Research

JAMA Psychiatry | Original Investigation

Cognitive Function and Variability in Antipsychotic Drug–Naive

Patients With First-Episode Psychosis
A Systematic Review and Meta-Analysis
Maria Lee, MScPsych; Martin Cernvall, PhD; Jacqueline Borg, PhD; Pontus Plavén-Sigray, PhD; Cornelia Larsson, MScPsych;
Sophie Erhardt, PhD; Carl M. Sellgren, MD, PhD; Helena Fatouros-Bergman, PhD; Simon Cervenka, MD, PhD

Supplemental content
IMPORTANCE Cognitive impairment contributes significantly to clinical outcome and level of
function in individuals with psychotic disorders. These impairments are present already at
psychosis onset at a group level; however, the question of heterogeneity in cognitive function
among patients has not been systematically investigated.

OBJECTIVE To provide an updated quantification of cognitive impairment at psychosis onset

before patients receive potentially confounding antipsychotic treatment, and to investigate
variability in cognitive function compared with healthy controls.

DATA SOURCES In this systematic review and meta-analysis, PubMed articles were searched
up to September 15, 2022.

STUDY SELECTION Original studies reporting data on cognitive function in antipsychotic

drug–naive patients with first-episode psychosis (FEP) were included.

DATA EXTRACTION AND SYNTHESIS Data were independently extracted by 2 researchers.

Cognitive tasks were clustered according to 6 domains of the Measurement and Treatment
Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery and
the domain of executive function. Random-effects model meta-analyses of mean differences
and coefficient of variation ratios (CVRs) were performed, as well as meta-regressions,
assessment of study quality, and publication bias.

MAIN OUTCOMES AND MEASURES The main outcome measure was Hedges g for mean
differences in cognition and CVR for within-group variability.

RESULTS Fifty studies were included in the analysis with a total of 2625 individuals with FEP
(mean [SD] age, 25.2 [3.6] years, 60% male; 40% female) and 2917 healthy controls (mean
[SD] age, 26.0 [4.6]; 55% male; 45% female). In all cognitive domains, the FEP group
displayed significant impairment compared with controls (speed of processing: Hedges
g = −1.16; 95% CI, −1.35 to −0.98; verbal learning: Hedges g = −1.08; 95% CI, −1.28 to −0.88;
visual learning: Hedges g = −1.05; 95% CI, −1.27 to −0.82; working memory: Hedges
g = −1.04; 95% CI, −1.35 to −0.73; attention: Hedges g = −1.03; 95% CI, −1.24 to −0.82;
reasoning/problem solving: Hedges g = −0.90; 95% CI, −1.12 to −0.68; executive function:
Hedges g = −0.88; 95% CI, −1.07 to −0.69). Individuals with FEP also exhibited a larger
variability across all domains (CVR range, 1.34-1.92).

CONCLUSIONS AND RELEVANCE Results of this systematic review and meta-analysis identified
cognitive impairment in FEP before the initiation of antipsychotic treatment, with large effect
sizes. The high variability within the FEP group suggests the need to identify those individuals
with more severe cognitive problems who risk worse outcomes and could benefit the most
from cognitive remediation.

Author Affiliations: Author

affiliations are listed at the end of this
article.
Corresponding Author: Maria Lee,
MScPsych, Department of Clinical
Neuroscience, Center for Psychiatry
Research, Akademiska Stråket 1,
JAMA Psychiatry. doi:10.1001/jamapsychiatry.2024.0016 Bioclinicum, J4-14, 17176 Stockholm
Published online February 28, 2024. (maria.lee@ki.se).

(Reprinted) E1

Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024

 Research Original Investigation
C
ognitive impairment is a central characteristic of
schizophrenia1-3 and other psychotic disorders and has
a strong association with functional4-6 and clinical
outcome.5,7 Cognitive symptoms may often precede illness
onset,8-10 and a previous meta-analysis11 showed signifi-
cantly impaired cognitive performance already present in an-
tipsychotic drug–naive patients with first-episode psychosis
(FEP), ie, before the introduction of medication effects or ad-
verse effects as potential confounders. In the study including
reports published up until 2012, cognition was found to be im-
paired in patients, with medium to large effect sizes (stan-
dardized mean difference [SMD] range, −0.74 to −1.03). These
results were comparable with earlier meta-analyses in FEP,
which also included patients taking medication (eg, Mesholam-
Gately and colleagues12 with an SMD range of −0.71 to −1.2),
whereas a more recent study 13 focusing only on Chinese
samples assessed with the Measurement and Treatment Re-
search to Improve Cognition in Schizophrenia (MATRICS) Con-
sensus Cognitive Battery14(MCCB) produced slightly larger
effect sizes (SMD range, −0.87 to −1.41). Importantly, all of these
studies focused on mean comparisons only, despite the fact
that psychotic disorders are generally considered to be highly
heterogeneous in terms of symptom profile and course of
illness.15
Attempts to describe variability in cognitive function
have been synthesized in a systematic review of studies in
schizophrenia spectrum and bipolar disorder16 and a narra-
tive review on schizophrenia.17 Both reviews focused on
studies investigating cognitive subgroups and concluded
that distinct subgroups exist ranging from relative intact
cognition to severe cognitive dysfunction. However, this
body of research has typically included patients with longer
duration of illness, and the question of heterogeneity early
in the illness trajectory and before initiation of antipsy-
chotic treatment is still unclear. Furthermore, due to the
methodological variability in studies examining potential
subgroups, meta-analytical evidence is still lacking regard-
ing cognitive heterogeneity.
One way to measure heterogeneity is through comparing
the amount of variability in the group of patients with psycho-
sis with the variability in the control group, eg, by using the co-
efficient of variation ratio (CVR).18 This method was recently
used by Catalan et al19 to demonstrate greater variability in cog-
nitive function in individuals at clinical high risk for psychosis
(CHR-P) compared with controls. Individuals at CHR-P are
defined using different combinations of risk factors, and cur-
rent estimates indicate that 25% go on to develop psychosis
within a 3-year period.20 The study examining cognitive vari-
ability in those at CHR-P also included a small group of pa-
tients with FEP who exhibited similar within-group variabil-
ity as those at CHR-P, but to our knowledge, cognitive variability
in FEP has not been analyzed at the meta-analytic level in larger
samples.
Since the previous meta-analysis of antipsychotic drug–
naive individuals with FEP,11 a large number of new studies as-
sessing cognitive deficits in patients have been published. The
field has also seen an increased trend toward data harmoni-
zation since MCCB has become the dominant test battery. In
E2 JAMA Psychiatry Published online February 28, 2024 (Reprinted)
Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024
Cognitive Function and Variability in Patients With First-Episode Psychosis
Key Points
Question How do patients with first-episode psychosis (FEP)
compare with controls in cognitive function and cognitive
variability, before initiation of antipsychotic medication?
Findings In this systematic review and meta-analysis including 50
studies and 2625 individuals with FEP, the mean cognitive
performance of antipsychotic drug–naive individuals with FEP was
significantly impaired compared with controls, across all cognitive
domains. At the same time, the within-group variability in FEP was
significantly greater, indicating a wider distribution of cognitive
ability in patients with FEP compared with controls.
Meaning Patients with psychosis display cognitive difficulties
very early in the disease process, and the considerable
heterogeneity in cognitive function should prompt individual
assessments in clinical settings in order to optimize care.
addition, guidelines and standards for the reporting of sys-
tematic reviews have been updated.21 In the present study, our
primary aim was, therefore, to update and extend the previ-
ous meta-analysis in antipsychotic drug–naive patients with
FEP. The secondary aim was to perform the first systematic
meta-analysis of within-group variability in cognitive ability
in FEP.
Methods
The meta-analysis was registered on PROSPERO and
followed the Preferred Reporting Items for Systematic Re-
views and Meta-analyses (PRISMA)21 (eTable 1 in Supplement 1)
and Meta-Analysis of Observational Studies in Epidemiology
(MOOSE)22 reporting guidelines (eTable 2 in Supplement 1).
Search Strategy and Selection Criteria
We applied the same search terms as in our previous meta-
analysis on antipsychotic drug–naive individuals with FEP11
(full search terms available in the eMethods 1 in Supple-
ment 1). A literature search was performed on September 15,
2022, using the PubMed database. All abstracts were screened
by 2 authors independently (M.L. and C.L.), using Rayyan
software.23 Studies included in the meta-analysis fulfilled the
following inclusion criteria: (1) available data on individuals
with psychosis spectrum disorder (corresponding to any di-
agnosis within the DSM-IV chapter on “Psychotic Disorders”
excluding a diagnosis of psychotic disorder due to a general
medical condition and substance-induced psychotic disor-
der), (2) individuals categorized as having FEP, (3) patients
reported as being antipsychotic-drug naive, (4) available data
on cognitive performance, (5) available cognitive data from a
healthy control group, and (6) ability to cluster cognitive tests
into 1 of 6 neurocognitive categories of the MCCB or execu-
tive function. Exclusion criteria were (1) articles in a language
other than English, (2) review articles or case reports, (3) over-
lapping sample and neurocognitive test, and (4) cohorts with
a mean duration of untreated psychosis (DUP) longer than 5
years. Studies deemed relevant were read in full length. Ref-
erence lists were screened, and known researchers who had
jamapsychiatry.com
 Cognitive Function and Variability in Patients With First-Episode Psychosis
Figure 1. Adapted Version of the PRISMA 2020 Flow Diagram for Updated Systematic Reviews Including Searches
of Databases, Registers, and Other Sources
23 Records included in previous
version of review
2 Excluded
1 DUP too long
1 Overlapping sample
DUP indicates duration of untreated psychosis; MCCB, Measurement and Treatment Research to Improve Cognition in Schizophrenia Consensus Cognitive Battery.
previously published on antipsychotic drug–naive samples
were contacted. Conflicts regarding inclusion were resolved
through discussion. For studies published before 2012, we re-
lied on data reported in Fatouros-Bergman et al11 (eMethods
2 in Supplement 1). Race and ethnicity data were not col-
lected in our study. Swedish records do not track race or eth-
nicity, and we did not consider it a primary aspect of our re-
search question.
Dataset From the Karolinska Schizophrenia Project
In addition to the literature search, we also included our own
cognitive data collected within the Karolinska Schizophrenia
Project (KaSP; antipsychotic drug–naive patients with FEP and
age- and sex-matched controls), parts of which have been pub-
lished elsewhere24,25 (eMethods 3 in Supplement 1).
Outcome Measures and Data Extraction
Cognitive tests were grouped in a similar way as in Fatouros-
Bergman et al11 (eMethods 2 in Supplement 1), following the
domains of the MCCB, with the exception of social cognition.
This was considered beyond the scope of this review, given the
many different aspects (and outcome measures) of social cog-
nition available.26 In addition, the domain of executive func-
tion was added. The executive function domain includes tests
and outcome measures with a focus on cognitive flexibility,
compared with the MCCB reasoning and problem-solving do-
main, which focuses more on organizational and planning
skills. In summary, tests are grouped according to 7 domains:
processing speed, attention, working memory, verbal learn-
ing, visual learning, reasoning and problem solving, and
executive function. A complete list of outcome measures clus-
jamapsychiatry.com
Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024
Original Investigation Research
497 Records identified from 3 Records identified through other
databases sources
2 Citation searching
1 Our unpublished data
348 Records removed 1 Excluded
319 Excluded 1 Not drug naive
32 Duplicate
146 Assessed for eligibility
119 Excluded
33 Not drug naive
27 Overlapping sample
13 Composite scores
11 No cognitive data
10 No control group
8 Subgroup-level data
6 Only social cognition
5 Cognition as correlation
4 DUP too long
2 No psychosis
1 No MCCB domain
50 Studies included in review
tered under their respective domain is available in the
eMethods 4 and eTable 9 in Supplement 1.
Two researchers (M.L. and M.C.) independently ex-
tracted data from the sample of studies into separate spread-
sheets. Cognitive performance (mean and SD) was extracted
for each eligible task, as well as demographic data (age, sex,
years of education, DUP, country). The independent spread-
sheets were then compared, and any inconsistencies were
resolved through discussion.
When additional information was required, correspond-
ing authors were contacted by email. If the FEP sample con-
sisted of both antipsychotic drug–naive patients and patients
either taking medication or drug free, data on only antipsy-
chotic drug–naive individuals was requested. Raw scores or un-
corrected T scores were requested when not already avail-
able. When several articles used patients from the same clinical
institution, authors were asked to clarify whether samples were
overlapping or not. If there was no response (after 1 addi-
tional reminder email), it was assumed that samples were over-
lapping (eTable 10 in Supplement 1). In all cases of overlap-
ping samples and cognitive tests, the study with the largest
sample size was chosen for inclusion in the meta-analysis.
Statistical Analysis
Our main meta-analytical effect size measure was Hedges g.
When applicable, test scores were transformed so that higher
values always indicated better performance. Random-effects
meta-analyses, using a restricted maximum likelihood random-
effects variance estimator, were conducted for each of the 7
cognitive domains, resulting in a pooled estimate. For several
of the domains (speed of processing, working memory, and ex-
(Reprinted) JAMA Psychiatry Published online February 28, 2024 E3
 Research Original Investigation Cognitive Function and Variability in Patients With First-Episode Psychosis

Figure 2. Combined Forest Plot of Mean Differences in Cognitive Ability for Speed of Processing, Verbal Learning,
and Visual Learning Domains in Drug-Naive Patients With First-Episode Psychosis (FEP) Compared With Controls

FEP, Controls, Studies, SMD

Test or outcome measure No. No. No. (95% CI)
Speed of processing
TMT A 860 938 17 −0.96 (−1.22 to −0.69)
Animal fluency 753 859 15 −1.07 (−1.35 to −0.79)
Fluency, other 222 209 6 −1.02 (−1.48 to −0.55)
WAIS coding 186 252 4 −1.55 (−2.12 to −0.99)
BACS 518 631 10 −1.63 (−1.98 to −1.29)
Speed of processing pooled estimate −1.16 (−1.35 to −0.98)
Verbal learning
HVLT-R 816 882 13 −1.18 (−1.42 to −0.93)
BSRT 48 48 1 −0.56 (−1.45 to 0.32)
SVLT 29 17 1 −1.69 (−2.73 to −0.64)
WMS logical memory 158 132 2 −0.92 (−1.52 to −0.31)
CVLT 82 86 2 −0.87 (−1.54 to −0.21)
RBANS immediate memory 214 132 1 −0.66 (−1.47 to 0.15)
Verbal learning pooled estimate −1.08 (−1.28 to −0.88)
Visual learning
BVMT−R 751 815 12 −1.20 (−1.46 to −0.95)
RCFT 48 48 1 −0.58 (−1.48 to 0.32)
RBANS figure recall 78 60 1 −0.41 (−1.28 to 0.46)
WMS visual reproduction 45 33 1 −0.94 (−1.87 to −0.01)
CANTAB PRM 80 72 1 −0.41 (−1.28 to 0.45)
Visual learning pooled estimate −1.04 (−1.27 to −0.82)

−2.5 −2 −1.5 −1 −0.5 0 0.5 1

SMD (95% CI)

The lines extending from the pooled estimates indicate prediction interval.
BACS indicates Basic Assessment of Cognition Scale; BSRT, Buschke Selective
Reminding Test; BVMT-R, Brief Visuospatial Memory Test–Revised;
CANTAB PRM, Cambridge Neuropsychological Test Automated Battery Pattern
Recognition Memory; CVLT, California Verbal Learning Test; HVLT-R, Hopkins
Verbal Learning Test–Revised; RBANS, Repeatable Battery for the Assessment
of Neuropsychological Status; RCFT, Rey-Osterrieth Complex Figure Test;
SMD, standardized mean difference; SVLT, Serial Verbal Learning Task;
TMT-A, Trail Making Test Part A; WAIS, Wechsler Adult Intelligence Scale;
WMS, Wechsler Memory Scale.

ecutive function), participants could contribute with several
outcome measures or test scores. To handle this unit of analy-
sis problem,27 multilevel meta-analyses were conducted using
the rmv.mv function in the metafor package, version 3.8-128
in R (R Core Team), accounting for nonindependent effects and
sampling errors (eMethods 5 in Supplement 1). Between-
study heterogeneity was assessed using the Q statistic, the I2
index, and prediction intervals. The presence of publication
bias was assessed by visually inspecting funnel plots. Study
quality was judged using a modified version of the Newcastle-
Ottawa Scale (eMethods 6 in Supplement 1). Two researchers
(M.L. and M.C.) independently rated all studies and, when rat-
ings differed, reached consensus through discussion. When 10
or more studies reported scores from the same cognitive test,
meta-regressions were performed to evaluate the impact of a
series of preregistered potential moderators. Age, sex, educa-
tion, and year of publication were selected in order to repli-
cate and extend the previous meta-analysis of cognitive func-
tion in antipsychotic drug–naive individuals with FEP by
Fatouros-Bergman et al,11 and in addition, we included study
quality (eMethods 8 in Supplement 1). To calculate and com-
pare the within-group variability in FEP and controls, we used
the CVR. The CVR is the natural logarithm of the ratio of esti-
mates of population coefficients of variation,18 which is the ra-
tio of the SDs normalized to the mean. Given the results of the
prior meta-analysis,11 large differences in mean values were ex-
pected, and to get an unbiased estimate of variability, the CVR
was deemed particularly appropriate in this case.
For robustness check, analyses of the log variability ratio
were also conducted and can be found in eTable 13 in Supple-
ment 1, along with CVR sensitivity analyses for domains con-
taining negative outcome measures (eMethods 7 in Supple-
ment 1). The CVR values presented in writing and plots have
been back transformed from the log scale, to aid interpretabil-
ity. All analyses were conducted using the metafor package28
in R, version 4.2.1 (R Core Team).29 Code for reproducing the
analyses and figures can be found at Github.30
Results
Sample
In addition to the literature search, we included our own cog-
nitive data collected within the Karolinska Schizophrenia Proj-
ect (KaSP; 42 antipsychotic drug–naive patients with FEP and
64 age- and sex-matched controls) (eTables 3-8 in Supple-
ment 1). The final sample consisted of 50 studies (Figure 1 and
eTable 11 in Supplement 1), with a total of 2625 patients with
FEP (mean [SD] age, 25.2 [3.6] years, 60% male; 40% female)
and 2917 healthy controls (mean [SD] age, 26.0 [4.6]; 55% male;

E4 JAMA Psychiatry Published online February 28, 2024 (Reprinted) jamapsychiatry.com

Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024

 Cognitive Function and Variability in Patients With First-Episode Psychosis Original Investigation Research

Figure 3. Combined Forest Plot of Mean Differences in Cognitive Ability for Working Memory, Attention, Reasoning/Problem Solving,
and Executive Function Domains in Drug-Naive Patients With First-Episode Psychosis (FEP) Compared With Controls

FEP, Controls, Studies, SMD

Test or outcome measure No. No. No. (95% CI)
Working memory
CANTAB SWM 124 162 3 −0.76 (−1.88 to 0.35)
BACS digit sequence 56 64 1 −0.67 (−2.22 to 0.89)
Spatial span 454 589 9 −0.82 (−1.34 to −0.29)
LNS 248 327 4 −0.85 (−1.62 to −0.07)
N−back, 2−back 217 100 3 −0.84 (−1.76 to 0.07)
WAIS digit span 364 460 7 −0.95 (−1.54 to −0.35)
AX−CPT long delay 50 53 1 −1.05 (−2.62 to 0.51)
PASAT 56 56 1 −2.27 (−3.86 to −0.69)
N−back, 1−back 35 47 2 −1.32 (−2.52 to −0.12)
Sternberg WM task 23 33 1 −5.29 (−7.17 to −3.41)
Working memory pooled estimate −1.04 (−1.35 to −0.73)
Attention
CPT correct trials 68 64 1 −0.42 (−1.33 to 0.50)
CPT−IP 639 730 11 −1.28 (−1.57 to −0.99)
CPT AX accuracy 12 15 1 −0.10 (−1.24 to 1.04)
CPT index 334 694 7 −0.84 (−1.20 to −0.47)
CAT hits 17 24 1 −0.91 (−1.98 to 0.16)
Attention pooled estimate −1.03 (−1.24 to −0.82)
Reasoning, problem-solving
Mazes 644 718 10 −0.90 (−1.12 to −0.67)
Reasoning problem-solving pooled estimate −0.90 (−1.12 to −0.67)
Executive function
TMT B 370 331 8 −1.12 (−1.45 to −0.79)
CANTAB IED 159 194 3 −0.44 (−0.95 to 0.07)
CANTAB SOC 159 194 3 −0.63 (−1.15 to −0.11)
WCST categories 514 457 11 −0.69 (−0.97 to −0.42)
WCST perseverative 520 450 10 −0.93 (−1.22 to −0.65)
WCST errors 155 152 4 −0.86 (−1.33 to −0.40)
Tower of London 13 15 1 −1.33 (−2.46 to −0.21)
Executive function pooled estimate −0.88 (−1.07 to −0.69)

−2.5 −2 −1.5 −1 −0.5 0 0.5 1

SMD (95% CI)

The lines extending from the pooled estimates indicate prediction interval.
CANTAB IED indicates Cambridge Neuropsychological Test Automated Battery
intradimensional/extradimensional set shifting; CANTAB SOC, CANTAB
Stockings of Cambridge; CANTAB SWM, CANTAB Spatial Working Memory;
CPT, Continuous Performance Test; CPT-IP, Continuous Performance
Test–Identical Pairs; LNS, Letter Number Sequencing; PASAT, Paced Auditory
Serial Addition Test; SMD, standardized mean difference; TMT-B, Trail Making
Test Part B; WAIS, Wechsler Adult Intelligence Scale; WCST, Wisconsin Card
Sorting Test.

45% female). The mean (SD) education level was 11.9 (1.3) years
for patients and 13.2 (1.8) years for controls. Mean (SD) DUP was
20.9 (16.8) months (the number of studies available for analy-
sis of DUP = 24).
Cognitive Functioning in Antipsychotic Drug–Naive
Patients With FEP as Compared With Controls
Individuals with FEP performed significantly worse than
control individuals across all cognitive domains, with large ef-
fect sizes ranging from −0.88 to −1.16 (Figure 2 and Figure 3).
In order of magnitude, the effect was largest for speed of pro-
cessing (Hedges g = −1.16; 95% CI, −1.35 to −0.98), followed by
verbal learning (Hedges g = −1.08; 95% CI, −1.28 to −0.88), vi-
sual learning (Hedges g = −1.05; 95% CI, −1.27 to −0.82), work-
ing memory (Hedges g = −1.04; 95% CI, −1.35 to −0.73), atten-
tion (Hedges g = −1.03; 95% CI, −1.24 to −0.82), reasoning/
problem solving (Hedges g = −0.90; 95% CI, −1.12 to −0.68),
and executive function (Hedges g = −0.88; 95% CI, −1.07 to
−0.69). Study-level forest plots are available in eFigures 1
through 7 in Supplement 1.
Heterogeneity, Study Quality, Publication Bias,
and Meta-Regression
There was substantial heterogeneity in effect sizes. All cog-
nitive domains displayed I2 estimates above 70%, suggest-
ing that the variance in observed effects was mostly due to
differences in true effects and not sampling error.31 Predic-
tion intervals were large, indicating that the mean group
difference could vary substantially depending on sample
(eTable 12 in Supplement 1). Quality rating of the studies
ranged from 3 to 8 (mean = 6.1, median = 6). Publication
biases are reported in eTable 12 and eFigures 15 to 21 in
Supplement 1. Meta-regressions revealed no strong effects
of the potential moderators (age, sex, education, publica-
tion year, and study quality) that were assessed (eTable 15 in
Supplement 1).

jamapsychiatry.com (Reprinted) JAMA Psychiatry Published online February 28, 2024 E5

Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024

 Research Original Investigation Cognitive Function and Variability in Patients With First-Episode Psychosis

Figure 4. Combined Forest Plot of Within-Group Variability in Cognitive Ability for Speed of Processing, Verbal Learning, and Visual Learning
Domains in Drug-Naive Patients With First-Episode Psychosis (FEP) Compared With Controls

Greater Greater
FEP, Controls, Studies, SMD variability variability
Test or outcome measure No. No. No. (95% CI) in controls in patients
Speed of processing
TMT-A 860 938 17 1.46 (1.24 to 1.71)
Animal fluency 753 859 15 1.20 (1.01 to 1.42)
Fluency, other 222 209 6 1.32 (0.99 to 1.75)
WAIS coding 186 252 4 1.67 (1.19 to 2.33)
BACS 518 631 10 1.67 (1.36 to 2.04)
Speed of processing pooled variability estimate 1.41 (1.29 to 1.55)
Verbal learning
HVLT-R 816 882 13 1.52 (1.34 to 1.73)
BSRT 48 48 1 1.79 (1.12 to 2.86)
SVLT 29 17 1 2.30 (1.27 to 4.18)
WMS Logical memory 158 132 2 1.69 (1.21 to 2.35)
CVLT 82 86 2 1.51 (1.05 to 2.16)
RBANS immediate memory 214 132 1 1.29 (0.86 to 1.94)
Verbal learning pooled variability estimate 1.55 (1.40 to 1.72)
Visual learning
BVMT−R 751 815 12 1.91 (1.63 to 2.24)
RCFT 48 48 1 1.35 (0.77 to 2.36)
RBANS figure recall 78 60 1 1.97 (1.13 to 3.42)
WMS Visual reproduction 45 33 1 2.65 (1.50 to 4.70)
CANTAB PRM 80 72 1 1.43 (0.85 to 2.42)
Visual learning pooled variability estimate 1.87 (1.63 to 2.15)

0 1 2 3 4 5
SMD (95% CI)

BACS indicates Basic Assessment of Cognition Scale; BSRT, Buschke Selective
Reminding Test; BVMT-R, Brief Visuospatial Memory Test-Revised;
CANTAB PRM, Cambridge Neuropsychological Test Automated Battery Pattern
Recognition Memory; CVLT, California Verbal Learning Test; HVLT-R, Hopkins
Verbal Learning Test–Revised; RBANS, Repeatable Battery for the Assessment
of Neuropsychological Status; RCFT, Rey-Osterrieth Complex Figure Test;
SMD, standardized mean difference; SVLT, Serial Verbal Learning Task;
TMT-A, Trail Making Test Part A; WAIS, Wechsler Adult Intelligence Scale;
WMS, Wechsler Memory Scale.

Variability of Cognitive Functioning in Antipsychotic Drug–
Naive Patients With FEP Compared With Controls
The patients showed significantly greater within-group vari-
ability in cognitive performance compared with the con-
trols. CVR values ranged between 1.34 to 1.92 (Figure 4 and
Figure 5), meaning that variability in the patient group was
about 30% as large to almost twice as large as that of the
controls. In order of magnitude, the largest CVR was
obtained for the attention domain (CVR = 1.92; 95% CI, 1.62-
2.27), followed by visual learning (CVR = 1.87; 95% CI, 1.63-
2.15), working memory (CVR = 1.61; 95% CI, 1.37-1.90), ver-
bal learning (CVR = 1.55; 95% CI, 1.40-1.72), reasoning/
problem solving (CVR = 1.46; 95% CI, 1.31-1.64), speed of
processing (CVR = 1.43; 95% CI, 1.27-1.61), and executive
function (CVR = 1.34; 95% CI, 1.13-1.58). Sensitivity analyses
of within-group variability are available in eTable 14 in
Supplement 1 as well as additional study-level forest plots of
CVR in eFigures 8 through 14 in Supplement 1.
Discussion
Results of this updated systematic review and meta-analysis
supported the presence of substantial cognitive impairment
in patients experiencing their FEP, before antipsychotic treat-
ment has been introduced. As a group, patients underper-
formed compared with controls, with similar effect sizes, across
all cognitive domains. Furthermore, we, for the first time (to
our knowledge), performed a comprehensive meta-analytic
comparison of within-group variability, demonstrating a greater
variability in cognitive performance within the patient group
compared with the healthy controls across all domains.
The present synthesis incorporated more than twice as
many studies and participants as the previous meta-analysis
on antipsychotic drug–naive individuals with FEP11 and was
conducted using updated reporting standards and meta-
analytic methods. Inclusion criteria followed the same prin-
ciples, apart from a stricter selection in regard to DUP as we
wanted to characterize the onset of the disorder and not the
chronic effects of unmedicated psychosis. The updated lit-
erature search yielded some additional differences in study
characteristics where a majority of studies (eTable 11 in
Supplement 1) used the MCCB, and one-half of the new stud-
ies (eTable 11 in Supplement 1) included were conducted in 1
country (China) as compared with 25% of included studies in
the previous meta-analysis.11 Effects detected in the current
study were slightly larger (a difference of 0.08-0.33 in Cohen
d). There are several possible explanations, including
increased harmonization across studies and more reliable
psychometrics, although meta-regression of publication year

E6 JAMA Psychiatry Published online February 28, 2024 (Reprinted) jamapsychiatry.com

Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024

 Cognitive Function and Variability in Patients With First-Episode Psychosis Original Investigation Research

Figure 5. Combined Forest Plot of Within-Group Variability in Cognitive Ability for Working Memory, Attention, Reasoning/Problem Solving,
and Executive Function Domains in Drug-Naive Patients With First-Episode Psychosis (FEP) Compared With Controls

Greater Greater
FEP, Controls, Studies, SMD variability variability
Test or outcome measure No. No. No. (95% CI) in controls in patients
Working memory
CANTAB SWM 124 162 3 0.78 (0.42 to 1.43)
BACS digit sequence 56 64 1 1.44 (0.64 to 3.22)
Spatial span 454 589 9 1.54 (1.17 to 2.02)
LNS 248 327 4 1.52 (1.02 to 2.28)
N−back, 2−back 217 100 3 1.79 (1.10 to 2.92)
WAIS digit span 364 460 7 1.48 (1.09 to 2.02)
AX−CPT long delay 50 53 1 2.55 (1.05 to 4.00)
PASAT 56 56 1 4.00 (2.33 to 4.00)
N−back, 1−back 35 47 2 2.99 (1.57 to 4.00)
Sternberg WM task 23 33 1 2.20 (0.94 to 4.00)
Working memory pooled variability estimate 1.64 (1.41 to 1.90)
Attention
CPT correct trials 68 64 1 1.09 (0.52 to 2.28)
CPT−IP 639 730 11 1.93 (1.53 to 2.42)
CPT AX accuracy 12 15 1 1.04 (0.41 to 2.62)
CPT index 334 694 7 1.87 (1.40 to 2.51)
CAT hits 17 24 1 4.00 (3.20 to 4.00)
Attention pooled variability estimate 1.92 (1.62 to 2.27)
Reasoning, problem-solving
Mazes 644 718 10 1.46 (1.31 to 1.64)
Reasoning problem-solving pooled variability estimate 1.46 (1.31 to 1.64)
Executive function
TMT B 370 331 8 1.48 (1.02 to 2.14)
CANTAB IED 159 194 3 1.02 (0.54 to 1.91)
CANTAB SOC 159 194 3 1.08 (0.60 to 1.95)
WCST categories 514 457 11 2.08 (1.51 to 2.86)
WCST perseverative 520 450 10 0.85 (0.61 to 1.20)
WCST errors 155 152 4 1.01 (0.59 to 1.75)
Tower of London 13 15 1 4.00 (1.93 to 4.00)
Executive function pooled variability estimate 1.34 (1.13 to 1.58)

0 1 2 3 4
CVR (95% CI)

CANTAB IED indicates Cambridge Neuropsychological Test Automated Battery Test–Identical Pairs; CVR, coefficient of variation ratio; LNS, Letter Number
intradimensional/extradimensional set shifting; CANTAB SOC, CANTAB Sequencing; PASAT, Paced Auditory Serial Addition Test; TMT-B, Trail Making
Stockings of Cambridge; CANTAB SWM, CANTAB Spatial Working Memory; Test Part B; WAIS, Wechsler Adult Intelligence Scale; WCST, Wisconsin Card
CPT, Continuous Performance Test; CPT-IP, Continuous Performance Sorting Test.

failed to show significance. The results are similar to those
reported in chronic schizophrenia samples1-3 and are slightly
larger compared with those observed in medicated FEP.12,13
Hence, our findings corroborate the notion of cognitive
impairment as a central feature already present in the early
stages of psychotic disorders,32,33 before the introduction of
antipsychotic medication.
Similar to the previous meta-analysis, we observed a large
proportion of variability between studies, where I2 values and
prediction intervals show that the mean effect size can vary
considerably depending on the sample. Meta-regressions did
not reveal any strong effects of our preregistered potential
moderators. However, recruiting antipsychotic drug–naive
individuals with FEP for research is notoriously complicated,
potentially resulting in different screening and recruitment pro-
cedures across studies, which could increase variability be-
tween studies.
Using CVR as an outcome, we found greater within-
group variability in the patient group compared with the con-
trol group, across all cognitive domains (eDiscussion in Supple-
ment 1). This indicates a wider distribution of cognitive ability
in those with FEP compared with the normal population, mean-
ing that some may experience severe cognitive problems and
some may be relatively spared or may even be performing
above average. The results can be viewed as support for the
suggestion that patients with psychotic disorders can be di-
vided into different cognitive subgroups,16,17 which may, in
turn, share underlying biological disease mechanisms.34
Several lines of data indicate that cognitive impairment in
psychotic disorders typically starts years before psychosis
onset8-10; however, it is not clear whether the transition from
the CHR-P or prodromal stage to a psychotic syndrome is as-
sociated with further cognitive deterioration.35 In a meta-
analysis by Catalan and colleagues36 of cognitive function in

jamapsychiatry.com (Reprinted) JAMA Psychiatry Published online February 28, 2024 E7

Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024

 Research Original Investigation Cognitive Function and Variability in Patients With First-Episode Psychosis

those at CHR-P compared with healthy controls, the effect sizes
obtained were in the medium range (−0.39 to −0.51), ie, con-
siderably smaller than our results in patients with FEP. At the
same time, the variability estimates19 were consistently lower
(CVR range, 1.15-1.45 vs 1.34-1.94). Given that individuals at
CHR-P represent a heterogeneous group where only a minor-
ity of individuals will go on to develop psychosis, less severe
deficits on a group level are to be expected, whereas it may be
seen as more surprising that cognitive variability is lower than
in FEP. In addition to the individuals at CHR-P who transition
to psychosis, the FEP group is made up of the estimated 10%
who do not experience a prodrome37,38 (possibly adding more
preserved cognitive ability) as well as those individuals who
refuse to seek help before emergency services are needed39,40
(possibly adding more impaired individuals). Although inter-
preting cross-sectional data in a longitudinal context should
be done with caution, it could be speculated that the in-
creased variability may be a consequence of this addition of
more spared, and more impaired, groups of patients. A fur-
ther explanation could be that within the FEP group, there are
individuals who experience considerably more cognitive de-
cline than others during the transition to psychosis, resulting
in this even wider distribution of cognitive performance (as
compared with controls). This lends indirect support to the idea
of different cognitive trajectories leading up to and during the
transition to psychosis; however, longitudinal studies are
needed to address these questions.
Limitations
There are several limitations to address. A general concern is
cognitive testing of unmedicated individuals with psychosis.
This procedure provides data free from possible confounding
effects of medication but may be associated with other fac-
tors affecting the test results, such as sleep deprivation, lack
of motivation, and disruptive psychotic symptoms (although
most studies fail to find a strong association between cogni-
tion and psychotic symptoms35). Regarding the CVR analysis,
this measure depends not only on the performance and vari-
ance of the patient group but also on the characteristics of the
control group. Controls who volunteer for research could theo-
retically be a selected group with less variability, meaning that
their similarity would inflate the variability of the patient group.
In the quality assessment of studies, the recruitment proce-
dure for controls was scrutinized, and the majority of studies
were reported using community controls. This would indi-
cate that most researchers have attempted to address this is-
sue. Furthermore, CVR values may have been influenced by
the factors listed previously regarding the testing of individu-
als not taking medication, potentially adding to the degree of
variability. Psychometric features of specific tasks used may
also have contributed to our results, and tasks with lower re-
liability would likely result in more variability. Finally, we aimed
to study cognitive function early in the psychotic illness pro-
cess and, therefore, limited inclusion to studies with a mean
sample DUP of 5 years or less (thereby excluding, eg, chroni-
cally unmedicated samples). However, a subgroup of the stud-
ies included did not report DUP. This is a possible limitation,
although all studies included defined their sample as having
FEP, and the mean age of nearly all samples would indicate a
DUP of less than 5 years.
Conclusions
Results of this systematic review and meta-analysis identi-
fied cognitive dysfunction already present at the onset of psy-
chotic illness and before initiation of antipsychotic treatment
and, for the first time (to our knowledge), revealed higher vari-
ability between individuals with FEP compared with con-
trols. The results highlight the importance of cognitive assess-
ments in clinical practice, in order to identify individuals at risk
for poor functional outcome, and to select individuals for
cognitive remediation,41,42 as part of a precision medicine
approach.

ARTICLE INFORMATION Pharmacology, Karolinska Institutet, Stockholm, clinical trial sponsored by Boehringer Ingelheim. No
Accepted for Publication: December 8, 2023. Sweden (Erhardt, Sellgren). other disclosures were reported.

Published Online: February 28, 2024.
doi:10.1001/jamapsychiatry.2024.0016
Open Access: This is an open access article
distributed under the terms of the CC-BY License.
© 2024 Lee M et al. JAMA Psychiatry.
Author Affiliations: Centre for Psychiatry
Research, Department of Clinical Neuroscience,
Karolinska Institutet, and Stockholm Health Care
Services, Region Stockholm, Sweden (Lee, Borg,
Plavén-Sigray, Larsson, Sellgren,
Fatouros-Bergman, Cervenka); Department of
Medical sciences, Psychiatry, Uppsala University,
Uppsala, Sweden (Cernvall, Cervenka); Center for
Cognitive and Computational Neuropsychiatry,
Department of Clinical Neuroscience, Karolinska
Institutet, & Stockholm Health Care Services,
Region Stockholm, Sweden (Borg); Neurobiology
Research Unit, Copenhagen University Hospital,
Rigshospitalet, Copenhagen, Denmark
(Plavén-Sigray); Department of Physiology and
Author Contributions: Ms Lee had full access to all
of the data in the study and takes responsibility for
the integrity of the data and the accuracy of the
data analysis.
Concept and design: Lee, Borg, Sellgren,
Fatouros-Bergman, Cervenka.
Acquisition, analysis, or interpretation of data:
Lee, Cernvall, Plavén-Sigray, Larsson, Erhardt,
Sellgren, Fatouros-Bergman, Cervenka.
Drafting of the manuscript: Lee, Plavén-Sigray,
Sellgren, Cervenka.
Critical review of the manuscript for important
intellectual content: All authors.
Statistical analysis: Lee, Plavén-Sigray.
Obtained funding: Erhardt, Sellgren, Cervenka.
Administrative, technical, or material support: Lee,
Larsson, Fatouros-Bergman.
Supervision: Borg, Plavén-Sigray, Sellgren,
Fatouros-Bergman, Cervenka.
Conflict of Interest Disclosures: Dr Cernvall
reported having served as a cognitive rater in a
Funding/Support: This work was funded by grant
523-2014-3467 from the Swedish Research Council
(Dr Cervenka).
Role of the Funder/Sponsor: The funder had no
role in the design and conduct of the study;
collection, management, analysis, and
interpretation of the data; preparation, review, or
approval of the manuscript; and decision to submit
the manuscript for publication.
Data Sharing Statement: See Supplement 2.
Additional Contributions: We thank Robin Emsley,
MD, DSc, and PhD candidate Retha Smit
(Stellenbosch University, Cape Town, South Africa);
Camilo de la Fuente-Sandoval, MD PhD,
Francisco-Reyes Madrigal, MD MSc, and Rodolfo
Solis Vivanco, PhD (Instituto Nacional de
Neurología y Neurocirugía, Mexico City, Mexico);
Monte S. Buschsbaum, MD, PhD, and Nora S. Vyas,
PhD (Kingston University, London, UK); Qian Guo,
MD PhD (Shanghai Mental Health Center, Shanghai
Jiao Tong University School of Medicine, Shanghai,

E8 JAMA Psychiatry Published online February 28, 2024 (Reprinted) jamapsychiatry.com

Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024

 Cognitive Function and Variability in Patients With First-Episode Psychosis
China); Yen Kuang Yang, MD, PhD, and Shih-Hsien
Lin, PhD (National Cheng Kung University Hospital,
Taiwan); and Li Dongwei, PhD (State Key
Laboratory of Cognitive Neuroscience and
Learning, Beijing Normal University, Beijing, China)
for graciously providing their data. All mentioned
parties contributed to data collection and received
no compensation for their work.
REFERENCES
1. Heinrichs RW, Zakzanis KK. Neurocognitive
deficit in schizophrenia: a quantitative review of the
evidence. Neuropsychology. 1998;12(3):426-445.
doi:10.1037/0894-4105.12.3.426
2. Dickinson D, Ramsey ME, Gold JM. Overlooking
the obvious: a meta-analytic comparison of digit
symbol coding tasks and other cognitive measures
in schizophrenia. Arch Gen Psychiatry. 2007;64(5):
532-542. doi:10.1001/archpsyc.64.5.532
3. Schaefer J, Giangrande E, Weinberger DR,
Dickinson D. The global cognitive impairment in
schizophrenia: consistent over decades and around
the world. Schizophr Res. 2013;150(1):42-50. doi:10.
1016/j.schres.2013.07.009
4. Green MF. What are the functional
consequences of neurocognitive deficits in
schizophrenia? Am J Psychiatry. 1996;153(3):321-330.
doi:10.1176/ajp.153.3.321
5. Lepage M, Bodnar M, Bowie CR. Neurocognition:
clinical and functional outcomes in schizophrenia.
Can J Psychiatry. 2014;59(1):5-12. doi:10.1177/
070674371405900103
6. Halverson TF, Orleans-Pobee M, Merritt C,
Sheeran P, Fett AK, Penn DL. Pathways to
functional outcomes in schizophrenia spectrum
disorders: meta-analysis of social cognitive and
neurocognitive predictors. Neurosci Biobehav Rev.
2019;105(July):212-219. doi:10.1016/j.neubiorev.
2019.07.020
7. Oomen PP, Begemann MJH, Brand BA, et al.
Longitudinal clinical and functional outcome in
distinct cognitive subgroups of first-episode
psychosis: a cluster analysis. Psychol Med. 2023;53
(6):2317-2327. doi:10.1017/S0033291721004153
8. Cannon TD, Bearden CE, Hollister JM, Rosso IM,
Sanchez LE, Hadley T. Childhood cognitive
functioning in schizophrenia patients and their
unaffected siblings: a prospective cohort study.
Schizophr Bull. 2000;26(2):379-393. doi:10.1093/
oxfordjournals.schbul.a033460
9. MacCabe JH, Wicks S, Löfving S, et al. Decline in
cognitive performance between ages 13 and 18
years and the risk for psychosis in adulthood:
a Swedish longitudinal cohort study in males. JAMA
Psychiatry. 2013;70(3):261-270. doi:10.1001/2013.
jamapsychiatry.43
10. Jonas K, Lian W, Callahan J, et al. The course of
general cognitive ability in individuals with
psychotic disorders. JAMA Psychiatry. 2022;79(7):
659-666. doi:10.1001/jamapsychiatry.2022.1142
11. Fatouros-Bergman H, Cervenka S, Flyckt L,
Edman G, Farde L. Meta-analysis of cognitive
performance in drug-naïve patients with
schizophrenia. Schizophr Res. 2014;158(1-3):156-162.
doi:10.1016/j.schres.2014.06.034
12. Mesholam-Gately RI, Giuliano AJ, Goff KP,
Faraone SV, Seidman LJ. Neurocognition in
first-episode schizophrenia: a meta-analytic review.
jamapsychiatry.com
Downloaded from jamanetwork.com by CONSEJERIA DE SALUD y SERVICIOS user on 02/29/2024
Neuropsychology. 2009;23(3):315-336. doi:10.
1037/a0014708
13. Zhang H, Wang Y, Hu Y, et al. Meta-analysis of
cognitive function in Chinese first-episode
schizophrenia: MATRICS Consensus Cognitive
Battery (MCCB) profile of impairment. Gen Psychiatr.
2019;32(3):e100043. doi:10.1136/gpsych-
2018-100043
14. Nuechterlein KH, Green MF, Kern RS, et al. The
MATRICS Consensus Cognitive Battery, part 1: test
selection, reliability, and validity. Am J Psychiatry.
2008;165(2):203-213. doi:10.1176/appi.ajp.2007.
07010042
15. Tandon R, Nasrallah HA, Keshavan MS.
Schizophrenia, “just the facts” 4. clinical features
and conceptualization. Schizophr Res. 2009;
110(1-3):1-23. doi:10.1016/j.schres.2009.03.005
16. Green MJ, Girshkin L, Kremerskothen K,
Watkeys O, Quidé Y. A systematic review of studies
reporting data-driven cognitive subtypes across the
psychosis spectrum. Neuropsychol Rev. 2020;30
(4):446-460. doi:10.1007/s11065-019-09422-7
17. Carruthers SP, Van Rheenen TE, Gurvich C,
Sumner PJ, Rossell SL. Characterising the structure
of cognitive heterogeneity in schizophrenia
spectrum disorders: a systematic review and
narrative synthesis. Neurosci Biobehav Rev. 2019;
107(August):252-278. doi:10.1016/j.neubiorev.2019.
09.006
18. Nakagawa S, Poulin R, Mengersen K, et al.
Meta-analysis of variation: ecological and
evolutionary applications and beyond. Methods
Ecol Evol. 2015;6(2):143-152. doi:10.1111/2041-210X.
12309
19. Catalan A, Radua J, McCutcheon R, et al.
Examining the variability of neurocognitive
functioning in individuals at clinical high risk for
psychosis: a meta-analysis. Transl Psychiatry. 2022;
12(1):198. doi:10.1038/s41398-022-01961-7
20. Salazar de Pablo G, Radua J, Pereira J, et al.
Probability of transition to psychosis in individuals
at clinical high risk: an updated meta-analysis. JAMA
Psychiatry. 2021;78(9):970-978. doi:10.1001/
jamapsychiatry.2021.0830
21. Page MJ, McKenzie JE, Bossuyt PM, et al. The
PRISMA 2020 statement: an updated guideline for
reporting systematic reviews. BMJ. 2021;372:n71.
doi:10.1136/bmj.n71
22. Brooke BS, Schwartz TA, Pawlik TM. MOOSE
reporting guidelines for meta-analyses of
observational studies. JAMA Surg. 2021;156(8):
787-788. doi:10.1001/jamasurg.2021.0522
23. Ouzzani M, Hammady H, Fedorowicz Z,
Elmagarmid A. Rayyan-a web and mobile app for
systematic reviews. Syst Rev. 2016;5(1):210. doi:10.
1186/s13643-016-0384-4
24. Orhan F, Fatouros-Bergman H, Goiny M, et al;
Karolinska Schizophrenia Project (KaSP)
Consortium. CSF GABA is reduced in first-episode
psychosis and associates to symptom severity. Mol
Psychiatry. 2018;23(5):1244-1250. doi:10.1038/
mp.2017.25
25. Lee M, Fatouros-Bergman H, Plavén-Sigray P,
et al. No association between cortical dopamine D2
receptor availability and cognition in
antipsychotic-naive first-episode psychosis. NPJ
Schizophr. 2021;7(1):46. doi:10.1038/s41537-021-
00176-x
(Reprinted) JAMA Psychiatry Published online February 28, 2024
Original Investigation Research
26. Healey KM, Bartholomeusz CF, Penn DL.
Deficits in social cognition in first episode
psychosis: a review of the literature. Clin Psychol Rev.
2016;50:108-137. doi:10.1016/j.cpr.2016.10.001
27. Harrer M, Cuijpers P, Furukawa TA, Ebert DD.
Doing Meta-Analysis With R: A Hands-On Guide.
Chapman & Hall/CRC Press; 2021.
28. Viechtbauer W. Conducting meta-analyses in R
with the metafor package. J Stat Softw. 2010;36(3):
1-48. doi:10.18637/jss.v036.i03
29. R Core Team. R: a language and environment
for statistical computing. Accessed November 10,
2023. https://www.R-project.org
30. Github. Meta_analysis_cogn_dnFEP. Accessed
November 10, 2023. https://github.com/
MariaLeeR/Meta_analysis_cogn_dnFEP
31. Borenstein M, Hedges LV, Higgins JPT,
Rothstein HR. Introduction to Meta-Analysis. 2nd ed.
Wiley; 2021. doi:10.1002/9781119558378
32. Kahn RS. Why Kraepelin was right:
schizophrenia as a cognitive disorder. In: Silverstein
SM, Moghaddam B, Wykes T. Schizophrenia:
Evolution and Synthesis. MIT Press; 2013.
33. Insel TR. Rethinking schizophrenia. Nature. 2010;
468(7321):187-193. doi:10.1038/nature09552
34. Lewandowski KE. Genetically,
developmentally, and clinically distinct cognitive
subtypes in schizophrenia: a tale of three
trajectories. Am J Psychiatry. 2020;177(4):282-284.
doi:10.1176/appi.ajp.2020.20020132
35. McCutcheon RA, Keefe RSE, McGuire PK.
Cognitive impairment in schizophrenia: aetiology,
pathophysiology, and treatment. Mol Psychiatry.
2023;28(5):1902-1918. doi:10.1038/
s41380-023-01949-9
36. Catalan A, Salazar de Pablo G, Aymerich C, et al.
Neurocognitive functioning in individuals at clinical
high risk for psychosis: a systematic review and
meta-analysis. JAMA Psychiatry. 2021;78(8):859-867.
doi:10.1001/jamapsychiatry.2021.1290
37. Addington J. The prodromal stage of psychotic
illness: observation, detection, or intervention?
J Psychiatry Neurosci. 2003;28(2):93-97.
38. Yung AR, McGorry PD. The initial prodrome in
psychosis: descriptive and qualitative aspects. Aust
N Z J Psychiatry. 1996;30(5):587-599. doi:10.
3109/00048679609062654
39. Norman RMG, Malla AK, Verdi MB, Hassall LD,
Fazekas C. Understanding delay in treatment for
first-episode psychosis. Psychol Med. 2004;34(2):
255-266. doi:10.1017/S0033291703001119
40. Fusar-Poli P, Oliver D, Spada G, Estrade A,
McGuire P. The case for improved transdiagnostic
detection of first-episode psychosis: electronic
health record cohort study. Schizophr Res. 2021;
228:547-554. doi:10.1016/j.schres.2020.11.031
41. Lejeune JA, Northrop A, Kurtz MM.
A meta-analysis of cognitive remediation for
schizophrenia: efficacy and the role of participant
and treatment factors. Schizophr Bull. 2021;47(4):
997-1006. doi:10.1093/schbul/sbab022
42. Wykes T, Huddy V, Cellard C, McGurk SR,
Czobor P. A meta-analysis of cognitive remediation
for schizophrenia: methodology and effect sizes.
Am J Psychiatry. 2011;168(5):472-485. doi:10.1176/
appi.ajp.2010.10060855
E9