A growing number of studies on non-human
animals have documented that stressors modulate
the expression of yawning. In particular, recent
experimental research shows that yawns are
initially inhibited following physical stress,
but then become potentiated thereafter. However,
stress-induced yawning in humans has yet to be
demonstrated experimentally. Here, we
investigated the temporal relationship between
self-reported contagious yawning and an acute
physical stressor in 141 human subjects in the
laboratory.
Using a 2 _ 2 between-subjects design,
participants either underwent the cold pressor
test (CPT) or a matched control condition prior
to viewing a contagious yawning stimulus that
was either displayed immediately thereafter or
following a 20-min delay. Consistent with the
comparative literature, we show an interaction
between stress and time conditions, whereby both
the incidence and frequency of yawning are
lowest in the immediate-CPT trials and highest
in the delayed- CPT trials. These findings
support a homologous effect of acute physical
stress on yawning across birds and mammals that
may be related to an adaptive thermoregulatory
and arousal function.
Résumé
Un nombre croissant d'études animales
ont documenté que les facteurs de stress
modulent l'expression du bâillement. En
particulier, des recherches
expérimentales récentes montrent
que les bâillements sont initialement
inhibés à la suite d'un stress
physique, mais qu'ils augmentent de
fréquence par la suite.
Le bâillement causé par le
stress chez l'homme n'a pas encore
été démontré
expérimentalement. Les auteurs ont
étudié la relation temporelle
entre le self-report de contagion du
bâillement et un stress physique
aiguë chez 141 sujets humains (en
laboratoire).
Les participants ont subi une exposition au
froid et une condition de contrôle c'est
à dire de visualisation de
bâillements qui a été
montrée immédiatement à la
suite du stress ou après un délai
de 20 minutes.
Après une revue de la
littérature, les auteurs montrent une
interaction entre les conditions de stress et de
temps, où l'incidence et la
fréquence du bâillement sont plus
faibles immédiatement après le
stress et plus fréquents à
distance. Ces résultats confirment
l'analogie des effets du stress physique sur le
bâillement des oiseaux et des
mammifères. Celui-ci pourrait être
lié à une fonction de
thermorégulation et de stimulation de la
vigilance d'après ces auteurs qui ne
mentionnent pas la nature parasympathique du
bâillement opposé au stress de
nature adrénergique.
Introduction
Yawning is characterized by a powerful
gaping of the jaw with inspiration, following by
a brief period of muscle contraction and a
passive closure of the jaw with expiration
(Barbizet 1958). Phylogenetically old, yawns or
similar yawn-like mandibular gaping patterns
have been observed across vertebrate classes
(e.g., Baenninger 1987). Numerous hypotheses
have been proposed to explain the adaptive
function of yawning (Smith 1999), and this topic
is still debated (Guggisberg et al. 2010; Gallup
2011). To date, however, empirical
investigations primarily support a role of
yawning in promoting arousal and state change
(Baenninger 1997; Provine 1986, 2005; Vick and
Paukner 2010; Walusinski 2006) through enhanced
intracranial circulation and brain cooling
(Gallup and Gallup 2007; Shoup-Knox et al. 2010;
Walusinski 2014). Consistent with this view, a
growing number of studies have documented a
close connection between yawning and various
stressors across diverse species of mammals and
birds. Research on non-human primates indicates
that yawning tends to increase during periods of
psychosocial stress (Maestripieri et al. 1992).
In particular, various studies have documented
that yawns occur during antagonist interactions
and hostile social situations (Macaca nigra,
Hadidian 1980; Theropithacus gelada, Leone et
al. 2014), whereby dominant males perform
directed threat yawns with canine displays and
subordinates may yawn in response to the
stressful interaction (Cercocebus albigena and
Macaca fascicularis, Deputte 1994; Macaca
fascicularis and Macaca fuscata, Troisi et al.
1990; Redican 1982).
However, unlike ordinary yawns (see
definition above), the display yawner rather
than closing their eyes at the peak of the yawn,
fixes their attention on the target during the
yawning episode to monitor the effect of the
threat. These social displays are typically
documented among non-human primate species with
sexual dimorphism in body size, canine size, and
aggressive competition, and were first described
by Charles Darwin in The Expression of the
Emotions in Man and Animals (Darwin 1872).
Importantly, directed threat yawns appear to be
fundamentally different from more ubiquitous
forms of yawning related to transitions in
arousal, and the sex differences in yawn
frequency among primates are lost within species
with limited sexual dimorphism in canine size
(Homo sapiens, Schino and Aureli 1989; Pan
troglodytes, Vick and Paukner 2010).
Nonetheless, spontaneous yawns have also been
linked with more general social and
environmental stressors in primates.
A study on the behavioral responses of
female olive baboons (Papio anubis) showed that
the mere presence of a dominant conspecific
increased yawning rates by 40% in comparison to
when the nearest neighbor was a subordinate
(Castles et al. 1999). Similarly, chimpanzees
(Pan troglodytes) in the wild have been observed
to yawn more frequently when in the presence of
humans (Goodall 1968). Chimpanzees in captivity
have also been shown to increase spontaneous
yawning and self-directed behaviors following
vocalizations and noisy displays from
neighboring groups of conspecifics, prompting
the interpretation of yawning as a behavioral
indicator of anxiety in this species (Baker and
Aureli 1997). Stress also appears to influence
the temporal expression of this response among
bonobos (Pan paniscus), with spontaneous yawns
occurring at highest frequency during
non-stressful situations and being least
frequent during periods immediately following
social stress (Demuru and Palagi 2012).
Furthermore, non-social stressors have been
documented to elicit spontaneous yawning in
other primates, including grey-cheeked mangabeys
(Cercocebus albigena) and, most recently, lemurs
(Lemur catta) (Zannella et al. 2015).
Specifically, yawns in these species tend to
increase following alarm calling in response to
predators or after predatory attacks or
aggression.
Yawning has been implicated as a behavioral
sign of stress and anxiety across other mammals
as well. For example, one study reported that
yawning increased in frequency among wild horses
following aggressive behaviors in semi-natural
conditions (Górecka-Bruzda et al. 2016).
Another recent study investigating the effects
of noseband tightening on horse behavior found
that heart rate and eye temperature rose
significantly during the tightest condition, and
that following the removal of the bands yawn
frequency increased (Fenner et al. 2016).
Furthermore, a mild rise in yawning has been
observed among domesticated dogs following
certain stressful stimuli (Beerda et al. 1998),
leading to various investigations of yawning as
a behavioral measure of anxiety. However, the
evidence connecting yawning and stress in dogs
is mixed, with subsequent studies providing
limited support for changes in yawn behavior
following various stressful conditions in dogs
(e.g., Dreschel and Granger 2005; Hennessy et
al. 1998; Part et al. 2014; Schilder and van der
Borg 2004). Furthermore, less than 10% of owners
identify yawning as a marker of stress in their
own pets (Mariti et al. 2012).
Controlled laboratory experiments on rodents
have allowed researchers to further examine a
potential causal link between stressors and the
expression of yawning. For example, a recent
study demonstrated that classical fear
conditioning trials reliably induced yawning,
along with anxiety-related behavior and activity
in the central nucleus of the amygdala among
male Wistar rats (Kubota et al. 2014). A
separate study revealed that the repeated
witnessing of painful shocks delivered to
conspecifics increased yawning frequency in male
Long-Evans rats (Carrillo et al. 2015). This
report also showed that the heightened yawning
following the observation of pain in others was
effectively inhibited by metyrapone, an
anti-stress drug and glucocorticoid synthesis
blocker, providing further support that yawns
were an affective response to the repeated
observation of distress in others. Other work
has shown that adrenalectomy, which stops the
production of glucocorticoids, abolished yawning
in male Winstar rats
(Anías-Calderón et al. 2004). This
same study also demonstrated that the
administration of a synthetic glucocorticoid,
dexamethasone, restored yawning within the same
adrenalectomized rats. An increase in yawning
has also been documented following rapid rises
in ambient temperature that produce thermal
stress in Sprague Dawley rats (Gallup et al.
2011). Furthermore, studies have shown that the
pattern of physical stressors experienced by
rats modulates the expression of yawning. For
example, constant swimming and foot shock
stressors appear to inhibit drug-induced yawning
in albino Wistar rats, while exposure to the
same stressors at an intermittent basis
significantly increased this response (Tufik et
al. 1995). Similarly, experiments on Sprague
Dawley rats showed that experiencing foot-shocks
at fixed 10-min intervals produced a gradual
increase in yawning that peaked at 40-min, while
yawning was less pronounced when rats
experienced foot shocks at random levels (Moyaho
and Valencia 2002).
Comparative experiments in birds have shown
similar findings pertaining to yawning and
stress, both in relation to frequency and
temporal effects. In a series of experiments
designed to induce thermal stress in captive
budgerigars (Melosittacus undulatus), increases
in yawning were reliably produced by rapid rises
in ambient temperature within a thermal chamber
(Gallup et al. 2009, 2010). The pattern of
yawning has also been measured in this species
following a 4-min handling stressor, which
simulates a predator encounter and produces a
significant physiological stressor as measured
by rises in body temperature (Miller et al.
2010). Within this study it was shown that, in
comparison to a control condition, handling
significantly modulated the temporal expression
of yawning following the encounter (Miller et
al. 2010). In particular, yawns were initially
inhibited after release (~1.5 yawns/h), but were
then potentiated 20&endash;40-min thereafter
(>5 yawns/h). A follow-up study also using
budgerigars examined the effects of a
non-specific stressor, a loud white noise that
elicited clear startle responses, on yawn
frequency within small groups (Miller et al.
2012). In comparison to handling stress, the
loud noise did not produce the same immediate
reduction in yawning, but a similar increase in
the frequency of this response occurred 20-min
thereafter. A more recent paper documents
similar temporal effects of stressors on yawning
in wild Nadza boobies (Sula granti) (Liang et
al. 2015). In the first of two studies, yawning
was measured in adult birds during and after an
extended human capture-restraint stressor that
produces a rise in corticosterone. Consistent
with other studies, yawns were absent during the
stressor itself, and remained at low frequency
from 0 to 30-min following release before
significantly climbing in rate 30&endash;60-min
thereafter. In the second study, the researchers
simply observed behavioral responses of Nadza
booby nestlings following maltreatment of
non-parental adults. None of the nestlings
yawned during the stressful event itself, but
nearly every nestling yawned within 15-min
afterwards.
Overall, a large and growing comparative
literature on diverse species of mammals and
birds generally implicates stress as a trigger
for overall increases in yawning, but that
one-time acute physical stressors appear to
initially inhibit, then subsequently potentiate
this response. The physiological effects of
stress are diverse, with a number of components
involved in the stress system that lead to
behavioral and peripheral changes that function
to equilibrate homeostasis and adaptive outcomes
(Tsigos and Chrousos 2002). The stress response
consists of two major components 1) the locus
coeruleus/norandrenergic (LC/NE) sympathetic
nervous system (SNS) pathway which releases
norepinephrine from the adrenal medulla
immediately after stress (Itoi and Sugimoto
2010) and 2) the hypothalamic-pituitary-adrenal
(HPA) axis whose end product is cortisol which
is released from the adrenal cortex at peak
levels approximately 20 min after the cessation
of a stressor (Spencer and Deak 2016). Stress
also produces rises in body temperature (e.g.,
Zethof et al. 1994, 1995; Van der Heyden et al.
1997; Olivier et al. 2003), which could be a
result of hyperthermia or fever. Hyperthermic
responses from stress could occur from increased
locomotor or muscular activity and cutaneous
vasoconstriction, while fever would be a
consequence of a raised thermoregulatory set
point (Oka et al. 2001).
Although yawning has been associated with
stress in humans, and even hypothesized to be
linked with rises in cortisol (see Thompson
Cortisol Hypothesis, e.g., Thompson 2011), to
this point the connection has only been
indirect. For example, increased yawning in
humans has been noted prior to prolonged
stressful or anxiety provoking situations, such
as among Olympians immediately prior to
competition, musicians waiting to perform, and
paratroopers leading up to their first free-fall
(Provine 2005). Frequent yawning has also been
documented among individuals with some anxiety
disorders (Daquin et al. 2001), as well as a
variety medical conditions and neurological
diseases (Gallup and Gallup 2008; Walusinski
2009). One study even found that a small
proportion of college students self-reported
that stressful situations are conducive to
yawning (Greco et al. 1993). To date, however,
we are unaware of any experimental studies
assessing this relationship.
Here, we describe the results of an
experiment investigating how the direct
manipulation of an acute physical stressor
alters the frequency and temporal expression of
yawning among human participants in the
laboratory. In particular, we examined how the
cold pressor test (CPT), which included the
immersion of the non-dominant hand into ice
water (2-4 °C) for at least 60-s, alters
yawning in comparison to a control condition
(immersion into room temperature water 20- 22
°C) at distinct time intervals: immediately
following the test and 20-min thereafter. As a
physical stressor, the CPT task is more likely
to preferentially activate early autonomic
(LC/NE) stress pathways over later-responding
psychological (HPA) stress pathways; however,
these systems are largely interconnected with a
large degree of cross-activation. Accordingly,
the CPT has been shown to not only increase a
biomarker (Banks et al. 2014) and peripheral
measures of autonomic arousal (Schwabe et al.
2008), but to also increase cortisol levels with
peak activity occurring after a delay (³10 min
post-stress) (Alomari et al. 2015; Schwabe et
al. 2008; Viena et al. 2012). Consistent with
previously documented temporal effects following
physical stressors in horses (Fenner et al.
2016), rats (Moyaho and Valencia 2002) and birds
(Miller et al. 2010, 2012; Liang et al. 2015),
we hypothesized that yawning would increase
following the CPT, but only in the delayed
condition. Because yawning has been documented
to be relatively infrequent among participants
in laboratory research (Baenninger and Greco
1991), contagious yawning was used as a proxy
for spontaneous yawning in this experiment.
Contagious yawns appear indistinguishable from
spontaneous yawns aside from their triggers
(i.e., social vs. physiological), and contagious
yawning can be easily manipulated under
laboratory conditions (e.g., Platek et al.
2003). Furthermore, growing research shows that
physiological variables that alter spontaneous
yawn frequency (i.e., those that influence brain
and body temperature) have the same effects on
yawn contagion, suggesting that both yawn types
share fundamental mechanistic pathways (Gallup
and Gallup 2007; Gallup and Gallup 2010; Gallup
2016).
Discussion
This is the first experiment, to our
knowledge, to directly assess the effects of
stress on yawning in humans. Consistent with the
comparative literature, we demonstrated that an
acute physical stressor significantly modulated
this response among participants in the
laboratory. Although brief ice water immersion
of the non-dominant hand (CPT) did not produce
an overall increase in the proportion or
frequency of yawning, it did alter the time
course of its expression. Specifically, we show
that within the CPT condition self-reported
contagious yawns occur at relatively low
frequency immediately following exposure to the
stressor, but then increase 20-min thereafter.
These findings are similar to previous results
reported in rats, horses and bonobos (Moyaho and
Valencia 2002; Fenner et al. 2016; Demuru and
Palagi 2012), and specifically match recently
documented effects resulting from handling and
capture restraint stress in two bird species
(Miller et al. 2010; Liang et al. 2015).
There is growing evidence indicating that
the motor action pattern of yawning functions to
cool the brain through increased intracranial
circulation and countercurrent heat exchange
with the ambient air (reviewed by Gallup and
Gallup 2008; Gallup and Eldakar 2013). For
example, predicted patterns of brain/skull
temperature change surround yawns in rats
(Shoup-Knox et al. 2010) and humans (Gallup and
Gallup 2010), and experimentally manipulated
behavioral brain cooling mechanisms diminish
spontaneous and contagious yawn frequency among
human participants in the laboratory (Gallup and
Gallup 2007; Gallup and Gallup 2010). Further
support for this hypothesis comes from the
repeated demonstration that spontaneous and
contagious yawn frequency can be effectively
increased or diminished as a function of ambient
temperature manipulation and variation across
diverse species (Gallup et al. 2009, 2010, 2011;
Gallup and Eldakar 2011; Massen et al. 2014;
Eldakar et al. 2015; for a review see Gallup
2016), as well as the close connection between
yawning and thermoregulatory dysfunction (Gallup
and Gallup 2008). Stress-induced increases in
yawning behavior may reflect LC/NE activation
since activation of this system quickly elevates
core body temperature (Chrousos 1998).
Different explanations have been posited to
explain the increase in body temperature
following stress. For example, the central
autonomic ganglia that are activated as part of
the LC/NE stress response trigger a widespread
cascade of Bfight or flight^ responses that
could produce hyperthermia due to increased
muscular activity and cutaneous
vasoconstriction. Alternatively, research
indicates that stress-induced rises in
temperature could also result from a febrile
response due to prostaglandin E2-dependent and
5- HT-mediated mechanisms (Oka et al. 2001).
Although we did not obtain physiological
measures in the current study, the CPT is a
physical stressor that clearly reduces skin
temperature of the submerged hand (Brusselmans
et al. 2015). Despite this localized reduction,
the CPT produces vasoconstriction and has been
shown to result in an associated rise in the
skin temperature of the non-cold exposed hand
(Frank and Raja 1994).
However, other studies show either no rise
in skin temperature (Edelson and Robertson
1986;Washington et al. 2000) or even a decrease
in skin temperature in spite of concomitantly
measured vasoconstriction (Watson and Nance
1994). This apparently contradictory
relationship can be explained by a cold-induced
constriction of hand arterioles that can produce
a redirection of blood through arteriovenous
anastomoses (AVA) resulting in heat loss from
the skin surface despite a rosy skin appearance
caused by constriction of the AVA (Chwa_czy_ska
et al. 2015; Mizeva et al. 2015; Walløe
2016). Accordingly, we argue that either model
of stress-induced temperature elevation can
effectively explain the temporal pattern of
yawning witnessed here. In terms of explaining
the hyperthermic response, animals would need to
balance the cost/benefit trade-offs associated
with yawning immediately following stressful
situations. For example, previous research on
budgerigars suggested that the initial
suppression of yawns following handing restraint
might adaptively reduce attention-getting
movements and promote effective antipredatory
behaviors (Miller et al. 2010).
However, birds in that study showing the
greatest increases in body temperature after
handling yawned sooner following release.
Furthermore, stress produces increases in heart
rate and blood pressure and adaptive behavioral
changes for heightening arousal, alertness and
vigilance (Chrousos 1998), which would
effectively inhibit the natural mechanisms
triggering yawns (Baenninger 1997; Gallup and
Gallup 2007). Following a brief recovery period,
however, an increase in yawning may reflect a
mechanism to provide thermoregulatory benefits
and effectively maintain waning vigilance and
arousal when external threats have subsided and
the environment becomes more predictable.
Alternatively, if the CPT stress induced a fever
through a rise in the thermoregulatory set
point, thermoregulatory warming mechanisms would
be triggered initially and cooling mechanisms
would be inhibited. Following the removal of the
stressor and reduced febrile response over time,
the thermoregulatory set point would then be
reduced, and cooling mechanisms, such as
yawning, would be activated to promote thermal
homeostasis. These effects on yawn frequency and
fever have previously been described for the use
of antipyretics (Gallup and Gallup 2013). Other
evolutionary explanations for the previously
observed temporal effects on yawn expression
following stress in birds include communicating
and signaling arousal reduction to group members
(Guggisberg et al. 2010; Liang et al. 2015).
Multiple functions are possible, whereby yawning
could provide thermoregulatory benefits and
serve as a signal to conspecifics. However,
aside from direct threat yawns with canine
displays in non-human primates, there is
currently no evidence that yawns provide a
meaningful signal to receivers (see Gallup and
Clark 2015). For example, group members do not
appear to orient towards or respond to yawns of
others, and it is not clear what communicative
benefits there would be to yawning. Although
contagious yawns are inherently social, in that
sensing yawns in others triggers them, this does
not mean the action pattern actually
communicates anything. Furthermore, since yawns
occur under a variety of contexts outside of
stress (i.e., during changes in arousal and
important events, before and after sleep, during
boredom, transitions in activity patterns), any
signal that is displayed remains nonspecific
(Gallup 2011). Therefore, experimental research
is needed to test the predictions of
social-communication hypotheses.
The results reported here could also be
interpreted under the recent Thompson Cortisol
Hypothesis (Thompson 2011), which posits that
yawns are triggered due to rises in cortisol.
While this hypothesis may not serve as a global
explanation for all forms of yawning, it could
provide insight into the specific association
between yawning and stress. In particular, the
20-min delay in yawn enhancement following the
CPT is consistent with the temporal rises of
cortisol observed following stress in humans
(Alomari et al. 2015; Cornelisse et al. 2011;
Klopp et al. 2012). However, it is important to
note that elevations in cortisol resulting from
the CPT are relatively low when compared to
other tests (reviewed by Schwabe et al. 2008).
It is possible that cortisol is not the direct
mediator of the stress-induced yawning response
since the activity of cortisol as part of the
delayed stress response recruits, and is
co-activated with, a host of many
neuromodulators (e.g. serotonin, dopamine) and
induces paracrine adrenal signaling that
potentially also influences stress-induced
changes in neural processing involved in the
yawning response. In addition, results from
experiments designed to test the Thompson
Cortisol Hypothesis have consistently showed
that rises in cortisol among human subjects
occur both among those that yawn and do not yawn
under normal laboratory conditions (Thompson and
Bishop 2012; Thompson et al. 2014a, 2014b). This
reproducible effect, i.e., that rises in
cortisol occur independent of yawning, indicates
that cortisol does not serve as a primary
mechanism for triggering this response.
The current findings add to the comparative
literature on stress and yawning, and provide
the first direct experimental demonstration of
stress-induced changes in yawning in humans.
However, there are some limitations to the
current study that should be acknowledged. For
one, the subjective nature of yawning may have
increased the chances of measurement error,
though we have no a priori reason to believe
evaluations of yawning would be biased as a
function of these conditions and previous
research indicates a strong congruence between
self-report and objective measures of yawning
under similar laboratory settings (Gallup and
Church 2015; Massen et al. 2015).
Follow-up experiments could obtain video
recordings of participants during and after
stress to measure both spontaneous and
contagious yawns, which would allow for more
detailed analyses such as how stress modulates
yawn latency or duration (Gallup et al. 2016b).
Another weakness to the current study is that no
physiological or subjective measures of stress
were recorded, which limits our interpretation
of the yawning response. However, our previous
work, which used the same CPT procedures as the
present study, has showed increased cortisol
after CPT induction with maximal concentration
30 min post-CPT (Alomari et al. 2015). We
recognize that the CPT is only one of many
possible stressors that could be applied within
a laboratory setting, and may produce a greater
physical rather than psychological stress
response. Given that yawning has also been
documented to occur prior to some stressful and
anxietyprovoking situations in humans (Provine
2005), future research should examine how the
temporal expression of this response varies as a
function of different forms of psychosocial and
physical stress. Furthermore, follow-up studies
should include selfreport surveys to assess
participants' subjective perception to the
stressor. In combination these measures could be
useful to analyze inter-individual variation in
physiological and behavioral responses to
stressful manipulations.
Understanding the role of yawning in
relation to behavioral and physiological
symptoms of stress may provide further insights
into the latency and degree of eventual recovery
from stressful encounters. The temporal changes
in human yawning following the CPT reported here
(i.e., an initial inhibition followed by an
enhancement) mirror recently documented effects
in budgerigars and Nadza boobies, suggesting a
potential evolutionarily conserved and
homologous behavioral/physiological response to
onetime acute physical stressors across birds
and mammals. In terms of an ultimate
functionality, this observed pattern is
consistent with previous research supporting an
adaptive thermoregulatory and arousal response,
though future experimental research could
explore a potential signaling function to
stress-induced changes in yawning among humans
and other animals.
