Yonsei Medical Journal

Vol. 47, No. 3, pp. 343 - 353, 2006

Greater Impairment in Negative Emotion Evaluation Ability

in Patients with Paranoid Schizophrenia
Suk Kyoon An,1,2 Eun Lee,3 Jae-Jin Kim,1,2 Kee Namkoong,1,2 Jee In Kang,1 Jong Hee Jeon,2 Jeong Ho Seok,4
and Sung Hwan Choi1

1
Department of Psychiatry, 2Institute of Behavioral Science in Medicine, Yonsei University College of Medicine, Seoul, Korea;
3
Ilsan Hospital, National Health Insurance Corporation, Gyeonggi-do, Korea;
4
Department of Psychiatry, Hallym University College of Medicine, Seoul, Korea.

To explore whether or not patients with schizophrenia INTRODUCTION

display a more profound impairment of negative emotion
processing, we assessed the implicit evaluation of positive A considerable number of reports have been
and negative emotional stimuli. Twenty patients with schizo- published on the subject of emotional evaluation
phrenia (9 paranoid, 11 non-paranoid) and 22 normal controls
impairment, which is a core component of the
were instructed to classify emotional pictures according to
the intrinsic valence if the pictures were black and white. If psychopathology in schizophrenia.1-3 Several re-
the stimuli were color-filtered, participants were instructed to searchers have suggested that there is greater
press the positive/negative response key according to the impairment in the ability to evaluate negative
extrinsic valence (assigned valence of color). The error rates emotions than there is for positive emotions.4-8
of the color-filtered stimuli were used as dependent measures. Others reported the existence of generalized
Normal controls made more errors on trials of the positive deficits which encompass all types of emotions.9-11
pictures when the correct response was the negative response The specific impairment of negative emotion
key than when the correct response was the positive response evaluation ability is thought to stem from the
key. The reverse was true on trials of the negative pictures.
repeated social withdrawal that results from
Patients with schizophrenia, especially paranoid schizo-
phrenia, committed more errors in trials of the positive
negatively- arousing stimuli; social withdrawal is
pictures when the correct response key was the negative one of the typical symptoms of schizophrenia.2,12,13
response key. However, the reverse was not true on trials of In contrast, the generalized emotion recognition
the negative pictures. These findings suggest that patients deficit is due to part of the global reduction in
with paranoid schizophrenia might suffer from an impaired cognitive function, which is also common in
ability to evaluate negative emotions and have a loosening of schizophrenia.2
association within their negative emotional networks. These inconsistent findings may be influenced
Key Words: Schizophrenia, positive emotion, negative emo- by the differences in emotional stimuli, experi-
tion mental and control tasks, subjects' characteristics,
and other variables. Several attempts have been
made to investigate the evaluation of emotions in
the specific subtypes of schizophrenia. Schizo-
phrenic patients with positive symptoms such as
Received August 19, 2005
Accepted November 21, 2005 delusion and hallucination were reported to show
This study was supported by a grant from the Korea Health 21 negative emotion-specific impairment, while
R D Project, Ministry of Health Welfare, Republic of Korea patients with negative symptoms such as affective
(03-PJ1-PG3-20500- 0048). flattening exhibited a generalized emotion-recog-
Reprint address: requests to Dr. Kee Namkoong, Department of nition deficit.14 Patients with paranoid schizo-
Psychiatry, Yonsei University College of Medicine, 134 Shinchon-
dong, Seodaemun-gu, Seoul 120-752, Korea. Tel: 82-2-2228-1625, phrenia have been reported to display less im-
Fax: 82-2-313-0891, E-mail: keen@yumc.yonsei.ac.kr pairment in identifying emotions,15 especially

Yonsei Med J Vol. 47, No. 3, 2006

 Suk Kyoon An, et al.

negative emotions,16 than patients with non-para- tional processing using the EAST, which is less
noid schizophrenia, although there has been a likely to be biased or distorted.
contrary report.17 Recently, paranoid schizophre- The aim of the present study was to determine
nics were reported to show negative (sad, fear) whether patients with schizophrenia had more
emotion- specific impairment in comparison with impairment in their ability to evaluate negative
residual schizophrenics as well as normal con- emotions than did normal controls. We hypothe-
trols, while disorganized schizophrenic patients sized that patients with schizophrenia would be
displayed an impairment in generalized emotion- more profoundly impaired in negative emotion
recognition.18 evaluation, and that the diagnostic subtype of
It is supposed that another possible source of schizophrenics would be related to the ability to
these inconsistent findings is due to the use of recognize emotion. We measured the emotional
self-reporting measures to assess the function of evaluation performance by using the EAST in
emotional evaluation. The self-report may be patients with schizophrenia and in normal con-
affected by a variety of distortions and biases due trols. We propose that this approach could be a
to the social norms, investigators' expectations, better way to assess the separate estimates of the
19
allocations of attention, and other factors. strength of associations within positive or nega-
To be free from these distortions and biases, an tive emotional networks than a traditional ap-
Implicit Association Test (IAT) was developed, proach, which may be influenced by subjects'
which is a well-known indirect measurement tech- intentions or control.19
nique used in social cognition research.20-24 In
addition, the IAT has already been used to study
various forms of psychopathology such as social MATERIALS AND METHODS
anxiety,25,26 depression,27 and alcoholism.28,29 The
IAT assesses the relative strength of an association Subjects
between a target and an attribute by considering
the accuracy or reaction time with which subjects A consecutive series of 21 stable outpatients
can employ two response keys when each has (Department of Psychiatry, Severance Mental
been assigned a dual meaning.24 For example, in Health Hospital, Yonsei University College of
the Greenwald et al.22 flower-insect IAT, subjects Medicine, Korea) with schizophrenia diagnosed
were asked to pair two concepts (e.g., flower and by the Diagnostic and Statistical Manual of Mental
good, or insect and good). The more closely-asso- Disorders, 4th edition (DSM-IV)32 were enrolled in
ciated the two concepts are, the easier it is to the present study. All schizophrenia patients were
respond to them as a single unit. Thus, if there clinically judged to have been stable for more than
was a stronger association between "flower" and two consecutive months prior to the study.
"good" than between "insect" and "good", it would Twenty-two healthy control subjects were re-
be easier to respond faster and with more accu- cruited through a local newspaper advertisement.
racy when subjects are requested to give the same All subjects fulfilled the inclusion criteria, which
response than it would be when "insect" and consisted of their being between the ages of 18
"good" are paired. Recently, the Extrinsic Affective and 50, having completed more than 12 years of
Simon Task (EAST) was developed by De Houwer education, and being right-handed. Handedness
30,31
as a modified version of the IAT. Compared to was assessed using the revised version of Annett's
the IAT, the EAST has a moderate or strong effect Hand Preference Questionnaire.33,34 All subjects
size (IAT: usually Cohen's d > 0.6). In addition, were assessed using a comprehensive psychiatric
the EAST can provide separate estimates of the evaluation by means of the Structured Clinical
emotional evaluations of positive and negative Interview for the DSM-IV (SCID).35,36 The diag-
emotional categories simultaneously.30,31 Based on nosis of schizophrenia and its subtypes was made
the above sources, we considered that it should be by psychiatrists (SJH) using the SCID according to
possible to investigate the separate estimates of the DSM-IV criteria. The psychopathology of
associations within positive and negative emo- patients with schizophrenia was assessed by a

Yonsei Med J Vol. 47, No. 3, 2006

 Impairment in Negative Emotion Evaluation in Paranoid Schizophrenia

psychiatrist (LE) using the Positive and Negative 5.28 ± 2.28, according to the ratings by another 20
Syndrome Scale (PANSS).37 Subjects with additio- subjects) from our own collection for the positive
nal Axis I or II disorders, neurological disorders, stimuli and 2455, 2700, 3170, 3350, 6212 for the
a past history of loss of consciousness, or relevant negative stimuli).
visual defects were excluded. This investigation In the color trials, the pictures had either a blue
was carried out in accordance with the Declara- or a green-colored filter placed over them. The
tion of Helsinki. The study was reviewed and green or blue filtering was created by setting the
approved by the Institutional Review Board of red, green and blue values at 0, 100 and 60 or at
Severance Mental Health Hospital. All subjects 0, 60 and 100, respectively, in the MGI Photo Suite
signed a written informed consent form. III program. Each stimulus was presented in ran-
One patient did not complete the session dom order in all blocks, with the restriction that
because of poor task compliance and was there- the same picture should not appear in two or
fore excluded, which resulted in the final study more consecutive trials and that the response keys
population consisting of 20 patients and 22 nor- should not be the same in four or more consecu-
mal controls. tive trials. The inter-trial interval was 1,500 ms.

Stimuli Procedures

Five positive and five negative emotional The procedure used for the EAST was essen-
stimuli were presented in the color trials, and tially the same as that used in De Houwer's
another five positive and five negative stimuli study,31 which demonstrates the reliability and
were presented in the black and white trials. All validity of the EAST as a tool for assessing emo-
stimuli were selected from the International tional evaluation ability. The subjects were seated
Affective Pictures System (IAPS)38 and from our in front of a computer monitor at an eye-to-screen
own collection, and were then transformed into 27 distance of approximately 40 cm. Subjects were
× 20 cm-sized black and white pictures. To use requested to classify the emotional pictures by
pictures of comparable levels of valence, arousal pressing the positive response key (i.e. the “p”
and gray color intensity, 60 black and white key) or the negative response key (i.e. the “q” key)
transformed pictures were selected across a whole according to the intrinsic valence of black and
range of valence and arousal. To avoid any dif- white stimuli or to the assigned valence of color
ferences in the filtering effect due to variable gray (extrinsic valence). In the black and white trials,
color intensity across these 60 pictures, 35 pictures they were told to press the positive response key
were chosen within the range of -0.8 - +0.8 stan- (p) in response to the positive pictures and to
dard deviation of the average calculated mean press the negative key (q) for the negative
level of gray color intensity of each picture using pictures. In the color trials, however, they were
a histogram constructed with Adobe Photoshop instructed to press the positive (p) or negative (q)
7.0. Then, 10 positive and 10 negative stimuli were key according to the color (Table 1). Half of the
selected according to the ranks of mean valence subjects were instructed to press the positive key
ratings of the 35 pictures by another 20 laypersons (p) for the green-filtered pictures, and to press the
using the Self-Assessment Manikin (SAM).39 negative key (q) for the blue-filtered pictures,
Finally, each positive and negative picture was while this color assignment was inversed for the
classified according to valence and used in the other half of the subjects. Next, subjects were
black and white trials (the following pictures from informed that if they performed an error, a white
the IAPS were used: 2091, 2352.1, 2550, 5831, 8496 ‘x’ would appear below the picture until the cor-
for the positive stimuli and 1301, 2053, 2375.1, rect response was performed. The subjects were
9320 for the negative stimuli), and in the color requested to respond as accurately, but also as
trials (the following pictures from the IAPS were quickly, as possible. Finally, subjects were infor-
used: 1710, 4626, 4653, 4599, as well as a roller med that they were to perform three practice
coaster picture (valence = 7.05 ± 1.53, arousal = blocks of 20 trials (one black and white block, two

Yonsei Med J Vol. 47, No. 3, 2006

 Suk Kyoon An, et al.

Table 1. Emotional Pictures and Tasks

Color trials (Extrinsic valence)

Intrinsic stimulus valence Black and white trials
Positive* Negative*
Green-filtered Blue-filtered
Positive Positive pictures
(Congruent trials) (Incongruent trials)
Green-filtered Blue-filtered
Negative Negative pictures
(Incongruent trials) (Congruent trials)
*Half of the subjects were instructed to press the positive key for the green-filtered pictures, and to press the negative key for the
blue-filtered pictures, while for the other half of the subjects this color assignment was inversed.

color blocks) and four test blocks of 30 trials (10 response when the correct response was the
black and white trials, 20 color trials). All experi- negative response key, and the incongruent trial
mental procedures were performed by a psy- of negative pictures was the response when the
chiatrist (KJI) or a psychologist (JJH), who sat correct response was the positive response key.
about one meter away from the subject. A repeated-measures analysis of variance
To verify the appropriateness of all emotional (ANOVA) was carried out on error rates and log-
stimuli after the EAST session, all subjects were transformed reaction times, with intrinsic valence
asked to rate the valence and arousal of all stimuli (positive and negative) and extrinsic valence
(five positive and five negative black and white (positive and negative; assigned color response) of
pictures, and five positive and five negative color stimuli being the repeated measures and
color-filtered pictures) using the SAM.39 groups (normal controls and schizophrenia
patients) as the between-group factors. The
Data reduction and statistical analysis Wilcoxon signed rank test was performed on the
percentage of error, and a paired t-test was
We analyzed the results of the test trials in carried out on the log-transformed reaction time
which the color-filtered stimuli were presented within- group differences of congruent trials and
and only took into account the accuracy of the incongruent trials for each positive and negative
initial response and the reaction time. The color- filtered picture, in each group as a post-hoc
reaction time was recorded as the elapsed time analysis. We also repeated these analyses on the
from the stimulus onset until the correct response error rates and reaction times using three groups
was given, which is a common measure strategy (normal controls, paranoid schizophrenics and
of the IAT.22 According to the handling procedure non-paranoid schizophrenics). The effect size of
of Greenwald et al.,22 latencies below 300 ms or each positive and negative trial in each group was
above 3,000 ms were rounded to 300 ms and 3,000 calculated by taking the difference between the
ms, respectively, and the reaction time was then mean error rates on congruent and incongruent
log-transformed. Next, the percentage of errors trials dividing them by the pooled standard
and the mean log-transformed reaction times were deviation of each positive and negative color trial,
calculated separately according to the intrinsic respectively.
valence and the extrinsic valence (assigned To verify the appropriateness of the emotional
valence of color) of the color stimuli. The con- stimuli, paired t-tests on the ratings of valence
gruent trial of positive pictures was the response and arousal of all color-filtered pictures within
when the correct response was the positive each group (controls, schizophrenics as a whole,
response key. The congruent trial of negative paranoid schizophrenics, and non-paranoid schiz-
pictures was the response when the correct ophrenics) were carried out. Independent t-tests
response was the negative response key. The between groups (normal controls and schizo-
incongruent trial of positive pictures was the phrenics) and a one-way ANOVA between groups

Yonsei Med J Vol. 47, No. 3, 2006

 Impairment in Negative Emotion Evaluation in Paranoid Schizophrenia

(normal controls, paranoid schizophrenics, and level between the normal controls, paranoid schi-
non-paranoid schizophrenics) were also perfor- zophrenics, and non-paranoid schizophrenics.
med. The demographic and clinical profiles were
compared using a one-way ANOVA, an indepen- Verification of the emotional stimuli
dent t-test, and a 2 test. The criterion for signi-
ficance was set to p < 0.05, and to indicate a trend The paired t-test showed that the valence
p had to be less than 0.10. ratings using the SAM of positive and negative
pictures used in the color-filtered trial were signi-
ficantly different in both the normal controls (6.3
RESULTS ± 2.3 for the positive pictures vs. 2.6 ± 1.5 for the
negative pictures, p < 0.001), and in patients with
Demographic and clinical profiles schizophrenia (for all patients, 6.3 ± 2.4 for the
positive pictures vs. 2.6 ± 1.7 for the negative
Demographic and clinical profiles are sum- pictures, p < 0.001; paranoid patients, 6.2 ± 2.4 for
marized in Table 2. The subjects in the control and the positive pictures vs. 2.9 ± 2.0 for the negative
patient groups were well-matched in age, sex and pictures, p < 0.001; non-paranoid patients, 6.4 ± 2.4
education level. All schizophrenic patients were for the positive pictures vs. 2.3 ± 1.5 for the
receiving antipsychotic medication (atypical = 12, negative pictures, p < 0.001). There was no signi-
typical = 5, and both typical and atypical = 3). ficant difference between the arousal ratings of the
Post-hoc groups based on specific type (para- positive and negative pictures in either the normal
noid and non-paranoid) were also broadly similar controls (6.3 ± 1.9 for the positive pictures vs. 6.2
in terms of demographic and clinical charac- ± 1.9 for the negative pictures, p = 0.969) or in the
teristics except for the positive symptoms score on schizophrenia group (for all patients, 6.1 ± 2.3 for
the PANSS37 (paranoid 13.9 ± 6.7, non-paranoid the positive pictures vs. 5.9 ± 2.4 for the negative
8.0 ± 1.6; t = 2.6, df = 8.7, p = 0.030). There was no pictures, p = 0.574; paranoid patients, 5.7 ± 2.5 for
significant difference in age, sex, or education the positive pictures vs. 5.4 ± 2.5 for the negative

Table 2. Sociodemographic and Clinical Profiles of Normal Controls and Patients with Schizophrenia

Normal (n = 22) Schizophrenia (n = 20) p value

Age (yr) 31.3 ± 6.2 30.3 ± 7.9 0.642
Sex (M/F) 10/12 10/10 0.768
Education (yr) 14.8 ± 2.5 14.1 ± 1.7 0.312
Onset age (yr) - 25.2 ± 6.2
Illness duration (Months) - 63.2 ± 49.4
Subtype (paranoid/undifferentiated/residual) - 9/3/8
PANSS* -
Positive - 10.7 ± 5.4
Negative - 12.1 ± 3.6
General psychopathology - 21.8 ± 5.2
Chlopromazine equivalent dose (mg/day) - 518.8 ± 330.4
Data are shown as mean ± standard deviation.
The p values are calculated by t-tests or χ2 tests.
*PANSS, positive and negative syndrome scale.
Doses are calculated according to the guidelines of the American Psychiatric Association.

Yonsei Med J Vol. 47, No. 3, 2006

 Suk Kyoon An, et al.

pictures, p = 0.599; non-paranoid patients, 6.3 ± was a trend-level interaction effect between ex-
2.1 for the positive pictures vs. 6.2 ± 2.2 for the trinsic valence and group [F(1,40) = 3.79, p = 0.059].
negative pictures, p = 0.780). There were no sig- In addition, there were no other significant or
nificant differences of valence and arousal of posi- trend-level interaction effects. The Wilcoxon
tive and negative pictures between normal con- signed rank tests were conducted to examine the
trols, patients with paranoid schizophrenia, and within-group differences of congruent trials and
patients with non-paranoid schizophrenia. These incongruent trials for each positive and negative
findings mean that controls and schizophrenia picture in each group (Table 3). These Wilcoxon
patients as a whole, as well as normal controls, signed rank tests revealed that normal controls
paranoid schizophrenics, and non-paranoid schiz- committed more errors on the incongruent trials
ophrenics recognized the positive and negative than the congruent trials for positive stimuli (8.18
pictures as positive and negative stimuli with ± 8.39 v.s. 2.05 ± 3.98, Z = 2.62, p = 0.009) and for
comparable arousal levels. negative stimuli (4.32 ± 5.83 v.s. 1.36 ± 2.75, Z =
2.36, p = 0.018). The effect sizes of positive and
Emotional task performance negative stimuli were 0.93 and 0.65, respectively,
for the normal controls (Fig. 1). All Patients with
A repeated-measures ANOVA was carried out schizophrenia did not make more errors on the
on error rates and reaction times, with intrinsic incongruent trials (17.75 ± 22.80) than in the con-
valence (positive and negative) and extrinsic va- gruent trials (13.50 ± 18.29) for negative stimuli (Z
lence (positive and negative; assigned color re- = 0.54, p = 0.591), while they made more signi-
sponse) of color stimuli as the repeated measures ficant errors on the incongruent trials (29.25 ±
and groups (normal controls and schizophrenia 26.12) than in the congruent trials (8.75 ± 9.58) for
patients) as the between-group factors. Results for positive stimuli (Z = 3.42, p = 0.001). The effect
error rates indicated a significant main effect of sizes of positive and negative stimuli were 1.04
intrinsic valence [F(1,40) = 5.45, p = 0.025], extrin- and 0.21, respectively, in patients with schizo-
sic valence [F(1,40) = 8.39, p = 0.006], groups (nor- phrenia (Fig. 1).
mal controls and schizophrenics) [F(1,40) = 14.63, Repeated-measure ANOVA results for reaction
p < 0.001], and significant interaction effects times indicated a significant main effect of in-
between intrinsic valence and extrinsic valence trinsic valence [F(1,40) = 21.50, p < 0.001], extrinsic
[F(1,40) = 19.74, p < 0.001], and more importantly, valence [F(1,40) = 17.0, p < 0.001], and group (nor-
intrinsic valence, extrinsic valence, and group in- mal controls and schizophrenia patients) [F(1,40)
teraction effects [F(1,40) = 4.23, p = 0.046]. There = 20.21, p < 0.001], and also suggested significant

Table 3. EAST Performances of Normal Controls and Patients with Schizophrenia

Normal (n = 22) Schizophrenia (n = 20)

Intrinsic valence Extrinsic valence Extrinsic valence
Positive Negative t/Z p value Positive Negative t/Z p value
Positive
Error rate (%) 2.05 ± 3.98 8.18 ± 8.39 2.62 0.009 8.75 ± 9.58 29.25 ± 26.12 3.42 0.001
Reaction times (ms) 839.4 ± 188.6 964.5 ± 298.3 3.21 0.004 1,153.5 ± 363.7 1,413.5 ± 387.5 4.66 < 0.001
Negative
Error rate (%) 4.32 ± 5.83 1.36 ± 2.75 2.36 0.018 17.75 ± 22.80 13.50 ± 18.29 0.54 0.591
Reaction times (ms) 819.0 ± 151.9 866.7 ± 250.3 -1.47 0.157 1,214.1 ± 499.7 1,168.9 ± 289.9 0.13 0.899
EAST, Extrinsic Affective Simon Task.
Data are shown as mean ± standard deviation.
ms, milliseconds.

Yonsei Med J Vol. 47, No. 3, 2006

 Impairment in Negative Emotion Evaluation in Paranoid Schizophrenia

Fig. 1. Effect size of positive and negative evaluations in normal controls, schizophrenia, paranoid schizophrenia, and
non-paranoid schizophrenia. The effect sizes of the positive and negative pictures in each group were calculated using the
difference between the mean error rates on congruent and incongruent trials dividing them by the pooled standard
deviation of each emotional stimulus.

interaction effects of intrinsic valence and extrinsic ferences for each positive and negative picture for
valence [F(1,40) = 13.47, p = 0.001]. There was also each group. These Wilcoxon signed rank tests
a trend-level interaction effect between intrinsic revealed that non-paranoid schizophrenics made
valence, extrinsic valence and group [F(1,40) = more errors on the incongruent trials than in the
3.28, p = 0.078]. There were no other significant or congruent trials for positive stimuli (27.27 ± 17.80
trend-level interaction effects. The paired t-tests vs. 7.78 ± 6.84, Z = 2.53, p = 0.012) and for nega-
were conducted to examine the within-group dif- tive stimuli (21.36 ± 24.50 vs. 8.64 ± 9.24, Z = 1.79,
ferences of congruent trials and incongruent trials p = 0.074). The effect sizes of positive and negative
for each positive and negative picture in each stimuli were 1.45 and 0.69, respectively, in
group (Table 3). These paired t-tests revealed that patients with non-paranoid schizophrenia (Fig. 1).
there was a greater delay on the incongruent trials However, patients with paranoid schizophrenia
than in the congruent trials for positive stimuli, did not make more errors on the incongruent
which was not seen for the negative stimuli in trials (13.33 ± 21.07) than in the congruent trials
normal controls and patients with schizophrenia. (19.44 ± 24.81) for negative stimuli (Z = -1.63, p =
When these analyses were repeated on the error 0.104), while they made more significant errors on
rates in the three groups (normal controls, para- the incongruent trials (31.67 ± 34.73) than in the
noid schizophrenics, and non-paranoid schizo- congruent trials (10.0 ± 12.5) for positive stimuli
phrenics), a significant main effect of intrinsic (Z = 2.25, p = 0.024). The effect sizes of positive
valence [F(1,39) = 5.50, p = 0.024], extrinsic valence and negative stimuli were 0.83 and -0.27, respec-
[F(1,39) = 13.47, p = 0.001], and group [F(2,39) = tively, in patients with paranoid schizophrenia
7.28, p = 0.002] was seen, and significant interac- (Fig. 1). Repeated-measure ANOVA results for
tion effects of intrinsic valence and extrinsic reaction times indicated a significant main effect
valence [F(1,39) = 22.27, p < 0.001], extrinsic valence of intrinsic valence [F(1,39) = 20.02, p < 0.001],
and group interaction effect [F(2,39) = 4.63, p = extrinsic valence [F(1,39) = 16.95, p < 0.001], and
0.016], and more importantly an interaction effect group (normal controls, paranoid patients and
between intrinsic valence, extrinsic valence and non-paranoid patients) [F(2,39) = 11.48, p < 0.001],
group [F(2,39) = 3.36, p = 0.045] was also observed. and also significant interaction effects of intrinsic
There was no other significant or trend-level valence and extrinsic valence [F(1,39) = 15.14, p <
interaction effect. The Wilcoxon signed rank tests 0.001]. There were no other significant or trend-
were conducted to examine the within-group dif- level interaction effects.

Yonsei Med J Vol. 47, No. 3, 2006

 Suk Kyoon An, et al.

DISCUSSION each positive and negative emotional category. In

addition, the effect sizes of positive and negative
This study revealed that normal controls made emotions in the present study were so robust that
more errors in trials of the positive pictures when they allowed for the comparison of the perfor-
the correct response was the negative response mances of normal controls and patients with
key. The reverse was true for trials of the negative schizophrenia.
pictures. Patients with schizophrenia, especially Patients with schizophrenia made significantly
paranoid schizophrenia, committed more errors more errors on the incongruent trials than con-
on trials of the positive pictures when the correct gruent trials for positive pictures, but not for
response key was the negative response key. negative pictures. This finding suggests that schi-
However, the reverse was not true in trials of the zophrenics may suffer from a greater impairment
negative pictures. These findings suggest that in their ability to evaluate negative emotions. This
patients with schizophrenia, especially paranoid is also compatible with previous studies in that
schizophrenia, might suffer from a greater impair- there was a greater impairment of the evaluation
ment in the ability to evaluate negative emotions. of negative emotions than positive emotions in
4-8
The performances of normal controls showed a schizophrenics.
more significant error rate on the incongruent When splitting schizophrenics into two groups
trials than the congruent trials for both positive (paranoid and non-paranoid), there was a rela-
and negative stimuli. In addition, the effect sizes tionship between the diagnostic subtype and the
in both positive and negative emotions were ability to evaluate emotions. Paranoid schizo-
robust according to Cohen's criteria.40 These phrenic patients made more significant errors for
findings were grossly compatible to the previous positive pictures but not for negative pictures,
EAST report using emotional words,31 which while non-paranoid patients had significantly
showed that the EAST could measure a separated more errors for positive pictures and a trend-level
estimate of emotional evaluation within each tar- of more errors for negative pictures in the incon-
get category and retain the robust effect size. gruent trials than in the congruent trials. These
However, when taking reaction time into ac- findings mean that paranoid schizophrenic
count, the present study found that there was a patients show more profound impairments in the
significantly longer delay in the incongruent trials evaluation of negative stimuli, while non-para-
than the congruent trials for positive pictures, but noid patients do not show emotional processing
not for the negative pictures in both subject impairment of the positive or negative stimuli.
groups, while De Houwer's31 findings were that The paranoid schizophrenics were reported to
there was a significant delay in the analysis of show negative emotion recognition impairment
reaction times for both positive and negative while residual patients displayed no impairment
trials. Since the stimuli in De Houwer's study31 in negative or positive emotion recognition.18 To
were not pictures but words, a direct comparison the extent that our non-paranoid patients were
is not possible. One of the probable reasons for residual schizophrenics (n = 8) and undifferenti-
the discrepancy in the reaction time analysis is ated schizophrenics (n = 3) who have nearly re-
that subjects may base their responses more on mitted positive symptoms, our main findings are
reaction time than on accuracy. This might be an compatible with previous findings.18
inherent characteristic of the procedural structures The main findings of greater impairment in
of the EAST with pictures used in the present negative emotion processing in paranoid schizo-
study, although all subjects were instructed to phrenics seem to be incompatible to those of the
41,42
press the key as accurately, but also as quickly, as previous emotional Stroop tests. However,
possible. Although the precise reason for this these previous Stroop studies to measure pre-
discrepancy is not yet clear, it seems reasonable attentive emotional processing found that delu-
to suppose that in taking the error rates of the sional schizophrenic patients demonstrated a
response into account, EAST using emotional selective delay in response time to the threatening
pictures provides separate evaluative estimates of words,41,42 which means that there is a heightened

Yonsei Med J Vol. 47, No. 3, 2006

 Impairment in Negative Emotion Evaluation in Paranoid Schizophrenia

pre-attentive bias to threatening, highly-arousing Imaging (fMRI) study47 using fearful faces as
negative stimuli. One of the possible explanations threatening stimuli suggested that there is a func-
of this apparent inconsistency is the difference in tional disconnection in the autonomic and central
stimuli. In the present study, most of the negative systems for the processing of threatening stimuli
stimuli were complex pictures depicting social in paranoid schizophrenia.
events, while in the previous studies41,42 the There were some limitations in the present
stimuli were simple emotional words depicting study. Firstly, all patients were on medication.
threatening ideas. Another more plausible ex- However, in previous studies on this issue, no
planation is the difference in construct of each evidence was found of any adverse effects caused
task. The emotional Stroop test measures the pre- by antipsychotics on the emotional task perform-
attentive bias to specific emotional contents, while ances in schizophrenics.1,15,48 Thus, the medication
the EAST measures the strength of an association effect seemed to be small, although the possibility
within each emotional category. of there being some such effect could not be ruled
The main finding of a negative emotion-specific out. Another more important limitation is that the
deficit in paranoid schizophrenics in the present color-filtered negative stimuli might simply be
study is compatible with the results of previous more difficult than the positive stimuli for
priming studies and is quite informative about the patients with schizophrenia,3 instead of there
issue of emotional impairment in schizophrenia. being any specific relationship to negative emo-
In previous emotional priming studies,43,44 there tion. However, in this EAST task the evaluation of
was a significant negative judgment shift of the the intrinsic emotion of color-filtered stimuli was
target stimuli, which suggested that patients with done implicitly, not explicitly. There has been no
schizophrenia could not control or suppress the accumulation of knowledge about whether it is
negative emotional information at the automatic more difficult to implicitly evaluate negative
processing stage. However, whether positive emo- stimuli than positive stimuli. In addition, it was
tion processing is impaired or not could not be found in the present study that between normal
determined because there was a strong tendency controls and patients with schizophrenia, and
to judge the stimuli to be positive even when between normal controls, paranoid patients, and
there had been priming with neutral stimuli in non-paranoid patients according to emotional
these previous studies. Based on the assumption valence, there was no significant difference of
that each estimate of the strength of association reaction time, which may reflect the difficulty
within each positive and negative emotional net- required to perform the task. Thus, the main
work could be indexed by the EAST methods,30,31 finding of the present study may not be well-ex-
our findings suggest that patients with paranoid plained by the greater difficulty in the recognition
schizophrenia might suffer from an impairment of negative stimuli in schizophrenia, although the
within the negative emotional association net- possibility could not be ruled out.
work, and were less likely to have an impairment In conclusion, our findings support the hypo-
within the positive emotional category. This loose thesis that patients with paranoid schizophrenia
association within the negative emotional network might suffer from a greater impairment in the
could result in the inability of patients with schiz- ability to evaluate negative emotions, and this
ophrenia to regulate the spreading of activation to deficit would imply that there may be a loosening
the negative emotional prime, which was mani- of association within the negative emotional
fested by a negative judgment shift to prime in network in patients with paranoid schizophrenia.
previous studies.43,44
This loose association within the negative emo-
tional category in paranoid schizophrenia might ACKNOWLEDGEMENTS
also have resulted from the disrupted connectivity
between nodes in the prefrontal-thalamic and We thank Hae-Jeong Park, Ph.D., for giving us
temporolimbic or cerebellar circuit.45,46 Further- permission to use his ViStim software to perform
more, a recent functional Magnetic Resonance the EAST.

Yonsei Med J Vol. 47, No. 3, 2006

 Suk Kyoon An, et al.

REFERENCES 17. Mandal MK, Rai A. Responses to facial emotion and

psychopathology. Psychiatry Res 1987;20:317-23.
1. Mueser KT, Penn DL, Blanchard JJ, Bellack AS. Affect 18. Weniger G, Lange C, Rüther E, Irle E. Differential
recognition in schizophrenia: a synthesis of findings impairments of facial affect recognition in schizo-
across three studies. Psychiatry 1997;60:301-8. phrenia subtypes and major depression. Psychiatry Res
2. Mandal MK, Pandey R, Prasad AB. Facial expressions 2004;128:135-46.
of emotions and schizophrenia: a review. Schizophr 19. Parrot WG, Hertel P. Research methods in cognition
Bull 1998;24:399-412. and emotion. In: Dalgeish T, Power MJ, editors. Hand-
3. Edwards J, Jackson HJ, Pattison PE. Emotion recogni- book of Cognition and Emotion. Chichester: Wiley;
tion via facial expression and affective prosody in 1999. p.61-82.
schizophrenia: a methodological review. Clin Psychol 20. Greenwald AG, Farnham SD. Using the implicit asso-
Rev 2002;22:789-832. ciation test to measure self-esteem and self-concept. J
4. Dougherty FE, Bartlett ES, Izard CE. Responses of Pers Soc Psychol 2000;79:1022-38.
schizophrenics to expressions of the fundamental 21. Greenwald AG, Banaji MR. Implicti social cognition:
emotions. J Clin Psychol 1974;30:243-6. attitudes, self-esteem, and stereotypes. Psychol Rev
5. Musekari LH, Bates ME. Judgment of emotion among 1995;102:4-27.
chronic schizophrenics. J Clin Psychol 1977;33:662-6. 22. Greenwald AG, McGhee DE, Schwartz JLK. Measuring
6. Heimberg C, Gur RE, Erwin RJ, Shtasel DL, Gur RC. individual differences in implicit cognition: the Implicit
Facial emotion discrimination: III. Behavioralfindings in Association test. J Pers Soc Psychol 1998;74:1464-80.
schizophrenia. Psychiatry Res 1992;42:253-65. 23. De Houwer J. A structural and process analysis of the
7. Bell M, Bryson G, Lysaker P. Positive and negative implicit association test. J Exp Soc Psychol 2001;37:443-
affect recognition in schizophrenia: a comparison with 51.
substance abuse and normal control subjects. 24. Fazio RH, Olson MA. Implicit measures in social cogni-
Psychiatry Res 1997;73:73-82. tion research: their meaning and use. Ann Rev Psychol
8. An SK, Lee SJ, Lee CH, Cho HS, Lee PG, Lee CI, et al. 2003;54:297-327.
Reduced P3 amplitudes by negative facial emotional 25. De Jong P, Pasman W, Kindt M, van den Hout MA.
photographs in schizophrenia. Schizophr Res 2003;64: A reaction time paradigm to assess (implicit) com-
125-35. plaint-specific dysfunctional beliefs. Behav Res Ther
9. Kerr S, Neale JM. Emotion perception in schizophrenia: 2001;39:101-13.
specific deficit or further evidence of generalized poor 26. De Jong PJ. Implicit self-esteem and social anxiety:
performance. J Abnorm Psychol 1993;102:312-8. differential self-positivity effects in high and low anx-
10. Kring AM, Neale JM. Do schizophrenic patients show ious individuals. Behav Res Ther 2002;40:501-8.
a disjunctive relationship among expressive, experien- 27. Germar MC, Segal ZV, Sagrati S, Kennedy SJ. Mood-
tial, and psychophysiological components of emotion. induced changes on the Implicit Association Test in
J Abnorm Psychol 1996;105:249-57. recovered depressed patients. J Abnorm Psychol 2001;
11. Salem JE, Kring AM, Kerr S. More evidence for 110:282-9.
generalized poor performance in facial emotion per- 28. Wiers RW, van Woerden N, Smulders FT, De Jong PJ.
ception in schizophrenia. J Abnorm Psychol 1996;105: Implicit and explicit alcohol-related cognition in heavy
480-3. and light drinker. J Abnorm Psychol 2002;111:648-58.
12. Walker EF, McGuire M, Bettes B. Recognition and 29. Lee E, An SK, Jung S-Y, Park JY, Lee SJ, Namkoong K.
identification of facial stimuli by schizophrenics and Evaluation of implicit emotion of alcohol-related cues
patients with affective disorders. Br J Clin Psychol 1984; in alcohol dependence. J Korean Neuropsychiatr Assoc
23:37-44. 2005;44:334-41.
13. Walker EF, Marwit SJ, Emory E. A cross-sectional study 30. De Houwer J. The implicit association test as a tool for
of emotion recognition in schizophrenics. J Abnorm studying dysfunctional associations in psychopatho-
Psychol 1980;89:428-36. logy: strength and limitations. J Behav Ther Exp Psy-
14. Mandal MK, Jain A, Haque-Nizamie S, Weiss U, chiatr 2002;33:115-33.
Schneider F. Generality and specificity of emotion- 31. De Houwer J. The extrinsic affective simon task. Exp
recognition deficit in schizophrenic patients with posi- Psychol 2003;50:77-85.
tive and negative symptoms. Psychiatry Res 1999;87:39- 32. American Psychiatric Association. Diagnostic and
46. Statistical Manual of Mental Disorders, 4th ed.
15. Lewis S, Garber D. Treatment and diagnostic subtype Washington, DC: American Psychiatric Press; 1994.
in facial affect recognition in schizophrenia. J Psy- 33. Annett M. A classification of hand preference by asso-
chiatry Res 1995;29:5-11. ciation analysis. Br J Psychol 1970;61:303-21.
16. Kline JS, Smith JE, Ellis HC. Paranoid and nonparanoid 34. Briggs GG, Nebes RD. Patterns of hand preference in
schizophrenic processing of facially displayed affect. J a student population. Cortex 1975;11:230-8.
Psychiatry Res 1992;26:169-82. 35. First MB, Spitzer RL, Gibbon M, Williams JBW. Struc-
tured Clinical Interview for DSM-IV Axis I Disorders:

Yonsei Med J Vol. 47, No. 3, 2006

 Impairment in Negative Emotion Evaluation in Paranoid Schizophrenia

Patient Edition (SCID-I/P). Version 2. New York: New 333-42.

York State Psychiatric Institute Biometric Research; 43. Lee SJ, An SK. (Korean language) The characteristics of
1996. emotional regulation of schizophrenics. Korean J Clin
36. First MB, Gibbon M, Spitzer RL, Williams JBW. Struc- Psychol 2000;19:269-79.
tured Clinical Interview for DSM-IV Axis I Disorders: 44. Höschel K, Irle E. Emotional priming of facial affect
Non-Patients Edition (SCID-I/PS). Version 2. New identification in schizophrenia. Schizophr Bull 2001;27:
York: New York State Psychiatric Institute Biometric 317-27.
Research; 1996. 45. Andreasen NC, Paradiso S, O'Leary DS. Cognitive dys-
37. Kay SR, Fiszhein A, Opler LA. The positive and nega- metria as an integrative theory of schizophrenia: a
tive syndrome scale for schizophrenia. Schizophr Bull dysfunction in cortico-subcortical-cerebellar circuitry.
1987;13:261-72. Schizophr Bull 1998;24:203-18.
38. Lang PJ, Bradley MM, Cuthbert BN. International affec- 46. Weinberger DR, Berman KF, Suddath R, Torrey EF.
tive picture system (IAPS): Instruction manual and Evidence of dysfunction of a prefrontal-limbic network
affective ratings. Technical Report A-5, The Center for in schizophrenia: a magnetic resonance imaging and
Research in Psychophysiology, University of Florida, regional cerebral blood flow study of discordant
2001. monozygotic twins. Am J Psychiatry 1999;46:908-20.
39. Bradley MM, Lang PJ. Measuring emotion: The self- 47. Williams LM, Das P, Harris AWF, Liddle BB, Brammer
assessment manikin and the semantic differential. J MJ, Olivieri G, et al. Dysregulation of arousal and
Behav Ther Exp Psychiatr 1994;25:49-59. amygdala-prefrontal systems in paranoid schizo-
40. Cohen J. Statistical power analysis for the behavioral phrenia. Am J Psychiatry 2004;161:480-9.
sciences. Hillsdale, NJ: Lawrence Erlbaum; 1988. 48. Schneider F, Gur RC, Gur RE, Shtasel DL. Emotional
41. Bentall RP, Kaney S. Content specific information processing in schizophrenia: neurobehavioral probes in
processing and persecutoiry delusions: an investigation relation to psychopathology. Schizophr Res 1995;17:67-
using the emotional stroop test. Br J Med Pschol 1989; 75.
62:355-64. 49. American Psychiatric Association. Practice guideline
42. Green MJ, Phillips ML. Social threat perception and the for the treatment of patients with schizophrenia. Am J
evolution of paranoia. Neurosci Biobehav Rev 2004;28: Psychiatry 1997;154:1-63.

Yonsei Med J Vol. 47, No. 3, 2006