MedBiolEngComput(2014)52:467473

DOI10.1007/s1151701411501

OriginalArticle

EffectofrespirationonKorotkoffsoundsandoscillometriccuff
pressurepulsesduringbloodpressuremeasurement
DingchangZhengLuigiYuriDiMarcoAlanMurray

Received:2April2013/Accepted:11March2014/Publishedonline:26March2014The
Author(s)2014.ThisarticleispublishedwithopenaccessatSpringerlink.com

Abstract Bloodpressure(BP)measurementaccuracy
depends on consistent changes in Korotkoff sounds
(KorS)formanualmeasurementandoscillometricpulses
forautomatedmeasurement,yetlittleisknownaboutthe
directeffectofrespirationonthesephysiologicalsignals.
Theaimofthisresearchwastoquantitativelyassessthe
modulationeffectofrespirationonKorotkoffsoundsand
oscillometric pulses. Systolic and diastolic blood
pressures were measured manually from 30 healthy
subjects(age41 12years).Threestaticcuffpressure
conditions werestudied fortwo respiratoryrates. Cuff
pressure[withoscillometricpulses(OscP)],ECG,chest
motion respiration [respiration signal (Resp), from
magnetometer] and Korotkoff sounds (KorS, from
digital stethoscope)were recordedtwice for20 s. The
physiologicaldatawereevenlyresampled.Respiratory
frequencywascalculatedfromResp(fR),OscP(fO)and
KorS(fK)frompeakspectralfrequency.Therewasno
statisticallysignificantdifferencebetweenfR andfO or
fK.Respiratorymodulationwasobservedinallsubjects.
OscP amplitude modulation changed significantly
betweenthetworespiratoryrates(p<0.05)andbetween
the three cuff pressures (p < 0.0001), and decreased
significantlywithdecreasingcuffpressure(p <0.05).
ThephaseshiftbetweenRespandmodulationofOscP
wasstatisticallysignificantwithrespiratoryrates(p <
0.05),butnotwithcuffpressures.ItisacceptedthatBP
in individuals is variable and that this relates to
respiration;wenowshowthatthisrespirationmodulates
oscillometricpulseandKorotkoffsoundamplitudesfrom
whichBPismeasured.

D.Zheng(*)L.Y.DiMarcoA.Murray
CardiovascularPhysicsandEngineeringResearch
Group,InstituteofCellularMedicine,Newcastle
University,NewcastleNE24HH,UK

email:dingchang.zheng@ncl.ac.uk

13

Keywords Bloodpressuremeasurement
OscillometricpulsesKorotkoffsoundsRespiratory
modulation

1 Introduction
Itiswellacceptedthatrespirationinfluencesbeatby
beat blood pressure (BP) changes [6]. During the
respiratorycycle,changesoccurinthecentralvenous
pressure as a consequence of chest expansion and
compression. During inspiration, the decrease in
centralvenouspressureincreasesvenousreturnand
right atrial filling and at the same time reduces
pulmonaryvenousflowtotheleftsideoftheheart,
leadingtoreducedstrokevolumeanddecreasedmean
systemic arterial pressure. During expiration, the
oppositeoccurs.
The physiological interactions during respiration
arecomplex.In1952,Dornhorstetal.[6]documented
the effect of respiration on direct BP and showed
phase differences between them. Later, Saul et al.
[13] extended these observations from invasive
recordings of the radial arterial pressure with
respiration at a constant mean respiratory rate, and
Laude et al. [9] studied these phases at different
respiratory frequencies. Many different techniques
have been used to study the phase relationship
betweentheinterbeatinterval(orinstantaneousheart
rate)andthesystolicbloodpressure(SBP)recorded
noninvasively with paced breathing. They include
frequencydomaincrossspectralanalysis[2, 14, 17,
19], windkessel model [5], time domain cross
correlation [3, 9],mutualinformationanalysis [20]
and a numeric model of the closed loop regulation
system[13].
There are currently two common noninvasive tech
niques (manual auscultatory and automated
oscillometric) to determine BPs. Manual auscultatory
techniqueisbased

\468

MedBiolEngComput(2014)52:467473

2Methods
on the auscultation of the Korotkoff sound, whereas auto
mated measurement is almost always based on the oscil
lometric pulse waveform. BP measurement accuracy there
foredependsonconsistentchangesinKorotkoffsoundsfor
manualmeasurementandoscillometricpulsesforautomated
measurement.
Althoughtheexactmechanismsunderlyingtherespiratory
influenceonBPmeasurementarenotfullyunderstood,itis
likely that oscillometric pulses recorded from cuff pressure
during BP measurement is influenced by respiration.
Moreover,ifrespirationmodulatestheoscillometricpulses,
this modulation could also affect the Korotkoff sounds
associatedwiththebloodpulseflowingthroughthebrachial
artery. One longstanding hypothesis on the genesis of
Korotkoffsoundsisthatthesharpaudibletappingsoundis
generatedbythedistensionofthearterialwallcausedbythe
changingtransmuralpressuregradient[4,10,16,18].Itmay
therefore be speculated that if respiration modulates the
pressurepulse,itmightalsoinfluencetheforcedeployedin
openingtheartery,whichinturnmayreflectontheamplitude
oftheKorotkoffsounds.
Ourpreviousstudyhasquantifiedtheclinicalimportance
oftheeffectofrespirationonBPmeasurement.Withregular
deeper breathing, both SBP and diastolic blood pressure
(DBP) changed significantly in comparison with normal
condition:decreasingby4.4and4.8mmHg,respectively[21].
A major review in the Journal of the American Medical
Association(JAMA)estimatedthata5mmHgerroreither
above or below the actual BP would result in 27 million
Americans being exposed to unnecessary treatment or 21
millionbeingdeniedtreatment[8].Therefore,anysmallBP
changes caused by respiration are clinically important and
worthfurtherinvestigation.
Since the respiratory modulation of the oscillometric
waveformandKorotkoffsoundcouldinfluencetheaccuracy
of BP measurement, it is therefore of clinical interest to
quantify the influence of respiration on both signals, and
furtherinvestigationcouldprovideadditionalinsightintoBP
measurement variability. However, to the best of our
knowledge,nodatahavebeenreportedonthedirecteffectof
respirationonoscillometricpulsesandKorotkoffsounds.
The aim of this research was to study such effects by
quantitatively assessing: (1) the presence of a modulating
effect of respiration on oscillometric pulse amplitude and
Korotkoff sound amplitude; (2) different amplitude
modulation of oscillometric pulses and Korotkoff sounds
betweendifferentrespiratoryratesandcuffpressures;(3)the
respiratory modulation of phase relationship between
respiration and changes in the oscillometric pulses and
Korotkoffsoundssignals.

2.1Subjectdata
Thirtyhealthysubjectswereenrolledinthestudy(14male,
16 female). Clinical information for those subjects is sum
marizedinTable1.Thestudywascarriedoutinaccordance
withtheDeclarationofHelsinki(1989)oftheWorldMedical
AssociationandwasapprovedbythelocallyappointedEthics
Committee.Allsubjectsgavetheirwritteninformedconsent
beforeparticipating.

2.2Experimentalprotocol
For each subject, SBP and DBP were measured manually
with the subject sitting quietly by a trained operator
according to the recommendations of the European
HypertensionSocieties[11].Physiologicalsignals,including
thecuffpressure,chestwallmovementandKorotkoffsounds
(KorS),wererecordedfor20sunderthreestaticcuffpressure
conditions:high(H)SBP10mmHg,medium
(M)(SBP+DBP)/2andlow(L)DBP+10mmHg.Theorder
ofthreecuffpressureswasrandomizedforeachsubject,and
theywerestudiedfortworespiratorymetronomerates):0.20
Hz(12breaths/min)and0.30Hz(18breaths/min).Thesesix
recordingswerethenrepeatedaftera1minrestperiod,giving
atotalof12recordingsforeachsubject.Repeatabilitywas
assessed from the two repeat measurements, and average
valuesfromthetwomeasurementswerecalculatedforboth
SBPandDBP.
Breathingwaspacedusingavisualmetronomewithequal
inspirationandexpirationperiods.Subjectswereinstructedon
howtofollowthemetronomeandweregivenfamiliarization
timebeforerecordingsstarted.

2.3Dataacquisition
Thecuffpressurecontainingtheoscillometricpulses(OscP)
was recorded together with a singlelead (lead II)
electrocardiogram(ECG).Achestmagnetometerwasusedto
recordchestwallmovement[7],andhence,areference
Table1 Summaryclinicaldetailsforthe30subjectsparticipatinginthestudy

MeanSD

Age (years)
Systolicbloodpressure(mmHg)
Meanarterialpressure(mmHg)
Diastolicbloodpressure(mmHg)
Armcircumference(cm)

DataarepresentedasmeanSD

13
MedBiolEngComput(2014)52:467473\

respirationsignal(Resp)wasobtained.ApiezoelectricmicrophonewasusedtorecordKorS,withthebellshapedstethoscope
terminalconnectedtothemicrophoneandplacedontheantecubitalfossaoftheforearminthepositionwherethebloodpulse
wasmostaudiblewiththecuffinflatedat(SBP+DBP)/2.Thestethoscopewassecuredtothearmwithadhesivemedicaltape.
Allsignalswererecordedsimultaneously,sampledat2kHz,16bit/sampleandstoredtoacomputerforofflineprocessing.Fig
ure1showstheexamplesofResp,OscPandKorSfortwoconsecutiverespiratorycyclesfromasubjectbreathingat0.3Hz.

2.4Signalprocessing
2.4.1 RRintervalwithdifferentrespiratoryfrequency
AnestablishedthresholdbasedQRSdetectionalgorithmwasusedtodetecttheECGRwave[12].ThemeanRRintervalacross
alldetectedheartbeatsforeachrecordingwascalculatedandusedforthecomparisonbetweenthetworespiratoryratesand
betweenthethreecuffpressures.

2.4.2 Oscillometricpulseamplitude
Thefootoftheoscillometricpulsewasautomaticallyidentifiedinafixed200mswindowfollowingtheRwave.Foreach
oscillometricpulse,thepulseamplitude(peaktonadir)wasmeasuredastheamplitudedifferencebetweenthepeakandthefoot.

2.4.3 Korotkoffsoundamplitude

TheKorSsignalwasbandpassfiltered(3dBpassband591,000Hz)toremovebaselinenoise.Thisbandencompassesthe
frequencybandofinterestinKorSanalysisusedin[1].Foreachpulse,thepeaktonadiramplitudein
Fig.1 Examplesofphysiologicaldatafortwoconsecutiverespiratorycyclesfromasubjectbreathingat0.3Hz(18
breaths/min).Toppanelchestmagnetometersignal.Centralpaneloscillometricpulse(solidtrace)andestimated
respiratorymodulation(dottedtrace,displacedforvisibility).BottompanelbandpassfilteredKorS(solidtrace)with
estimatedrespiratorymodulation(dottedtrace).Allverticalscalesareinarbitraryunits

thecorrespondingKorSwasretainedtogeneratethetimeseriesofKorSamplitude.
2.4.4 EstimationofrespiratoryratefromthetimeseriesofoscillometricandKorotkoffamplitudes

AstheOscPandKorSamplitudeswereseparatedbyvaryingpulseintervals,theywereevenlyresampledat4Hzbycubic
splineinterpolation.ThissampleratesatisfiestheNyquistconditionastherespiratoryrateinrestingconditionsisgenerally
containedintherange0.10.5Hz.Thisinterpolationandresamplingmethodsarecommonlyusedinpreliminaryprocessingfor
thespectralanalysisoftheRRintervaltimeseries,whichisalsomodulatedbyrespiration(respiratorysinusarrhythmia)[ 15].
ThemagnetometerRespwasthenalsoresampledat4HzthatwastimealignedtotheOscPandKorSamplitudedatawiththe
respiratorydepthcalculatedfromthepeaktonadiroftherespiratorywaveform.

TherespiratorymodulationsignalsfromResp,OscPandKorSwerethenzeromeaned(meanvaluewassubtractedfromsignals),
detrended(1stordertrendwasremoved)andzeropaddedto64s.TherespiratoryfrequencywascalculatedfromResp(f R),OscP(fO)
andKorS(fK)asthepeakfrequencyinthepowerspectraldistributionestimatedbytheWelchperiodogramusingaHammingwindow,
between0.1and0.5Hz,witharesolutionof16mHz.
ThedistributionsofthefrequencydifferencebetweenRespandthatcalculatedfromKorSamplitude(f RfK),andbetweenRespandOscP
amplitude(fRfO)werecalculatedacrossallrecordings.The presenceofrespiratory modulationwasacceptedwhenthe modulatedpeak
frequencyinOscPorKorSamplitudewasdetectedwithinthespectralresolutionof16mHz.Thehistogramoftheirfrequencydifferenceswas
plottedwiththebinwidthof16mHz.

Inordertotesttheshortrepeatabilityofmeasuredrespiratorymodulationfrequency,therootmeansquare

13

\470

MedBiolEngComput(2014)52:467473

(RMS)differencebetweenrepeatswascalculatedforOscPand
KorS.

(The Mathworks, Natick, MA, USA). Post hoc multiple

comparisonwasdoneusingthetwotail t testforpaired
samples.Asignificancelevel=0.05wasadopted.

2.4.5 Respiratoryamplitudemodulation

3Results

The respiratory amplitude modulation of OscP and KorS was

definedasthepeaktonadirdifferencenormalizedtothemean
amplitudeoverthe20sanalysisperiod.

3.1 Presenceofrespiratorymodulationofoscillometric
pulsesandKorotkoffsounds

2.4.6Phaseshift

As shown in Fig. 2, there was no statistically significant

difference in respiratory rate from Resp and that estimated
fromeitherOscPamplitudeorKorSamplitude,indicatingthat
thesesignalsweremodulatedbyrespiration.Table 2 shows
the number of recordings with respiratory modulation.
RespiratorymodulationofOscPandKorSamplitudeswas
observedinall30subjects.ForOscPamplitude,modulation
wasobservedinallsubjectsineightormorerecordings,and
in17subjectsmodulationwasobservedinallrecordings.For
KorSamplitude,theyweresixormorerecordingsandnine
subjects,respectively.

ThephaseshiftsbetweenRespandthemodulatedamplitudes
ofOscPandKorSwerecalculatedasthephasedifference
betweenthefirstpeakofthecrosscorrelationfunctionand
thezerolagpoint.Similarly,thephasedifferencebetweenthe
amplitudesofOscPandKorSwascalculated.

2.5Statisticalanalysis
TheeffectofrespiratoryrateandcuffpressureonKorSand
OscP amplitude, phase shift, RR interval and respiratory
depth was studied by threeway analysis of variance
(ANOVA)usingMATLABStatisticalToolboxsoftware
Fig.2 Distributionofrespiratory
rateestimationerroracrossall
recordings(N=360).Oscillometric
pulseamplitudemodulation(top
panel)andKorotkoffsound
amplitudemodulation(bottom).Bin
width(thesameasthespectral
resolution)is16mHz

Table2 Estimationof
respiratoryrate

CuffH,M,Lareforcuff
pressureshigh,medium,low
Thepresenceofrespiratory
modulationwasacceptedwhenthe
modulatedpeakfrequencyinOscP
orKorSamplitudewasdetected
withinthespectralresolutionof16
mHz

There was also no statistically significant difference in

measuredrespiratorymodulationfrequencybetweenrepeated

All
Resp0.2Hz
Resp0.3Hz
CuffH
CuffM
CuffL

13
MedBiolEngComput(2014)52:467473\

recordingsforeitherOscPorKorS.TheRMSdifferencebetweenrepeatswas10mHzforOscPand11mHzforKorS.
3.2 RespiratorymodulationofamplitudeforoscillometricpulsesandKorotkoffsounds

OscPamplitudemodulationchangedsignificantlybetweenthetworespiratoryrates(withameanandSDofthedifferenceof
0.040.09,p<0.05)andalsobetweenthethreecuffpressures(p<0.0001),whereasnostatisticallysignificantchangeswere
seenwithKorSamplitude(Table 3).ThestatisticallysignificantchangesforOscPmodulationwithdifferentcuffpressures
showedadecreasewithdecreasingcuffpressure,whichwassignificantlyreducedatboththemedium(M)andlow(L)cuff
pressures(p<0.05)(Fig.3),withameanandSDofdifferenceof0.080.14and0.16 0.15whenreferencedtothehigh(H)
cuffpressure.
3.3 RespiratorymodulationofphaserelationshipsforoscillometricpulsesandKorotkoffsounds

ThephaseshiftbetweenRespandtheOscPamplitudemodulationwasstatisticallysignificantbetweenthetworespiratoryrates
(withameanandSDofthedifferenceof0.030.09cycle,p<0.05)(Fig.3andTable 4).However,thedifferencewasnot
statisticallysignificantfordifferentcuffpressures(Table4).ThephasedifferencebetweenOscPandKorSamplitudeswasalso
notstatisticallysignificantforeitherrespiratoryrateorcuffpressure(Table4).
3.4 Heartrateandrespiratorydepthchangeswithrespiratoryfrequency

TherewasnostatisticallysignificantdifferenceinRRinterval(meanSDof0.880.12sforalltheconditions)
Table3 Oscillometricpulse(OscP)andKorotkoffsound(KorS)amplitudemodulationatdifferentrespiratoryratesandcuffpressures
Respiratoryrate
0.2Hz
OscP

0.340.13

KorS

1.870.92

DataarepresentedasmeanSD

Amplitudemodulationisdefinedasthechangefromminimumtomaximumdividedbythemeanamplitudeoverthe20sanalysisperiod

Fig.3 Toppanels oscillometricamplitudemodulation(left)and phaseshift

from respiratory magnetometer (right) with two respiratory rates (black
squarefor0.2Hz,whitesquarefor0.3Hz)andthreecuffpressures(H,Mand
L).Themeanvaluesand95%confidenceintervalsaregiven.Centralpanels
comparisonofoscillometricamplitudemodulation(left)andphaseshiftfrom
respiratorymag

netometer(right)betweenthetworespiratoryrates(referencedtothevalues
from 0.3 Hz). Bottom panels comparison of oscillometric amplitude
modulation (left) and phase shift from respiratory magnetometer (right)
betweenthethreecuffpressures[referencedtohighcuffpressure(H)].The
mean difference and 95 % confidence intervals of difference are given.
Asteriskindicatesp<0.05

13
472\
Table4 Phaseshiftchangeswithrespiratoryrateandcuffpressure
Respiratoryrate
0.2Hz
RespOscP
RespKorS

0.370.12
0.370.12

OscPKorS

0.000.08

DataarepresentedasmeanSD
Phaseshiftisgivenasafractionoftherespiratorycycle.Positivevaluesindicatedelayofthesecondsignalwithrespecttothefirst

respiratory rate which is associated with larger respiratory

movementofthechest.

betweenthetworespiratoryratesandbetweenthethreecuff
pressures(p>0.05).Respiratorydepthsignificantlyincreased
whenbreathingattheslowerrate(0.2Hz)comparedwiththe
higherrate(0.3Hz)by25 10%(p <0.0001).Thisisas
expected physiologically as respiratory depth needs to
increasewithslowerrespirationtoallowtransferofsimilar
oxygenlevels.

4Discussion
Thisstudyhasshownsignificantrespiratorymodulationofthe
oscillometric pulse and Korotkoff sound amplitudes from
whichBPismeasured.IftheamplitudeofKorSandOscP
signals was modulated by respiration, the respiratory
frequencycalculatedfromKorSandOscPwouldbeexpected
to be the same at that from Resp. Our results of the
nonsignificant difference of respiratory rate from Resp and
thatestimatedfromeitherOscPamplitudeorKorSamplitude
confirmed this hypothesis. In this study, respiratory
modulation was discovered in all subjects, but in a higher
proportion of recordings for oscillometric pulses (92 %
comparedwith84%).Thiscouldhavebeencausedbypoorer
signalqualityoftherecordedKorotkoffsound,perhapsdueto
greater depth of the brachial artery in some subjects or
physiologicalpropertiesofthearterialwall,orduetopoor
contactbetweenthestethoscopeandskin,orhighersensitivity
toslightmovementsofthearmduringmeasurement.
The respiratory depth was characterized by significantly
larger amplitude at the lower respiratory rate (0.2 Hz, 12
breaths/min), and this was associated with higher OscP
amplitudemodulation.Thisresultsupportsthehypothesisofa
direct mechanical coupling between respiration and the
vasculaturesuggestedbySauletal.[13].
TherespiratorymodulationofOscPamplitudedecreased
with cuff pressure at both respiratory rates (Fig. 3), sug
gesting mechanical coupling delivers higher energy at
increased load (resistance to arterial vessel distension
occurringathighercuffpressures).Thishypothesisseemsto
be confirmed by the significantly higher OscP amplitude
modulationatthehighcuffpressure,forthelower

13
MedBiolEngComput(2014)52:467473\

Astatisticallysignificantlargerphaseshiftwasobserved
between the respiratory signal recorded by the chest mag
netometerandtheoscillometricpulseamplitudemodulation.
Sin and colleagues [14] studied the phase relationship
betweenrespiratorydepthandtheSBPenvelope.IntheirFig.
6, the lower left panel shows a phase delay of 0.2 of a
respiratorycyclefortheSBPmaximumwithrespecttothe
beginning of inspiration for a respiratory rate of 0.2 Hz.
Assuming a regular respiratory pattern, 0.25 cycles are
requiredtoreachtheinspirationpeak,givingatotalof0.35
cycles delay between respiratory peak and SBP maximum.
Thisresultwasconsistentwithoursforthesamerespiratory
rate(Table4).
Sinetal.[14]alsoexaminedlowerrespiratoryrates,down
to0.1Hzandfoundadecreasingtrendinphasedelaywith
increasingrespiratoryrate,whichsuggeststhatalargertidal
volume(lowerrespiratoryrate)isassociatedwithincreased
inertialdelayinthemechanicalcouplingbetweenrespiration
and the arterial hemodynamic response. However, they did
notanalyzerespiratoryrateshigherthan0.2Hz.Inourstudy,
for a respiratory rate of 0.3 Hz, we found a significantly
increaseinphasedelaywithrespecttothelowerrateof0.2
Hz;namely,aninversionofthetrendobservedbySinand
colleaguesforlowerratesupto0.2Hz.Thiscouldbecaused
by the effects of nonmechanical processes such as
hyperventilationthatcomeintoplayonlyathigherrespiratory
rates.
Ourresultsalsoappearedconsistentwiththephaseshift
analysisofSauletal.[13].Althoughnumericvalueswerenot
reportedintheirstudy,theirFig.3showedabetweensubject
meanphasedelayofthepulsepressurewithrespecttothe
respiratory signal (instantaneous lung volume) of
approximately110,or31%oftherespiratorycycle,which
was similar for the respiratory rates of 0.2 and 0.3 Hz.
However,theirfigureillustratedonlygroupmeansforeach
frequency, and therefore, it is uncertain whether the phase
shiftvariedsignificantlywiththerespiratoryrate.

One limitation of this research was that only 30 healthy

subjectswithnormalBPswerestudied,andtheresults,although
providingphysiologicalinsight,werebasedon

statisticalsignificanceandnotclinicalsignificance.Afutureclinicalstudywithalargesamplesize,includingbothhypotensive
andhypertensivesubjects,wouldbeausefulinvestigationofthepotentialeffectofdiseaseandwouldhelpdetermineclinical
significance.However,thisworkhasalreadymadeanimportantstepinunderstandingthemodulationeffectofrespirationon
oscillometricpulsesandKorotkoffsounds.Animportantclinicalimplicationisthatamplitudechangesforoscillometricpulses
andKorotkoffsoundsarecommonlyusedforBPdetermination;andhenceunderstandingthatadditionalamplitudechanges
causedbyrespirationcouldhelpimprovetheaccuracyofclinicalBPmeasurementbyadvisingpatientstolimitdeeperbreathing
duringBPmeasurement.AlsoofrelevanceisthattherespiratoryratecanbederivedfromtheoscillometricpulseandKorotkoff
soundamplitude.ThiswouldaddanadditionalphysiologicalparameterduringnormalBPmeasurement.
In conclusion, this study has shown quantitative evidence of a respiratory modulation of the oscillometric pulse and
KorotkoffsoundamplitudeinnoninvasiveBPmeasurementinhealthynormotensivesubjects,implicatingtheinfluenceof
respirationonclinicalBPmeasurement.
Acknowledgments ThisstudywassupportedbytheEngineering andPhysicalSciencesResearchCouncil(EPSRC)HealthcarePartnershipAward(reference
numberEP/I027270/1),andEPSRCstandardgrant(referencenumberEP/F012764/1).
OpenAccess ThisarticleisdistributedunderthetermsoftheCreativeCommonsAttributionLicensewhichpermitsanyuse,distribution,andreproductionin
anymedium,providedtheoriginalauthor(s)andthesourcearecredited.

References
\1.\ llenJ,GehrkeT,OSullivanJJ,KingST,MurrayA(2004)CharacterizationoftheKorotkoffsoundsusingjointtimefrequencyanalysis.PhysiolMeas
25:107117
\2.\BlaberAP,YamamotoY,HughsonRL(1995)ChangeinphaserelationshipbetweenSBPandRRintervalduringlowerbodynegativepressure.AmJPhysiol
268(4Pt2):H1688H1693
\3.\BowersEJ,MurrayA(2004)Interactionbetweencardiacbeattobeatintervalchangesandsystolicbloodpressurechanges.ClinAutonRes14:9298
\4.\ChungcharoenD(1964)GenesisofKorotkoffsounds.AmJPhysiol207:190194
\ 5.\ deBoer RW, Karemaker JM, Strackee J (1987) Hemodynamic fluctuations and baroreflex sensitivity in humans: a beattobeat model. Am J Physiol
253:H680H689
\6.\DornhorstAC,HowardP,LeathartGL(1952)Respiratoryvariationsinbloodpressure.Circulation6:553558
\7.\ riffithsCJ,GilmartinJJ,GibsonGJ,MurrayA(1983)Measurementofchestwallmovement;design,performanceandclinicaluseofafourchannel
magnetometerinstrument.ClinPhysPhysiolMeas4:363371
\8.\JonesDW,AppelLJ,ShepsSG,RoccellaEJ,LenfantC(2003)Measuringbloodpressureaccurately:newandpersistentchallenges.JAMA289:10271030
\9.\ audeD,GoldmanM,EscourrouP,ElghoziJL(1993)Effectofbreathingpatternonbloodpressureandheartrateoscillationsin
humans.ClinExpPharmacolPhysiol20:619626
10\.\McCutcheonEP,RushmerRF(1967)Korotkoffsounds:anexperimentalcritique.CircRes20:149161
11\.\OBrienE,AsmarR,BeilinL,ImaiY,MallionJM,ManciaG,MengdenT,MyersM,PadfieldP,PalatiniP,ParatiG,PickeringT,RedonJ,StaessenJ,
Stergiou G, Verdecchia P (2003) European society of hypertension working group on blood pressure monitoring. European society of hypertension
recommendationsforconventional,ambulatoryandhomebloodpressuremeasurement.JHypertens21:821848

12\.\PanJ,TompkinsWJ(1985)ArealtimeQRSdetectionalgorithm.

IEEETransBiomedEng32:230236
13\.\SaulJP,BergerRD,Albrecht
RJ(1991)Transferfunction
uniqueinsightsintocardiovascularregulation.AmJPhysiol261:H1231H1245
\14.\SinPY,GalletlyDC,TzengYC(2010)Influenceofbreathingfrequencyonthepatternofrespiratorysinusarrhythmiaandbloodpressure:oldquestions
revisited.AmJPhysiolHeartCircPhysiol298:H1588H1599
15\.\ askForceoftheEuropeanSocietyofCardiologyandtheNorthAmericanSocietyofPacingandElectrophysiology(1996)Heartratevariability:standards
ofmeasurement,physiologicalinterpretation,andclinicaluse.Circulation93:10431065
\16.\ avelME,FarisJ,NasserWK,FeigebaumH,FischC(1969)Korotkoffsounds.Observationsonpressurepulsechangesunderlyingtheirformation.
Circulation39:465474
17\.\ aylorJA,EckbergDL(1996)FundamentalrelationsbetweenshorttermRRintervalandarterialpressureoscillationsinhumans.Circulation93:15271532
\18.\VenetR,MiricD,PavieA,LachehebD(2000)Korotkoffsound:thecavitationhypothesis.MedHypothese55:141146
19\.\WeiseF,LondonGM,GuerinAP,PannierBM,ElghoziJL(1995)Effectofheaddowntiltoncardiovascularcontrolinhealthysubjects:aspectralanalytic
approach.ClinSci(Lond)88:8793
\20.\ZhaoY,YamamotoM,MunakataM,NakaoM,KatayamaN(1999)Investigationofthetimedelaybetweenvariationsinheartrateandbloodpressure.Med
BiolEngComput37:344347
21\.\ZhengD,GiovanniniR,MurrayA(2012)Effectofrespiration,talkingandsmallbodymovementsonbloodpressuremeasurement.JHumHypertens
26:458462

13