SCHRES-05178; No of Pages 7

Schizophrenia Research xxx (2012) xxxxxx

Contents lists available at SciVerse ScienceDirect

Schizophrenia Research
journal homepage: www.elsevier.com/locate/schres

Cognitive reserve as a predictor of two year neuropsychological performance in early

onset rst-episode schizophrenia
Elena de la Serna a, b,, Susana Andrs-Perpi a, b, e, Olga Puig b, Inmaculada Baeza a, b, c, e, Igor Bombin a, d,
David Bartrs-Faz e, f, Celso Arango a, d, Ana Gonzalez-Pinto a, g, Mara Parellada a, d, Mara Mayoral a, d,
Montserrat Graell a, h, Soraya Otero i, Joan Guardia j, Josena Castro-Fornieles a, b, e, f
a
Centro de Investigacin Biomdica en Red de Salud Mental, CIBERSAM, Spain
b
Department of Child and Adolescent Psychiatry and Psychology, Institut Clinic de Neurocincies, Hospital Clnic Universitari, Barcelona, Spain
c
Programa Esquizofrenia Clinic, Spain
d
Adolescent Unit, Department of Psychiatry, Hospital General Universitario Gregorio Maran, Madrid, Spain
e
Institut d'Investigaci Biomdica August Pi i Sunyer, IDIBAPS, Barcelona, Spain
f
Departament de Psiquiatria i Psicobiologia Clnica, Facultat de Medicina, Universitat de Barcelona, Barcelona, Spain
g
Department of Psychiatry, Hospital Santiago Apstol, EHU/University of the Basque County, Vitoria, Spain
h
Section of Child and Adolescent Psychiatry and Psychology, Hospital Infantil Universitario Nio Jess, Madrid, Spain
i
Child and Adolescent Psychiatry Unit, Department of Psychiatry, Hospital Universitario Marques de Valdecilla, Santander, Cantabria, Spain
j
Methodology and Behavioral Sciences, University of Barcelona, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: The concept of cognitive reserve (CR) has been dened as individual differences in the efcient
Received 14 October 2011 utilization of brain networks which allow some people to cope better than others with brain pathology. CR
Received in revised form 3 October 2012 has been developed mainly in the eld of aging and dementia after it was observed that there appears to
Accepted 25 October 2012 be no direct relationship between the degree of brain pathology and the severity of clinical manifestations
Available online xxxx
of this damage. The present study applies the concept of CR to a sample of children and adolescents with a
rst episode of schizophrenia, aiming to assess the possible inuence of CR on neuropsychological perfor-
Keywords:
Cognitive reserve
mance after two year follow-up, controlling for the inuence of clinical psychopathology.
Schizophrenia Methods: 35 patients meeting DSM-IV criteria for schizophrenia or schizoaffective disorder (SSD) and 98
First episode healthy controls (HC) matched for age and gender were included. CR was assessed at baseline, taking into
Psychosis account premorbid IQ, educationaloccupational level and leisure activities. Clinical and neuropsychological
Cognition assessments were completed by all patients at two year follow-up.
Results: The CR proxy was able to predict working memory and attention at two year follow-up. Verbal
memory and cognitive exibility were not predicted by any of the variables included in the regression model.
The SSD group obtained lower scores than HC on CR. CR measures correctly classied 79.8% of the sample as
being SSD or HC.
Conclusions: Lower scores on CR were observed in SSD than in HC and the CR measure correctly classied a high
percentage of the sample into the two groups. CR may predict SSD performance on working memory and atten-
tion tasks.
2012 Elsevier B.V. All rights reserved.

1. Introduction
The concept of cognitive reserve (CR) has developed in the eld of
aging and dementia after it was observed that people with the same
amount of brain damage could show different clinical expressions
depending on their compensatory capacity. CR has been dened as
individual differences in how people process tasks which allow some
Corresponding author at: Department of Child and Adolescent Psychiatry and
Psychology, Hospital Clnic Universitari de Barcelona, C/ Villarroel, 170, Barcelona 08036,
Spain. Tel./fax: +34 93 2279974.
E-mail address: eserna@clinic.ub.es (E. de la Serna).
to cope better than others with brain pathology (Stern, 2009). This
concept is focused on the ability to optimize performance based on
more efcient brain network utilization, and may vary depending on
environment and lifetime exposure to certain environmental factors.
Although it has been used primarily in the elds of dementia and
brain injury, this concept could potentially be applicable to a whole
range of neurological and psychiatric conditions. For example, CR
could play an important role in the expression of symptoms and func-
tional outcome of schizophrenia patients (Barnett et al., 2006). Studies
have shown that patients with schizophrenia with better performance
on cognitive tests have better social and functional outcomes (Green
et al., 2000; Munro et al., 2002). Holthausen et al. (2002) studied a

0920-9964/$ see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.schres.2012.10.026

Please cite this article as: de la Serna, E., et al., Cognitive reserve as a predictor of two year neuropsychological performance in early onset rst-
episode schizophrenia, Schizophr. Res. (2012), http://dx.doi.org/10.1016/j.schres.2012.10.026
2 E. de la Serna et al. / Schizophrenia Research xxx (2012) xxxxxx
sample of 118 rst episode of psychosis patients and classied them
according to the presence or absence of cognitive decits. Results
showed that patients with normal cognitive functioning had higher
scores on IQ measures and higher educational levels than patients
with cognitive difculties. The authors suggested that the observed dif-
ferences may have been due to the higher compensation capacity or
higher CR of the group with normal cognitive functioning.
An important caveat when investigating CR is the absence of any
single measure of this concept. Variables such as occupational and ed-
ucational attainment, leisure activity and IQ have been used as CR
proxies (Stern, 2009). Staff et al. (2004) tested three possible CR
proxies (head size, education and occupational attainment) and ob-
served that while education and occupational attainment contributed
to CR, intracranial volume did not. There is no consensus as to which
CR measures (premorbid IQ, education, occupation or leisure activi-
ties) could be the most representative. In the present study, CR
was measured with the main CR proxies used in previous literature
(i.e. Scarmeas and Stern, 2003; Stern et al., 2005), premorbid IQ,
which is thought to partially reect an innate reserve capacity, and
educationaloccupational attainment and leisure activity, which
more directly reect lifetime exposure to particular environmental
factors that help develop mental capacities.
Studies in child and adolescent samples with a rst episode of
schizophrenia have observed a broad range of neuropsychological dif-
culties in cognitive domains such as attention, working memory, exec-
utive function, verbal learning and memory (Kenny et al., 1997;
Fagerlund et al., 2006; Mayoral et al., 2008; Zabala et al., 2010). Most
of these decits have been associated in young adults with premorbid
functioning and educational and occupational levels (Silverstein et al.,
2002; Norman et al., 2005; Rund et al., 2007), which are considered
part of CR. However, studies in young patients have also observed that
clinical manifestations such as negative symptoms play an important
role in neuropsychological performance (Bilder et al., 2000; Fitzgerald
et al., 2004). Taking this previous research into account, the aim of the
present study was, rst, to employ the concept of CR with a sample of
children and adolescents with a rst episode of schizophrenia in order
to compare them to control subjects. The second aim was to assess
the inuence of CR on neuropsychological performance after a two
year follow-up of schizophrenia patients while controlling for the inu-
ence of clinical psychopathology.
We hypothesized that the schizophrenia spectrum disorder group
(SSD) would show lower CR measures than healthy controls (HC).
Moreover, we expected that neuropsychological variables after two
years could be predicted by baseline CR measures.
2. Methods
This research was part of the Child and Adolescent First-episode
Psychosis study (CAFEPS), a multi-center, longitudinal study aimed at
evaluating different clinical, neuropsychological and biological factors,
as well as treatment and prognostic factors in these patients; the
corresponding methodology has been described previously in detail
(Castro-Fornieles et al., 2007). The CAFEPS study included 110 patients
aged between 9 and 17 years diagnosed with a rst episode of psycho-
sis, and 98 matched healthy controls. Patients were recruited from child
and adolescent psychiatry units at six university hospitals and assessed
by mental health professionals with experience diagnosing and evaluat-
ing subjects with semi-structured interviews and clinical scales.
Inclusion criteria for patients were: a) an onset of positive psychosis
symptoms less than 6 months prior to baseline assessment and b) age
between 7 and 17 years. Exclusion criteria included: a) presence of an-
other concomitant Axis I disorder at the time of assessment that could
account for the psychotic symptoms, including substance-induced psy-
chotic disorder, post-traumatic stress disorder, or acute stress disorder;
b) IQ below 70 with impaired functioning; c) pervasive developmental
disorder d) neurological disorders, including history of head trauma
Please cite this article as: de la Serna, E., et al., Cognitive reserve as a predictor of two year neuropsychological performance in early onset rst-
episode schizophrenia, Schizophr. Res. (2012), http://dx.doi.org/10.1016/j.schres.2012.10.026
with loss of consciousness; and e) pregnancy. Occasional and regular
substance use was not an exclusion criterion if positive symptoms
persisted for more than two weeks after a negative urine toxicology
test and a substance-induced psychotic disorder was not diagnosed.
Socioeconomic status of the whole sample was estimated with the
Hollingshead Redlich Scale (Hollingshead and Redlich, 1958).
The healthy control group was matched by age and gender to pa-
tients. Sample recruitment and description of both patients and healthy
controls have been detailed previously (Castro-Fornieles et al., 2007).
The study was approved by the Ethical Review Board of each hos-
pital. All patients and controls and their parents or legal guardians
provided written informed consent.
2.1. Subjects
Patients were given clinical and CR assessments at baseline and
neuropsychological assessments at two year follow-up. To homoge-
nize the sample and guarantee the stability of the diagnosis, only sub-
jects diagnosed with schizophrenia or schizoaffective disorder (SSD)
at the two-year follow-up were included in the study. This group
comprised only 35 patients from the total CAFEPS sample. A healthy
control group (HC) of 98 subjects matched for age and gender was
also recruited.
2.2. Clinical assessment
Clinical assessment at baseline consisted of:
- Kiddie Schedule for Affective Disorders and Schizophrenia, Present
and Lifetime version (K-SADS-PL) (Kaufman et al., 1997) using the
Spanish validated adaptation; This is a semi-structured interview,
which was administered by psychiatrists trained in the use of the
instrument and in the assessment of children and adolescents.
Parents and children from both the patients group and control
group were interviewed separately.
- Positive and Negative Syndrome Scale (PANSS): This is a 30 item
rating scale which aims to assess the symptom severity of patients
with psychosis. It is subdivided into three subscales-positive, nega-
tive and general psychopathology- and a total score (Kay et al.,
1987; Peralta and Cuesta, 1994). Each subscale is evaluated from 1
to 7 according to the severity of the symptoms.
2.3. Assessment of cognitive reserve
Our determination of the main proxies of CR was based on recent lit-
erature in the eld, with special consideration given to the areas most
commonly included in cognitive reserve questionnaires (Bartres-Faz
et al., 2009; Sole-Padulles et al., 2009; Bosch et al., 2010). As a result,
the CR measure was composed of an estimation of premorbid IQ and a
measure of educationoccupation levels and lifetime leisure-social
activities.
- Premorbid IQ was assessed between 4 and 8 weeks after admission
when patients had reached clinical stability. Premorbid IQ was esti-
mated using the Vocabulary subtest of the Wechsler Adult Intelligence
ScaleIII Revised (WAIS III) (Wechsler, 2001) or the Wechsler Intel-
ligence Scale for ChildrenRevised (WISC-R) (Wechsler, 1974),
depending on the subject's age. Direct scores of the subscale were
translated to standard scores, which have a mean of 10 and a stan-
dard deviation of 3.
- Educationoccupation (EO). Educationoccupation was assessed
taking into account the number of years of obligatory education
that subjects had completed, school performance before the begin-
ning of the disorder, parents' educational level and questions about
the children's development in terms of language, reading, writing
and motor functions. EO could have values between 0 (low level
of CR) and 33 (high CR).
 E. de la Serna et al. / Schizophrenia Research xxx (2012) xxxxxx
- Lifetime leisure and social activities (LS) were assessed by asking
the parents about the intellectual pursuits, hobbies, peer relation-
ships and sociability of their children. LS could have scores between
0 (low level of CR) and 30 (high level of CR).
A composite CR score which included premorbid IQ, educational
occupational and leisure activities was obtained for each subject
using a conrmatory factor analysis.
2.4. Neuropsychological assessment
Cognitive assessment was performed at two-year follow-up. All
direct scores were standardized and transformed into Z scores,
which have a mean of 0 and a standard deviation of 1 based on the
performance of the control group at baseline. The cognitive domains
assessed were working memory, attention, verbal memory and cogni-
tive exibility, which have been considered as separable cognitive
factors in previous studies (Nuechterlein et al., 2004).
- The working memory domain included the scores obtained on
digits forward, digits backward and LetterNumber Sequencing
of the Wechsler Adult Intelligence ScaleIIIRevised (WAIS III)
(Wechsler, 2001) and the Trail Making Test part B (Reitan and
Wolfson, 1985). In Digits forward the subject must repeat a series
of numbers in the same order as has been read by the examiner.
The second task requires the subject to say in reverse order the digits
that have been read by the examiner. In the letternumber sequenc-
ing task, the examiner reads a list of letters and numbers and asks the
subject to say rst the numbers, from lowest to highest, and then the
letters, in alphabetical order. In the Trail Making Test part B (TMT-B)
subjects have to draw lines connecting, in alternating order, numbers
(from low to high) and letters (in alphabetical order) in the shortest
possible time. Time required to complete the task was used as a mea-
sure of working memory.
- The attention domain included the TMT-A (Reitan and Wolfson,
1985), and the Continuous Performance Test (CPT) (Conners,
2000). In the rst task the subject is given a sheet of paper with
the numbers 1 through 25 on it, and is instructed to draw lines
connecting the numbers as quickly as possible. The time taken to
complete the task is recorded. In the CPT test, the subject is required
to press the space bar every time any letter other than X appears on
the computer screen. When an X appears, the subject must inhibit
his/her response. Reaction time and correct responses were used in
the statistical analysis.
- The verbal memory and learning domain was assessed with the
TAVEC (Verbal Learning Test-Complutense Spain), the Spanish
adaptation of the California Verbal Learning Test (CVLT), which pro-
vides a learning curve and immediate and delayed memory scores
(Benedet and Alejandre, 1998).
- Lastly, the cognitive exibility domain included the number of mis-
takes, perseverative errors and perseverations on the Wisconsin
Card Sorting Test (WCST) (Heaton et al., 1997) and the interference
part of the Stroop test (Golden, 1978). The WCST is a measure of
executive function that requires planning strategies and cognitive
exibility in order to change the use of learned strategies. The inter-
ference part of the Stroop test measures the subject's ability to in-
hibit an automatic predominant response.
2.5. Statistical analysis
Categorical socio-demographic variables were analyzed using the
Pearson's chi-square test, whereas continuous variables were com-
pared between the two groups using the Student's t test.
To test the normality of the sample distribution, the Kolmogorov
Smirnov test was used, together with the Levene test to assess the
equality of variances.
Please cite this article as: de la Serna, E., et al., Cognitive reserve as a predictor of two year neuropsychological performance in early onset rst-
episode schizophrenia, Schizophr. Res. (2012), http://dx.doi.org/10.1016/j.schres.2012.10.026
3
Firstly, in order to conrm the factorial structure of CR and the
architecture of cognitive domains, two conrmatory factor analysis
(CFA) were performed in the control sample. Both analysis were
conducted using EQS 6.1 with the asymptotic free distribution meth-
od for the analysis of CR and with maximum likelihood estimation
with robust estimation for the cognitive domains analysis. Six differ-
ent goodness-of-t statistics were used to assess the t between the
hypothesized models and the data, including 2, the 2/df ratio, the
Akaike Information Criterion (AIC), BentlerBonet Normed Fit Index
(BBNFI) and Root Mean-Square Error of Approximation (RMSEA).
Secondly, we aimed to test the predictive capacity of CR for each diag-
nostic group (SSD or HC), so a logistic regression analysis (enter-method),
was performed.
Finally, to assess the predictive value of CR on cognitive domains at
two-year follow-up, a linear regression analysis was performed in the
SSD group with CR as a potential predictor. To test whether our CR
proxy was a better predictor of cognitive functions than premorbid IQ,
a new linear regression was conducted with premorbid IQ as a predic-
tor. Taking into account the inuence of clinical variables on cognitive
performance (Basso et al., 1998; Bilder et al., 2000; Fitzgerald et al.,
2004; Good et al., 2004), total scores of PANSS were also included in
both regression models. All analysis were performed using SPSS 18.0
and signicance was set at the p b 0.05 level.
3. Results
3.1. Socio-demographic characteristics of the sample
No signicant differences were found between SSD and healthy
controls in age, gender or parental socioeconomic status. Table 1
shows socio-demographic characteristics of the sample.
3.2. Conrmatory factor analysis of CR
The CFA of CR was based on data from 98 healthy controls. We
hypothesized the existence of a three factor structure for CR which in-
cludes: premorbid IQ (F1), EducationOccupation (F2) and Lifetime
leisure activities (F3). The values obtained on the goodness-of-t
statistics indicate a good t between the model and the observed
data. Thus, the CFA conrm the presence of a three factor structure
of CR in our sample. The correlation between latent variables showed
signicant results specically for rF1F2 = 0.920 (p b 0.001) and rF2F3 =
0.592 (p b 0.001). No other correlation was signicant.
Tables 2 and 3 summarize the goodness-of-t index and the re-
sults of conrmatory factor analysis of CR.
3.3. Conrmatory factor analysis of cognitive domains
As in CR, the CFA of cognitive domains was performed with the
results of the cognitive assessment of 98 healthy controls. The CFA
showed the presence of four factors, namely: working memory (F1), at-
tention (F2), verbal memory (F3) and cognitive exibility (F4). The
values obtained on the goodness-of-t statistics indicated a good t
between the model and the observed data. The correlation between
Table 1
Socio-demographic characteristics of SSD and HC.
SSD HC 2 p-value
N= 35(%) N= 98(%)
Gender: Male 23 (65.71%) 64 (65.3%) 0.002 0.965
Mean SD Mean SD t p-value
Age 15.00 1.9 15.17 1.93 0.454 0.651
Socio-economic statusa 35.22 17.1 38.35 15.75 0.721 0.327
a
Hollingshead Redlich Scale (Hollingshead and Redlich, 1958).
4 E. de la Serna et al. / Schizophrenia Research xxx (2012) xxxxxx

Table 2 signicant 2 (1, N = 133) = 29.15, p > 0.001 indicating that it was
CFA model t summary for CR and cognitive domains. able to distinguish between SSD patients and HC. The model as a
2 df Ratio IND Model BBNFI RMSEA (CI 95%) whole explained between 21.7% (Cox and Snell R square) and 31.8%
2/df AIC AIC (Nagelkerke R squared) of the variance and correctly classied
Cognitive 149.344 24 6.22 202.009 43.942 .909 0.045 (0.0270.061) 79.8% of the cases (B = 0.152; p b 0.001; Exp (B) = 0.859; CI
reserve 95% = 0.8040.917). The specicity (percentage of the group without
Cognitive 117.75 51 2.30 239.78 44.24 .921 0.039 (0.020.051) the characteristic of interest) of the model was 92.1% and the sensitiv-
domains
ity (percentage of the group that has the characteristic of interest)
AIC: Akaike Information Criterion; BBFNI: BentlerBonet Normed Fit Index; RMSEA: was 58.1%. The positive predictive value of the model was 75%.
Root Mean-Square Error of Approximation.

3.5. Predictive value of CR on cognitive domains

Table 3
Standardized parameters. Factor loading of CR variables on CR model.
Variable
Vocabulary subtest of
WAIS-III/WISC-R
Years of formal education
School performance before the
beginning of the disorder
Parents' educational level
Questions about the development of
language, motor functions and
writing and reading
Intellectual hobbies
Family hobbies
Peer relationships
Sociability
p b 0.001.
p b 0.01.
The linear regression analysis to test the predictive capacity of CR
on cognitive domains was based on results obtained in patients two
Premorbid Education/ Lifetime leisure years after onset of illness. To assess the effects of clinical variables
IQ (F1) occupation (F2) activities (F3)
on neuropsychological performance at two year follow-up, total
0.573 score of PANSS at baseline was included in the regression analysis.
CR predicted the performance at two year follow-up of the SSD sub-
0.487
0.348 jects on working memory and attention. Verbal memory and cogni-
tive exibility, however, were not predicted by any of the variables
0.337 included in the regression model.
0.953
To test whether premorbid IQ had greater predictive value than
global CR, a new linear regression was conducted with premorbid
0.229 IQ and total score of PANSS at baseline as potential predictors. Only
0.836 working memory at two-year follow-up was predicted by the
0.900 premorbid IQ. However, premorbid IQ and total score of PANSS
0.901
explained 26.7% of the variance of working memory whereas CR
and total score of PANSS explained 40.30%. Table 5 shows the results
of linear regressions in cognitive domains.

4. Discussion
all latent variables showed signicant results as follows: rF1F2 = 0.559
(pb 0.001); rF1F3 = 0.344 (pb 0.001); rF1F4 =0.235 (pb 0.05); rF2F3 =
The main nding of the present study is that our proxy of CR based
0.496 (pb 0.001); rF2F4 = 0.240 (pb 0.05); rF3F4 = 0.294 (pb 0.05).
on premorbid IQ, educationaloccupational level and lifetime leisure
Standardized parameters for cognitive domains and the goodness-of-
and social activities, was able to predict some cognitive domains
t index are provided in Tables 2 and 4.
such as working memory and attention at two year follow-up. More-
over SSD patients were shown to have lower scores than control sub-
3.4. Differences in CR between SSD and HC jects in CR and the CR proxy correctly classied 79.8% of the sample as
SSD or HC.
Signicant differences between groups were found in composite The concept of CR has previously been used in dementia research to
scores of CR (t=5.474; pb 0.001), with the SSD ( x =48.387.23) help explain the compensatory capacity that some individuals exhibit in
group obtaining lower composite scores for CR than HC ( x =57.85 terms of their ability to minimize the clinical impact of acquired brain
8.09). pathology. In this regard, environmental and lifetime factors such as ed-
A logistic regression was performed to assess the predictive value ucational level, occupation, leisure activities and premorbid IQ have fre-
of CR for each clinical group (SSD or HC). The model was statistically quently been considered proxies reecting CR, since these variables are
related to a decreased risk of developing dementia symptoms in epide-
Table 4
miological studies (Scarmeas and Stern, 2003; Valenzuela and Sachdev,
Standardized parameters. Factor loading of neuropsychological variables on cognitive 2006; Helzner et al., 2007). In schizophrenia, CR has been proposed as a
domains. factor related to better prognosis, higher adherence and less psychotic
symptoms in schizophrenia (Leeson et al., 2009).
Variable Working Attention Verbal Cognitive
memory (F2) memory exibility Taking previous literature (Stern, 2009) into account, our CR mea-
(F1) (F3) (F4) sure included premorbid IQ, educational level, occupational attainment,
Digits forward 0.943 and leisure activity, all variables which have been independently linked
Digits backward 0.681 to schizophrenia (Jones et al., 1994; Zammit et al., 2004; Reichenberg
Letter-Number Sequencing 0.650 et al., 2005; Chong et al., 2009). Based on these studies, we hypothe-
Trail Making Test part B 0.799 sized that the SSD group would attain lower scores than HC in our CR
Trail Making Test part A 0.646
Reaction time of CPT 0.211
composite measure and our results supported this hypothesis, with
Correct responses on CPT 0.185 CR correctly classifying 79.8% of the sample as SSD or HC.
TAVEC immediate memory score 0.910 Our study also found that performance on working memory was
TAVEC delayed memory score 0.780 predicted by CR. Because we aimed to test whether our CR proxy
TAVEC 0.751
(which included premorbid IQ, educational and occupational level
WCST errors 0.882
WCST perseverative errors 0.833 and leisure activities) was a better predictor than premorbid IQ
WCST perseverative responses 0.992 (which had been commonly used as CR proxy in previous literature),
Stroop Interference score 0.872 the analysis was repeated with Premorbid IQ and total PANSS score as
p b 0.001. predictors. The results showed that premorbid IQ also predicted
p b 0.05. working memory. However premorbid IQ only explained 26.7% of

Please cite this article as: de la Serna, E., et al., Cognitive reserve as a predictor of two year neuropsychological performance in early onset rst-
episode schizophrenia, Schizophr. Res. (2012), http://dx.doi.org/10.1016/j.schres.2012.10.026
 E. de la Serna et al. / Schizophrenia Research xxx (2012) xxxxxx 5

Table 5
Results of linear regression analysis assessing the predictive capacity of CR and premorbid IQ on neuropsychological variables after two years.

Cognitive domain Predictor B S.E. t P

Linear regression with cognitive Working memory Cognitive reserve 0.255 0.061 0.609 4.143 b0.001
reserve and PANSS total score R = 0.635 PANSS total score 0.020 0.025 0.118 0.804 0.428
R2 = 0.403
Corrected R2 = 0.360
F = 9.438
p = 0.001

Attention Cognitive Reserve 0.047 0.019 0.443 2.402 0.025

R = 0.510 PANSS total score 0.009 0.008 0.205 1.111 0.279
R2 = 0.260
Corrected R2 = 0.193
F = 3.874
p = 0.036

Verbal memory Cognitive reserve 0.186 0.103 0.326 1.812 0.081

R = 0.326 PANSS total score 0.001 0.041 0.002 0.013 0.990
R2 = 0.107
Corrected R2 = 0.043
F = 1.669
p = 0.207

Cognitive exibility Cognitive reserve 0.049 0.054 0.166 0.921 0.365

R = 0.327 PANSS total score 0.036 0.021 0.302 1.679 0.104
R2 = 0.107
Corrected R2 = 0.043
F = 1.672
p = 0.206

Linear regression with premorbid Working memory Premorbid IQ 0.610 0.192 0.480 3.173 0.003
IQ and PANSS total score R = 0.517 PANSS total score 0.030 0.025 0.179 1.183 0.245
R2 = 0.267
Corrected R2 = 0.221
F = 5.831
p = 0.007

Attention Premorbid IQ 0.091 0.061 0.285 1.481 0.152

R = 0.385 PANSS total score 0.011 0.008 0.251 1.303 0.205
R2 = 0.148
Corrected R2 = 0.074
F = 1.996
p = 0.159

Verbal memory Premorbid IQ 0.711 0.280 0.410 2.544 0.016

R = 0.412 PANSS total score 0.007 0.037 0.031 0.193 0.848
R2 = 0.170
Corrected R2 = 0.118
F = 3.269
p = 0.051

Cognitive exibility Premorbid IQ 0.291 0.145 0.321 2.010 0.053

R = 0.428 PANSS total score 0.035 0.019 0.290 1.815 0.079
R2 = 0.183
Corrected R2 = 0.132
F = 3.580
p = 0.040

Signicant differences (p b 0.05) marked in bold.

the variance of the cognitive domain compared to 40.3% of the
variance explained by the CR proxy. Thus, CR, a concept which joins to-
gether different variables such as premorbid IQ, educationaloccupational
level and leisure activities, was shown to have greater predictive capacity.
Regarding attention, our CR proxy predicted SSD performance on at-
tention tasks. These results are partially supported by previous literature
with adult samples (Meguro et al., 2001; Roldan-Tapia et al., 2012). To
our knowledge, there are no studies in child and adolescent patients
with SSD which assess the inuence of CR on neuropsychological perfor-
mance; however, previous studies have investigated separately the inu-
ence of premorbid IQ and educational level on cognitive performance in
SSD samples (Silverstein et al., 2002; Norman et al., 2005; Rund et al.,
2007). Rund et al. (2007) and found that patients with good levels of
premorbid academic functioning had better scores in some neuro-
psychological areas than those with lower levels of premorbid func-
tioning; and Silverstein et al. (2002) suggested that social and
academic adjustment could inuence the cognitive performance
seen in adult schizophrenia. Moreover, childhood IQ has been
shown to be a strong predictor of social outcome in schizophrenia
patients in a 21 year follow-up study (Munro et al., 2002), while a num-
ber of studies show an inverse relationship between schizophrenia and
lower IQ, academicoccupational level and leisure activities (Chong
et al., 2009; Koenen et al., 2009; Pitkanen et al., 2009). These variables
are usually studied separately, but taken together they could reect
the compensation capacity of patients and have a strong inuence on
clinical and neuropsychological outcomes.
The role of CR on neuropsychiatric disorders has been previously
explained by Barnett et al. (2006), who found some difculties in the
application of this concept in psychiatry. CR and clinical symptoms are
not independent from each other, above all in neurodevelopmental dis-
orders such as schizophrenia where the accumulation of CR could be
impeded by the disease. Differences at baseline between patients' and

Please cite this article as: de la Serna, E., et al., Cognitive reserve as a predictor of two year neuropsychological performance in early onset rst-
episode schizophrenia, Schizophr. Res. (2012), http://dx.doi.org/10.1016/j.schres.2012.10.026
6 E. de la Serna et al. / Schizophrenia Research xxx (2012) xxxxxx
controls' CR help conrm this. Nevertheless, the concept could be useful
to explain some of the differences between schizophrenia patients in
clinical and neuropsychological outcomes.
Regarding the limitations of this study, the main one is that no
validated instrument for measuring CR was used. However, this was
due to the fact that no such instrument has yet been developed for
use with children and adolescents. Another limitation is the relatively
small sample size. Additionally, the effects of antipsychotic drugs
have not been controlled for. Although previous studies have related
antipsychotics and cognitive performance (Keefe et al., 1999), the in-
uence of different antipsychotics on different cognitive functions is
not clear: some studies have found an improvement (Andersen
et al., 2011; Pardo et al., 2011) whereas others have found no differ-
ences (Stip, 2006; Robles et al., 2011).
Nevertheless, our study found lower scores on CR in SSD than in
HC and the CR measure correctly classied a high percentage of the
sample into SSD or HC. Moreover, CR predicted SSD performance on
working memory and attention at two year follow-up. CR could be a
useful concept in schizophrenia research to help explain differences
in outcome among SSD patients. However, further research is needed
to investigate the inuence of CR on neuropsychological performance
in schizophrenia patients, and the development of a specic instru-
ment for assessing the CR of SSD children and adolescents could facil-
itate such research.
Role of funding source
This study was supported by a grant from ISCIII (Ministry of Health, Spain). This insti-
tution had no further role in the study design; in the collection, analysis and interpretation
of data; in the writing of the report; and in the decision to submit the paper for publication.
Contributors
The contributions of each author to the paper are the following: Dr. E. de la Serna, J.
Castro, O. Puig and S. Andres designed the current study. All authors contributed to the
acquisition of data. Dr. E. de la Serna managed the literature searches. Dr. E. de la Serna,
Dr. I Baeza, Dr. J. Castro and Dr. J. Guardia undertook the statistical analysis. Dr. E de la
Serna wrote the rst draft of the manuscript. Dr. I. Baeza and Dr. J. Castro contributed to
the earlier versions of the manuscript. All authors contributed to and have approved
the nal manuscript.
Conict of interest
Elena de la Serna, Immaculada Baeza, Susana Andrs, Olga Puig, Igor Bombin, David
Bartrs-Faz, Celso Arango, Ana Gonzalez-Pinto, Mara Parellada, Mara Mayoral, Montserrat
Graell and Soraya Otero afrm that they have no conicts of interest.
Dr. J. Castro has had the following relationships which may represent a conict of
interest:
Consultant: Eli Lilly and Pzer.
Travel support: Eli Lilly.
Acknowledgments
The authors of this report would like to thank the support of the Spanish Ministry of
Health, Instituto de Salud Carlos III and the Catalonia Government, DIUE (Departament d'
Innovaci, Universitat i Empresa) 2009SGR1119 and CIBER-SAM CB07/09/0005.
References
Andersen, R., Fagerlund, B., Rasmussen, H., Ebdrup, B.H., Aggernaes, B., Gade, A., Oranje,
B., Glenthoj, B., 2011. Cognitive effects of six months of treatment with quetiapine
in antipsychotic-naive rst-episode schizophrenia. Psychiatry Res. 187, 4954.
Barnett, J.H., Salmond, C.H., Jones, P.B., Sahakian, B.J., 2006. Cognitive reserve in neuro-
psychiatry. Psychol. Med. 36, 10531064.
Bartres-Faz, D., Sole-Padulles, C., Junque, C., Rami, L., Bosch, B., Bargallo, N., Falcon, C.,
Sanchez-Valle, R., Molinuevo, J.L., 2009. Interactions of cognitive reserve with re-
gional brain anatomy and brain function during a working memory task in healthy
elders. Biol. Psychol. 80, 256259.
Basso, M.R., Nasrallah, H.A., Olson, S.C., Bornstein, R.A., 1998. Neuropsychological corre-
lates of negative, disorganized and psychotic symptoms in schizophrenia. Schizophr.
Res. 31, 99111.
Benedet, M., Alejandre, A., 1998. TAVEC, Test de aprendizaje Verbal Espaa-Complutense.
TEA Ediciones, Madrid.
Bilder, R.M., Goldman, R.S., Robinson, D., Reiter, G., Bell, L., Bates, J.A., Pappadopulos, E.,
Willson, D.F., Alvir, J.M., Woerner, M.G., Geisler, S., Kane, J.M., Lieberman, J.A., 2000.
Neuropsychology of rst-episode schizophrenia: initial characterization and clini-
cal correlates. Am. J. Psychiatry 157, 549559.
Please cite this article as: de la Serna, E., et al., Cognitive reserve as a predictor of two year neuropsychological performance in early onset rst-
episode schizophrenia, Schizophr. Res. (2012), http://dx.doi.org/10.1016/j.schres.2012.10.026
Bosch, B., Bartres-Faz, D., Rami, L., Arenaza-Urquijo, E.M., Fernandez-Espejo, D., Junque,
C., Sole-Padulles, C., Sanchez-Valle, R., Bargallo, N., Falcon, C., Molinuevo, J.L., 2010.
Cognitive reserve modulates task-induced activations and deactivations in healthy
elders, amnestic mild cognitive impairment and mild Alzheimer's disease. Cortex
46, 451461.
Castro-Fornieles, J., Parellada, M., Gonzalez-Pinto, A., Moreno, D., Graell, M., Baeza, I., Otero, S.,
Soutullo, C.A., Crespo-Facorro, B., Ruiz-Sancho, A., Desco, M., Rojas-Corrales, O., Patino, A.,
Carrasco-Marin, E., Arango, C., 2007. The child and adolescent rst-episode psychosis
study (CAFEPS): design and baseline results. Schizophr. Res. 91, 226237.
Chong, S.A., Subramaniam, M., Lee, I.M., Pek, E., Cheok, C., Verma, S., Wong, J., 2009.
Academic attainment: a predictor of psychiatric disorders? Soc. Psychiatry Psychiatr.
Epidemiol. 44, 9991004.
Conners, K., 2000. Conners' Continuous Performance Test, CPT-II. MHS.
Fagerlund, B., Pagsberg, A.K., Hemmingsen, R.P., 2006. Cognitive decits and levels of IQ
in adolescent onset schizophrenia and other psychotic disorders. Schizophr. Res.
85, 3039.
Fitzgerald, D., Lucas, S., Redoblado, M.A., Winter, V., Brennan, J., Anderson, J., Harris, A.,
2004. Cognitive functioning in young people with rst episode psychosis: relation-
ship to diagnosis and clinical characteristics. Aust. N. Z. J. Psychiatry 38, 501510.
Golden, C.J., 1978. Stroop Color and Word Test. Stoelting Co, Wood Dale, IL.
Good, K.P., Rabinowitz, J., Whitehorn, D., Harvey, P.D., DeSmedt, G., Kopala, L.C., 2004.
The relationship of neuropsychological test performance with the PANSS in anti-
psychotic naive, rst-episode psychosis patients. Schizophr. Res. 68, 1119.
Green, M.F., Kern, R.S., Braff, D.L., Mintz, J., 2000. Neurocognitive decits and functional out-
come in schizophrenia: are we measuring the right stuff? Schizophr. Bull. 26, 119136.
Heaton, R.K., Chelune, G.J., Talley, J.L., Kay, G., Curtiss, G., 1997. Wisconsin Card Sorting
Test (WCST). TEA Ediciones, Madrid.
Helzner, E.P., Scarmeas, N., Cosentino, S., Portet, F., Stern, Y., 2007. Leisure activity and
cognitive decline in incident Alzheimer disease. Arch. Neurol. 64, 17491754.
Hollingshead, A., Redlich, F., 1958. Social Class and Mental Illness. Wiley, New York.
Holthausen, E.A., Wiersma, D., Sitskoorn, M.M., Hijman, R., Dingemans, P.M., Schene, A.H.,
van den Bosch, R.J., 2002. Schizophrenic patients without neuropsychological decits:
subgroup, disease severity or cognitive compensation? Psychiatry Res. 112, 111.
Jones, P., Rodgers, B., Murray, R., Marmot, M., 1994. Child development risk factors for
adult schizophrenia in the British 1946 birth cohort. Lancet 344, 13981402.
Kaufman, J., Birmaher, B., Brent, D., Rao, U., Flynn, C., Moreci, P., Williamson, D., Ryan,
N., 1997. Schedule for Affective Disorders and Schizophrenia for School-Age
ChildrenPresent and Lifetime Version (K-SADS-PL): initial reliability and validity
data. J. Am. Acad. Child Adolesc. Psychiatry 36, 980988.
Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The Positive and Negative Syndrome Scale
(PANSS) for schizophrenia. Schizophr. Bull. 13, 261276.
Keefe, R.S., Silva, S.G., Perkins, D.O., Lieberman, J.A., 1999. The effects of atypical anti-
psychotic drugs on neurocognitive impairment in schizophrenia: a review and
meta-analysis. Schizophr. Bull. 25, 201222.
Kenny, J.T., Friedman, L., Findling, R.L., Swales, T.P., Strauss, M.E., Jesberger, J.A., Schulz,
S.C., 1997. Cognitive impairment in adolescents with schizophrenia. Am. J. Psychiatry
154, 16131615.
Koenen, K.C., Moftt, T.E., Roberts, A.L., Martin, L.T., Kubzansky, L., Harrington, H.,
Poulton, R., Caspi, A., 2009. Childhood IQ and adult mental disorders: a test of
the cognitive reserve hypothesis. Am. J. Psychiatry 166, 5057.
Leeson, V.C., Barnes, T.R., Hutton, S.B., Ron, M.A., Joyce, E.M., 2009. IQ as a predictor of
functional outcome in schizophrenia: a longitudinal, four-year study of rst-episode
psychosis. Schizophr. Res. 107, 5560.
Mayoral, M., Zabala, A., Robles, O., Bombin, I., Andres, P., Parellada, M., Moreno, D.,
Graell, M., Medina, O., Arango, C., 2008. Neuropsychological functioning in adolescents
with rst episode psychosis: a two-year follow-up study. Eur. Psychiatry 23, 375383.
Meguro, K., Shimada, M., Yamaguchi, S., Ishizaki, J., Ishii, H., Shimada, Y., Sato, M.,
Yamadori, A., Sekita, Y., 2001. Cognitive function and frontal lobe atrophy in nor-
mal elderly adults: implications for dementia not as aging-related disorders and
the reserve hypothesis. Psychiatry Clin. Neurosci. 55, 565572.
Munro, J.C., Russell, A.J., Murray, R.M., Kerwin, R.W., Jones, P.B., 2002. IQ in childhood
psychiatric attendees predicts outcome of later schizophrenia at 21 year follow-
up. Acta Psychiatr. Scand. 106, 139142.
Norman, R.M., Malla, A.K., Manchanda, R., Townsend, L., 2005. Premorbid adjustment
in rst episode schizophrenia and schizoaffective disorders: a comparison of social
and academic domains. Acta Psychiatr. Scand. 112, 3039.
Nuechterlein, K.H., Barch, D.M., Gold, J.M., Goldberg, T.E., Green, M.F., Heaton, R.K.,
2004. Identication of separable cognitive factors in schizophrenia. Schizophr. Res.
72, 2939.
Pardo, B.M., Garolera, M., Ariza, M., Pareto, D., Salamero, M., Valles, V., Delgado, L.,
Alberni, J., 2011. Improvement of cognitive exibility and cingulate blood ow
correlates after atypical antipsychotic treatment in drug-naive patients with
rst-episode schizophrenia. Psychiatry Res. 194, 205211.
Peralta, V., Cuesta, M.J., 1994. Psychometric properties of the Positive and Negative
Syndrome Scale (PANSS) in schizophrenia. Psychiatry Res. 53, 3140.
Pitkanen, A., Hatonen, H., Kuosmanen, L., Valimaki, M., 2009. Individual quality of life of
people with severe mental disorders. J. Psychiatr. Ment. Health Nurs. 16, 39.
Reichenberg, A., Weiser, M., Rapp, M.A., Rabinowitz, J., Caspi, A., Schmeidler, J., Knobler,
H.Y., Lubin, G., Nahon, D., Harvey, P.D., Davidson, M., 2005. Elaboration on premorbid
intellectual performance in schizophrenia: premorbid intellectual decline and risk for
schizophrenia. Arch. Gen. Psychiatry 62, 12971304.
Reitan, R.M., Wolfson, D., 1985. The HalsteadReitan Neuropsychological Test Battery .
(Tucson).
Robles, O., Zabala, A., Bombin, I., Parellada, M., Moreno, D., Ruiz-Sancho, A., Arango, C.,
2011. Cognitive efcacy of quetiapine and olanzapine in early-onset rst-episode
psychosis. Schizophr. Bull. 37, 405415.
 E. de la Serna et al. / Schizophrenia Research xxx (2012) xxxxxx
Roldan-Tapia, L., Garcia, J., Canovas, R., Leon, I., 2012. Cognitive reserve, age, and their
relation to attentional and executive functions. Appl. Neuropsychol. 19, 28.
Rund, B.R., Melle, I., Friis, S., Johannessen, J.O., Larsen, T.K., Midboe, L.J., Opjordsmoen, S.,
Simonsen, E., Vaglum, P., McGlashan, T., 2007. The course of neurocognitive function-
ing in rst-episode psychosis and its relation to premorbid adjustment, duration of
untreated psychosis, and relapse. Schizophr. Res. 91, 132140.
Scarmeas, N., Stern, Y., 2003. Cognitive reserve and lifestyle. J. Clin. Exp. Neuropsychol.
25, 625633.
Silverstein, M.L., Mavrolefteros, G., Close, D., 2002. Premorbid adjustment and neuro-
psychological performance in schizophrenia. Schizophr. Bull. 28, 157165.
Sole-Padulles, C., Bartres-Faz, D., Junque, C., Vendrell, P., Rami, L., Clemente, I.C., Bosch,
B., Villar, A., Bargallo, N., Jurado, M.A., Barrios, M., Molinuevo, J.L., 2009. Brain struc-
ture and function related to cognitive reserve variables in normal aging, mild cog-
nitive impairment and Alzheimer's disease. Neurobiol. Aging. 30, 11141124.
Staff, R.T., Murray, A.D., Deary, I.J., Whalley, L.J., 2004. What provides cerebral reserve?
Brain 127, 11911199.
Stern, Y., 2009. Cognitive reserve. Neuropsychologia 47, 20152028.
Stern, Y., Habeck, C., Moeller, J., Scarmeas, N., Anderson, K.E., Hilton, H.J., Flynn, J.,
Sackeim, H., van Heertum, R., 2005. Brain networks associated with cognitive
reserve in healthy young and old adults. Cereb. Cortex 15, 394402.
Please cite this article as: de la Serna, E., et al., Cognitive reserve as a predictor of two year neuropsychological performance in early onset rst-
episode schizophrenia, Schizophr. Res. (2012), http://dx.doi.org/10.1016/j.schres.2012.10.026
7
Stip, E., 2006. Cognition, schizophrenia and the effect of antipsychotics. Encephale 32,
341350.
Valenzuela, M.J., Sachdev, P., 2006. Brain reserve and dementia: a systematic review.
Psychol. Med. 36, 441454.
Wechsler, D., 1974. Wechsler Intelligence Scale for ChildrenRevised (WISC-R). TEA
Ediciones, Madrid.
Wechsler, D., 2001. WAIS-III escala de inteligencia de Wechsler para adultos. TEA
Ediciones, Madrid.
Zabala, A., Rapado, M., Arango, C., Robles, O., de la Serna, E., Gonzalez, C., Rodriguez-
Sanchez, J.M., Andres, P., Mayoral, M., Bombin, I., 2010. Neuropsychological func-
tioning in early-onset rst-episode psychosis: comparison of diagnostic subgroups.
Eur. Arch. Psychiatry Clin. Neurosci. 260, 225233.
Zammit, S., Allebeck, P., David, A.S., Dalman, C., Hemmingsson, T., Lundberg, I., Lewis, G.,
2004. A longitudinal study of premorbid IQ Score and risk of developing schizophrenia,
bipolar disorder, severe depression, and other nonaffective psychoses. Arch. Gen. Psychi-
atry 61, 354360.