includes change scores, interaction terms, and treat-
ment dose, and indicates how those factors work together
over time.
In the Spaulding et al. (1998) data, 36 residualized
change scores can be computed including interactions
between treatment and pretreatment measures, and inter-

282
Cognitive Functioning Schizophrenia Bulletin, Vol. 25, No. 2, 1999

actions between treatment and residualized change scores.
Five of these were found to have significant bivariate cor-
relations with the AIPSS change score, and four are inter-
actions between pretreatment cognitive functioning and
treatment. The fifth is an interaction between changing
cognitive functioning and treatment. A multiple regres-
sion equation including these, the card sorting and Trails
B change scores, and treatment dose accounts for about
28 percent of the change in the AIPSS (adjusted R2 =
0.275, p < 0.0003). The regression statistics are summa-
rized in table 5.
Table 5. Multiple regression statistics for
memory functioning in situations demanding social com-
petence.
We repeated these analyses using a different outcome
variable, a composite measure of performance on four
assessments of information acquisition associated with
social and living skills training. These consisted of paper-
and-pencil quizzes and brief role plays, designed specifi-
cally to assess assimilation of information presented in
skill training modalities. Such assessments lack the eco-
logical validity of the AIPSS, but they are commonly used
to monitor progress in treatment and may reflect more
directly patients' ability to benefit from particular modali-
ties. The path model constructed from measures showing

Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016

prediction of AIPSS change (residualized to significant bivariate correlations with the skill acquisition
remove initial value effects) by Interaction terms, change score is shown in figure 2. Unlike the AIPSS
cognitive change, and treatment dose model, no pretreatment measures predicted the outcome
Variable Beta Significance1 measure, and cognitive treatment did not affect outcome.
Behavioral functioning had a significant impact on out-
Dose of cognitive treatment 0.25 0.018
come, part of which was mediated by reductions in the
Change on COG LAB Card
Sort task 0.18 0.090 BPRS disorganization and paranoia dimensions and part
Change on Trails B 0.13 0.214 by change on the Denman verbal memory score (Denman
Change on Rey VLT auditory 1984). The amount of change in paranoia and verbal
form x treatment dose 0.27 0.010 memory was associated in turn with behavioral function-
Pretreatment COGLAB false ing as measured by the NOSIE-30. The size of the beta
alarms x treatment dose 0.20 0.067 weight of the Denman score is sufficient to allow some
Pretreatment Denman confidence that it reflects a real benefit of cognitive
Nonverbal x treatment dose 0.11 0.383 change for performance on the skill measure.
Pretreatment NOSIE-30 x
Taken together, these analyses suggest that cognitive
treatment dose 0.09 0.486
Pretreatment Rey VLT visual recovery, especially of verbal learning and memory, has
form x treatment dose 0.02 0.904 beneficial effects on various aspects of acquisition of
Note.AIPSS o Assessment of Interpersonal Problem Solving social skill and competence in the course of rehabilitation.
Skills; Trails B - Halstead Reltan Trailmaking B; VLT = Verbal In some cases, exposure to cognitive treatment may be
Learning Test; NOSIE = Nurses Observation Scale for Inpatient necessary for patients to fully realize these benefits.
Evaluation.
However, substantial portions of the improvement in
'Significance of T, reflecting the probability that the variable's social skill and competence remain unexplained, whether
unique contribution to the overall prediction of the dependent vari-
able is zero. or not the improvement is associated with cognitive treat-
ment. It is hoped that further application of this analytic
The sizes of most of the interaction terms' beta strategy, with increasingly sophisticated cognitive assess-
weights do not allow confidence that they contribute ment protocols and with other outcome measures, will
uniquely to prediction of AIPSS change, but the size for identify additional mechanisms and will articulate the
the interaction of cognitive treatment with change on the importance of additional aspects of cognitive recovery.
auditory form of the Rey Verbal Learning Test allows For the time being, it appears that verbal learning and
confidencejhat at least one interaction term is associated memory are particularly importanLtargets fortreatment.
with outcome. The Rey VLT is also one of the measures
that showed nonspecific improvement (i.e., unrelated to
treatment condition) in the outcome study. While learning Possible Mechanisms of the Effects of
and memory performance as measured by the Rey Psychosocial Interventions in Chronic
improves independently of cognitive treatment, the signif- Schizophrenia
icant interaction term indicates that the improvement is
associated with outcome only if cognitive treatment is The foregoing results and discussion indicate a need for
provided. This suggests that cognitive therapy works in theoretical models to account for how and why cognitive
part by helping patients to use improving learning and improvements occur in response to psychosocial treat-

283
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
Figure 2. Path model constructed from bivariate correlations between the skill acquisition outcome
measure and other variables In the assessment protocol
Pre-treatment status Changes during treatment
Change In BPRS
Disorganization
Change In BPRS
Paranoia
.33**
/ N 0 S I E - 3 B thru
V treatment period
.34**
f Change in Denman
V Uerbal Memory
Note.Numbers on paths indicate beta weights. BPRS = Brief Psychiatric Rating Scale; NOSIE = Nurses Observation Scale for Inpatient
Evaluation.
*p < 0.05.
"p < 0.01.
ment in chronic schizophrenia. Discussions of cognitive
treatment effects (e.g., Bellack 1992; Corrigan and
Storzbach 1993; Brenner et al. 1994; Spaulding et al.
1998; Bellack et al. 1999, this issue) raise the possibility
of at least three types of mechanisms: prosthetic mecha-
nisms, in which new skills are established to replace or
compensate for impaired abilities; remediational mecha-
nisms, in which the impaired processes undergo actual
repair; and reorganizational mechanisms, in which envi-
ronmental conditions enhance functional reorganization of
processes disrupted by acute psychosis. In addition, edu-
cational, attributional, and self-instructional mechanisms
have been proposed, although these are usually associated
with cognitive-behavioral therapy aimed at the content of
cognition, independent of neurocognitive impairments
(for a review, see Alford and Correia 1994).
Prosthetic, educational, attributional, and self-instruc-
tional mechanisms could operate according to familiar
models of learning, conditioning, and cognitive develop-
ment. Remediational mechanisms pose a greater concep-
tual problem, and it is unclear what kind of model could
284
W.D. Spaulding et al.
Outcome
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
Change In composite
skill acquisition
assessment
account for normalization of an otherwise stable neu-
rocognitive impairment. Mechanisms to account for cog-
nitive disorganization and reorganization in schizophrenia
have been proposed, in terms of learned behavioral
response hierarchies (e.g., Broen and Storms 1966). At
least one specific brain mechanism is involved in con-
struction and maintenance of such organizational hierar-
chies (Houk and Wise 1993; Houk 1995), and it lies in
brain structures implicated in the etiology of schizophre-
nia. Transient physiological dysregulation of such a brain
mechanism, as in acute psychosis, could produce a linger-
ing disorganization of response hierarchies, whose reorga-
nization is enhanced by the environmental contingencies
and demands associated with psychosocial treatment.
Thus, the prospects seem good for credible models to
explain prosthetic and reorganizational cognitive treat-
ment effects, especially models that integrate neurophysi-
ological and neuropsychological mechanisms of learning,
conditioning, and behavioral organization.
A possible neuroendocrine mechanism of chronic-
course cognitive recovery has emerged from studies of
Cognitive Functioning
hypothalamic-pituitary-adrenal (HPA) functioning in
schizophrenia. The HPA axis has been increasingly impli-
cated in the etiology of schizophrenia, most extensively in
Walker and Diforio's (1997) neural diathesis-stress elabo-
ration of Walker's (1994) more general neurodevelopmen-
tal model. In this model, prenatal stress leads to high lev-
els of cortisol, causing hippocampal cellular dysplasia.
This renders the individual vulnerable to environmental
stress by creating a predisposition to HPA axis hyperactiv-
ity. HPA hyperactivity continues to interact with hip-
pocampal lesions, leading to further cellular damage.
HPA hyperactivity also contributes to expression of psy-
chotic symptoms through activation of subcortical
dopamine systems. Cortisol hyperactivity is associated
with acute schizophrenia (Albus et al. 1982; Christie et al.
1986; Copolov et al. 1989).
The neural diathesis-stress model describes the HPA's
role in onset and acute psychosis, but the model's implica-
tions for the chronic course are largely unexplored. HPA
activity is affected by antipsychotic drugs. Treatment and
rehabilitation environments are usually designed to reduce
"stress" and thereby affect HPA activity. Patients' beliefs,
attributions, and appraisals about themselves and their
control over their environment may undergo dramatic
changes after onset, and these factors are known to have
significant impact on HPA activity (Lazarus and Folkman
1984). HPA dysregulation can produce a failure of the
activating function of cortisol, resulting in a hypoactiva-
tion syndrome different in expression from the more
familiar hyperactivation/stress syndrome (Chrousos and
Gold 1992). It would not be inconsistent with the neural
diathesis-stress model to expect that a poorly regulated
HPA axis could result in cortisol hypoactivity in the
chronic, residual phase of schizophrenia, at least in some
patients. In fact, abnormally low levels of cortisol, some-
times to the point of obliterating normal diurnal cycles,
have been observed in chronic, stable schizophrenia
patients in institutional settings (Bhadrinath and
Girijashanker 1984; Spaulding et al. 1990; Partridge
1992; Fleming 1998).
The recent discovery of cortisol receptors in cortical
neurons (for a review, see Newcomer et al. 1994) indi-
cates that the activating function of cortisol involves cog-
nitive as well as physiological resources. Experimental
suppression of cortisol with dexamethasone in normal
subjects is associated with temporary impairments in
memory (Newcomer et al. 1994). Cortisol deficiencies
produce potentially reversible cognitive deficits in elderly
patients (Basavaraju and Phillips 1989).
Of the three studies in our lab that identified institu-
tionalized schizophrenia patients with cortisol hypoactiv-
ity, two (Spaulding et al. 1990; Partridge 1992) showed
some normalization of cortisol levels and fluctuations
285
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
associated with intensive psychiatric rehabilitation. The
third (Fleming 1998) showed that patients with low corti-
sol activity also show especially severe impairment in
some aspects of cognitive functioning. Taken together,
these findings suggest that in the chronic course of schizo-
phrenia, some patients experience a disruption of the reg-
ulatory relationship between cognitive and HPA function-
ing that produces a hypoactivation syndrome. This could
be the result of medication, environmental controls,
and/or the patient's appraisals and attributions all interact-
ing with an already compromised HPA system. The con-
sequences include residual cognitive impairment. A reha-
bilitative environment, with systematic and graduated
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
environmental demands for physiological and cognitive
activation and attention to the patient's attributions and
appraisals, may help reverse this process.
The data showing recovery of cognitive functioning
in schizophrenia, and the models that account for that
data, suggest a number of conditions that should be
expected to maximize recovery. These include environ-
mental engineering that maximizes the clarity and pre-
dictability of behavioral performance demands and conse-
quences, psychosocial interventions that exercise key
behavioral repertoires and their supporting executive-
level cognitive abilities, interventions that address
patients' appraisals of stressful events and their control
over them, and daily routines that enhance regulatory
cycles of organismic activation. Significant cognitive
benefits appear to accrue from conventional rehabilitation
modalities, including contingency management and skill
training, but modalities that specifically target cognitive
abilities have added value. However, the heterogeneity of
cognitive impairment suggests that individual differences
must be taken into account for optimal effect. Despite the
overriding role of executive-level cognitive reorganiza-
tion, different patients rely on different cognitive
resources and cope with different cognitive liabilities in
the reorganization process. A cognitive technology is
needed that is sensitive to such individual differences and
allows their consideration in construction of individual-
ized rehabilitation strategies.
Conclusions
Potential applications of cognitive science and technology
to schizophrenia go far beyond elucidation of develop-
mental etiological processes. Cognitive functioning is a
significant source of individual variation among patients
and is importantly related to outcome. Cognitive impair-
ments can change over the acute, post-acute, and chronic
course of schizophrenia, in response to psychosocial as
well as pharmacological treatment and rehabilitation. The
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
findings of research using the paradigms of experimental
psychopathology and neuropsychology suggest that a cog-
nitive technology can be perfected that would contribute
significantly to diagnosis, treatment and rehabilitation
planning, evaluation of patients' response to treatment,
and the design of future treatment modalities. For some
applications, it may be important to revise traditional
approaches to behavioral, psychological, and neuropsy-
chological assessment to accommodate the need for longi-
tudinal tracking of recovery over the time frames in which
pharmacological and psychosocial treatments exert their
effects. Models for systematically applying this technol-
ogy to clinical assessment and decision making are
already beginning to appear (see Spaulding et al. 1994).
To fully realize these potentials, we need research on the
cognitive psychopathology and neuropsychology of schiz-
ophrenia that systematically addresses the longitudinal
course of the disorder, within and across its acute, post-
acute, and chronic residual phases, in the context of state-
of-the-art biopsychosocial treatment and rehabilitation.
References
Albus, M.; Ackenheil, M.; Engel, R.; and Mueller, F.
Situational reactivity of autonomic functions in schizo-
phrenic patients. Psychiatry Research, 6:361-370, 1982.
Alford, B., and Correia, C. Cognitive therapy of schizo-
phrenia: Theory and empirical status. Behavior Therapy,
25:17-33, 1994.
Basavaraju, N., and Phillips, S. Cortisol deficient state:
A cause of reversible cognitive impairment and delirium
in the elderly. Journal of the American Geriatric Society,
37:49-51, 1989.
Bellack, A.S. Cognitive rehabilitation for schizophrenia:
Is it possible? Is it necessary? Schizophrenia Bulletin,
18:43-50, 1992.
Bellack, A.S.; Gold, J.M.; and Buchanan, R.W. Cognitive
rehabilitation for schizophrenia: Problems, prospects, and
strategies. Schizophrenia Bulletin, 25:257-274, 1999.
Bhadrinath, B., and Girijashanker, K. DST in mania:
Prolonged suppression? British Journal of Psychiatry,
145:555-559, 1984.
Bowen, L.; Wallace, C ; Glynn, S.; Nuechterlein, K.;
Lutzger, J.; and Kuehnel, T. Schizophrenics' cognitive
functioning and performance in interpersonal interactions
and skills training procedures. Journal of Psychiatry
Research, 28:289-301, 1994.
Braff, D.L., and Sacuzzo, D.P. The effect of antipsychotic
medication on speed of information processing in schizo-
phrenia. American Journal of Psychiatry, 139:1127-1130,
1982.
286
W.D. Spaulding et al.
Brenner, H.D.; Hodel, B.; Roder, V.; and Corrigan, P.
Treatment of cognitive dysfunctions and behavioral
deficits in schizophrenia. Schizophrenia Bulletin,
18:21-26,1992.
Brenner, H.; Roder, V.; Hodel, B.; Kienzle, N.; Reed, D.;
and Liberman, R. Integrated Psychological Therapy for
Schizophrenic Patients. Toronto, Ontario: Hogrefe &
Huber, 1994.
Brenner, H.D.; Stramke, W.; Hodel, B.; and Rui, C. "A
Treatment Program for Impaired Cognitive Functions
Aimed at Preventing Chronic Disability Among
Schizophrenic Patients." Paper presented at World
Psychiatric Association Symposium, Baltimore, MD,
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
1983.
Broen, W., and Storms, L. Lawful disorganization: The
process underlying a schizophrenic syndrome.
Psychological Review, 73:265-279, 1966.
Christie, J.; Whalley, L.; Blackwood, H.; Bennie, J.;
Blackburn, I.; and Fink, G. Disturbed endocrine function
in psychosis. British Journal of Psychiatry, 155:455^61,
1986.
Chrousos, G., and Gold, P. The concepts of stress and
stress system disorders. Journal of the American Medical
Association, 267:1244-1252, 1992.
Copolov, D.; Rubin, R.; Stuart, G.; Poland, R.; Mander,
A.; Sashidharan, S.; Whitehouse, A.; Blackburn, I.;
Freeman, C ; and Blackwood, D. Specificity of the sali-
vary dexamethesaone suppression test across psychiatric
diagnoses. Biological Psychiatry, 25:879-893, 1989.
Corrigan, P., and Storzbach, D. The ecological validity of
cognitive rehabilitation for schizophrenia. Journal of
Cognitive Rehabilitation, 11:14-21, 1993.
Cortina, J. Interaction, nonlinearity, and multicollinearity:
Implications for multiple regression. Journal of
Management, 19:915-922, 1993.
Denman, S. Denman Neuropsychological Memory Scale:
A clinical assessment of immediate recall, short-term
memory and long-term memory in verbal and nonverbal
areas. Charleston, SC: Sydney B. Denman, 1984.
Donohoe, C ; Carter, M.; Bloem, W.; Hirsch, G.; Laasi,
N.; and Wallace, C.J. Assessment of interpersonal prob-
lem solving skills. Psychiatry, 53(4):329-339, 1990.
Downs, W., and Robertson, J. Random versus clinical
samples: A question of inference. Journal of Social
Service Research, 14:57-83, 1991.
Erickson, R.C. Neuropsychological assessment and the
rehabilitation of persons with severe psychiatric disabili-
ties. Rehabilitation Psychology, 33:15-25, 1988.
Erickson, R. Neuropsychological assessment and consul-
tation in psychiatric rehabilitation. In: Spaulding, W., ed.
Cognitive Functioning
Cognitive Technology in Psychiatric Rehabilitation.
Lincoln, NE: University of Nebraska Press, 1994.
pp. 2 7 ^ 8 .
Erickson, R.C., and Binder, L.M. Cognitive deficits
among functionally psychotic patients: A rehabilitative
perspective. Journal of Clinical and Experimental
Neuropsychology, 8:257-274, 1986.
Fleming, S. "Neuroendocrine Concomitants of Neuro-
cognitive Impairment in Chronic Schizophrenia."
Doctoral dissertation, University of NebraskaLincoln,
1998.
Gilbertson, M.; Yao, J.; and Van Kammen, D. Memory
and plasma HVA changes in schizophrenia: Are they
episode markers? Biological Psychiatry, 35:203-206,
1994.
Gjerde, P.F. Attentional capacity dysfunction and arousal
in schizophrenia. Psychological Bulletin, 93:57-72, 1983.
Gold, J.; Hermann, B.; Randolph, C ; and Wyler, A.
Schizophrenia and temporal lobe epilepsy: A neuropsy-
chological analysis. Archives of General Psychiatry,
51:265-272, 1994.
Goldstein, G. Comprehensive neuropsychological test
batteries and research in schizophrenia. In: Steinhauer, S.;
Gruzelier, J.; and Zubin, J., eds. Handbook of Schizo-
phrenia: Neuropsychology, Psychopathology and In-
formation Processing. Vol. 5. Amsterdam, The Nether-
lands: Elsevier, 1991.
Green, M. What are the functional consequences of neu-
rocognitive deficits in schizophrenia? American Journal
of Psychiatry, 153:321-330, 1996.
Green, M. Schizophrenia From a Neurocognitive Per-
spective: Probing the Impenetrable Darkness. Boston,
MA: Allyn & Bacon, 1998.
Gruzelier, J. Synthesis and critical review of the evidence
for hemisphere asymmetries of function in psychopathol-
ogy. In: Gruzelier, J., and Flor-Henry, P., eds. Hemisphere
Asymmetries of Function in Psychopathology. New York,
NY: Elsevier/North-Holland, 1979.
Gruzelier, J.H. Hemispheric imbalance: Syndromes of
schizophrenia, premorbid personality, and neurodevelop-
mentaHnfluences.-In; Steinhauerr S.R.; Gruzelier, J.H.;
and Zubin, J., eds. Handbook of Schizophrenia: Vol. 5.
Neuropsychology, Psychopathology, and Information
Processing. Amsterdam, The Netherlands: Elsevier
Science, 1991. pp. 599-650.
Harding, C ; Zubin, J.; and Strauss, J. Chronicity in
schizophrenia: Revisited. British Journal of Psychiatry,
161(Suppl. 18):27-37, 1992.
Hogarty, G.; Anderson, C ; and Reiss, D. Family psy-
choeducation, social skills training, and medication in
287
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
schizophrenia: The long and short of it. Psychopharma-
cology Bulletin, 23:12-13, 19.87.
Honigfeld, G.; Gillis, R.; and Klett, J. NOSIE-30: A treat-
ment-sensitive ward behavior scale. Psychological
Review, 19:180-182, 1966.
Houk, J. Information processing in modular circuits link-
ing basal ganglia and cerebral cortex. In: Houk, J.; Davis,
J.; and Bieser, D., eds. Models of Information Processing
in the Basal Ganglia. Cambridge, MA: MIT Press, 1995.
pp. 3-9.
Houk, J., and Wise, S. Outline for a theory of motor
behavior. In: Rudomin, P.; Arbib, M.; and Cervantes-
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
Perez, F., eds. From Neural Networks to Artificial
Intelligence. Heidelberg, Germany: Springer-Verlag,
1993. pp. 452-470.
Kern, R.S., and Green, M.F. Cognitive prerequisites of
skill acquisition in schizophrenia: Bridging micro- and
macro-levels of processing. In: Spaulding, W., ed.
Cognitive Technology in Psychiatric Rehabilitation.
Lincoln, NE: University of Nebraska Press, 1994.
pp. 49-66.
Knight, R.G., and Russell, P.N. Global capacity reduction
and schizophrenia. British Journal of Social and Clinical
Psychology, 17:275-280, 1978.
Lam, M. "Evaluating the Utility of a Card Sorting Task
as a Prognostic Indicator for Psychiatric Rehabilitation
Over the Course of Schizophrenia." Doctoral disserta-
tion, Department of Psychology, University of
NebraskaLincoln, 1997.
Lazarus, R.S., and Folkman, S. Stress, Appraisal, and
Coping. New York, NY: Springer, 1984.
Lee, M.; Thompson, P.; and Meltzer, H. Effects of cloza-
pine in cognitive function in schizophrenia. Journal of
Clinical Psychiatry, 55(Suppl. B):82-87, 1994.
Liberman, R.P.; Nuechterlein, K.H.; and Wallace, C.J.
Social skills training and the nature of schizophrenia. In:
Curran, J.P., and Monti, P.M., eds. Social Skills Training:
A Practical Handbook for Assessment and Treatment.
New York, NY: Guilford, 1982. pp. 5-56.
Meltzer, H.; Thompson, P.; Lee, M.; and Ranjan, R.
Neuropsychological deficits in schizophrenia: Relation to
social function and effect of antipsychotic drug treatment.
Neuropsychopharmacology, 14(Suppl. 3):27-33, 1996.
Newcomer, J.; Craft, S.; Hershey, T.; Askins, K.; and
Bardgett, M. Glucocorticoid-induced impairment in
declarative memory performance in adult humans.
Journal of Neuro science, 14:2047-2053, 1994.
Partridge, S. "Cortisol Functioning and the Treatment
Environment in Chronic Schizophrenia." Doctoral disser-
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
tation, Department of Psychology, University of
NebraskaLincoln, 1992.
Penn, D. Cognitive rehabilitation of social deficits in
schizophrenia: A direction of promise or following a
primrose path? Psychosocial Rehabilitation Journal,
15:27-41, 1991.
Penn, D.; Mueser, K.; Spaulding, W.; Hope, D.; and Reed,
D. Information processing and social competence in
chronic schizophrenia. Schizophrenia Bulletin, 21:269-
281, 1995.
Penn, D.; Spaulding, W.; Reed, D.; and Sullivan, M. The
relationship of social cognition to ward behavior in
chronic schizophrenia. Schizophrenia Research, 20:327-
335, 1996.
Peveler, R., and Fairburn, C. Measurement of neurotic
symptoms by self-report questionnaire: Validity of the
SCL-90R. Psychological Medicine, 20:873-879, 1990.
Plasky, P. Antidepressant usage in schizophrenia.
Schizophrenia Bulletin, 17:649-657, 1991.
Reed, D.; Sullivan, M.; Penn, D.; Stuve, P.; and
Spaulding, W. Assessment and treatment of cognitive
impairments. In: Liberman, R.P., ed. Effective Psychiatric
Rehabilitation. San Francisco, CA: Jossey-Bass, 1992.
pp. 7-20.
Silver, H., and Nassar, A. Fluvoxamine improves nega-
tive symptoms in treated chronic schizophrenia.
Biological Psychiatry, 31:698-704, 1992.
Silverstein, M.; Harrow, M.; and Bryson, G. Neuro-
psychological prognosis and clinical recovery. Psychiatry
Research, 52:265-272, 1994.
Silverstein, S.; Knight, R.; Schwartzkopf, S.; West, L.;
Osborne, L.; and Kamim, D. Configuration and context
effects in perceptual organization in schizophrenia.
Journal of Abnormal Psychology, 105:410-420, 1996.
Siris, S. Assessment and treatment of depression in schiz-
ophrenia. Psychiatric Annals, 24:463467, 1994.
Spaulding, W. The relationships of some information pro-
cessing factors to severely disturbed behaviors. Journal of
Nervous and Mental Disease, 166:417^28, 1978.
Spaulding, W. Spontaneous and induced changes in cog-
nition during psychiatric rehabilitation. In: Cromwell,
R.L., and Snyder, C.R., eds. Schizophrenia: Origins,
Processes, Treatment, and Outcome. New York, NY:
Oxford Press, 1993. pp. 299-312.
Spaulding, W.; Garbin, C ; and Dras, S. Cognitive func-
tioning in chronic psychiatric patients and schizotypal
college students. Journal of Nervous and Mental Disease,
177:717-728, 1989.
288
W.D. Spaulding et al.
Spaulding, W.; Reed, D.; Storzbach, D.; Sullivan, M.;
Weiler, M; and Richardson, C. The effects of a remedia-
tional approach to cognitive therapy for schizophrenia. In:
Wykes, T, ed. Outcome and Innovation in Psychological
Treatment of Schizophrenia. London, England: John
Wiley, 1998. pp. 145-160.
Spaulding, W.D.; Storms, L.; Goodrich, V.; and Sullivan,
M. Applications of experimental psychopathology in psy-
chiatric rehabilitation. Schizophrenia Bulletin, 12:560-
577, 1986.
Spaulding, W.; Sullivan, M.; Weiler, M.; Reed, D.;
Richardson, C ; and Storzbach, D. Changing cognitive
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
functioning in rehabilitation of schizophrenia. Acta
Psychiatrica Scandinavica, 90(Suppl. 384): 116-124,
1994.
Spaulding, W.; Wyss, H.; and Littrell, R. Training
psychophysiological self-regulation skills in psychiatric
rehabilitation. Psychosocial Rehabilitation Journal,
13:37-39,1990.
Ventura, J.; Green, M.; Shaner, A.; and Liberman, R.
Training and quality assurance with the BPRS.
International Journal of Methods in Psychiatric Research,
3:221-244, 1993.
Wagman, A., and Wagman, W. On the Wisconsin. In:
Walker, E.; Dworkin, R.; and Cornblatt, B., eds. Progress
in Experimental Personality Research. New York, NY:
Springer, 1992. pp. 162-182.
Walker, E. Developmentally moderated expressions of
the neuropathology underlying schizophrenia.
Schizophrenia Bulletin, 20:453^180, 1994.
Walker, E., and Diforio, D. Schizophrenia: A neural
diathesis-stress model. Psychological Review, 104:667-
684, 1997.
Wallace, C.J., and Liberman, R.P. Social skills training
for patients with schizophrenia: A controlled clinical trial.
Psychiatry Research, 15:239-247, 1985.
Weilage, M. "Cost Effectiveness of an Intensive
Psychiatric Rehabilitation Program." Doctoral disserta-
tion, Department of Psychology, University of
NebraskaLincoln, 1997.
Weinberger, D. On the plausibility of "the neurodevelop-
mental hypothesis" of schizophrenia. A new understand-
ing: Neurological basis and long-term outcome of schizo-
phrenia. Neuropsychopharmacology, 14:1-11, 1996.
Weinberger, D., and Lipska, B. Cortical maldevelopment,
antipsychotic drugs, and schizophrenia: A search for
common ground. Schizophrenia Research, 16:87-110,
1995.
Cognitive Functioning
Wykes, R., and Dunn, G. Cognitive deficit and the pre-
diction of rehabilitation success in a chronic psychiatric
group. Psychological Medicine, 22:389-398, 1992.
Wykes, T.; Start, E.; and Katz, R. The prediction of reha-
bilitative success after three years: The use of social,
symptom, and cognitive variables. British Journal of
Psychiatry, 157:865-870, 1990.
Acknowledgments
This article was supported in part by National Institute of
Mental Health grant R01 MH-44756-01. The authors also
acknowledge the cooperation and support of the adminis-
tration of Lincoln Regional Center and the staff of the
Center's Community Transition Program and Short-term
Care Program.
289
Schizophrenia Bulletin, Vol. 25, No. 2, 1999
The Authors
William D. Spaulding, Ph.D., is Professor of Psychology,
University of NebraskaLincoln, and Clinical Psycholo-
gist, Community Transition Program, Lincoln Regional
Center, Lincoln, NE. Shelley K. Fleming, M.A., is a grad-
uate student in the Department of Psychology, University
of NebraskaLincoln. Done Reed, Ph.D., is a clinical
research associate, Department of Psychology, University
of NebraskaLincoln, and clinical psychologist, Com-
munity Transition Program, Lincoln Regional Center.
Mary Sullivan, M.S.W., is a clinical associate, Department
of Psychology, University of NebraskaLincoln, and
Program Director, Community Transition Program,
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016
Lincoln Regional Center. Daniel Storzbach, Ph.D., is a
research fellow, Department of Psychiatry and Behavioral
Science, Veterans Affairs Medical Center, Portland, OR.
Mona Lam, M.A., is a graduate student in the Department
of Psychology, University of NebraskaLincoln.
Downloaded from http://schizophreniabulletin.oxfordjournals.org/ by guest on December 12, 2016