75

Graduate Student Journal of Psychology
2022, Vol. 19

Copyright 2022 by the Department of Counseling and Clinical Psychology
Teachers College, Columbia University

Mask-wearing and Facial Emotion Recognition:            
A Preliminary Analysis of the Relevance of Depressive 

Symptoms
Mateus Morante Mazzaferro, 

Teachers College, Columbia University

Depression is a psychiatric condition that impairs the life of millions of people around the globe. Previous re-
șǓƺȖljǦ� Ǧƺș� șǦȅΛǿ� ȠǦƺȠ� ǏǓȒȖǓșșǓǏ� ǩǿǏǩΚǩǏȣƺǹș� ȠǓǿǏ� Ƞȅ� ȒȖǓșǓǿȠ� ǏǓЙljǩȠș� ǩǿ� ǟƺljǩƺǹ� ǓǾȅȠǩȅǿ� ȒǓȖljǓȒȠǩȅǿঀ� 'ȅȖ� ǩǿșȠƺǿljǓॹ�
perception accuracy may be reduced and biases in perceived intensity may be enhanced. Mask-wearing practic-
es initiated in response to the COVID-19 pandemic have become a new social norm often enforced by local man-
dates. Preliminary studies have shown that mask-wearing may impair facial emotion recognition. In this study, we 
aimed at understanding how facial emotion recognition impairment interacts with depressive symptoms in a sam-
ȒǹǓ� ȅǟ�(ǓȖǾƺǿ� ƺǏȅǹǓșljǓǿȠș� ƺǿǏ� ƺǏȣǹȠș� শF�઀� �ॹࢲࢺ �ઔࢷࢶ ǟǓǾƺǹǓॹ�ǾǓƺǿ� ƺǠǓ� ȅǟ� �ࢸঀࢳࢴ ΡǓƺȖșষ� LjΡ� ȣȠǩǹǩΦǩǿǠ� ƺ�ǾǩΠǓǏেǓАǓljȠș�
linear regression analysis. We found evidence that mask-wearing may be a limiting factor for facial emotion rec-
ȅǠǿǩȠǩȅǿॹ� ƺș�ΛǓǹǹ� ƺș� ǟȅȖ� ǓǾȅȠǩȅǿ� ǩǿȠǓǿșǩȠΡ� ȖƺȠǩǿǠșঀ�/ȅΛǓΚǓȖॹ� ƺ� șǩǠǿǩЙljƺǿȠ� ƺșșȅljǩƺȠǩȅǿ�ȅǟ�ǏǓȒȖǓșșǩΚǓ� șΡǾȒȠȅǾș�ΛǩȠǦ�
these outcome variables was not detected. Still, larger sample sizes may have the potential to substantiate a trend 
toward an interaction of depressive symptoms and mask-wearing for the rating of happy faces. Future research 
should be committed to psychophysiological processes and to improving the quality of the stimulus material. 

Theoretical Background
Depression and Facial Emotion Recognition
 In 2017, the WHO issued a report estimating 
ȠǦƺȠॹ� ǠǹȅLjƺǹǹΡॹ� ȅΚǓȖ� �ࢱࢱࢴ Ǿǩǹǹǩȅǿ� ȒǓȅȒǹǓ� șȣАǓȖ� ǟȖȅǾ�
depression, equivalent to 4.4% of the world’s popula-
tion (WHO, 2017). A review from 2014 shows that 
the average 12-month prevalence estimate of DSM-IV 
major depressive episodes varies from 5.5% to 5.9% 
in most countries (Kessler et al., 2014). In the U.S. 
alone, around 30 million adults have met the criteria 
for major depressive disorder at least once in their life-
time (Haro et al., 2006). The core symptoms of clin-
ǩljƺǹǹΡ� ȖǓǹǓΚƺǿȠ� ǏǓȒȖǓșșǩȅǿ� ǩǿljǹȣǏǓ� ЙȖșȠ� ƺǿǏ� ǟȅȖǓǾȅșȠ�
a persistently depressed mood and loss of interest or 
pleasure in almost all activities (APA, 2013). In ad-
ǏǩȠǩȅǿॹ� ȠǦǓȖǓ� ƺȖǓ� șǓΚǓȖƺǹ�ȅȠǦǓȖ� ƺАǓljȠǩΚǓॹ� ljȅǠǿǩȠǩΚǓॹ� șȅ-
matic, and motivational-behavioral features that play 
a role in depression, such as sleep disturbances, a sense 
of guilt, and impaired executive function (Berking 
& Rief, 2011). The presentation of each symptom 
can vary greatly in nature and intensity so the over-
ƺǹǹ� ȒǩljȠȣȖǓ� ljƺǿ� ǏǩАǓȖ� ljȅǿșǩǏǓȖƺLjǹΡ� ǟȖȅǾ� ljƺșǓ� Ƞȅ� ljƺșǓঀ
 Early theories of depression, such as Beck’s (1976) 
schema and Bower’s (1981) network theory, proposed 
that, in depressive patients, cognitive biases operate 
throughout information processing, including per-
ception, attention, and memory. Beck (1976) propos-
es that those with depression develop cognitive distor-
tions and tend to overlook positive attributes of reality 
and selectively attend to the negative. The associative 
network theory of emotion and memory, outlined by 

Bower (1981), suggests that depressed mood leads to 
biases favoring the perception of mood-congruent in-
ǟȅȖǾƺȠǩȅǿঀ��ȅȠǦ�ǾȅǏǓǹș�ƺșșȣǾǓ�ȠǦƺȠ�ƺАǓljȠǓǏ�ǩǿǏǩΚǩǏ-
uals selectively process information related to sadness, 
loss, and failure. Furthermore, poor interpersonal 
relationships have been proposed as a critical aspect 
in both the etiology and maintenance of depression 
(Finch & Zautra, 1992; Platt et al. 2013). Impaired 
emotion recognition may contribute to the interper-
șȅǿƺǹ� ǏǩГljȣǹȠǩǓș� ƺǿǏ� ƺΚȅǩǏƺǿljǓ� LjǓǦƺΚǩȅȖ� șǓǓǿ� ǩǿ� ǏǓ-
ȒȖǓșșǓǏ� ȒƺȠǩǓǿȠș� শXǓȖșƺǏ�૭� XȅǹǩΚΡॹ� �ষঀࢴࢺࢺࢲ ^ǩǿljǓ� ǏǓЙ-
cits in emotion recognition could contribute to the 
maintenance of depressive symptoms, investigating 
the relationship between these two variables has im-
portant implications for both existing and the devel-
opment of novel cognitive-behavioral interventions.
 For a long time now, research has shown that peo-
ȒǹǓ� șȣАǓȖǩǿǠ� ǟȖȅǾ�ƺАǓljȠǩΚǓ�ǏǩșȅȖǏǓȖș�ǦƺΚǓ�ǏǩГljȣǹȠǩǓș�
correctly interpreting human emotions from the per-
ception of facial expressions. In adults, the facial emo-
tion recognition (FER) impairments have been associ-
ated, for example, with bipolar disorder (Derntl et al. 
2009) and also very commonly with depression (e.g., 
Demenescu et al., 2010; Bistricky et al 2011, Rubinow 
૭�XȅșȠॹࢳࢺࢺࢲ�ষঀ��/ȅΛǓΚǓȖॹ�ЙǿǏǩǿǠș�ǩǿ�ȠǦǩș�ƺȖǓƺ�șǓǓǾ�Ƞȅ�
vary strongly based on symptom severity, psychiatric 
comorbidity, and the nature of stimuli (Bistricky et 
al., 2011). A meta-analysis conducted by Dalili et al. 
(2015) using 22 independent samples found signif-
ǩljƺǿȠ� ȖǓljȅǠǿǩȠǩȅǿ� ǏǓЙljǩȠș� ǩǿ� ǏǓȒȖǓșșǓǏ� ȒƺȖȠǩljǩȒƺǿȠș�
in all basic emotions except sadness. But the main 



76

feature of facial emotion recognition in depression is 
ǿȅȠ�ƺ�ǏǓЙljǩȠ�ǩǿ�ȠǦǓ�ƺLjǩǹǩȠΡ�Ƞȅ�ǩǏǓǿȠǩǟΡ�ǓǾȅȠǩȅǿș�ȒǓȖ�șǓॹ�
it is rather a bias in emotional valence rating. Many 
studies have reported a negative bias in depressed pa-
tients (e.g., Schmid & Schmid Mast, 2010; Milders et 
ƺǹঀॹࢱࢲࢱࢳ�ॹ�?ƺǿ�ǓȠ�ƺǹঀॹࢵࢱࢱࢳ�ষॹ�ΛǦǩljǦ�ǾǓƺǿș�ȠǦǓ�ƺАǓljȠǓǏ�
populations sometimes perform better in recognizing 
sad faces, tend to interpret neutral or ambiguous faces 
as sad more often and label happy faces as less happy.
 The literature presents similar evidence for ad-
olescents and younger children. Nyquist & Luebbe, 
শࢱࢳࢱࢳষ�ȖǓΚǩǓΛǓǏࢷࢳ��șȠȣǏǩǓș�ƺǿǏ�ǩǏǓǿȠǩЙǓǏ�ȠǦȖǓǓ�ȠȖǓǿǏș�
for FER in youth that were either depressed or quali-
ЙǓǏ�ƺș�ǦǩǠǦেȖǩșǷ�ȒȅȒȣǹƺȠǩȅǿș�ǟȅȖ�ǏǓȒȖǓșșǩȅǿॸ�șǓǿșǩȠǩΚǩȠΡ�
to sadness (higher accuracy in recognizing sadness and 
biased rating towards sad faces), over-perception of 
anger (falsely recognizing ambiguous or neutral faces 
as angry), and under-perception of happiness (lower 
accuracy in recognizing happiness and biased rating 
towards happiness). Investigating features of depres-
sion in families, Lopez-Duran et al. (2013) observed 
sensitivity to sadness as a potential mechanism of risk 
among boys at familial risk for depression. In a similar 
vein, Kluczniok et al. (2015) found that children from 
mothers with remitted depression displayed depres-
sive-like emotion recognition bias that correlated with 
their mothers’ performance. As suggested by the liter-
ature, emotion recognition bias is a common correlate 
of depressive symptomatology in all age groups. It is, 
ȠǦǓȖǓǟȅȖǓॹ�ǓșșǓǿȠǩƺǹ�Ƞȅ�ǩǿΚǓșȠǩǠƺȠǓ�ǦȅΛ�ȠǦǓșǓ�ǏǓЙljǩȠș�ljƺǿ�
impair social functioning and contribute to the patho-
logical behavior of clinically relevant populations.

Masks and Emotion Recognition
 Since March of 2020, the COVID-19 pandemic 
Ǧƺș�LjȖȅȣǠǦȠ�șǩǠǿǩЙljƺǿȠ�ljǦƺǿǠǓș�Ƞȅ�ȠǦǓ�ǏƺǩǹΡ�ǹǩΚǓș�ȅǟ�ǓΚ-
eryday citizens. To help stop the spread of the virus, 
the wearing of a protective face mask, which has been 
shown to drastically reduce viral transmission (Cheng 
et al., 2020), has become a largely accepted norm (e.g., 
Cheng et al., 2020; Hong et al., 2020; Howard et al., 
2020). Albeit necessary for keeping the virus under 
control, masks may have important psychological im-
ȒƺljȠș�ȅǿ�șȅljǩƺǹ� ǩǿȠǓȖƺljȠǩȅǿșॹ�ǟȅȖ�ǓΠƺǾȒǹǓॹ�LjΡ�ǾȣЖǩǿǠ�
speech and other forms of communicative vocalization 
(Mheidly et al., 2020). Social distancing and mask-wear-
ing can also impair interpersonal communication by 
limiting physical touch and body language, as well as 

by hindering visibility of the lower half of the face.
 Regarding facial emotional recognition, recent 
studies have found that mask-wearing can make emo-
tions less well recognized or interpreted (e.g., Carbon, 
2020a; Carbon, 2020b), faces less well-remembered, 
or critically impair holistic processing (Freud et al., 
2020). New data, nonetheless, shows that emotions 
can still be recognized both in adults (e.g., Calbi et 
al., 2021; Kastendieck, et al., 2021) and in children 
(e.g., Ruba & Pollak, 2020), even if the intensity of 
the emotion decreases (Kastendieck et al., 2021). 
 Studies also show that sad and angry faces cause 
ǾȅȖǓ� ЙΠƺȠǩȅǿ� ȅǿ� ȠǦǓ� ǓΡǓșॹ� ΛǦǩǹǓ� ǦƺȒȒΡ� ǟƺljǓș� ƺȠȠȖƺljȠ�
ǾȅȖǓ� ЙΠƺȠǩȅǿ� ȅǿ� ȠǦǓ�ǾȅȣȠǦॹ� ǏȣǓ� Ƞȅ� ȠǦǓ� ΛƺΡ� ǏǩАǓȖ-
ǓǿȠ� ǓǾȅȠǩȅǿș� ǾȅLjǩǹǩΦǓ� ǏǩАǓȖǓǿȠ� ǟƺljǩƺǹ� ǾȣșljȣǹƺȠȣȖǓ�
(Eisenbarth, 2011; Schurgin, 2014) - curiously, psy-
ljǦǩƺȠȖǩlj� ȒƺȠǩǓǿȠș� șǓǓǾ� Ƞȅ� ǦƺΚǓ� ǹǓșș� ЙΠƺȠǩȅǿ� ȅΚǓȖƺǹǹ�
(Eisenbarth, 2011). These facts may account for 
ЙǿǏǩǿǠș� ȖǓȒȅȖȠǩǿǠ� ȠǦƺȠ�ΛǦǓǿ�ȅǿǹΡ� ȠǦǓ�ȣȒȒǓȖ�ȒƺȖȠ� ȅǟ�
the face is visible, participants perceive and recog-
nize negative emotions, like anger and fear, better 
than positive ones (Marta et al., 2021; Fischer et al., 
2012). In sum, it is possible that mask-wearing does 
not impair or even relatively increase the perception 
of negative facial expressions and, at the same time, 
diminishes the perception of positive ones (Spitzer, 
�ষঀࢱࢳࢱࢳ eǦǓȖǓǟȅȖǓॹ� ǩȠ� ǾƺΡ� ǦƺΚǓ� șǩǠǿǩЙljƺǿȠ� ǓАǓljȠș� ȅǿ�
day-to-day activities that rely on interpersonal com-
munication and social interaction (Freud et al., 2020).
� eǦǓ� ǟƺljȠȅȖș� ȠǦƺȠ� ǩǿМȣǓǿljǓ� ǟƺljǩƺǹ� ǓǾȅȠǩȅǿ� ȖǓljȅǠ-
nition in laboratory settings remain unclear when 
individuals wear masks. The type of stimuli, for in-
stance, is discussed as a possible moderator: so far, 
mainly adult faces have been used (e.g., Ruba & Pol-
lak, 2020), whereas in real-life interactions children 
as well as adolescents and adults represent their coun-
terpart. For this reason, emotional child and adult 
faces will be used in the current study. Moreover, 
static photos are mostly used (e.g., Carbon, 2020a, 
Carbon, 2020b; Ruba & Pollak, 2020), whereas clos-
er proximity to reality can be assumed when videos 
are used (Rymarczyk et al., 2016). For that reason, 
this study is going to use video stimulus material.

Research Gap and Current Research Purpose
 As to this point, we are unaware of any research 
ȒȖȅǴǓljȠș� ȠȖΡǩǿǠ� Ƞȅ� Йǹǹ� ȠǦǓ� ǿǓΛǹΡ� ljȖǓƺȠǓǏ� ȖǓșǓƺȖljǦ� ǠƺȒ�
of how mask-wearing interacts with psychopatholo

MAZZAFERRO



77

MASK-WEARING AND FACIAL EMOTION RECOGNITION

gy in impairing social function. Considering that the 
COVID-19 pandemic has introduced a new norm in 
face-to-face interactions (which now almost always 
include the wearing of face masks), research on FER 
must rise to the challenge of understanding how 
this phenomenon can impact the nature of social ex-
ljǦƺǿǠǓॹ� ǓșȒǓljǩƺǹǹΡ� ǟȅȖ�ȒȅȒȣǹƺȠǩȅǿș� ȠǦƺȠ� ƺǹȖǓƺǏΡ� șȣАǓȖ�
ǟȖȅǾ� șȅljǩƺǹ� ǏǩГljȣǹȠǩǓșॹ� ǹǩǷǓ� ƺ� șǩǠǿǩЙljƺǿȠ� ƺǾȅȣǿȠ� ȅǟ�
patients diagnosed with mental disorders (Lehmann 
et al., 2019) – since these people may be particular-
ly challenged by the pandemic (Druss, 2020). This 
preliminary study was conducted as a part of the Ge-
sichter lesen1 project, an initiative seeking to investi-
ǠƺȠǓ� ǦȅΛ� ǾƺșǷেΛǓƺȖǩǿǠ� ǏǩАǓȖǓǿȠǩƺǹǹΡ� ǩǾȒƺljȠș� șȅljǩƺǹ�
exchange in children, adolescents, and adults. Here, 
we aim to understand how depressive symptoms in-
ȠǓȖƺljȠ�ΛǩȠǦ�ȠǦǓ�ǓАǓljȠș�ȅǟ�ǾƺșǷেΛǓƺȖǩǿǠ�ǩǿ�ǟƺljǩƺǹ�ǓǾȅ-
tion recognition processes in adolescents and adults.
 In the future, the Gesichter lesen project intends 
to analyze data from multiple psychopathological vari-
ables and attitudes in emotion recognition and facial 
mimicry in children and adolescents. Considering 
that children’s and early adolescents’ development are 
ǹƺȖǠǓǹΡ� ǩǿМȣǓǿljǓǏ�LjΡ�ȠǦǓ�șȅljǩƺǹ�ljȅǿȠǓΠȠ�ΛǩȠǦǩǿ�ΛǦǩljǦ�
they interact (Richards & Light, 1986), the impact 
ȅǟ�ǾƺșǷেΛǓƺȖǩǿǠ�ǾǩǠǦȠ�LjǓ�ȒƺȖȠǩljȣǹƺȖǹΡ� șǩǠǿǩЙljƺǿȠ� ǟȅȖ�
them, notably if they struggle with any form of psycho-
ȒƺȠǦȅǹȅǠΡঀ�'ȅȖ�ȠǦǩș�ЙȖșȠ�ǓΠȒǓȖǩǾǓǿȠॹ�ΛǓ�șǓǹǓljȠǓǏ�ƺ�ǾȅȖǓ�
accessible sample of adults recruited online and decided 
to focus on a single psychopathology-related construct 
(depressive symptoms, continuous) and mask-wearing 
(binary) as independent variables. We aim to examine 
how the wearing of face masks interacts with the biases 
in facial emotion reading found to be associated with 
depressive symptoms in adolescents and adults (see 
above). The dependent variables of interest are, there-
fore, emotion recognition and emotion intensity rating.

Hypotheses
� eǦǓ�ЙȖșȠ�ȅȣȠljȅǾǓ�LjǓǩǿǠ�ǓΠƺǾǩǿǓǏ�ǩș�ǓǾȅȠǩȅǿ�ȖǓlj-
ognition.  We hypothesize that participants’ emotion 
recognition of anger, sadness, and happiness is nega-
ȠǩΚǓǹΡ�ƺАǓljȠǓǏ�LjΡ�ȠǦǓ�ȒȖǓșǓǿljǓ�ȅǟ�ƺ�ǟƺljǓ�ǾƺșǷ�শ/ࢲঀࢲষঀ�2ǿ�
addition, recognition of anger, sadness, and happiness 
are associated with the participant’s depressive symp-

toms (H1.2). In line with the presented literature, 
we expect sadness and anger to be overperceived and 
happiness to be underperceived by participants with 
elevated depressive symptoms. Finally, there is an in-
ȠǓȖƺljȠǩȅǿ� ǓАǓljȠ� LjǓȠΛǓǓǿ� ȠǦǓ� ȒȖǓșǓǿljǓ� ȅǟ� ƺ� ǟƺljǓ�ǾƺșǷ�
and the participant’s depressive symptoms in emotion 
recognition for anger, sadness, and happiness (H1.3).
 For the second outcome, emotion intensity rating, 
we hypothesize that all emotions (anger, sadness, and 
happiness) are rated as less intense in masked faces than 
emotions in unmasked faces (H2.1). Additionally, the 
intensity rating of all perceived emotions (anger, sad-
ness, and happiness) is associated with the participant’s 
depressive symptoms (H2.2). In line with the present-
ed literature, we expect sadness and anger to be rated 
more intensively and happiness to be rated less inten-
sively by participants with elevated depressive symp-
ȠȅǾșঀ�@ƺșȠǹΡॹ�ȠǦǓȖǓ�ǩș�ƺǿ�ǩǿȠǓȖƺljȠǩȅǿ�ǓАǓljȠ�LjǓȠΛǓǓǿ�ȠǦǓ�
presence of a face mask and the participant’s depres-
sive symptoms in the rating of emotion intensity for 
all emotions (anger, sadness, and happiness) (H2.3). 

Methods
Sample. Inclusion criteria for the online experiment 
ljȅǿșǩșȠǓǏ�ȅǟ�ljȣȖȖǓǿȠǹΡ�ǹǩΚǩǿǠ�ǩǿ�(ǓȖǾƺǿΡॹ�șȒǓƺǷǩǿǠ�Мȣ-
ent German, and being at least 14 years old. Recruiting 
was conducted between January and June 2021 through 
ǾȣǹȠǩȒǹǓ�ljǦƺǿǿǓǹșॹ�ǩǿljǹȣǏǩǿǠ�XȖȅǹǩЙlj2ॹ�ǏǩАǓȖǓǿȠ�șȅljǩƺǹ�
media platforms (Facebook and Twitter), and student 
recruiting resources from the Institute of Psychology 
at the Humboldt University Berlin, which consist of 
ǾƺǩǹǩǿǠ� ǹǩșȠș� ƺǿǏ� МΡǓȖ� ǏǩșȠȖǩLjȣȠǩȅǿঀ� XƺȖȠǩljǩȒƺǿȠș� ȖǓ-
ceived 5 EUR on average for their participation, which 
ǩș� ljǹƺșșǩЙǓǏ� ƺș� ǠȅȅǏ� ȒƺΡǾǓǿȠ� ǩǿ� XȖȅǹǩЙljঀ��ȅǾȒǓǿșƺ-
Ƞǩȅǿ�Λƺș�ljƺȖȖǩǓǏ�ȅȣȠ�ǓǩȠǦǓȖ�Κǩƺ�XȖȅǹǩЙljॹ�LjƺǿǷ�ǏǓȒȅșǩȠॹ�ȅȖ�
as test person points for undergraduate credit. A total 
of 108 participants (adults and adolescents 14 years 
and over) completed the online experiment and gave 
their consent via Sosci Survey. After the exclusion of 
subjects that did not complete psychopathology mea-
sures or dropped out before the end of the experiment, 
ȠǦǓ�Йǿƺǹ�șƺǾȒǹǓ�ljȅǿșǩșȠǓǏ�ȅǟ�ǏƺȠƺ�ǟȖȅǾࢲࢺ��ȒƺȖȠǩljǩȒƺǿȠș�
(51 women, 56%) with a mean age of 32.7 years (SD 
= 15.6, range 14-71), who reported having seen the 
video stimuli and for whom at least one video upload 

 1 1 For more information on the Gesichter lesen project, see https://www.kinderprojekte-psychologie.de/projekte/gesichter-lesen/ For more information on the Gesichter lesen project, see https://www.kinderprojekte-psychologie.de/projekte/gesichter-lesen/
 2  2 'ȅȖ�ǾȅȖǓ�ǩǿǟȅȖǾƺȠǩȅǿ�ȅǿ�ȠǦǓ�ȖǓljȖȣǩȠǩǿǠ�ǓǿǠǩǿǓ�XȖȅǹǩЙljॹ�șǓǓ�ǦȠȠȒșॸইইΛΛΛঀȒȖȅǹǩЙljঀljȅইॾǦȅΛেǩȠেΛȅȖǷș'ȅȖ�ǾȅȖǓ�ǩǿǟȅȖǾƺȠǩȅǿ�ȅǿ�ȠǦǓ�ȖǓljȖȣǩȠǩǿǠ�ǓǿǠǩǿǓ�XȖȅǹǩЙljॹ�șǓǓ�ǦȠȠȒșॸইইΛΛΛঀȒȖȅǹǩЙljঀljȅইॾǦȅΛেǩȠেΛȅȖǷș



78

MAZZAFERRO

was successful. We did not collect information on race, 
ethnicity, income level, or educational attainment.

Material. For the diagnostic material, we selected 
the PHQ-9 for the recording of depressive symptoms 
(Patient Health Questionnaire - 9; Richardson et al., 
2010). The PHQ-9 is a versatile instrument for screen-
ing and measuring the severity of depression using only 
9 items. It incorporates diagnostic criteria from the 
DSM-IV and other important major depressive symp-
toms into a short self-report tool (Kroenke et al., 2001). 
This instrument has been shown to have adequate in-
ternal consistency (>0.8; Kim & Lee, 2019; Titov et al., 
2010), a one-factor replicable structure (Kim & Lee, 
2019; Titov et al., 2010), and satisfactory convergent 
validity (Kim & Lee, 2019; Beard et al., 2016).  In this 
study, the PHQ-9 had a Cronbach’s alpha of 0.86.
'ȅȖ� ƺ� ǏǩАǓȖǓǿȠ� șǓȠ� ȅǟ� ƺǿƺǹΡșǓș� ǿȅȠ� ȖǓǹǓΚƺǿȠ� Ƞȅ� ȠǦǓ�
present study, the experiment materials also includ-
ed the Inclusion of Other in the Self scale (IOS, 
Aron et al., 1992), the GAD-7 (Spitzer et al., 2006), 
and a short questionnaire on the participant’s at-
titudes towards mask-wearing (An et al, 2021).
 As video stimuli, we selected validated child and 
adult faces from the online Radbound Faces Data-
base (Bijsterbosch et al., 2020), which were presented 
as videos through a morphing process created using 
FantaMorph5. Facial expressions ranged from neutral 
to emotional. Selected examples of the stimuli are pre-
sented in Figure 1. A surgical mask was added to half 
of the stimuli with the help of the video editing soft-
ΛƺȖǓ�@Ǔǿș�^ȠȣǏǩȅ�LjΡ�^ǿƺȒ�2ǿljঀ�eǦǓ�Йǿƺǹ�ΚǩǏǓȅș�ΛǓȖǓ�
ȠǦǓǿ� ljȅǾȒȅșǓǏ� ΛǩȠǦ� �ǏȅLjǓ� �ǟȠǓȖ�АǓljȠș� ƺș� ǟȅǹǹȅΛșॸ�
-�șǓljȅǿǏș�ǟȅȖ�ȠǦǓ�ǾȅȖȒǦ�șǓࢶ��șǓljȅǿǏș�ЙΠƺȠǩȅǿ�ljȖȅșșॹࢶঀࢲ
quence, 1.5 seconds post-stimuli neutral background. 
For the sad videos, the morph sequence is 6 seconds 
long instead of 5 because research has shown that the 
onset and development of a sad facial expression take 
somewhat longer than happiness or anger (Fayolle 
& Droit-Volet, 2014). The pool of stimuli consists 
of 48 videos of 16 agents (8 adults, 8 children; 50% 
of all agents being female) showing three selected 
emotions (happiness, anger, and sadness). Half of all 
ΚǩǏǓȅș� șǦȅΛ�ƺǠǓǿȠș�ΛǓƺȖǩǿǠ� ǟƺljǓ�ǾƺșǷșঀ��� șȣȒǓȖЙljǩƺǹ�
qualitative evaluation of how well a face mask would 
ЙȠ�ȠǦǓ�ǾȅȖȒǦ�Λƺș�ljȅǿǏȣljȠǓǏ�Ƞȅ�ǏǓȠǓȖǾǩǿǓ�ΛǦǩljǦ�ΚǩǏ-
eos would feature face masks. Each agent appears on 
3 videos (one for each emotion), with either 1 out of 

3 or 2 out of 3 videos being masked. The stimuli were 
randomized using a block design to prevent the same 
gender from being presented more than twice in a row.

Procedure. The study was conducted entirely online. 
For adolescents 14 years and over and adults, partic-
ǩȒƺȠǩȅǿ�Λƺș�ljȅǾȒǹǓȠǓǹΡ�șǓǹǟেșȣГljǩǓǿȠॹ�ƺș�ƺȒȒȖȅΚǓǏ�LjΡ�
the Ethics Committee of the Humboldt University 
Berlin. Participants were informed that they would 
be required to watch 48 short videos, that the task 
would take about 40 minutes, and that participation 
in the experiment would only be possible if they had 
a webcam-enabled computer/laptop and agree to a 
webcam recording of their face during the experi-
ment. Informed consent included standard details 
ȅǿ� ljȅǾȒǓǿșƺȠǩȅǿॹ� ljȅǿЙǏǓǿȠǩƺǹǩȠΡॹ� ƺǿǏ� ljȅǿȠƺljȠ� ǩǿǟȅȖ-
mation. Participants who agreed to participate were 
instructed to set up their webcam to allow record-
ǩǿǠॹ�Ƞȅ�ƺȖȖƺǿǠǓ�șȣГljǩǓǿȠ�ǹǩǠǦȠǩǿǠॹ�ƺǿǏ�Ƞȅ�ȖǓǟȖƺǩǿ�ǟȖȅǾ�
eating or covering their face during the experiment.
� eǦǓ� ЙȖșȠ� șǓȠ� ȅǟ� ȕȣǓșȠǩȅǿș� ƺșǷǓǏ� ǟȅȖ� ǏǓǾȅǠȖƺȒǦ-
ic information. Participants then watched the video 
stimuli while their facial activity was recorded. Fol-
lowing each video, using 7-point Likert-scales, par-
ticipants rated the targets’ emotion expressions using 
ƺǿ� ǓǾȅȠǩȅǿ� ȒȖȅЙǹǓ� শǦƺȒȒǩǿǓșșॹ� șƺǏǿǓșșॹ� ǟǓƺȖॹ� ƺǿǠǓȖॹ�
disgust, and surprise) and were asked to indicate how 
close they felt to the person shown using the IOS Scale 
and how much they would like to meet the displayed 
ƺǠǓǿȠঀ�'ȅǹǹȅΛǩǿǠ�ȠǦǓ�ΚǩǏǓȅ�ȠƺșǷॹ�ȒƺȖȠǩljǩȒƺǿȠș�ЙǹǹǓǏ�ȅȣȠ�

Figure 1: On the left, a woman with a mask expressing happiness. 
On the right a child without a mask expressing anger.



79

MASK-WEARING AND FACIAL EMOTION RECOGNITION

psychopathology measures (PHQ-9 and GAD-7) and 
were asked about their attitudes toward mask-wearing 
ƺǿǏ�ȕȣǓșȠǩȅǿǓǏ�ƺLjȅȣȠ�ǦȅΛ�ȠǦǓ�ȒƺǿǏǓǾǩlj�ǦƺǏ�ƺАǓljȠǓǏ�
their lives with the scale adapted from An et al. (2021). 
Finally, participants were given the opportunity to 
opt-out by requiring their videos to be erased (no par-
ticipant used this option), informed about the purpose 
of the experiment, thanked for their participation, 
and received the necessary instructions for payment.

Statistical Analysis. Hypothesis for both outcome 
variables (emotion recognition and emotion inten-
sity rating) were analyzed with linear mixed models 
(LMMs) using the statistical programming language 
[�ƺǿǏ�ǩȠș�ȒƺljǷƺǠǓș�ǹǾǓࢵ�ƺǿǏ�ǹǾǓȖeǓșȠঀ�eǦǓ�Ǿƺǩǿ�ǓАǓljȠ�
of mask, depression score and the interaction between 
the two were estimated by random intercept/ random 
șǹȅȒǓ�ǾȅǏǓǹșঀ� �ȅǦǓǿঢ়ș� Ǐ� ǓАǓljȠ� șǩΦǓș� ΛǓȖǓ� ljȅǾȒȣȠǓǏ�
ǟȅȖ� ǓΚǓȖΡ� șǩǠǿǩЙljƺǿȠ� ljȅǓГljǩǓǿȠঀ� eǦǓ� ǏƺȠƺ� șȠȖȣljȠȣȖǓ�
characterized a repeated measures design since we 
collected many observations on the same variables 
ǟȖȅǾ� ǓΚǓȖΡ� ȒƺȖȠǩljǩȒƺǿȠঀ� eǦȣșॹ� ȒƺȖȠǩljǩȒƺǿȠ� ǩǏǓǿȠǩЙljƺ-
tion (ID) was included in the model as a cluster vari-
able. Hence, the full model was computed as follows: 
Outcome ~ Mask * Depression + (1+Mask+Depres-

sion | ID).

Results
� �ǾȅȠǩȅǿ� ǩǿȠǓǿșǩȠΡ� ȖƺȠǩǿǠঀ� eǦǓ� Ǿƺǩǿ� ǓАǓljȠ� ȅǟ�
the mask factor on emotion intensity rating can be 
observed in the descriptive bar plot in Figure 2. The 
mask factor (after controlling for depression score) 
ΡǩǓǹǏǓǏ�șǩǠǿǩЙljƺǿȠ�LjǓȠƺ�ljȅǓГljǩǓǿȠș�ǟȅȖ�ǦƺȒȒǩǿǓșș�শݾ�઀�
-1.57, p < 0.001, Cohen’s d = -0.41, CI95% = [-0.47; 
েࢱঀࢶࢴযষॹ�ƺǿǏ�ǟȅȖ�șƺǏǿǓșș�শݾ�઀�েࢱঀࢺࢺॹ�Ȓ�ઃࢱ�ঀࢲࢱࢱॹ��ȅǦǓǿঢ়ș�
d = -0.24, CI95% = [-0.28; -0.19]), but not for anger. 
eǦǓ�Ǿƺǩǿ�ǓАǓljȠ�ǟȅȖ�ǦƺȒȒǩǿǓșșॹ�ǦȅΛǓΚǓȖॹ�Λƺș�ȕȣƺǹǩЙǓǏ�
by the interaction mask*depression score and thus 
Λƺș� ǩǿȠǓȖȒȖǓȠǓǏ� ǩǿ� ȠǦǩș� ljȅǿȠǓΠȠঀ� eǦǓșǓ� ǓАǓljȠș� ΛǓȖǓ�
also tested via likelihood-ratio test model comparison 
of intercept-only models and the results were main-
ȠƺǩǿǓǏঀ� �ǓȒȖǓșșǩȅǿ� șljȅȖǓș� ǏǩǏ� ǿȅȠ� ΡǩǓǹǏ� șǩǠǿǩЙljƺǿȠ�
Ǿƺǩǿ�ǓАǓljȠ�ȖǓǠȖǓșșǩȅǿ�ljȅǓГljǩǓǿȠș�ǟȅȖ�ƺǿΡ�ȅǟ�ȠǦǓ�șǓǹǓljȠ-
ed emotions in the LMMs. It is important to mention 
ȠǦƺȠॹ� ǟȅȖ� șƺǏǿǓșșॹ� ƺ� ǿȅǿেșǩǠǿǩЙljƺǿȠ� șǾƺǹǹ� ȒȅșǩȠǩΚǓ� Ǔǟ-
ǟǓljȠ�ȅǟ�ǏǓȒȖǓșșǩΚǓ� șΡǾȒȠȅǾș�Λƺș�ȅLjșǓȖΚǓǏ� শݾ�઀ࢱ�ঀࢵࢱॹ�
p < 0.1, Cohen’s d = 0.11, CI95% = [-0.02; -0.24]).
 

� �ș�ǟȅȖ�ȠǦǓ�ǩǿȠǓȖƺljȠǩȅǿ�ǓАǓljȠșॹ�ȠǦǓ�ǾƺșǷॴǏǓȒȖǓșșǩȅǿ�
șljȅȖǓ�ǩǿȠǓȖƺljȠǩȅǿ�Λƺș�șǩǠǿǩЙljƺǿȠ�ƺȠ�ȠǦǓࢶ�ઔ�ǹǓΚǓǹ�ǟȅȖ�ǦƺȒ-
ȒǩǿǓșș�শݾ�઀�েࢱঀࢶࢱॹ�Ȓ�ઃࢱ�ঀࢶࢱॹ��ȅǦǓǿঢ়ș�Ǐ�઀�েࢱঀࢸࢱॹ��2ࢶࢺઔ�
= [-0.13; -0.01]) but not for sadness or anger. The sig-
ǿǩЙljƺǿȠ� ǩǿȠǓȖƺljȠǩȅǿ�ljȅǓГljǩǓǿȠș�ΛǓȖǓ�ȒǹȅȠȠǓǏ�Ƞȅ�ƺǹǹȅΛ�
ƺ� LjǓȠȠǓȖ� ǩǿȠǓȖȒȖǓȠƺȠǩȅǿ� ȅǟ� ȠǦǓ� ǓАǓljȠ� ƺǿǏ� ljƺǿ� LjǓ� șǓǓǿ�
in Figure 5. The model for happiness, including the 
mask, depression scores, and the interaction between 
the two produced a marginal R2 of 0.174 (calculated 
according to Nakagawa & Schielzeth, 2013). A mod-
ǓȖƺȠǩȅǿ�ƺǿƺǹΡșǩș�ȅǟ�ǏǓȒȖǓșșǩΚǓ�șljȅȖǓ� ǩǿ�ȠǦǓ�ǾƺșǷ�ǓАǓljȠ�
was conducted with the help of simple slopes (with the 
ȒƺljǷƺǠǓ�ǩǿȠǓȖƺljȠǩȅǿș�ǩǿ�[ষ�ƺȠ�ȠǦȖǓǓ�ǏǩАǓȖǓǿȠ�ȒȅǩǿȠș�ȅǟ�
the depression scale (-1SD, mean, +1SD). The mask 
ǓАǓljȠ�Λƺș�șǩǠǿǩЙljƺǿȠ�ȠǦȖȅȣǠǦȅȣȠ�ȠǦǓ�șȒǓljȠȖȣǾ�ȅǟ�ǏǓ-
pressive symptoms (p < 0.001). However, the 95% con-
ЙǏǓǿljǓ�ǩǿȠǓȖΚƺǹș�ȅǟ�ȠǦǓ�ljȅǓГljǩǓǿȠș�ǏǩǏ�ȅΚǓȖǹƺȒॹ�ȠǦƺȠ�ǩșॹ�
ȠǦǓ�șǩǾȒǹǓ�șǹȅȒǓș�ΛǓȖǓ�ǿȅȠ�șǩǠǿǩЙljƺǿȠǹΡ�ǏǩАǓȖǓǿȠ�ǟȖȅǾ�
ȅǿǓ� ƺǿȅȠǦǓȖঀ�eǦȣșॹ� ȠǦǓ�ǾȅǏǓȖƺȠǩȅǿ� ǓАǓljȠ� ljȅȣǹǏ� ǿȅȠ�
LjǓ� ǩǿȠǓȖȒȖǓȠǓǏ� ƺș� șǩǠǿǩЙljƺǿȠঀ� �șȠǩǾƺȠǓș� ǟȅȖ� ȠǦǓ� ǓАǓljȠ�
of mask at the three respective levels of depression 
ΛǓȖǓݾ��઀� েࢲঀࢲࢴ�ΛǩȠǦ��2ࢶࢺઔ�઀� মেࢲঀࢳࢷআ� েࢲঀࢱࢱয�ƺȠ� েࢲ^�ॹ�
�ݾ ઀� েࢲঀࢸࢶ� ΛǩȠǦ� �ઔࢶࢺ�2 ઀� মেࢲঀࢺࢸআ� েࢲঀࢷࢴয� ƺȠ� ȠǦǓ� ǾǓƺǿ�
ƺǿǏݾ��઀� েࢲঀࢵࢹ�ΛǩȠǦ��2ࢶࢺઔ�઀� মেࢳঀࢵࢲআ� েࢲঀࢴࢶয� ƺȠ�੼ࢲ^�

Emotion recognition.�eǦǓ�Ǿƺǩǿ�ǓАǓljȠ�ȅǟ�ȠǦǓ�ǾƺșǷ�ǟƺlj-
tor for emotion recognition can be observed in the de-
scriptive bar plot in Figure 3. A hit (=1) was registered 
when the intensity rating of the target emotion was 
higher than the intensity rating of all other emotions. 
A miss (=0) was registered when one of the distractor 
emotions was rated higher than or equal to the target 
emotion. The second outcome variable, emotion rec-
ognition, produced similar results to emotion intensi-
ȠΡ�ȖƺȠǩǿǠșঀ�eǦǓ�Ǿƺǩǿ�ǓАǓljȠ�ȅǟ�ǾƺșǷ�শƺǟȠǓȖ�ljȅǿȠȖȅǹǹǩǿǠ�
ǟȅȖ�ǏǓȒȖǓșșǩȅǿ�șljȅȖǓșষ�Λƺș�ȒȖǓșǓǿȠ�ǟȅȖ�ǦƺȒȒǩǿǓșș�শݾ�઀�
-0.17, p < 0.001, Cohen’s d = -0.27, CI95% = [-0.34; 
েࢱঀࢱࢳযষ�ƺǿǏ�șƺǏǿǓșș� শݾ�઀� েࢱঀࢺࢲॹ�Ȓ�ઃࢱ�ঀࢲࢱࢱॹ��ȅǦǓǿঢ়ș�Ǐ�
= -0.21, CI95% = [-0.26; -0.16]) but not for anger. 
Model comparison tests were consistent with this 
șǩǠǿǩЙljƺǿljǓ� ȒƺȠȠǓȖǿঀ� eǦǓ� ΚƺȖǩƺLjǹǓ� ǏǓȒȖǓșșǩΚǓ� șΡǾȒ-
ȠȅǾș�ǏǩǏ�ǿȅȠ�șǩǠǿǩЙljƺǿȠǹΡ�ȒȖǓǏǩljȠ�ȠǦǓ�ǓǾȅȠǩȅǿ�ȖǓljȅǠ-
nition of any of the emotions displayed in this study.  



80

MAZZAFERRO

Figures 2 (top) and 3 (bottom):  Emotion intensity rating and 
recognition for each emotion as a function of mask. Error bars are 
ઔ�ljȅǿЙǏǓǿljǓ�ǩǿȠǓȖΚƺǹșঀࢶࢺ

Interactions between mask and depression scores also 
ǟƺǩǹǓǏ�Ƞȅ�ΡǩǓǹǏ�șǩǠǿǩЙljƺǿȠ�ȖǓșȣǹȠșঀ�^Ƞǩǹǹॹ�ǩȠ�ǩș�ΛȅȖȠǦ�ǾǓǿ-
ȠǩȅǿǩǿǠ�ȠǦƺȠॹ�ǩǿ�ȠǦǓ�ljƺșǓ�ȅǟ�ǦƺȒȒǩǿǓșșॹ�ƺ�ǿȅǿেșǩǠǿǩЙ-
ljƺǿȠ�ǩǿȠǓȖƺljȠǩȅǿ�ǓАǓljȠ�ljȅǿșǩșȠǓǿȠ�ΛǩȠǦ�ȠǦǓ�ȅǿǓ�ǟȅȣǿǏ�
in the emotion intensity rating variable was reported 
শݾ�઀�েࢱঀࢷࢱࢱॹ�Ȓઃࢱঀࢲॹ��ȅǦǓǿঢ়ș�Ǐ�઀�েࢱঀࢶࢱॹ��2ࢶࢺઔ�઀�
[-0.12; 0.01]), which is plotted in Figure 5. The model 
for emotion recognition of happiness including the 
mask, depression scores, and the interaction between 
the two produced a marginal R2 of 0.093 (calculated 
according to Nakagawa & Schielzeth, 2013). Moder-
ation analysis through simple slopes was conducted 
for this outcome variable and showed similar results 
to the one of emotion intensity rating, meaning that 
ȠǦǓ�ǓАǓljȠ�ȅǟ�ȠǦǓ�ǾƺșǷ�Λƺș�șǩǠǿǩЙljƺǿȠ�ȠǦȖȅȣǠǦȅȣȠ�ȠǦǓ�
depression scale and tended to increase as partici-
ȒƺǿȠș�ȖǓȒȅȖȠǓǏ�ǾȅȖǓ�șΡǾȒȠȅǾșॹ�LjȣȠࢶࢺ�ઔ�ljȅǿЙǏǓǿljǓ�
ǩǿȠǓȖΚƺǹș�ȅΚǓȖǹƺȒȒǓǏ�șȅ�ȠǦƺȠ�ȠǦǓ�ǏǩАǓȖǓǿljǓ�LjǓȠΛǓǓǿ�ȠǦǓ�
simple slopes 

ljȅȣǹǏ�ǿȅȠ�LjǓ�ǩǿȠǓȖȒȖǓȠǓǏ�ƺș�șǩǠǿǩЙljƺǿȠঀ��ȅǓГljǩǓǿȠș�ƺȠ�
ȠǦǓ�ȠǦȖǓǓ�ȖǓșȒǓljȠǩΚǓ�ȒȅǩǿȠș�ΛǓȖǓݾ��઀�েࢱঀࢴࢲ�ΛǩȠǦ��2ࢶࢺઔ�
઀�মেࢱঀࢺࢲআ�েࢱঀࢸࢱয�ƺȠ�েࢲ^�ॹݾ��઀�েࢱঀࢸࢲ�ΛǩȠǦ��2ࢶࢺઔ�઀�
মেࢱঀࢲࢳআ�েࢱঀࢳࢲয�ƺȠ�ȠǦǓ�ǾǓƺǿ�ƺǿǏݾ��઀�েࢱঀࢱࢳ�ΛǩȠǦ��2ࢶࢺઔ�઀�
[-0.26; -0.14] at +1SD.

 Within the framework of the Gesichter lesen proj-
ect, this study utilized morphed face stimuli to assess 
ȠǦǓ�ǓАǓljȠ�ȅǟ� ǟƺljǓ�ǾƺșǷș�ƺǿǏ�ǏǓȒȖǓșșǩΚǓ� șΡǾȒȠȅǾș�ȅǿ�
emotion recognition and emotion intensity rating.

eǦǓ��АǓljȠ�ȅǟ�EƺșǷেvǓƺȖǩǿǠ
 In partial accordance with H1.1 and H2.1, the 
results suggest that mask-wearing has, according to 
Cohen’s standards (Cohen, 1988), small to medium 

Figures 4 (top) and 5 (bottom): Regression slopes of emotion 
intensity rating and emotion recognition of happiness by de-
pression score grouped by mask (0: without mask, 1: with mask). 
Error bars are SDs.



81

MASK-WEARING AND FACIAL EMOTION RECOGNITION

ǓАǓljȠș�ȅǿ�LjȅȠǦ�ȖǓljȅǠǿǩȠǩȅǿ�ƺǿǏ�ǩǿȠǓǿșǩȠΡ�ȖƺȠǩǿǠ�ȅǟ�ǦƺȒ-
ȒΡ�ƺǿǏ�șƺǏ�ǟƺljǓșॹ�LjȣȠ�ǿȅ�șǩǠǿǩЙljƺǿȠ�ǓАǓljȠș�ΛǓȖǓ�ǟȅȣǿǏ�
for angry faces. This pattern could, at least partially, be 
explained by the notion that, in comparison to happy 
and sad faces, the diagnostic face region of anger tends 
to be in the upper half of the observed face (Bassili, 
1979). On a similar note, Calvo et al. (2018) evaluat-
ed eye movement patterns of participants looking at 
ǏǩАǓȖǓǿȠ�ǟƺljǩƺǹ�ǓǾȅȠǩȅǿș�ƺǿǏ�ǟȅȣǿǏ�șǩǠǿǩЙljƺǿȠǹΡ�ǾȅȖǓ�
ЙΠƺȠǩȅǿ�ȅǿ�ȠǦǓ�ǓΡǓ�ȖǓǠǩȅǿ�ǟȅȖ�ƺǿǠȖΡ�ǟƺljǓșঀ�eǦȣșॹ�ƺ�ǟƺljǓ�
mask that covers only the lower region of the face may 
not impair the recognition of anger at all. However, it 
is important to point out that the fact that our study 
ǏǩǏ�ǿȅȠ�ЙǿǏ�ƺ�ǏǩАǓȖǓǿljǓ�LjǓȠΛǓǓǿ�ȠǦǓ�ǾƺșǷǓǏ�Κșঀ�ȣǿ-
masked condition for anger does not mean that this 
ǏǩАǓȖǓǿljǓ�ǏȅǓș�ǿȅȠ�ǓΠǩșȠঀ�2ǿ�ȠǦǩș�șǓǿșǓॹ�ǾȅȖǓ�ȖǓșǓƺȖljǦ�
with larger samples is needed to further investigate 
the role of masks in the facial expression of anger.
 Still, considering that the mask may, for example, 
also impair the understanding of speech (Mheidly 
ǓȠ� ƺǹঀॹ� �ষॹࢱࢳࢱࢳ ȅȣȖ� ЙǿǏǩǿǠș� ǟȅȖ� ǦƺȒȒǩǿǓșș� ƺǿǏ� șƺǏǿǓșș�
may have important consequences for social interac-
tions that include face masks. Social interaction part-
ners may have to express their sadness or happiness in 
less ambiguous ways to avoid being misunderstood, 
for example, by speaking louder and incorporating 
gestures and body language when displaying these 
ǓǾȅȠǩȅǿșঀ� KǿǓ� ȠΡȒǓ� ȅǟ� ǩǿȠǓȖƺljȠǩȅǿ� ȠǦƺȠ� ljȅȣǹǏ� șȣАǓȖ�
ljȅǿșǩǏǓȖƺLjǹΡ�ǟȖȅǾ�ȠǦǓșǓ�ǏǓЙljǩȠș�ǩșॹ�ǟȅȖ�ǩǿșȠƺǿljǓॹ�ƺ�ǹǩΚǓ�
psychotherapy session in which the client and the 
therapist are wearing face masks. Psychotherapists 
should pay attention to these impairments when try-
ing to read the facial expression of patients and when 
expressing emotional reactions themselves. Failure in 
accounting for this shortfall may worsen interperson-
al communication and hinder therapeutic progress.

The Relevance of Depressive Symptoms
 The analysis of a possible�Ǿƺǩǿ� ǓАǓljȠ� ȅǟ� ǏǓȒȖǓș-
sive symptoms on emotion recognition and intensity 
rating as postulated by hypotheses H1.2 and H2.2 
ǏǩǏ�ǿȅȠ�ΡǩǓǹǏ�ƺǿΡ�șǩǠǿǩЙljƺǿȠ�ȖǓșȣǹȠșঀ�eǦǩș�ЙǿǏǩǿǠ�ǠȅǓș�
against some of the literature presented in the theoret-
ical background of the present study. We could spec-
ulate in the following ways about the non-emergence 
ȅǟ�ƺǿ�ǓАǓljȠঀ�'ǩȖșȠॹ�ljȅǿșǩǏǓȖǩǿǠ�ȠǦƺȠ�ΛǓ�ΛǓȖǓ�ǓΠȒǓljȠǩǿǠ�
șǾƺǹǹ�ǓАǓljȠșॹ�ȠǦǓ�șȠȣǏΡ�ǾƺΡ�ǦƺΚǓ�ǹƺljǷǓǏ�ȠǦǓ�ȒȅΛǓȖ�Ƞȅ�
ЙǿǏ� ǩȠঀ� /ȅΛǓΚǓȖॹ� ƺ� ȒȅΛǓȖ� ƺǿƺǹΡșǩș� ǟȅȖ� ǹǩǿǓƺȖ� ǾǩΠǓǏ�

models, which needs simulation studies, was beyond 
the scope of this article. Second, our study did not 
apply psychiatric diagnostic of participants, and the 
number of participants presenting moderately severe 
to severe (>15) PHQ-9 scores was relatively low (11 
ȅȣȠ�ȅǟࢲࢺ�ॹࢳࢲ�ઔষঀ�2Ƞ�ljȅȣǹǏ�LjǓ�ȠǦƺȠ�ȠǦǓ�ǓАǓljȠ�ǩǿ�ȕȣǓșȠǩȅǿ�
manifests itself in clinical depression cases, as observed 
for example by Bistricky et al. (2011) and Bourke et al. 
(2010) but disappears when evaluating the non-clin-
ical spectrum of depressive symptoms. Furthermore, 
ǾƺǿΡ�șȠȣǏǩǓș�ȠǦƺȠ�ȅLjșǓȖΚǓǏ�'�[�ǏǓЙljǩȠș�ǩǿ�ǏǓȒȖǓșșǩΚǓ�
patients utilized ambiguous or neutral facial stimuli 
(e.g., Beevers et al., 2009; Kan et al., 2004, Bourke et 
al., 2010) whereas this study focused solely on a less 
ambiguous positive (happy) or less unambiguous neg-
ative (angry and sad) facial expressions. At the same 
time, it is relevant to note that there is still a debate 
in the literature as to whether facial emotion recog-
nition biases are indeed a characteristic of depressive 
disorders. Wu et al. (2012), for example, reported 
normal performance by highly depressive patients 
when testing for accuracy in recognizing emotions. 

2ǿȠǓȖƺljȠǩȅǿ�ȅǟ�EƺșǷেvǓƺȖǩǿǠ�ƺǿǏ��ǓȒȖǓșșǩȅǿ
 Concerning H1.3 and H2.3, the data produced 
mixed results. For happiness, sadness, and anger, no 
ljȅǿșǩșȠǓǿȠ� șǩǠǿǩЙljƺǿȠ� ǩǿȠǓȖƺljȠǩȅǿ� Λƺș� ȅLjșǓȖΚǓǏ� LjǓ-
tween the mask factor and depressive scores for any 
of the dependent variables. If at all, the pattern of the 
results was most suggestive in the case of happiness. 
Given the simple slope pattern, there may be some in-
ǏǩljƺȠǩȅǿ�Ƞȅ�șȒǓljȣǹƺȠǓ�ȠǦƺȠ�ȠǦǓȖǓ�ljȅȣǹǏ�LjǓ�ƺ�șǩǠǿǩЙljƺǿȠ�
ǏǩАǓȖǓǿljǓ�ȅǿljǓ�ȠǦǓ�șƺǾȒǹǓ�șǩΦǓ�Λƺș�ǹƺȖǠǓȖঀ��Ρ�ȠƺǷǩǿǠ�ƺ�
ljǹȅșǓȖ�ǹȅȅǷ�ƺȠ�'ǩǠȣȖǓșࢵ��ƺǿǏࢶ��ƺǿǏ�ȠǦǓ�ǩǿȠǓȖƺljȠǩȅǿ�ljȅǓГ-
cients, it seems that the presence of a face mask triggers 
a depressive response to the evaluation of happiness. In 
other words, when happy faces are masked, depressive 
scores correlate negatively in a slightly stronger fash-
ion with emotion recognition and emotion intensity 
ratings. However, it is imperative to note that moder-
ation analysis with the help of simple slopes revealed 
ȠǦƺȠ�ȠǦǓࢶࢺ�ઔ�ljȅǿЙǏǓǿljǓ�ǩǿȠǓȖΚƺǹș�ȅǟ�ȠǦǓșǓ�ljȅǓГljǩǓǿȠș�
ΛǓȖǓ�ǿȅȠ�șǩǠǿǩЙljƺǿȠǹΡ�ǏǩАǓȖǓǿȠ�ǟȖȅǾ�ȅǿǓ�ƺǿȅȠǦǓȖঀ�KǿǓ�
could speculate that with increased power and more 
ȖǓǹǩƺLjǹǓ�ǾǓƺșȣȖǓǾǓǿȠॹ� ȠǦǩș� ǓАǓljȠ� ljȅȣǹǏ�LjǓ� ǟȅȣǿǏ� șǩǠ-
ǿǩЙljƺǿȠ� ǩǿ� ǟȣȠȣȖǓ� șȠȣǏǩǓș� ȅȖ� ȠǦǓ�ȅǿǠȅǩǿǠ�ȒȖȅǴǓljȠ�Ge-
sichter lesenঀ�^Ƞǩǹǹॹ�ǓАǓljȠș�ƺȖǓ�ǓΠȒǓljȠǓǏ�Ƞȅ�LjǓ�șǾƺǹǹॹ�ƺǿǏ�
no conclusions can be drawn from the present sample.



82

MAZZAFERRO

Implications of Findings
 2ǟ�ȅǿǓ�ǏǓȠǓljȠǓǏ�ƺ�șǩǠǿǩЙljƺǿȠ�ǩǿȠǓȖƺljȠǩȅǿ�ǓАǓljȠॹ�ǟȅȖ�
example, with a larger dataset, it could be interpret-
ǓǏ�ǩǿ�ƺ�ǟǓΛ�ǏǩАǓȖǓǿȠ�ΛƺΡșঀ�'ǩȖșȠॹ� ǩȠ� ǩș�ȒȅșșǩLjǹǓ�ȠǦƺȠॹ�ƺș�
reported by Leyman et al. (2008), participants with 
higher rates of depressive symptoms have more trou-
LjǹǓ� ЙΠƺȠǩǿǠ� ȠǦǓǩȖ� ǠƺΦǓ� ȅǿ� ǦƺȒȒΡ� ǟƺljǓș� ƺǿǏ� ȠǦǓȖǓǟȅȖǓ�
ȒȖǓșǓǿȠ� ǾȅȖǓ� ǏǩГljȣǹȠΡ� ǩǿ� ȖǓljȅǠǿǩΦǩǿǠ� ȠǦǩș� ǓǾȅȠǩȅǿ�
when the mask is present. Duque & Vázquez (2015) 
also observed attentional bias in depressed patients, 
in the sense that positive emotions attracted less at-
tention than negative ones. The eye-tracking meth-
odology could be utilized with the presence of a mask 
in future studies to further investigate this rationale. 
Second, one could argue that the mask may act as a 
social signal that triggers pandemic-related psycholog-
ical burdens, and therefore impacts emotion interpre-
tation performance. It is noteworthy to mention that 
the second explanation does not clarify why the inter-
ƺljȠǩȅǿ�ǓАǓljȠ�ΛȅȣǹǏ�LjǓ�ǾǩșșǩǿǠ�ǟȅȖ�șƺǏǿǓșș�ƺǿǏ�ƺǿǠǓȖঀ��
� eǦǓ�ЙǿǏǩǿǠ�ȅǟ�ƺǿ�ǩǿȠǓȖƺljȠǩȅǿ�ǓАǓljȠ�ȅǿ�ǦƺȒȒǩǿǓșș�
and other emotions would have especially important 
implications. Many recently published studies high-
light how pandemic-related safety measures interact 
with the mental health of vulnerable populations. 
One study conducted in Germany by Benke et al. 
(2020), for example, found that higher restrictions 
due to lockdown measures, a greater reduction of 
social contacts, and greater perceived changes in life 
were associated with higher mental health impair-
ments. In Italy, Fiorenzato et al. (2021) found sub-
jective cognitive functioning and mental health were 
strongly associated with enforcing social distancing 
measures. In China, Lai et al. (2020) reported par-
ticularly bad mental health outcomes for front-line 
professions such as doctors and nurses, people that 
consistently use face masks and protective equipment 
in their daily tasks. In this context, this preliminary 
study is a further account of how the impairments 
in facial emotion recognition brought about by the 
wearing of face masks can have a distinctive impact 
on people with a disposition to depressive symptoms.
 The present investigation sheds light on one spe-
ljǩЙlj� ȒǦǓǿȅǾǓǿȅǿ� ȠǦƺȠ� ljƺǿ� ǓΠǩșȠ� ǩǿ� ȠǦǓ� ȣǿǩΚǓȖșǓ� ȅǟ�
interactions between psychopathology and the bur-
dens produced by the current pandemic, namely, that 
ǾƺșǷেΛǓƺȖǩǿǠ�ǾƺΡ�șǩǠǿǩЙljƺǿȠǹΡ�ǩǾȒƺǩȖ�ȒǓȅȒǹǓঢ়ș�ƺLjǩǹǩȠΡ�
to interpret happy and sad facial expressions and that 

this impairment may present itself more strongly for 
individuals scoring higher on a depressive symptoms 
scale. Following the studies cited in this paragraph, 
the present data provides some evidence that the pan-
demic and the safety measures related to it do impact 
ȒșΡljǦȅǹȅǠǩljƺǹ�ǟȣǿljȠǩȅǿǩǿǠ�ƺǿǏ�ǾƺΡ�Ǐȅ�șȅ�ǏǩАǓȖǓǿȠǹΡ�ǩǿ�
dependence on people’s mental health. Governments 
ƺǿǏ� ȖǓșȒȅǿșǩLjǹǓ� ƺȣȠǦȅȖǩȠǩǓș� șǦȅȣǹǏ� ȠƺǷǓ� ȠǦǓșǓ� ЙǿǏ-
ings into consideration when dealing with the spread 
of the Coronavirus and planning future restrictions.

Limitations and Outlook
 Although one should consider the time and ac-
cessibility constraints under which this study was 
conducted and evaluated, there are a few strengths 
ΛȅȖȠǦ� ǾǓǿȠǩȅǿǩǿǠঀ� 'ǩȖșȠॹ� ΛǓ� LjȣǩǹȠ� ƺ� ȠǩǾǓেǓАǓljȠǩΚǓ�
online study that could be completed by anyone that 
spoke German and had a tablet or computer with up-
to-date software and a webcam. Second, considering 
the complexity and time limitation usually associat-
ed with this investigation, this study conducted fair-
ly sophisticated data analysis with the help of LMMs 
and its possibilities within the statistical software R, 
which help prevent false-positive associations due 
to population or relatedness structure and increase 
ȒȅΛǓȖ�LjΡ�ƺȒȒǹΡǩǿǠ�ƺ�ljȅȖȖǓljȠǩȅǿ�ȠǦƺȠ�ǩș�șȒǓljǩЙlj�Ƞȅ�ȠǦǩș�
structure. In addition, the preliminary character of 
ȠǦǩș� ǩǿΚǓșȠǩǠƺȠǩȅǿ�ƺǹǹȅΛș� ǟȅȖ� ǟȣȖȠǦǓȖ� șljǩǓǿȠǩЙlj� ǩǿȕȣǩȖΡ�
of the psychological processes addressed in this paper, 
both within and outside of the Gesichter lesen project.
 Nonetheless, the present study contains sever-
al limitations worthy of critical evaluation. Because 
of the short period of time within which the data 
had to be collected for this thesis, we conducted 
the analysis on a relatively small and homogeneous 
șƺǾȒǹǓঀ� /ƺΚǩǿǠ� ǓΠȒǓljȠǓǏ� șǾƺǹǹ� ǓАǓljȠ� șǩΦǓșॹ� ƺ� ǹƺȖǠ-
er and more heterogeneous sample could account 
ǟȅȖ� ǾȅȖǓ� ȖǓǹǩƺLjǹǓ� ȖǓșȣǹȠș� ƺǿǏ� ȒȅșșǩLjǹΡ� ЙǿǏ� ǓАǓljȠș�
that remained undisclosed in the present sample.
 Dealing with the phenomenon of emotion percep-
tion and processing, the fact that this study was based 
solely upon psychological rating data is also a limita-
tion. Including other forms of measurement to assess 
response to stimuli, such as psychophysiological data, 
can increase validity and further contribute to the un-
derstanding of the psychological response to emotions. 
The Gesichter lesen project is already in the process of 
collecting and analyzing this type of data. In a relat



83

MASK-WEARING AND FACIAL EMOTION RECOGNITION

ed online experiment, Kastendieck et al. (2021) have 
found that facial mimicry - the perceiver’s imitation of 
the other’s emotional display - was reduced or absent in 
response to happy but preserved for sad mask-covered 
expressions. In the future, it would be interesting to see 
how mimicry and other physiological processes such 
as skin conductance and heart rate relate to psychopa-
thology in emotion processing. For that, a study set up 
in laboratory settings would be the better approach. 
 Because of the COVID-19 pandemic, university 
facilities and therefore laboratories had their access 
restricted and did not allow studies to take place. In 
order to increase standardization (for example, using 
electromyography instead of OpenFace3 video anal-
ysis), future studies should utilize superior meth-
ods for mimicry assessment and, if possible, go back 
to being conducted in laboratory environments. 
 As a last point, one could argue that there is still 
room for improvement in the quality of the stim-
uli. Here, we utilized morphed videos composed 
of static photos from neutral faces to faces display-
ing full emotions and added a face mask to it. To 
increase ecological validity, future stimuli should 
consist of recorded videos of people wearing face 
masks and expressing their respective emotions.

Conclusion
 In summary, the present study found evidence that 
masks impair facial emotion recognition and bias the 
ǩǿȠǓǿșǩȠΡ�ȖƺȠǩǿǠ�ȅǟ�ǦƺȒȒΡ�ƺǿǏ�șƺǏ�ǟƺljǓș�LjȣȠ�ǦƺΚǓ�ǿȅ�ǓАǓljȠ�
on angry faces. Depression scores were not associated 
ΛǩȠǦ�ȠǦǓ�ȅȣȠljȅǾǓș�ǩǿ�ȠǦǩș�șƺǾȒǹǓঀ��ǿ�ǩǿȠǓȖƺljȠǩȅǿ�ǓАǓljȠ�
between mask-wearing and depressive symptoms may 
ǓΠǩșȠ�ǟȅȖ�ǦƺȒȒΡ�ǟƺljǩƺǹ�ǓΠȒȖǓșșǩȅǿșঀ�eǦǓ�ЙǿǏǩǿǠș�ǩǾȒǹǩ-
cate that social interaction partners wearing face masks 
should pay additional attention to facial expressions to 
avoid misinterpreting emotions. In light of the results 
of this preliminary study and the relevance of precau-
tion measures to contain the spread of COVID-19 
at the current moment, authorities should consider 
ȠǦǓ� șȣǠǠǓșȠǓǏ�ǏǓЙljǩȠș� ǩǿ�ǓǾȅȠǩȅǿ�ȖǓljȅǠǿǩȠǩȅǿ�ƺǿǏॹ� ǩǿ�
particular, how they associate with psychopathology 
when planning future pandemic-related public policy.

 3 3 The  The Gesichter lesenGesichter lesen project is currently analyzing mimicry data using the OpenFace software. For more information on OpenFace, see Baltrusaitis et. al.,   project is currently analyzing mimicry data using the OpenFace software. For more information on OpenFace, see Baltrusaitis et. al.,  
(2016). (2016). 

References
American Psychiatric Association. (2013). Diag-

nostic and statistical manual of mental disor-
ders (5th ed.). https://doi.org/10.1176/appi.
books.9780890425596

An, L., Hawley, S., Van Horn, M. L., Bacon, E., Yang, 
P., & Resnicow, K. (2021). Development of a 
coronavirus social distance attitudes scale. Patient 
Education and Counseling, 104(6), 1451–1459. 
https://doi.org/10.1016/j.pec.2020.11.027

Aron, A., Aron, E. N., & Smollan, D. (1992). Inclu-
sion of Other in the Self Scale and the structure 
of interpersonal closeness. Journal of Personality 
and Social Psychology, 63(4), 596–612. https://
doi.org/10.1037/0022-3514.63.4.596

Bassili, J. N. (1979). Emotion recognition: The role of 
facial movement and the relative importance of 
upper and lower areas of the face. Journal of Per-
sonality and Social Psychology, 37(11), 2049–2058. 
https://doi.org/10.1037/0022-3514.37.11.2049

Baltrusaitis, T., Robinson, P., & Morency, L. P. 
(2016). OpenFace: An open-source facial behavior 
analysis toolkit (p. 10). https://doi.org/10.1109/
WACV.2016.7477553

Beard, C., Hsu, K. J., Rifkin, L. S., Busch, A. B., & 
Björgvinsson, T. (2016). Validation of the PHQ-9 
in a psychiatric sample.� =ȤɂȵȞǙȘ� ȤǾ��ωDzǨȿȈ͠Dz��Ȉȸ-
orders, 193, 267–273. https://doi.org/10.1016/j.
jad.2015.12.075

Beevers, C., Wells, T., Ellis, A., & Fischer, K. (2009). 
2ǏǓǿȠǩЙljƺȠǩȅǿ� ȅǟ� �ǾȅȠǩȅǿƺǹǹΡ��ǾLjǩǠȣȅȣș� 2ǿȠǓȖ-
personal Stimuli Among Dysphoric and Non-
dysphoric Individuals. Cognitive Therapy and 
Research, 33, 283–290. https://doi.org/10.1007/
s10608-008-9198-6

Beck, A. T. (1976). Cognitive therapy and the emotion-
al disorders. New York: Meridian.

Benke, C., Autenrieth, L. K., Asselmann, E., & 
Pané-Farré, C. A. (2020). Lockdown, quarantine 
measures, and social distancing: Associations 
with depression, anxiety, and distress at the begin-
ning of the COVID-19 pandemic among adults 
from Germany. Psychiatry Research, 293, 113462. 
https://doi.org/10.1016/j.psychres.2020.113462

Berking, M., & Rief, W. (Eds.). (2012). Klinische Psy-
chologie und Psychotherapie für Bachelor: 



84

MAZZAFERRO

 Band I: Grundlagen und Störungswissen. Lesen, 
Hören, Lernen im Web. Springer-Verlag. https://
doi.org/10.1007/978-3-642-16974-8

Bijsterbosch, G., Mobach, L., Verpaalen, I. A. 
M., Bijlstra, G., Hudson, J. L., Rinck, M., & 
Klein, A. M. (2021). Validation of the child 
models of the Radboud Faces Database by 
children. International Journal of Behavior-
al Development, 45(2), 146–152. https://doi.
org/10.1177/0165025420935631

Bistricky, S. L., Ingram, R. E., & Atchley, R. A. (2011). 
'ƺljǩƺǹ�ƺАǓljȠ�ȒȖȅljǓșșǩǿǠ�ƺǿǏ�ǏǓȒȖǓșșǩȅǿ�șȣșljǓȒȠǩLjǩǹ-
ity: Cognitive biases and cognitive neuroscience. 
Psychological Bulletin, 137(6), 998–1028. https://
doi.org/10.1037/a0025348

Bourke, C., Douglas, K., & Porter, R. (2010). Process-
ing of Facial Emotion Expression in Major De-
pression: A Review. Australian & New Zealand 
Journal of Psychiatry, 44(8), 681–696. https://doi.
org/10.3109/00048674.2010.496359

Bower, G. H. (1981). Mood and memory. Ameri-
can Psychologist, 36(2), 129–148. https://doi.
org/10.1037/0003-066X.36.2.129

Calbi, M., Langiulli, N., Ferroni, F., Montalti, M., 
Kolesnikov, A., Gallese, V., & Umiltà, M. A. 
(2021). The consequences of COVID-19 on so-
cial interactions: An online study on face cover-
ing. ^ǨȈDzȞȿȈϫǨ� [DzȱȤȵȿȸ࣓� .https://doi .2601 ࣓ओࠌऒࠌࠌ
org/10.1038/s41598-021-81780-w

Calvo, M. G., Fernández-Martín, A., Gutiérrez-García, 
�ঀॹ�૭�@ȣǿǏȕΚǩșȠॹ��ঀ�শࢹࢲࢱࢳষঀ�^ǓǹǓljȠǩΚǓ�ǓΡǓ�ЙΠƺȠǩȅǿș�
on diagnostic face regions of dynamic emotion-
al expressions: KDEF-dyn database. SǨȈDzȞȿȈϫǨ�
Reports, 8(1), 17039. https://doi.org/10.1038/
s41598-018-35259-w

Carbon, C. C. (2020a). Wearing Face Masks Strong-
ly Confuses Counterparts in Reading Emo-
tions. Frontiers in Psychology, 11. https://doi.
org/10.3389/fpsyg.2020.566886

Carbon, C. C. (2020b). The Psychology of Wearing 
Face Masks in Times of the COVID-19 Pan-
demic. SSRN Electronic Journal. https://doi.
org/10.2139/ssrn.3584834

Cheng, V. C.C., Wong, S.C., Chuang, V. W.M., So, 
S. Y.C., Chen, J. H.K., Sridhar, S., To, K. K.W., 
Chan, J. F.W., Hung, I. F.N., Ho, P.L., & Yuen, 
K.Y. (2020). The role of community-wide wear-
ing of face mask for control of coronavirus dis-

ease 2019 (COVID-19) epidemic due to SARS-
CoV-2. The Journal of Infection, 81(1), 107–114. 
https://doi.org/10.1016/j.jinf.2020.04.024

Cohen, J. (1988). Statistical Power Analysis for the Be-
havioral Sciences. New York, NY: Routledge Aca-
demic

Dalili, M. N., Penton-Voak, I. S., Harmer, C. J., & Mu-
nafò, M. R. (2015). Meta-analysis of emotion rec-
ȅǠǿǩȠǩȅǿ�ǏǓЙljǩȠș�ǩǿ�ǾƺǴȅȖ�ǏǓȒȖǓșșǩΚǓ�ǏǩșȅȖǏǓȖঀ�Psy-
chological Medicine, 45(6), 1135–1144. https://
doi.org/10.1017/S0033291714002591

Demenescu, L. R., Kortekaas, R., den Boer, J. A., & 
Aleman, A. (2010). Impaired Attribution of Emo-
tion to Facial Expressions in Anxiety and Major 
Depression. PLoS ONE, 5(12), e15058. https://
doi.org/10.1371/journal.pone.0015058

Derntl, B., Seidel, E.-M., Kryspin-Exner, I., Has-
mann, A., & Dobmeier, M. (2009). Facial emo-
tion recognition in patients with bipolar I and 
bipolar II disorder. The British Journal of Clini-
cal Psychology, 48(Pt 4), 363–375. https://doi.
org/10.1348/014466509X404845

Druss, B. G. (2020). Addressing the COVID-19 Pan-
demic in Populations With Serious Mental Illness. 
JAMA Psychiatry, 77(9), 891–892. https://doi.
org/10.1001/jamapsychiatry.2020.0894

Duque, A., & Vázquez, C. (2015). Double attention 
bias for positive and negative emotional faces in 
clinical depression: Evidence from an eye-track-
ing study. Journal of Behavior Therapy and Ex-
perimental Psychiatry, 46, 107–114. https://doi.
org/10.1016/j.jbtep.2014.09.005

Eisenbarth, H., & Alpers, G. W. (2011). Happy mouth 
and sad eyes: Scanning emotional facial expres-
sions. Emotion, 11(4), 860–865. https://doi.
org/10.1037/a0022758

Fayolle, S. L., & Droit-Volet, S. (2014). Time Per-
ception and Dynamics of Facial Expressions of 
Emotions. PLoS ONE, 9(5), e97944. https://doi.
org/10.1371/journal.pone.0097944

Fiorenzato, E., Zabberoni, S., Costa, A., & Cona, G. 
(2021). Cognitive and mental health changes and 
their vulnerability factors related to COVID-19 
lockdown in Italy. PLOS ONE, 16(1), e0246204. 
https://doi.org/10.1371/journal.pone.0246204

Finch, J. F., & Zautra, A. J. (1992). Testing latent 
longitudinal models of social ties and depression 
among the elderly: A comparison of distri



85

MASK-WEARING AND FACIAL EMOTION RECOGNITION

 bution-free and maximum likelihood estimates 
with nonnormal data. Psychology and Aging, 
7(1), 107–118. https://doi.org/10.1037/0882-
7974.7.1.107

Fischer, A. H., Gillebaart, M., Rotteveel, M., Beck-
er, D., & Vliek, M. (2012). Veiled Emotions: 
eǦǓ� �АǓljȠ� ȅǟ� �ȅΚǓȖǓǏ� 'ƺljǓș� ȅǿ� �ǾȅȠǩȅǿ� XǓȖ-
ception and Attitudes. Social Psychological and 
Personality Science, 3(3), 266–273. https://doi.
org/10.1177/1948550611418534

Freud, E., Stajduhar, A., Rosenbaum, R. S., Avidan, 
G., & Ganel, T. (2020). The COVID-19 pandem-
ic masks the way people perceive faces.� ^ǨȈDzȞȿȈϫǨ�
Reports, 10(1), 22344. https://doi.org/10.1038/
s41598-020-78986-9

Haro, J. M., Arbabzadeh-Bouchez, S., Brugha, T. S., 
De Girolamo, G., Guyer, M. E., Jin, R., Lepine, J. 
P., Mazzi, F., Reneses, B., Vilagut, G., Sampson, 
N. A., & Kessler, R. C. (2006). Concordance of 
the Composite International Diagnostic Inter-
view Version 3.0 (CIDI 3.0) with standardized 
clinical assessments in the WHO World Mental 
Health Surveys. International Journal of Methods 
in Psychiatric Research, 15(4), 167–180. https://
doi.org/10.1002/mpr.196

Hong, L.X., Lin, A., He, Z.B., Zhao, H.H., Zhang, 
J.G., Zhang, C., Ying, L.J., Ge, Z.M., Zhang, X., 
Han, Q.Y., Chen, Q.Y., Ye, Y.H., Zhu, J.S., Chen, 
H.X., & Yan, W.H. (2020). Mask wearing in 
pre-symptomatic patients prevents SARS-CoV-2 
transmission: An epidemiological analysis. Trav-
el Medicine and Infectious Disease, 36, 101803. 
https://doi.org/10.1016/j.tmaid.2020.101803

Howard, J., Huang, A., Li, Z., Tufekci, Z., Zdimal, V., 
Westhuizen, H.M. van der, Delft, A. von, Price, 
A., Fridman, L., Tang, L.H., Tang, V., Watson, G. 
L., Bax, C. E., Shaikh, R., Questier, F., Hernan-
dez, D., Chu, L. F., Ramirez, C. M., & Rimoin, 
A. W. (2020). Face Masks Against COVID-19: An 
Evidence Review. https://doi.org/10.20944/pre-
prints202004.0203.v1

Kan, Y. (2004). Recognition of emotion from moving 
facial and prosodic stimuli in depressed patients. 
Journal of Neurology, Neurosurgery & Psychiatry, 
75(12), 1667–1671. 

Kim, Y. E., & Lee, B. (2019). The Psychometric Prop-
erties of the Patient Health Questionnaire-9 in a 
Sample of Korean University Students. Psychia-

try Investigation, 16(12), 904–910. https://doi.
org/10.30773/pi.2019.0226

Kessler, R. C., de Jonge, P., Shahly, V., van Loo, H. 
M., Wang, P. S. E., & Wilcox, M. A. (2014). Ep-
idemiology of depression. In I. H. Gotlib & C. L. 
Hammen (Eds.), Handbook of depression (7–24). 
The Guilford Press.  https://doi.org/10.1136/
jnnp.2004.036079

Kastendieck, T., Zillmer, S., & Hess, U. (2021). (Un)
ǾƺșǷ� ΡȅȣȖșǓǹǟॻ� �АǓljȠș� ȅǟ� ǟƺljǓ� ǾƺșǷș� ȅǿ� ǟƺljǩƺǹ�
mimicry and emotion perception during the 
COVID-19 pandemic. Cognition and Emotion, 
36(1), 59–69. https://doi.org/10.1080/0269993
1.2021.1950639

Kluczniok, D., Hindi Attar, C., Fydrich, T., Fuehrer, 
D., Jaite, C., Domes, G., Winter, S., Herpertz, S. 
C., Brunner, R., Boedeker, K., & Bermpohl, F. 
শࢷࢲࢱࢳষঀ�eȖƺǿșǠǓǿǓȖƺȠǩȅǿƺǹ�ǓАǓljȠș�ȅǟ�ǾƺȠǓȖǿƺǹ�ǏǓ-
ȒȖǓșșǩȅǿ�ȅǿ�ƺАǓljȠ�ȖǓljȅǠǿǩȠǩȅǿ�ǩǿ�ljǦǩǹǏȖǓǿঀ Journal 
ȤǾ��ωDzǨȿȈ͠Dz��ȈȸȤȵǮDzȵȸ239–233 ࣓ࠔࠓࠌ�࣓. https://doi.
org/10.1016/j.jad.2015.09.051

Kroenke, K., & Spitzer, R. L. (2002). The PHQ-9: A 
new depression diagnostic and severity measure. 
Psychiatric Annals, 32(9), 509–515. https://doi.
org/10.3928/0048-5713-20020901-06

Lai, J., Ma, S., Wang, Y., Cai, Z., Hu, J., Wei, N., Wu, 
J., Du, H., Chen, T., Li, R., Tan, H., Kang, L., 
Yao, L., Huang, M., Wang, H., Wang, G., Liu, 
Z., & Hu, S. (2020). Factors Associated With 
Mental Health Outcomes Among Health Care 
Workers Exposed to Coronavirus Disease 2019. 
JAMA Network Open, 3(3), e203976. https://doi.
org/10.1001/jamanetworkopen.2020.3976

Leyman, L., Raedt, R. D., Vaeyens, R., & Philippaerts, 
R. M. (2011). Attention for emotional facial expres-
sions in dysphoria: An eye-movement registration 
study. Cognition and Emotion, 25(1), 111–120. 
https://doi.org/10.1080/02699931003593827

Lopez-Duran, N. L., Kuhlman, K. R., George, C., 
& Kovacs, M. (2013). Facial emotion expression 
recognition by children at familial risk for depres-
sion: High risk boys are oversensitive to sadness. 
Journal of Child Psychology and Psychiatry, and 
Allied Disciplines, 54(5), 565–574. https://doi.
org/10.1111/jcpp.12005

Lehmann, K., Maliske, L., Böckler, A., & Kanske, P. 
(2019). Social Impairments in Mental Disorders: 
Recent Developments in Studying the Mecha



8686

MAZZAFERRO

 nisms of Interactive Behavior. Clinical Psychology 
in Europe, 1(2), 1–15. https://doi.org/10.32872/
cpe.v1i2.33143

Marta, C., Nunzio, L., Francesca, F., Martina, M., 
Anna, K., Vittorio, G., & Alessandra, U. M. 
(2020). I see how you feel: Facial expressions’ rec-
ognition and distancing in the time of COVID-19 
[Preprint]. In Review. https://doi.org/10.21203/
rs.3.rs-107145/v1

Mheidly, N., Fares, M. Y., Zalzale, H., & Fares, J. 
শࢱࢳࢱࢳষঀ� �АǓljȠ� ȅǟ� 'ƺljǓ� EƺșǷș� ȅǿ� 2ǿȠǓȖȒǓȖșȅǿƺǹ�
Communication During the COVID-19 Pan-
demic. Frontiers in Public Health, 8:582191. 
https://doi.org/10.3389/fpubh.2020.582191

Milders, M., Bell, S., Platt, J., Serrano, R., & Runcie, 
O. (2010). Stable expression recognition abnor-
malities in unipolar depression. Psychiatry Re-
search, 179(1), 38–42. https://doi.org/10.1016/j.
psychres.2009.05.015

Nakagawa, S., & Schielzeth, H. (2013). A general 
and simple method for obtaining R2 from gen-
ǓȖƺǹǩΦǓǏ� ǹǩǿǓƺȖ� ǾǩΠǓǏেǓАǓljȠș� ǾȅǏǓǹșঀ�Methods in 
Ecology and Evolution, 4(2), 133–142. https://doi.
org/10.1111/j.2041-210x.2012.00261.x

Nyquist, A. C., & Luebbe, A. M. (2020). An Emotion 
Recognition-Awareness Vulnerability Hypothe-
sis for Depression in Adolescence: A Systematic 
Review. Clinical Child and Family Psychology 
Review, 23(1), 27–53. https://doi.org/10.1007/
s10567-019-00302-3

XǓȖșƺǏॹ�^ঀ�Eঀॹ�૭�XȅǹǩΚΡॹ�=ঀ�শࢴࢺࢺࢲষঀ��ǩАǓȖǓǿljǓș�LjǓȠΛǓǓǿ�
depressed and nondepressed individuals in the rec-
ognition of and response to facial emotional cues. 
Journal of Abnormal Psychology, 102(3), 358–368. 
https://doi.org/10.1037//0021-843x.102.3.358

Platt, B., Kadosh, K. C., & Lau, J. Y. F. (2013). The 
Role of Peer Rejection in Adolescent Depression. 
Depression and Anxiety, 30(9), 809–821. https://
doi.org/10.1002/da.22120

Richards, M., & Light, P. (Eds.). (1986). Children of 
social worlds: Development in a social context (p. 
327). Harvard University Press.

Richardson, L. P., Rockhill, C., Russo, J. E., Gross-
man, D. C., Richards, J., McCarty, C., McCauley, 
E., & Katon, W. (2010). Evaluation of the PHQ-
2 as a brief screen for detecting major depression 
among adolescents. Pediatrics, 125(5), e1097-
1103. https://doi.org/10.1542/peds.2009-2712

Ruba, A. L., & Pollak, S. D. (2020). Children’s emotion 
inferences from masked faces: Implications for so-
cial interactions during COVID-19. PLOS ONE, 
15(12), e0243708. https://doi.org/10.1371/jour-
nal.pone.0243708

Rubinow, D. R., & Post, R. M. (1992). Impaired rec-
ȅǠǿǩȠǩȅǿ�ȅǟ�ƺАǓljȠ�ǩǿ�ǟƺljǩƺǹ�ǓΠȒȖǓșșǩȅǿ�ǩǿ�ǏǓȒȖǓșșǓǏ�
patients. Biological Psychiatry, 31(9), 947–953. 
https://doi.org/10.1016/0006-3223(92)90120-o

[ΡǾƺȖljΦΡǷॹ�?ঀॹ��ȣȖƺΛșǷǩॹ�Dঀॹ�=ƺǿǷȅΛǩƺǷে^ǩȣǏƺॹ�?ঀॹ�૭�
Szatkowska, I. (2016). Emotional Empathy and 
Facial Mimicry for Static and Dynamic Facial Ex-
pressions of Fear and Disgust. Frontiers in Psychol-
ogy, 7. https://doi.org/10.3389/fpsyg.2016.01853

Schmid, P. C., & Schmid Mast, M. (2010). Mood 
ǓАǓljȠș� ȅǿ� ǓǾȅȠǩȅǿ� ȖǓljȅǠǿǩȠǩȅǿঀ� Motivation 
and Emotion, 34(3), 288–292. https://doi.
org/10.1007/s11031-010-9170-0

Schurgin, M. W., Nelson, J., Iida, S., Ohira, H., 
Chiao, J. Y., & Franconeri, S. L. (2014). Eye 
movements during emotion recognition in fac-
es. Journal of Vision, 14(13), 14–14. https://doi.
org/10.1167/14.13.14

Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Löwe, 
B. (2006). A Brief Measure for Assessing General-
ized Anxiety Disorder: The GAD-7. Archives of 
Internal Medicine, 166(10), 1092–1097. https://
doi.org/10.1001/archinte.166.10.1092

^ȒǩȠΦǓȖॹ�Eঀ� শࢱࢳࢱࢳষঀ�EƺșǷǓǏ� ǓǏȣljƺȠǩȅǿঁ�eǦǓ� LjǓǿǓЙȠș�
and burdens of wearing face masks in schools 
during the current Corona pandemic. Trends in 
Neuroscience and Education, 20, 100138. https://
doi.org/10.1016/j.tine.2020.100138

Titov, N., Andrews, G., Davies, M., McIntyre, K., 
Robinson, E., & Solley, K. (2010). Internet treat-
ment for depression: A randomized controlled 
trial comparing clinician vs. technician assistance. 
PloS One, 5(6), e10939. https://doi.org/10.1371/
journal.pone.0010939

World Health Organization (2017). Depression 
and Other Common Mental Disorders. Global 
Health Estimates. Geneva: World Health organi-
zation.Retrieved from: https://apps.who.int/iris/
handle/10665/254610 Licence: CC BY-NC-SA 
3.0 IGO.