mise à jour du
22 août 2009
Cognitive, affective, behavioral neurosci
An investigation of auditory contagious yawning
Stephen R. Arnott , Anthony Singhal, Melvyn A. Goodale
Rotman Research Institute, Toronto, Ontario, Canada
and University of Western Ontario, London, Ontario, Canada


Tous les articles sur la contagion du bâillement
All articles about contagious yawning
One of the more curious aspects observed in human behavior is the tendency to adopt the gestures and pos- tures of others (e.g., the chameleon effect; Chartrand & Bargh, 1999). Perhaps the most well-known example of this is contagious yawning (i.e., the act of yawning or the urge to yawn in the presence of another person yawning; Provine, 1986). The widespread familiarity of contagious yawning is due in part to the fact that almost everyone has experienced the phenomenon and that it occurs seem- ingly without volition. Indeed, even the mere thought of yawning (e.g., reading about it) can be enough to incite contagious yawning (Provine, 1986). Despite the near uni- versality of this phenomenon, however, research inves- tigating the cortical mechanisms underlying contagious yawning, although providing many advances, has yet to fully explain its neural basis (Provine, 2005).
A thorough understanding of the neural underpinnings of contagious yawning is important for an understanding of the behavioral and perhaps the communication systems of human and non human species, and it may be particularly relevant to revealing why certain human populations demonstrate absent or aberrant forms of mimicry (e.g., autism; see Rogers, Hepburn, Stackhouse, & Wehner, 2003). Recently, a potential advance in the understanding of contagious yawning has arisen with the discovery of mir- ror neurons (MNs). These cells, located in the posterior inferior frontal gyrus (pIFG) of monkeys and assumedly humans (Fadiga, Fogassi, Pavesi, & Rizzolatti, 1995; Riz- zolatti & Craighero, 2004), become active when an animal performs a particular action as well as when it perceives another perform that same action (di Pellegrino, Fadiga, Fogassi, Gallese, & Rizzolatti, 1992; Gallese, Fadiga, Fo- gassi, & Rizzolatti, 1996; Rizzolatti & Craighero, 2004; Rizzolatti, Fadiga, Gallese, & Fogassi, 1996).
Accordingly, MNs provide a basis for action understanding (Rizzolatti & Craighero, 2004) and may form part of the neural sub- strate underlying imitative actions (Jeannerod, 1994; but see Makuuchi, 2005)—capacities both relevant to the phenomenon of contagious yawning. There is also speculation that the MN system extends into visceromotor brain cen- ters, thereby enabling the experiential understanding of others, emotional states (Gallese, Keysers, & Rizzolatti, 2004; Keysers & Gazzola, 2006; Wicker et al., 2003). Here too, a putative link with contagious yawning exists, because con- tagious yawning appears to be part of the neural network involved in empathy (Platek, Critton, Myers, & Gallup, 2003; Platek, Mohamed, & Gallup, 2005). Accordingly, there is reason to speculate that IFG cells may contribute to the phenomenon of contagious yawning (Provine, 2005). That said, null findings from two fMRI investigations of contagious yawning (Platek et al., 2005; Schiirmann et al., 2005) have called this hypothesis into question, leading to speculation that cells in the pIFG seem to be circumvented when a yawn is observed. It is possible, however, that contagious yawn-related hemodynamic activity in those studies was obscured by issues of experimental design (e.g., block designs and task requirements) that rendered it impossible to selectively average only those yawn stimuli that actually elicited an urge to yawn in each participant. In addition, it is arguable that some of the control stimuli used in those experiments were also MN stimulating, further reducing the likelihood of observing pIFG activity in response to observing yawns.
In the present auditory fMRI study, we reexamined the neural underpinnings of contagious yawning, this time using an event-related experimental design in which participants rated how much each stimulus made them feel like yawning. In this way, it was possible to isolate hemodynamic activity associated with those yawn stimuli that elicited strong urges to yawn (i.e., contagious yawning) within each participant. An auditory approach was adopted primarily because a formal investigation was needed to support the widely held, although largely anecdotal (Moore, 1942), notion that the sound of a yawn elicits contagious yawning. Like visual images, sounds are also effective MN stimulators (Gazzola, Aziz-Zadeh, & Keysers, 2006; Keysers et al., 2003; Kohier et al., 2002; Pazzaglia, Pizzamiglio, Pes, & Aglioti, 2008; Rizzolatti & Craighero, 2004), and so an auditory paradigm was considered to be a valid test of the MN hypothesis.
 auditory contagion
Our results clearly show that hearing someone yawn not only increases a person's urge to yawn, but that it activates brain areas that (1) have been shown to be involved in hearing and executing mouth actions (Ga.zzola et al., 2006), and (2) are necessary for recognizing the actions of others (Pazzaglia et al., 2008). Accordingly, our findings link contagious yawning to the human MN network (Rizzolatti & Craighero, 2004). In fact, the more contagious the yawn stimulus was, the more these regions (i.e., pIFG and pSTG) were activated. Such activity cannot easily be explained by appealing to notions of yawn suppression (e.g., the activation was the result of participants inhibiting the urge to yawn while undergoing a movement-sensitive fMRI scanning procedure), because the yawn stimuli associated with high urge-to-yawn ratings still generated more BOLD activity in these regions than did the control stimuli associated with equivalently high urge-to-yawn ratings.
These findings are all the more significant given that prior fMRI studies argued against pIFG involvement in contagious yawning (Platek et al., 2005; Schtirrnann et al., 2005). As in our study, yawh-specific activation had been observed in the right posterior superior temporal lobe, perhaps related to the processing of socially relevant biological motion and its imitation (lacoboni, 2005; Schiirmann et al., 2005) or to the fact that the stimulus was a mouth action sound (Gazzola et al., 2006). But neither of the previous contagious yawning studies found that watching people yawn resulted in differential pIFG activation when compared with watching people laugh (Platek et al., 2005) or move their tongue in their cheek (Schürmann et al., 2005). This has prompted speculation that contagious yawning is an automatically released behavior that circumvents the essential parts of the MN system (Schürmann et al., 2005).
On reflection, however, it seems possible that yawnrelated pIFG activity in those studies could have been masked by activity elicited by the control conditions. Laughing, for instance, is also a contagious behavior (Provine, 2005) that can invoke right pIFG activity, during both its perception and expression (Meyer, Baumann, Wildgruber, & Alter, 2007). Similarly, the tongue-in-cheek control condition used by Schurmann et al. (2005) may also have elicited an imitation neural response, given that human MNs are believed to respond even to movements that are not goal directed (Rizzolatti & Craighero, 2004). Indeed, when the results from that study are reexamined, pIFG activation is noticeably present for both the yawn and tongue stimuli (Schürmann et al., 2005, Figure 2, p. 1262). Moreover, the activation appears to be stronger in the right IFG and also appears greatest for the yawn stimuli. In the present study, the control stimuli (i.e., the scrambled yawn sounds) may be considered more "action benign" in the sense that they were artificially created, physically impossible stimuli.
Second, it is very likely that the block design nature of the previous studies reduced the ability to isolate contagious yawn hemodynamic activity. Clearly, contagious yawning does not always occur every time a yawn stimulus has been perceived. Although some yawn stimuli are more effective than others, a stimulus that is effective for one person may not be so for another. Furthermore, there is most likely variation even within a given individual from one occasion to the next. Unlike the previous designs, the online rating technique of the present eventrelated experiment enabled us to identify and isolate brain responses to yawn sounds that were experienced as contagious compared with those that were not. That said, it is unclear to what extent the yawning judgments themselves (i.e., the urge-to-yawn ratings) affected the activity observed in response to contagious yawns.
Although we do not deny that being cognizant of yawning may prime a person to be more susceptible to contagious yawning, this should not invalidate the present results. Indeed, people are known to be more susceptible to yawning under different circumstances (Provine & Hamernik, 1986), and many would agree that there are occasions in which one is more susceptible to contagious yawning than others (e.g., in situations were attention is not strongly focused on a task at hand). This makes the point that, in some respects, contagious yawning may rely on some form of embodiment or mental reflection of the yawn that has been perceived, be it at its most basic, perhaps subconscious, level (see also Walusinski, 2006).
Contagious Yawning and MNs
Our finding of contagious yawn-related activity in what is thought to be the hub of the MN system (i.e., the pIFG; Rizzolatti & Craighero, 2004) is not the only datum consistent with an MN hypothesis of contagious yawning. First, it has been argued that the human MN system is not functional at birth but develops some time after 6 months of age (Falck-Ytter, Gredebäck, & von Hofsten, 2006). It is noteworthy that, as an antecedent to any MN-mediated behavior, human contagious yawning is reportedly not exhibited until some time after the second year of life (Anderson & Meno, 2003; Piaget, 1951), despite the fact that spontaneous yawning is readily exhibited by infants, newborns, and even fetuses (Egerman & Emerson, 1996; Sepulveda & Mangiamarchi, 1995). This incidentally implicates a perceptual and/or cognitive component as being central to contagious yawning (Lejimann, 1979; Platek et al., 2005).
Also, children with autism, a disorder that may be related to MN-system abnormalities (Oberman & Ra machandran, 2007), show a marked reduction in contagious yawning despite exhibiting "normal" spontaneous yawning behavior (Senju et al., 2007). Our findings show a positive correlation between a person's susceptibility to contagious yawning and how empathic they are. Furthermore, individuals who were more empathic showed greater right pIFG activations to contagious yawn sounds, consistent with the idea that the MN system is central to social cognition (Gazzola et al., 2006; Kaplan & lacoboni, 2006; Pfeifer et al., 2008). Lastly, akin to the fact that MNs are suspected only in certain populations (e.g., primates; Rizzolatti & Craighero, 2004), it is interesting that contagious yawning has been documented in only humans,
chimpanzees (Anderson, Myowa-Yamakoshi, & Matsuzawa, 2004), and (questionably) macaques (Paukner & Anderson, 2006),' despite the fact that a vast array of vertebrate species are known to exhibit spontaneous yawning (Baenninger, 1987).
Given that classically defined MN-mediated behaviors are considered to be those that invoke mirror cell activity during the observation as well as the execution of the relevant behavior (Rizzolatti & Craighero, 2004), one might expect the pIFG also to be active when a person yawns. However, because of the stereotypical head movements associated with the act of yawning (Deputte, 1994; Provine, 1986; Redican, 1975), this question is difficult to directly assess with functional neuroimaging. Nevertheless, there is good reason to believe that the pIFG may be associated with yawning. First, it has been reported that a patient with Foix-Chavany-Marie syndrome, a form of pseudobulbar palsy resulting from bilateral opercular lesions, was unable to yawn voluntarily (whether spared involuntary yawning was spontaneous and/or contagious is unclear; Laurent-Vannier, Fadda, Laigle, Dusser, & Leroy-Malherbe, 1999). Second, and more convincingly, recent fMRI investigations of mouth movements have shown that very proximal, if not the same, regions of the pIFG in BA44 are activated when participants are asked to smile (Warren et al., 2006), imitate facial expressions (Lee, Josephs, Dolan, & Critchley, 2006; van der Gang, Minderaa, & Keysers, 2007), manipulate small objects with their lips (Gazzola et al., 2006), listen to pure mouth sounds (e.g., crunching a piece of candy with the teeth, kissing, gurgling, crunching potato chips, finishing a can of soft drink with a straw; Gazzola et al., 2006), or listen to nonverbal human vocalizations (Warren et al., 2006).
Although what functional or evolutionary advantage that contagious yawning confers is unknown, one popular hypothesis is that it serves as some type of communicative gesture-that is, to synchronize moods in gregarious animals (Deputte, 1994), or to increase vigilance (Daquin, Micallef, & Blin, 2001; Gallup & Gallup, 2007). Certainly, the left pIFG is important for the production of verbal communication signals (Pa.zzaglia et al., 2008), and the homologous region in the right hemisphere appears to be important for nonverbal acoustic signals (Meyer & Jäncke, 2006). In fact, both brain regions activated during the perception of contagious yawns in the present study (i.e., right pIFG and pSTG) are thought to be part of a gesture-recognition network (Villarreal et al., 2008). In this regard, the right pSTG activation that correlated with yawn contagiousness ratings could also be viewed as a node for abstract (i.e., nonlexical) language comprehension (Bookheimer, 2002; Gemsbacher & Kaschak, 2003; Jung-Beeman, 2005).
pIFG Neurons: Triggers for Contagious Behavior?
Inferior prefrontal hemodynamic activity previously has been shown to increase with the amount of explicit imitation that a person exerts (e.g., using a facial imitation task; Lee et al., 2006), and more recently, this region has been shown to become active when unconscious word primes representing actions are presented to an observer (Galati et al., 2008). Furthermore, Warren et al. (2006) not only reported increased pIFG activity when people passively listened to nonverbal human vocalizations, they also found concurrent increases in the electromyographic activity of lower facial muscles, suggesting that orofacial gestural responses are engaged automatically when such stimuli are perceived. In this context, our novel finding of ventral prefrontal activity concurrently increasing with contagious impulses (i.e., the urge to carry out the behavior) leads to speculation that pIFG activation may play a role in eliciting the actual contagious event. Such conjecture will, of course, require further research.
Finally, as alluded to earlier, it is likely that attention also has a role in modulating contagious behaviors. For example, Lee et al. (2006) found that the inferior prefrontal activity associated with the imitation of facial expressions was essentially eliminated when participants were asked to perform a gender discrimination task on the same facial stimuli. This, combined with similar findings from other studies, prompted Lee et al. to speculate that the MN system may be attenuated when attention is focused on anything other than the posture or behavior to be imitated. Although this too will require further investigation, one can certainly imagine a need for such a mechanism. Indeed, without a way of modulating urges to imitate, goaldirected activities would be subject to constant disruption in environments ripe with social cues.