Auditory
Contagious Yawning Is Highest Between Friends
and Family Members: Support to the Emotional
Bias Hypothesis
Norscia I, Zanoli A, Gamba M, Palagi
E.
Contagious yawning differs from spontaneous
yawning because it occurs when an individual
yawns in response to someone else's yawn. In
Homo sapiens and some non-human primates
contagious yawning is higher between strongly
than weakly bonded individuals. Up to date, it
is still unclear whether this social asymmetry
underlies emotional contagion (a basic form of
empathy preferentially involving familiar
individuals) as predicted by the Emotional Bias
Hypothesis (EBH) or is linked to a top-down,
selective visual attention bias (with selective
attention being preferentially directed toward
familiar faces) as predicted by the Attentional
Bias Hypothesis (ABH). To verify whether the
visual attentional bias explained the yawn
contagion bias or not, in this study, we
considered only yawns that could be heard but
not seen by potential responders (auditory
yawns). Around 294 of auditory yawning
occurrences were extrapolated from over 2000
yawning bouts collected in free ranging humans
for over nine years. Via GLMM, we tested the
effect of intrinsic features (i.e., gender and
age) and social bond (from strangers to family
members) on yawn. The individual identity of the
subjects (trigger and potential responder) was
included as random factor. The social bond
significantly predicted the occurrence of
auditory yawn contagion, which was highest
between friends and family members. A gender
bias was also observed, with women responding
most frequently to others' yawns and men being
responded to most frequently by others. These
results confirm that social bond is per se one
of the main drivers of the differences in yawn
contagion rates between individuals in support
of the EBH of yawn contagion.
Le bâillement contagieux
diffère du bâillement
spontané car il se produit lorsqu'un
individu bâille en réponse au
bâillement de quelqu'un d'autre. Chez Homo
sapiens et certains primates non humains, le
bâillement contagieux est plus
fréquent entre les individus fortement
que faiblement liés. À ce jour, on
ne sait toujours pas si cette asymétrie
sociale sous-tend la contagion
émotionnelle (une forme de base
d'empathie impliquant
préférentiellement des individus
familiers) comme le prédit
l'hypothèse de biais émotionnel
(EBH) ou si elle est liée à un
biais d'attention visuelle sélective
(avec une attention sélective
dirigée de préférence vers
des visages familiers) comme le prédit
l'hypothèse de biais attentionnel (ABH).
Pour vérifier si le biais visuel
d'attention expliquait ou non le biais de
contagion du bâillement, dans cette
étude, les auteurs ont
considéré uniquement les
bâillements qui pouvaient être
entendus mais non vus par les intervenants
potentiels (bâillements auditifs). Environ
294 occurrences de bâillements auditifs
ont été extrapolées
à partir de plus de 2000 épisodes
de bâillements collectés chez des
humains pendant plus de neuf ans. Via GLMM, ils
ont testé l'effet des
caractéristiques intrinsèques
(c'est-à-dire le sexe et l'âge) et
le lien social (des étrangers aux membres
de la famille) sur le bâillement.
L'identité individuelle des sujets
(déclencheur et répondeur
potentiel) a été incluse comme
facteur aléatoire. Le lien social a
prédit de manière significative la
survenue d'une contagion auditive du
bâillement, qui était la plus
élevée entre amis et membres de la
famille. Un parti pris sexiste a
également été
observé, les femmes répondant le
plus souvent aux bâillements des autres et
les hommes déclenchant plus souvent le
bâillement des autres. Ces
résultats confirment que le lien social
est en soi l'un des principaux moteurs des
différences de taux de contagion du
bâillement entre les individus à
l'appui de l'EBH de la contagion du
bâillement.
-Palagi E,
Leone A, Mancini G, Ferrari PF. Contagious
yawning in gelada baboons as a possible
expression of empathy. Proc Natl Acad Sci USA.
2009;106(46):19262-19267
-Palagi E,
Norscia I, Demuru E. Yawn contagion in
humans and bonobos: emotional affinity matters
more than species Peer J 2:e519
-Palagi E,
Norscia I, Cordoni G. Lowland gorillas
(Gorilla gorilla gorilla) failed to respond to
others' yawn: Experimental and naturalistic
evidence. J Comp Psychol. 2019
-Palagi
E, Cordoni G. Intraspecific Motor and
Emotional Alignment in Dogs and Wolves: The
Basic Building Blocks of
Dog&endash;Human
Affective Connectedness. Animals2020;10(2):241
Yawning is an involuntary sequence of mouth
opening, deep inspiration, brief apnea, and more
or less slow expiration (Baenninger, 1997;
Walusinski and Deputte, 2004; Guggisberg et al.,
2010; Krestel et al., 2018). When elicited, a
yawn cannot be totally suppressed. Therefore, it
has been defined as a stereotyped or reflex-like
pattern (Lehmann, 1979; Provine, 1986). In Homo
sapiens, several hypotheses have been put forth
with variable support to explain mechanisms and
functions of spontaneous yawning, such as
oxygenation (respiratory function caused by
hypoxia), stress-related behavior (caused by
arousal), or thermoregulation (caused by
hyperthermia; Guggisberg et al., 2010; Massen et
al., 2014; Gallup and Gallup, 2019). Being a
physiological response, yawning can be affected
by internal and external factors such as the
time of the day (Giganti and Zilli, 2011) or
intracranial/brain temperature (Gallup and
Eldakar, 2013).
Yawning can be self-directed and/or
displayed to others (Moyaho et al., 2017; Palagi
et al., 2019). In human and non-human primates,
depending on the species, when yawning is shown
to others, it can communicate threat (Troisi et
al., 1990; Deputte, 1994) and/or physiological
and behavioral changes (Provine et al., 1987;
Leone et al., 2015; Zannella et al., 2015). In
humans, yawning is a socially modulated response
because it can be inhibited by actual&emdash;and
not virtual&emdash;social presence (Gallup et
al., 2019) and because a yawn can be triggered
by someone else's yawn, as a result of a
phenomenon known as contagious yawning (Provine,
1989, 2005). Yawn contagion can be elicited even
if the yawn is heard but not seen (Arnott et
al., 2009; Massen et al., 2015).
In humans, their phylogenetically closest
ape species (chimpanzees: Pan paniscus; bonobos:
Pan troglodytes) and the African monkey
Theropithecus gelada, contagious yawning is not
only present (Provine, 1986; Palagi et al.,
2009; Tan et al., 2017; but see: Amici et al.,
2014) but also socially modulated because the
yawning response is highest when certain
categories of individuals are involved (e.g.,
kin, group-members, dominants; Palagi et al.,
2009; Campbell and de Waal, 2011, 2014; Norscia
and Palagi, 2011; Demuru and Palagi, 2012;
Massen et al., 2012). Two main arguments have
been presented to explain this social asymmetry
in contagious yawning, which have been grouped
into two main hypotheses: the Emotional Bias
Hypothesis (EBH), linking contagious yawning to
emotional transfer, and the Attentional Bias
Hypothesis (ABH), which considers contagious
yawning as a motor response that is subject to
differences in top-down attentional processes
(Palagi et al., 2020).
The EBH predicts that the social asymmetry
observed in yawn contagion rates reflects
differences in the different social bonding, a
proxy of emotional bonding, between individuals.
This hypothesis is supported by evidence that
yawn contagion rates follow an empathic trend
(sensu Preston and de Waal, 2002), being highest
between individuals sharing a strongest
emotional bond. Specifically, Norscia and Palagi
(2011) found that in humans yawn contagion rates
are greatest in response to kin and friends than
in response to acquaintances and strangers. In
adult chimpanzees yawn contagion is higher
between in-group compared to out-group members
(Campbell and de Waal, 2011) and in bonobos yawn
contagion rates are greatest between individuals
that affiliate more with one another (Demuru and
Palagi, 2012). In a comparative investigation
including both humans and bonobos, Palagi et al.
(2014) found that the yawn contagion rates were
affected by the relationship quality between
individuals more than by the species the
subjects belonged to. Additionally, in humans
yawn contagion increases with age when the
ability to identify others' emotions increases
and declines with old age when such ability
declines (Wiggers and van Lieshout, 1985;
Anderson and Meno, 2003; Saxe et al., 2004;
Singer, 2006; Millen and Anderson, 2011;
Bartholomew and Cirulli, 2014). Yawn contagion
rates increase from infancy to adulthood also in
chimpanzees (Madsen and Persson, 2013).
The ABH predicts that the social asymmetry
observed in yawn contagion can be due to
differences in social, visual attention (Massen
and Gallup, 2017). In particular, highest levels
of contagious yawning would be due to the extra
top-down, selective visual attention paid to
individuals that are more relevant to the
observer, such as familiar subjects, as it
occurs in humans and geladas, or dominants, as
it occurs in chimpanzees or bonobos (Yoon and
Tennie, 2010; Massen et al., 2012; Massen and
Gallup, 2017). According to Massen and Gallup
(2017), ABH would be backed up by the existing
evidence on the different visual detection and
visual perceptive encoding of faces of familiar
and/or in-group subjects compared to unfamiliar
ones (e.g., Buttle and Raymond, 2003; Ganel and
Goshen-Gottstein, 2004; Jackson and Raymond,
2006; Michel et al., 2006).
In this study, we analyzed data on yawning
collected over 9 years on humans in their
natural settings and we extrapolated the cases
in which the yawn emitted by a subject could be
heard but not seen by a potential responder
(auditory yawn). By considering only the cases
in which the visual cue of the yawning stimulus
was not detectable, we verified whether the
social asymmetry previously observed in yawn
contagion rates persisted or not. In particular,
we tested the following alternative predictions
derived from the two hypotheses presented above
(EBH and ABH).
Prediction 1a: according to the EBH, the
rates of yawn contagion are influenced by the
strength of the inter-individual social
bond&emdash;a proxy of the emotional
bond&emdash;per se and not by a different
top-down, selective visual attention paid to
certain individuals in particular. If this
hypothesis is supported, we expect to observe
the social bias also when the visual cue of the
yawning stimulus is excluded and the rates of
auditory contagious yawning to be higher between
strongly bonded compared to weakly bonded
individuals.
Prediction 1b: according to the ABH, the
higher levels of yawn contagion between strongly
bonded compared to weakly bonded individuals
would be linked to the closest top-down,
selective visual attention that individuals pay
to individuals that are relevant to them, e.g.,
family and friends. If this hypothesis is
supported, the social bias observed in the yawn
contagion should disappear when only auditory
yawns are considered because the visual cue
cannot be attended by the potential
responder.
Discussion
This study shows for the first time that
yawn contagion is significantly affected by the
social bond between individuals (Table 1) even
when the triggering stimuli are auditory yawns,
which we defined as vocalized yawns that could
be heard but not seen (visual cue undetectable,
auditory cue detectable). In particular,
auditory contagious yawning is significantly
more frequent between kin and friends than
between strangers and acquaintances (Figure 1).
This finding supports prediction 1a based on the
EBH and not prediction 1b based on the ABH,
leading to the conclusion that in humans
top-down, selective visual attention cannot be
the main driver of the social asymmetry observed
in yawn contagion rates (Norscia and Palagi,
2011; Norscia et al., 2016a). Also the sex of
the trigger and the sex of the receiver had a
significant effect on yawn contagion rates, with
men&emdash;as triggers&emdash;being responded to
by others more frequently than women (Figure 2)
and women responding more frequently to others'
yawns than men (Figure 3).
Contrary to Bartholomew and Cirulli (2014),
we found no age effect on yawn contagion, most
probably because our database on auditory yawns
had a strong prevalence of adults (25&endash;64
years old). The highest levels of auditory yawn
contagion in women compared to men confirm the
gender bias observed in naturalistic conditions
on humans susceptible to yawn contagion by
Norscia et al. (2016a, b) when considering a
larger dataset that also included yawns that
could be seen by the potential responder (with
yawning sensory modality&emdash;vision, hearing,
or both&emdash;not affecting the response).
The gender bias is also in partial agreement
with previous results obtained in controlled
settings, including the visual cue (Chan and
Tseng, 2017; but see Norscia and Palagi, 2011;
Bartholomew and Cirulli, 2014). It has been
hypothesized that the high degree of yawn
contagion in women might inform emotional
contagion (Norscia et al., 2016a), in the light
of reportedly higher empathic
abilities&emdash;related to maternity&emdash;of
women compared to men (Christov-Moore et al.,
2014). However, this issue is still under debate
because cultural differences across human
societies can mold social bonding dynamics in a
different way. It is therefore complicate, at
this stage of knowledge, to disentangle cultural
factors, inter-personal relationship quality,
and gender influence in the distribution of yawn
contagion. Our results also show that men
perform better than women as triggers and the
most parsimonious hypothesis for this might be
that men's vocalizations can be better heard in
natural settings, often characterized by
background noises. Indeed, the perception of
voice gender primarily relies on the fundamental
frequency that is on average lower by an octave
in male than female voices, with lower frequency
vocalizations traveling further than high
frequency ones (Marten and Marler, 1977; Latinus
and Taylor, 2012). However, to our knowledge,
there is no specific study addressing the
possible gender bias in yawn audibility and
further investigation with experimental trials
in controlled condition is therefore necessary
to verify this speculation.
In this study, we also found that the
differences in yawn contagion rates across
categories (family and friends, strangers, and
acquaintances) cannot be explained by
differences in top-down, selective visual
attention. This finding is in line with previous
literature. Contagious yawning appears to
involve brain areas that are more related to the
orienting-bottom up network [temporoparietal
junction (TPJ), brainstem nuclei, ventrolateral
prefrontal cortex (vlPC)] than top-down
related areas [frontal eye fields (FEFs),
intraparietal sulcus (IPS), parietal areas; for
a review: Palagi et al., 2020]. Moreover,
yawn contagion is neither sensitive to the
sensory cues present in the signal (auditory,
visual, or audio-visual) (Arnott et al., 2009;
Norscia and Palagi, 2011) nor affected by the
visual perspective of the triggering stimulus
(yawns in orientations of 90°, 180°,
and 270° are able to trigger yawning
responses as frontal, 0° yawns; Provine,
1989, 1996). Chan and Tseng (2017) found that
the ability to detect a yawn as such (perceptual
detection sensitivity) was related to the
duration of gaze to the eyes of the stimulus
releasing face, but eye-gaze patterns were not
able modulate contagious yawning. In
chimpanzees, contagious yawning frequencies were
highest between same-group than different-group
individuals, even if the responders looked
longer at out-group chimpanzee videos (Campbell
and de Waal, 2011). The argument that visual
selective attention can bias yawn contagion
rates in a specific direction (subjects
responding more to family than strangers) is
also undermined by the absence of any specific
pattern of social attention in human and
non-human primate. Via eye-tracking (applied to
measure the viewing time) and by showing unknown
faces to their experimental subjects,
Méary et al. (2014) observed that humans
were skewed toward own-race faces whereas rhesus
macaques' attention was more attracted by new
than by same species faces. Kawakami et al.
(2014) observed that human subjects paid more
attention to the eyes of ethnic in-group members
and to nose and mouth of ethnic out-group
members.
The same study also revealed that visual
attention did not depend on the target race. By
measuring how long the experimental subjects
gazed at the screen, Whitehouse et al. (2016)
observed that Barbary macaques paid more
attention to scratching videos of non-stranger
than stranger individuals but also noted that
within the non-strangers, macaques paid most
attention to those individuals with which they
shared a weak social bonding. By measuring
glance rates, Schino and Sciarretta (2016)
observed that mandrills looked more at their own
kin than at non-kin but also more at dominant
than at subordinate group mates. Therefore,
these studies (used to support ABH) describe no
single pattern of selective attention. One
further important point to consider is the very
definition of familiarity and group-membership
adopted by most of the studies used to support
ABH (Massen and Gallup, 2017). These studies
showed better visual detection and visual
perceptive encoding of faces of
familiar/in-group subjects compared to
unfamiliar ones but defined familiarity and
group membership not on the basis of the
personal relationships between individuals.
Instead, familiarity or group-membership were
defined on the ground of indirect knowledge
(e.g., photo of famous people or of a subject
already shown in pre-trial phases) or common
race (e.g., Buttle and Raymond, 2003; Ganel and
Goshen-Gottstein, 2004; Jackson and Raymond,
2006; Michel et al., 2006). This definition is
fine for the purposes of these studies but it is
not as much fine if the results are used to
propose alternative explanations for the
influence that real social bonding&emdash;based
on real relationships&emdash;may have on a
phenomenon, in this case yawn contagion. For
example, Michel et al. (2006) observed that
Caucasian and Asian subjects could better
recognize same-race faces but this difference
was not present in Asian subjects who had been
living for about a year among Caucasians.
Another point of discussion concerns the
presence of yawn contagion in children with
autism spectrum disorder (ASD), which frequently
show alterations in visual attention (Richard
and Lajiness-O'Neill, 2015). In ASD children,
yawn contagion can be absent (Senju et al.,
2007), impaired (Helt et al., 2010), or similar
to typically developing children when the
subjects are induced to redirect their attention
the video stimulus during the experimental
trials (Usui et al., 2013). In a recent study,
Mariscal et al. (2019) found that yawn contagion
in ASD children was positively related to the
blood concentration of oxytocin, the hormone
involved in parental and social attachment
(Decety et al., 2016) and posited that yawn
contagion in ASD children may be related to
variable mean oxytocin concentrations across
different study cohorts (Mariscal et al., 2019).
This finding is in line with the EBH hypothesis
that links yawn contagion rates to social
bonding, which can reflect emotional
bonding.
Conclusion
Our study adds to the discussion over the
mechanisms underlying the social asymmetry in
yawn contagion (for a critical reviews: see
Adriaense et al., 2020; Palagi et al., 2020), by
showing that yawn contagion is probably
associated with bottom-up, rather than with
top-down, selective attention. Bottom-up
attention is primarily lead by the sensory
perception of the eliciting stimulus whereas
top-down, selective attention is a voluntary,
sustained process in which a particular item is
selected internally and focused upon or examined
(Katsuki and Constantinidis, 2014). In this
respect, the acoustic stimulus (auditory yawn)
emitted by the trigger was heard and could
elicit a yawning response in the receiver, even
though the receiver was not paying any voluntary
visual attention to the trigger. Moreover, the
yawning response rates were socially modulated,
with auditory yawn contagion being highest in
individuals that were most strongly bonded to
one another. Hence, top-down selective attention
is not the main driver of the social asymmetry
observed in yawn contagion, which appears to be
a stimulus driven phenomenon-related to
bottom-up attention processes. Further
investigation is necessary to understand whether
and in what way other forms of attention or
pre-attentive stages are able to affect yawn
contagion.
