Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

Contents lists available at ScienceDirect

Progress in Neuropsychopharmacology
& Biological Psychiatry
journal homepage: www.elsevier.com/locate/pnp

Affective symptoms in schizophrenia are strongly associated with MARK

neurocognitive deficits indicating disorders in executive functions, visual
memory, attention and social cognition
Buranee Kanchanatawana, Supaksorn Thikaa, George Andersonb, Piotr Galeckic,
Michael Maesa,d,e,f,g,⁎,1
a
Department of Psychiatry, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand
b
CRC Scotland & London, Eccleston Square, London, UK
c
Department of Adult Psychiatry, Medical University of Lodz, Poland
d
Department of Psychiatry, Medical University of Plovdiv, Plovdiv, Bulgaria
e
Department of Psychiatry, Faculty of Medicine, State University of Londrina, Londrina, Brazil
f
Revitalis, Waalre, The Netherlands
g
IMPACT Strategic Research Center, Deakin University, Geelong, Australia

A R T I C L E I N F O A B S T R A C T

Keywords: The aim of this study was to assess the neurocognitive correlates of affective symptoms in schizophrenia.
Major depression Towards this end, 40 healthy controls and 80 schizophrenia patients were investigated with six tests of the
Bipolar Cambridge Neuropsychological Test Automated Battery (CANTAB), assessing spatial working memory, paired-
Anxiety association learning, one touch stocking, rapid visual information (RVP), emotional recognition test and intra/
Schizophrenia
extradimensional set shifting. The Hamilton Depression (HDRS) and Anxiety (HAMA) Rating Scales and the
CANTAB
Calgary Depression Scale for Schizophrenia (CDSS) as well as the Positive and Negative Syndrome Scale (PANSS)
Cognition
were also used.
There were highly significant associations between all 6 CANTAB tests and HDRS, HAMA and CDSS (except
RVP) scores. The most significant items associating with neurocognitive impairments in schizophrenia were self-
depreciation (CDSS), fatigue, psychomotor retardation and agitation, psychic and somatic anxiety (HDRS), fears,
cognitive symptoms, somatic-muscular, genito-urinary and autonomic symptoms and anxious behavior (HAMA).
The selected HDRS and HAMA symptoms indicate fatigue, fears, anxiety, agitation, retardation, somatization
and subjective cognitive complaints (SCC) and are therefore labeled “FAARS”. Up to 28.8% of the variance in the
6 CANTAB measurements was explained by FAARS, which are better predictors of neurocognitive impairments
than the PANSS negative subscale score. Neurocognitive deficits in schizophrenia are best predicted by FAARS
combined with difficulties in abstract thinking.
In conclusion, depression and anxiety symptoms accompanying the negative and positive symptoms of
schizophrenia are associated with neurocognitive deficits indicating disorders in executive functions, attention,
visual memory, and social cognition. Neurocognitive deficits in schizophrenia reflect difficulties in abstract
thinking and FAARS, including subjective cognitive complaints.

1. Introduction aggressive and catatonic behaviors (Andreasen and Olsen, 1982;

Lieberman et al., 1991; Mellor, 1991), whilst negative symptoms in-
Although widely accepted as a heterogeneous disorder with com- clude alogia, anhedonia, social inhibition, flattened affect, loss of in-
plex etiologies and courses (Kanchanatawan et al., 2017a), schizo- terest and anhedonia (Andreasen and Olsen, 1982; Lieberman et al.,
phrenia is classically associated with positive, negative and neurocog- 1991; Marneros et al., 1991).
nitive symptoms. Positive symptoms include delusions, hallucinations, Neurocognitive deficits are often evident in schizophrenia and in-
thought disorder and disorganized thinking, as well as hostile, clude impairments in higher-level thinking, decision-making,

⁎
Corresponding author at: IMPACT Strategic Research Center, Barwon Health, Deakin University, Geelong, Victoria, Australia.
E-mail address: dr.michaelmaes@chula.md (M. Maes).
1
https://scholar.google.co.th/citations?user=1wzMZ7UAAAAJ&hl=th&oi=ao.

http://dx.doi.org/10.1016/j.pnpbp.2017.06.031
Received 16 February 2017; Received in revised form 21 June 2017; Accepted 26 June 2017
Available online 27 June 2017
0278-5846/ © 2017 Elsevier Inc. All rights reserved.
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

reasoning, planning, visual memory, working memory, executive priori” hypothesis states that affective symptoms in schizophrenia are
functions, emotional recognition and attention (Reichenberg, 2010; Yu accompanied by neurocognitive dysfunction, independently of classical
et al., 2015; Keefe and Harvey, 2012; Cuesta and Peralta, 1995; negative and positive symptoms.
Schaefer et al., 2013; Seidman et al., 2003; Eng et al., 2013). Evidence
shows that schizophrenia is accompanied by impairments in different 2. Subjects and methods
components of working memory, including spatial working memory,
long-term strategic memory and strategy use (Zilles et al., 2010; 2.1. Participants
Pantelis et al., 1997; Stone et al., 1998; Badcock et al., 2005). Patients
with schizophrenia score lower in tests of visual memory, characterized In this cross-sectional study, Thai individuals of both sexes and aged
by impairments in organizational processing and visual working 18–65 years were recruited, comprising 80 outpatients with schizo-
memory, as well as in spatial visual perceptual discrimination, retention phrenia and 40 healthy controls. The patients were clinically stable and
and retrieval (Seidman et al., 2003; Kim et al., 2014; Tek et al., 2002). met the DSM-IV-TR diagnostic criteria for schizophrenia. All patients
Schizophrenia patients also show impaired spatial planning, as in- were admitted to the Department of Psychiatry, Faculty of Medicine,
dicated by deficits in spatial serial ordering, planning ability and ac- Chulalongkorn University, Bangkok, Thailand. Healthy individuals
curacy (Fraser et al., 2004; Holt et al., 2013; Badcock et al., 2005; Hilti were recruited by word of mouth from the same catchment area.
et al., 2010). These neurocognitive deficits in schizophrenia can con- Exclusion criteria included patients with a lifetime diagnosis of any
tribute to occupational and social disabilities (Keefe and Harvey, 2012) other axis-I DSM-IV/DSM-V disorder, including major depressive dis-
and are strongly associated with the negative symptoms of schizo- order, bipolar disorder, schizoaffective disorder, major anxiety dis-
phrenia (Yu et al., 2015; Kanchanatawan et al., 2017b). orders (obsessive compulsive disorder, post-traumatic stress disorder,
Many individuals with schizophrenia also experience depression panic disorder), substance use disorders, alcohol or substance abuse and
and anxiety symptoms (Emsley et al., 1999). Affective dysregulation in psycho-organic disorders. Healthy individuals were excluded if a life-
schizophrenia has been proposed to be secondary to positive symptoms time diagnosis of any axis-I DSM-IV/DSM-V disorder was evident, in-
rather than negative symptoms, although there are mixed results cluding schizophrenia, or if a positive family history of schizophrenia
(Emsley et al., 1999; Kirschner et al., 2016). In an acute phase, both was present. We excluded all individuals suffering from major medical
unipolar and bipolar depression patients can show cognitive deficits, illness, including neurodegenerative and neuroinflammatory disorders,
including in complex problem solving, executive functions, paired as- such as multiple sclerosis, Parkinson's disease, Alzheimer's disease, and
sociate learning, spatial recognition memory, rapid visual processing, immune/autoimmune disorders, such as inflammatory bowel disease,
visual planning, speed of processing and sustained attention (Egerházi rheumatoid arthritis, diabetes, chronic obstructive pulmonary disease
et al., 2013; Weiland-Fiedler et al., 2004; Mohn and Rund, 2016), some and psoriasis. Individuals treated with immunomodulatory drugs or
of which may persist after clinical remission (Trichard et al., 1995; antioxidant supplements were also excluded. All participants gave
Sweeney et al., 2000). When present in mood disorder patients, cog- written informed consent prior to study participation. The study was
nitive deficits can be associated with impaired psychosocial functioning conducted according to Thai and international ethics and privacy laws.
(Evans et al., 2014), which is highly prevalent in schizophrenia. Other Approval for the study was obtained from the Institutional Review
psychiatric conditions can also be associated with cognitive impair- Board of the Faculty of Medicine, Chulalongkorn University, Bangkok,
ments, such as planning, decision-making, response inhibition, and set- Thailand, which is in compliance with the International Guideline for
shifting-cognitive flexibility, including obsessive compulsive disorder Human Research protection as required by the Declaration of Helsinki,
(OCD) and other anxiety associated classifications (Muller et al., 2015). The Belmont Report, CIOMS Guideline and International Conference on
Neurocognitive deficits can also be associated with trait anxiety in Harmonization in Good Clinical Practice (ICH-GCP).
schizophrenia patients (Lysaker et al., 2005). Bipolar disorder patients
show cognitive difficulties in executive functions that are similar to 2.2. Methods
schizophrenia patients (Pradhan et al., 2008). Such data indicates an
association of cognitive deficits with mood and/or anxiety disorders. Psychometric assessments and other study variables were all col-
Möser et al. (2006) reported that the neurocognitive impairments in lected on the same day as the clinical interview and CANTAB mea-
schizophrenia, including attention and memory, are associated with surements. A senior research psychiatrist and a trained clinical research
depressive symptoms. Other results suggest that there is a lowered at- assistant (master degree in mental health) performed a comprehensive
tention to affective information in patients with schizophrenia (Martin clinical interview (BK) and CANTAB testing (ST), respectively. The
et al., 2013). However, no research in schizophrenia has investigated clinical interview was semi-structured, collecting data on socio-demo-
the affective, including depressive, anxiety and hypomania, correlates graphic and clinical variables, as well as the Positive and Negative
of neurocognitive deficits using the Cambridge Neuropsychological Test Syndrome Scale (PANSS), summed from postive (PANSS +) and nega-
Automated Battery (CANTAB). tive (PANSS −) symptoms (Kay et al., 1986), the Calgary Depression
The CANTAB tests neurocognitive functioning, including attention, Scale for Schizophrenia (CDSS) (Addington et al., 1992), the Hamilton
working memory, planning, executive function and emotional re- Depression Rating Scale (HDRS) (Hamilton, 1960), the Hamilton An-
cognition. The CANTAB tests have been extensively utilized in the xiety Rating Sclale (HAMA) (Hamilton, 1959), the Young Mania Rating
cognitive assessment of a range of neuropsychiatric patients (Kim et al., Scale (YMRS) (Young et al., 1978) and the Mini-International Neu-
2014; Levaux et al., 2007; Kanchanatawan et al., 2017b). The CANTAB ropsychiatric Interview (M.I.N.I.) in a validated Thai translation
tests are computerized and standardized, independent of language, non- (Kittirathanapaiboon and Khamwongpin, 2005). Tobacco use disorder
sensitive for gender while automatically comparing results from pa- (TUD) was assessed using DSM-IV-TR criteria, while the Fagerström
tients with a normative database. Previous use of CANTAB with schi- Nicotine Dependence Scale was used to assess severity of nicotine de-
zophrenia patients indicates widespread impairments in executive pendence (Heatherton et al., 1991). The same day, body weight and
functions, decision-making, working memory, visual memory, attention height were measured, allowing the computation of the body mass
and attentional set-shifting (Levaux et al., 2007; Kanchanatawan et al., index (BMI).
2017b). However, the influence of mood and anxiety in the modulation Six CANTAB tests were used to assess neurocognitive functions in
of CANTAB scores in schizophrenia patients has not been previously schizophrenia patients and controls, namely:
investigated.
The present study investigated the neurocognitive correlates of de- 1) Paired-association learning (PAL) to assess episodic and visual
pressive, anxiety and hypomanic symptoms in schizophrenia. The “a memory and learning. We analyzed the following PAL factors: PAL

169
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

Table 1
Socio-demographic, clinical and neurocognitive data in healthy controls (HCs) and schizophrenia (SCZ) patients divided into those with and without affective symptoms ( ± AFFECT).

Variables HCa SCZ − AFFECTa SCZ + AFFECTa F/X2/KWT df p

Age (years) 37.4 (12.8) 41.1 (10.8) 41.1 (11.6) 1.36 2/117 0.262
Sex (M/F) 10/30a 30/22a 13/15 9.88 2 0.007
Education (years) 14.3 (4.9) 12.6 (3.8) 11.7 (4.8) 2.98 2/117 0.055
TUD (N/Y) 38/2 47/5 28/0 – – –
BMI (kg/m2) 24.0 (4.3) 24.2 (4.6) 24.9 (6.1) 0.32 2/122 0.730
DSM 5 (first/multiple) − 6/29 4/18 0.01 1 0.920
Duration illness (years) − 13.7 (8.9) 16.8 (12.5) 1.51 1/72 0.223
Number psychoses − 3.1 (2.7) 3.4 (2.9) 0.22 1/73 0.638
HDRS 0.6 (2.0)a 4.8 (3.6)a 12.4 (5.4)a 84.49 2/117 < 0.001
CDSS 0.2 (0.9)a 1.4 (1.5)a 6.4 (5.4)a 4.346 2/116 < 0.001
HAMA 2.6 (5.4)a 6.2 (4.0)a 21.0 (8.2)a 94.14 2/117 < 0.001
YMRS 0.1 (0.5)a 1.5 (2.0)a 3.7 (4.8)a 14.60 2/117 < 0.001
PANSS+ 7.0 (0.0)a 11.8 (4.9)a 19.5 (9.1)a KWT − < 0.001
PANSS− 7.0 (0.0)a 17.5 (9.8)a 22.6 (10.9)a KWT − < 0.001
PC PAL − 0.51(0.67)a +0.14 (0.83)a + 0.48 (1.33)a 10.33 2/116 < 0.001
PC RVP − 0.51 (0.62)a +0.24 (1.05)a + 0.30(1.09)a 8.88 2/115 < 0.001
PC SWM − 0.61 (0.88)a +0.31 (0.75)a + 0.43 (1.00)a 16.40 2/115 < 0.001
PC OTS − 0.60 (0.99)a +0.34 (0.74)a + 0.24 (1.06)a 13.44 2/117 < 0.001
PC IED − 0.49 (0.62)a +0.11 (0.88)a + 0.51(1.31)a 10.02 2/117 < 0.001
PC ERT − 0.64 (1.02)a +0.25 (0.80)a + 0.45 (0.84)a 16.04 2/117 < 0.001

All results are shown as mean ( ± SD); F: analyses of variance; X2: analyses of contingency tables; KWT: Kruskal-Wallis test.
TUD: tobacco use disorder.
BMI: body mass index.
HDRS: Hamilton-Depression Rating Scale.
CDSS: Calgary Depression Scale for Schizophrenia.
HAMA: Hamilton-Anxiety Rating Scale.
YMRS: Young Mania Rating Scale.
PANSS +: Positive and Negative Syndrome Scale, sum of the positive symptom subscale.
PANSS −: Positive and Negative Syndrome Scale, sum of the negative symptom subscale.
PC PAL: Paired-association learning to assess visual memory, episodic memory and learning.
PC RVP: Rapid visual information process test to assess visual sustained attention.
PC SWM: Spatial working memory to assess working memory and strategy use.
PC OTS: One touch stockings of Cambridge to assess spatial planning.
PC IED: Intra/extradimensional set shifting to assess rule acquisition and attention set shifting.
PC ERT: Emotion recognition task to test the ability to identify emotions in facial expressions.
a
Results of post-hoc tests comparing the three groups.

number of patterns succeeded, PAL mean errors to success, PAL emotions in facial expressions, providing the following factors: ERT
mean trials to success, PAL first trial memory score, PAL stages mean overall response latency, ERT percent correct and ERT total
completed, PAL number of patterns reached, PAL total errors, PAL number correct.
stages completed on first trial, PAL total trials, PAL total errors
adjusted and PAL total trials adjusted.
2) Spatial working memory (SWM) to measure strategy use and 2.3. Statistics
working memory. Factors analyzed were: SWM strategy, SWM mean
time to first response, SWM between errors, SWM strategy 4–10 Analyses of variance (ANOVAs) or Kruskal-Wallis test were used to
boxes, SWM between errors 4 boxes, SWM total errors, SWM mean assess differences among groups in continuous or ordinal variables,
token preparation time, SWM mean time to last response and SWM while Χ2-tests were used to check associations between categorical
total error 4 boxes. variables. Multivariate general linear model (GLM) analyses was used
3) One touch stockings of Cambridge (OTS) to check spatial planning. to assess the effects of explanatory variables (including the affective
Factors analyzed were: OTS mean latency to first choice, OTS mean rating scales) on dependent variables (e.g. CANTAB measurements)
latency to correct, OTS probability of error given correct, OTS mean while controlling for background variables, such as education, age and
choices to correct, OTS probability solved on first choice and OTS sex. In case there were significant multivariate effects, we used tests for
probability of error given error. between-subject effects to check the (protected) effects of explanatory
4) Intra/extradimensional set shifting (IED) to check attention set- variables on dependent variables. We used model-predicted estimated
shifting and rule acquisition, utilizing the following factors: IED marginal means (SE), which were obtained from the multivariate GLM
complete stage trials, IED completed stage errors, IED Pre-ED errors, analyses, coupled with protected least significant difference to assess
IED errors block 1, IED total errors, IED stages completed, IED total post-hoc differences among treatment means. Parameter estimates were
latency, IED total errors adjusted, IED total trials adjusted and IED used to assess the sign of the association and the impact of the ex-
total trials. planatory on the dependent variables. Cluster analysis was performed
5) Rapid visual information process test (RVP) to check visual sus- to classify schizophrenic patients into relevant groups (clusters) with
tained attention, measuring the following factors: RVP A, RVP respect to affective symptoms using the K-means clustering method
probability of false alarm, RVP probability of hit blocks 1–7, RVP performed on all subjects and the CDSS, HDRS, HAMA and YMRS
probability of hit, RVP total correct rejections, RVP total hits, RVP scores. We used this method to divide the patient group in two sub-
total false alarms, RVP mean latency, RVP total misses, RVP total groups, namely those with increased affective symptoms versus those
hits blocks 1–7 and RVP total misses blocks 1–7. without, and to display the clustering characteristics. All participants
6) Emotion recognition task (ERT) assesses the ability to identify were consequently allocated to three groups, i.e. healthy controls and
the cluster-analysis-generated subgroups of schizophrenic patients with

170
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

and without affective symptoms. Consequently, ANOVAs and Χ2-tests Table 2

were used to assess the differences between the three groups. Factor Results of multivariate GLM analysis with the six first principal components (PCs) sub-
tracted from the 6 CANTAB domains as dependent variables and the scores on the
analysis, namely principal component (PC) method, was used as a data
Hamilton-Depression Rating Scale (HDRS), Calgary Depression Scale for Schizophrenia
reduction method to reduce the number of CANTAB variables into a few (CDSS), Hamilton-Anxiety Rating Scale (HAMA) and Young Mania Rating Scale (YMRS)
interpretable PCs. We subtracted the first PC of each set of CANTAB as explanatory variables.
domains and used the 6 PC scores in subsequent analyses, as indexes of
the variability in each of the 6 CANTAB subtests. The first PC subtracted Type test Dependent Explanatory F df p
variables variables
from the PAL test results (PC PAL) explained 76.6% of the variance in
the PAL data; the first PC subtracted from the SWM data (PC SWM) Multivariate PC PAL, PC SWM, HDRS 5.13 6/102 < 0.001
64.5%; PC OTS: 68.5%; PC IED: 48.7%; PC RVP: 76.8% and PC ERT: ≠1 PC OTS, PC IED,
75.3%. Factor analysis was also used to explore the factor structure of PC RVP, PC ERT
Between- PC PAL HDRS 16.05 1/107 < 0.001
variables in the data set. Towards this end, a quartimax rotation of the
subject PC SWM HDRS 16.02 1/107 < 0.001
relevant PCs was used as determined by means of eigenvalues > 1.0. effect PC OTS HDRS 15.72 1/107 < 0.001
Loadings > 0.450 were considered to be relevant for the interpreta- PC IED HDRS 10.29 1/107 0.002
tion of the data structure. Kaiser-Meyer-Olkin (KMO) Measure of PC RVP HDRS 10.57 1/107 0.002
PC ERT HDRS 22.58 1/107 < 0.001
Sampling Adequacy was checked. Results of regression analyses were
checked for collinearity using the variance inflation factor and toler- Multivariate PC PAL, PC SWM, CDSS 2.88 6/102 0.012
ance. Results of multiple (post-hoc) comparisons were p corrected for ≠2 PC OTS, PC IED,
PC RVP, PC ERT
false discovery rate according to Benjamini and Hochberg (1995). Z- Between- PC PAL CDSS 9.91 1/107 0.002
unit weighted composite scores were computed on z transformed subject PC SWM CDSS 11.46 1/107 0.001
variables. Tests were 2-tailed, and an alpha level of 0.05 indicated a effect PC OTS CDSS 12.79 1/107 0.001
statistically significant effect. All statistical analyses were performed PC IED CDSS 4.55 1/107 0.028
PC RVP CDSS 3.52 1/107 0.063
using IBM SPSS Windows version 22.
PC ERT CDSS 6.98 1/107 0.010

Multivariate PC PAL, PC SWM, HAMA 3.85 6/102 0.002

≠3 PC OTS, PC IED,
3. Results PC RVP, PC ERT
Between- PC PAL HAMA 13.08 1/107 < 0.001
3.1. Descriptive statistics subject PC SWM HAMA 9.08 1/107 0.003
effect PC OTS HAMA 13.19 1/107 < 0.001
PC IED HAMA 13.14 1/107 < 0.001
Table 1 shows the demographic, clinical and CANTAB data in the
PC RVP HAMA 8.23 1/107 0.005
healthy controls and schizophrenia patients, the latter subdivided into PC ERT HAMA 13.04 1/107 < 0.001
those with and without affective symptoms. There were no significant
Multivariate PC PAL, PC SWM, YMRS 1.70 6/102 0.128
differences in age, education, BMI and TUD among the study groups. ≠4 PC OTS, PC IED,
There were no significant differences between the two schizophrenia PC RVP, PC ERT
subgroups in DSM-5 diagnosis into first episode and multiple episodes,
duration of illness or number of psychotic episodes. The total HDRS, HDRS: Hamilton-Depression rating Scale.
CDSS: Calgary Depression Scale for Schizophrenia.
HAMA, YMRS, PANSS + and PANSS − scores were significant different
HAMA: Hamilton-Anxiety Rating Scale.
among the 3 categories and increased from healthy controls to schizo-
YMRS: Young Mania Rating Scale.
phrenic patients without affective symptoms to those with affective PC PAL: Paired-association learning to assess visual memory, episodic memory and
symptoms. The latter group had higher total CDSS scores versus controls learning.
and schizophrenia without affective symptoms. PC PAL, PC SWM, PC PC RVP: Rapid visual information process test to assess visual sustained attention.
OTS, PC IED, PC RVP and PC ERT were significantly higher in both PC SWM: Spatial working memory to assess working memory and strategy use.
schizophrenia subgroups than in controls, whilst there were no sig- PC OTS: One touch stockings of Cambridge to assess spatial planning.
PC IED: Intra/extradimensional set shifting to assess rule acquisition and attention set
nificant differences between schizophrenia subgroups.
shifting.
PC ERT: Emotion recognition task to test the ability to identify emotions in facial ex-
pressions.
3.2. Associations between affective scores and CANTAB results
3.3. Putative effects of confounding variables
Table 2 shows the results of three multivariate GLM analyses with
the 6 CANTAB PCs as dependent variables and the HDRS (regression As shown in Table 2, the effects of the affective rating scales on the
#1), CDSS (regression #2), HAMA (regression #3) or YMRS (regression PC scores were significant even after adjusting for age (inversely as-
#4) scores as primary explanatory variables, while adjusting for age sociated with PC scores) and education (positively associated with the
and education and sex. We found significant multivariate effects of total PC scores), while sex was not significantly associated with the PC
HDRS, CDSS and HAMA, but not YMRS, scores on the 6 CANTAB PCs. scores. For example, regression #1 shows that the effects of the HDRS
Correction for false discovery rates showed that HDRS (p < 0.001), are significant after adjusting for age (F = 4.16, df = 6/102,
CDSS (p = 0.013) and HAM-A (p = 0.002) remained significant. Tests p < 0.001) and education (F = 8.24, df = 6/102, p < 0.001). The
for between-subject effects showed that the HDRS was significantly and effects of HDRS, CDSS and HAM-A remained significant after adjusting
positively associated with all 6 PCs. The CDSS showed a significant for other confounding variables, including TUD (and Fagerström score),
effect on the 6 CANTAB tests but less significant than the HDRS. In BMI, history of electroconvulsive therapy and use of medications. For
addition, tests for between-subject effects showed that the CDSS was example, GLM analyses controlling for these putative background
significantly associated with all PCs, except RVP. The effects on PC IED variables showed that the impact of HDRS on the 6 CANTAB PCs re-
remained significant after p correction (p = 0.034). The HAMA had a mained significant and that none of these variables (except clozapine)
significant effect on the 6 PC scores, while tests for between-subject had significant effects, namely sex (F = 1.46, df = 6/97, p = 0.201),
effects showed that the HAMA was significantly and positively asso- TUD (F = 1.15, df = 6/97, p = 0.337), Fagerström score (F = 1.19,
ciated with all 6 PCs. df = 6/97, p = 0.189), BMI (F = 1.20, df = 6/94, p = 0.313), a

171
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

history of electro-convulsive treatment (F = 1.44, df = 6/95, Table 3

p = 0.207), use of clozapine (F = 3.2, df = 6/98, p = 0.002), halo- Results of multivariate GLM analysis with the six first principal (PCs) components sub-
tracted from the 6 CANTAB domains as dependent variables and the scores on selected
peridol (F = 2.16, df = 6/98, p = 0.054), anxiolytics/hypnotics
items of the Hamilton-Depression Rating Scale (HDRSs), item 3 Calgary Depression Scale
(F = 1.13, df = 6/98, p = 0.351), antidepressants (F = 0.99, df = 6/ for Schizophrenia (CDSS3) or 6 Hamilton-Anxiety Rating Scale (HAMAs) as explanatory
98, p = 0.440), risperidone (F = 0.93, df = 6/97, p = 0.478), fluphe- variables.
nazine (F = 0.66, df = 6/97, p = 0.687) and perphenazine
(F = 0.66,df = 6/97, p = 0.687). Type test Dependent Explanatory F df p
variables variables

3.4. Item analysis in relation to neurocognitive deficits Multivariate PC PAL, PC SWM, HDRSs 5.94 6/102 < 0.001
≠1 PC OTS, PC IED,
In order to define the most important items which are correlated PC RVP, PC ERT
Between- PC PAL HDRSs 19.33 1/107 < 0.001
with the 6 CANTAB PCs we performed multivariate GLM analysis with
subject PC SWM HDRSs 21.92 1/107 < 0.001
the 6 CANTAB variables as dependent variables and the items of the effect PC OTS HDRSs 15.77 1/107 < 0.001
HDRS, CDSS and HAMA (and age and education) as explanatory vari- PC IED HDRSs 15.41 1/107 < 0.002
ables. Results show that 5 HDRS items had significant effects on the 6 PC RVP HDRSs 11.26 1/107 0.001
PC ERT HDRSs 20.62 1/107 < 0.001
PCs, namely HDRS item 7 (HDRS7 or fatigue, F = 3.11, df = 6/102,
p = 0.008), HDRS8 (psychomotor retardation, F = 3.99, df = 6/102, Multivariate PC PAL, PC SWM, CDSS3 2.75 6/102 0.016
p = 0.001), HDRS9 (psychomotor agitation, F = 3.84, df = 6/102, ≠2 PC OTS, PC IED,
PC RVP, PC ERT
p = 0.002), HDRS10 (psychic anxiety, F = 3.60, df = 6/102, Between- PC PAL CDSS3 7.22 1/107 0.008
p = 0.003) and HDRS11 (somatic anxiety, F = 3.35, df = 6/102, subject PC SWM CDSS3 10.99 1/107 0.001
p = 0.005). Only one of the CDSS items had a significant multivariate effect PC OTS CDSS3 11.24 1/107 0.001
effect, namely item CDSS3 (self-depreciation, F = 2.75, df = 6/102, PC IED CDSS3 6.34 1/107 0.013
p = 0.016). Results show that 6 HAMA items had significant effects on Multivariate PC PAL, PC SWM, HAMAs 5.83 6/102 < 0.001
the 6 CANTAB PCs, namely HAMA3 (fears, F = 3.90, df = 6/102, ≠3 PC OTS, PC IED,
PC RVP, PC ERT
p = 0.001), HAMA5 (cognitive symptoms, F = 3.91, df = 6/102,
Between- PC PAL HAMAs 20.49 1/107 < 0.001
p = 0.001), HAMA7 (somatic muscular symptoms, F = 2.58, df = 6/ subject PC SWM HAMAs 13.13 1/107 < 0.001
102, p = 0.023), HAMA12 (genito-urinary symptoms, F = 3.07, effect PC OTS HAMAs 18.87 1/107 < 0.001
df = 6/102, p = 0.008), HAMA13 (autonomic symptoms, F = 3.43, PC IED HAMAs 18.59 1/107 < 0.001
df = 6/102, p = 0.004) and HAMA14 (anxious behavior at interview, PC RVP HAMAs 14.58 1/107 < 0.001
PC ERT HAMAs 17.07 1/107 < 0.001
F = 4.72, df = 6/102, p < 0.001).
The sums of these 5 HDRS items (HDRSs) and 6 HAMA items HDRSs: sum of HDRS item 7 (HDRS7) or fatigue + HDRS8 or psychomotor retardation
(HAMAs) were computed and consequently we used their z transformed + HDRS9 or psychomotor agitation + HDRS10 or psychic anxiety + HDRS11 or somatic
values (zHDRSs and zHAMAs) in statistical analyses. Table 3 shows the anxiety.
results of multivariate GLM analyses with the 6 CANTAB PC scores as HAMAs: sum of HAMA3 or fears + HAMA5 or cognitive symptoms + HAMA7 or somatic
dependent variables and HDRSs, CDSS3 or HAMAs as dependent vari- muscular symptoms + HAMA12 or genito-urinary symptoms + HAMA13 or autonomic
symptoms + HAMA14 or anxious behavior at interview.
ables. We found highly significant effects of HDRSs, CDSS3 and HAMAs
PC PAL: Paired-association learning to assess visual memory, episodic memory and
on the 6 CANTAB PC scores (also after p correction). Tests for between- learning.
subject effects showed significant positive associations between HDRSs PC RVP: Rapid visual information process test to assess visual sustained attention.
and HAMAs and all 6 PCs, while CDSS3 was significantly and positively PC SWM: Spatial working memory to assess working memory and strategy use.
associated with PC PAL, PC SWM, PC OTS and PC IED (also after p PC OTS: One touch stockings of Cambridge to assess spatial planning.
correction). PC IED: Intra/extradimensional set shifting to assess rule acquisition and attention set
shifting.
PC ERT: Emotion recognition task to test the ability to identify emotions in facial ex-
3.5. Effects of positive and negative symptoms on affective symptoms pressions.

Based on the results presented in Tables 2 and 3 it is, however, the selected HDRS and HAMA items, which are significantly correlated
difficult to conclude whether the associations between CANTAB PC with cognitive deficits. We did not include CDSS3 in this composite
scores and the affective symptoms are independent from positive and score as the inclusion of this item lowered the relevance of the zFAARS
negative symptoms of schizophrenia. There are indeed strong associa- score. Subsequently, statistical analyses were carried out to examine
tions between affective and positive or negative symptoms of schizo- whether the zFAARS index and PANSS +, PANSS − or CDSS have in-
phrenia as displayed in Table 4. Table 4 (regression #1) shows the dependent and cumulative effects on the 6 CANTAB PC scores. Table 5
outcome of multivariate GLM analyses with the 4 affective rating scale shows the outcome of different multivariate GLM analyses with the 6
scores as dependent variables and the PANSS − and PANSS + subscale CANTAB PC scores as dependent variables and zFAARS, PANSS +,
scores as explanatory variables. Both PANSS subscales have significant PANSS − or CDSS as explanatory variables. Regressions #1 shows that
effects on the 4 affective rating scales, whereby HDRS and HAMA are zFAARS had a highly significant effect on the 6 CANTAB data with a
associated with both PANSS subscales, while the YMRS was associated
partial eta squared of 0.288. Tests for between-subject effects showed
with PANSS + subscale only. The same table (regression #2) shows that that all 6 CANTAB PCs were highly associated with zFAARS with partial
PANSS + and PANSS − subscale scores have significant effects on
eta squared between 0.123 (PC RVP) to 0.180 (PC PAL) (all results
HDRSs and HAMAs subscale scores. Tests for between subject effects remained significant at p = 0.001 after p correction). Regressions #2
show that both PANSS subscale rating scores were significantly asso-
and #3 show that zFAARS, but not PANSS − or PANSS +, had a sig-
ciated with the HDRSs and HAMAs subscales. nificant effect on the CANTAB data. Thus, adding PANSS − to zFAARS
(regression #2) showed that the former was not significant, while
3.6. Construction of the FAARS composite score predicting neurocognitive zFAARS was significant (p = 0.024 after p correction). Adding PANSS
deficits + to zFAARS showed that the former was not significant, while zFAARS
was significant (p = 0.002 after p correction).
A z-unit weighted composite score was computed as z score of Regression #4 shows that PANSS − had significant effects on all 6
HDRSs (zHDRSs) + zHAMAs = zFAARS. Thus, zFAARS is an index of

172
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

Table 4 impairments in episodic and visual memory and learning, strategy use,
Results of multivariate GLM analyses with the affective rating scale scores as dependent working memory, spatial planning, attention set-shifting, rule acquisi-
variables and the Positive and Negative Syndrome Scale, positive symptom subscale
tion, visual sustained attention and interpretation of emotional pro-
(PANSS +) and negative symptom subscale (PANSS−) as explanatory variables.
cesses. Previous studies showed associations between neurocognitive
Tests Dependent Explanatory F df p impairments and depressive symptoms in schizophrenia (Möser et al.,
variables variables 2006) and affective symptoms in depression, bipolar disorder and an-
xiety disorders. For example, using the MATRICS Consensus Cognitive
Multivariate #1 HDRS, CDSS, PANSS + 14.17 4/107 < 0.001
HAMA YMRS PANSS − 11.84 4/107 < 0.001 Battery, depressed patients showed cognitive impairments in working
Between-subject HDRS PANSS + 15.57 1/110 < 0.001 memory, reasoning/problem solving and speed of processing (Mohn
effect PANSS − 30.14 1/110 < 0.001 and Rund, 2016). During an acute phase of major depression, neuro-
CDSS PANSS + 1.86 1/110 0.176 cognitive impairments are detected in various domains, most notably in
PANSS − 3.34 1/110 0.070
paired associate learning, speed of processing, executive function,
HAMA PANSS + 14.21 1/110 < 0.001
PANSS − 6.21 1/110 0.014 working memory, and learning/short-term memory (Egerházi et al.,
YMRS PANSS + 47.17 1/110 < 0.001 2013; Mohn and Rund, 2016). Overall, depressed and bipolar in-
PANSS − 3.19 1/110 0.077 dividuals show an array of neurocognitive deficits using different cog-
Multivariate #2 HDRSs, HAMAs PANSS + 11.43 2/109 < 0.001 nitive batteries (Evans et al., 2014; Pradhan et al., 2008).
PANSS − 34.77 2/109 < 0.001 In disorders where anxiety is predominant, as in OCD, cognitive
Between-subject HDRSs PANSS + 14.97 1/110 < 0.001 impairments are also commonly evident, including deficits in response
effect PANSS − 70.09 1/110 < 0.001
HAMAs PANSS + 18.82 1/110 < 0.001
inhibition, decision-making, behavioral reversal, planning, set-shifting
PANSS − 13.71 1/110 < 0.001 and cognitive flexibility (Muller et al., 2015). Trait anxiety in schizo-
phrenia patients is also associated with neurocognitive deficits (Lysaker
HDRS: Hamilton-Depression Rating Scale. et al., 2005). However, in order to limit heterogeneity, the present
CDSS: Calgary Depression Scale for Schizophrenia. study excluded all participants with a life-time diagnosis of depression,
HAMA: Hamilton-Anxiety Rating Scale.
bipolar disorder, dysthymic disorder and anxiety disorders, including
YMRS: Young Mania Rating Scale.
PANSS +: Positive and Negative Syndrome Scale, sum of the positive symptom subscale.
schizophrenia patients for whom affective symptoms were core symp-
PANSS −: Positive and Negative Syndrome Scale, sum of the negative symptom subscale. toms, thereby excluding patients with schizoaffective disorder. In-
HDRSs: sum of HDRS item 7 (HDRS7) or fatigue + HDRS8 or psychomotor retardation specting Table 1 shows that the mean values of the HDRS and HAMA in
+ HDRS9 or psychomotor agitation + HDRS10 or psychic anxiety + HDRS11 or somatic the patients are lower than the scores observed in acute major de-
anxiety. pression and severe anxiety disorders. Therefore, it is difficult to com-
HAMAs: sum of HAMA3 or fears + HAMA5 or cognitive symptoms + HAMA7 or somatic
pare our results on affective symptoms in schizophrenia with results
muscular symptoms + HAMA12 or genito-urinary symptoms + HAMA13 or autonomic
symptoms + HAMA14 or anxious behavior at interview.
obtained in patients with primary affective disorders.
Even with more extreme affective presentations being excluded, the
CANTAB PCs, but less significant than zFAARS. Additional item-by-item present results reinforce the role of affective symptoms in schizophrenia
analyses showed that the most significant PANSS − predictors of the 6 as modulators of core cognitive changes (Emsley et al., 1999). The
CANTAB PCs was N5 (difficulty in abstract thinking: F = 8.74, df = 6/ present results also suggest that affective symptoms may be secondary
102, p < 0.001; partial eta squared: 0.339). The best PANSS + pre- to negative and positive symptoms of schizophrenia, as indicated by
dictor of the 6 CANTAB PCs was item P2 (conceptual disorganization: previous work showing that positive symptoms can increase mild hy-
F = 5.35, df = 6/102, p < 0.001, partial eta squared: 0.239). pomanic symptoms, whilst both negative and positive symptoms con-
Regression #5 shows that zFAARS (p = 0.016 after correction) and tribute to anxiety and depression (Emsley et al., 1999; Kirschner et al.,
PANSS − item N5 (difficulty in abstract thinking; p = 0.002 after cor- 2016). Individuals with schizophrenia may be primed to develop af-
rection) had both significant effects on the 6 CANTAB PCs, zFAARS fective symptoms, especially depression, in part via the early develop-
being associated with PC SWM, PC OTS, PC IED and PC ERT, while N5 mental influences on immune-inflammatory and oxidative and ni-
was associated with PC RVP, PC PAL and PC ERT. Regression #6 and trosative processes (Anderson et al., 2013), some of the effects of which
#7 show that after considering the effects of zFAARS, the total CDSS may be mediated via changes in the gut-brain axis (Anderson and Maes,
score and CDSS item 3 did not have significant multivariate effects on 2016). Future research should delineate the developmental and pa-
the 6 CANTAB data. thophysiological underpinnings of the signs and symptoms of positive,
negative, cognitive and affective dimensions in people currently diag-
nosed with schizophrenia.
3.7. Results of factor analysis This study did not observe significant correlations between hypo-
manic symptoms in schizophrenia and neurocognitive impairments.
Table 6 shows the results of PC analysis performed on the 6 C- Previous research found that neuropsychological impairments are more
ANTAB tests, PANSS +, PANSS −, zFAARS and YMRS. KMO measure of widely distributed in the mixed/manic states, namely in temporal,
sampling adequacy was 0.846 (Bartlett's test of sphericity: X2 = 626.8, parietal and frontostriatal systems, than in the depressive state
df = 45, p < 0.001). All variables (except YMRS) loaded highly on the (Weiland-Fiedler et al., 2004). However, the present YMRS results in
first unrotated PC, which explained 41.6% of the variance in the data. schizophrenia patients cannot be compared to those in bipolar disorder
Two PCs were subtracted with eigenvalues > 1.0. Inspection of the patients, given that bipolar disorder was an exclusion factor and the
quartimax-rotated PCs (explaining in total 65.8% of the variance in the investigated patients only had a very mild degree of hypomanic, and
data) showed that the 6 CANTAB PCs loaded highly on PC1 together not manic, symptoms. Interestingly, Weiland-Fiedler et al. (2004) found
with PANSS − and zFAARS. PC2 loaded highly on PANSS +, PANSS −, neurocognitive deficits in clinically depressed patients to be more re-
zFAARS and YMRS. stricted than in manic patients, whereas the present results indicate
strong associations with depressive symptoms, but not with hypomanic
4. Discussion symptoms in schizophrenia. The exclusion of clinical mania is therefore
likely to have limited the heterogeneity of cognitive deficits in schizo-
The first major finding of this study is that affective symptoms in phrenia patients.
schizophrenia are strongly related to different neurocognitive deficits, Notably, previous research in depression indicates that marked
as measured by the six CANTAB tests. These results indicate neuropsychological deficits are more commonly evident in midlife

173
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

Table 5
Results of multivariate GLM analysis with the six first principal components (PCs) subtracted from the 6 CANTAB tests as dependent variables and the scores of the positive (PANSS +)
and negative (PANSS −) subscales of the Positive and Negative Syndrome Scale, the Calgary Depression Scale of Schizophrenia (CDSS) and a new index based on selected affective
symptoms in schizophrenia (zFAARS) as explanatory variables.

Type test Dependent variables Explanatory variables F df p

Multivariate ≠ 1 PC PAL, PC SWM, PC OTS, PC IED, PC RVP, PC ERT zFAARS symptoms 6.87 6/102 < 0.001
Between-subject effects PC PAL zFAARS symptoms 23.43 1/107 < 0.001
PC SWM zFAARS symptoms 19.95 1/107 < 0.001
PC OTS zFAARS symptoms 20.28 1/107 < 0.001
PC IED zFAARS symptoms 19.95 1/107 < 0.002
PC RVP zFAARS symptoms 15.01 1/107 < 0.001
PC ERT zFAARS symptoms 22.02 1/107 < 0.001

Multivariate ≠ 2 PC PAL, PC SWM, PC OTS, PC IED, PC RVP, PC ERT zFAARS symptoms 2.90 6/101 0.012
PANSS − 1.45 6/101 0.205

Multivariate ≠ 3 PC PAL, PC SWM, PC OTS, PC IED, PC RVP, PC ERT zFAARS symptoms 4.92 6/101 < 0.001
PANSS + 1.41 6/101 0.217

Multivariate ≠ 4 PC PAL, PC SWM, PC OTS, PC IED, PC RVP, PC ERT PANSS − 5.10 6/102 < 0.001
Between-subject effects PC PAL PANSS − 18.77 1/107 < 0.001
PC SWM PANSS − 16.50 1/107 < 0.001
PC OTS PANSS − 9.80 1/107 0.002
PC IED PANSS − 7.09 1/107 0.009
PC RVP PANSS− 17.59 1/107 < 0.001
PC ERT PANSS− 19.49 1/107 < 0.001

Multivariate ≠ 5 PC PAL, PC SWM, PC OTS, PC IED, PC RVP, PC ERT zFAARS symptoms 2.77 6/101 0.016
N5 PANSS − 4.86 6/101 0.001

Multivariate ≠ 6 PC PAL, PC SWM, PC OTS, PC IED, PC RVP, PC ERT zFAARS symptoms 4.77 6/101 < 0.001
CDSStotal 1.04 6/101 0.406

Multivariate ≠ 7 PC PAL, PC SWM, PC OTS, PC IED, PC RVP, PC ERT zFAARS symptoms 7.06 6/101 < 0.001
CDSS3 1.45 6/101 0.205

PC PAL: Paired-association learning to assess visual memory, episodic memory and learning.
PC RVP: Rapid visual information process test to assess visual sustained attention.
PC SWM: Spatial working memory to assess working memory and strategy use.
PC OTS: One touch stockings of Cambridge to assess spatial planning.
PC IED: Intra/extradimensional set shifting to assess rule acquisition and attention set shifting.
PC ERT: Emotion recognition task to test the ability to identify emotions in facial expressions.
PANSS +: Positive and Negative Syndrome Scale, sum of the positive symptom subscale.
PANSS −: Positive and Negative Syndrome Scale, sum of the negative symptom subscale.
zFAARS: computed as z value of sum of items 7 + 8 + 9 + 10 + 11 of the Hamilton Depression Rating Scale + z value of items 3 + 5 + 7 + 12 + 13 + 14 of the Hamilton Anxiety
Rating Scale. FAARS symptoms indicate symptoms of fatigue, fears, anxiety, agitation, retardation, somatization and subjective cognitive complaints.
N5 PANSS −: item 5 of the PANSS negative subscale, namely “difficulty in abstract thinking”.
CDSS3: item 3 of the CDSS, namely “self-depreciation”.

patients with severe depression and in elderly patients, versus younger 2002), namely concentration disorders and memory disturbances, were
ambulatory depressed patients (Grant et al., 2001). However, the pre- strongly associated with CANTAB measurements in controls and in
sent results in schizophrenia shows that affective symptoms modulation patients with schizophrenia (Kanchanatawan et al., 2017a,b). However,
of neurocognition is independent of age, although a general age-asso- other studies sometimes showed a discrepancy between SCCs and ob-
ciated decline in neurocognitive functions was evident. jective neurocognitive test results in healthy individuals and patients
The second major finding of this study is that selected affective with psychiatric disorders (Stenfors et al., 2013).
symptoms, belonging to HDRS and HAMA rating scales, were more The results of the current study also contribute to the understanding
strongly related to neurocognitive deficits in schizophrenia, versus the on the relationship between negative symptoms and neurocognitive
effects of negative symptoms on neurocognition. Thus, FAARS symp- impairments in schizophrenia. Harvey et al. (2006) posited two general
toms, namely fears, fatigue, anxiety, agitation, retardation and auto- models in regard to this: 1) negative symptoms and neurocognitive
nomic, somatic and subjective cognitive symptoms, are strongly asso- deficits are identical features of schizophrenia; and 2) negative symp-
ciated with cognitive impairments in schizophrenia, with FAARS being toms and neurocognitive impairments are distinct dimensions with a
statistically better predictors of neurocognitive deficits than negative different pathophysiology. However, the present results indicate that
symptoms. However, FAARS (and affective) symptoms are strongly the relationship is even more complex, with neurocognitive symptoms
predicted by negative and positive symptoms, making it difficult to more strongly related to FAARS symptoms. This could suggest that
delineate whether FAARS or negative symptoms (or both) underpin the neurocognitive impairments in schizophrenia are explained by FAARS
neurocognitive deficits in schizophrenia. symptoms, which may develop as a consequence of negative as well as
Interestingly, one of the FAARS symptoms is the item “difficulty in positive symptoms. Or neurocognitive impairments are partly side ef-
concentration and poor memory” of the HAMA, indicating that such fects of FAARS symptoms, such as agitation, retardation, fatigue, and
subjective cognitive complaints (SCCs) are strongly associated with the autonomic symptoms, which may interfere with normal cognitive
CANTAB's neurocognitive measurements. Also of note, the two most functioning. Alternatively, and perhaps more likely, is that neurocog-
important PANSS predictors of neurocognitive impairments were con- nitive symptoms in schizophrenia are in part objective sings of SCCs
ceptual disorganization and difficulties in abstract thinking. We re- (including concentration and memory dysfunctions), conceptual dis-
ported previously that two items of the FibroFatigue scale, a rating organization, difficulties in abstract thinking, and in part side effects of
scale for fibromyalgia and chronic fatigue syndrome (Zachrisson et al., some FAARS symptoms. Given the great heterogeneity in the

174
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

Table 6 prospective, in order to examine the onset of affective and FAARS

Results of factor analysis (principal component method) performed on the 6 CANTAB symptoms in relation to positive and negative symptoms in a cohort of
measurements, the scores of the positive (PANSS +) and negative (PANSS −) subscales of
individuals at risk of developing schizophrenia. Given that affective
the Positive and Negative Syndrome Scale, the Young mania Rating Sale (YMRS) and a
new index of selected affective symptoms in schizophrenia (zFAARS). dysregulation may be primed by physiological changes in early devel-
opment (Anderson et al., 2013), it will be important to investigate the
Variables Unrotated PC (first) Rotated PCsa biological correlates of FAARS and how these link to positive, negative
and neurocognitive changes. It could be argued that the CDSS is the
PC1 rotated PC2 rotated
most validated questionnaire to measure depressive symptoms in
PC SWM 0.766 0.808 0.034 schizophrenia, while including HDRS may cause a multiple testing
PC RVP 0.762 0.794 0.063 problem. Nevertheless, results were p corrected for false discovery rate
PC PAL 0.810 0.853 0.042 (Benjamini and Hochberg, 1995). In addition, a recent systematic re-
PC OTS 0.759 0.813 0.004
view showed that the HDRS can reliably measure depressive psycho-
PC IED 0.732 0.726 0.152
PC ERT 0.785 0.832 0.026 pathology in schizophrenia (Lako et al., 2012), while this study shows
PANSS+ 0.620 0.333 0.850 that the HDRS score is significantly better associated with objective
PANSS− 0.685 0.539 0.503 neurocognitive deficits than the CDSS.
zFAARS 0.754 0.574 0.601
Fig. 1 summarizes this study's main findings. Neurocognitive
YMRS 0.391 0.124 0.758
Explained variance (%) 41.6% 46.4% 19.4% symptoms are strongly related to FAARS (fatigue, fears, anxiety, agi-
tation, retardation, automatic and somatic symptoms and subjective
PC PAL: Paired-association learning to assess visual memory, episodic memory and cognitive complaints), which are part of depression and anxiety rating
learning. scales and are determined by negative symptoms and to a lesser extent
PC RVP: Rapid visual information process test to assess visual sustained attention. also by positive symptoms. Neurocognitive deficits are not associated
PC SWM: Spatial working memory to assess working memory and strategy use.
with positive and hypomanic symptoms, although positive, negative,
PC OTS: One touch stockings of Cambridge to assess spatial planning.
PC IED: Intra/extradimensional set shifting to assess rule acquisition and attention set
affective, FAARS and hypomanic symptoms are strongly inter-
shifting. correlated. These variable symptom dimension presentations, which
PC ERT: Emotion recognition task to test the ability to identify emotions in facial ex- additionally may change over time, are likely to contribute to a fuller
pressions. understanding of wider aspects of the pathophysiology and classifica-
PANSS +: Positive and Negative Syndrome Scale, sum of the positive symptom subscale. tion of individuals with schizophrenia.
PANSS −: Positive and Negative Syndrome Scale, sum of the negative symptom subscale.
zFAARS: computed as z value of sum of items 7 + 8 + 9 + 10 + 11 of the Hamilton
Depression Rating Scale + z value of items 3 + 5 + 7 + 12 + 13 + 14 of the Hamilton
Conflict of interest
Anxiety Rating Scale. FAARS symptoms indicate symptoms of fatigue, fears, anxiety,
agitation, retardation, somatization and subjective cognitive complaints. The authors have no conflict of interest with any commercial or
YMRS: Young Mania Rating Scale. other association in connection with the submitted article.
a
Shown are the results of quartimax rotation of the first two PCs. Relevant loadings
(> 0.450) are shown in bold.
Author's contributions

BK and MM designed the study. BK recruited patients and com-

pleted diagnostic interviews and rating scales measurements. ST per-
formed CANTAB measurements. MM carried out the statistical analyses.
All authors contributed to interpretation of the data and writing of the
manuscript.

Ethical statement

The study was conducted according to Thai and international ethics

and privacy laws. Approval for the study was obtained from the
Institutional Review Board of the Faculty of Medicine, Chulalongkorn
University, Bangkok, Thailand, which is in compliance with the
International Guideline for Human Research protection as required by
the Declaration of Helsinki, The Belmont Report, CIOMS Guideline and
International Conference on Harmonization in Good Clinical Practice
(ICH-GCP).

Acknowledgement

Fig. 1. Symptom dimensions of schizophrenia, including neurocognitive deficits, positive, This research has been supported by the Asahi Glass Foundation,
negative, hypomanic and FAARS symptoms. Chulalongkorn University Centenary Academic Development Project.

presentations and developmental etiologies of people currently diag- References

nosed with schizophrenia, future research on pathophysiological pro-
cesses of underlying such heterogeneity should better clarify such Addington, D., Addington, J., Maticka-Tyndale, E., Joyce, J., 1992. Reliability and va-
lidity of a depression rating scale for schizophrenics. Schizophr. Res. 6, 201–208.
classical measures as positive, negative and neurocognitive symptoms,
Anderson, G., Maes, M., 2016. How immune-inflammatory processes link CNS and psy-
including as to how FAARS links to this, and indeed, if FAARS is bio- chiatric disorders: classification and treatment implications. CNS Neurol. Disord.
logically meaningful. Drug Targets (Nov 22 [Epub ahead of print] PubMed PMID: 27875954).
A limitation of this study is its cross-sectional design, which pre- Anderson, G., Maes, M., Berk, M., 2013. Schizophrenia is primed for an increased ex-
pression of depression through activation of immuno-inflammatory, oxidative and
vents causal modeling. Therefore, future research should be nitrosative stress, and tryptophan catabolite pathways. Prog. Neuro-

175
B. Kanchanatawan et al. Progress in Neuropsychopharmacology & Biological Psychiatry 80 (2018) 168–176

Psychopharmacol. Biol. Psychiatry 42, 101–114. Lieberman, J.A., Jody, D., Alvir, M.J., Borenstein, M., Mayerhoff, D.I., Geisler, S.,
Andreasen, N.C., Olsen, S., 1982. Negative v positive schizophrenia. Definition and va- Szymanski, S., Gonzales, A., Bogerts, B., 1991. Negative symptoms in schizophrenia:
lidation. Arch. Gen. Psychiatry 39, 789–794. relationship to positive symptoms and outcome. In: Marneros, A., Andreasen, N.,
Badcock, J.C., Michiel, P.T., Rock, D., 2005. Spatial working memory and planning Tsuang, M.T. (Eds.), Negative Versus Positive Schizophrenia. Springer-Verlag, Berlin
ability: contrasts between schizophrenia and bipolar I disorder. Cortex 41, 753–763. Heidelberg, pp. 109–125.
Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: a practical and Lysaker, P.H., Davis, L.W., Lightfoot, J., Hunter, N., Stasburger, A., 2005. Association of
powerful approach to multiple testing. J. Royal Statist. Soc. Series B (Methodological) neurocognition, anxiety, positive and negative symptoms with coping preference in
57, 289–300. schizophrenia spectrum disorders. Schizophr. Res. 80, 163–171.
Cuesta, M.J., Peralta, V., 1995. Cognitive disorders in the positive, negative, and dis- Marneros, A., Deister, A., Rohde, A., 1991. Long-term monomorphism of negative and
organization syndromes of schizophrenia. Psychiatry Res. 58, 227–235. positive schizophrenic episodes. In: Marneros, A., Andreasen, N., Tsuang, M.T. (Eds.),
Egerházi, A., Balla, P., Ritzl, A., Varga, Z., Frecska, E., Berecz, R., 2013. Automated Negative Versus Positive Schizophrenia. Springer-Verlag, Berlin Heidelberg, pp.
neuropsychological test battery in depression — preliminary data. 183–196.
Neuropsychopharmacol. Hung. 15, 5–11. Martin, E.A., Becker, T.M., Cicero, D.C., Kerns, J.G., 2013. Examination of affective and
Emsley, R.A., Oosthuizen, P.P., Joubert, A.F., Roberts, M.C., Stein, D.J., 1999. Depressive cognitive interference in schizophrenia and relation to symptoms. J. Abnorm.
and anxiety symptoms in patients with schizophrenia and schizophreniform disorder. Psychol. 122, 733–744.
J. Clin. Psychiatry. 60, 747–751. Mellor, C.S., 1991. Methodological problems in identifying and measuring first-rank
Eng, G.K., Lam, M., Bong, Y.L., Subramaniam, M., Bautista, D., Rapisarda, A., Kraus, M., symptoms of schizophrenia. In: Marneros, A., Andreasen, N., Tsuang, M.T. (Eds.),
Lee, J., Collinson, S.L., Chong, S.A., Keefe, R.S., 2013. Brief assessment of cognition in Negative Versus Positive Schizophrenia. Springer-Verlag, Berlin Heidelberg, pp.
schizophrenia: normative data in an English-speaking ethnic Chinese sample. Arch. 71–78.
Clin. Neuropsychol. 28, 845–858. Mohn, C., Rund, B.R., 2016. Neurocognitive profile in major depressive disorders:
Evans, V.C., Iverson, G.L., Yatham, L.N., Lam, R.W., 2014. The relationship between Relationship to symptom level and subjective memory complaints. BMC Psychiatry
neurocognitive and psychosocial functioning in major depressive disorder: a sys- 16, 108.
tematic review. J. Clin. Psychiatry. 75, 1359–1370. Möser, C., Krieg, J.C., Zihl, J., Lautenbacher, S., 2006. Attention and memory deficits in
Fraser, D., Park, S., Clark, G., Yohanna, D., Houk, J.C., 2004. Spatial serial order pro- schizophrenia: the role of symptoms of depression. Cogn. Behav. Neurol. 19,
cessing in schizophrenia. Schizophr. Res. 70, 203–213. 150–156.
Grant, M.M., Thase, M.E., Sweeney, J.A., 2001. Cognitive disturbance in outpatient de- Muller, J. de Lima, Torquato, K.I., Manfro, G.G., Trentini, C.M., 2015. Executive functions
pressed younger adults: evidence of modest impairment. Biol. Psychiatry 50, 35–43. as a potential neurocognitive endophenotype in anxiety disorders: a systematic re-
Hamilton, M., 1959. The assessment of anxiety states by rating. Br. J. Med. Psychol. 32, view considering DSM-IV and DSM-5 diagnostic criteria classification. Dement.
50–55. Neuropsychol. 9 (3), 285–294.
Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry 23, Pantelis, C., Barnes, T.R., Nelson, H.E., Tanner, S., Weatherley, L., Owen, A.M., Robbins,
56–62. T.W., 1997. Frontal-striatal cognitive deficits in patients with chronic schizophrenia.
Harvey, P.D., Koren, D., Reihenberg, A., Bowie, C.R., 2006. Negative symptoms and Brain 120, 1823–1843.
cognitive deficits: what is the nature of their relationship? Schizophr. Bull. 32, Pradhan, B.K., Chakrabarti, S., Nehra, R., Mankotia, A., 2008. Cognitive functions in
250–258. bipolar affective disorder and schizophrenia: comparison. Psychiatry Clin. Neurosci.
Heatherton, T.F., Kozlowski, L.T., Frecker, R.C., Fagerström, K.O., 1991. The Fagerström 62, 515–525.
test for nicotine dependence: a revision of the Fagerström tolerance questionnaire. Br. Reichenberg, A., 2010. The assessment of neuropsychological functioning in schizo-
J. Addict. 86, 1119–1127. phrenia. Dialogues Clin. Neurosci. 12, 383–392.
Hilti, C.C., Delko, T., Orosz, A.T., Thomann, K., Ludewig, S., Geyer, M.A., Vollenweider, Schaefer, J., Giangrande, E., Weinberger, D.R., Dickinson, D., 2013. The global cognitive
F.X., Feldon, J., Cattapan-Ludewig, K., 2010. Sustained attention and planning def- impairment in schizophrenia: Consistent over decades and around the world.
icits but intact attentional set-shifting in neuroleptic-naïve first-episode schizo- Schizophr. Res. 150, 42–50.
phrenia patients. Neuropsychobiology 61, 79–86. Seidman, L.J., Lanca, M., Kremen, W.S., Faraone, S.V., Tsuang, M.T., 2003.
Holt, D.V., Wolf, J., Funke, J., Weisbrod, M., Kaiser, S., 2013. Planning impairments in Organizational and visual memory deficits in schizophrenia and bipolar psychoses
schizophrenia: specificity, task independence and functional relevance. Schizophr. using the Rey-Osterrieth complex figure: effects of duration of illness. J. Clin. Exp.
Res. 149, 174–179. Neuropsychol. 25, 949–964.
Kanchanatawan, B., Sirivichayakul, S., Ruxrungtham, K., Carvalho, A.F., Geffard, M., Stenfors, C.U., Marklund, P., Magnusson Hanson, L.L., Theorell, T., Nilsson, L.G., 2013.
Ormstad, H., Anderson, G., Maes, M., 2017a. Deficit, but not nondeficit, schizo- Subjective cognitive complaints and the role of executive cognitive functioning in the
phrenia is characterized by mucosa-associated activation of the tryptophan catabolite working population: a case-control study. PLoS One 8, e83351.
(TRYCAT) pathway with highly specific increases in IgA responses directed to pico- Stone, M., Gabrieli, J.D., Stebbins, G.T., Sullivan, E.V., 1998. Working and strategic
linic, xanthurenic, and quinolinic acid. Mol. Neurobiol.. http://dx.doi.org/10.1007/ memory deficits in schizophrenia. Neuropsychology 12, 278–288.
s12035-017-0417-6. ([Epub ahead of print] PubMed PMID: 28181189). Sweeney, J.A., Kmiec, J.A., Kupfer, D.J., 2000. Neuropsychologic impairments in bipolar
Kanchanatawan, B., Thika, S., Sirivichayakul, S., Carvalho, A.F., Anderson, G., Maes, M., and unipolar mood disorders on the CANTAB neurocognitive battery. Biol. Psychiatry
2017b. Neurocognitive impairments in schizophrenia are strongly associated with 48, 674–684.
negative symptoms and discriminate deficit from nondeficit schizophrenia, (sub- Tek, C., Gold, J., Blaxton, T., Wilk, C., McMahon, R.P., Buchanan, R.W., 2002. Visual
mitted). perceptual and working memory impairments in schizophrenia. Arch. Gen.
Kay, S.R., Fiszbein, A., Opler, L.A., 1986. Negative symptom rating scale: limitations in Psychiatry 59, 146–153.
psychometric and research methodology. Psychiatry Res. 19, 169–173. Trichard, J.L., Martinot, M., Alagille, M.C., Masure, P., Hardy, D., Ginest, E., et al., 1995.
Keefe, R.S., Harvey, P.D., 2012. Cognitive impairment in schizophrenia. Handb. Exp. The course of prefrontal lobe dysfunction in severely depressed in-patients: a long-
Pharmacol. 213, 11–37. itudinal neuropsychological study. Psychol. Med. 25, 79–85.
Kim, H.S., An, Y.M., Kwon, J.S., Shin, M.S., 2014. A preliminary validity study of the Weiland-Fiedler, P., Erickson, K., Waldeck, T., Luckenbaugh, D.A., Pike, D., Bonne, O.,
Cambridge neuropsychological test automated battery for the assessment of executive Charney, D.S., Neumeister, A., 2004. Evidence for continuing neuropsychological
function in schizophrenia and bipolar disorder. Psychiatr. Investig. 11, 394–401. impairments in depression. J. Affect. Disord. 82, 253–258.
Kirschner, M., Aleman, A., Kaiser, S., 2016. Secondary negative symptoms - a review of Young, R.C., Biggs, J.T., Ziegler, V.E., Meyer, D.A., 1978. A rating scale for mania: re-
mechanisms, assessment and treatment. Schizophr. Res.. http://dx.doi.org/10.1016/ liability, validity and sensitivity. Br. J. Psychiatry 133, 429–435.
j.schres.2016.05.003. (May 23 pii:S0920-9964(16)30224-9 [Epub ahead of print] Yu, M., Tang, X., Wang, X., Zhang, X., Zhang, X., Sha, W., Yao, S., Shu, N., Zhang, X.,
Review. PubMed PMID: 27230288). Zhang, Z., 2015. Neurocognitive impairments in deficit and non-deficit schizophrenia
Kittirathanapaiboon, P., Khamwongpin, M., 2005. The Validity of the Mini International and their relationships with symptom dimensions and other clinical variables. PLoS
Neuropsychiatric Interview (M.I.N.I.). Thai Version Suanprung Hospital, Department One 10, e0138357.
of Mental Health. Zachrisson, O., Regland, B., Jahreskog, M., Kron, M., Gottfries, C.G., 2002. A rating scale
Lako, I.M., Bruggeman, R., Knegtering, H., Wiersma, D., Schoevers, R.A., Slooff, C.J., for fibromyalgia and chronic fatigue syndrome (the FibroFatigue scale). J.
Taxis, K., 2012. A systematic review of instruments to measure depressive symptoms Psychosom. Res. 52, 501–509.
in patients with schizophrenia. J. Affect. Disord. 140, 38–47. Zilles, D., Gruber, E., Falkai, P., Gruber, O., 2010. Patients with schizophrenia show
Levaux, M.N., Potvin, S., Sepehry, A.A., Sablier, J., Mendrek, A., Stip, E., 2007. deficits of working memory maintenance components in circuit-specific tasks. Eur.
Computerized assessment of cognition in schizophrenia: promises and pitfalls of Arch. Psychiatry Clin. Neurosci. 260, 519–525.
CANTAB. Eur. Psychiatry. 22, 104–115.

176