Open Journal of Depression
2013. Vol.2, No.4, 64-71
Published Online November 2013 in SciRes (
Open Access
Mood-Related Negative Bias in Response to Affective Stimuli in
Patients with Major Depression
Rottraut Ille1,2*, Peter Hofmann2, Christoph Ebner2, Hans-Peter Kapfhammer2,
Anne Schienle1
1Institute of Psychology, Karl-Franzens-University of Graz, Graz, Austria
2University Hospital of Psychiatry, Medical University of Graz, Graz, Austria
Email: *
Received July 25th, 2013; revised August 30th, 2013; accepted September 8th, 2013
Copyright © 2013 Rottraut Ille et al. This is an open access article distributed under the Creative Commons At-
tribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the
original work is properly cited.
Background: The study focuses on the type and degree of impairment in the processing of affective faces
and scenes in patients afflicted with major depression (MD). We investigated effects of emotional traits,
gender, depression severity, and cognitive performance. Method: Thirty MD patients (15 men, 15 women)
and 30 healthy controls were presented with pictures of emotional facial expressions and affective scenes.
They were asked to estimate the intensity and allocation of the emotions expressed by the faces as well as
the elicited emotions by the scenes. Results: MD patients showed a broad impairment of emotion recogni-
tion. Patients’ responses to happy faces suggested a negativity bias, which also became evident in the
perception of emotional scenes. The negativity bias was stronger in male than female patients. Depression
severity was negatively related to experience of happiness. Patient’s lower cognitive performance was
associated with allocation accuracy of angry and disgusted faces. Conclusions: Our findings show accor-
dance with the mood-congruency hypothesis. Depression treatment should put increased focus to the as-
sociation between negative mood bias and social functioning.
Keywords: Emotion Recognition; Emotion Experience; Impairment; Negativity Bias; Cognitive
Performance; Severity of Depression
Major Depression (MD) is characterized by low mood ac-
companied by lowered self-esteem, and by loss of interest or
pleasure in normally enjoyable activities (APA, 2000). With in-
creasing severity of illness MD is associated with impaired so-
cial functioning. The afflicted patients have a reduced social
competence, emit fewer interpersonal behaviors and are less
skillful in solving interpersonal problems (Persad & Polivy,
1993). Furthermore, several studies could verify biased proc-
essing of emotional information (Bylsma et al., 2008; Persad &
Polivy, 1993).
Depressive disorder involves several types of emotional ab-
normalities, most notably increased propensity to negative af-
fect and reduced capacity to experience pleasure (Drevets,
2001). In order to explain how MD alters emotional reactivity
three alternative approaches have been put forward (Bylsma et
al., 2008): MD may alter emotional reactivity by increased
negative reactivity, by reduced positive reactivity or by “emo-
tion context insensitivity” (ECI; reduced positive and negative
reactivity; Rottenberg et al., 2005).
The pronounced tendency to experience negative emotions
and the reduced tendency to experience positive emotions by
patients with depression have been explained by the mood con-
gruency hypothesis (Bower, 1981). It suggests that depressed
mood may enhance the processing of mood congruent material
and impair the processing of mood incongruent material. But
several studies on depression (e.g., Persad & Polivy, 1993;
Mikhailova et al., 1996) found no evidence for mood congru-
ency effects.
The ECI model states that individuals with depression will
exhibit decreased reactivity to all emotion cues, regardless of
valence (Rottenberg, 2005; Rottenberg et al., 2005; Rottenberg,
2007). For this reason, individuals with MD should display
lowered responses to both positive and negative stimuli, in-
cluding sadness. Data of a meta-analysis by Bylsma et al. (2008)
confirmed the ECI hypothesis, suggesting an overall reduced
emotional reactivity in MD, with the reduction larger for posi-
tive stimuli. Results were comparable for self-reported experi-
ence, expressive behavior, and peripheral physiology. Rotten-
berg et al. (2005) found most pronounced attenuation in emo-
tion experience reports. Furthermore, MD patients with the
most pronounced ECI showed the most severe symptoms, the
longest episodes of depression, and the poorest overall psycho-
social functioning.
Depression is also associated with abnormalities in the iden-
tification of emotional facial expressions (Surguladze et al.,
2004). Affective facial expressions are salient features in social
interactions (Persad & Polivy, 1993). Therefore, the ability to
accurately identify others’ emotional facial expressions is of
considerable importance. According to previous data MD is
characterized by a reduced ability to correctly identify affective
facial expressions.
*Corresponding author.
Open Access 65
There are some inconsistencies, if deficits in facial expres-
sion decoding are emotion-specific or more general. Findings
from Mikhailova et al. (1996) attested poor recognition of hap-
piness and sadness in patients with depression. Mendlewicz et
al. (2005) reported a lower decoding accuracy of depressed
patients only for angry facial expressions, while for happy,
disgusted, sad, and fearful expressions they did not differ from
healthy controls. Several authors also found an enhanced ten-
dency of depressed patients to judge facial expressions as dis-
playing negative affect (“negativity bias”; Hale, 1998; Mik-
hailova et al., 1996; Milders et al., 2010).
It could also be shown that severity of depressive symp-
toms plays a role in recognition accuracy of facial expressions.
Hale (1998) reported a positive relationship between judgments
of negative emotions in facial expressions and depression se-
verity. Results by Leppänen et al. (2004) revealed a correlation
between depression symptom score and sadness bias in neutral
As most studies tested recognition performance in depression
only for several emotions, it is not clear, whether this disorder
is associated with a general deficit in affecting identification or
with impairments in the recognition of specific emotions. Fur-
thermore, it remains unresolved, whether this deficit can be
seen as a mood-congruent interpretation bias (e.g., towards
sadness) or as an overall deficient decoding of emotional facial
To summarize, results concerning decoding performance in
MD are inconsistent. This may be a consequence of different
intensities of facial expressions presented. Furthermore, MD
patients must be seen as a heterogeneous population, with co-
morbidity of several mental disorders like anxiety disorders,
posttraumatic stress disorder or addiction. As a consequence,
patients’ emotional reactivity may be different.
Finally, depressive symptoms differ between genders (Angst
et al., 2002; Kockler & Heun, 2002; Scheibe et al., 2003). Ag-
gression and especially anger attacks are more prevalent in
males compared to females (Winkler et al., 2005), while fe-
males report more anxiety (Scheibe et al., 2003). This may
result in a gender-specific difference of emotion experience and
perception of emotions in others. Furthermore, several authors
have reported that in healthy subjects females show greater
identification accuracy of affective facial expressions compared
to men (e.g., Cellerino et al., 2004; Kring & Gordon, 1998). It
would be of interest if women suffering from depression can
benefit from this advantage.
To our knowledge, no previous studies have investigated
emotion experience and allocation accuracy of emotional facial
expressions simultaneously in patients afflicted with MD. Es-
timation of affective scenes may provide information about
patients’ mood state.
The present study focuses on the type and degree of impair-
ment in the processing of affective faces and scenes in patients
with depressive disorder. We investigated effects of emo-
tional traits, gender, depression severity, and cognitive per-
We studied 30 patients with the diagnosis Major Depression
(criteria according to DSM-IV), 15 men and 15 women (M =
48.1 years, SD = 10.5), who were inpatients at the University
Hospital of Graz (Austria). Socio-economic status was based on
the highest educational level completed. Mean years of educa-
tion were 11.1 years (SD = 3.3). Pharmacologic therapy was
applied to all patients (antidepressants: 23 patients, antidepres-
sants and antipsychotics: 7 patients). We further tested 30 men-
tally healthy subjects, 15 men and 15 women, matched for age
and socio-economic status. They had been recruited by adver-
tisements in a local newspaper. The control group underwent a
standardized clinical interview (Mini-Dips; Margraf, 1994) to
exclude the presence of mental disorders. Mean age of the con-
trols was 45.4 years (SD = 9.8). Mean education level was 11.9
years (SD = 3.6). Groups did not differ in age (t (58) = 1.04, p
= .300) and years of education (t (58) = .90, p = .370).
Cognitive performance was assessed using the Test for Early
Detection of Dementia (TFDD; Ihl et al., 2000). This scale
ranges between 0 and 50 points and allows the detection of
early signs of cognitive impairment. A score lower than 35
points indicates a tentative dementia diagnosis (exclusion crite-
rion). The Cronbach’s alpha is .88.
Habitual emotional reactivity was assessed by several self-
report inventories:
The Beck Depression Inventory (BDI; German version: Haut-
zinger et al., 1994) assesses depressive symptomatology. Cron-
bach’s alpha is .88.
The Questionnaire for the Assessment of Disgust Proneness
(QADS; Schienle et al., 2002a) measures disgust propensity
and describes 37 situations, which have to be judged on 5-point
scales with regard to the experienced disgust (0 = “not disgust-
ing”; 4 = “very disgusting”; e.g., “You are just about to drink a
glass of milk as you notice that it is spoiled”). The Cronbach’s
alpha of the total scale is .90.
The disgust sensitivity scale of the Disgust Propensity and
Sensitivity Scale-Revised (DPSS-R; Van Overveld et al., 2006)
assesses a person’s tendency to evaluate disgust experiences as
negative (e.g., “It embarrasses me when I feel disgusted”). Pos-
sible mean sores range from 1 (“never”) to 5 (“always”). The
Cronbach’s alpha of the sensitivity scale is .77.
The Trait scale of the State-Trait Anxiety Inventory (STAI;
Laux et al., 1981) measures the frequency of anxious feel-
ings on a 4-point scale. The Cronbach’s alpha of the scale
is .88.
The Trait scales of the State-Trait-Anger Inventory (STAXI;
Schwenkmezger et al., 1992) assess trait anger as well as anger
expression. All items are rated on 4-point scales. Internal con-
sistence of the STAXI is .90.
Stimuli for the Picture Perception Tasks
All participants viewed emotional scenes and facial expres-
sions on a computer screen (notebook, 15 inches). The partici-
pants sat at about 50 cm from the screen. Before starting the
experiment participants were asked for their understanding of
basic emotions by a short verbal description.
Forty-two pictures with emotional facial expressions depict-
ing happiness (6), fear (6), sadness (6), anger (6), disgust (6),
and surprise (6) from the Karolinska-Set (Lundquist et al., 1998)
were presented. Half of the posers were female, half were male.
Participants were asked to view the faces as long as necessary
for getting an impression of the emotion displayed. Maximum
presentation time for each picture was 10 seconds.
Open Access
Twenty-four emotion-relevant scenes for the induction of
happiness (6), fear (6), and disgust (6) were presented. Most
scenes were taken from the International Affective Picture Sys-
tem (IAPS; Lang et al., 2001). Disgust-inducing pictures were
developed by Schienle et al. (2002b) and included scenes with
animals (maggots, bluebottles, and slugs), a dirty toilet, carrion
and an eczematous face. The fear-inducing pictures showed
threatening situations either through attacks of animals (“dog
with its teeth bared”, IAPS 1300; “white shark”, developed by
the authors) or human attacks (“man threatening a woman with
a knife”, IAPS 6350; “men with pistol”, IAPS 6230; “war
scene”, IAPS 6940; “masked robber”, IAPS 6370). Happy pic-
tures included animals (“baby seal”, IAPS 1440; “young rab-
bits”, IAPS 1750; “playing dolphins”, IAPS 1920) and food
(“roast chicken”, IAPS 7230; “gateau”, IAPS 7282; “ice cream”,
IAPS 7330). The stimulus material had been matched for
item difficulty, complexity, brightness and colour. Since the
IAPS does not include pictures which reliably induce anger,
sadness and surprise these categories were omitted. It is
known that IAPS scenes which should induce sadness or
anger usually produce mixed emotions (e.g., 50% anger,
50% sadness). Maximum presentation time for each picture
was 10 seconds.
Then, the subject was asked to rate the pictures on a 9-point
scale within 10 seconds. For each facial expression subjects
rated how intense the depicted person experienced the six basic
emotions (e.g., “Please indicate how intense the depicted per-
son experienced disgust”: 1 = very little; 9 = very intense). For
the scenes, subjects rated how intense the six basic emotions
were induced by a particular picture (e.g., “Please indicate how
intense you experienced disgust while viewing the picture”: 1 =
very little, 9 = very intense).
Asking the participants to rate emotion intensities of facial
expressions and scenes for all basic emotions (graded choice)
allowed the analysis of quantitative (intensity) as well as quali-
tative (classification accuracy) emotion processing deficits.
To avoid position effects, the order of the two-picture-per-
caption-tasks (recognition vs. experience), order of pictures,
and order of basic emotions to rate was randomised.
Statistical Analysis
All statistical analyses were carried out using SPSS Statistics
19.0 for Windows. Descriptive statistics were performed using
Student’s t-tests, and for group comparisons ANOVAS were
calculated. We computed mean intensity ratings of all six basic
emotions for affective facial expressions. For estimation of af-
fective faces repeated measures analyses of variance were car-
ried out for each emotion category separately (2 × 6 ANOVAs),
with group and rater’s sex as between-subjects factors. For
pairwise comparisons of significant interactions Student t-tests
were used. F-values were Greenhouse-Geisser-adjusted if nec-
essary. All ε were > .50. Classification accuracy of target emo-
tions was calculated as difference between rated target emotion
intensity (e.g., rated disgust intensity for disgusted facial ex-
pression) and mean intensity of all non-target emotions (fear,
anger, sadness, happiness, and surprise). Effect sizes were cal-
culated by Cohen’s d.
Group Comparisons for the Questionnaires
MD patients scored higher for depressive symptoms (BDI; t
(58) = 13.68, p < .001), trait anxiety (STAI; t (58) = 11.21, p
< .001), anger expression (STAXI; trait anger: F (1, 58) = 18.26,
p < .001; “Anger in”: F (1, 58) = 75.37, p < .001), disgust pro-
pensity (QADS; t (58) = 5.16, p < .001), and disgust sensi-
tivity (DPSS-R; t (58) = 4.16, p < .001). Patients also showed
lower cognitive performance (TFDD) than controls (t (58) =
5.52, p < .001) (Table 1). We found a sex x group interaction
for trait anger (F (1, 58) = 2.83, p = .045). Males with depres-
sion reported higher trait anger (M = 24.27, SD = 3.73) com-
pared to females with MD (M = 20.21, SD = 5.40; t (28) = 2.37,
p = .025).
Group Comparisons for the Picture Perception Tasks
Facial Expressions
1) Perceived intensities of target emotions (Table 2): Ana-
lyzing the intensity ratings of the six target emotions (anger,
disgust, sadness, fear, happiness, and surprise; e.g., disgust
intensity in faces depicting disgust) we revealed a significant
interaction for emotion x group (F (5, 280) = 2.53, p = .042).
Patients reported lower intensities than controls for surprised (t
(58) = 2.45, p = .017; d = .58) and for angry faces (t (58) = 2.12,
p = .038; d = .78).
Table 1.
Means (Standard Deviations) of questionnaires in MD patients and con-
MD patients Controls p
(n = 30) (n = 30)
TFDD 41.41 (3.97) 45.97 (2.03) <.001
BDI 28.42 (8.66) 3.07 (4.95) <.001
Mean QADS 2.88 (.70) 2.02 (.59) <.001
DPSS-R 23.17 (7.23) 16.10 (5.88) <.001
STAI 57.68 (7.54) 32.03 (9.67) <.001
Trait Anger 21.81 (4.96) 16.67 (4.04) <.001
Anger-In 21.46 (4.74) 12.37 (3.02) <.001
Anger-Out 12.58 (3.43) 11.37 (2.66) .143
Anger Control 23.08 (4.39) 24.03 (5.01) .454
Note: TFDD (Test for Early Detection of Dementia); BDI (The Beck Depression
Inventory); QADS (The Questionnaire for the Assessment of Disgust Proneness);
DPSS-R (Disgust Propensity and Sensitivity Scale-Revised); STAI (State-Trait
Anxiety Inventory); STAXI (State-Trait-Anger Inventory).
Table 2.
Means and standard deviations (SD) for target emotion intensities of
facial expressions and scenes.
Faces Scenes
Patients Controls MD
Patients Controls
Fear 6.47 (1.34)5.85 (1.95) 7.67 (1.50) 5.97 (2.43)
Disgust 6.59 (1.83)6.61 (2.15) 6.91 (1.58) 5.67 (1.59)
Happiness 7.32 (1.69)7.91 (1.04) 6.52 (2.01) 6.59 (1.84)
Anger 6.93 (1.49)7.65 (1.09)
Sadness 6.39 (1.61)6.59 (1.35)
Surprise 7.42 (1.05)7.96 (.79)
Open Access 67
2) Classification accuracy of target emotions (Figure 1): For
assessing participants’ classification accuracy of facial expres-
sions we calculated the difference between the target emotion
intensity and the mean of all non-target intensity ratings.
Analyses revealed a significant main effect for group (F (1, 50)
= 16.24, p < .001) and for sex (F (1, 50) = 15.21, p < .001).
Patients displayed lower allocation accuracy of affective faces
compared to controls, and males showed lower classification
accuracy compared to females. Furthermore, we found a sig-
nificant emotion x group interaction (F (5, 280) = 3.31, p
= .006). Patients showed lower allocation accuracy than con-
trols for all emotions with exception of fear: surprise (t (58) =
3.48, p = .001, d = .90), anger (t (58) = 3.36, p = .001, d = .86),
happiness (t (58) = 3.32, p = .002, d = .86), sadness (t (58) =
2.91, p = .005, d = .75), and disgust (t (58) = 2.40, p = .020, d
= .62). Patients assessed all negative non-target emotions in
surprised faces more intense than controls (F (5, 280) = 4.69,
p< .001; anger: t (58) = 2.79, p = .007; sadness: t (58) = 2.44, p
= .018; fear: t (58) = 2.39, p = .020; disgust (t (58) = 2.24, p
= .029). In sad faces patients also estimated all negative
non-target emotions more intense than controls (F (5, 280) =
4.71, p = .002; anger: t (58) = 3.24, p = .002; fear: t (58) = 3.18,
p = .002; disgust (t (58) = 2.34, p = .023). In fearful faces pa-
tients rated more anger than controls (F (5, 280) = 4.87, P
< .001; (t(58) = 2.64, P =.011). Moreover, compared to con-
trols, patients rated higher intensities of disgust for angry faces
(patients: M = 2.03, SD = 2.40; controls: M = 3.63, SD = 2.48; t
(58) = 2.53, p = .014). Also for happy faces patients reported
higher intensities for all negative emotions (F (5, 280) = 5.13, p
=.011; fear: t (58) = 3.14, p = .003; disgust: t (58) = 3.13, p
= .003; anger: t (58) = 2.68, p = .010; sadness: t (58) = 2.90, p
= .005). Furthermore, we found an emotion x group x sex in-
teraction for intensity ratings of happy faces (F (5, 280) = 3.34,
p = .047). Compared to females, male MD patients estimated
all negative emotions more intense (anger: t (28) = 2.18, p
= .038; sadness: t (28) = 2.27, p = .035; fear: t (28) = 2.45, p
= .021; disgust: (t (28) = 2.24, p = .035). Furthermore, male
DisgustAngerFearSadness HappinessSurprise
Decoding accuracy
p = 0.020
p = 0.001
p = 0.264
p = 0.005
p = 0.002
p = 0.001
Figure 1.
Group comparison of the mean scores (standard errors) for classifica-
tion accuracy of affective faces.
patients rated happy faces less happy (t (28) = 2.49, p = .021)
than females. The controls’ happiness ratings did not differ
between males and females (all p > .106).
3) Perception of neutral faces (Figure 2): We found a sig-
nificant effect for group (F (1, 56) = 15.21, p < .001) and for
sex (F (1, 56) = 7.47, p = .008). Patients rated neutral faces
more intense than controls, and males rated neutral faces more
intense compared to females.
Affective Scenes
1) Intensity of target emotions (Table 2): Intensity analyses
for the three target emotions (happiness, fear, disgust) showed a
significant main effect for group (F (1, 58) = 10.86, p = .002),
and an emotion x group interaction (F (2, 110) = 5.99, p = .003).
Compared to controls MD patients reported higher intensities
for the scenes eliciting fear (t (58) = 3.22, p = .002; d = .84) and
disgust (t (58) = 3.00, p = .004; d = .78). Furthermore, we
found a significant emotion x group interaction for happiness
experience (F (5, 280) = 4.97, p = .012). Patients experienced
happy scenes sadder than controls (t (58) = 2.09, p = .048).
2) Experience of neutral scenes (Figure 3): Ratings of the
neutral scenes displayed a main effect for group (F (1, 58) =
26.52, p < .001). Patients rated neutral scenes more intense than
Correlative Analyses
MD patients showed a significant negative correlation be-
tween severity of depressive symptoms (BDI score) and ex-
perience of happiness in happiness-inducing (r = .47, p = .019)
as well as neutral scenes (r = .55, p = .005). Severity of de-
pressive symptoms was also correlated to experience of sadness
in fearful (r = .42, p = .039) and neutral (r = .43, p = .031)
scenes (Figure 4). But we found no association between de-
pression severity and cognitive performance (r = .13, p = .522).
Patients’ cognitive performance was associated with allocation
accuracy of anger (r = .55, p = .002) and disgust (r = .49,
DisgustAngerFearSadness HappinessSurprise
p = 0.002
p = 0.001p = 0.001
p = 0.028
p = 0.058
p = 0.002
Figure 2.
Group comparison of the mean scores (standard errors) for intensity
ratings of neutral faces.
Open Access
p = .007). But there was no relationship between cognitive per-
formance and recognition accuracy in controls (all p > .290).
“Anger in” score of patients was negatively related to experi-
ence of happiness (r = .42, p = .032). Patients’ disgust pro-
pensity (mean QADS) was correlated with disgust ratings of
fearful faces (r = .53, p = .002), angry faces (r = .47, p = .010),
and sad faces (r = .38, p = .041) (Figure 5) as well as with dis-
gust perception in fear-inducing (r = .49, p = .007) and disgust-
ing (r = .64, p < .001) scenes (Figure 6). In controls we found a
positive relationship between disgust propensity and intensity
rating of disgusting scenes (r = .64, p < .001).
Our results point to a general discrimination impairment of
negative facial emotions in MD patients. This is contrary to
several previous studies that report emotion-specific deficits in
this mental disorder (e.g., Mendlewicz et al., 2005). Patients
showed lower classification accuracy than controls for all nega-
tive facial emotions with exception of fear. Fearful faces were
poorly recognized also in healthy controls, indicating high task
difficulty. The patients’ lower allocation accuracy was due to
higher intensity ratings of all negative non-target emotions.
But patients rated negative non-target emotions more intense
also in happy faces, showing a negativity bias that cannot be
explained by task difficulty. Happiness usually yields greatest
accuracy, while negative emotions, especially fear, are less
easily discriminable (Du & Martinez, 2013; Ekman &
Friesen, 1976; Johnston et al., 2001). Our data are consistent
with Surguladze et al., (2004) who reported a response bias in
patients with depressive disorder judging happy faces as less
happy compared to healthy volunteers.
DisgustAngerFearSadness HappinessSurprise
p < 0.001
p < 0.001p < 0.001p < 0.001
p = 0.063
p < 0.001
Figure 3.
Group comparison of the mean scores (standard errors) for intensity
ratings of neutral scenes.
Happiness in happy−inducing scenes
Depression severity
r = −0.467
p = 0.019
Happiness in neutral scenes
r = −0.547
p = 0.005
Sadness in fear−inducing scenes
Depression severity
r = 0.415
p = 0.039
Sadness in neutral scenes
r = 0.433
p = 0.031
Figure 4.
MD patients: Correlations between depression severity (BDI score) and happiness ratings in
happy-inducing and neutral scenes (above) respectively sadness ratings in fear-inducing and
neutral scenes (below).
Open Access 69
Disgust in fearful faces
Disgust propensity
r = 0.534
p = 0.002
Disgust in angry faces
r = 0.465
p = 0.010
Disgust in sad faces
Disgust propensity
r = 0.376
p = 0.041
Figure 5.
MD patients: Correlations between disgust propensity (mean QADS score) and disgust ratings in
fearful and angry faces (above), and in sad faces (below).
Disgust in fear−inducing scenes
Disgust propensity
r = 0.487
p = 0.007
Disgust in disgust−inducing scenes
r = 0.636
p < 0.001
Figure 6.
MD patients: Correlations between disgust propensity (mean QADS score) and disgust ratings
in fear-inducing and disgust-inducing scenes.
The patients’ negativity bias in judging happy faces provide
support for the mood congruency hypothesis (e.g., Leppänen et
al., 2004) that states people would respond in accordance with
their mood. Previous studies found out that subjects with de-
pressed mood tended to judge positive emotions as neutral and
neutral faces as negative (Csukly et al., 2008; Hale, 1998).
Patients’ negativity bias became also evident in the rating of
affective scenes. They perceived fear-inducing and disgust-
inducing pictures more intense than controls. In addition, they
discerned happy scenes sadder than controls. Dominated by
negative mood, patients with MD rated emotional stimuli more
negative and perceived a more negative mood also in other
Undoubtedly, depression is characterized by a rather negative
mood, with sadness being one key symptom. But as shown by
our data, patients with MD also display elevated anxiety, anxi-
ety sensitivity, disgust propensity, disgust sensitivity, and anger
suppression compared to controls. We found associations be-
Open Access
tween patients’ disgust propensity and disgust ratings of fearful,
angry and sad faces pointing to a disgust bias in patients with
elevated disgust proneness. Moreover, anger suppression was
negatively related to patients’ happiness experience. These
findings also coincide with the mood congruency effect.
Our emotion recognition data are in accordance with Asthana
et al. (1998) who found a general impairment of emotion de-
coding performance in patients with MD. Comparable to our
patient sample, patients in this study met the diagnostic criteria
of DSM-IV for major depressive disorder that is associated
with a general cognitive disorder. Patients of our study dis-
played reduced cognitive performance compared to controls.
Furthermore, we found a relationship between patients’ cogni-
tive performance and identification accuracy of angry and dis-
gusted facial expressions. These facial emotions are similar-
looking and therefore are particularly difficult to discriminate.
Patients rated higher intensities of disgust for angry faces than
Furthermore, severity of depressive symptoms was nega-
tively correlated to experience of happiness in pleasure-induc-
ing and neutral scenes. Depression severity was also positively
associated to experience of sadness in fearful and neutral
However, the degree of depressive symptoms had no effect
on decoding accuracy of affective faces. Previous findings are
contradictory. Milders et al. (2010) found no association be-
tween accuracy performance of sad faces and BDI ratings.
However, Hale (1998) reported a positive relationship between
the judgment of negative emotions in the facial expressions and
depression severity.
We also found no association between depression severity
and cognitive performance. It is possible that cognitive inter-
ference is associated with major depression but does not sig-
nificantly vary with its symptom severity.
Patients’ elevated negative perception of emotion-inducing
scenes and emotional facial expressions as well as their higher
intensity ratings of neutral stimuli are contrary to suggestions of
the ECI hypothesis. Rottenberg (2005) found that patients with
the most pronounced ECI showed the most severe depression.
Gender-specific differences of depressive symptoms, with
males showing more anger compared to females (e.g., Rutz et
al., 1997) were confirmed by our data. Males with MD reported
higher trait anger, and they also rated happy faces more nega-
tive than female patients. Although males and females with MD
did not differ in depression score, this difference in judgment of
happy faces may be the consequence of a more negative mood
state in males suffering from depression. Items of depression
inventories like the BDI do not sufficiently consider male de-
pression symptoms like anger. We also could replicate the
findings of previous studies that pointed a female advantage in
decoding affective faces (e.g., Cellerino et al., 2004). Women
in both groups displayed a better decoding of facial expressions
than males. Therefore, compared to men, women with MD had
an advantage by a less negative mood state and a more correct
estimate of other subjects’ feelings.
Results indicate that in depression treatment increased focus
should put to the association between negative mood bias and
social functioning. Especially in male depression this aspect
may be underestimated.
As a limitation of our study it must be stated that all patients
were treated with antidepressants and/or antipsychotics. Tranter
et al. (2009) could show that treatment of depressed patients
with antidepressants resulted in improved recognition accuracy
of disgusted, happy, and surprised facial expressions, being
strongest for disgust. But it remains unclear, if this effect is
directly associated with antidepressant treatment or a result of
mood improvement. Therefore, antidepressant treatment could
have resulted in a shift of our patients’ emotion decoding and
In conclusion, the present findings indicate that MD is char-
acterized by a broad impairment of emotion recognition, con-
cerning all negative facial emotions with exception of fear.
Further on, patients’ responses to happy faces suggest a nega-
tivity bias in perceiving other peoples’ facial expressions,
which also became evident in the perception of emotional
scenes. The negativity bias was stronger in male than female
patients. Depression severity was negatively related to experi-
ence of happiness, and positively associated to a sadness-bias in
fear-eliciting and neutral scenes. Compared to controls MD
patients showed a lower cognitive performance, which was
associated with identification accuracy of angry and disgusted
facial expressions. However, we found no association between
depression severity and decoding accuracy of facial emotions
and to patients’ cognitive performance. Thus, our findings on
negativity bias show accordance with the mood-congruency
hypothesis, suggesting facilitated processing of negative emo-
tional cues and deficient processing of positive emotional stim-
uli. By contrast, our outcomes are contrary to suggestions of the
ECI hypothesis, supposing decreased reactivity to all emotion
cues in individuals with depression, regardless of valence. Re-
sults indicate that in depression treatment increased focus
should put to the association between negative mood bias and
social functioning.
American Psychiatric Association (APA) (2000). Diagnostic and statis-
tical manual of mental disorders (4th ed.).Washington DC: Author.
Angst, J., Gamma, A., Gastpar, M., Lepine, J. P., Mendlewicz, J., &
Tylee, A. (2002). Gender differences in depression. Epidemiological
findings from the European DEPRES I and II studies. European Ar-
chives of Psychiatry and Clinical Neuroscience, 25 2, 20-29.
Asthana, H. S., Mandal, M. K., Khurana, H., & Haque-Nizamie, S. (1998).
Visuospatial and affect recognition deficit in depression. Journal of
Affective Disorders, 48, 57-62.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36,
Bylsma, L. M., Morris, B. H., & Rottenberg, J. (2008). A meta-analysis
of emotional reactivity in major depressive disorder. Clinical Psy-
chology Review, 28, 676-691.
Cellerino, A., Borghetti, D., & Sartucci, F. (2004). Sex differences in
face gender recognition in humans. Brain Research Bulletin, 63, 443-
Csukly, G., Czobor, P., Simon, L., & Takács, B. (2008). Basic emotions
and psychological distress: Association between recognition of facial
expressions and symptom checklist-90 subscales. Comprehensive Psy-
chiatry, 49, 177-183.
Drevets, W. C. (2001). Neuroimaging and neuropathological studies of
depression: Implications for the cognitive-emotional features of
mood disorders. Current Opinion in Neurobiology, 11, 240-249.
Du, S., & Martinez, A. M. (2013). Wait, are you sad or angry? Large
exposure time differences required for the categorization of facial
expressions of emotion. Journal of Vision, 13, 1-14.
Open Access 71
Ekman, P., & Friesen, W. V. (1976). Pictures of facial affect. Palo Alto,
CA: Consulting Psychological Press.
Hale, W. W. (1998). Judgment of facial expressions and depression persis-
tence. Psychiatry Research, 80, 265-274.
Hautzinger, M., Bailer, M., Worall, H., & Keller, F. (1994). Beck-
Depressionis-Inventar (BDI). Testhandbuch (2nd ed.). Bern: Hans Hu-
Ihl, R., Grass-Kapanke, B., Lahrem, P., Brinkmeyer, J., Fischer, S., &
Gaab, N. (2000). Development and validation of a test for early di-
agnosis of dementia with differentiation from depression (TFDD).
Fortschritte der Neurologie—P sychiatrie, 68, 413-422.
Johnston, P. J., Katsikitis, M., & Carr, V. J. (2001). A generalised defi-
cit can account for problems in facial emotion recognition in schizo-
phrenia. Biological Psychology, 58, 203-227.
Kockler, M., & Heun, R. (2002). Gender differences of depressive
symptoms in depressed and non-depressed elderly persons. Interna-
tional Journal of Geriatric Psychiatry, 17 , 65-72.
Kring, A. M., & Gordon, A. H. (1998). Sex differences in emotion:
Expression, experience, and physiology. Journal of Personality and
Social Psychology, 74, 686-703.
Lang, P. J., Bradley, M., & Cuthbert, B. (2001). International affective
picture system. Center for research in psychophysiology. Gainsville,
Florida: University of Florida.
Laux, L., Glanzmann, P., & Spielberger, C. D. (1981). State Trait Ang-
stinventar (STAI). Weinheim: Beltz Testgesellschaft.
Leppänen, J. M., Milders, M., Bell, J. S., Terriere, E., & Hietanen, J. K.
(2004). Depression biases the recognition of emotionally neutral
faces. Psychiatry Research, 128, 123-133.
Lundquist, D., Flykt, A., & Öhman, A. (1998). The Karolinska directed
emotional faces of emotion and attention. Lawrence, Stockholm: Ka-
rolinska Hospital.
Margraf, J. (1994). Mini-DIPS—Diagnostisches Kurz-interview bei psy-
chischen Störungen. Berlin, Heidelberg, New York: Springer.
Mendlewicz, L., Linkowski, P., Bazelmans, C., & Philippot, P. (2005).
Decoding emotional facial expressions in depressed and anorexic pa-
tients. Journal of Affective Disorders, 89, 195-199.
Mikhailova, E. S., Vladimirova, T. V., Iznak, A. F., Tsusulkovskaya, E.
J., & Sushko, N. V. (1996). Abnormal recognition of facial expres-
sion of emotions in depressed patients with major depression disor-
der and Schizotypal personality disorder. Biological Psychiatry, 40,
Milders, M., Bell, S., Platt, J., Serrano, R., & Runcie, O. (2010). Stable
expression recognition abnormalities in unipolar depression. Psy-
chiatry Research, 179, 38-42.
Persad, S. M., & Polivy, J. (1993). Differences between depressed and
nondepressed individuals in the recognition of and response to facial
emotional cues. Journal of Abnormal Psychology, 102, 358-368.
Rottenberg, J. (2005). Mood and emotion in major depression. Current
Directions in Psychological Science, 14, 167-170.
Rottenberg, J. (2007). Major depressive disorder: Emerging evidence
for emotion context insensitivity. In J. Rottenberg, & S. L. Johnson
(Eds.), Emotion and psychopathology: Bridging affective and clinical
science. Washington DC: American Psychological Society.
Rottenberg, J., Gross, J. J., & Gotlib, I. H. (2005). Emotion context
insensitivity in major depressive disorder. Journal of Abnormal Psy-
chology, 114, 627-639.
Rutz, W., Wålinder, J., Von Knorring, L., Rihmer, Z., & Pihlgren, H.
(1997). Prevention of depression and suicide by education and
medication: Impact of male suicidality. An update of the Gotland
study. International Journal of Psychiatry in Clinical Practice, 1,
Scheibe, S., Preuschhof, C., Cristi, C., & Bagby, R. M. (2003). Are
there gender differences in major depression and its response to an-
tidepressants? Journal of Affective Disorders, 75, 223-235.
Schienle, A., Stark, R., Walter, B., Blecker, C., Ott, U., & Sammer, G.
(2002b). The insula is not specifically involved in disgust processing:
An fMRI study. Neuroreport, 13, 2023-2026.
Schienle, A., Walter, B., Stark, R., & Vaitl, D. (2002a). A questionnaire
for the assessment of disgust sensitivity. Zeitschrift für Klinische
Psychologie und Psychotherapie, 31, 110-120.
Schwenkmezger, P., Hodapp, V., & Spielberger, C. D. (1992). State-trait-
Ärgerausdrucksinventar STAXI. Bern: Huber.
Surguladze, S. A., Young, A. W., Senior, C., Brebion, G., Travis, M. J.,
& Phillips, M. L. (2004). Recognition accuracy and response bias to
happy and sad facial expressions in patients with major depression.
Neuropsychology, 18, 212-218.
Tranter, R., Bell, D., Gutting, P., Harmer, C., Healy, D., & Anderson, I.
M. (2009). The effect of serotonergic and noradrenergic antidepres-
sants on face emotion processing in depressed patients. Journal of
Affective Disorders, 118, 87-93.
Van Overveld, W. J. M., DeJong, P. J., Peters, M. L., Cavanagh, K., &
Davey, G. C. L. (2006). Disgust propensity and disgust sensitivity:
Separate constructs that are differentially related to specific fears.
Personality and Individual D i fferences, 41, 1241-1252.
Winkler, D., Pjrek, E., & Kasper, S. (2005). Anger attacks in depress-
sion—Evidence for a male depressive syndrome. Psychotherapy and
Psychosomatics, 74, 303-307.