Psychological Research

https://doi.org/10.1007/s00426-018-0991-x

ORIGINAL ARTICLE

Startle reflex modulation by affective face “Emoji” pictographs

Anton Aluja1,2 · Ferran Balada2,3 · Eduardo Blanco1,2 · Ignacio Lucas1,2 · Angel Blanch1,2

Received: 25 September 2017 / Accepted: 12 February 2018

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract
The current research was designed to assess possible differences in the emotional content of pleasant and unpleasant face
emoji using acoustically evoked eyeblink startle reflex response. Stimuli were selected from Emojipedia Webpage. First, we
assessed these stimuli with a previous independent sample of 190 undergraduate students (46 males and 144 females) mean
age of 21.43 years (SD 3.89). A principal axis method was performed using the 30 selected emoji faces, extracting two factors
(15 pleasant and 15 unpleasant emoji). Second, we measured the acoustic startle reflex modulation in 53 young adult women
[mean age 22.13 years (SD 4.3)] during the viewing of each of the 30 emoji emotional faces in the context of the theory of
motivation and emotion proposed by Lang (1995), but considering only the valence dimension. We expected to find higher
acoustically evoked startle responses when viewing unpleasant emoji and lower responses for pleasant ones, similarly to the
results obtained in the studies using human faces as emotional stimulus. An ANOVA was conducted to compare acoustic
startle responses associated with pleasant and unpleasant emoji. Results yielded main effects for picture valence (λ = 0.80,
F(1, 50) = 12.80, p = .001, η2 = 0.20). Post-hoc t test analysis indicated significant differences in the startle response between
unpleasant (50.95 ± 1.75) and pleasant (49.14 ± 2.49) emoji (t (52) = 3.59, p = .001), with a Cohen’s d = 0.495. Viewing affec-
tive facial emoji expressions modulates the acoustic startle reflex response according to their emotional content.

Introduction
Acoustic startle reflex (ASR) modulation by affective emoji
is a useful paradigm for studying emotion. Eyeblink mag-
nitude modulated by affective state has become one of the
most widely used physiological measures in the study of
motivation and emotion (Filion, Dawson, & Schell, 1998).
Lang and collaborators proposed a theory of emotion, where
appetitive and aversive systems activate in accordance with
the subject’s current emotional state (Bradley & Lang, 1999;
Bradley & Lang, 2000a). ASR is then modulated depend-
ing on how they are activated. Several studies have demon-
strated that acoustic startle responses are attenuated during
passive viewing of affective pleasant pictures, whereas they
are increased when viewing unpleasant pictures (Aluja,
* Anton Aluja
aluja@pip.udl.cat
1
University of Lleida, Lleida, Catalonia, Spain
2
Institute of Biomedical Research of Lleida (IRB Lleida),
University of Lleida, Avda. de l’Estudi General, 4,
25001 Lleida, Catalonia, Spain
3
Autonomous University of Barcelona, Lleida, Catalonia,
Spain
Blanch, & Balada, 2015; Blanch, Aluja, Blanco, & Balada.,
2016; Bradley & Lang, 2000a; Cuthbert, Bradley, & Lang,
1996; Lang, Bradley, & Cuthbert, 1998; Vrana, Spence,
& Lang, 1988). Affective startle modulation demonstrates
that neuronal networks underlying emotions include direct
connections to the brain’s primary motivational systems
(Lang, Davis, & Ohman, 2000). Lang and collaborators
also proposed three factors to measure affective reactions:
(a) “emotional valence”, indicating the hedonic value of a
specific emotion as either positive or negative; (b) “emo-
tional arousal”, indicating its intensity; and (c) “dominance”,
indicating the feeling of being in control vs being controlled
(Bradley & Lang, 1994).
Emotional expressions of human faces are relevant in
non-verbal communication and can produce a significant
modulation of the ASR according to their emotional valence
(Anokhin & Golosheykin, 2010; Duval, Lovelace, Aarant,
& Filion, 2013; Paulus, Musial, & Renn, 2014). This effect
is especially potentiated when viewing angry faces (Dun-
ning, Auriemmo, Castille, & Hajcak, 2010). Nevertheless,
other studies have failed to find relationships among the
valence of facial expressions and ASR responses (Alpers &
Adolph, 2006; Waters, Neumann, Henry, Craske, & Ornitz,
2008). In a study by Balaban (1995), emoji of unfamiliar

13
Vol.:(0123456789)

adult faces with happy, neutral, and angry expressions were
administered to 5-month-old infants and affected the ASR
amplitude. It increased when they viewed angry faces and
decreased when happy faces were presented. In another
study, Springer, Rosas, McGetrick, and Bowers (2007) found
a significant effect of facial expressions on startle amplitude,
where angry but not fearful faces produced an increased eye-
blink response compared to all other expressions.
Emoji are pictographic symbols widely used in Internet,
electronic, and text communication such as instant messag-
ing applications and social media sharing (see real-time
emoji use on Twitter http://www.emoji​track​er.com/). Emoji
are available on different operative systems (e.g., iOS,
Android…). On the Internet, web pages provide several
emoji collections. The Emoji_Report_20151 indicate that
emoji are being used by 92% of the online population and
over 60% have used emoji several times a week or several
times a day. Some reasons for using emoji according to U.S.
Internet users as of August 20152 were: (a) “they help me
more accurately express what I am talking” (70.4%); (b) “it
makes it easy for other people to understand me” (64.7%);
and (c) “they help create a more personal connection with
the other persons” (49.7%). Traditional emoji faces dominate
the top two categories, with nearly 60% of all emoji sent.
Face emoji convey their meaning through graphic resem-
blance to human faces (e.g., a smiling face) (Miller et al.,
2016). Emoji can be considered examples of how people
have created surrogates for non-verbal cues. All these data
emphasize the importance of studying this form of commu-
nication and their emotional impact further.
Neuroimaging studies report that human facial expres-
sions activate emotion-related brain circuits in the limbic
system, particularly the amygdala (Breiter et al., 1996;
Hariri, Tessitore, Mattay, Fera, & Weinberger, 2002; Vuil-
leumier, Armony, Driver, & Dolan, 2001; Whalen et al.,
2001). The amygdala plays an important role in the pro-
cessing of facial emotional expressions by humans and
non-human primates as is demonstrated by lesion and
direct electrical stimulation studies (Adolphs et al., 2005;
Adolphs, Tranel, Damasio, & Damasio, 1995). However,
schematic faces appear to evoke greater amygdala activa-
tion on arousal-based contrasts (negative vs neutral) than
in valence-based ones (negative vs positive) (Evans et al.,
2008). Emoji faces express feelings and emotions, both posi-
tive and negative, although are more usual in a positive than
in a negative context (Derks, Bos, & von Grumbkow, 2008).
These authors also suggest that people seem to use emoji
1
https​ : //www.aarga​ u erze ​ i tung ​ . ch/asset ​ _ docum ​ e nt/i/12906 ​ 7 827/
downl​oad.
2
https​://www.stati​sta.com/stati​stics​/47635​4/reaso​ns-usage​-emoji​
s-inter​net-users​-us/.
13
Psychological Research
faces in a similar way to face-to-face communication with
regard to the social context and interaction partner. Never-
theless, an event-related potential (ERP) study with emoji
showed a reduced amplitude of N170 when the emoji were
inverted, the opposite pattern showed by faces, indicating
differences in recognition of the physiognomic features of
emoji compared to human faces, but not through configu-
ral processing (Churches, Nicholls, Thiessen, Kohler, &
Keage, 2014). Participants were better at categorizing face
cues as fearful or neutral with upright than with inverted
faces (Khalid, Horstmann, Ditye, & Ansorge, 2017). In fact,
some studies have suggested that human faces and schematic
faces (similar to emoji, but not in color) activate the same
emotional/motivational system.
As far as we know, no study has attempted to explore the
associations between emotional face emoji and ASR modu-
lation. The main objective of this study was to assess the
emotional impact of facial expression emoji via the effect
of pleasant and unpleasant emoji faces to the acoustically
evoked startle response. In accordance with the theory of
motivation and emotion proposed by Lang, Bradley, Cuth-
bert, and Patrick (1993), we hypothesized that the ASR
should be potentiated by emoji expressions with negative
emotional valence (fearful, angry, crying face, tired face,
weary face, etc.), whereas the startle response should be
attenuated by positive emotional valence faces (face with
tears of joy, smiling face with smiling eyes, etc.). The cur-
rent research was conducted in two different studies. The
first one, to assess the subjective emotional valence of the
emoji, and the second one to assess their emotional impact
with a psychophysiological measure. It was expected that
emoji with different emotional expressions would modulate
the ASR in a similar way as human faces with emotional
expressions.
Study 1: emoji valence assessment
The aim of this study was to validate the valence classifi-
cation of the previous emoji selected from the emojipedia
descriptions.
Method
Participants In Study 1, valence scores were obtained with
a sample of 190 undergraduate students (46 males and 144
females) with a mean age of 21.43 years old (SD 3.89). Par-
ticipation was voluntary and anonymous.
Instrument and procedure
We developed a paper and pencil protocol with 30 emoji (15
pleasant and 15 unpleasant), according to the description
provided by emojipedia (e.g., angry, fearful, smiling; http://
Psychological Research

Table 1  Pleasant (15) and unpleasant emoji (15) by presentation group (startle reflex condition) and operative system

Group A Group B Group C

Emoji Valence Startle Picture Operative System Picture Operative System Picture Operative System
Angry Unpleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
Face Throwing A Kiss Pleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
Face With TearsofJoy Pleasant N Twemoji 2.0 iOS 9.3 Android 6.0.1
Fearful Unpleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
Smiling Face With Heart-Shaped Eyes Pleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
Weary Face Unpleasant N Twemoji 2.0 iOS 9.3 Android 6.0.1
Tired Face Unpleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
Smiling Face With Smiling Eyes Pleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
Crying Face Unpleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
White Smiling Face Pleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1
Angry Unpleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
Face Throwing A Kiss Pleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1
Face With TearsofJoy Pleasant N iOS 9.3 Android 6.0.1 Twemoji 2.0
Fearful Unpleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
Smiling Face With Heart-Shaped Eyes Pleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1
Weary Face Unpleasant N iOS 9.3 Android 6.0.1 Twemoji 2.0
Tired Face Unpleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
Smiling Face With Smiling Eyes Pleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1
Crying Face Unpleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
White Smiling Face Pleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
Angry Unpleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1
Face Throwing A Kiss Pleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
Face With TearsofJoy Pleasant N Android 6.0.1 Twemoji 2.0 iOS 9.3
Fearful Unpleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1
Smiling Face With Heart-Shaped Eyes Pleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
Weary Face Unpleasant N Android 6.0.1 Twemoji 2.0 iOS 9.3

Tired Face Unpleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1

Smiling Face With Smiling Eyes Pleasant Y iOS 9.3 Android 6.0.1 Twemoji 2.0
Crying Face Unpleasant Y Twemoji 2.0 iOS 9.3 Android 6.0.1

White Smiling Face Pleasant Y Android 6.0.1 Twemoji 2.0 iOS 9.3
The 30 emoji are the same in the three groups, but ordered in different ways
Y yes, N no, Operative system iOS 9.3, Twemoji 2.0, Android 6.0.1

emoji​pedia​.org). Each emoji description was selected and
represented in three operating systems: iOS 9.3, Android
6.0.1 and Twemoji 2.0 (Table 1). Participants were asked to
evaluate each emoji according to its affective valence from
1 (unpleasant) to 9 (pleasant).
Results
Table 2 shows the factorial structure of the 30 emoji using
the principal axis analysis (PA) with oblimin rotation
extracting two factors using similar procedure conducted by
Aluja and Blanch (2011). The Kaiser–Meyer–Olkin meas-
ures of sample adequacy were 0.89 and a Bartletts’s test
of sphericity yielded an approximate Chi square value of
4061,992 (degrees of freedom 435; p < .001). The two fac-
tors explained 52.76% of the total variance using the struc-
ture matrix. The first factor was composed by the 15 pleasant
emoji and the second factor by the 15 unpleasant ones. Cor-
relation between factors was − 0.41. Table 2 also shows the
means and standard deviations of the 30-emoji subjective

13
 Psychological Research

Table 2  Principal axis analysis I II M SD

with oblimin rotation (structure) Smiling Face With Heart-Shaped Eyes Android 6.0.1 .84 -.34 2.20 2.20
of pleasant (I) and unpleasant Smiling Face With Heart-Shaped Eyes Twemoji 2.0 .83 -.36 1.82 1.77
(II) emoji and valence (mean White Smiling Face Twemoji 2.0 .81 -.34 2.38 1.86
and standard deviation)
Smiling Face With Smiling Eyes Twemoji 2.0 .80 -.29 2.46 1.98
Smiling Face With Smiling Eyes Android 6.0.1 .77 -.26 2.36 1.91
Face Throwing A Kiss Twemoji 2.0 .76 -.32 2.05 1.86
Face Throwing A Kiss Android 6.0.1 .76 -.27 2.25 2.00
Face Throwing A Kiss iOS 9.3 .66 -.34 1.79 1.62
White Smiling Face Android 6.0.1 .64 -.26 2.97 2.20
Face With Tears of Joy Android 6.0.1 .64 -.20 2.65 2.10
Smiling Face With Heart-Shaped Eyes iOS 9.3 .61 -.41 1.66 1.48
White Smiling Face iOS 9.3 .61 -.36 1.91 1.46
Face With Tears of Joy Twemoji 2.0 .59 -.14 2.09 1.89
Smiling Face With Smiling Eyes iOS 9.3 .53 -.36 1.89 1.45
Face With Tears of Joy iOS 9.3 .53 -.25 1.83 1.46
Crying Face Android 6.0.1 -.27 .76 6.57 1.81
Tired Face iOS 9.3 -.42 .74 6.15 1.78
Weary Face Android 6.0.1 -.25 .73 6.60 1.77
Weary Face Twemoji 2.0 -.25 .72 6.18 1.75
Crying Face Twemoji 2.0 -.28 .72 6.29 1.95
Angry Android 6.0.1 -.44 .71 7.57 1.92
Weary Face Android 6.0.1 -.25 .71 6.44 1.78
Fearful Twemoji 2.0 -.25 .70 5.99 1.58
Fearful Android 6.0.1 -.13 .69 6.29 1.89
Crying Face iOS 9.3 -.49 .69 6.12 1.97
Angry Twemoji 2.0 -.52 .66 6.95 1.90
Tired Face Android 6.0.1 -.20 .63 6.48 1.71
Angry iOS 9.3 -.54 .62 6.90 1.84
Weary Face iOS 9.3 -.42 .62 5.93 1.89
Fearful iOS 9.3 -.31 .56 5.57 1.63
M mean, SD standard deviation, Operative system iOS 9.3, Twemoji 2.0, Android
6.0.1

assessments. Significant differences in subjective valence
assessment were found between positive (7.85 ± 1.31) and
negative (3.61 ± 1.30) emoji (t test (190) = 26.33, p < .001).
Discussion
Following the Ekman & Friesen proposal (1975), other pre-
vious study has categorized emotions of emoji according to
six categories (happy, sad, fear, surprise, angry, and disgust),
but emoji used in cellular phones can also be divided into
two major categories of pleasant (smiling, laughing, etc.)
and unpleasant (sad, anger, fear, etc.) (Toratani & Hirayama,
2011).
The present factorial analyses confirm that all the
30-selected emoji from the emojipedia description loaded
into two independent factors according to affective valence
(pleasant vs unpleasant). An additional advantage of the
factorial analysis is that each factor is ordered according to
items with high to low factorial loadings. All emoji facto-
rial loadings were above 0.50 for their respective factor. The
factorial structure was robust and orthogonal. The pleas-
ant emoji with high loadings were represented by “smiles”,
while the unpleasant emoji with high loadings were repre-
sented by “crying”, “tired”, and “weary” faces. This clas-
sification supported the bipolar emotional content of these
emoji and its use as emotional material in the ASR study.
Study 2: startle response modulation
The aim of this study was to assess the differences in ASR
modulation associated to pleasant or unpleasant emoji.
Method
Participants The participants were 53 young adult women
composed the sample (mean age 22.13 years; SD 4.3). Par-
ticipation was also voluntary and anonymous. Both stud-
ies were authorized by the ethical committee and were in
accordance with the ethical standards laid down in the 1964
Declaration of Helsinki.

13
Psychological Research
Materials We composed three blocks of 10 emoji from of
the 30 obtained in study 1.
Each block contained five negative and five positive emoji
from three different operating systems (iOS 9.3, Android
6.0.1, Twemoji 2.0). The emoji were presented in the same
emotional valence sequence for each block (unpleasant,
pleasant, pleasant, unpleasant, pleasant, unpleasant, unpleas-
ant, pleasant, unpleasant, and pleasant), but alternating
the operative system for each emoji. Once determined the
sequence of stimuli, the operating system corresponding to
each stimulus was randomized, assuming that these ten stim-
uli should have four stimuli of an operating system and three
of each one of the other two operating systems. Participants
were assigned randomly to three groups, varying the order of
block presentation. After defining the first set of stimuli, the
other two sets were designed by changing operating systems.
For example, if the first block had an emoji from iOS 9.3, in
the second block, the same emoji was from Android 6.0.1,
and in the third block from Twemoji 2.0 (Table 1).
Procedure A clinical psychologist with open questions
interviewed all subjects. None of them reported a history
of psychopathological disorders, current medication, audi-
tory impairment or any uncorrected visual problems. Par-
ticipants were strongly advised to refrain from smoking and
drinking alcoholic or stimulant drinks for at least 12 h prior
to the experiment. The participants signed a written con-
sent form. The recording sessions took place in an isolated
acoustic and electromagnetic room (Faraday cage). Emoji
had a diameter of 15 cm and were presented in a 32-inch
screen, while the subject was sitting at a distance of 60 cm.
Prior to the stimuli presentation, the physiological sensors
were attached and impedances checked. To habituate partic-
ipants to the acoustic probe and minimize the novelty effect,
three acoustic stimuli were delivered before the emoji pres-
entations. Emoji were presented over a black background
for 6 s with an inter-trial interval (ITI) varying from 16 to
18 s; each emoji was preceded by a white cross (1 s dura-
tion). An acoustic startle probe was presented associated
with each emoji, within 2.5 to 3.5 s after emoji onset. Each
block included one of the five emoji of each emotional
valence without sound. Two ITI startle probes were pre-
sented in each block to reduce the temporal predictability
of the startle sound and to compare modulation of ASR by
emotional valence with a “baseline” level of reactivity. All
participants in this study assessed the 30 emoji according
to valence values with the same procedure used in Study 1.
Psychophysiological recording The startle probe was a
105 dB white noise, with 50 ms of duration and nearly instan-
taneous rise time delivered binaurally through headphones
(Maico, TDH-39-P). All data were sampled at 1000 Hz and
amplified with a gain of 5000 Hz using the electromyogra-
phy (EMG) module of the Biopac system and Acqkowledge
software. The startle blink reflex was recorded electromyo-
graphically at the right orbicularis oculi muscle with two
6-mm Ag–AgCl electrodes filled with electrolyte gel; one
electrode was placed below the right lower eyelid in line
with the pupil in forward gaze, and the other approximately
1–2 cm next to the first one (center-to-center) (Balada,
Blanch, & Aluja, 2014;Blanch, Balada, & Aluja, 2013; Blu-
menthal et al., 2005). Impedance levels were below 5 kΩ for
each participant.
EMG eyeblink responses were collected using a
10–500 Hz band pass analogic filter. The signal was then
integrated and smoothed with a five-point moving average
off-line filter. Peak amplitude was measured with the high-
est point in the EMG response between 20 and 120 ms after
probe onset, using the average of the 30 ms before probe
administrations as baseline. No response was considered
when peak amplitude was lowest than the average baseline
plus three standard deviations for this period. Responses
with onset latency below 20 ms and/or rise time above than
95 ms were rejected. Raw data on ASR magnitude showed
high between-subjects variability. To reduce this variability,
magnitude was transformed to within-subjects T-scores.
Results
A 2 × 3 ANOVA performed with group as between-group
factor and valence (pleasant and unpleasant) as within-group
factor yielded main effects for emoji valence [λ = 0.80, F(1,
50) = 12.80, p = .001, η2 = 0.20], but not for the interaction
effect valence × group [λ = 0.98, F(2, 50) = 0.58, p = .56]. In
addition, no significant effects were observed in between-
subject effects of groups [F(2, 50) = 0.78, p = .47].
Figure 1 shows that there were significant differences
in the startle response between unpleasant (50.95 ± 1.75)
and pleasant (49.14 ± 2.49) emoji [t (52) = 3.59, p = .001],
with an intermediate Cohen’s effect d = 0.495 (Cohen,
1988). No statistical differences appeared when we com-
pared ITI startle magnitude (49.83 ± 4.47) with startle
response during viewing positive [t (52) = − 0.78, p = .44]
and negative [t (52) = 1.57, p = .12] emoji’s. Significant
differences in subjective valence assessment were found
between positive (8.08 ± 1.38) and negative (3.46 ± 1.31)
emoji (t (52) = 13.31, p < .001). These data did not differ
from the data obtained in the first study for both positive [t
(242) = 1.11, p = .2 7] and negative emoji [t (242) = 0.73,
p = .47].
Discussion
This study assessed if pleasant and unpleasant emotional
face emoji modulated the ASR as the past research using
human faces in accordance with Lang’s emotional theory.
13

Fig. 1  Comparison of startle responses to pleasant and unpleasant
emoji (T-scores). ***p < .001. Bars represent standard deviations
As expected, the results showed that the emotional emoji
were associated with a significant modulation of acousti-
cally evoked startle eyeblink. The magnitude of the ASR
ITI response had an intermediate value with respect to the
magnitude observed during the vision of negative and posi-
tive emoji. These data would reaffirm that the valence of the
stimulus is relevant for the response, independently of the
possible enhancer effect of emoji arousal. The ASR was sen-
sitive to differences in emoji facial expressions as observed
in past research with human expressions (Alpers & Adolph,
2006; Anokhin & Golosheykin, 2010). The ASRs were
larger to unpleasant compared to pleasant emoji.
General discussion
The purpose of this study was to assess the emotional infor-
mation contained in emoji faces. Differences observed on
subjective ratings for these stimuli were in line with the
psychophysiological findings in startle-evoked responses
according to Lang’s motivational priming hypothesis, where
two different motivational systems are activated according
to the stimulus valence (Lang, 1995). The startle reaction
enhancement would be an effect of the organism preparation
13
Psychological Research
to aversive events and the response inhibition would be asso-
ciated with the relaxation to pleasant events. These differ-
ences in startle probe responses confirmed real valence dis-
crimination between positive and negative emoji faces. The
emoji faces linked to different pleasant and unpleasant emo-
tions were related to the ASR modulation in a similar way as
reported using human faces (Anokhin & Golosheykin, 2010;
Balaban, 1995; Guerra et al., 2012).
The expressions of emotions in face-to-face interaction
are non-verbal cues, which are not present in text-based
computer-mediated communication. Emoji play a role
as emotional substitutes in non-verbal communication.
Simulating facial expressions emoji can improve the emo-
tional information conveyed in this kind of communication
(Derks et al., 2008). The effects of emoji resembling the
human faces indicate that the valence of the emoji affects
the corresponding impression formation (Ganster, Eimler,
& Kramer, 2012). Besides, face emoji are processed in a
similar way as real faces (Churches et al., 2014). Although
the results suggest that differences in modulation of pleasant
and unpleasant emoji ASR are similar to those produced by
human faces, this finding should be replicated. It would also
be desirable to design more studies that simultaneously use
both human and emoji faces to assess their similarities and
differences.
In general, this study opens an interesting line for future
investigations in the field of emotion. This is the first study
assessing the emotional content of emoji using psycho-
physiological measures. For future investigations, it would
be interesting to investigate other peripheral indicators as
electrodermal activity or heart rate. The selected emoji for
this study were rounded faces with nearly the same color,
differing fundamentally in eyes and mouth configurations.
It is important to note that the emoji presented in the experi-
ment were bigger than the ones we usually view in our com-
munication devices. Size image effect has been studied with
different kinds of visual stimuli. According to Detenber and
Reeves (1996): “image size should affect arousal ratings, but
may not change evaluations of valence” (p. 70). Assuming
that this study is focused only in the valence dimension, the
size effect would not substantially affect the findings. How-
ever, the size image effect of the emoji should be carefully
investigated. Furthermore, emoji with different colors asso-
ciated with emotional content could cause different startle
modulation and could be incorporated as emotional mate-
rial in future studies. It is possible that the incorporation of
additional elements in the emoji pictogram, such as nose
or ears, could potentiate the emotional impact. Moreover,
due to emoji are often linked to written words, it would also
be interesting to study the emotional impact of this kind of
stimuli over different written messages.
The study had certain limitations as the low num-
ber of subjects and the fact that there were only female
Psychological Research
participants. In addition, a possible withdrawal effects
(headache, irritability) that may have occurred if partici-
pants were regular nicotine/caffeine users. As participants
could not smoke or take stimulants 12 h before the study.
Another limitation of this study is the lack of neutral
emoji. While the pleasant and unpleasant emoji seem easy
to identify subjectively, the affective “neutrality” from the
pool of emoji was more difficult to recognize; for this rea-
son, it was not included in this study. We have included
the ITIs startle amplitudes to the analysis to compare the
obtained differences between pleasant and unpleasant
pictures with a neutral condition or “baseline” level of
reactivity. If we consider that ITIs are equivalent to the
neutral responses to pictures, the ASR magnitude should
be lower than the magnitude of the negative pictures and
higher than positive pictures. Our results of the average
ASR of the ITIs were in this line, as they were placed
between the positive and negative pictures, although the
differences were not significant. This lack of differentia-
tion could be due to the lower valence effect produced
by emoji in comparison with other emotional stimuli like
IAPS pictures (International Affective Picture System). It
would be interesting to compare the emoji with other com-
monly used emotional stimulus (Aluja, Rossier, Blanch
Blanco Martí-Guiu, & Balada, 2015).
In summary, this research demonstrates that emoji
induce changes in the ASR similar to the findings reported
for human faces in line with Lang’s theory (Lang et al.,
1990). Our intention was to be cautious and first inves-
tigate the possible ASR difference depending on emoji
affective valence and later, in future studies, incorporate
the arousal and dominance dimensions. These results
represent a starting point, which other researchers can
build on, using different emoji types in research involving
emotions and ASR modulation or other neurophysiology
techniques and to be replicated in samples of men. The
emoji affective content should be considered as an object
of study for future investigations in the field of emotion.
Funding This research was supported by a grant from the “Plan
Nacional” number: PSI2015-63551-P. Ministerio de Ciencia e Inno-
vación. Spain. This research was performed within the Catalonian
Consolidated Research Group SGR 111. 2014–2016 period. Eduardo
Blanco is a “Serra Húnter Fellowship”.
Compliance with ethical standards
Conflict of interest The authors report no conflicts of interest. The au-
thors alone are responsible for the content and writing of the article.
Ethical approval All procedures performed were in accordance with
the ethical standards of the institutional research committee “Comité
Ètic de l’Hospital Aranu de Vilanova” of the University of Lleida
(Spain), and with the 1964 Helsinki declaration and its later amend-
ments or comparable ethical standards.
Informed consent Informed consent was obtained from all individual
participants included in the study.
References
Adolphs, R., Gosselin, F., Buchanan, T. W., Tranel, D., Schyns, P., &
Damasio, A. R. (2005). A mechanism for impaired fear recogni-
tion after amygdala damage. Nature, 433(7021), 68–72. https​://
doi.org/10.1038/natur​e0308​6.
Adolphs, R., Tranel, D., Damasio, H., & Damasio, A. R. (1995). Fear
and the human amygdala. The Journal of neuroscience: the official
journal of the Society for Neuroscience, 15(9), 5879–5891.
Alpers, G. W., & Adolph, D. (2006). Startle and autonomic nervous
system modulation while viewing emotional scenes or emotional
facial expressions. Psychophysiology, 43(Supplement 1), S7. https​
://doi.org/10.1016/j.biops​ycho.2009.10.001.
Aluja, A., & Blanch, A. (2011). Neuropsychological behavioral inhibi-
tion system (BIS) and behavioral approach system (BAS) assess-
ment: A shortened sensitivity to punishment and sensitivity to
reward questionnaire version (SPSRQ–20). Journal of Person-
ality Assessment, 93(6), 1–10. https​://doi.org/10.1080/00223​
891.2011.60876​0.
Aluja, A., Blanch, A., & Balada, F. (2015). Affective modulation of the
startle reflex and the reinforcement sensitivity theory of personal-
ity in females. Physiology and Behavior, 138, 332–339. https​://
doi.org/10.1016/j.physb​eh.2014.09.009.
Aluja, A., Rossier, L., Blanch Á., Blanco, E., Martí-Guiu, M., & Bal-
ada, F. (2015). Personality effects and sex differences on the Inter-
national Affective Picture System (IAPS): A Spanish and Swiss
study. Personality and Individual Differences, 77, 143–148. https​
://doi.org/10.1016/j.paid.2014.12.058.
Anokhin, A. P., & Golosheykin, S. (2010). Startle modulation by
affective faces. Biological Psychology, 83(1), 37–40. https​://doi.
org/10.1016/j.biops​ycho.2009.10.001.
Balaban, M. T. (1995). Affective influences on startle in five-month-
old infants: Reactions to facial expressions of emotions. Child
Dev, 66(1), 28–36. https​://doi.org/10.1111/j.1467-8624.1995.
tb008​53.x.
Balada, F., Blanch, A., & Aluja, A. (2014). Arousal and habituation
effects (excitability) on startle responses to the international affec-
tive picture systems (IAPS). Journal of Psychophysiology, 28,
233–241. https​://doi.org/10.1027/0269-8803/a0001​15.
Blanch, A., Aluja, A., Blanco, E., & Balada, F. (2016). Examining
habituation of the startle reflex with the reinforcement sensitivity
theory of personality: Habituation and RST. Psychophysiology,
53(10), 1535–1541. https​://doi.org/10.1111/psyp.12725​.
Blanch, A., Balada, F., & Aluja, A. (2013). Presentation and Acq-
Knowledge: An application of software to study human emotions
and individual differences. Comput Methods Programs Biomed,
110(1), 89–98. https​://doi.org/10.1016/j.cmpb.2012.10.013.
Blumenthal, T. D., Cuthbert, B. N., Filion, D. L., Hackley, S., Lipp, O.
V., & Van Boxtel, A. (2005). Committee report: Guidelines for
human startle eyeblink electromyographic studies. Psychophysiol-
ogy, 42(1), 1–15. https:​ //doi.org/10.1111/j.1469-8986.2005.00271​
.x.
Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The self-
assessment manikin and the semantic differential. Journal of
behavior therapy and experimental psychiatry, 25(1), 49–59. https​
://doi.org/10.1016/0005-7916(94)90063​-9.
Bradley, M. M., & Lang, P. J. (1999). Affective norms for English
words (ANEW): Stimuli, instruction manual and affective ratings
(Tech. Rep. No. C-1). Gainesville: University of Florida, Center
for Research in Psychophysiology.
13

Bradley, M. M., & Lang, P. J. (2000a). Affective reactions to acous-
tic stimuli. Psychophysiology, 37(2), 204–215. https ​ : //doi.
org/10.1111/1469-8986.37202​04.
Bradley, M. M., & Lang, P. J. (2000b). Measuring emotion: Behavior
feeling, and physiology. In R. D. Lane & L. Nadel (Eds.), Cogni-
tive neuroscience of emotion (pp. 242–276). New York: Oxford
University Press.
Breiter, H. C., Etcoff, N. L., Whalen, P. J., Kennedy, W. A., Rauch,
S. L., Buckner, R. L. et al. (1996). Response and habituation of
the human amygdala during visual processing of facial expres-
sion. Neuron, 17(5), 875–887. https​://doi.org/10.1016/S0896​
-6273(00)80219​-6.
Churches, O., Nicholls, M., Thiessen, M., Kohler, M., & Keage, H.
(2014). Emoticons in mind: An event-related potential study.
Soc Neurosci, 9(2), 196–202. https​://doi.org/10.1080/17470​
919.2013.87373​7.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences.
Hillsdale, NJ: Erlbaum.
Cuthbert, B. N., Bradley, M. M., & Lang, P. J. (1996). Probing picture
perception: Activation and emotion. Psychophysiology, 33(2),
103–111. https​://doi.org/10.1111/j.1469-8986.1996.tb021​14.x.
Derks, D., Bos, A. E., & von Grumbkow, J. (2008). Emoticons in
computer-mediated communication: Social motives and social
context. Cyberpsychology Behavior, 11(1), 99–101. https​://doi.
org/10.1089/cpb.2007.9926.
Detenber, B. H., & Reeves, B. (1996). A bio-informational the-
ory of emotion: Motion and image size effects on view-
ers. Journal of Communication, 46(3), 66–84. https ​ : //doi.
org/10.1111/j.1460-2466.1996.tb014​89.x.
Dunning, J. P., Auriemmo, A., Castille, C., & Hajcak, G. (2010). In
the face of anger: Startle modulation to graded facial expres-
sions. Psychophysiology, 47(5), 874–878. https​://doi.org/10.111
1/j.1469-8986.2010.01007​.x.
Duval, E. R., Lovelace, C. T., Aarant, J., & Filion, D. L. (2013). The
time course of face processing: Startle eyeblink response modu-
lation by face gender and expression. International Journal of
Psychophysiology, 90(3), 354–357. https:​ //doi.org/10.1016/j.ijpsy​
cho.2013.08.006.
Ekman, P., & Friesen, W. V. (1975). Unmasking the face; a guide to
recognizing emotions from facial clues. N.J.: Prentice-Hall.
Evans, K. C., Wright, C. I., Wedig, M. M., Gold, A. L., Pollack, M.
H., & Rauch, S. L. (2008). A functional MRI study of amygdala
responses to angry schematic faces in social anxiety disorder.
Depression and anxiety, 25(6), 496–505. https​://doi.org/10.1002/
da.20347​.
Filion, D. L., Dawson, M. E., & Schell, A. M. (1998). The psychologi-
cal significance of human startle eyeblink modification: A review.
Biological Psychology, 47(1), 1–43. https​://doi.org/10.1016/
S0301​-0511(97)00020​-3.
Ganster, T., Eimler, S. C., & Kramer, N. C. (2012). Same same but
different!? The differential influence of smilies and emoticons on
person perception. Cyberpsychology Behaviour and Social Net-
work, 15(4), 226–230. https​://doi.org/10.1089/cyber​.2011.0179.
Guerra, P., Sanchez-Adam, A., Anllo-Vento, L., Ramirez, I., & Vila, J.
(2012). Viewing loved faces inhibits defense reactions: A health-
promotion mechanism? Plos One. https​://doi.org/10.1371/journ​
al.pone.00416​31 (ARTN e41631).
Hariri, A. R., Tessitore, A., Mattay, V. S., Fera, F., & Weinberger,
D. R. (2002). The amygdala response to emotional stimuli: A
13
Psychological Research
comparison of faces and scenes. Neuroimage, 17(1), 317–323.
https​://doi.org/10.1006/nimg.2002.1179.
Khalid, S., Horstmann, G., Ditye, T., & Ansorge, U. (2017). Measur-
ing the emotion-specificity of rapid stimulus-driven attraction of
attention to fearful faces: Evidence from emotion categorization
and a comparison with disgusted faces. Psychological Research,
81(2), 508–523. https​://doi.org/10.1007/s0042​6-016-0743-8.
Lang, P. J. (1995). The emotion probe—studies of motivation and
attention. American Psychologist, 50(5), 372–385. https​://doi.
org/10.1037/0003-066x.50.5.372.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion, atten-
tion, and the startle reflex. Psychological Review, 97(3), 377–395.
https​://doi.org/10.1037/0033-295x.97.3.377.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1998). Emotion, moti-
vation, and anxiety: Brain mechanisms and psychophysiology.
Biol Psychiatry, 44(12), 1248–1263. https​://doi.org/10.1016/
S0006​-3223(98)00275​-3.
Lang, P. J., Bradley, M. M., Cuthbert, B. N., & Patrick, C. J. (1993).
Emotion and psychopathology: A startle probe analysis. Progress
in Experimental Personality & Psychopathology Research, 16,
163–199.
Lang, P. J., Davis, M., & Ohman, A. (2000). Fear and anxiety: Animal
models and human cognitive psychophysiology. Journal of Affec-
tive Disorders, 61(3), 137–159. https​://doi.org/10.1016/S0165​
-0327(00)00343​-8.
Miller, H., Thebault-Spieker, J., Chang, S., Johnson, I., Terveen, L., &
Hecht, B. (2016). “Blissfully happy” or “ready to fight”: Varying
Interpretations of Emoji. In Paper presented at the proceedings
of the international AAAI conference on web and social media
(ICWSM-2016), Cologne, Germany.
Paulus, A., Musial, E., & Renn, K. (2014). Gender of the expresser
moderates the effect of emotional faces on the startle reflex.
Cognition and Emotion, 28(8), 1493–1501. https ​ : //doi.
org/10.1080/02699​931.2014.88655​7.
Springer, U. S., Rosas, A., McGetrick, J., & Bowers, D. (2007).
Differences in startle reactivity during the perception of
angry and fearful faces. Emotion, 7(3), 516–525. https​://doi.
org/10.1037/1528-3542.7.3.516.
Toratani, Y., & Hirayama, M. J. (2011, September). Psychological
analysis of emoticons used for e-mails on cellular phones. In
Mobile IT Convergence (ICMIC), 2011 International Conference
on IEEE (pp. 49–53).
Vrana, S. R., Spence, E. L., & Lang, P. J. (1988). The star-
tle probe response: A new measure of emotion? Jour-
nal of Abnormal Psychology, 97(4), 487–491. https ​ : //doi.
org/10.1037/0021-843X.97.4.487.
Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J. (2001). Effects
of attention and emotion on face processing in the human brain:
An event-related fMRI study. Neuron, 30(3), 829–841. https:​ //doi.
org/10.1016/S0896​-6273(01)00328​-2.
Waters, A. M., Neumann, D. L., Henry, J., Craske, M. G., & Ornitz, E.
M. (2008). Baseline and affective startle modulation by angry and
neutral faces in 4–8-year-old anxious and non-anxious children.
Biological Psychology, 78(1), 10–19. https​://doi.org/10.1016/j.
biops​ycho.2007.12.005.
Whalen, P. J., Shin, L. M., McInerney, S. C., Fischer, H., Wright, C. I.,
& Rauch, S. L. (2001). A functional MRI study of human amyg-
dala responses to facial expressions of fear versus anger. Emotion,
1(1), 70–83. https​://doi.org/10.1037/1528-3542.1.1.70.