Yawning — Its anatomy, chemistry, role, and pathological considerations Progress in Neurobiology

(1)

Review article

Yawning—Its anatomy, chemistry, role, and pathological considerations

Heinz Krestel

^a,

*, Claudio L. Bassetti

^a

, Olivier Walusinski

^b

aDepartmentofNeurology,BernUniversityHospitalandUniversityofBern,Switzerland

b20RuedeChartres,28160Brou,France

ARTICLE INFO Articlehistory:

Received9January2017

Receivedinrevisedform29October2017 Accepted28November2017

Availableonlinexxx Keywords:

Yawning Ocytocin Dopamine NMDA GABA Serotonin Respiration Communication Arousal Braincooling Stroke Braintumor Coma Epilepsy Migraine Neurodegenerative Multiplesclerosis

ABSTRACT

Yawningisaclinicalsignoftheactivityofvarioussupra-andinfratentorialbrainregionsincludingthe putativebrainstemmotorpattern,hypothalamicparaventricularnucleus,probablytheinsulaandlimbic structuresthatareinterconnectedviaafibernetwork.Thisinteractioncanbeseeninanalogytoother cerebralfunctionsarisingfromanetworkorzonesuchaslanguage.Withinthisnetwork,yawningfulfills itsfunctioninastereotype,reflex-likemanner;aphylogeneticallyoldfunction,preservedacrossspecies barriers,withthepurposeofarousal,communication,andmaybeotherfunctionsincludingrespiration.

Abnormalyawningwith!3 yawns/15minwithoutobviouscausearisesfromlesionsofbrainareas involvedintheyawningzone,itstrajectoriescausingadisconnectionsyndrome,orfromalterationof networkactivitybyphysicalormetabolicetiologiesincludingmedication.

Contents

1. Introduction:whatisayawn ... 00

2. Anatomyofyawning ... 00

2.1. Hypothesesregardingtheneuroanatomicallocationoftheyawningcenter ... 00

2.2. Hypothalamicparaventricularnucleuscontrolsbrainstemyawningcenter ... 00

2.3. TrajectoriesprojectingtothePVN ... 00

3. Neurochemistryofyawning ... 00

Abbreviations: ACTH, adrenocorticotropic hormone; AChR, acetylcholine receptor; ALS, amyotrophic lateral sclerosis; AMPA, a-amino-3-hydroxy-5-methyl-4- isoxazolepropionicacid;Ca²⁺,acalciumion;CA1,cornuammonisarea1,aregioninthehippocampusanatomy;CO2,carbondioxide;CSF,cerebrospinalﬂuid;CT,computed tomography;D1receptor,dopaminereceptorD1,(D1R);D2receptor,dopaminereceptorD2,(D2R);D3receptor,dopaminereceptorD3,(D3R);D4receptor,dopaminereceptor D4,(D4R);EEG,electroencephalography;fMRI,functionalmagneticresonanceimaging;GABA,gamma-aminobutyricacid;Gprotein,guaninenucleotide-bindingprotein;K⁺, apotassiumion;L-DOPA,levodopaorL-3,4-dihydroxyphenylalanine;MS,multiplesclerosis;MSH,melanocyte-stimulatinghormone;NADPH,nicotinamideadenine dinucleotidephosphate;NIHSS,NationalInstitutesofHealthStrokeScale;NMDA,N-Methyl-D-asparticacid;NO,nitricoxide;NOS,nitricoxidesynthase;pO2,Partialpressure ofOxygen;PD,Parkinson’sdisease;O2,oxygen,bymolecularformula(O2);pIFG,posteriorinferiorfrontalgyrus;PVN,paraventricularnucleus;RAS,reticularactivating system;REM,Rapideyemovementsleep,aphaseofsleep;RLS,Restlesslegssyndrome;RT-PCR,Reversetranscriptionpolymerasechainreaction;STS,superiortemporal sulcus;TSH,thyroideastimulatinghormone;T2,T2-weightedmagneticresonanceimaging;vmPFC,ventromedialprefrontalcortex;5-HT1Areceptor,Gprotein-coupled receptorthatbindsserotonin(5-HT)andmediatesinhibitoryneurotransmission;5-HT2Creceptor,Gprotein-coupledreceptorthatbindsserotonin(5-HT)andmediates excitatoryneurotransmission;5-HT6receptor,anotherGprotein-coupledreceptorthatbindsserotonin(5-HT)andmediatesexcitatoryneurotransmission.

* Correspondingauthorat:DepartmentofNeurology,InselspitalUniversityHospital,Freiburgstrasse10,3010Bern,Switzerland.

E-mailaddresses:heinz.krestel@insel.ch,heinz-krestel@bluewin.ch(H.Krestel).

ContentslistsavailableatScienceDirect

Progress in Neurobiology

j o u r n a lh o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / p n e u r o b i o

(2)

3.1. Basicmodelwithyawningasoxytocinergicbehavior ... 00

3.2. Dopaminergicmodulationofthebasicmodel ... 00

3.3. Oxytocinnetworkextendingtosupra-andinfratentorialbrain ... 00

3.4. N-Methyl-D-asparticacid(NMDA)andgamma-Aminobutyricacid(GABA) ... 00

3.5. Endogenousopioids ... 00

3.6. Serotonergicyawning ... 00

3.7. Acetylcholineandnoradrenaline ... 00

3.8. Adrenocorticotropichormone(ACTH)andmelanocyte-stimulatinghormones ... 00

4. Potentialrolesofyawning ... 00

4.1. Respiratoryhypothesis ... 00

4.2. Communicationhypothesis ... 00

4.2.1. Contagiousyawning ... 00

4.3. Arousalhypothesis ... 00

4.4. Brain-coolinghypothesis ... 00

5. Abnormalyawningassociatedwithdiseasesormedication ... 00

5.1. Yawningduringandaftercerebrovascularinsult ... 00

5.1.1. Anteriorcirculationstroke ... 00

5.1.2. Posteriorcirculationstroke ... 00

5.2. Locked-insyndromesandyawning ... 00

5.2.1. Locked-insyndromeinvascularpathologyofthebrain ... 00

5.2.2. Locked-insyndromecausedbytumor ... 00

5.3. Braintumors ... 00

5.4. Traumaticbraininjury ... 00

5.5. Persistentvegetativestate ... 00

5.6. Epilepsy ... 00

5.7. Migraine ... 00

5.8. Neurodegenerativedisease ... 00

5.8.1. Amyotrophiclateralsclerosis ... 00

5.8.2. Yawninginuntreatedextrapyramidaldisease ... 00

5.8.3. Yawningintreatedextrapyramidaldisease ... 00

5.9. Multiplesclerosis ... 00

6. Conclusionandperspectives ... 00

References ... 00

1.Introduction:whatisayawn

Physiologicalyawning,i.e.yawningwithoutacauseinhealthy individuals, is a ubiquitous phenomenon that can be observed acrossspeciesbarriersatleastinmostmammals(Fig.1A,B)ifnot inmostclasses ofvertebrates(fordetailed reviewof published observationsseee.g.Baenninger,1997).Yawnsconsistofatypical sequenceofrespiratoryphasessuchaslonginspiration,briefpeak or acme, and rapid expiration, accompanied by a coordinated motorpatternincludingopeningofthejaw,closureoftheeyes, contractionoffacialmusclesandsometimesstretchingoftrunk, neck and arms. Changes in autonomic function frequently accompany yawning (in more detail e.g. Baenninger, 1997).

Yawningcanbemodulatedinitsfrequencybyoperantcondition- ing, better so in animals such as monkeys (Louboungou and Anderson,1987; Anderson and Wunderlich,1988) but also in humans(Baenninger,1997).Itsvisibleactionpattern–thatisthe coordinatedrepiratoryandmotorbehavior–hasbeendescribedas stereotypedorreﬂex-like(Lehmann,1979;Provine,1986),because once elicited it cannot be completelysuppressed. However,its appearancecanbemodulatedaswell,e.g.byinhibitingopeningof thejaw(Fig.1C)andsuppressingfacialandthoracicinnervation (Goessleretal.,2005).Yawningistermedpathological,abnormal, orexcessive if it is spontaneous, more frequentthan generally perceivedasnormal,compulsive,andnottriggeredbyappropriate stimuliincludingfatigueorboredom.Chasm(Latinforgap,cavity, cleftasreviewedinlayandmedicaldictionariesbyHeusner,1946) asasynonymispersedescriptive,butratherusedinabnormalthan physiologicalconditions.Noconsensusdeﬁnitionexistsconcern- ingthefrequencyofyawns.Descriptionsvarybetween2and30 yawnsper10min(Singeretal.,2007;Cattaneoetal.,2006).We recentlyadoptedtheabnormalyawningfrequencyto!3yawns/

15mintodecrease thelikelihood that2 subsequent, accidental

yawns were counted as one episode with abnormal yawning (Krestel et al., 2015). Abnormal yawning seems to be a rare neurobiologicalphenomenon.Itscauseinhumansisunknown,but itcanbeobservedinavarietyofmedicalconditionsthatwillbe coveredinmoredetaillateron.

2.Anatomyofyawning

2.1.Hypothesesregardingtheneuroanatomicallocationofthe yawningcenter

The neuroanatomical location of the motor pattern that orchestratesyawningin humansisnot known,respectivelyhas notbeenexperimentallyproven.However,fromearlyanalysesof malformed infantswho lacked a telencephalonbut in whoma yawn-stretchactcouldbeobserved,itwasconcludedthatifayawn centerexists,“itistobefoundinthemedullaoblongatainthe immediate vicinity of the respiratory and vasomotor centers”

(Fig.2A;Heusner,1946).Thisstatementwasdeducedfromtwo earlieraccounts.Theﬁrstdescribedananencephalicfemalewho survived3 monthsand inwhom –postmortally– themedulla spinalis&oblongata, ponsand mesencephalonwereseemingly welldeveloped(Gamper,1926).The telencephalonwasmissing and the most cranial brain tissue was a severely malformed diencephalon. A certain residual function of the hypothalamus could not be excluded because the substantia grisea centralis behindthe3rdventriclecontainedwell-formedneuronalnests– howeverwithoutresemblancetoknownneuronalnuclei.Asthis child showed phase switches betweensleep and wake, a fully developedyawn-stretchactandnoyawningnorstretchingduring sleep–henceaseeminglynormalyawningbehavior–,Gamper concludedthattheyawn-stretchcenterisasubcorticalmechanism and should be located in the substantia grisea centralis or

(3)

periaqueductalgrey(Fig.2B)butmightwellbeinﬂuencedbythe telencephalon [in healthy subjects]. The second noteworthy accountwasbyCatelandKrauspe(1930)whoobservedyawning in a seven days surviving female anencephalus. The nervous systemwasapparentlywelldevelopeduptotheareaaroundthe trigeminalnerveintheponto-mesencephalictransition,however, thepyramidal tract and parts of ventral pons weremissing or largelyrudimentary.CranialnervesXIItoV,includingtheirnuclei werewell formed. The rudimentary diencephalon consisted of irregularlyformednon-coherentcellularconglomerates,cystsand plexus-likestructures.Averyimmatureoptictractoriginatedfrom 2ofthesecellularconglomerates,which hencemusthavebeen parts of the thalamus. However, no further anatomical details couldbedelineatedinthisbrainpart.Anintactconnectiontothe medulla oblongatawas not found.More cranial(i.e. forebrain) venousplexus, mesenchymal cysts, partly isolated cavities and dispersedcellnests,remindingofrudimentarybasalganglia,were found (Catel and Krauspe, 1930). In more recent publications, Askenasy(1989)locatedtheyawningcenter“atthelevelofthe reticularbrainstemclosetothereticularactivatingsystem”(RAS) and substantiatedhis suggestion bycase reports of patientsin whomthecorticobulbarpathwayswereinterruptede.g.bytumors andwhoweretetraplegicbutcouldyawn.TheRASiscomposedof severalneuronalcircuitsconnectingthebrainstemtothecortex.

The most caudal circuits are located in the midbrain reticular formation and in mesencephalic and pontine nuclei. Thus, according to today’s knowledge, Ashkenasy’s yawning center wouldresidesomewhereinthemidbrain(Fig.2A).Third,itwas proposedthattheyawningcenterislocatedamongpontomedul- larycentralpatterngenerators(Walusinski,2006;Fig.2A).Central pattern generators are neuronal circuits in the medulla and subserveinnate,repetitivemotorbehaviorsthatareessentialfor survivalsuch ascough,swallowing,breathing (e.g.Marder and Rehm,2005).Walusinski(2006)suggestedthattheyawningmotor

patternrecruitscontrolsystemsincludingthelocuscoeruleusand theparaventricularnucleus.

2.2.Hypothalamicparaventricularnucleuscontrolsbrainstem yawningcenter

In principle, the central nervous system is composed of a

“primary” and “secondary neuron”. For efferent systems this frequentlymeansthatthe“ﬁrstneuron”(ornetwork,frequently locatedinthesupratentorialbrain)exertssomecontroloverthe

“secondneuron”(ornetwork,frequentlylocatedintheinfraten- torial central nervous system). This central nervous system organizationissometimesalsocalledtop-downcontrol.Inregard toyawning,theparaventricularnucleus(PVN)inthehypothala- mus(Fig.3)isfrequentlyregardedasthesupratentorialcontrolof theyawningcenterinthebrainstem.Thisleadsustoafrequently cited studyin anesthetized rats, inwhich stereotyped yawning couldbeelicitedbychemicalorelectricalstimulationofthePVNin thehypothalamus(Sato-Suzukietal.,1998).Thestudyproposed thatparvocellularoxytocinergicneuronsinthePVNprojectingto thelowerbrainstemmediatetheyawningresponseandthatnitric oxide is an important player. This concept had already been suggestedelsewhere(seee.g.ArgiolasandMelis,1998,in3.1.Basic model). Asthe PVN might be involvedin yawning not only in rodentsbutalsoinhumans,basedonconservationoftheactof yawningacrossspeciesbarriersatleastinmammals,wewouldlike to review its neuroanatomy including its connections to other brainareasandthebrainstemnuclei.Inhumans,theanteriorpart ofthehypothalamuscontainsneuronalnucleiofwhichthemost importantarethepreoptical,supraopticalandtheparaventricular nuclei(Fig.3).ThePVNcontainsneurosecretorymagnocellularand parvocellularneuronsaswellascentrallyprojecting(i.e.toother brainregions)neurons.Magnocellularneurosecretoryneuronsare known for their production and transport of oxytocin and

Fig.1.Physiologicalyawning.Thisisanubiquitousphenomenonacrossspeciesbarriersatleastinmammals;heredepictedinahorse(A)andatiger(B).Yawningcannotbe completelysuppressed,onceelicited.However,itsappearancecanbemodulatedtocertainextente.g.by(partially)inhibitedopeningofthejaw(C).

(4)

antidiuretic hormone via their axons (supraopticohypophyseal tract) to the posterior pituitary gland where these peptide- hormones are secreted and exert their neuro-endocrinological function such as oxytocin does in regulating menorrhea, con- tractions of the pregnant uterus, milk secretion from female mammaeandtrustbehavior(Kosfeldetal.,2005).Viareleasingand possiblyinhibitoryhormones,parvocellularneurosecretoryneu- ronsindirectlyregulatetheproductionofhormonesintheanterior pituitaryglandsuchasgrowthhormone,gonadotropins,TSH,and ACTHetc.Finally,thePVNcontainsinterneuronsandpopulations of neurons including parvocellular, oxytocinergic neurons that projecttootherbrainregions.Oneoftheseprojectionsprobably occurs via the medial forebrain bundle to structures in the brainstemimplicatedinregulationofrespiratory, cardiovascular and autonomic functions including thelocus coeruleus, solitary nucleus, ventrolateralmedulla, the motor nucleus of thevagal

nerveandviasynapsingtofurthermotornucleisuchastheonesof the trigeminal, facial and hypoglossal nerve (Duus, 1995).

Especially the ventrolateral medulla contains a cluster of interneurons – named Pre-Bötzinger complex – that seems to beessentialforthegenerationofrespiratoryrhythmsinmammals (Smithetal.,1991).Theexactmechanism ofrhythmgeneration andtransmissiontomotornucleisuchasthephrenicnerve(C1-C4) innervating the diaphragm remains controversial (Rybak et al., 2007; Abdala et al., 2009). Further neuronal effector centers involvedinyawningarethemotornucleiofcranialnervesV,VII,IX, XI[andXII,moreinanimals?],andnervesinnervatingaccessory respiratory muscles (Goessler et al., 2005). Additional data, generatedinanimals,abouttheintranuclearorganizationofthe PVN and its projections is presented here, as it might be of relevance to better understand the neuroanatomical basis of yawning.Cellsintheparvo-andmagnocellulardivisionofthePVN

Fig.2.Hypothesesregardingtheneuroanatomicallocationoftheyawningmotorcenter.A)Showsasagittalsectionofthebrainstem,rostrallyincludingthehypothalamus andthepituitarygland(Duus,1995).Heusner(1946)proposedthattheyawningmotorcenterislocatedinthevicinityofrespiratoryandvasomotorcentersinthemedulla oblongata.Askenasy(1989)locatedtheyawningcenterclosetothereticularactivatingsystem(RAS),whosemostcaudalcircuitswouldtodayresideinthemidbrainand upperpons.Third,someauthorsbelievethattheyawningcenterexistsamongpontomedullary,centralpatterngeneratorsforcough,swallowing,breathing.B)Showsanaxial sectionthroughthemesencephalonincludingtheaqueductandcentralgray(Duus,1995).Gamper(1926)deducedthelocationoftheyawningcenterfromstudiesof malformedkidswithmissingtelencephalontoasubcorticallocationinthebrainstemwithinthecentralorperiaqueductalgray(highlighted).III,oculomotornerve;IV, trochlearnerve;VIIfacialnerve;IX,glossopharyngealnerve;X,vagusnerve;XII,hypoglossalnerve.

(5)

haveaxons thatbranchsometimes– socalledaxoncollaterals.

Severalaxoncollateralsofparvocellularneuronsramifylocallyand appear to contact dendrites of cells in both the parvo- and magnocellulardivision.Thisevidencesuggestsonepossibilityhow the output of parvo- and magnocellular neurons might be integrated (Swanson and Sawchenko, 1983, and references therein)andcouldbeimportantinregardtothehypothesisthat cellsintheparvocellulardivision[only]mediatethecontrolling effects on yawning (see 3.3. oxytocin network). In addition, magnocellularneuronsseem tocommunicatebygapjunctions.

The predominant projection to the brainstem is the so called paraventriculo-spinal pathway, discovered in 1975, that was originally described to directly connect the PVN and other hypothalamic areas with preganglionic cell groups of both the parasympathetic and sympathetic divisions of the autonomic nervoussysteminthedorsalvagalcomplexand thoracicspinal cord,respectively.Subsequentworkbyseveralgroupsaddedmore detailtoitscourse.Incowandrat,fibersoriginatinginthePVN descendinitiallythroughthemedianforebrainbundleandthen, afterpassingbetweensubstantianigraandrednucleus,continue throughventrolateralpartsofthereticularformationtoenterthe dorsolateralfuniculusofthespinalcord.Inthepons,fibersleave thispathwaytoinnervatetheparabrachialnucleusandthelocus coeruleus, while in the medulla, some fibers arch dorsally to innervatethedorsal motor nucleus of thevagal nerve and the nucleustractussolitarius.Inaddition,asecondpathwaydescends through the central grayand appears toinnervate the Edinger Westphal nucleus,locus coeruleus,and perhaps thecentral gray (substantiagriseacentralis),aswell(SwansonandSawchenko,1983 and references therein). This detailed description also contains reported fibers to neuronal networks suggested to home the putative yawning motor pattern including the central or periaqueductalgrey(Gamper,1926),themidbrainreticularformation near the reticular activating system (Askenasy,1989), and the ponto-medullarybrainstem nearthereticularformation (Walu- sinski,2006).

2.3.TrajectoriesprojectingtothePVN

Most of neural afferents to the PVN origin from a rather small number of cell groups in the brainstem, hypothalamus, and limbic regionsof the telencephalon. The brainstemsends noradrenergicandadrenergicﬁberstothePVN,whichformone

of the most dense catecholamineterminal ﬁelds in the brain.

Mainly noradrenergic ﬁbers project from A2 [noradrenergic]

(andC2[adrenergic])cellgroupsofthenucleustractussolitarius and from the A6 group of locus coeruleus tothe parvocellular division,andfromA1(andC1)cellgroupsoftheventralmedulla (locateddorsaltobutnotwithinthelateralreticularnucleus)to both the parvo- and magnocellular division (Swanson and Sawchenko,1983 and referencestherein). Sparseserotoninergic afferents reach the PVN mainly from mesencephalic raphe nuclei.ThehypothalamusprojectstothePVNfromtheanterior and lateral hypothalamic area, the ventromedialand dorsomedial nucleus, and the preoptic area which all end in the parvocellular division. Only projections from the dorsomedial nucleus and themedian preopticnucleusappear toend in the magnocellular division, as well. The suprachiasmatic nucleus, which receives direct input from the retina, projects to the parvocellularpart.ACTHimmunoreactiveﬁbersarefoundinthe parvocellular divison and in the magnocellular division where oxytocinergic cells are concentrated. These projections arise from immunoreactiveneuronsin and nearthe arcuate nucleus of the hypothalamus. Because

b

-endorphin has been co- localized within many ACTH-stained neurons,

b

-endorphin- containingfibers alsoreachthePVN (Swanson andSawchenko, 1983 and references therein).The telencephalon mayinfluence PVNactivity,butratherthroughitslimbicsystem.Thereisasyet noevidencefordirect projectionsfromneocorticalareastothe PVN. Retrograde transport studies in the rat suggest that the lateral nucleus of the septal region, the amygdala (medial nucleus), and thehippocampus(specifically theventral partof thesubiculum)projecttothePVN.However,theseresultsdonot seem tobewithout controversy(see Swansonand Sawchenko, 1983 and references therein). Other sources support [oxytocinergic] connectionsbetweenthehippocampus,tuberomam- millarybodies,andthePVN(e.g.Walusinski,2006;Argiolasand Gessa,1991).Thebednucleusofthestriaterminalismaypresent one additionalrouteby whichthe limbicregionmayinfluence parvocellular and oxytocinergic magnocellular neurons in the PVN (Swanson and Sawchenko, 1983 and references therein).

Future workshouldbedirectedintovalidatingthebiochemical natureofﬁberstotheputativeyawningmotorpattern,giventhe factthatcirca30differentputativeneurotransmittershavebeen identiﬁed in cell bodies or in presumed terminals within the PVN (Swanson andSawchenko,1983).

Fig.3.Supratentorialcontroloftheyawningcenterinthebrainstem.Shownisacoronalbrainsectionthroughthe3rd(III)ventricleandthesurroundinghypothalamus (Duus,1995).Shownarealsovariousnucleiwithinthehypothalamus.Notetheparaventricularnucleus(highlighted,PVN)onthemoreanteriororrostral,right,brainsection (accordingtoneuroradiologicalconvention).Electricaland/orchemicalstimulationofthePVNcanexperimentallyelicityawning.AsimilarPVNfunctioninhumansis deducedfromthepreservedyawn-stretchacrossspeciesbarriersinmammals.

(6)

3.Neurochemistryofyawning

3.1.Basicmodelwithyawningasoxytocinergicbehavior

A lot of work has been invested into elucidating the neurochemistryinvolvedinyawning.Thesubsequentlypresented overviewcentersonamodel,proposedbyagroupthatcontributed significantlytothisfieldandalsointegratedworkofmanyothers (ArgiolasandMelis,1998).Theirmodelisextendedandupdated withunmentioned and more recent findings. The basic model suggeststhatoxytocinergicneuronsoriginatinginthePVNofthe hypothalamusand projecting toextra-hypothalamicbrainareas mediatetheexpressionofyawninginanimalsinmanybutnotall circumstances. Activation of neurons in the PVN by dopamine receptoragonists,excitatoryaminoacids,andoxytocinresultsin yawning, while their inhibition by e.g. opioids prevents the behavioralresponse.Dopaminereceptoragonistsbindmainlyto D2receptorsandoxytocintouterine-typeoxytocinergicreceptors andthisleadstoactivationof[omega-conotoxin-sensitive]N-type Ca²⁺channels via[pertussistoxin-sensitive] G0/Gq proteins and subsequentlytocalciuminflux.Toaminorextent,Ca²⁺isreleased fromintracellularstoresviathePhosphoinositid-PhospholipaseC (PLC)pathway.Excitatoryneurotransmittersincludingglutamate and N-Methyl-D-aspartic acid (NMDA), but not

a

-amino-3- hydroxy-5-methyl-4-isoxazolepropionicacid(AMPA),wereorigi- nallysuggestedtoactivateNMDAreceptors,whicharecoupledto Ca²⁺ channels. This was conﬁrmed by subsequent agonist- antagonist studies showing that NMDA-induced yawning was antagonized by (+)-MK-801, a selective antagonist of NMDA receptors, but not by omega-conotoxin, a potent blocker of N- typeCa²⁺channels(Succuetal., 1998),anditisthereforelikelythat NMDA receptors mediate Ca²⁺ inﬂux themselves. Increased cytoplasmicCa²⁺concentrationinoxytocinergicneuronsactivates nitricoxide synthase(NOS)toproducenitricoxide(NO) thatis suggestedtoact asintracellular messenger.Themechanism by which NO activates oxytocinergic transmission is not known.

Guanylatecyclase, one favored target of NO, seemed then not involved.The inhibitory effectof opioids suchas morphine on yawninginducedbydopamineD2receptoragonists,oxytocin,or NMDAisapparentlymediatedby

m

-typemorphinereceptorssince themorphineeffectisblockedbytheprioradministrationofthe respectiveantagonistnaloxone.Inaddition,averypotentagonist actingatkappaopioidreceptorsisineffectiveinelicitingyawning.

Signaltransmissionfromextra–tointracellularmaybemediated bydecreasedintracellularCa²⁺andreducedNOSactivity,asless NOSmetabolitesweremeasuredbyinvivomicrodialysisinthe PVNafteradministrationofmorphine.Thelinkbetween

m

-receptor activation and decreased intracellular Ca²⁺ was stated as unknownbyArgiolasand Melis (1998),but is probably due to closureofvoltage-gatedCa²⁺channels(e.g.Schroederetal.,1991).

Finally,itwassuggestedthatthePVNisapparentlynotinvolvedin yawninginducedby5-HT2CagonistsorACTH/MSH.Itwasfurther proposedthatthelattercompoundsactatsiteslocated[priortoor]

afteroxytocinergicneurons,but there existprobablyadditional neuronalpathwaysthatinﬂuenceyawning.Thiswasalsoproposed forcholinergic, noradrenergic&GABA-ergicsystemsinvolvedin yawning(ArgiolasandMelis,1998).

3.2.Dopaminergicmodulationofthebasicmodel

The PVN receives dopaminergic afferents, which was concluded from the presence of dopaminergic ﬁbers in the vicinityof parvo- and magnocellular neurons of the PVN and from dopaminergic boutons synapsing on magnocellular neurons.Theoriginoftheseﬁberswasnotpreciselylocatedbutthe ventral tegmental areawas suggested (Buijs et al.,1984). The

basicmodelbyAgriolasetal.suggestedthatthedopaminergic effectofyawningismediatedbypostsynapticdopamineD2but not D1 receptors. The postsynaptic site of dopamine receptor activation was later conﬁrmed (Collins and Eguibar, 2010).

Recently, the dopamine D2 receptor was questioned and the dopamine D3 receptorwas suggestedinstead. This wasbased onagonist-antagoniststudiesinrats(CollinsandWoods,2008;

Collinsetal.,2008;Baladietal.,2010).Thesamegroupwenton to propose speciﬁc roles for the D3 and D2 receptor in dopamineagonist-inducedyawning(CollinsandEguibar,2010).

Otherexperimentaldata(againinrats)didnotsupportamajor roleof dopamine D3and D4receptors, butof D2receptors in dopamine agonist-induced yawning (Depoortère et al., 2009;

Sannaetal.,2012a).ThereisevidencethattheD2receptorstill might playa rolein dopamine agonist-induced yawning. This comesfromtreatmentofhumanswithParkinson’sdisease(PD).

In these patients, apomorphine, a relatively non-selective dopaminereceptoragonist,withpossiblyslightlyhigherafﬁnity for D2-like dopamine receptors, is used to test dopaminergic responsiveness or intermittently treat parkinsonian motor ﬂuctuations. Yawning was quite commonly observed after apomorphine administration (Frankel et al., 1990; O’Sullivan et al.,1999 and references therein). Four major dopaminergic pathwayshave beendescribed in mammalianbrain. Theseare the mesocortical and "limbic pathwayswith their dopamine- producing somata in the ventral tegmental area, the nigros- triatal pathway and the tuberoinfundibular pathway with its dopamine-containing somata in the hypothalamus. To what extentthenigrostriatalsystemisinvolvedinyawning, remains tobeanswered.However,microinjectionofnanogramamounts ofapomorphine andotherdopamine D2agonistsintothePVN butnot in the striatumof rats induced yawning(Melis et al., 1987). Apart from possible involvement of several receptor subtypes in yawning, dopamine receptors seem to underlie diurnal variations in sensitivity, since yawning, induced with identical doses of apomorphine, could be elicited more frequently in the morning than in the evening in healthy men(Lalet al.,2000).Dopamine-agonistinducedyawningcan be modulated by other transmitters. These include nicotine (Tizabiet al.,1999;Brown et al., 2006), adenosine(Rimondini et al., 1998), endocannabinoids (Beltramo et al., 2000;

Nakamura-Palaciosetal.,2002),and noradrenaline(Sawchenko and Swanson, 1981; Nowak et al., 2006). A dense catechol- aminergic innervationwas showntosynapse onPVNneurons (e.g. Swanson et al., 1981).Of interestis also the modulatory effect ofdexamethasone, sinceyawning hasbeenpostulated– amongst others- as a signal for stress (see Communication hypothesis, below). Dexamethasone modulated (increased as single and decreased as repeated administration) pilocarpine- butnot apomorphine-inducedyawningin rats(Hipólide etal., 1999).

3.3.Oxytocinnetworkextendingtosupra-andinfratentorialbrain Oxytocin immunoreactivity is found in magnocellular (e.g.

Buijs,1978;Hatakeyamaetal.,1996)andinparvocellularneurons (e.g.Buijsetal.,1978).ThepresenceofNOS,indicatedbypositive NADPHdiaphorasestaining,colocalizeswithoxytocinergicmag- nocellularneurons(e.g.Sánchezetal.,1994)andisalsofoundin therostral partof thePVN where ratherparvocellular neurons reside (e.g. Sato-Suzuki et al., 1998). However, a direct co- localizationofNOSandoxytocininparvocellularneuronsescaped our literature review. The type of oxytocineric PVN neuron involved in yawning has not been unequivocally proven, but reportsofrecentyearsfavorparvocellularneurons(seee.g.Sato- Suzuki et al.,1998; Kita et al., 2006). Neuroanatomically, both

(7)

parvo- and magnocellular neurons were shown to have extra- hypothalamicprojections.Oxytocin-containingfibersfrommag- nocellular PVN neurons were traced to the dorsal/ventral hippocampus, amydgala, substantia nigra, central gray, nucleus tractus solitarius, and nucleus ambiguus in rats. In contrast to projections to rostral brain areas, far more oxytocin- than vasopressin-containingfibers werefoundinthemedullaoblon- gata(andspinalcord)(Buijs,1978).Theauthorevensuggestedthat some of the bi- and multipolar PVN cells may project to the neurohypophysis,aswellastoextrahypothalamicareasandadded thatitisnotpreciselyknownwhetheralloftheseextrahypotha- lamicfibersareaxonsorwhethersomemightbe“dendriteswitha receptivefunction” (Buijs,1978).Incontrast,retrogradedouble- labelingstudiesfromanothergroupindicatedthathypothalamic projectionstothebrainstemorspinalcorddonot,forthemost part,arise ascollateralsfrommagnocellularPVN cellsthat also projecttotheposteriorpituitarygland(reviewed inSawchenko and Swanson, 1982). Their retrograde labeling studies found roughly 10 times more cells in the parvo- than magnocelluar disvisionofthePVNthatcontainedaretrogradetracer(injected intothe dorsalmotor nucleus of thevagus and nucleus of the solitarytract)andwereimmunoreactiveforoxytocin(Sawchenko and Swanson, 1982). In a follow-up study – by injecting a retrogradetracerintothedorsalvagalcomplexandthoracicspinal cordandstainingthePVNimmunocytochemicallyagainstoxyto- cin and vasopressin – fibers within the paraventriculo-spinal pathwaycouldbeidentifiedthatoriginatedpredominantlyfrom oxytocinergic cells, located in the parvo- and magnocellular divisionofthePVNbutmorefrequentlyinthemedialandlateral parvocellulardivision(SwansonandSawchenko,1983).However, only20%ofretrogradelylabeledcells stainedpositiveforeither oxytocin or vasopressin suggesting that additional cell types contributed to this tract. Interestingly, additional 5% of cells containingtheretrogradetracerstainedimmunocytochemically positivefortyrosinehydroxylase(suggestedbytheauthorstobe dopaminergic),somatostatin,andmet-enkephalin.Apartfromthis tractconnectingthePVNwithpreganglioniccellsofparasympa- theticandsympatheticdivisionsoftheautonomicnervoussystem, nostatementwasmadeaboutthenatureoffibersthatbranchfrom theparaventriculo-spinalpathwaytoputativeneuronalnetworks containingtheyawningmotorpattern.Still,fibersfrommagno- cellular PVN cells also project to the vicinity of putative cardiovascular and respiratory nuclei. Several experimental studiesinratsshowedthatelectricalorchemicalstimulationof theparvocellulardivisionofthePVNaffectedyawning(e.g.Sato- Suzukietal.,1998;Kitaetal.,2000).However,thePVNissosmall intheseanimalsand eveniftargetingwasprecise,it cannotbe excludedthatmagnocellularneuronsinthevicinityareexcitedas well.Therefore,theissuemightnotbefinallysettledwhichtypeof PVNneuronisinvolvedinyawningcontrol.Thehippocampuswas repetitivelyimplicatedinyawningasoxytocin injectionsintoit inducedyawning(e.g.ArgiolasandGessa,1991).Theseobserva- tionsmightreflectanindirecteffectusingthePVNassortofrelay station, ashippocampal-hypothalamic oxytocinergic projections wereshowntoexist.Similarily,oxytocinmayinduceyawningif injectednotonlyinthePVNandhippocampus,butalsoinother brainareas,suchastheventraltegmentalareaandtheposterior nucleus of the amygdala that, like the hippocampus, receive oxytocinergic projections from the PVN (Sanna et al., 2012b).

Oxytocinseemstoact asneurotransmitter inthe brainstemas iontophoreticallyappliedoxytocinisabletochangetheﬁringrate ofsomeneuronsinthedorsalmedullaandoxytocinisreleasedin Ca²⁺ dependent manner from tissue slices of medulla upon potassium application (Swanson and Sawchenko, 1983). The putativelinkbetweentheactionofNOasintracellularmessenger andoxytocinergictransmissioninyawningisstillmissing.

3.4.N-Methyl-D-asparticacid(NMDA)andgamma-Aminobutyric acid(GABA)

AcorrelationbetweenNMDA-inducedyawningandNOSinthe PVN was suggested. First, NMDA-induced yawning in freely movingconsciousratscouldbeblockedbyaNOSinhibitor(Melis et al.,1994a).Second,ratsthathad beenrendereddiabeticand impairedin theirNMDA-inducedyawningresponse bystrepto- zotocintreatmentwereinjectedadenovirusescarryingagenefor neuronal NOSinto thePVN. This couldrestore NMDA-induced yawning (Zheng et al., 2007).Excitatory (glutamatergic) amino acidactivationisrathermediatedviaNMDA-thanAMPAreceptors, asadministrationofNMDAbutnotAMPAintothePVNinduced yawning(Melisetal.,1994b;CollinsandEguibar,2010).Yawning induced byinjectionof a NO-releasingcompoundor glutamate into the medial parvocellular PVN of anesthetized rats was paralleled in its effect byinjectionof cyanide, a mitochondrial cytochromecoxidaseinhibitor.Thisledtothesuggestionthatthe medialparvocellulardivisionissensitivetochemicalhypoxiaor ischemiaandevencontainsanoxygensensorthatcausesyawning (Kitaetal.,2000;seealso4.1.Respiratoryhypothesis).GABAmay also modulate yawning by affecting NOS activity. Injection of GABA-A(muscimol)butnotGABA-B(baclofen)receptoragonists intothePVNreducedyawninginducedbyapomorphine,NMDA, andoxytocinandwasparalleledbyadecreaseinNOSmetabolites as measured in a paraventricular microdialysate (Melis and Argiolas, 2002).However,this effectis controversal (see Doger etal.,1989).

3.5.Endogenousopioids

Enkephalins,beingendogenousopioids,arefoundinmagno- cellularandparvocellularPVNdivisions,andtheparaventriculo- hypophyseal tract mayalso contain enkephalins (Swanson and Sawchenko, 1983). Yawning is a prevalent sign of the opiate withdrawal syndrome in human opiate addicts (O’Brien, 1996;

reviewed in Argiolas and Melis, 1998). Adenosine and opiate systems were shown to modulate each other, resulting in a withdrawalsyndromeincludingyawning,ifrats–physicallymade dependent with adenosine agonists – were challenged with adenosineantagonists,orifadenosineantagonistsweregivento morphinepre-treatedrats(CouparandTran,2001).Intracellular phosphorylation, convertingtheopioidreceptorintoaconstitu- tivelyactiveform,playsaroleintheopiatewithdrawalsyndrome in dogs.Administrationofa proteinkinaseC inhibitor, priorto renderingdogsopioiddependent,attenuatedacutesymptoms,if withdrawalwasinduced(FreyeandLevy,2005).

3.6.Serotonergicyawning

Serotonin(5-HT)2Creceptoragonistsareeffectiveatinducing yawning inratsand humans,which canbeblockedby 5-HT2C antagonists. Stimulation of 5-HT1A receptors can also block dopamine induced yawning (Argiolas and Melis, 1998). The relationship between serotonin and yawning is however not straightforwardinrats:5-HT2Creceptoragonistsdonotinduce yawning whenthey aredirectly injectedintothePVN; 5-HT2C receptorantagonistsdonotpreventdopamine/oxytocininduced yawning and 5-HT2C mediated yawning is not prevented by oxytocinreceptorantagonistsbutbyNOSinhibitors,whengiven intolateralventricles(ArgiolasandMelis, 1998).Inaddition,5-HT6 receptors play a role in the control of yawning, as respective antagonistscould increasethenumber of yawns(Sleight et al., 1998), butnot incontinuous treatment forsevendays (Marcos etal.,2008).Inhumans,yawningwasreportedassideeffectduring treatment with selective serotonin reuptake inhibitors such as

(8)

paroxetine(Harada,2006),escitalopram(dosagereductiontermi- natedyawning)(Gutiérrez-Alvarez,2007),thecombinedserotonin andnorepinephrinereuptakeinhibitorDuloxetine(DeLasCuevas and Sanz,2007) and withvenlafaxine, a reuptake-inhibitor of serotonin,norepinephrine and weaklydopamine(Chen and Lu, 2009).ItssiteofactionwasproposedtolayoutsidethePVNandwe recentlyproposedthattheinsulamightbethelongsought-after brain region for serotonin-mediated yawning (see also 5.1.1.

Anteriorcirculationstroke;Kresteletal.,2015).

3.7.Acetylcholineandnoradrenaline

Cholinergic or "mimetic drugs such as the acetylcholine esterase inhibitor physiostigmine and the muscarinic receptor agonistpilocarpine can induce yawning in rodents.This effect seemstobe mediated by M1-and M2 subtypes of muscarinic acetylcholine receptors (AChR) as the muscarinic receptor antagonists atropine and scopolamine, but not nicotinic AChR antagonistspreventedyawninginducedbyACTH,dopaminergic agonists,andoxytocin(ArgiolasandMelis,1998).Itisnotknownat whichneuroanatomicalsiteacetylcholineand itsagonistsaffect yawning.Possibilities include modulationof the PVN afferents, efferentsandatsomataofPVNneuronsthemselves.Lesionstudies suggestedacentralroleofsepto-hippocampalcholinergicneurons intheinductionofcholinergicyawning,butitremainedunclear whichtrajectoriesreachedthebrainstemtoinducetheyawning motorpattern(CollinsandEguibar,2010).Theabovementioned observation that muscarinic receptor antagonists prevented yawninginducedbyabovementionedtransmittersandhormones suggeststhatmodulationtakesplaceatthelevelofPVNefferents ordirectlyin thebrainstem.Indeed, muscarinicAChRsubtypes M1-M5 have been found in pons and medulla of rats with biochemicalmethods(Weietal.,1994).Eventhenucleustractus solitarius(Endoh,2007;patchclampexperiments)andtherostral ventrolateralmedulla(Kumar etal.,2009; quantitativeRT-PCR) wereshowntoexpressM2typemuscarinicAChRinrat.Theproof howeverisstillmissingwhethercholinergictrajectoriesfromthe PVNtobrainstemexistthatmaydirectlyinduceyawning.Onthe other hand, as intense bidirectional noradrenergic projections between the parvocellular division and the brainstem (locus coeruleus,nucleus tractussolitaries, dosal vagalcomplex) exist (e.g.Sawchenkoand Swanson,1981), adrenergic modulationof yawningviadirectPVNbrainstemprojectionsseemspossible.

3.8.Adrenocorticotropichormone(ACTH)andmelanocyte- stimulatinghormones

Adrenocorticotropichormone(ACTH)andmelanocyte-stimu- lating hormones (MSH) can induce a stretching-yawning syn- dromethat was considered different fromthe classic yawning syndrome (induced e.g. by dopamine agonists or oxytocin), becausethebehaviorstartsonly20to30minafterintraventricular injection and lasts for hours. Some evidence concerning the neuroanatomical site of action comes from a study in which injectionofACTHintotheperi-[notpara-]ventricularareaaround the3rdventriclecouldinduceyawningandbeblockedbyprior installation of a melanocortin 4 receptor antagonist. However, yawning could not be elicited by injection into CA1 area of hippocampus, caudatenucleus or preopticarea(Argiolaset al., 2000).TheinvolvementofPVN–brainstemprojectionsinACTH/

MSHinduced yawningremains controversial.It shows insome instances analogies to dopamine or oxytocin-induced yawning such as prevention of yawning by omega-conotoxin and NOS inhibitorsand hypophysectomy.On theother hand ACTH/MSH inducedyawningisnotpreventedbyPVNlesionsnorbydopamine oroxytocinreceptorantagonists(reviewedinArgiolasandGessa,

1991).Therefore,bothpossibilities-i.e.involvementanddetouring thePVNinyawning–seempossible.

4.Potentialrolesofyawning

4.1.Respiratoryhypothesis

Besides unraveling the neuroanatomical and chemical blue- printofyawning,ithasalwaysbeenofinteresttounderstandits roleinbiologyandwhyithasbeenevolutionaryconserved.Oneof theearlierandstillwidelyheldbeliefsisthatyawningisdueto changesinrespiratoryfunctionorbrainperfusion(withoxygen), respectively.TherespiratoryhypothesisdatesbacktoJohannesde Gorter(1689–1762),aprolificDutchauthorwho,inhisbook“De Perspirationeinsensibili”,attributedyawning“toaneedforfaster bloodcirculationandtocerebralanemia”.Morerecently,H.Russell’s monograph from 1891, which is one of the earliest and most extensiveprintedexcerptsaboutyawningintheEnglishlanguage, suggestssuchphysiologicalpowerforyawning.Herein,yawningis anautomaticimpulsecausedbybadairinthelungs,designedby natureasa gymnasticandintendedbothtoawakenrespiratory organsintoactivityandtoeffectastimulationofthebrainthrough increasedactivityofthecirculation(Russell,1891).Althoughthis concepthasbeenfrequentlyattackedandseemstodatetheleast experimentally validated hypothesis about the purpose of yawning,itisstillnotwithoutfavoramongclinicians.Oftencited together is the theory that yawning is caused by anemia or insufficientperfusionofthebrain.Oneofitsearlyprotagonistswas ValentinDumpertwhosuggestedthatyawningisnotonlycaused byanemiaorpoorbraincirculation,butinadditionatendencyto defendimpaired consciousness is required[For himthe yawn- stretchactwasanelementaryindirect vascularreflex(“elemen- tarer indirekter Gefässreflex, dem das ganze Blutgefässsystem unterstelltist”)]thatcontrolledallbloodvessels(Dumpert,1921).

Hence,thetwotheoriestogethercanbeformulatedthesedaysas thehypothesisofanimpactof(cerebral)hypoxemiaorhypercap- niaonyawningfrequencyandduration.Observationsinfavorof this hypothesishavebeen multiplesuchas: i)Enhancementof venousbackflowtotheheartand activation oftheparasympa- theticnervoussystemwithdilation ofarterioles andbronchioli duringyawnsresultsinincreasedrespiratoryeliminationofcarbon dioxideand uptake ofoxygen (reviewedin Askenasy,1989).ii) Deepinspirationremovedatelectasisthataccumulatedduring2h ofnormalventilationinexcisedratlungsbyasinglelargeinflation and wasassumed toparalleltheinvivo situationwithshallow breathing andsubsequent yawning(Thetetal.,1979).iii)Brain hypoperfusion bystepwise reduction of arterialblood pressure triggered yawning and decrease in pO2 and EEG frequency (Karasawa et al., 1982). iv) Experimentally induced chemical hypoxiabylocalinjectionofcyanide,amitochondrialcytochrome coxidaseinhibitor,intotheparvocellularpartofthePVNelicited yawninginanesthetized rats,promptingtheauthorstosuggest thatthisareacontainsanoxygensensor(Kitaetal.,2000).Indeed, current understanding of candidates in the carotid body – supposed to register oxygen fluctuations in blood – include NADPH oxidases generating reactive oxygen species in O2

dependentmanner,oxygen-regulatedplasmalemmalK⁺channels, andcytochromecoxidaseswhichcanrespondwithdepolarization of mitochondrial membranes and Ca²⁺ release (Kummer and Yamamoto, 2002). However, the currently favored model for oxygensensinginthecarotidbodyisinhibitionofO2sensitiveK⁺ channels in glomus cells with subsequent depolarisation, Ca²⁺

entry, and transmitter release activating afferent nerve ﬁbers (López-Barneo et al., 2008). Evidence against the respiratory hypothesisinclude observationsofyawns –mouthmovements, grimaces,tongueprotrusions–asearlyasinhumanandrodent

(9)

fetuses(VanWoerdenetal.,1988;Shereretal.,1991;Walusinski, 2010). Althoughfetuses require oxygen and must expelcarbon dioxide, they clearly do not use the pulmonary respiration mechanism. Fetalmouth movements, grimacing and stretching mightbetheexerciseofamotorpatternbywhichneonatesare abletoproducenormallyintegratedyawnswithinminutesafter beingborn.Second,astudyinyoung,healthystudentsinwhom neitheryawningrateand/ordurationwasalteredbybreathinggas mixtureswithhigherthannormallevelsofCO2(3or5%)or100%

O2, although both affected breathing rate. Physical exercise sufﬁcient to double breathing rate had no effect on yawning, too.TheauthorsconcludedthatalteredpartialpressuresofO2and CO2inblooddidnotaffectyawning(Provineetal.,1987).However, thereportdoesnotmentionanymeasurementsofbloodgaspartial pressures(neithertranscutaneousnorarterial).Third,periodsof apneaarenotfollowedbycompensatoryyawningafterbreathing isresumed (reviewed inBaenninger,1997).The measurements, recordedbyoneofususingdigitaloxymetricmonitoring,showa light drop of oxygen saturation immediately after a yawn (unpublisheddata).Theconclusionseemstobethatﬂuctuations inpartialpressureofoxygenorcarbondioxidedonotseemtobea drivingforce for yawningin humans. Still, tofullyexclude the respiratoryhypothesis,thestudybyProvineetal.(1987)should probablyberepeatedwithmeasurementsofO2andCO2partial pressures.

4.2.Communicationhypothesis

Thesecondso-called communicationhypothesisisaboutthe purpose of yawning as a communicative tool that serves to synchronizethebehaviorofagroup(Daquinetal.,2001).Oneof theconclusionsinananalysisofanovelbyJean-PaulSartreabout profoundboredomgetsittothepoint:“theyawnisaholewitha difference[...]itdoesnotinviteﬁlling[...]butitisfulland expressive” (Bell, 1980). It is generally accepted that yawning signalizes sleepiness.This association is obviousand has been proven as yawning occurs more frequently in mornings and evenings during transitions between activity and sleep (Greco et al.,1993; Baenninger, 1997).Furthermore, it is presumed to signalize lack of interest/boredom, as increased yawning rates wereobservedinstudentssittinginaclass,inpeopledrivingacar orduringstudying(whichisgenerallybutnotnecessarilybelieved to be boring; Greco et al., 1993). In addition, more frequent yawning is associated with viewing uninteresting, repetitive stimuli than with viewing interesting stimuli (Provine and Hamernik,1986).Yawningasasignalforstressorthreat/conﬂict is more established in the animal kingdom (reviewed in Baenninger, 1997). In fact, increased yawning frequency as a signalforstressinhumanswasreportedonlyonceinasurveyof students(Grecoetal.,1993).

4.2.1.Contagiousyawning

Apartfromyawningwithoutobviousexternaltrigger,itcanbe elicitedbyobservingorlisteningotherindividualsyawnandby readingorthinkingaboutit.Theurgetoyawnisoftenirresistible and only partially suppressible and has hence been termed contagious.Itsneuroanatomicalsubstratehasbeenattemptedto beunraveledwithfunctionalmagneticresonanceimaging(fMRI) inhumans.DifferencesinfMRI-activatedbrainareaswerefound among experimental studies. According to the predominantly activatedbrainregion–andresultsofnon-fMRIstudies–different theorieshavebeenproposed. First,contagious yawningmaybe mediatedbycerebralmechanismsinvolvedinempathyandself- processing because brains of probands observing other people yawn versuslaugh were selectively activated in areas that are believedtobeinvolvedinprocessingofempathy(theoryofmind)

and information about oneself such as theposterior cingulate, precuneus,bilateralthalamus,andparahippocampalgyrus(Platek etal.,2005).Thesamegroupalsoshowedthattheurgetoyawnin probandsobservingothers yawnwasamongst otherspositively correlatedwithscoresofpenciltestsfor empathy(Plateketal., 2003).Supportforthistheorycamefromanothergroup,showing that the ventromedial prefrontal cortex (vmPFC) is selectively activatedinfMRIscansofbrainsofpeopleobservingyawnversus cough/gapeconditions(Nahabetal.,2009).ThevmPFCwasshown tobeinvolvedin empathicprocessing(Eslinger,1998; Shamay- Tsooryetal.,2003),butalsoinweighingorbiasingfuturechoices andminimizingdecision-makingtime(Becharaetal.,1999,2000;

Fellows and Farah,2007).A differentmechanistichypothesis is thatcontagiousyawningismediatedbyamechanismcalledaction understandingorimitation.Thistheorywasputforwardbyastudy that compared yawning and intransitive orofacial movement conditions.Here,selectivefMRIactivationwasseeninthesuperior temporalsulcus(STS;Schürmannetal.,2005).NeuronsintheSTS dischargewhenanindividualobservescertainmovementsbutnot when theindividualperforms thecorresponding motoractions him-/herself. As one characteristic of mirror neurons is to dischargeduringperceptionofamotoractionbut–astheyare motor neurons– alsoduringperformance ofthat actionbythe individual him-/herself and because STS neurons lack motor neuronproperties,theyareconsideredtobestrictlyrelatedtothe mirrorneuronsystembutnotpartofit(RizzolattiandCraighero, 2004).TheSTSisbelievedtobetheareawheretheactionmadeby anotherindividualiscomparedwithsensory-motorconsequences ofthesameactionmadebyoneselfinordertocopyanewactionor toadjustanactionpresentinone’smotorrepertoiretoadifferent action(Iacobonietal.,2001).Therefore,theuniqueSTS-activation inSchürmann’sstudymightreﬂecta[pattern]recognitionprocess of the yawning versus orofacial movement conditions. Taken together,theabove-mentionedfMRIstudieshaveincommonthat their yawning study conditions were based onseeing ﬁlms or picturesofothersubjectsandevokeduniqueactivationinbrain areasthatdonotbelongtothemirrorneuronsystem.Therefore, theauthorsunivocallyconcludedthatcontagiousyawningisnot relatedtoimitationthatismediatedbythemirrorneuronsystem.

Theyrathersuggestthecorticalreleaseofastereotypicalmotor pattern(inparticularSchürmannetal.,2005;Nahabetal.,2009).

Incontrast,investigationofauditorycontagiousyawningbyfMRI showedthatyawnsoundsnotonlywereeffectiveatelicitingan urgetoyawnbutyawnsoundswithhighurgetoyawnsigniﬁcantly activated essentialparts ofthemirror neuronsystem including rightposteriorinferiorfrontalgyrus(pIFG)andposteriorsuperior temporalgyrus(STS;Arnottetal.,2009).Howcanthiscontroversy about involvement of the mirror neuron system in contagious yawningbeexplained?First,thestimulusmaymakeadifference, asprocessingofsoundtakesplaceinadifferentcorticalareathan processingvisualstimuli.However,whenitcomestorecognition of a stereotypical pattern, acoustically and visually processed information should converge. Here, the mirror neuron system offers the theoretical advantage of containing so called visuo- motor and echo-neurons; cells that discharge upon seeing or hearingaparticularaction(RizzolattiandCraighero,2004)which makesthem inprinciplesuitable torecognize differentsensory informationofoneaction.Ontheotherhand,fMRIactivationin classical mirror neuron areas including the posterior inferior frontalgyruscanbedetectedinthestudiesbySchürmannetal.

(2005)andNahabetal.(2009)especiallyonsecondexaminationof thoseﬁguresinwhichyawningwascomparedtoablankscreenor white cross on grey screen respectively, and not to orofacial movements or gape/coughconditions. Eventhestudyof Platek etal.(2005)showedﬁgureswithfMRIactivationinthevicinityof the superior/inferior parietal gyrus – being part of the mirror

(10)

neuron network – if yawning was compared to neutral face conditions and not to laugh. Finally, the studies about visual contagious yawning did not grade the urge to yawn of their probandsin response to visual stimuli (exept for Nahab et al.

whoseprobandsansweredabinaryquestionnairewhetherthey yawned while seeingvisual stimuli and whether they yawned more when seeing repeated stimuli). Therefore correlations of subjectivehigh-urge-to-yawnfeelingswithfMRIactivitypatterns werenotreported.Insummary,themirrorneuronnetworkcan stillbeinvolvedintherecognitionandprocessingofcontagious yawning.The question remains whethercontagiousyawning is based on imitation, action understanding, empathy, or some

“inborn” capacity of the brain. One relative argument against imitationis anobservation byMoore (1942) that blindpeople yawninresponsetoanaudiorecordingofyawns.Howcanblind peopleperformtheactofyawninginacorrectandrecognizable way,iftheyhaven’teverobserveditandifitisnotamotorpattern belonging to the basic repertoire of the brain? It is therefore suggested that the efferent partof contagious yawning is not elicitedbycorticalmirrorneuronmotorareas.Ratherthehighly stereotypicalexpressionisrecognizedbythecortex(mightitbe STS, pIFG or vmPFC) and might activate intermediate control centersincludinge.g.theinsula,hippocampus,andthePVNwhich then induce execution of the motor program by brainstem structuresthatcoordinateinnervationoffacial,laryngeal,pharyn- gealstructures,thediaphragmandaccessoryrespiratorymuscles.

Asmentioned,contagiousyawningdoesnotnecessarilydependon visualoracousticinputonly.Thinkingorreadingaboutyawning canbeenoughtotriggertheyawnact(BaenningerandGreco,1991;

Greco et al., 1993; Provine, 1986). The mirror neuron network wouldbehandytoexplainthissortofobservationinitsfunctionas pattern recognition system. The theory of speech evolution is basedonatransferofgesturalmeaningtoabstractsoundmeaning andontheexistenceofacommonneuralsubstrateforhand/arm andspeechgestures,aswellasforsoundandlanguageprocessing (RizzolattiandCraighero,2004).Itwasshownthatduringreading and spontaneous speech without further motor action, mirror neuronareasincludingthehandmotorcortexwereactivatedas well(Meisteretal.,2003).Therefore, readingorthinkingabout yawningmightactivatebrainareasthatstarted evolutionaryas motorneuronsandhavebecomespecializedtorecognize,thatisto dischargeto,thewrittenperceptionofanactionoritsretrieval frommemorybut maynotnecessarily contributetothemotor performanceof thatactionanymore.Theconclusionisthatthe communicationhypothesiscanonlyexplainapartofthebehavior bygreatapes,perhapsparrotsandotherbirds,andmammals,but notbyreptilians.Thesauropsidsincludereptiles(poikilotherm) and birds (homeotherm).But, in fact, relative brainsize varies greatlyamongsauropsids,withturtlesandsnakesoccupyingthe lowendoftherange.Allreptilespresentatrilaminarcortexina largedorsalventricularridgethatprotrudesintotheventricle.Itis difﬁculttohomologizetoanythinginmammalianbrains.These differencesin cerebralarchitecture probablyexplainwhysocial processessuchasunconsciousmimicrylikecontagiousyawning cannotbeobservedbyreptilesortortoises(Wilkinsonetal.,2011).

4.3.Arousalhypothesis

The third major hypothesis favors arousalas theprominent biological purpose of yawning. Supporters of this hypothesis includeRonaldBaenningerwhorecapitulatedthat“thethreadused to tie all the diverse data and observations together is the basic hypothesisthatamajorfunctionofyawningistoregulatelevelsof arousal”(Baenninger,1997).Forhisconclusion,healsorevertedto owndatasuchasthestudyshowingsigniﬁcantlyincreasedwrist activity"asmeasuredwithwristmonitors"afteryawningasan

indicator for elevated arousal (Baenninger et al., 1996). While analyzing the biochemistry of yawning, the constellation of neurotransmitters favoring yawning such as serotonin and dopaminesuggestedtoJAskenasy(1989),anotherprotagonistof the arousal hypothesis, that yawning has an “antisleep effect”.

OlivierWalusinskisuggestedthattheyawning-stretchsyndrome maybeanafferentproprioceptivefeedbackcontributingtoone’s cerebralbodyscheme.Buthealsoadvocateditscontributionto arousalbytriggeringdissolutionofthecerebralnetworkcontrol- ling sleep – and especially REM sleep – and facilitating the establishment of a network that controls the wake status (Walusinski, 2006). He adds recently to his theory a more complete explanation by involving the brain network that is functional during the resting state, that is, the default mode network.Whenthisnetworkisactive,yawningmanifestsaprocess of switching to the attentional system through its capacity to increasecirculationofcerebrospinalfluid(CSF),therebyincreasing clearanceofsomnogenicfactorsaccumulatinginthecerebrospinal fluid(Walusinski, 2014). A further pro-arousal argument came fromSato-Suzukiandco-workerswhoshowedthatcorticalEEG frequency – recorded experimentally with 2 screws from the vertexofanesthetizedrats–increasedpriortoyawnselicitedby electrical orchemical stimulationof thePVN and subsequently reversed toslowEEGactivityafteryawning(Sato-Suzukiet al., 1998).ThesebehavioralandEEGphenomenacouldberepeated withdifferentstimulisuchasorexins(Sato-Suzukietal.,2002)and histamine(Sekietal.,2002).Ontheotherhand,Guggisbergandco- workers showed in 16 patients who suffered from daytime sleepiness that yawning is triggered by states of low vigilance but doesnot leadtoan arousingeffectwhen compared tothe performanceofvoluntaryisolatedbodymovementsbythesame groupofpatients.Indetail,yawningwasprecededbysignificantly greaterdeltaactivityinEEGthanvoluntarymovements,indicating drowsinessbeforetheyawnevent.Afteryawning,powerinthe deltafrequencyrangewasstill significantlygreaterthanaftera voluntary movement. In addition, mean alpha frequency over central regions increased after isolated movements while it decreased after yawningindicating arousal rather byvoluntary body movements (Guggisberg et al., 2007). Other studies investigatingtheassociationofyawningwitharousalinhumans withEEG,skinconductance,ormeasurementsofotherautonomic parameters foundcontroversialresults(Guggisberget al.,2010, 2011). In conclusion, yawning correlates with low states of vigilance and is associated with transitions betweenwake and sleep,periodsofexposuretouninterestingrepetitivestimuli, or withparticularstressfulevents.Thefailuretoassociateyawning withimproved vigilance orincreasedautonomic tonedoes not precludeyawningtobeaclinicalsignforparticularactivitystates ofthosebrainregionsthatareinvolvedinregulationorexecution ofyawning.Ifthiscanbedemonstratedinfutureexperiments,the arousal hypothesis may be rephrased as focal brain activity hypothesis.

4.4.Brain-coolinghypothesis

One research group promoted brain cooling as a biological functionofyawning.Dataofarousalandautonomicactivationin humans(Coreyetal.,2012)andtransientbraintemperaturepeaks aroundboutsofyawninginrats(Shoup-Knoxetal.,2010)were presentedasevidenceforthishypothesis.Abnormalyawningwas suggestedtobeindicativeforthermoregulatorydysfunction,i.e.

disorders temporarilyassociated with abnormal yawning were associatedwithdifﬁcultiesinbody temperaturecontrol(Gallup andGallup, 2008;Prasad,2008;Gallup,2014).Itcanbeagreed uponthat(physiological)yawningandtongueprotrusioncanhelp inthecontrolofbody temperatureinspeciessuchasfelidaeor

(11)

canines,whichdonotexchangeheataseffectivelyviabodysurface andperspirationasdohumansbecauseoftheirfur.Inhumans, heatexchangeislikelytobemoreeffectiveatthebodysurface, promoted by perspiration, than in lungs due to their surface difference.Fromaneuroanatomicalpointofviewthehypothala- musisbelieved tobethecenter forthermoregulation, and the sympathetic nervous system, which controls amongst others sweatglandsinskin,hasoriginsinthehypothalamus,aswell.In addition,hypothalamicnucleiarebelievedbeonesuperordinate controlsystemoftheyawningmotorpattern.Therefore,werather suggest that neurological diseases or neurotropic medication, affecting input/output or function of the hypothalamus itself, mightconjointlyaffectthermoregulationandyawningbehavior.

This does not necessarily mean that these 2 mechanisms are causallyrelatedandoneisperformedtoservetheother.Oneneeds to take into consideration that a physiologist, Hannu Elo, has shownbycalculationsthatthetemperaturedecreasesclaimedto occurduringyawningarephysicallyimpossible(Elo,2010).

5.Abnormalyawningassociatedwithdiseasesormedication Yawningmayoccurnotonlybecauseofboredom,drowsiness, orbycontagionbutalsoinassociationwithvariouspathologies.

Wetriedtogivearepresentativereviewofdiseasesassociatedwith noticeable yawning, but may not have covered all aspects.

Abnormalyawningwastermedpathologicalorexcessiveinonly afewreports.Foritsdeﬁnitionincludingfrequency,thereaderis referredtotheintroduction.Severalreportsofnoticeableyawning couldsometimesbebettergroupedaccordingtoaneurological syndrome of various etiologies. Otherwise, reports were listed accordingtotherespectiveneurologicaldisease.

5.1.Yawningduringandaftercerebrovascularinsult 5.1.1.Anteriorcirculationstroke

Anterior circulation stroke with unilateral, supratentorial lesions can be associated with abnormal yawning. In the ﬁrst report,3ormoreyawns per15minwereobservedineachof7 patientswhopresentedwithacutemiddlecerebralarterystroke andsignsofcorticaldysfunctionsuchasaphasia,neglectandgaze

palsy.Symptomonsetwaswithin12h,averageNIHSSscorewas17 and1/3ormoreoftheMCAterritorywasaffected.Theauthors’

conclusion was that supratentorial lesions may release the hypothalamicPVNfromneocorticalcontrolmechanisms,thereby increasing its activity and leading to excessive yawning. As temporallobestructureswerepartiallydamagedintheirpatients, theyspeculatedthathippocampal/periamygdalar–hypothalamic connections, shown before to exist, might have been affected (Singeretal.,2007).Thesecondreportwasaboutapatientwith leftinternalcapsulestrokeandright-sidedweakness(Upperand lowerlimb0/5and2/5respectively),whocouldnotmovehisarm voluntarilybut was able todo soduringyawning. Theauthors speculatedthatintactnon-pyramidalprojectionsfromaputative, non-lesioned yawning centerin thebrainstem might sharethe common lower motor neuron pathwayto innervate brainstem nucleiandspinalcordanteriorhorncells.It wasnotmentioned whetherthe patientexhibitedexcessive yawning(Wimalaratna andCapildeo,1988).Involuntaryyawning-associatedmovements in paralyzed limbsweretermed parakinesiabrachialis oscitans (Walusinskietal.,2005).Theywerealsoobservedinalargecase seriesofpatientswithacuteanteriorcirculationstroke.Interest- ingly,theseyawning-associatedmovementsdonotseemtohave any prognostic value (Meenakshisundaram et al., 2010). We publishedaseriesof10patientswithacuteanteriorcirculation strokeand!3yawns/15minwithoutobviouscause.Weshowed that thestrongestlesionoverlapwas intheinsula andcaudate nucleus(Fig.4)andthatdurationofabnormalyawningcorrelated withclinicalstrokeseverity (NIHSS)and withneuroradiological strokeintensity(inverseapparentdiffusioncoefﬁcientcorrelation) inthestrategicallylesionedbrainareas(Kresteletal.,2015).Based on the insula’s intense brainstem connections via the cortico- striato-thalamic network and corticobulbar pathways, we sug- gestedfollowingmechanismsuponitslesioning:i)adisconnection syndromeofinsulartargetsincludingtheentorhinalcortex,lateral hypothalamus, ormono-/oligosynaptictrajectoriestotheRaphe nucleus/nucleus tractussolitarius with(GABA-ergic?) disinhibi- tion of the nearby pre-Bötzinger complex respiratory rhythm generator and cranial nerve nuclei V, VII, IX, X, and XII. ii) A serotonin overspill theory upon partial insular lesioning with activation of the brainstem yawning pattern in analogy to

Fig.4.Hotspotsofabnormalyawning.Areasofmostintenselesionoverlap(purple),depictedincoronalT1-weightedMRIsections,andoriginatingfrom10patientswith anteriorcirculationstrokeassociatedwithabnormalyawning.Forinter-individualcomparisons,imagesofpatientswithleft-sidedlesionswereﬂippedtothecontralateral hemisphere.Group-speciﬁclesionoverlayplotswerecreatedusingMRIcroN.Fortechnicaldetails,seeKresteletal.(2015).Unpublisheddata.