Breathing as a Fundamental Rhythm

of Brain Function
Ongoing fluctuations of neuronal activity have long been considered intrinsic noise that introduces
unavoidable and unwanted variability into neuronal processing, which the brain eliminates by averaging
across population activity (Georgopoulos et al., 1986; Lee et al., 1988; Shadlen and Newsome, 1994;
Maynard et al., 1999). It is now understood, that the seemingly random fluctuations of cortical activity form
highly structured patterns, including oscillations at various frequencies, that modulate evoked neuronal
responses (Arieli et al., 1996; Poulet and Petersen, 2008; He, 2013) and affect sensory perception
(Linkenkaer-Hansen et al., 2004; Boly et al., 2007; Sadaghiani et al., 2009; Vinnik et al., 2012; Palva et al.,
2013). Ongoing cortical activity is driven by proprioceptive and interoceptive inputs. In addition, it is partially
intrinsically generated in which case it may be related to mental processes (Fox and Raichle, 2007; Deco et
al., 2011). Here we argue that respiration, via multiple sensory pathways, contributes a rhythmic component
to the ongoing cortical activity. We suggest that this rhythmic activity modulates the temporal organization
of cortical neurodynamics, thereby linking higher cortical functions to the process of breathing. Detlef H.
Heck1*, Samuel S. McAfee1, Yu Liu1, Abbas Babajani-Feremi1,2, Roozbeh Rezaie2, Walter J. Freeman3, James W. Wheless2,
Andrew C. Papanicolaou1,2, Miklós Ruszinkó4, Yury Sokolov5 and Robert Kozma5,6 1Department of Anatomy and Neurobiology,
University of Tennessee Health Science Center, Memphis, TN, USA, 2Department of Pediatrics, Division of Pediatric Neurology, University of Tennessee
Health Science Center and Le Bonheur Children’s Hospital Neuroscience Institute, Memphis, TN, USA, 3Department of Molecular and Cell Biology, Division of
Neurobiology, University of California at Berkeley, Berkeley, CA, USA, 4Rényi Institute of Mathematics, Hungarian Academy of Sciences, Budapest, Hungary,
5Department of Mathematical Sciences, University of Memphis, Memphis, TN, USA, 6Department Computer Sciences, University of Massachusetts Amherst,

Amherst, MA, USA Edited by: Alexey Semyanov, University of Nizhny Novgorod, Russia Reviewed by: Beate Rassler, Leipzig University, Germany Leslie M.
Kay, University of Chicago, USA *Correspondence: Detlef H. Heck dheck@uthsc.edu Received: 19 September 2016 Accepted: 26 December 2016 Published:
12 January 2017 Citation: Heck DH, McAfee SS, Liu Y, Babajani-Feremi A, Rezaie R, Freeman WJ, Wheless JW, Papanicolaou AC, Ruszinkó M, Sokolov Y and
Kozma R (2017) Breathing as a Fundamental Rhythm of Brain Function. Front. Neural Circuits 10:115. doi: 10.3389/fncir.2016.00115 Keywords: mind-body,
cortical oscillations, respiration, embodied cognition, phase transitions, phase amplitude coupling, proprioception, graph theory We
have
recently shown that respiration-locked olfactory bulb activity in awake, head restrained
mice causes respiration-locked delta oscillations and gamma power modulations in the
somatosensory cortex (Ito et al., 2014). This unexpected direct ability of respiration-locked
sensory activity to modulate oscillatory neuronal activity in the neocortex led us to consider
the potential wider implications of a link between breathing and brain activity, particularly
with respect to the possibility that respiration influences cortical neuronal activity
underlying cognitive function. Based on our own experimental findings, results from our
modeling studies using a simple graph theory model and a review of the literature, we
argue that respiration, via multiple sensory pathways, provides a subtle but continuous
rhythmic modulation of cortical neuronal activity that modulates sensory, motor, emotional
and cognitive processes. Specifically, we hypothesize that: (1) respiration causes
respiration-locked oscillations that are synchronized across large areas of neocortex at the
species-specific respiratory rhythm; (2) that increases in the power of gamma oscillations
(40–100 Hz) occur preferably during certain phases of (i.e., are phase-locked to) the
respiratory cycle. Both hypotheses are supported by solid experimental results in the
somatosensory Frontiers in Neural Circuits | www.frontiersin.org 1 January 2017 | Volume 10 | Article 115Heck etal.
BreathingasaFundamentalRhythmofBrainFunction barrel cortexinawakemice(Ito etal., 2014; Figure1) andby our
modelingstudies(seebelow).Additionalresults,published in
abstractform,supportthepossibilitythatrespiration-locked oscillations
arealsopresentinseveralotherareasofmouse neocortex (Liu etal., 2015),
includingthevisualcortex(McAfee et al., 2016). A
thirdpredictionissupportedbypreliminaryexperimental findings
onsuddenchangesinsynchronizationpatternsof neural
activitypublishedinabstractform(Kozma etal., 2015). We
predictthat(3)thetimingofthesuddenchangesin the
networkactivity,i.e.,fasttransitionsbetweensynchronized and de-
synchronizednetworkstates,arephase-lockedtothe respiratory
rhythm.Thesetransitionscanbedetectedasjumps in
theanalyticphaseofoscillatorypopulationactivityusing a
Hilberttransformbasedanalysisoflocalfieldpotential (LFP)
orelectroencephalographic(EEG)activity(Freeman and Rogers, 2002; Freeman etal., 2006;
Freeman, 2015). Graph theoreticalargumentsprovideamodelingframework to
describetheexperimentallyobservedsuddenchanges as ‘‘phasetransitions’’(Puljic andKozma,
2008; Kozma and Puljic, 2015), atermwewillusethroughoutthis article. As
welayoutinmoredetailbelow,gammaoscillations are
formsofcorticalactivitywidelylinkedtocognitiveand other
highercorticalfunctions.Ourhypothesespredictthat a
consciouslycontrolledchangeinrespiratorybehaviorwill cause
achangeincognitiveandemotionalstates,whichisa common observationinyogicbreathing(Jella
andShannahoff- Khalsa, 1993; Stancák andKuna, 1994; Brown andGerbarg, 2005)
andstressreducingrespiratoryexercisessuchascombat tactical
breathingemployedbymilitaryandspecialforces (Grossman andChristensen, 2011).
Asecondkeyprediction is thatrespiration-lockedmodulationofcorticalgamma activity
andphasetransitiontimingdirectlylinksrespiratory behavior
tohighercorticalprocesses,includingcognitiveand limbic
functions,sensoryperceptionandmotorcontrol.The respiration-locked
modulationofneocorticalactivitywepropose here
wouldthusprovideaneuronalmechanismandcausallink between
respirationandpainperception(Arsenault etal., 2013; Iwabe etal., 2014), motorcontrol(Ebert
etal., 2002; Rassler and Raabe, 2003; Li andLaskin, 2006; Iwamoto etal., 2010; Cao et al.,
2012; Krupnik etal., 2015), attention(Gallego etal., 1991; Krupnik etal., 2015)
andemotion(Benson etal., 1974; Arch and Craske, 2006; Homma andMasaoka, 2008). FIGURE 1|
From Ito etal. (2014): respiratorymodulationofthepowerofgammafrequencyoscillationsinmousewhiskerbarrelcortex. Phase–amplitude
couplingbetweenrespiration-lockeddeltaandgammabandoscillationsinthebarrelcorticallocalfieldpotential(LFP)activityofanawakeintact and
anawakebulbectomizedmouse,followedbypopulationstatistics. (A) Respiratory activity(toptrace),amplitudeofgammabandoscillations(middletrace)and delta
oscillations(lightgreenbottomtrace)anditsphase(darkgreenbottomtrace)inanintactmouse.Gammaoscillation(75Hz)amplitudepeaksrhythmically phase
lockedtothedeltacycle. (B) Gamma oscillationamplitudeasafunctionofdeltaphase(red).Thesolidanddottedblacklinesindicatethemeanandthe 2.5
and97.5percentileboundariesofthesurrogateamplitudedistributionestimatedfrom1000phase-randomizedsurrogates.Gammaamplitudemodulationis
significant atphase0ofthedeltacycle. (C,D) Same as (A,B), respectively,butforabulbectomizedmouse.Afterremovaloftheolfactorybulb,theamplitude
modulation ofthegammabandoscillationsisnolongerphaselockedtorespiration. FrontiersinNeuralCircuits | www.frontiersin.org 2 January 2017|Volume10|
Article115Heck etal. BreathingasaFundamentalRhythmofBrainFunction Oscillations
ofneocorticalactivityinthegamma(30–
100Hz) frequency range,havebeenstronglyimplicatedinaffectiveand cognitive
brainfunctionssuchasattention(Fries etal., 2001; Laufs etal., 2003; Tallon-Baudry, 2004),
sensoryperception (Engel etal., 2001; Tallon-Baudry, 2003; Gould etal., 2012), decision
making(Kay andBeshel, 2010; Siegel etal., 2011; Gould et al., 2012; van Vugtetal., 2012;
Wyart etal., 2012; Nácher etal., 2013), problemsolving(Sheth etal., 2009), memoryformation
(Marshall etal., 2006; Tort etal., 2009; Chauvette etal., 2012) and
languageprocessing(Crone etal., 2001; Towle etal., 2008; Babajani-Feremi etal., 2014).
Sudden changesinnetworksynchronizationarecharacteristic features
ofcorticalactivitythathavebeenwidelylinkedto cognitive processes(Kozma andFreeman,
2016). Detailed analysis ofrabbitandhumanintracranialelectrocorticography (ECoG)
signalsrevealeddiscontinuitiesintheanalyticphase determined byHilbertanalysis(Freeman
andRogers, 2002; Freeman etal., 2006; Freeman, 2015). Experimentswithrabbits trained
withaclassicalconditioningparadigmshowedthat discontinuities
oftheanalyticphasehavecognitiverelevance (Freeman, 2004; Kozma andFreeman, 2008).
Namely,after delivering theconditionedstimulus,theoccurrenceofthe phase
discontinuitycorrelateswiththestimulus,suggesting that
thesediscontinuitiescanbeviewedasmarkersofthe cognitive
activity(stimulusclassification)performedbythe rabbits. Schölvinck etal. (2015) observed
thatvariabilityofneuronal responses intheprimaryvisualcortextorepeatedidentical stimuli
wascausedbylargescalenetworkactivity,which was
morevariablewhenthenetworkwasinasynchronized state vs.anasynchronousstate.Recently,
Tan etal. (2014) also showedthatvisualstimulationshiftedtheactivitystates of
themacaqueprimaryvisualcortexfromsynchronousto asynchronous
activity.Thesefindingsarefundamentallyinline with
ourhypothesisthatthetimingofsuchphasetransitionsis linked
totherhythmicsensorystimulationcausedbyrespiration. We
haveobtainedpreliminarysupportingevidenceforphase- locking
betweenrespirationandphasediscontinuitiesinhuman cortical
activityfromananalysisofECoGsignalsfromahuman subject.
Weinterpretedtheresultsasphasetransitionsincortical population
activitybetweensynchronizedandde-synchronized states; see Kozma etal. (2015). A
smallgroupofresearchershaveenvisionedthepossibilityof respiration influencinglarge-
scalebrainactivityviatheolfactory system.
Freemanandcolleaguesperformedpioneeringstudieson the
influenceofrespirationthrougholfactionontheratbrain (Eeckman andFreeman, 1990; Kay
andFreeman, 1998). Effects of theta-modulationofsaccadicsignalshavebeendescribed as
visualsniffing(Kozma andFreeman, 2001). Fontanini and Bower (2006) speculated
thatolfactorybulbrespiration- locked oscillationsinrodentsmaypropagatethroughtheentire
cortex. However,noneoftheseearlierstudiesanticipatedthat respiration
couldmodulatethepowerofgammaoscillations or
consideredarespiratoryinfluenceonthetimingofphase transitions
incorticalpopulationactivityasamechanism that
directlylinksrespiratorybehaviorandcognitivebrain processes. Respiration
createsbothconsciousandunconsciousstreams of
rhythmicsensoryinputstothebrain.Consciouslyaccessible sensations
ofnormal,unobstructedbreathingincludeodor perception,
themechanicalandthermalsensationofairflowing through
nose,mouthandupperairways,andtheproprioception of
movementsofthechestandabdomen.Unconscioussensory signals
causedbyrespirationincludeinteroceptivesignalsfrom the
lungs,diaphragmandinternalorgans,whichrepresent the
mechanicalconsequencesofrespiratorymovements,and the
chemosensitivesignalsfromthecardiovascularsystem, which representbreath-by-
breathfluctuationsofCO2 and O2 levels intheblood.Thesensationsandbrainactivity patterns
associatedwithhunger-for-air(Liotti etal., 2001; Macey etal., 2005)
arenotconsideredhere,astheyrepresent an
emergencyresponsenotrelatedtonormal,unobstructed breathing. There
arealsoanumberofindirectwayscorticalareas receive respiration-
lockedsensoryinput.Eyemovements,for example, havebeenshowntobetransientlyphase-
lockedto respiration duringsleep(Rittweger andPöpel, 1998) aswellas in
theawakestate(Rassler andRaabe, 2003). Recently, Ito etal. (2013) reported
saccaderelatedchangesinthepowerofneuronal oscillatory
activityinfourfrequencybands,includinggamma, in
primatesthatwerefreelyviewingtheirenvironment.This suggests
thattheretinalflowassociatedwitheyemovements causes
amodulationofpowerinvisualcorticaloscillations that
ispartiallycorrelatedwithrespiration.Anotherindirect respiration-locked
sensoryinputcomesfromtheauditorycortex, which
receivesrhythmicauditoryinputrelatedtorespiration caused
bythesoundofairflowingthroughthenoseormouth. Finally,
neuronsinthebrainstemprojectbroadlytothalamic nuclei (Carstens etal., 1990; Krout etal.,
2002). Theseprojections likely providerespiration-lockedinputtothethalamus(Chen et al.,
1992), introducinganon-sensoryrespiratoryrhythmtothe thalamo-cortical network. While
therearemanysourcesofrespiration-lockedactivity, the
olfactorysystemdeservesspecialattention,becauseearly mammals
reliedstronglyontheirolfactorysenseandhad proportionately largeolfactorybulbs(Rowe etal.,
2011). Furthermore, neuronaloscillations,particularlygamma oscillations,
areauniversalelementofodorprocessingin animals
asfarremovedfromjointevolutionaryancestorsas mammals andinsectsare(Kay, 2015).
Eventhoughinprimates the olfactorysenselosttheprimeimportanceithasformost other
mammalsinfavorofvision(Gilad etal., 2004). EEG studies
comparingnasalandoralbreathingofroomairfound that
nasalbreathingelicitedsignificantlydifferentpatternsof EEG
activitythanmouthbreathing(Servít etal., 1977; Lorig et al., 1988).
Thisisinlinewithourfindingsofnasalairflowin mice
drivingdeltaoscillationsandgammapowermodulationsin a non-olfactoryareaofneocortex(Ito
etal., 2014) andsuggests that theolfactorybulbactivationexertssimilarinfluenceon human
corticalactivity. The detectionandanalysisofrespirationlockedcortical activity
requiresthesimultaneousmeasurementofrespiration and
brainactivity.Suchsimultaneousmeasurementsarenot commonly
performed.Anotableexceptionisarecentstudyof FrontiersinNeuralCircuits | www.frontiersin.org 3 January 2017|Volume10|
Article115Heck etal. BreathingasaFundamentalRhythmofBrainFunction the
effectsofsleepdisorderedbreathing(SDB)inchildrenon cortical oscillatoryactivity(Immanuel
etal., 2014). Immanuel et al. (2014) showed thattheaveragepoweroftheEEGsignal decreased
duringinspirationandincreasedduringexpiration, in
afrequencybandandsleepstagedependentmanner,inboth healthy
subjectsandsubjectssufferingfromSDB.Thisstudydid, however, notevaluatephase-
lockingbetweenEEGoscillations and therespiratorycycle. Respiration
relatedsensoryactivityduringunobstructed breathing mainlyreachesthreeareasofthecortex:
(1)the olfactory cortexandsurroundingareasreceiveolfactorybulb input;
(2)thesomatosensorycortexreceivesinputsfrom mechanoreceptors
ofchest,theabdominalskinandmusclesthat are
stretchedandmovedbyrespiration;and(3)theinsularcortex receives
inputfromchemoreceptorsandmechanoreceptorsin the
lungs,diaphragmandinternalorgans.Ourrecordingsof olfactory bulbdependentrespiration-
lockedoscillationsinthe mouse somatosensorycortexsuggestthatrespiration-locked activity
propagatesfromprimarysensoryareastopartsofthe cortex
thatdonotreceivedirectrespirationrelatedsensory inputs.
Alikelymodeofpropagationisthroughthecortico- cortical
networkitself,possiblyinvolvingalsocortico-thalamic connections.
However,theanatomyofaxonalconnections within
theparabulbarandlimbicareassuggestanumberof subcortical
regionsandneuromodulatorsystemsmayalsobe influenced byrespiration-
drivensensoryinput.Forexample, widely projectingserotonergicandcholinergicneuronswithin
the ratbasalforebrainhavebeenshowntorhythmically discharge
inphasewithrespiration(Manns etal., 2003; Mason etal., 2007), witholfactorybulbrespiration-
locked activity asalikelydrivingforce(Linster andHasselmo, 2000). Stimulation
ofcholinergicneuronsinparticularisassociated with
increasedneocorticalgammaoscillations(Cape andJones, 2000)
amechanismthatmightcontributetotherespiration- locked
modulationofgammapowerinmousesomatosensory whisker barrelcortex(Ito etal., 2014).
However,aswe argue below,respiration-lockedgammapowermodulation may
resultfromintrinsicpropertiesofthecorticalnetwork itself. The linkbetweenrespiration-
lockedcorticaloscillations and respiration-relatedsensoryinputstothecortex is
straightforward:respiration-lockedrhythmicinputs drive
corticalneuronstofirerhythmicallyatthesame frequency.
Experimentsinanesthetizedrodentsshowthat respiration-locked
oscillationsinthepiriformcortexaredriven this waybyrespiration-
lockedactivityofolfactorybulbafferents (Fontanini andBower, 2005; Uchida etal., 2014),
whichalso drive respiration-lockedactivityinthehippocampusofmice, both
underanesthesia(Yanovsky etal., 2014) andwhileawake and walkingonatreadmill(Nguyen
Chietal., 2016). However, the mechanismsbehindrespiration-lockedmodulationsof gamma
power,whichweobservedinthemousesomatosensory cortex (Figure1), arelessobvious. To
investigatetheprocessesleadingtorespirationlocked increases
inthepowerofgammaoscillationsweuseda simple
graphtheorymodelinspiredbycorticalnetwork architecture,
withabiologicallyappropriatebalanceofexcitatory and inhibitoryneuronsandmixofshort-
andlong-range connections. Expandingonpreviouswork(Reijneveld etal., 2007; Turova
andVilla, 2007; Gallos etal., 2012; Janson etal.,