Directory UMM :Data Elmu:jurnal:B:Biological Psichatry:Vol48.Issue2.2000:

Emotion Recognition Deficit in Schizophrenia:
Association with Symptomatology and Cognition
Christian G. Kohler, Warren Bilker, Michael Hagendoorn, Raquel E. Gur, and
Ruben C. Gur
Background: Previous investigations have found impaired
recognition of facial affect in schizophrenia. Controversy
exists as to whether this impairment represents a specific
emotion recognition deficit when compared with other face
recognition control tasks. Regardless of whether the emotion
processing deficit is differential, it may uniquely influence
other manifestations of schizophrenia. We compared patients
and healthy control subjects on computerized tasks of emotion and age recognition. Performances on emotion and age
recognition tasks were correlated with cognitive functioning
and with symptomatology.
Methods: Thirty-five patients with schizophrenia and 45
healthy people underwent computerized testing for emotion and age recognition. Participants were assessed
neuropsychologically, and patients were rated for positive
and negative symptoms.
Results: The patients with schizophrenia performed worse
than control subjects on emotion and age recognition
without differential deficit. In both groups, we found

higher error rates for identification of emotion in female
faces and for identification of sad versus happy faces. In
schizophrenic patients, emotion but not age recognition
correlated with severity of negative and positive symptoms. In healthy control subjects, neither task correlated
with cognitive functions. In schizophrenic patients, emotion but not age recognition correlated with attention,
verbal and spatial memory, and language abilities.
Conclusions: This study did not reveal a specific deficit for
emotion recognition in schizophrenia; however, our findings
lend support to the concept that emotion recognition is
uniquely associated in schizophrenia with core symptomatology and cognitive domains. Biol Psychiatry 2000;48:
127–136 © 2000 Society of Biological Psychiatry
Key Words: Emotion recognition, schizophrenia, gender,
cognition

From the Schizophrenia Research Center, Neuropsychiatry Section, Department of
Psychiatry (CGK, MH, REG, RCG) and the Department of Biostatistics and
Epidemiology (WB), University of Pennsylvania School of Medicine,
Philadelphia.
Address reprint requests to Christian G. Kohler, M.D., Mental Health Clinical
Research Center, 3400 Spruce Street, Department of Psychiatry, 10 Gates Bldg,

Philadelphia PA 19104.
Received June 14, 1999; revised December 7, 1999; accepted February 15, 2000.

© 2000 Society of Biological Psychiatry

Introduction

P

rocessing of affect is an important component of social
interaction (Anthony 1978; Charlesworth 1982; Izard
1971) and is supported by distributed neural systems
(Damasio 1998; Heilman and Gilmore 1998; Lang et al 1998;
Leventhal and Tomarken 1986; Mesulam 1998). Impaired
processing of emotion has been noted in a range of disorders,
including focal brain lesions (Adolphs et al 1994, 1996;
Blonder et al 1991; Borod et al 1993; Bowers et al 1991;
Sackeim et al 1982), Huntington’s Disease (Jacobs et al
1995), Parkinson’s Disease (Adolphs et al 1998), and neuropsychiatric disorders (Feinberg et al 1986; Gur et al 1992;
Mikhailova et al 1996; Walker et al 1984). The study of

emotion processing domains, including recognition, experience and expression, has benefitted from the use of standardized facial stimuli (Ekman et al 1972).
Abnormalities in emotional expression and experience
have been reported since the earliest descriptions of
schizophrenia (Bleuler 1911). More recently, deficits in
facial-affect recognition also have been noted; however,
an issue has been raised as to whether a differential deficit
can be demonstrated against the more general impairment
in facial processing (Kerr and Neale 1993; Salem et al
1996). Some studies support a differential deficit. Cutting
(1981) reported that acutely ill schizophrenic patients
performed poorly relative to patients in remission and
patients with psychotic depression on an affect discrimination task, which compared “friendliness” and “meanness,” but not on color discrimination and age perception
tasks. Walker et al (1984) studied patients with chronic
schizophrenia, affective disorder, and healthy control subjects. Three emotional tasks of 16 faces each were used to
probe emotion discrimination, emotion labeling, and multiple choice assignment of six different emotional valences
(happiness, anger, fear, sadness, surprise, and shame) but
not neutral faces. Although patients with schizophrenia
performed as well as control subjects on the Benton Facial
Recognition Test (Benton and Van Allen 1968), their
performance on recognition of affect was significantly

below the level of both control subjects and patients with
affective disorders. Heimberg et al (1992) likewise re0006-3223/00/$20.00
PII S0006-3223(00)00847-7

128

BIOL PSYCHIATRY
2000;48:127–136

ported differential impairment in discrimination of happy
and sad facial expression relative to age discrimination
using a set of neutral, happy, and sad faces. Patients
performed worse on emotion than on age recognition,
whereas healthy subjects performed comparably on these
tasks. Using the same testing paradigm (which included
254 faces, enabling comparison of happy versus neutral
and sad versus neutral in a blocked design), Gur et al
(1992) found that the deficit associated with schizophrenia
was more marked compared with the performance of
patients with depression. The large number of facial

stimuli used, which permitted blocking sad and happy
discrimination, enabled separate determination of false
positive and negative rates for each emotion; however, the
interspersed stimulus presentation of the other investigations is more naturalistic. Borod et al (1993) administered
facial emotion identification and discrimination tasks of
six different emotions and neutral, ranging from 21 to 32
facial stimuli to patients with schizophrenia, patients with
right-brain damage, and healthy control subjects and found
that both patient groups were impaired. When performances on the Benton Facial Recognition Test and a
visuospatial task were used as covariates, this effect
disappeared in both groups for the emotion discrimination,
but not for the more difficult identification task.
In contrast, other studies indicate a generalized facial
processing deficit in schizophrenia. Novic et al (1984)
reported that patients performed more poorly than control
subjects on a task requiring participants to match pictures
with identical emotions, but when they controlled for
performance on the Benton Facial Recognition Test, the
difference became insignificant. The validity of the results
is limited by the small number of presented stimuli (six out

of 16) used in the analysis. Likewise, Feinberg et al (1986)
found that schizophrenic patients were impaired on emotion matching and labeling tasks, but not on facial recognition tasks relative to depressed patients. Patients with
schizophrenia were also impaired on emotion and facial
recognition tasks compared with healthy control subjects.
Archer et al (1992) reported that patients with schizophrenia were as impaired on facial expression recognition of
six different emotions as they were on two facial recognition tasks, relative to patients with major depression and
normal control subjects. For patient and control groups,
performances on the tasks ranged between 89 –98%, indicating that the experimental tasks may have been too easy
to detect differences between the tasks. Schneider et al
(1995) described similar results in that patients with
schizophrenia were impaired relative to healthy control
subjects on identification of happy, sad, and neutral
expressions, as well as assessment of age; however, this
study did not evaluate whether the deficit was generalized
or specific. Kerr and Neale (1993) utilized a differential

C.G. Kohler et al

deficit design based on the requirements put forth by
Chapman and Chapman (1978) as described below. This

study compared a group of institutionalized and unmedicated with patients with chronic schizophrenia with a
control group on tests of perception and discrimination of
different emotions (happiness, sadness, fear, anger, shame,
and surprise) and Benton’s Facial Recognition Test.
Schizophrenic patients performed worse on all tasks without a differential deficit, indicating that they display
impaired facial processing, rather than a specific emotion
recognition deficit. Performance data were not presented.
Using an abbreviated test design, Salem et al (1996)
replicated these findings in a group of 23 medicated, male,
chronic schizophrenia patients compared with a male
control group. Although the differences in performance
between patient and control groups reached significance,
the magnitude of differences between tasks may not have
been large enough to detect differential deficits for emotion recognition.
The lack of consensus on whether a generalized face
processing deficit can account for the affect recognition
deficit may have several reasons. Except for the last two
studies mentioned, all investigators utilized different emotion recognition paradigms. The use of different facial
control tasks may further have contributed to the disparity
of findings; however, many studies utilized the Test of

Facial Recognition as the control task. This test, in an
original 54-face (Benton and Van Allen 1968) and shortened 27-face (Benton et al 1983) version, measures the
ability to recognize faces without a memory component.
Performance relies on visuospatial and linguistic processing and is impaired in patients with right parietal lesions,
aphasia, and dementias (Lezak 1995). Most emotion recognition studies attempted to conform to the Chapman and
Chapman (1978) differential deficit design. To establish
differential deficit, the target and control tasks should be
equated for difficulty and true score variance in healthy
people; however, the methodologically important question
of generalized face processing versus specific affect recognition deficit should be considered within the context of
efforts to elucidate the pathophysiology of schizophrenia
and treatment of its behavioral manifestations. Impaired
emotion recognition has considerable impact on a person’s
abilities to communicate and comprehend nonverbal cues.
As such, impaired emotion recognition in the person with
schizophrenia may exert adverse effects on psychosocial
functioning independent of the presence and severity of
positive and negative symptoms and cognitive difficulties.
Although this issue was not explored in our present study,
we believe emotion recognition is relevant for interpersonal and social adjustment and has impact regardless of

its severity relative to cognitive deficits that serve as a
control task.

Emotion Recognition in Schizophrenia

Gender differences have been reported in the clinical
presentation and course of schizophrenia. Men tend to
have an earlier age of onset (Beratis et al 1994; Jayaswal
et al 1987; Vazquez-Barquero et al 1995), poorer premorbid adjustment (Shtasel et al 1992), lower level of daily
functioning, and worse negative symptomatology (Goldstein et al 1989; Salem and Kring 1998; Shtasel et al 1992)
than women. No previous study has addressed the issue of
gender differences in emotion perception in schizophrenia.
In healthy subjects, women performed better than men
(reviewed in Kring and Gordon 1998). Erwin et al (1992)
reported an interaction of gender of poser with the gender
of observer: women were more sensitive to male faces,
whereas men showed a lack of sensitivity to sad expressions in women.
In our study, patients completed computerized emotion
and age recognition tasks, along with ratings of symptomatology and neuropsychologic testing. The computerized
tasks consisted of male and female faces with randomized

neutral, sad, and happy affect. This was modified from our
previous blocked stimulus design to allow for a more naturalistic methodology and better comparability with other
studies. The control task used the same type of stimuli and
differed only in the requirement to rate age rather than affect.
The goals of the present study were to measure the severity
of emotion recognition deficits in schizophrenia compared
with age recognition performance. We also evaluated
whether schizophrenic patients exhibit gender-related differences in emotion recognition similar to healthy subjects.
Finally, we examined the association of emotion and age
recognition performance with clinical and neurocognitive
manifestations of schizophrenia.
We hypothesized the following: 1) Emotion and age
recognition would be impaired in schizophrenia compared
with healthy control subjects. 2) Male patients would
exhibit a greater deficit than female patients. This finding
would be consistent with worse symptomatology in men.
3) Impaired emotion, rather than age recognition, would
be associated with severity of positive and negative
symptoms. 4) Emotion recognition performance correlates
with different cognitive functions than age discrimination

performance. Support for hypotheses 3 and 4 would
indicate that the emotion recognition deficit in schizophrenia relates uniquely to core features of the disorder.

Methods and Materials
Subjects
The sample consisted of 45 healthy people (25 men, 20 women)
and 35 people with schizophrenia (20 men, 15 women). Recruitment of participants followed established procedures (Gur et al
1991; Shtasel et al 1991). Patients underwent diagnostic examination using the Structured Clinical Interview for DSM-IV

BIOL PSYCHIATRY
2000;48:127–136

129

(SCID-P; First et al 1995a) and met criteria for schizophrenia or
schizophreniform disorder. Patients with the initial diagnosis of
schizophreniform disorder (n 5 7) were followed longitudinally
and met criteria for schizophrenia. Patients with the diagnosis of
schizoaffective disorder, any concomitant Axis I disorder, or
history of any other disorder affecting brain function were
excluded. Testing of age and emotion recognition and symptom
and cognitive assessments were performed in patients who were
symptomatic but clinically stable at the time of evaluation. At the
time of emotion recognition testing, 31 patients were treated with
standard neuroleptic medication, whereas four patients were
medication naive (diagnosis schizophrenia:schizophreniform53:
1). Patients were treated with typical neuroleptics (n 5 22),
atypical neuroleptics (n 5 4) or both (n 5 5) at an average
dose of 454 chlorpromazine equivalents per day. Medication
exposure ranged from 1 week to 14.8 years. Healthy control
subjects underwent comprehensive assessment including Structured Clinical Interview for DSM-IV nonpatient version (SCIDNP; First et al 1995b), physical examination, and laboratory tests
to exclude concurrent psychiatric or other medical illness. Those
with a family history of schizophrenia or affective illness were
excluded. Sociodemographic data are presented in Table 1. Male
and female control subjects did not differ in age, education,
ethnicity, and handedness. Likewise, male and female patients
did not differ in these variables. Clinically, they differed in age
of onset of disease with men (20.6 years 6 4.7) having an earlier
onset than women (25.3 years 6 7.9). They did not differ in
number of hospitalizations. Patient and control groups differed in
ethnicity, handedness, and education. As expected, patients
attained fewer years of education than control subjects but did
not differ in parental education.

Procedures
EMOTION RECOGNITION TASK. A computerized task,
consisting of emotion and age identification subtests, was constructed using Authorware Professional 2.0 (Macromedia, San
Francisco). The test is a computerized version of the emotion
discrimination task developed in our laboratory (Erwin et al
1992). In the emotion discrimination condition, participants were
presented with 40 faces displaying happy (five male, five
female), sad (five male, six female), or neutral (nine male, 10
female) expressions (Figure 1). The age perception task included
40 faces of the same people with similar distribution of happy
(eight male, six female), sad (five male, five female), or neutral
(eight male, eight female) expressions. The order of stimuli was
randomized but fixed across subjects.
Participants were asked to rate the emotional valence of the
expression on each face on a seven-point scale ranging from very
sad, moderately sad, somewhat sad, neutral, somewhat happy,
moderately happy, to very happy. Choices were displayed on the
screen along with the stimuli, and subjects responded by clicking
the mouse button. As soon as participants had responded to all 40
emotion stimuli, the task proceeded to the age perception
component. Subjects were asked to rate the age of the 40 facial
stimuli on a seven-point scale ranging from “70s” on top to
“teens” on the bottom. Average testing time for the entire task
was 10 min for control subjects and 20 min for patients. Tests

130

C.G. Kohler et al

BIOL PSYCHIATRY
2000;48:127–136

Table 1. Sociodemographic Characteristics

Gender (men/women)
Age
Handedness
Ethnicity

Subject education
Parental education (paternal 1
maternal:2)
Age at onset
men
women
Duration (years)
No. of hospitalizations

Healthy group

Schizophrenic group

Significance

25/20
27.5 6 8.5
Right: 42
Left: 3
White: 32
African American: 12
Asian: 1
14.9 6 2.0

20/15
30.6 6 9.5
Right: 26
Left: 8
White: 16
African American: 19

ns
ns
x2 5 4.59
p 5 .03
x2 5 4.65
p 5 .02

13.4 6 2.4

11.7 6 1.8

12.5 6 2.5

t 5 3.1
p 5 .003
n.s.

—
—
—
—

20.6 6 4.7
25.3 6 7.9
5.6 6 7.8
1.45 6 .51

t 5 2.1
p 5 .05
—
—

were administered by a trained neuropsychologic tester, and
participants were provided with a demonstration of tasks and
time to practice.
SYMPTOM RATINGS. Symptom severity ratings were performed by trained investigators (ICC . .85) and included the
Scale for the Assessment of Negative Symptoms (SANS; Andreasen 1984a), the Scale for the Assessment of Positive Symptoms (SAPS; Andreasen 1984b), and the 21-item Hamilton
Depression Rating Scale (HDRS; Hamilton 1960). Ratings were
obtained within 3 months of testing in 27 of the 35 patients and
within 8 months for the remaining patients. Patients remained
clinically stable during this period.
NEUROPSYCHOLOGIC BATTERY. A standardized, previously detailed (Saykin et al 1991, 1994) neuropsychologic
battery was administered to participants. Briefly, the battery
consisted of tests evaluating eight functional domains:

1) Abstraction–flexibility: Wisconsin Card Sorting Test
(Heaton 1981)
2) Attention-vigilance: Continuous Performance Test (Gordon 1986), Seashore Rhythm Test (Halstead-Reitan Neuropsychological Test Battery HRB] Reitan and Wolfson
1985), letter and symbol cancellation (Mesulam 1985)
3) Verbal memory: story recall (Wechsler Memory Scale,
Wechsler 1945; Wechsler Memory Scale-Revised, Wechsler 1987), California Verbal Learning Test (Delis et al
1983)

4) Spatial memory: design reproduction (Wechsler Memory
Scale, Wechsler 1945; Wechsler Memory Scale-Revised,
Wechsler 1987), Benton Facial Recognition (Benton and
Van Allen 1968), Continuous Visual Memory Test Trahan
and Larrabee 1988)
5) Verbal intelligence and language function: vocabulary
(Wechsler Adult Intelligence Scale-Revised [WAIS-R],
Wechsler 1981), controlled oral word association (Multilingual Aphasia Examination [MAE], Benton and Hamsher 1976), animal naming (Boston Diagnostic Aphasia
Exam [BDAE], Goodglas and Kaplan 1983), MAE Visual
Naming (Benton and Hamsher 1976), Reading Sentences
and Paragraphs subtest from the BDAE (Goodglass and
Kaplan 1983)
6) Spatial abilities: Benton Judgment of Line Orientation
(Benton et al 1978), block design (WAIS-R, Wechsler
1981)
7) Sensory: stereognosis (Luria-Nebraska Neuropsychological Battery, Golden et al 1980)
8) Fine manual motor functions: finger tapping (HRB, Reitan
and Wolfson 1985).
The battery was administered by trained neuropsychologic
technicians, who are supervised by neuropsychologists, and is
scored according to the established procedures for each test. For
18 of the 35 patients and 23 of the 45 control subjects, the battery
was administered within 3 months of the facial recognition tasks.
For the remaining subjects, the battery was administered within
a year.

Figure 1. Neutral, sad, and happy facial
expressions.

Emotion Recognition in Schizophrenia

BIOL PSYCHIATRY
2000;48:127–136

131

Figure 2. Means of error rates for the age discrimination and emotion discrimination tasks for male and female posers (facial stimuli)
in men (CNT_M) and women (CNT_F) who are healthy and in men (SCH_M) and women (SCH_F) who are ill with schizophrenia.

Data Analysis
EMOTION AND AGE RECOGNITION. The dependent measure for performance was error rate defined as percentage of
responses that are more than one point away from the correct
response on the seven-point scale. In the emotion discrimination
task, the correct response was defined by the mean of the
normative sample used to validate the stimuli (Erwin et al 1992).
For the age discrimination task, birth dates were available for
each poser. These percentages were determined separately for
male and female faces. To satisfy normality assumptions for the
repeated measures multivariate analysis of variance
(MANOVA), these percentages were transformed using the
arcsin transformation. The model used was a repeated measures
MANOVA, with error rates (for emotion and age discrimination
tasks) as outcome measures, diagnosis (patients, control subjects), and gender of the observer as between-group factors, and
gender of facial stimuli as the within-group factor. Handedness,
ethnicity, and education were considered as covariates in our
model. Additionally, a model was fit to the patient group to
assess the impact of duration.
SYMPTOM RATINGS. Spearman correlation coefficients
were computed for emotion and age recognition errors (arcsine
transformed for normality), and the SANS, SAPS, and HDRS
ratings. For SANS, we used the global ratings for affective
flattening, alogia, avolition, and anhedonia; for SAPS, we used
the global ratings for hallucinations, delusions, bizarre behavior,
and thought disorder; for HDRS we used the total score.

NEUROPSYCHOLOGICAL
PERFORMANCE. Spearman
correlation coefficients were computed for emotion and age
recognition errors (arcsine transformed for normality) and for
cognitive domains of abstraction-flexibility, attention, verbal and
spatial memory, language and spatial abilities, and motor and
sensory abilities. For the neuropsychologic battery, raw test
scores were transformed to standard equivalents (z scores) using
the means and SEM of the normal control group, as previously
described (Saykin et al 1991, 1994).

Results
Emotion and Age Recognition
The mean error rates for male and female faces in patients
and controls (Figure 2) indicate impaired performance in
schizophrenia that is moderated by gender of poser and
gender of subject. The MANOVA comparing error rates in
the age and emotion discrimination tasks showed that
patients performed worse than comparison participants
across tasks [F(1,76) 5 24.00, p , .001]; however,
there was no task 3 diagnosis interaction [F(1,76) 5
1.30, p 5 .25]. Thus, the hypothesis of differential
deficit in emotion discrimination was not supported. Follow
up ANOVAs for the age discrimination tasks revealed a
main effect of diagnosis [F(1,76) 5 18.04, p 5 .001].
For emotion discrimination, there were main effects of

132

C.G. Kohler et al

BIOL PSYCHIATRY
2000;48:127–136

both diagnosis [F(1,76) 5 10.95, p 5 .001] and
gender [F(1,76) 5 4.78, p 5 .03], with women having
fewer errors.
A MANOVA breaking the scores by poser’s gender
showed a task 3 poser gender interaction, [F(1,76) 5
4.39, p , .05], reflecting increased error rate for
judging emotional displays of female posers in contrast to
equal error rate on age discrimination. To test this hypothesis more directly, we calculated an opposite-gender
emotion detection index, defined as same-gender errors
minus opposite-gender errors. As predicted, this score was
negative for men [20.04 6 0.06, t(23) 5 2.95, p ,
.01] and positive for women [0.05 6 0.07, t(18) 5
2.59, p , .01] in the healthy group (between group
difference was significant, t 5 3.92, p , .001). The
effect was absent in men with schizophrenia [20.02 6
0.03, t(18) 5 0.46, p 5 .65] and not significant in
women with schizophrenia [0.04 6 0.03, t(13) 5
1.41, p 5 .18] probably attributable to low power
(because the effect size is similar to that of the healthy
women).
The inclusion of handedness, education, and ethnicity as
covariates did not alter our findings. Additionally, in a
separate model including only patients, when duration was
added as a covariate, none of these findings were altered.
A MANOVA contrasting emotion discrimination performance for happy with sad faces showed a main effect of
emotional valence, with more errors for sad than happy
faces [F 5 44.07, df 5 1,76, p , .001]. When
performance was examined separately for happy and sad
expressions, we found different effects for the two tasks
depending on diagnosis and gender. Patients performed
more poorly than control subjects on recognition of happy
faces [F 5 8.16; df 5 1,76; p 5 .005], whereas the
effect is weaker for sad faces [F 5 4.54, df 5 1,76;
p 5 .036]. Men performed more poorly on identification
of sad faces regardless of diagnosis [F 5 5.84; df 5
1,76; p 5 .018]. Whereas fewer errors were made
across groups in identifying happy faces than identifying
sad faces, men across groups identified sad female expressions most poorly [F 5 12.15; df 5 1,76; p , .001].

Emotion and Age Recognition and Symptomatology
Overall, age recognition performance did not correlate
with severity of symptoms. In contrast, errors on the
emotion recognition task were associated with greater
severity of alogia [r(33) 5 .46, p , .01] from the
SANS, and hallucinations [r(33) 5 .29, p , .05] and
thought disorder [r(33) 5 .48, p , .01] from the SAPS
in the whole group of patients and the patients who
completed testing of age and emotion recognition and
symptom ratings within 3 months. Severity of depressive

symptoms was not associated with emotion or age recognition performance. Results for subjects who completed
emotion and symptom ratings within 3 months and more
than 3 months apart were similar for negative, but not
positive symptoms. In the group that completed testing
and ratings within 3 to 8 months, similar correlations were
found for alogia, but not for hallucinations and thought
disorder, indicating relative stability of negative and not
positive symptoms over time.

Emotion and Age Recognition and Cognition
Results for subjects who completed emotions and neuropsychologic testing within 3 months and more than 3
months apart were similar and are reported for whole
patient and control groups. In healthy control subjects,
neither emotion nor age recognition performances showed
significant correlations with any neuropsychologic domain
(Figure 3). In patients, age recognition performance was
not correlated with neurocognitive performance, but better
emotion recognition performance was correlated specifically with better abstraction-flexibility [r 5 .36, df 5
33, p , .05], attention [r 5 .60, df 5 33, p , .001],
verbal [r 5 .59, df 5 33, p , .001], and spatial
memory [r 5 .65, df 5 33, p , .001], and language
abilities [r 5 .55, df 5 33, p , .001].

Discussion
Patients performed more poorly than control subjects on
both facial processing tasks, without an indication of
differential impairment. Chapman and Chapman (1978)
proposed that to probe for the presence of a differential
deficit, performance of an experimental task needs to be
compared with performance on a psychometrically
matched control task. Over the past 15 years, studies that
have used a differential deficit design have yielded findings both supporting (Borod et al 1993; Heimberg et al
1992; Walker et al 1984) and failing to support (Archer et
al 1992; Feinberg et al 1986; Kerr and Neale 1993; Novic
et al 1984; Salem et al 1996) the presence of such a
differential or specific emotion recognition impairment.
We reported differential deficit in an earlier study (Heimberg et al 1992), which used a larger set of the same facial
stimuli but presented them in a blocked rather than
interspersed or randomized design as was done here.
Whether blocked presentation makes the task more sensitive to differential impairment in schizophrenia merits
further study.
Although impairment in age discrimination was similar
in men and women with schizophrenia, men were disproportionately impaired in emotion recognition relative to
women with schizophrenia. When we evaluated recogni-

Emotion Recognition in Schizophrenia

BIOL PSYCHIATRY
2000;48:127–136

133

Figure 3. Spearman correlation coefficients between age (shaded columns) and emotion (solid columns) recognition errors in
schizophrenia subjects. SANS, Scale for the Assessment of Negative Symptoms; AFF, affect; ALO, alogia; AVO, avolition; ANH,
anhedonia; SAPS, Scale for the Assessment of Positive Symptoms; HAL, hallucinations; DEL, delusions; BIZ, bizarre behavior; THD,
thought disorder; NP, neuropsychological; ABF, abstraction–flexibility; ATT, attention; VMEM, verbal memory; SMEM, spatial
memory; LAN, language; SPA, spatial; SEN, sensory; MOT, motor.

tion of happy and sad faces, these effects too were
moderated by the observers’ gender and diagnosis. All
groups had fewer errors for happy than for sad discrimination, replicating our earlier reports (Erwin et al 1992;
Heimberg et al 1992). Also replicating an earlier finding
(Erwin et al 1992), women were more accurate in discriminating emotions on male than on female faces, whereas
men did not show the corollary effect. This result could be
interpreted to suggest that women are less facially expressive, but recent evidence indicates that women are more
facially expressive in their emotions (Kring and Gordon
1998); however, this finding is consistent with the “huntergatherer” hypothesis (Cherney and Ryalls 1999; Hill 1984;
Lee 1974). This hypothesis stipulates that from an evolutionary standpoint, it was much more important for individuals of both genders to interpret emotional displays of
the physically powerful and socially dominant male than
to detect emotional expression in females. It is noteworthy
that both male and female patients made more errors than
control subjects in recognizing happy expressions. For sad
expressions, men but not women with schizophrenia
showed impairment relative to control subjects.
Beyond the question of differential deficits, the aim of
this study was to examine whether emotion processing
deficits are uniquely associated with neurocognitive performance and symptom severity. The results supported
this notion. Although neither symptom severity nor neuropsychologic performance was associated with age recognition errors, both were associated with errors in emo-

tion discrimination. The correlations with symptoms were
significant for the negative symptoms of alogia and for the
positive symptoms of hallucinations and thought disorder.
Positive symptoms (Barta et al 1990; Shenton et al 1992)
have been linked to neural circuitry of the temporal lobes,
whereas negative symptoms have been associated with
frontal (Buchanan et al 1994) and to a lesser extent
temporal (Turetsky et al 1995) structures. It was surprising
to find that the emotion processing deficit was associated
with alogia, rather than other negative symptoms such as
affective flattening, avolition, or anhedonia. This suggests
that emotion recognition has at least partially independent
effects on symptomatology that do not necessarily relate to
disturbed affect and psychosocial functioning. Furthermore, the association of emotion recognition, but not age
recognition, with core symptoms of schizophrenia offers
indirect evidence that emotion recognition may help examine temporolimbic features of schizophrenia. Longitudinal designs are needed to establish whether change in
performance on emotion discrimination tasks is predictive
of symptomatic improvement.
The were no correlations between performance on
either age discrimination or emotion discrimination and
neuropsychologic domains for the healthy participants.
This is unusual because most cognitive tasks share common variance related to sensory input, attention to the task,
facial processing, and linguistic ability (the “g factor”).
Nonetheless, there is increased evidence that face processing has distinct neural substrates (see review in Mesulam

134

C.G. Kohler et al

BIOL PSYCHIATRY
2000;48:127–136

1998). In patients, although there were no correlations
between age discrimination and neuropsychologic performance, significant and large correlations were observed
between performance on emotion discrimination and several neurocognitive domains. These were specific to abstraction and mental flexibility, attention, memory (verbal
and spatial), and language, all related to frontal and
temporal functioning. Several studies have indicated differential deficits in verbal memory in patients with schizophrenia, superimposed on generally lower cognitive performance (Saykin et al 1991, 1994). Thus, the correlations
suggest that the impairment in frontotemporal functions
extends to limbic regions related to emotional processing.
Our study is limited in sample size and the cross-sectional
design; it does not offer correlations between emotion
processing and regional brain function. Based on our
findings, however, we propose that emotion recognition,
as a primarily temporal lobe function, is impaired in
schizophrenia and contributes to the deficit load of patients suffering from this disorder. Even though the deficit
is not more severe for emotion recognition than for age
recognition, the emotion processing deficit is uniquely
associated with symptom severity and neurocognitive
deficits and therefore merits further scrutiny.
Our study has several limitations related to the task and
patient sample. Compared with our previous study of
emotion and age recognition in schizophrenia, the present
study uses only 40 stimuli, which may yield insufficient
statistical power to probe for differential impairment
regarding positive and negative emotions. The age recognition task, although carefully constructed to serve as a
control task that differs only in the component of recognition, may represent another limitation. Facial display of
emotions may influence the assessment of age. Furthermore, the requirement to assess emotional faces for age
rather than affect could be distracting to the subjects. The
number of moderately emotional faces in the age recognition task was insufficient to probe for this effect. These
test-related limitations may have contributed to the failure
to establish a differential deficit for emotion recognition.
On the other hand, the association of emotion recognition,
but not age recognition, with severity of symptoms and to
a lesser extent with cognitive impairment, can be considered as indirectly supporting the presence of specific
associations for the emotion processing deficit. Limitations regarding the patient group included patients being at
different stages of the illness, the use of different neuroleptics, and relating experimental measures that were not
performed at the same time. Patients were tested at
different times of illness, including seven patients with
schizophreniform disorder who were later diagnosed with
schizophrenia. Testing of age and emotion recognition and
symptom and cognitive assessments were performed in

patients who were symptomatic but clinically stable at the
time of evaluation. Unfortunately, the sample of patients
with first onset of schizophrenia was too small to allow for
comparison with more chronically ill patients. The use of
different medications, including atypical, typical, and
combined typical and atypical neuroleptics, as well as the
presence of neuroleptic-naive patients, precluded comparison of medication effects. It is noteworthy that comparing
the patients who underwent symptom rating within 3
months of testing to those with larger intervals were
remarkably similar. Nonetheless, it would have been
preferable if all assessments had been be performed within
shorter intervals.
Comparison of patients with schizophrenia with those
who have depression or mania with psychotic features
may help elucidate the disease specificity of deficits in
emotion recognition. Comparison of schizophrenia with
patients suffering from temporal lobe epilepsy would yield
information regarding emotional processing associated
with impairment in temporolimbic processing. To study
the effects of medication and stage of illness, we need to
assess emotion recognition longitudinally in the setting of
standardized neuroleptic medication treatment. Evaluation
of emotion recognition using functional imaging will help
delineate the extent of neuronal impairment related to
emotional processing that may contribute the pathophysiology of schizophrenia.

This research was supported by the National Institute of Health Grants
Nos. MH-43880, MH-42191, MH-01336, and MO1RR0040.

References
Adolphs R, Damasio H, Tranel D, Damasio AR (1996): Cortical
systems for the recognition of emotion in facial expressions.
J Neurosci 16:7678 – 687.
Adolphs R, Schul R, Tranel D (1998): Intact recognition of facial
emotion in Parkinson’s disease. Neuropsychology 12:253–
258.
Adolphs R, Tranel D, Damasio H, Damasio A (1994): Impaired
recognition of emotion in facial expressions following bilateral damage to the human amygdala. Nature 372:669 – 672.
Andreasen NC (1984a): The Scale for the Assessment of Negative
Symptoms (SANS). Iowa City: The University of Iowa.
Andreasen NC (1984b): The Scale for the Assessment of Positive
Symptoms (SAPS). Iowa City: The University of Iowa.
Anthony BJ (1978): Piagetian egocentrism, empathy, and affect
discrimination in children at high risk for psychosis. In:
Anthony EJ, Koupernik C, Chiland C, editors. The Child and
His Family: Vulnerable Children, Vol 4. New York: Wiley,
359 –379.
Archer J, Hay DC, Young AW (1994): Movement, face processing, and schizophrenia: Evidence of a differential deficit in
expression analysis. Br J Clin Psychol 33:517–528.

Emotion Recognition in Schizophrenia

Barta PE, Pearlson GD, Powers RE, Richards SS, Tune LE
(1990): Auditory hallucinations and smaller superior temporal gyral volume in schizophrenia. Am J Psychiatry 147:
1457–1462.
Benton AL, Hamsher K (1976): Multilingual Aphasia Examination. Iowa City, Iowa: AJA Associates.
Benton AL, Hamsher KD, Varney NR, Spreen O (1983):
Contributions to Neuropsychological Assessment. New York:
Oxford University Press.
Benton AL, Van Allen MW (1968): Impairment in facial recognition in patients with cerebral disease. Cortex 4:344 –358.
Benton AL, Varney NR, Hamsher K (1978): Visuospatial judgement: A clinical test. Arch Neurol 35:364 –367.
Beratis S, Gabriel J, Hoidas S (1994): Age at onset in subtypes
of schizophrenic disorders. Schizophr Bull 20:287–296.
Bleuler E (1911): Dementia Praeox oder die Gruppe der Schizophrenien, Aschaffenburgs Handbuch. Leipzig: Deutike.
Blonder LX, Bowers D, Heilman KM (1991): The role of the
right hemisphere in emotional communication. Brain 114:
1115–1127.
Borod JC, Martin CC, Alpert M, Brozgold A, Welkowitz J
(1993): Perception of facial emotion in schizophrenic and
right-brain damaged patients. J Nerv Ment Dis 181:494 –502.
Bowers D, Blonder LX, Feinberg T, Heilman KM (1991):
Differential impact of right and left hemisphere lesions on
facial emotion and object imagery. Brain 114:2593–2609.
Buchanan RW, Strauss ME, Kirkpatrick B, Holstein C, Breier A,
Carpenter WT (1994): Neuropsychological impairments in
deficit vs nondeficit forms of schizophrenia. Arch Gen
Psychiatry 51:804 – 811.
Chapman LJ, Chapman JP (1978): The measurement of differential deficit. J Psychiatr Res 14:303–311.
Charlesworth WR (1982): An ethological approach to research
on facial expressions. In: Izard CE, editor. Measuring Emotions in Infants and Children. New York: Cambridge University Press, 317–334.
Cherney ID, Ryalls BO (1999): Gender-linked differences in the
incidental memory of children and adults. J Exp Child
Psychol 72:305–328.
Cutting J (1981): Judgment of emotional expression in schizophrenics. Br J Psychiatry 139:1– 6.
Damasio AR (1998): Emotion in the perspective of an integrated
nervous system. Brain Res Brain Res Rev 26:83– 86.
Delis C, Kramer JH, Kaplan E, Ober BA (1983): California
Verbal Learning Test, Research Edition. Cleveland: Psychological Corp.
Ekman P, Friesen WV, Ellsworth P (1972): Emotion in the
Human Face: Guidelines for Research and an Integration of
Findings. New York: Pergamon.
Erwin RJ, Gur RC, Gur RE, Skolnick BE, Mawhinney-Hee M,
Smailis J (1992): Facial emotion discrimination: I. Task
construction and behavioral findings in normals. Psychiatry
Res 42:231–240.
Feinberg TE, Rifkin A, Schaffer C, Walker E (1986): Facial
discrimination and emotional recognition in schizophrenia
and affective disorders. Arch Gen Psychiatry 43:276 –279.
First MG, Spitzer RL, Gibbon M, Williams JB (1995a): Structured Clinical Interview for DSM-IV Patient Edition

BIOL PSYCHIATRY
2000;48:127–136

135

(SCID-P). New York: New York State Psychiatric Institute,
Biometrics Research Department.
First MB, Spitzer RL, Gibbon M, Williams JBW (1995b):
Structured Clinical Interview for DSM-IV Axis I Disorders,
Non-Patient Edition (SCID-NP). New York: New York State
Psychiatric Institute, Biometrics Research Department.
Golden CJ, Hammeke TA, Puritsch AD (1980): Luria-Nebraska
Neuropsychological Battery, Manual. Los Angeles: Western
Psychological Services.
Goldstein JM, Tsuang MT, Faraone SV (1989): Gender and
schizophrenia: Implications for understanding the heterogeneity of the illness. Psychiatry Res 28:243–253.
Goodglass H, Kaplan E (1983): The Assessment of Aphasia and
Related Disorders, 2nd ed. Philadelphia: Lea and Febiger.
Gordon M (1986): Manual, Gordon Diagnostic System. New
York: Gordon Diagnostic Systems.
Gur RC, Erwin RJ, Gur RE, Zwil AS, Heimberg C, Kraemer HC
(1992): Facial emotion discrimination: II. Behavioral findings
in depression. Psychiatry Res 42:241–251.
Gur RE, Mozley PD, Resnick SM, Levick S, Erwin R, Saykin AJ,
Gur RC (1991): Relations among clinical scales in schizophrenia: Overlap and subtypes. Am J Psychiatry 148:472– 478.
Hamilton M (1960): A rating scale for depression. J Neurol
Neurosurg Psychiatry 23:56 – 62.
Heaton RK (1981): Wisconsin Card Sorting Test, Manual.
Odessa, FL: Psychological Assessment Resources.
Heilman KM, Gilmore RL (1998): Cortical influences in emotion. J Clin Neurophysiol 15:409 – 423.
Heimberg C, Gur RE, Erwin RJ, Shtasel DL, Gur RC (1992):
Facial emotion discrimination: III. Behavioral findings in
schizophrenia. Psychiatry Res 42:253–265.
Hill J (1984): Prestige and reproductive success in man. Ethol
Sociobiol 5:77–95.
Izard CE (1971): The Face of Emotion. New York: AppletonCentury.
Jacobs DH, Shuren J, Heilman KH (1995): Impaired perception
of facial identity and facial affect in Huntington’s disease.
Neurology 45:1217–1218.
Jayaswal SK, Paraveenlal K, Khandelwal SK, Mohan D (1987):
Sex difference in the age at symptom onset in schizophrenia.
Indian J Psychol Med 10:68 –74.
Kerr SL, Neale JM (1993): Emotion perception in schizophrenia:
Specific deficit or further evidence of generalized poor
performance? J Abnorm Psychol 102:312–318.
Kring AM, Gordon AH (1998): Sex differences in emotion:
Expression, experience, and physiology. J Pers Soc Psychol
74:686 –703.
Kring AM, Neale JM (1996): Do schizophrenia patients show a
disjunctive relationship among expressive, experiential, and
psychophysiological components of emotion? J Abnorm Psychol 105:249 –257.
Lang PJ, Bradley MM, Cuthbert BN (1998): Emotion, motivation, and anxiety: brain mechanisms and psychophysiology.
Biol Psychiatry 44:1248 –1263.
Lee RB (1974): Male-female residence arrangements and political power in human hunter-gatherers. Arch Sex Behav
3:167–173.
Leventhal H, Tomarken AJ (1986): Emotions; today’s problems.
Ann Rev Psychol 3:565– 610.

136

BIOL PSYCHIATRY
2000;48:127–136

Lezak MD (1995): Perception, 3rd ed. Oxford, UK: Oxford
University Press.
Mesulam M (1985): Principles of Behavioral Neurology. Philadelphia: F.A. Davis.
Mesulam MM (1998): From sensation to cognition. Brain
121:1013–1052.
Mikhailova ES, Vladimirova TV, Iznack AF, Tsusulkovskaya
EJ, Sushko NV (1996): Abnormal recognition of facial
expression of emotions in depressed patients with major
depression disorder and schizotypal personality disorder. Biol
Psychiatry 40:697–705.
Novic J, Luchins DJ, Perline R (1984): Facial affect recognition
in schizophrenia. Is there a differential deficit? Br J Psychiatry 144:533–537.
Reitan RM, Wolfson D (1985): The Halstead Reitan Neuropsychological Test Battery: Theory and Clinical Interpretation.
Tucson: Neuropsychology Press.
Sackeim HA, Greenberg M, Weiman A, Gur RC, Hungerbuhler
J, Geschwind N (1982): Hemispheric asymmetry in the
expression of positive and negative emotions. Neurologic
evidence. Arch Neurol 39:210 –218.
Salem JE, Kring AM (1998): The role of gender differences in
the reduction of etiologic heterogeneity in schizophrenia. Clin
Psychol Rev 18:795– 819.
Salem JE, Kring AM, Kerr SL (1996): More evidence for
generalized poor performance in facial emotion perception in
schizophrenia. J Abnorm Psychol 105:480 – 483.
Saykin AJ, Gur RC, Gur RE, Mozley D, Mozley LH, Resnick
SM, et al (1991): Neuropsychological function in schizophrenia: selective impairment in memory and learning. Arch Gen
Psychiatry 48:618 – 624.
Saykin AJ, Shtasel DL, Gur RE, Kester DB, Mozley LH,
Stafiniak P, Gur RC (1994): Neuropsychological deficits in
neuroleptic naive patients with first-episode schizophrenia.
Arch Gen Psychiatry 51:124 –131.

C.G. Kohler et al

Schneider F, Gur RC, Gur RE, Shtasel DL (1995): Emotional
processing in schizophrenia: Neurobehavioral probes in relation to psychopathology. Schizophr Res 17:67–75.
Shenton ME, Kikinis R, Jolesz FA, Pollak SD, LeMay M, Wible
CG, et al (1992): Abnormalities of the left temporal lobe and
thought disorder in schizophrenia: a quantitative magnetic
resonance imaging study. N Engl J Med 327:604 – 612.
Shtasel DL, Gur RE, Gallacher F, Heimberg C (1992): Gender
differences in the clinical expression of schizophrenia.
Schizophr Res 7:225–231.
Shtasel DL, Gur RE, Mozley PD, Richards J, Taleff MM,
Heimberg C, et al (1991): Volunteers for biomedical research:
Recruitment and screening of normal controls. Arch Gen
Psychiatry 48:1022–1025.
Trahan DE, Larrabee GJ (1988): Continuous Visual Memory
Test: Professional Manual. Odessa, FL: Psychological Assessment Resources.
Turetsky BT, Cowell PE, Gur RC, Grossman RI, Shtasel DL,
Gur RE (1995): Frontal and temporal lobe brain volumes in
schizophrenia: Relationship to symptomatology and clinical
subtype. Arch Gen Psychiatry 52:1061–1070.
Vazquez-Barquero JL, Cuesta Nunez MJ, De La Varga M,
Herrera Castanedo S (1995): The Cantabria first episode
schizophrenia study: A summary of general findings. Acta
Psychiatr Scand 9:156 –162.
Walker E, McGuire M, Bettes B (1984): Recognition and
identification of facial stimuli by schizophrenics and patients
with affective disorders. Br J Clin Psychol 23:37– 44.
Wechsler D (1945): A standardized memory scale for clinical
use. J Psychol 19:87–95.
Wechsler D (1981): Wechsler Adult Intelligence Scale-Revised,
Manual. Cleveland: Psychological Corp.
Wechsler D (1987): Wechsler Memory Scale-Revised, Manual.
San Antonio: Psychological Corp.