Yawning can be observed in most vertebrate
species from foetal stages to old age. In
mammals, it consists of an involuntary sequence
of mouth opening, deep inspiration, brief apnea,
and slow expiration (Walusinski and Deputte,
2004). It can be accompanied by other
facultative motor acts such as stretching
(Provine et al., 1987a). In humans, yawns last
on average about 6 s, and the individual yawn
duration and frequency remains remarkably stable
over weeks (Provine, 1986). In birds and fish
species, a mouth gaping similar to yawning can
be observed, and yawning as opposed to other
forms of mouth openings has been defined as a
slow opening of the mouth, maintenance of the
open position for more than 3 s, followed by a
more rapid closure of the mouth (Baenninger,
1987).
The homology of yawning between different
species is controversial, but at least similar
movement sequences and similar conditions of
occurrence can be observed (Baenninger, 1987;
Deputte, 1994). Since yawning seems to be a
phylogenically old and frequent phenomenon, one
would expect that it provides some evolutionary
advantage, i.e., that is has a certain useful
function. Indeed, numerous hypotheses on the
function of yawing have been posited throughout
the centuries. They were usually derived from
behavioural observations of yawns. In mammals,
it has been observed that more than 90% of yawns
occur at rest whereas the remaining yawns seem
to be triggered by social or emotional stimuli.
These contextual differences have motivated a
classification of yawning into "physiological"
and "social" yawns, although the phenomenology
of yawns does not depend on the context
(Deputte, 1994; Walusinski and Deputte,
2004).
In accordance with the distinction of
physiological and social yawn contexts, the
hypotheses on the function of yawning have
emphasised either a physiological or a social
role of yawning. In contrast to the abundance of
theoretical considerations, experimental data is
relatively scarce. Yet, in the last few decades,
an increasing number of studies have shed some
light on its conditions and effects. Although
the available data is still far from providing a
complete or generally accepted account of the
mechanisms and consequences of yawning, it does
allow confronting some of the theoretical models
with empirical observations. In this
review,wewill try to classify existing
hypotheses according to their current
experimental evidence. All hypotheses
postulating a physiological role of yawning
share the common assumption that yawning
regulates a particular body function, e.g., the
blood oxygen level or the brain arousal level.
Thus, the mechanisms of yawning are
characterised as a homeostatic system with
negative feedback regulation.
Accordingly, physiological models
necessarily make at least two different
predictions that can be empirically tested: (i)
yawning is triggered by up- or downturns of a
given body state and, (ii) yawning acts on the
corresponding body function.Wewill therefore
review the evidence of each physiological
hypothesis based on its predictions with regards
to triggers and effects of yawning. In the case
of social models of yawning, the postulated
regulating function of yawning would not concern
body functions of individuals but rather the
communication within social groups. The
predictions of this model as well as the
corresponding evidence will also be reviewed.
This article will focus on normal yawning; a
recent review on pathological yawns can be found
elsewhere (Walusinski, 2009).
2. Anatomy and
pharmacology
Numerous neurotransmitters, neuropeptides,
and hormones have been found to be implicated in
the control of yawning. Neuroendocrine
substances as diverse as, among others,
dopamine, acetylcholine, glutamate, serotonin,
nitric oxide, adrenocorticotropic hormone (ACTH)
related peptides, oxytocin, and steroid hormones
facilitate yawning whereas opioid peptides have
an inhibitory effect. Some of these mediators
(e.g., dopamine, glutamate, oxytocin) interact
in the paraventricular nucleus of the
hypothalamus (PVN) and induce yawning via
oxytoninergic projections to the hippocampus,
the pons, and the medulla oblongata. Other
pathways seem to be effective for serotonin,
acetylcholine, and ACTH related peptides
(Argiolas et al., 1987; Argiolas and Melis,
1998; Sato-Suzuki et al., 1998). It would be
crucial in our search for a purpose of yawning
to understand the interaction of these
pharmacological pathways with vigilance and
respiration centres or with the mechanisms of
communication and empathy. However, studies
using an interdisciplinary approach of this kind
are currently lacking.
-Sanna
F, Succu S, Melis MR, Argiolas A. Dopamine
agonist-induced penile erection and yawning:
Differential role of D(2)-like receptor subtypes
and correlation with nitric oxide production in
the paraventricular nucleus of the hypothalamus
of male rats. Behav Brain Res. 2012
3. Physiological
hypotheses
3.1. Respiratory and circulatory
hypotheses For several centuries, at least
since Hippocrates in the 4th century BC,
scholars have thought that yawning might remove
"bad air" from the lungs and increase oxygen
circulation in the brain (Trautmann, 1901;
Schiller, 2002; Matikainen and Elo, 2008).
3.1.1. Oxygen need and hypercapnia do not induce
yawning This hypothesis predicts that yawning is
triggered when blood or brain oxygenation is
insufficient, i.e., when oxygen (O2) levels drop
and the CO2 concentration rises. However, from
self-observation most people will confirm that
they do not yawn more frequently when they do
exercise and need more oxygen than when they are
at rest (Provine et al., 1987b). In accordance
with this notion, experiments by Provine et al.
(1987b) demonstrated that healthy subjects who
are exposed to gas mixtures with high levels of
CO2 or physical exercise, do not yawn more
frequently.
Similarly, high levels of O2 had no
influence on the yawning rate. The study has
some limitations, since the subjects had to use
hand-held masks prone to leakage and had to
count their yawns themselves by pressing a
button to activate an event recorder. A
potential effect of blood gas concentration
might therefore have been hidden by confounding
effects. Moreover, the effect of breathing low
oxygen concentrations on the yawning rate has
not been evaluated due to safety concerns.
Nevertheless, the study clearly found
significant effects of blood gases and exercise
on breathing rates, which demonstrates that
breathing and not yawning is the primary
&endash; if not only &endash; physiological
mechanism used for regulation of blood
oxygenation. The breathing rate and yawning rate
were found to vary independently, indicating
that different central mechanisms are effective.
If yawning were critical for brain oxygenation,
one would expect that infrequent yawners have to
perform longer yawns to ensure similar
oxygenation. However, no relationship between
yawn frequency and duration has been observed in
humans (Provine, 1986). Although hypoxia is
frequent in patients with heart or lung disease,
no increased yawning is usually observed in
these patients. Conversely, prolonged
psychogenic hyperventilation with consecutive
hypocapnia has been reported to be associated
with automatic movements including yawns in some
patients (Walusinski, 2009).
In anesthetised rats, local hypoxia in the
paraventricular nucleus of the hypothalamus
(PVN) &endash; induced by injection of a
chemical agent &endash; did indeed produce a
yawning response, which was interpreted as
evidence for the respiratory hypothesis (Kita et
al., 2000). However, thePVNdoes not respond to
local hypoxia only but induces the same
stereotyped yawning response also after
stimulation with several other chemical agents
and even after electrical stimulation
(Sato-Suzuki et al., 1998; Seki et al., 2002).
Thus, the observed yawns during local PVN
hypoxia cannot be interpreted as specific
hypoxia sensitivity of PVN neurons. Rather, they
seem to result from an unspecific irritation of
these cells. The study does therefore not
provide convincing evidence for a causal link
between hypoxia and yawning.
Fish species exposed to low water oxygen
concentrations were found to respond with
opening of the gill operculum (Hasler et al.,
2009). Although this gill flaring response was
named yawning in this study, it is not
homologous to human yawning but rather seems to
be a respiratory act. Taken together, the
occurrence of yawning during periods with too
much blood oxygenation but not during periods
with oxygen need is exactly the opposite of what
would have been predicted by the respiration
hypothesis and thus casts severe doubts on its
correctness.
3.1.2. Yawning does probably not increase
brain oxygenation There are, to our
knowledge, no studies that measured the change
in blood oxygenation induced by yawning.
However, yawning would be amuch less effective
way of increasing oxygen intake than rapid
breathing, especially since the deep inspiration
during yawning is followed by a period of
relative apnoea (Baenninger, 1997). Indeed, the
subjects in the study of Provine et al. (1987b)
used increased breathing rates rather than
increased yawning rates to compensate for high
CO2 concentrations and exercise. Another
mechanism by which yawning could theoretically
increase tissue oxygenation is by increasing
blood circulation. Indeed, yawning has been
found to be associated with an activation of the
autonomic nervous system (Greco and Baenninger,
1991; Askenasy and Askenasy, 1996; Guggisberg et
al., 2007) which, by means of an increased heart
rate and vasodilatation, might result in
increased oxygen circulation. However, autonomic
changes following yawning occur to the same
amount also after simple body movements or after
deep breaths (Greco and Baenninger, 1991;
Guggisberg et al., 2007).
They are thus unspecific and obviously due
to the jaw movement and respiration rather than
the yawning as such. In other words, the act of
yawn does not induce more autonomic changes than
the ones that already occur hundreds of times
throughout the day due to simple breathing or
moving. Hence, from an evolutionary perspective,
yawning does not provide an advantage with
regards to autonomic activity, and it therefore
does not make sense to attribute a circulatory
function to yawning. Provine advanced a further
argument against the respiratory hypothesis
based on his analysis of the routes of
inhalation and exhalation during yawning
(Provine, 1986; Provine et al., 1987a,b). Unlike
normal breathing, yawns cannot be performed
through the nose if subjects have their mouth
taped shut, which indicates that yawning does
not have the degree of behavioural freedom of
normal breathing. Furthermore, oral inhalation
by itself was insufficient for a satisfactory
yawn. The subjects in Provine's study reported a
feeling of satisfaction only if they were
allowed to open their jaw during yawns.Apleasant
yawn depended therefore on the mouth gaping
component but not on the respiratory component
of yawning, which was interpreted as indirect
evidence against a respiratory function of
yawning.
3.1.3. Conclusions The predictions of
the respiratory hypothesis are not supported by
current experimental data. Additional research
is needed to test the effects of hypoxia on the
yawning rate under more controlled conditions.
Studies investigating the effects of yawning on
blood and brain oxygenation are also missing.
Given current evidence, it seems unlikely that
yawning has respiratory or circulatory
functions. 3.2. The arousal hypothesis The idea
that yawning might play an important role in
regulating physiological brain processes has
remained in the literature also after the
appearance of evidence against the respiratory
hypotheses. A widely expressed proposition now
speculated that yawning might be responsible for
the homeostatic regulation of vigilance and
brain arousal level (Baenninger, 1997; Giganti
et al., 2002; Walusinski and Deputte, 2004;
Matikainen and Elo, 2008; Vick and Paukner,
2010).
3.2.1. Drowsiness induces yawning
Yawning occurs preferentially during periods of
drowsiness, as it is predicted by the arousal
hypothesis. Behavioural studies consistently
reported that yawns occur most frequently before
and after sleep, i.e., during periods with lower
levels of alertness (Greco et al., 1993; Provine
et al., 1987a). The circadian distribution of
yawns precisely reflects the individual
sleep-wake rhythm (Giganti et al., 2007; Zilli
et al., 2007, 2008). Furthermore, the individual
subjective feeling of drowsiness correlates with
increased yawning rates (Zilli et al.,
2008).
Weused electroencephalography (EEG) to
objectively assess the vigilance of human
subjects before and after yawns (Guggisberg et
al., 2007). Spontaneous brain activity produces
electromagnetic oscillations in a variety of
frequencies which can be recorded byEEG and
which in turn correlate with specific aspects
ofhumanvigilance and arousal. EEG recordings
were obtained during Maintenance of Wakefulness
Tests (MWT). The MWT is a standardized
diagnostic tool that is widely used to assess
the ability to stay awake in patients with
excessive daytime sleepiness (Doghramji et al.,
1997; Littner et al., 2005). During this test,
the subjects must try to stay awake while
sitting alone in a quiet and darkened room, a
situation which frequently leads to spontaneous
yawning. EEG segments of 16 subjects who had
yawned at least 4 times during the test were
analyzed. Fig. 1 (left panel) shows that delta
(<3 Hz) power density over central midline
brain areas was significantly greater before
yawns than before control movements produced by
the same subjects during the same test (t(15) =
3.1, p = 0.008) (Guggisberg et al., 2007). These
control movements consisted in postural
adjustments without yawning. What does this
mean? Delta frequencies are known to increase
with the duration of wakefulness and to decrease
during sleep, and are therefore interpreted as
an indicator of an individual's sleep pressure
(Borbely et al., 1981). Thus, sleep pressure and
drowsiness proved significantly greater when
subjects yawned than when they moved only.
3.2.2. Yawning does not produce an
arousal Arousals are defined as a global
activation of brain activity that progresses
from brain stem structures to centres of the
autonomic nervous system and to distributed
cortical areas (Moruzzi and Magoun, 1949; Sforza
et al., 2000). They are accompanied by a typical
acceleration of EEG activity. Several studies
have therefore analyzed spectral EEG changes
after yawns in humans to test the hypothesis
that yawning has an arousing effect. However,
the results were negative.
Two studies looking at 30 s samples of EEG
before and after yawns were unable to find
significant and lasting changes in EEG activity
related to yawns (Laing and Ogilvie, 1988;
Regehr et al., 1992). One of these studies
reported transient increases in theta, spindle,
and beta activity, but they only reached
significance when the analysis was a priori
limited to data segments between 10 and 20 s
before and after yawning (Regehr et al., 1992).
Furthermore, EEG power after yawning was not
significantly different from EEG power after
postural adjustments without yawning (Laing and
Ogilvie, 1988).
In our analyses of EEG power spectra from
patients undergoing MWTs,weobserved that the
increase in delta power over the vertex
thatwasfound before yawning (as compared to
delta activity before postural adjustments
without yawning, see Section 3.2.1) persisted to
the same amount also after yawning (Fig. 1,
right panel). Thus, yawning did not reverse the
increased sleep pressure and drowsiness that
seemed to have triggered it.
Besides delta power, alpha oscillations
(~7.5&endash;12.5 Hz) also reflect the
individual vigilance level. They become faster
and smaller in amplitude when the arousal level
increases. Fig. 2A gives an example of the EEG
power spectrum 30 min after oral ingestion of
250mg caffeine (Barry et al., 2005). In
contrast, drowsiness is associated with a
slowing of alpha oscillations, and with a shift
of alpha oscillations from mainly occipital
towards central brain regions (Tanaka et al.,
1997; De Gennaro et al., 2001a,b). Fig. 2B shows
that alpha power after yawning showed a pattern
that is typical for sleepiness: alpha rhythms
decelerated, and shifted towards central brain
regions after yawning, as compared to the data
segments before yawning. Conversely, we did
observe EEG markers of increased arousal levels
after simple postural adjustments, as shown in
Fig. 2C: alpha rhythms became faster and smaller
after body movements. Hence, if yawning had an
arousing effect &endash; even if it were as
small as the effect of simple postural
adjustments &endash;we would have detected it
with our EEG analyses. Instead,weobserved signs
of progressive drowsiness after yawning.
Arousals are also accompanied by activations
of the autonomic system. As already discussed in
Section 3.1.2, yawning is indeed followed by
activations of the autonomic system, which might
indicate some elementary form of arousal.
However, this automatic activation is entirely
unspecific and related to the associated
movement and respiration rather than yawning as
such (Greco and Baenninger, 1991; Guggisberg et
al., 2007). Other studies have assessed the
arousal level after yawning by measuring the
skin conductance, which was shown to reflect
both autonomic and cortical activities (Barry et
al., 2005; Lawrence et al., 2005). Again, no
specific increase in skin conductance could be
observed after yawning (Greco and Baenninger,
1991). One study 3.2.2. Yawning does not produce
an arousal Arousals are defined as a global
activation of brain activity that progresses
from brain stem structures to centres of the
autonomic nervous system and to distributed
cortical areas (Moruzzi and Magoun, 1949; Sforza
et al., 2000). They are accompanied by a typical
acceleration of EEG activity. Several studies
have therefore analyzed spectral EEG changes
after yawns in humans to test the hypothesis
that yawning has an arousing effect. However,
the results were negative.
Two studies looking at 30 s samples of EEG
before and after yawns were unable to find
significant and lasting changes in EEG activity
related to yawns (Laing and Ogilvie, 1988;
Regehr et al., 1992). One of these studies
reported transient increases in theta, spindle,
and beta activity, but they only reached
significance when the analysis was a priori
limited to data segments between 10 and 20 s
before and after yawning (Regehr et al., 1992).
Furthermore, EEG power after yawning was not
significantly different from EEG power after
postural adjustments without yawning (Laing and
Ogilvie, 1988).
In our analyses of EEG power spectra from
patients undergoing MWTs,weobserved that the
increase in delta power over the vertex
thatwasfound before yawning (as compared to
delta activity before postural adjustments
without yawning, see Section 3.2.1) persisted to
the same amount also after yawning (Fig. 1,
right panel). Thus, yawning did not reverse the
increased sleep pressure and drowsiness that
seemed to have triggered it.
Besides delta power, alpha oscillations
(~7.5&endash;12.5 Hz) also reflect the
individual vigilance level. They become faster
and smaller in amplitude when the arousal level
increases. Fig. 2A gives an example of the EEG
power spectrum 30 min after oral ingestion of
250mg caffeine (Barry et al., 2005). In
contrast, drowsiness is associated with a
slowing of alpha oscillations, and with a shift
of alpha oscillations from mainly occipital
towards central brain regions (Tanaka et al.,
1997; De Gennaro et al., 2001a,b). Fig. 2B shows
that alpha power after yawning showed a pattern
that is typical for sleepiness: alpha rhythms
decelerated, and shifted towards central brain
regions after yawning, as compared to the data
segments before yawning. Conversely, we did
observe EEG markers of increased arousal levels
after simple postural adjustments, as shown in
Fig. 2C: alpha rhythms became faster and smaller
after body movements. Hence, if yawning had an
arousing effect &endash; even if it were as
small as the effect of simple postural
adjustments &endash;we would have detected it
with our EEG analyses. Instead,weobserved signs
of progressive drowsiness after yawning.
Arousals are also accompanied by activations
of the autonomic system. As already discussed in
Section 3.1.2, yawning is indeed followed by
activations of the autonomic system, which might
indicate some elementary form of arousal.
However, this automatic activation is entirely
unspecific and related to the associated
movement and respiration rather than yawning as
such (Greco and Baenninger, 1991; Guggisberg et
al., 2007). Other studies have assessed the
arousal level after yawning by measuring the
skin conductance, which was shown to reflect
both autonomic and cortical activities (Barry et
al., 2005; Lawrence et al., 2005). Again, no
specific increase in skin conductance could be
observed after yawning (Greco and Baenninger,
1991). One study even observed a yawning induced
decrease in skin conductance, which would
suggest a yawn-related decrease in arousal level
(Baenninger and Greco, 1991). One of the main
arguments for an arousing effect of yawning has
been derived from the observation that yawns are
followed by a significant increase in motor
activity (Baenninger, 1997; Giganti et al.,
2002; Vick and Paukner, 2010).
However, motor activity depends on numerous
factors and more motor activity does not
necessary point to an increased cerebral arousal
level. Sleepy individuals trying to stay awake
(e.g., during a boring meeting or during
vigilance tests) typically present amotor
restlessness with frequent changes of the body
position (Guggisberg et al., 2007). Fig. 2C
shows that these movements indeed do have an
arousing effect measurable by EEG. Hence, the
increased motor activity observed after yawns is
probably not an indicator of an arousing effect
of yawning, but an effective countermeasure
against the underlying drowsiness. An
association of yawns and arousals is regularly
observed in animal models of yawning. In
anesthetized rats, electrical stimulation of the
PVN or the application certain drugs leads to a
stereotyped sequence of arousal reaction
followed by yawning (Sato-Suzuki et al., 1998;
Kita et al., 2000; Seki et al., 2002).
However, since this arousal occurs before,
not after, the actual yawning, it cannot be
interpreted as a consequence of yawning, but
rather corresponds to a requirement for yawning
to occur during anaesthesia. Indeed, yawning
almost never spontaneously occurs during sleep
(Provine et al., 1987a; Greco et al., 1993;
Giganti et al., 2007). Yawning during induction
of anesthesia has also been observed in humans,
and recordings of the bispectral index have also
suggested an accompanying arousal reaction
(Kasuya et al., 2005). However, the bispectral
index is sensitive to artifacts from cranial
muscle activity which is abundant during
yawning. Even if the observations of the study
did not result from muscle artifacts, the same
interpretation holds as for the data obtained in
rats.
3.2.3. Conclusions The experimental
data suggests that yawning indeed occurs during
progressive drowsiness, which is compatible with
the notion that it is triggered by states of low
vigilance. However, no specific arousing effect
of yawning on the brain or the autonomic nervous
system could be observed. Experimental evidence
therefore suggests a rejection of the arousal
hypothesis. The absence of an arousing effect of
yawning does obviously not exclude that it might
have some other form of activating function on
brain metabolism or neuropharmacology, but these
effects should not be named arousal.
3.3. The sleepiness hypothesis Rather
than attributing an arousing effect to yawning,
some authors have suggested that it might lower
the arousal level (Deputte, 1994). Studies
assessing the arousal level after yawing have
indeed found signs of decreasing wakefulness
(see Section 3.2.2), which would be compatible
with this notion. However, the observations
could simply represent the drowsiness underlying
yawning that continues to progress also after
yawning. Thus, there is no established causal
link between yawning and subsequent drowsiness.
Moreover, if yawning had a soporific effect
apart from being induced by drowsiness, it would
be a self-reinforcing mechanism and would need
to be controlled by other processes in order to
ensure stability of the sleep-wake balance.
3.4. The thermoregulation hypothesis
Recently, another physiological function of
yawning has been proposed: the regulation of
brain temperature. It is postulated that yawning
might cool down the brain when its temperature
increases. The advocates of this model give a
detailed description of their arguments in
(Gallup and Gallup, 2008). Here, we provide a
brief critique of the corresponding experimental
evidence.
3.4.1. Does brain hyperthermia trigger
yawning? Yawning has a well-known contagious
effect. In a recent experiment, the frequency of
these contagious yawns (which were induced by
having the subjects watch videos of yawning
people) could be manipulated when the subjects
held temperature packs on their forehead or when
they breathed rapidly (Gallup and Gallup, 2007).
For example, a cold pack on the forehead was
associated with decreased contagious yawning
whereas a warm pack increased the occurrence of
contagious yawns. This was interpreted as
evidence for a role of brain temperature in the
generation of yawning. However, the experiment
did not control for potential confounding
factors. For instance, having an ice pack on
one's forehead likely has a profound arousing
effect whereas a nice and warm pack will promote
sleepiness. It is therefore impossible to
differentiate between effects of temperature and
sleepiness in this experiment. The authors of
the study acknowledge a correlation between the
circadian rhythms of temperature and vigilance,
but maintain that temperature is the decisive
parameter in yawning generation. However, there
is no evidence for the latter claim. On the
contrary, there is evidence from behavioural and
EEG studies that vigilance is one of the primary
yawn triggering factors. The same concern also
applies to a second study of the same group
performed in birds which were exposed to
different ambient temperature
conditions.Arapidly increasingroomtemperature
was associated with more frequent yawns than
relatively stable cold or warm temperatures
(Gallup et al., 2009), which may again be due to
uncontrolled factors such as differences in
drowsiness or related to rapidly changing vs.
stable temperatures. The proponents of the
thermoregulation hypothesis also advance
anecdotal data of yawning frequency in patients
with different brain diseases, but in the
absence of direct comparisons and controls, the
evidence remains inconclusive.
3.4.2. Yawning does probably not cool
down the brain The greatest challenge for
the proponents of the thermoregulation
hypothesis lies in demonstrating how yawning
would be able to cool down the brain. It is
suggested that the inflow of cool air during
yawning ventilates heat off the brain. However,
the proposition faces similar problems as the
respiratory hypotheses discussed above. Yawning
actually interrupts normal nasal breathing which
seems to be a more efficient way of
ventilation.
3.4.3. Conclusions There is currently
insufficient evidence for a thermoregulatory
effect of yawning. The thermoregulation
hypothesis seems to be counterintuitive and has
important explanatory gaps which seem to be
difficult to close.
3.5. The ear pressure hypothesis
Yawning has the much appreciated capacity to
equalize air pressure in the middle ear with
outside air pressure. It can thus relieve
discomfort in the ear and hearing problems due
to rapid altitude changes in air planes or
elevators. This is achieved through contraction
and relaxation of tensor tympani and stapedius
muscles during yawning, which results in an
opening of the Eustachian tubes and the aeration
of the tympanal cavities (Laskiewicz, 1953;
Winther et al., 2005). This observation has led
to the postulation that yawning might be a
"defence reflex" of the ear, which is triggered
by rapid altitude changes or other conditions
leading to air trapping in the middle ear
(Laskiewicz, 1953).
However, there is to our knowledge no
systematic investigation that would confirm
increased yawning rates under rapidly changing
ear pressure conditions. Also, yawning is not
the only mechanism to open the Eustachian tube;
swallowing, chewing, and the Valsalva manoeuvre
have the same effect (Laskiewicz, 1953; Winther
et al., 2005). The middle ear pressure release
of yawning does therefore not represent by
itself an indispensable evolutionary advantage.
Equalization of ear pressure seems to be a
useful effect that yawns have in common with
other contractions of oropharyngeal muscles
rather than the primary purpose of yawning.
3.6. The state change hypothesis
Rather than suggesting a single
physiological function of yawning, Provine
attempted to combine the multiple behavioural
state changes associated with yawning
(wakefulness to sleep, sleep to wakefulness,
alertness to boredom, etc.) within a single
framework. He proposed that "yawning is a
vigorous, widespread act that may stir up our
physiology and facilitate these transitions"
(Provine, 1986, 2005).
This approach has the advantage that it
might integrate findings from different research
fields. However, the proposition does not go
beyond amere description of the behavioural
changes associated with yawning and does not
give insights into how or why the proposed state
changes might be achieved. Given the current
scarcity of experimental evidence for any
physiological function of yawning, the
combination of several physiological states
within a single concept also lacks empirical
support.
3.7. Other physiological hypotheses
Several other variants of a regulatory function
of yawning on body physiology have been proposed
(Smith, 1999). To name only a few: yawning
prevents lung atelectasis (Cahill, 1978);
yawning renews surfactant films in lungs
(Forrester, 1988); yawning ensures intermittent
evacuation of the palatine tonsillar fossae
(McKenzie, 1994). None of these propositions has
been experimentally tested.
4. The social/communication
hypothesis In many cultures, yawning is
interpreted as a sign of boredom and sleepiness
and is therefore considered to be rude
(Schiller, 2002). Thus, yawning seems to
communicate a message that is almost universally
understood. Moreover, yawning frequently occurs
in social contexts. A communicative function of
yawning has therefore long been suspected. The
hypothesis states that yawning is a non-verbal
form of communication that synchronizes the
behaviour of a group (Barbizet, 1958; Provine,
1986; Weller, 1988; Deputte, 1994).
4.1. Yawning has physiological and social
triggers Yawning can be triggered by several
different physiological body states as well as
social contexts. Drowsiness (see above) and
boredom (Provine and Hamernik, 1986) are well
documented precursors of yawning. Observations
in animals further suggest that yawns may be
facilitated by hunger or mild psychological
stress (Deputte, 1994). The communication
hypothesis accounts for all these inductors by
stating that they generate yawning to transmit
the corresponding information to other members
of a social group. The number of possible
yawning triggers must of course not be
unlimited; otherwise the transmitted message
would be too ambiguous. Indeed, all triggers of
yawning mentioned above have in common that they
are mildly to moderately unpleasant while not
presenting an immediate threat.
4.2. Social effects of yawning The social
hypothesis predicts that yawning has some impact
on the behavioural organization of a social
group. Communication should result in better
synchronization of group behaviour. Such effects
have indeed been observed in Ostriches (Sauer
and Sauer, 1967), but studies that test the
prediction in a controlled fashion are
lacking.
4.3. Contagious yawning Yawning has a
well-known contagious effect in humans
(Baenninger, 1987; Provine et al., 1987b;
Provine, 1989a,b; Platek et al., 2003) and this
effect is now frequently used to induce yawning
for research purposes. Recent studies have
accumulated evidence that this contagiousness
depends on an intact social competence of the
yawning individual. The susceptibility to
contagious yawning correlates with empathic
skills in healthy humans (Platek et al., 2003)
and is reduced in patients with disorders
affecting the ability of social interaction,
such as autism (Senju et al., 2007) and
schizophrenia (Lehmann, 1979; Haker and Rossler,
2009). In patients with schizophrenia, the
occurrence of yawns has been interpreted as a
positive sign indicating that the patient is in
an accessible mood (Lehmann, 1979). Watching or
hearing other persons yawn activates a complex
network of brain regions related to motor
imitation, empathy, and social behaviour. Fig. 3
illustrates the brain regions that have been
reported to activate in different functional
magnetic resonance imaging (fMRI) studies when
human subjects observe yawns of others.
The so-called mirror neuron system is
important for action understanding and imitation
(Rizzolatti and Craighero, 2004) and mirror
neurons in the right posterior inferior frontal
gyrus also seem to be recruited for contagious
yawning (Arnott et al., 2009). The mirror neuron
activity is however not specific to yawning but
occurs to the same amount also during
observation of other movements (Nahab et al.,
2009; Arnott et al., 2009). Activations that are
more specific to contagious yawns have been
observed in the posterior cingulate (Platek et
al., 2005), the bilateral superior temporal
sulcus (Schurmann et al., 2005), or the
ventromedial prefrontal cortex (Nahab et al.,
2009). The fMRI activations in these areas were
significantly greater when the study subjects
watched other persons yawn than when they
watched control face movements of others.
Although different studies have reported
divergent areas to be implicated in contagious
yawning, all of them seem to be part of a
distributed neural network related to empathy
and social behaviour (Saxe et al., 2004;
Carrington and Bailey, 2009). In children, no
contagious yawning can be induced before the age
of five (Anderson and Meno, 2003), suggesting
that the contagiousness of yawning depends on
mechanisms that have to develop during childhood
in parallel with the empathic capacity to
understand mental states of others (Saxe et al.,
2004). In animals, contagious yawning has been
consistently observed in chimpanzees (Anderson
et al., 2004; Campbell et al., 2009; Vick and
Paukner, 2010), whereas it seems to be absent in
lions (Baenninger, 1987). In old-world monkeys
(Baenninger, 1987; Paukner and Anderson, 2006;
Palagi et al., 2009) and dogs (Joly- Mascheroni
et al., 2008; Harr et al., 2009), different
studies showed divergent results, but contagious
yawning occurs at least in some
individuals.
The findings from animal studies therefore
also support the notion that contagious yawning
mostly occurs in individuals and species with
advanced empathic and social skills. In monkeys,
the contagiousness of yawning correlates with
the level of grooming contact between
individuals (Palagi et al., 2009), i.e., it is
higher in animals that are socially and
emotionally close to each other.
In summary, research on contagious yawning
has revealed that yawns are part of the action
repertoire of empathic and communicative
processes in adult humans and some other
mammals, which provides strong evidence for a
social role of yawns in these species.
4.4. Other social modulators of yawning
Social contexts were found to have an important
impact on the yawning rate. In animals, the
hierarchical position within a social group
influences the frequency of yawning: group
leaders initiate more yawns than subordinates
(Hadidian, 1980). This difference in the yawning
rate may correspond to the greater importance of
communications from leaders than from other
individuals for the synchronized behaviour of
the group (Sauer and Sauer, 1967), and may thus
also be explained within the framework of the
communication hypothesis.
There are however also yawns that are
independent of social modulation. Yawning also
occurs when individuals are alone and in
non-social animals. This might be used as an
argument against the communication hypothesis
and for the need to postulate an additional
physiological effect of yawning. However, the
existence of yawns during aloneness does not
contradict the communication hypothesis in
general; it merely shows that the generators of
yawning lack a negative feedback mechanism
checking for the presence of other individuals.
Hence, the message of yawning seems to be
triggered by certain body states and "sent out",
no matter whether there are other individuals
that might actually receive it. In humans, the
presence of other humans may even have a
suppressive effect on the yawning rate. If human
subjects feel socially observed, they completely
stop yawning even if the usual conditions of
yawning are met (Baenninger and Greco, 1991;
Provine, 2005). This suppression may result from
arousing effects inherent to social observation.
Alternatively, the negative connotation of
yawning in human society may push the
individuals to hide or inhibit yawns when they
are felt to be inappropriate.
4.5. Conclusions The communication
hypothesis has the best experimental evidence
among all propositions on the function of
yawning. It is the only model that can account
for social effects of yawning such as
contagiousness and for the different
physiological states and social contexts that
can trigger it. Missing elements of this model
include controlled studies observing a
regulating effect of yawning on synchronized
group behaviour and data on the
neuropharmacological mechanisms underlying the
social inductors and effects of yawning. It is
also far from clear whether the findings of
contagious yawns derived mostly from studies in
humans and primates can be generalized to other
forms of yawns and to yawns in other species.
The social aspects of spontaneous
(non-contagious) yawns, particularly in species
and individuals who are not susceptible to
contagious yawning, have received little
research interest so far.
5. Discussion
In 1986, Robert R. Provine, the pioneer in
yawning research, wrote that "yawning may have
the dubious distinction of being Today, more
than two decades later, this may well still be
the case. In particular, the centuries-old
question ofwhyweyawnstill awaits a corroborated
answer. None of the numerous propositions on the
function of yawning has currently sufficient
experimental support or links to
neuropharmacological mechanisms. Nevertheless,
the preceding sections (which are summarized in
Table 1) may have demonstrated that the emphasis
of models on yawning has changed. Whereas
traditional hypotheses were mostly characterized
by the quest for a physiological function of
yawning in individuals, these propositions now
face severe explanatory problems or lack
empirical evidence. In contrast, the idea that
yawning might rather serve a social function in
groups of individuals receives increasing
support from studies in different fields. It
emerges that yawning might communicate
unpleasant but not immediately threatening
states to other members of a group in order to
enhance behavioural synchronization. This social
hypothesis of yawning is also the only model
that can account by itself for all elements
associated with yawns. For instance, contagious
effects or social contexts of yawning cannot be
explained when assuming a purely physiological
function. Physiological hypotheses therefore
have to postulate social effects in addition to
a physiological effect of yawning, whereas the
physiological triggers of yawning form an
integral part of social models. Hence, the
social hypothesis has not only the best
experimental support but is also the most
parsimonious model. From an evolutionary
perspective, the communicative value of yawning
may yield sufficient advantage to explain its
persistence and frequent usage in many
vertebrate species. The capacity to exchange
information about the physical and mental state
of each individual seems indeed to be crucial
for the survival of a group. There is therefore
no need to postulate additional physiological
functions of yawning to explain its selection
during evolution. One may argue that the
difficulties with physiological models results
from an oversimplification of a complex
phenomenon.
There might be different types of yawning
that assume different functions which are
unrecognized if all yawns are inappropriately
pooled. However, the data from observational
studies does not support this notion. Although
numerous yawn morphologies and contexts have
been described (Provine, 1986; Deputte, 1994;
Baenninger, 1987; Palagi et al., 2009; Vick and
Paukner, 2010), the different studies did not
converge on a consistent classification into
well-delimited types. Furthermore, most studies
found no functional or contextual differences
among the different yawning morphologies
(Provine, 1986; Deputte, 1994; Baenninger, 1987;
Palagi et al., 2009). Vick and Paukner (2010)
interpreted differences in the scratching rate
after "full yawns" vs. "modified yawns" with
additional voluntary face movements of
chimpanzees as evidence for a selective arousal
effect of modified yawns only, but we have seen
above that indirect behavioural markers of
arousals are problematic. The current limited
data therefore seems to suggest that yawning is
a single mechanism associated with a continuum
of behavioural manifestations rather than a
discrete set of functional entities.
On the species level, the generators and
functions of yawning may have evolved
differently in different species, and yawns may
even be a residual of earlier life forms with no
remaining function at all in some species.
However, in the absence of evidence for
systematic differences in the mechanisms and
functions of yawning between species or yawn
morphologies, this call for more complexity does
not withstand the simplicity and elegance of the
social model of yawning.
In conclusion, current data suggests that we
might have to get used to the idea that yawns
have a primarily social rather than
physiological function.
6. Future research
directions Several lessons can be
learned from research of the last three decades.
Experience with the respiratory and arousal
hypotheses demonstrates that one must be careful
when interpreting indirect or anecdotal
evidence. Although both hypotheses had some
arguments and indirect evidence on their side,
direct measurements showed negative results. In
order to differentiate between specific features
of yawning and nonspecific coexisting elements,
it is important to include control groups or
conditions during experiments. The lack of
controlled experimental studies on yawning
illustrates the need for research programs in
all related fields. Some of the specific
questions that could be addressed are listed in
Table 2.
All current models on the function of
yawning are derived from observations of the
phenomenology and contexts of yawning, which may
result in a negligence of aspects that are not
behaviourally evident. An exploration of the
neural and metabolic mechanisms may give new
hints on the functions of yawning that were
hitherto unsuspected or on the mechanisms of
existing concepts. Future research should
therefore systematically assess behavioural,
physiological, and social features of yawning
and combine observational with interventional
techniques. This requires interdisciplinary
strategies that would overcome limitations of
traditional techniques. For example, a
combination of interventional approaches
[e.g., administration of yawn-inducing or
-inhibiting drugs (Argiolas and Melis, 1998),
experimental lesions of brain structures
involved in yawn-generation such as the PVN
(Argiolas et al., 1987), manipulation of
environmental conditions] with systematic
behavioural observations during wakefulness may
increase the value of both animal models and
observational approaches.
A multimodal approach of this kind also
seems to be necessary to resolve long-standing
controversies on whether different types of
yawning exist and on whether yawns in different
species are homologous. Future studies
addressing these issues should systematically
compare not only behavioural but also social,
functional, and physiological parameters when
trying to classify yawns within and across
species. Besides this explorative approach,
there is also a need for hypothesis-driven
research based on the current models of yawning.
Numerous open questions related to the
hypotheses discussed above remain unanswered;
Table 2 lists only a few.
