
心率与年龄对照表
幼儿学习的特点-老公我想要
2023年3月18日发(作者:廉洁承诺一句话)大连理工大学硕士学位论文
摘要
HRv是指逐次窦性心动周期之间的微小变异,反映心脏自主神经系统的功能状态。
这种心搏间的微小差异,可以被计算机心电检测系统记录、测量和计算出来,作为临床
应用指导。已经公认HRV是定量分析心脏自主神经系统张力的方法。频阈分析是近年
来受到重视的分析方法之~,即对心率变异的速度和幅度进行心率功率谱的分析,亦称
心率能谱分析。在频域分析指标中,HF主要反映迷走神经张力变化;LF主要反映交感
神经张力变化,与外周血管温度调节、肾素、血管紧张素系统活动和心脏泵血功能等多
种因素有关;LF/HF则可评估心脏交感神经和迷走神经活动均衡性。
在最近的二十年,自主神经系统与心血管衰竭率,包括心脏的突然衰竭之问的密切
关系已经得到了大家的公认。实验证明,心率不齐与增强的交感神经或者减弱的迷走神
经活动的并发,刺激了自律活动在数量上的里程碑式发展。对心率变化的研究已经成为
评定自主神经系统的常用的工具。心率变化分析是以“快速的波动必将影响交感神经和
迷走神经的变化”这一理论为基础的。该分析指出,结构生成的信号并不是简单的线性
变化,而是包含非线性的变化。在本文中,作者计算出了不同年龄段人群的ECG信号
的最大Lyaptmov指数,结果表明,随着年龄的增长,心率呈递减的趋势;在论文中,
作者还给出了常态下的四组不同年龄段人群的非线性参数的范围。
关键词;tlltV;自主神经系统;迷走神经;心宰;Lyapunov指数
大连理工大学硕士学位论文
Heart
rate
variationinnormalsubjectsamong
various
agegroups
Abstract
HRVis
negligible
variationbetlⅣ咖the
periods
ofsuccessivesinusheartrate.reflects
thestatusoffunctionofeardial
automaticilervoussystem.This
kindof
negligible
variation
betweentheheart
ratecanbe
record,measure
and
calculate
by
computer
ECGDetection
System
anduseasguides
forClinical
application.It
is
well-known
thatHRVisthemethod
about
quantitative
analysisofeardialautomaticnervoussystemtension.frequency
threshold
analysis
is
OBeoftheanalysismethods
have
been
given
moreattentionthese
years,It
is
analysis
the
heartratepowerspectnanofthe
speed
and
amplitudeofheartratevariation,that
isheartratepowerspectnananalysis.Infrequency
domam
analysis.HF
refects
thetension
vailation
ofvagus;LF
reflects
thetensionvariation
ofsympathetic
llorVe.Itrelateswithouter
bloodvessel
temperamreregulatiothrennin,an#otensinsystemand
cardiac
pumping
function
andSOon;LF/HF啪evaluatetheproportionalitybetween
cardiacsympatheticnerve
and
vagus
activity.
Thelasttwodecades
havewimessedthe
recognition
ofa
significantrclatiomMp
between
theautonomicIIejt'VOus
syStem
andcardiovascular
mortality,including
sudden
cardiacdeath.
Experimental
evidenceforallassociationbeCweellapropensity
forlethal
arrhythmias
and
signs
of
eitherin.easedsympatheticorreduced
vagal
activity
has
encouragedthe
development
of
quantitative
markers
ofautonomic
activity.Analysis
ofheart
rate
variation
(HRV)hasbecomeapopularnoninvasivetool
for
assessing
theactivitiesoftheautonomic
nervous
system(ANS).HRVanalysis
isbasedonthe
concept
that
fast
fluctuations
may
specifically
reflect
changes
ofsympatheticand
vagalactivity.It
showsthatthestructure
generating
the
signal
isnot
simply
linear,but
alsoinvolvesnonlinearcontributions.Inthis
paper
the
largestLyapunovexponentofECGsignalsfordifferent
age
groups
hasbeen
calculated,Theresultsshow
that,with
aging
theheartrate
variability
decreases.Inthis
work,
the
ranges
ofnonlinear
parameters
for
four
agegroup
normalsubjects
are
presented.
KeyWords:HRV;Autonomicnervous
system;vagal;Heartrate;LyapunovExponent
—-III—·
独创性说明
作者郑重声明:本硕士学位论文是我个人在导师指导下进行的研究工
作及取得研究成果.尽我所知,除了文中特别加以标注和致谢的地方外,
论文中不包含其他人已经发表或撰写的研究成果,也不包含为获得大连理
工大学或者其他单位的学位或证书所使用过的材料.与我一同工作的同志
对本研究所做的贡献均已在论文中做了明确的说明并表示了谢意.
作者签名:复妞日期:幽晕12日加日
人连理工人学硕士研究生学位论文
大连理工大学学位论文版权使用授权书
本学位论文作者及指导教师完全了解“大连理工大学硕士、博士学位
论文版权使用规定”,同意大连理工大学保留并向国家有关部门或机构送
交学位论文的复印件和电子版,允许论文被查阅和借阅。本人授权大连理
工大学可以将本学位论文的全部或部分内容编入有关数据库进行检索,也
可采用影印、缩印或扫描等复制手段保存和汇编学位论文。
作者签名:望纽
聊躲≥兰毛
幽年卫月掣日
ntroduction
ne1asttwodecades
havewimessedthe
recognition
ofa
significant
relationship
betweell
theautonomicnervoussystem
andcardiovascular
mortality,mcluding
suddencardiac
death.
Experimentalevidenceforan
associationbetweenapropensity
for
lethalarrhythmiasand
signs
of
eitherincreased
sympatheticorreduced
vagai
activity
hasencouragedthe
development
ofquantitative
markersofautonomic
activity.
Heartratevariability(HRV)repr髓%tsoneofthemostpromisingsuchmafkcrs.1f1犯
apparentlyeasy
derivationofthis
me,asurehas
popularized
itsuse.Asmany
commercial
devices
nOW
provide
automatedmeasurementof
HRV,the
cardiologist
hasbeen邮vided
withaseeminglysimple
toolforbothresearch
and
elinica]studies.However,the
significance
and
meaning
ofthe
many
differentmeas蝴ofHRVaremore
complex
thangenerally
appreciatedandthere
isapotemialforincorrectconclusions
andforexcessiveorunfounded
extrapolations.
In
thiswn出the
nodme盯parameter
evaluatedshowheartvadation
among
different
age
groups·
111e
paperarrangedin
following,the
first
chapter
has
introducedthe
background
of
this
paper,thenithasdescribedvariousmethodsfor
heartratemeasurements,theclinical
USeof
heartratevariability,In
the
later
chapters
ithasdescribed
LargerLyapunovexponent
and
laterthedatafrom
varioussubjectshasbeen
analyzed.
望竺璺墨坐!堂垫塑竺翌型墅堕竺垫竺些竖!竺塑竺垒壁塑
1Introduction
Thelasttwodecadeshavewitn髓sedthe
recogrlition
ofa
significant
relationship
between
theautonomicnervoussystem
andcardiovascular
mortality,including
suddencardiacdeath.
Experimental
evidenceforallassociationbetweenapropensity
for
lethalarrhythmias
and
signs
of
eitherInereesed
sympatheticorreducod
vagal
activity
hasencouragedthe
developmentofquantitative
markersofautonomic
activity.
Heartratevariability(HRV)representsoneofthemostpromisingsuchmark@l'S.ne
apparentlyeasy
derivationofthismeasurehas
popularized
its
use.Asmany
commercial
devicesnow芦odde
automated
measur9111ent
of
HRV,the
cardiologist
has
been
provided
withaseeminglysimple
toolforbomresearchandclinical
studies.However,thesignificance
and
meaning
ofthemany
differentmeasures
of
HRVare
more
complex
thangenerally
appreciated
andthereisapotentialforIncorrectconclusionsandforexcessiveorunfounded
extrapolations.
Recognition
ofthese
problems
ledthe
European
Society
of
Cardiology
andtheNorth
American
Society
of
Pacing
and
Eleetrophysiology
toconstituteaTaskForechargedwitlI
the
responsibility
of
developingappropriate
standards.The
specificgoals
ofthisTaskForce
wereto:standardizenomenclatureand
develop
definitions
ofterms;specify
standardmethods
of
measurement;define
physiological
and
pathophysiological
correlates;describe
currently
appropriate
clinical
applications,and
identify
areasforfutureresearch.
Inordertoachievethese
goals.the
membersofthe
T勰kForceworedrawn
from
the
fieldsofmathematics,engineering,physiology,andclinicalmedicine.Thestandardsand
proposals
off盯edinthistextshouldnotlimitfurtherdevelopmentbut,rather,should
allow
appropriatecomparisons,pronlotccircumspectInterpretations,andleadtofurther
progress
in
thefield.
一2一
大连理工大学硕士学位论文
2
Background
Theclinicalrelevanceofheartratevailabillty
wasfirst
appreciated
in1965whenHon
andLee【1】notedthatfetal
distresswasprecededby
alterationsininterbentintervalsbefore
anyappreciablechange
occurred
intheheartrateitself.Twenty
yearsago,Sayers
andothers
focusedattentionontheexiatenee
ofphysiologicalrhythms
imbeddedinthebeat-to-beatheart
ratesignat【2巧】.
During
the1970s,Ewingcta1.【6】devisedanumberof
simple
bedsidetestsof
short-termRRdifferencestodetect
autonomic
neuropathy
indiabeticpatients.The
associationof
higher
riskof
post-infarction
mortality
withreducedHRV
wasfirstshown
by
Wolfeta1.in1977【7】.In1981,Aksekodeta1.introducedpowerspectralanalysis
ofheart
ratefluctuationstoquantitativdy
evaluatebeat-to-beatcardiovascular
control【8】.
These
frequency-domainanalyses
contributedtothe
understanding
oftheautonomic
background
ofRR
intervalfluctuations
intheheartraterecord【9,lO].The
clinical
importanceofHRVb{N2a.iile
apparent
inthelate1980swhenitWasconfirmedthatHRVWasa
strong
and
independent
predictorofmortality
following
anacutemyocardialinfarction【1l,
12].Withtheavailabilityofnew,digital,highfrequency,24-h
multi-channelelectrocardio
graphicrecorders,HRV
hasthe
potential
to
provide
additionalvaluable
insight
into
physiological
and
pathological
conditions
andtoelhllanceriskstratification.
TheHeartratevariability(rmv)isanon-invasiveindexoftheneuralcontrolofthe
heart.HRV
Canbequantifiedby
the
simple
calculationofthestandarddeviationoftheRR
intervals.Furthermore,in
the
frequencydomain,spectralanalysis
ofHRVrevealsthree
distincttiequencyregionsinthemodulationofheartrateinhumans.The
typicalspectral
pattern,in
normalconditionsshowsthepresence
ofthree
frequency
bands:a
very
low
frequency(VLF)band
from0.00toO.03
Hz,a
low
frequency(LF)band
fromO.03to0.15Hz
anda11ighfrequency(HF)bandin
respiratoryrange
generally
morethanO.15Hz.
ThepowerofLFcomponentSeeInstoberelatedtothe
vagal
and
sympathetic
activities
(LF
componentincreaseswith
every
form
ofsympatheticstimulation),whereastheareaof
bighfrequencycomponent(HF)providesaquantitative
indexoftheinfluenceof
respiration
onECGsignal
and
may
beconnected
tothe
vagalactivity.Thus
LF,HFratioisan
important
markerofsympatheticmodulationorsympatho·vagalbalanceonheartratevariability
contr01.
Physiologicalsignals
often
vary
inacomplex
and
irregularmanner.Analysis
oflinear
statisticssuch弛mean
values,variabilitymeasures,andspectra
ofsuch
signalsgenerallydoes
not
address
directly
their
complexity
andthus
may
miss
potentiallyusefulinformation.Since
the
underlying
mechanismsinvolved
in
thecontrol
ofheartratearemainly
nonlinear,the
application
ofnonlineartechniques踮咖sappropriate【13-16].
3一
Recently,new
dynamicmethodsofHRV
quantification
havebeenused
tounoover
nonlinearfluctuationsinheartratethatarenototherwise
apparent.Several
methodshavebeen
proposed:Lyapunovexponents【17],l/fslope【18】,approximateentropy(ApEn)[19】and
detrendodfluctuation
analysis【20].Comparedtomen,womenareatlowerriskof
coronary
heartdisease[21】andofserious
arrhythmias【22],suggestingabeneficialdifferenc宅in
autonomiccon协olofheartrate.Previousstudieshaveassessedgenderand
age-telatod
diffcrencesintimeandfi-equencydomainindicesf23】and80menonlinearcomponentof
HRV【241.Therealsoseemedtobeasignificantdifferencebetween
day
and
night
hours
when
studying
HRVindicesusingspectral
andtime
domain
methods【25].Terryet
al,have
shownthatthemiddle.aged
womenandmenhaveadominant
parasympathetic
and
sympatheticregulation
oftheheartrate,respectively【26].Itisproved
that,theheartrate
v鲥abilitydependsonthes懿also.TheheartrateV撕abilityiSmoreiuthe
physically
active
young
and
oldwomen
f271.The
reduc,edHRV
rcsultsinnewcardiaceventsarestudied
by
Hisakoeta1.Itisprovedby
Emeseeta1.thatthealertnewboresheartratevailationislOWgr
in
thecaseof
boys
thanin
the
c笛eofgirls[281.TheHeartratevariationfor
healthys蛔cas
from20.70
yrs
isstudied
by
Hendriketa1.andfound
that
the
HRVdecreaseswith
age
and
vailatiOlliSmoreinthecaseoffemalethanmen.Galeeveta1.haveanalyzedtheheartrate
variationfor
healthysubjectsof
age
fi'om6to16
yrs
and
observed
the
statistical
and
fi'equency
domainvariation奶m
aging
and
gender【29].Inthis
work,we
have
analyzedthe
heartrate
variationofnormalsubjectsfrom
23-90
yrs
in
fourgroups(204-10,30士10,41
4-
29,71士19)by
nonlinearmethod.
一4一
大连理工大学硕士学位论文
3Measurementofheart
ratevar
iabiI
ity
3.1Tiritedomainmethods
Variationsinheartrate
may
be
evaluated
byanumber
ofmethods.Perhaps
the
simplest
toperformarethetimedomainmoasllrcs.Withthesemethodseither
the
heartrate
atany
∞intintimeortheintervalsbetween鞠-1.j嘲sivenormal
complexesaredetermined.Ina
continuous
electrocardiographic(ECG)record,eachQRScomplexis
detected,and
the
∞-called
normal-to-normal(NN)intervals(thatisallintervalsbetweenadjacentQRS
complexesresulting
fromsinus
node
depolarizations,or
the
instantaneoushea.qrateis
determined.Simoletime-domainvariablesthatcanbecalculatedincludethemeanNN
interval,the
meanheart
rate,the
differencebetween
the
longest
andshortestNN
interval,the
differencebetween
night
and
day
heart
rate.ete.Other
time-domainmeasurements
that眦be
usedarevariationsininstantaneousheartratesgcondarytorespiration,tilt,Valsalva
malloeuvre,orsecondaryto
Dheilylephrine
infusion.Thesedifferencescabbedescribed勰
eitherdifferencesinheartrateorcyclelengtlL
3.2StatisticaImethods
FromaseriesofinstanhalleoBsheartratesorcycleintervals,particularly
thoserecorded
over
longerperiods,traditionally
2411,more
complexstatisticaltime-domainmeasurescanbe
calculated.These
may
bedividedinto
twoclasses,①those
deriredfromdirect
measurementsoftheNNintervalsorinstantaneousheartrate,(Dthose
derived
from
the
differencesbetweenNNintervals.Thesevuriables
may
bederivedfrom
analysis
ofthetotal
electrocardiographicrecordingormay
becalculated
using
smaller
segmentsofthe
recording
period.The
latter
method
allowscomparisonofHRVtobemade
during
varymgactivities,e.g.
rest,sleep.etc.Thesimplest
vailabletocaleulateisthestandarddeviationof
meNNinterval
(SDNN),i.e。thesquare
rootof
variance.Since
varianceis
mathematicallyequal
lototal
power
of
spectralanalysis,SDNN
reflectsall
the
cycliccomponentsresponsiblefor
variabilityinthe
periodofrecording.Inmanystudies,SDNNiscalculatedovera24-h
period
andthus
encomp船ses
bothshort-term
highfrequencyvariations,as
wellasthelowest
frequency
componentsseen
ina24-h
period.As
theperiodof
monitoring
decreases,SDNN
estimatesshorterandshorter
cyclelengths,Itshould
alsobenotedthatthetotalvarianceof
HRVincreaseswiththelength
ofanalyzedrecording[30].Thus,onarbitrarily
selected
ECGs,
SDNN
isnotawelldefinedstatistical
quantity
becauseofitsdependenceon
thelengthof
recordingpaiod.Thus.inpractice,it
is
inappropriate
tocompareSDNNmeasuresobtained
from
recordings
ofdifferent
durations.However,durationsofthe
recordings
usedto
一5一
determineSDNNvalues(andsimilady
otherHRVmeasures)shouldbe
standardized.
Short—term5-min
recordings
andnominal24h
long-terrarecordings
senititobe
appropriate
options.Otherconm30nlY
usedstaffsticalvariablescalculatedfromsegments
ofthetotal
monitotingpedod
includeSDANN.thestandard
deviation
ofthe
average
NNinterval
calculatedOVershort
periods'usually
5
rain,which
iSanestimateofthechangesinheartrate
duet0
cycleslonger
than5min,and
theSDNN
index.the
moanofthe5·rainstandard
deviationoftheNNintervalcalculatedOVeI*2411.whichmeasures
the
variability
dueto
cycles
shotterthan5minutes.Themost
commonly
used
measures
derivedfTominterval
differencesincludeRMSSD.the
square
rootofthemean
squared
differencesofsuccessive
NN
intervals,NN50,the
number
ofintervaldifferances
of
successiveNNintervals
greater
than50ms.and
pNN50
the
proportion
derived
bydividing
NN50
by
thetotalnumberofNN
intervals.AII
thesemeasurementsofshort-term
variation
estimate
highfrequencyvariations
inheartrateandthusare
highlycorrelated(F嘻3.1).
3.3GeometricaImethods
ThesefiesofNNintervalscanalsobeconvertedintoageometricpattem,suchasthe
sampledeusitydistribution
ofNN
intervaldurations,sampledensity
distribufionof
differencesbetweenadjacentNN
intervals,Lorenz
plot
ofNNor
RRintervals,etc.,anda
simple
formulaisusedwhich
judgesthevariabilitybased
on
the
geometric
and/or
graphic
properties
of
the
resultingpattern.Threegeneralapproachesareusedin
geometric
methods:
①abasicmeasurementofthegeometricpattern(e.g.thewidthofthedistribution
histogram
atthe
specifiedlevel)is
convertedintothemeasureofHRV,②the
genmetriepaRem
iS
interpolatedbyamathematically
de丘ncdshape(e.昌approximationofthedistribution
histogram
byatriangle,orapproximationofthedifferentialhistogram
byallexponential
curve)and
thentheparametersofthismathematical
shape
areused,and③thegeometric
shapeis
一6一
大连理工大学硕士学位论文
F培3.1Relationship
bcIfwccntheRMSSD∞d
pNN50(a),andpNN50
andNN50(”m日日哪ofHRV
assessedfrom
857
nominal
24-hHolter
Pattern-basⅨl
categories
whichrepresent
different
classesofHRV(e.辱elliptic,2inear
and
triangularshapes
ofLorenzplots).Mostgeometric
methods
require
thcRR(orNN)
intervalsequeIlcetobemeasuredonorconvertedtoadiscrete
scalewhichiSnottoofineor
toocoarseand
whichpermitstheconstructionofsmoothed
histograms.Mostexperience
has
beenobtained
with
bins
approximately
8mslong(precisely7·8125ms=1/128S1which
corresponds
tothe
precision
ofcurreotcommcTcial
equipment.TheHRV
triangular
index
measurementiSthe
integral
ofthedeositydistribution亿e.thenumber
ofa11
NNintervals)
divided
by
themaximum
ofthe
density
distribution.UsingameasurementofNN
intervalson
adiscrete
scale,the
measureis
approximatedby
thevalue:(totalnumberof
NN
intcrvals)/(numbcr
ofNN
intervals
in
themodalbin)whichiSdependentOHthe
length
ofthe
bin,i.e.On
the
precision
ofthediscretescaleofmcasurcmont.Thus.ifthediscrete
一7一
墅坠!堂垫型型墅塑型!堂塑堕
approximation
ofthemeBsureisused、析tllNNintervalmeasuremefltona∞aledifferentto
themost
frequentsampling
of128H乙thesizeofthebinsshould
be
quoted.
Duratlma0fnormalRRtntervals
Fig.3.2
The
U'ia.gularinterpolation
ofNN
Tope触geometricalmeasures
ontheNN
intervalhistogrmn,the
sampledensity
distributionDis
constructed
which
assigns
thenumberof
ezluallylong
NNintervalstoench
valueoftheir
lengths.111e
most
frequent
NN
interval
length
X
is
established,thatisJr=D∞
isthemaximumofthe
sampledensity
distributionD.TheHRV
triangular
indexiSt11evalue
obtained
bydividing
theareaintegral
ofD
by
themaximum
y.When
constructing
the
distributionDwithadiscretescaleonthehorizontalaxis.thevalueisobtained
according
to
theformulaHRViadex=(totalnumberofallNNintervals)/Y.
ForthecomputationoftheTINNmeasure.thevalUesNandMareestablishedonthe
timeaxisandamultilinearfunction
q
constructedsuchthat4(f):Ofor匿Nandf≥Mand
ga产Eandsuchthatthe
integral
』(础)一g(f))2dt,M_<tgN,g(石)=y
0
istheminimum
anlong
allselections
ofallvaluesNandM.mTINNme.aSBl*e
is
expressed
inmsand
givenby
theformulaTINN=M.N.interval
histogram(TINN)is
the
baselinewidth
ofthedistributionmeasuredas
abaseofatriangle,approximatingthe
NN
intervaldistribution(the
minimum
square
differenceisusedtofindsuchatrian西e).Detailsof
computing
the
HRV
triangular
indexandTINNareshown
in
Fig.2.Both
these
measures
express
overallHRVmeasuredover24handaremoreinfluenced
by
thelOWer
than
by
the
higIlel"fi-equencies【31].Othergeometric
methodsarestill
inthe
phase
of
exploration
and
explanation.Themajoradvantageof
geometric
methodsliesintheirrelative
insensitivity
to
一8一
胃;Ⅳ髫
igl|o
大连理工大学硕士学位论文
the
analytical
quality
oftheseriesofNNintervals【32].The
majordisadvantage
istheneed
forareasonablenumberofNNintervalsto
constructthe
geometricpattern.Inpractice,
recordings
ofatleast20min(butpreferably
24h)should
beusedtoousurethecorrect
performanceofthe
geometric
methods,i.e.thecurrent
geometric
methodsageinappropriate
to
assessshort-term
changesinHRV.
3.4
Frequency
domeinmethods
Various
spectralmethods【33】for
theanalysisofthe
tachogram
have
been
applied
since
thelate1960s.Powerspectraldensity伊SD)analysisprovides
thebasicinformationofhow
power(i.e.variance)distributesasafunctionof呐uency.Mdcpendentofthemethod
employed,only
anestimateofthe
trim
PSDofthe
signalscanbeobtained
by
proper
mathematical
algorithms.
MethodsforthecalculationofPSD
may
be
generally
classifiedas
non-parametric
and
parametric.In
mostinstances.botllmethods
providecomparable
results.Theadvantagesof
the
non-parametric
methodsage."①the
simplicity
ofthe
algorithmemployed(Fast
Fourier
Transform—FFT-in
mostofthecases)and(萤melligh
processing
speed.whilst
the
advantagesofparametricmethodsare:smoother
spectralcomponentswhichcanbe
distinguishedindependently
ofpreselectedffequency
bands;easypost-processing
ofthe
spoctmmwithanautomaticcalculationoflowandhi曲呐uencypowercomponents
and
cosy
identificationofthecentralffequencyofeachcomponent;anaccurateestimationofPSD
evenonasmallnumberof
samplesonwhichthe
signal
issupposedtomaintain
stationarity.
Thebasic
disadvantageof
parametric
methodsistheneedtovmfythe
suitability
ofthe
chosenmodelandits
complexity“e.the
orderofthemodel).
Only
the
LFandHFcomponents
correspondto
peaks
ofthe
spcgtrum
whilethevLFand
ULFCanbeapproximatedbyalineinthis
plot
with
logarithmic
scalesonbothaxes.The
slope
ofsuchaline
isthemeasureofHRV.
Spectralcomponents
一9
墅墼!型竺兰型型塑壁咝!麴塑哑
Frequency砰酗
Fig.3.3
Example
ofaneatimate
ofpowerspectral
demityobtainedfrom
the
entire
24-hintervalofa
Ionsterm
Holter
recording.
Short-term
recordingsthreemain
spectralcomponentsaredistinguished
inasp戗:tnm
calculatedfrom
shortterm
recordings
of2to5mill:verylOWfrequency(VLF),lowfrequency
(LF),and
high厅equency(HF)components.Thedistributionofthe
power
andthecentral
丹equencyofLFandHFarenotfixedbut
mayvary
in
relationto
changes
inautonomic
modulationsoftheheartperiod.nephysiologicalexplanation
oftheVLF
component
is
muchlessdefinedandtheexistenceofaspecificphysiologicalprocess
attributabletothese
heart
period
changesmight
evenbe
questioned.The
non-harmonic
component
whichdoesnot
havecoherent
properties
andwhichisaffected
byalgorithmsofbaselineortrendrernovalis
commonlyaccepted笛amajorconstituentof
VLF.Thus
VLFassessedfromshort-term
recordings
fc.g.5
min)isadubiousmeasuleandshouldbeavoidedwhen
interpreting
the
PSDofshort-termECGs.MeasurementofVLF.ISandHF
powercomponentsiS
usually
madeinabsolutevaluesofpower(ms2),butLF
andHF
mayalsobemeasuredinnormalized
units(n.u.)which
represent
therelativevalue
of
each
powercomponentinproportiontothe
totalpower
minustheVLFcomponent.TherepresentationofLFandHFin
n.u.emphasizes
thecontrolledandbalancedbehaviourofthetwobranchesoftheautonomic
nervoussystem.
Moreover。normalizationtendstominimizethe
effectonthevaluesofLFandHFcomponents
ofthe
changesintotalpower.Nevertheless.n.u.shouldalways
bequotedwithabsolutevalues
of
LF
andHF
power
inordert0describeintotalthedistributionof
power
in
spectral
components.
Long-term
recordingsSpectral
analysismay
also
be
usedtoanalysethe
sequence
ofNN
intervalsintheentire24-h
period.The
resultthenincludesallultra-low台equencycomponent
(UL】哆,inadditionlo
v】LF,LF
andHFcomponents.1kslope
ofthe24-h
spectrum
Canalso
10一
大连理工大学硕士学位论文
be鹊g髂scdonalog-log
scale
by
linear
fitting
thespec嘲values.Table21istsselected
frequency-domain
measures.The
problemof‘stationarity'isfrequently
discussed
with
long-term
recordings.If
mechanisms
responsible
for
heart
period
modulationsofa
certain
frequency
remain
unchangedduring
thewhole
period
of
recording
the
corresponding
frequencycomponentofHRV
may
beusedasameasugeofthesemodulations.Ifthe
modulationsarenot
stable,interpretation
oftheresultsoffrequencyanalysis
islesswell
defined.In
particular,physiological
mechanismsofheartperiod
modulations
responsible
for
LFandHF
powercomponentscannotbeconsidered
stationaryduring
the24-h
period.Thus,
spectral
analysisperformedintheentire24-h
periodaswellasspectral
resultsobtainedfrom
shortersegments(e.g.5rain)averaged
OVertheentire2仙period(theLFandHFresultsof
thesetwocomputationsarenotdifferent)provideaverages
ofthemodulations
attributableto
theLFandHFcomponents(Fig.3.31.Suchaverages
obscuredetailedinformationabout
autonomicmodulationof
RR
intervalsavailablein
sbOrt盯recordings.It
should
be
rememberedthatthecomponentsofHRV
provide
measurementsofthe
degree
ofautonomic
modulationsratherthanofthelevelofautonomictoneand
averages
ofmodulationsdonot
repgesentallaveraged
leveloftone.
3.4.1Ieehni
cQI
roqu
irementsandreooemendations.
Becauseofthe
important
differencesinthe
interpretation
ofthe
results,thespectral
analyses
ofshortand
long-termeleetrocardiograms
should
always
be
strictlydistinguished
The强aIysedECG
signal
should
satisfy
several
requirementsinordertoobtainareliable
spectralestimation.Anydeparture
fromthe
followingrequirementsmay
leadto
unreproducible
resultsthataredifficultto
interpret.In
ordertoattributeindividual
spectral
componentstowell
definephysiologicalmechanisms,such
mechanismsmodulatingtheheart
rateshouldnotchangeduring
therecording.Transientphysiologicalphenomenamayperhaps
be
analysedby
specific
methodswhichcurrently
constituteachallengingresearch
topic,but
whichagenot
yetready
tobeusedinappliedresearch.Tocheckthe
stability
ofthe
signal
in
termsofcertain
spectral
components.traditionalstatistical
testsmaybe
employed.11坞
samplingrate
has
tobe
properly
chosen.Alow
samplingratemayproduceajitterin
the
estimationoftheRwavetiduciai
point
whichaltersthe
spectrumconsiderably.Theoptimal
rangeis250-500
Mz
orperhaps
even
higher,whilealowersamplingrate(in
any
case100Hz)
may
behave
satisfactorilyonly
ifanalgorithm
ofinterpolation(c.鲁parabolic)isusedto
refine
theRwave
fidudalpoint.Baseline
andtrendremoval(ifused)may
affectthe
lower
componentsin
the
spectrum.It
isad“sabletocheckthefrequencyresponseofthefilterorthe
behaviourofthe
regressionalgorithm
andtoveilfythatthe
spectralcomponentsofinterest
ale
notsignificantly
affected.
旦竺璺坠!堂垫婪竺型塾堕塑型!型竺坐鲤
ThechoiceofQRSfiducial
pointmay
becritical.Itis
neo晦sary
to
useawelltested
algorithm(i.e.derivative+threshold,template,correlationmethod,etc.)inordertolocatea
stableand
noise-independent
reference
point.A
fiducialpointlocaliz司farwithintheORS
complexmay
also
beinfluenced
byvarying
ventricularconduction
disturbances.Ectopic
beats.arrhythmic
events,missing
data
andnoiseeffects
may
alterthe
esfimationofthePSD
of
HRV.Properinterpolation(or
linear
recessionorsimilar
algorithms)Oil
preceding/successivebeatson
theHRV
signalsoron
itsautocorrelationfunctionmayreduce
this
error.Preferentially,short-termrecordings
whicharefleeof
ectopy,missing
data,and
noiseshouldbeused。Insomecircumstances,however,acceptanceof
orgyectopic-free
short·termrecordingsmayintroduce
significantselection
bias.Insuch
cas翩.proper
interpolation
shouldbeusedandthepossibilityoftheresults
being
influenced
byeetopy
shouldbeconsidered.Thetelativenumberandrelativedurationoflmintervalswhich、糯
omittedand
interpolated
shouldalsobequoted.
3.4.200rrelatlonanddifferencesbetweentimand
frequency
domalnmeasures
Ⅵm髓analysingstationary
short-term
recordings.moreexperienceandthenretical
knowledge
existsonthe
physiologicalinterpretation
oftlleffequenc州omainmeasures
comparcdtothetime-domain
measlesderivedfromthesame
recordings.However,many
timeandfrequency-domainvariablesmeasuredOVertheentire24-h
periodarestrongly
correlatedwitheachother.These
strong
correlationsexistbecanseof
bothmathematicaland
physiologicalrelationships,m
addition.thephysiologicalinterpretation
ofthe
spectral
components—ca—lculatedOVer24his
difficult,for
thereasonsmentioned(section
entitled
Long-term
recordings).Thus,tmlessspccial
investigations
areperformedwhichusethe24-h
HRV
signal
toextractInformationotherthanthe
usualfrequencycomponents(e.g.the
log—logslope
ofspectrogram),the
resultsof
frequency-domainanalysis
areequivalentto
thoseoftime,--domain
analysis,which
iseasierto
perform.
3。5
RhythmpatternanaIysiS
Asillustratedin
Fig.3.4,the
time-domainand
spectral
methodssharelimitations
imposedby
theirregnlantyoftheRR
series.Clearly
different
profilesanalysed
by
these
techniquesmaygive
identical
results.Trends
of
decreasingorincreasingcyclelength
arein
realitynotsymmetrical【34,35】勰heartrateaccelerationsareusually
followed
byafaster
decrease,In
spectralresults,this
tendstoreducethe
peak
atthefundamental丘equency,andto
enlarge
itSbasis.T1lisleads
to
theideaof
measuring
blocksof
RR
intervalsdetermined
by
properties
ofthe
rhythm
and
investigating
the
relationship
ofsuchblockswithoutconsidering
theintemal
variability.
一12一
大连理工大学硕士学位论文
Approaches
defiMedfromthetime-domainandthefrequency-domainhasbeen
proposed
inordertoreducethesedifficulties.Theintervalsp_cl锄嗍andsp_。ctmm
ofcoul怄methods
leadto
equivalentresults(Fig.3mandarewellsuitedtoinvestigate
the
relationship
betweeal
HRVandthe
variability
ofother
physiological
me&sures.Theinterval
spectrum
iswell
adapted
tolinkRRintervalstovariablesdefinedonabeat-to-beat
basis(e.g.bloodpressure).
11lc
spectrum
ofcountsis
preferable
ifRRintervalsarerelatedtoacontinuoussignal(e.吕
respiration),ortotheocculTeqllCeofspecial
eveuts(e.g.arrhythna曲.The‘peak-valley’
proceduresarebasedeitheron
the
detectionofthesummitandthenadirofoseillations『36,
37】orOllthedetectionoftrendsofhenrt
rate【38].Thedetection
may
belimitedtoshort-term
changes
butitcanbeextendedto
longer
variations:secondandthirdorder
peaks
and
troughs
orstcpwise
increaseofasequ黜ofconsecutiveincreasingordecreasingcycles
surro岫ded
byopposite
trends.11怆variousoscillationsoanbeeharaetcrizedonthebasisoftheheartrate
acceleratingorslowing,thewavelength
and/or
the
amplitude.Inamajorityofshort-to
mid-term
recordings,the
resultsarecorrelated丽th呐uencycomponentsofHRV【39].
nle
correlations.however,tendtodiminish勰the
wavelengthoftheoscillationsandthe
recording
duration
increase.Complex
demodulation
BSesthe
techniques
of
interpolation
and
detrending【40】andprovides
thetime
resolution
necessary
todetectshort-termheartrate
changes.as
wellastodescribetheamplitudeandphaseofparticular台equencycomponentsas
functionsoftime.At
present,thenon-linearmethods
representpotentiallypromising
toolsfor
HRVassessment,butstandardsare
lacking
andthefull
scope
ofthesemethodscannotbe
assessed.Advancesin
technology
andthe
interpretation
oftheresultsofnon-lineaxmethods
areneededbe如rethesemethodsare
ready
for
physiological
andclinical
smdies.
一13一
坚塑坠!墅墅垫型竺鲤坠墅壁螋!塑竺壁唑
(c)
05∞
Fig.3.4
Example
offour
synthesised
timeserieswith
identicalme卸s,standarddeviations。and
ranges.
Series(c)and(d)also
haveidenticalautocorrelationfenctionsandthereforeidentical
powerspec仃:a.
3.6Non-I.nearmethods
Non-linear
phenomenaalecortainlyinvolvedin
the
genesis
of
HRV.111cyare
determined
bycomplex
interactionsof
haemodynamic,electrophysiological
andhumoral
variables,aswellas
by
autonomic
andcentralneqt-vousregulations.It
has
beenspeculatedthat
analysis
ofHRVbasedonthemethodsofliOn—linear
dynamicsmight
elicitvaluable
informationforthe
physiological
interpretation
ofHRVandfortheassessmentoftheriskof
suddendeath.Theparameterswhichhavebeenusedtomeasurenon-linenr
properties
ofHRV
includel/f
scaling
ofFourierspectra【41,42],Hscaling
exponent,andCoarse
Graining
SpectralAnalysis(CGSA)Fordatarepresentation,Poincaresections,10W-dimensionattractor
plots,singularvalue
decomposition,and
attractortrajectorieshavebeenused.Forother
quantitativedcscriptiOIlS,the
D2correlation
dimension,Lyapunovexponents,and
Kohnogoroventropyhavebeenemployed[43].
Although
in
principle
these
techniques
havebeenshowntobepowerfultoolsfor
characterization
ofvariouscomplexsystems,nomajorbreakthrough
has
yet
beenachieved
by
their
application
tobio-medicaldata
including
HRV
analysis.It
ispossiblethat
integral
complexity
measuresalenot
adequate
to
analyzebiologicalsystems
and
thus,aretoo
insensitivetodetectthenon—linear
perturbations
of
RR
intervalwhichwouldbeof
physiologicalorpracticalimportanco.More
encouragingresultshavebeenobtained
using
differential,ratherthan
integralcomplexitym翩sur髓,e.g.thescaling
indexmethod
However;,
nosystematic
study
hasbeen
conducted
to
investigate
large
patientpopulationsusing
these
—14一
methods.At
present,the
liOn·linearmethods
represent
potentiallypromising
tools
for
HRV
assessment,but
standardsarclacking
andthefull
scope
ofthesemethods
cannotbeassessed.
Advancesintechnologyandtheinterpretationoft.he
resultsofnon-linearmethodsarcneeded
before
these
methodsaieready
for
physiological
andclinical
studies.
一15一
4
Physi0IogicaI
correIares
ofheartrateVariabi
I
ity
4.1AutonomioinfIuences
ofheartrate
Although
cardiac
automaticity
isintrinsictovarious
pacemakertissues,heart
rateand
rhythmarelargely
underthe
con订ol
oftheautonomicnervoussyst唧.
Thc
parasympathetic
influenceonheartrateismediatedviareleaseof
acctylcholineby
the
vagus
nerve.Muscarinic
acetylcholinereceptors
respondto
this
release
mostlyby
an
increaseincellmembrane
K+conductance.Acetyicholine
alsoinkibitsthe
hyperpolarization
activated‘pagemaker’current
IfThe‘lk
decay'hypothesisproposes
that
pacemaker
depolarization
resultsfromslowdeactivationofthedelayed
rectifier
current,m,which,dueto
atime-independent
backgroundinwardcurrent,causes
diastolic
depolarization.Conversely,
the‘Ifactivation’hypothesissuggest
that
following
actionpotentialtermination,Ifprovidesa
slowlyactivating
inwardcurrent
predominating
OVer
decaying
IL
thus
initiating
slow
diastolic
depolalization.
The
sympathetic
influenceOnheartrateismediated
by
releaseof
epinephrineand
norepinephrine.Activationofa-adrenergicreceptors
resultsin
cyclic
AMP
mediated
phosphorilatiOil
ofmembrane
proteins
andincreasesinICaLandinIf.n增endresultiSan
acceleration
oftheslow
diastolic
depolarization.Underrestingconditions,vagaltone
prevails
andvariationsinheart
periodarelargelydependentonvagal
modulation.111e
vagal
and
sympatheticaetivityconstantly
intcracts.Asthesinus
node
iSdchiIl
acetylcholinesterase,the
effectof
any
vagalimpulseis
briefbecausethe
acetylcholineis
rapidlyhydrolysed.
Parasympatheticinfluencesexceedsympatheticeffects
probably
viatwoindependent
mechanisms:a
cholincrgically
inducedreductionofnorepinepbrine
releasedin
responseto
sympatheticactivity,andacholinergic
attenuationofthe
responsetoaadrenergic
stimulus.
Thel浪iiltervalvariations
present
duringresting
conditionsrepresentafine
tuning
of
thebeat-to-beatcontrolmechanisms.Vagal
afferentstimulation1cadstoreflexexcitationof
vagal
efferent
activity
andinhibitionofsympatheticefferentacti“锣.nleopposite
reflex
effectsare
mediated
by
thestimulationofsympatheticafferent
activity.
Efferent
vagalactivity
also
appears
tobeunder‘tonic’restraint
by
cardiacafferent
sympatheticactivity.Efferentsympatheticand删activitiesdirec'tedtothesinusnodeale
characterized
bydischargelargelysynchronous
witheachcardiac
cycle
whichc强be
modulated
bycentral(e.g.vasomotor
and
respiratoryc_,cntres)andperipheral(e.g.oscillation
inarterial
pressure
and
respiratorymovements)oscillators【44].Theseoscillators
generate
—16一
大连理工大学硕士学位论文
rhythmic
fluctuationsinefferent
n伽'al
discharge
whichmanifestasshortand
long-term
oscillation
inthe
heart
period.Analysis
ofthese
rhythmsmaypermit
inferencesonthestate
andfunctionof①thecentraloscillators,②thesympathetic
and
vagal
eferentactivity,③
humoralfactors,and④thesinusnode.An
understanding
ofthemodulatoryeffects
ofneural
mechanisms011thesinusnodehasbeenenhanced
by
spectralanal)'sisofHRV.
111eefferent
vagal
activity
isamajorcontributortotheHF
component,asseeninclinical
and
experimental
observationsofautonomicmanoeuvr龆suchaselectrical
vagal
stimulation,
muscarinic
receptorblockade,and
vagotomy.More
conll"oversialisthe
interpretation
ofthe
LFcomponentwhichisconsidered
by
some嬲amarkerofsympatheticmodulation
(especially
when
expressing
it
innormalized
units)andby
othersasaparameter
thatincludes
bo血sympatheticand
vagal
influences.
Ⅲs
discrepancyisduetothefactthatin
someconditions.associatedwithsympathetic
excitation,a
decrease
intheabsolute
power
oftheLFcomponentisobserved.Itis
important
torecallthat
duringsympatheticactivationthe
resultingtachycardia
is
usuallyaccompanied
byamarkedreductionintotal
pOWer,whereas
thereverseoccursduringvagal
activation.
Whenthe
spectralcomponentsageexpressed
inabsolute
units(ms2),the
changesin
total
power
influenceLFandHFinthesame
direction
and
preventthe
appreciation
ofthe
fractionaldistributionoftheenergy.TIlis
explainswhy
in
supinesubjectsundercontrolled
respirationatropine
reducesbotllLFand
HF[14】andwhyduring
exerciseLFismarkedly
reduced.
Thisconceptisexemplifiedin
Fig.3.3
showing
the
spectralanalysisofHRV
in
anormal
subjectduring
control
supine
conditionsand90
head-up
tilt.Duetothereductionintotal
power.LFappears鹤unchangedifconsideredinabsolute
units.However,after
normalization
anincrease
inLFbecomesevident.Similar
results
apply
tothe
LF/HFratio.
Spectralanalysis
of24.h
recordings
showsthatinnormalsubjectsLFand
HI:%Vessed
innormalizedunitsexhibitacircadian
pattern
and
reciprocalfluctuations,withhJ曲Ca"VMUes
ofLF
inthe
daytime
andofHFat
night.
These
patterns
becomeundetectablewhenasinglespectrum
oftheentire24·h
period
is
usedorwhen
spectra
of
subsequent
shorter
segmentsAreaveraged.Inlong-termrecordings,
theHFandLFcomponentsaccountforapproximately5%oftotalpower.AlthoughtheULF
andVLF
componentsaccountforthe
remaining
95%oftotalpoWer,theirphysiological
correlaresarestillunimown.LFandHFcaninereaseunderdifferentconditions.Anin.eased
LF(expressedin
normalizedunits)isobserved
during
90tilt,standing,mentalstressand
moderateexercise
inhealthysubjects,andduring
moderate
hypotension,physical
activity
and
occlusionofacoronaryarteryorcommoncarotidarteriesinconscious
dcIgs.Conversely,an
—17一
塑墼!些垫望型塑壁婴!墅生唑
increase
inHFisinduc蝴bycontrolled
respiration,cold
stimulationofthefaceandrotational
stimuli.
4.3
Summary
andrecommendat
i
ons
fori
nterpretat
ion
of
HRV
Vagalactivity
isthemajorcontributortotheHFcomponent.Disagreementexists
in
respect
oftheLF
component.Some
studies
suggest
that
LF,when
expressed
in
normalized
units,isaquantitative
markerfor
sympathetic
modulations,other
stodicsviewLFas
reflexing
both
sympathetic
and
vagalactivity.Consequently,theLF/HFratioisconsidcrod
by
some
investigators
to111:Iil-1"01"sympatho/vagal
balanceortoreflect
sympathetic
modulations.
Physiological
interpretation
oflower
frequencycomponentsofHRV(thatisoftheVLFand
ULFcomponentslwarrants
fI】rⅡler
elucidation.ItisimportanttonotethatHRVmcasnr鼯
fluctuationsinautonomic
inputs
totheheartratherthanthemeanlevelofautonomic
inputs.
Thusbotllautonomicwithdrawalandasamratmglyhigh
levelof
sympathetic
inpm
leadsto
diminishedHRV.
一18—
大连理工大学硕士学位论文
5
Changes
ofHRVrelatedto
specific
pathoIogios
Areduction
ofHRVhasbeenrqmrtedin
several
cardiological
and
non-caxdiological
di剐鞫瞎韶.
5.1
gyocardial
infarotion
Dq’ressedHRVafterMI
may
reflexztadecDeasein
vagalactivity
dbectedtothehcan
whichleadstoprevalen∞of
sympathetic
mechanismsandtocaxdiacelectrical
instability.In
the
acute
phase
ofMLthereductionin24·hSDNNis
significantly
relatedtoleftventriculax
dysfunction,peak
creatine
kinase,andKillip
class.
The
mechanism
by
whichHRVis
transiently
reducedafterMIand
by
whichadepressed
HRVis
predictive
oftheneural
response
to
acuteMIisnot
yet
defined.butitis
likelyto
involve
derangements
intheneural
activity
ofcardiaconion.One
hypothesis
involves
cardio—cardiac
sympatho-sympathetic
and
sympatho—vagal
reflexesand
suggests
thatthe
changes
inthe
geometry
ofa
beating
heartduetonecroticandnon.eontractingsegmentsmay
abnormallyincrease
the
firing
of
sympathetic
afferentfibres
by
mechanicaldistortionofthe
sensoryending.
This
sympathetic
excitationattenuatesthe
activity
of
vagal
fibresdirectedtothesinus
node.Another
explanation,especially
applicabletomarkedreductionof
HRV,is
thereduced
responsivenessofsinusnodalcellstoneural
modulations.Spectralanalysis
of
HRVin
patientssurviving锄acute
MIrevealedareductionintotalandintheindividualpowerof
spectralcomponents.However,whenthepowerofLFandHFWascalculatedinnormalized
units.allincreasedLFandadiminishedHFwereobserved
during
botll
resting
controlled
conditionsand24-h
recordings
analysedovermultiple
5min
periods.These
changesmay
indicateashiftofsympatho-vagal
balancetowards
sympatheticpredominance
andreduced
vagal
tone.Similarconclusionswereobtained
byconsidering
the
changes
intheLF/HFratio.
The
presence
ofanalterationinneuralcontrolmechanismsWasalso
reflected
by
the
blunting
ofthe
day.Ilight
variationsoftheRRintervalandLFandHF
spectral
components
presentinaperiodrangingfrom
daystoafewweeksaftertheacuteevent.In
post
MI
patients
witllaverydepressed
HRV.mostof
the
residualenergy
isdistributedintheVLFfrequency
range
below0.03Hz,with
onlyasmall
respiration-related
HF.Thesecharacteristicsofthe
spectralprofilearesimilartothoseobservedinadvancedcardiacfailureoraider
cardiac
transplant,andaxelikely
toreflecteitherdiminished
responsiveness
ofthe
targetorgan
to
neural
modulatoryinputsorasaturating
influenceonthesinusnodeofapersistentlyhigh
sympathetictone.
一19一
堕坠!些垫婪型墅壁咝!螋塑堕
5.2Diabetic
neuropathy
h
neuropathy
aasociatedwithdiabet鹳mdlitus,characterizedby
alterationof锄all
nervefibres.areduetionintime-domain
parameters
ofHRV
sgemsnot
only
to
carrynegative
prognostic
valLiebutalsoto
precede
theelinical
expression
ofautonomicn帆ropathy.In
diabetic
patients
without
evidenceofautonomic
neuropathy,reductiOll
oftheabsolute
poWelt"
ofLFandHF
during
controlledconditionswas
als4)reportedHowever,when
theLFmFratio
wasconsideredorwhenLFandHFwereanalysedinnormalized
units。riosignificant
differencein
comparisontonormalswas
present.Thus.the
initialmanifestationof
this
neuropathy
is
likely
toinvolve
both
efferentlimbs
ofthe
autonomic
nervous
system.
5.30ardiao
transpIantation
A
very
reducedHRVwithnodefinite
spectralcomponents
was
reported
in
patients
wim
arecenthearttranspl锄.111eappearanceofdiscrete
spectralcomponentsinafewpatientsis
consideredtoreflectcardiacre-innervation.11畦sreAnnervation
mayoccur
asearly
as1to2
yearsposttransplantationandis
usually
of
sympatheticorigin,Indeed,the
correlation
betweenthe
respiratory
rateandtheHFcomponentofHRVobservedinsometransplanted
patients
indicatesthatanon-neuralmechanism
may
alsocontributeto
generate
respiration.relatedrhythmicoscillation.The
initialobservationofidentif弭ngpatients
developing
an
allograflrcjectionaccording
tochanges
in
HRV
couldbcofclinicalinterestbut
ng@dsfurtherconfirmation.
AreducedHRVhasbeenconsistentlyobservediu
patients
with
cardiacfailure.Inthis
conditioncharacterized
bysignsofsympathetic
activation,such勰faster
heartratesand
high
levelsof
circulatingcathe
colamines,a
relationship
between
changes
inHRVandtheextent
of1疆ventricular
dysfunction
was
controversiallyreported,Infact,whereas
thereduetionin
timedomainmeasuresofHRVs#ernedto
parallel
theseverityofthe
disease,therelationship
between
spectralcomponents
and
indicesofventrieular
dysfunction
appears
to
bemore
complex.Inparticular,in
most
patients
withavery
advanced
phase
ofthe
diseaseandwitha
drasticreductionin
HRV.a
LFcomponentcouldnotbedetected
despite
clinlcaI
signs
of
sympatheticactivation.Thus,inconditions
characterized
by
markedand
unopposed
persistent
sympatheticexcitation,the
sinusnodeseanstodrastically
diminishitsresponsivenessto
neural
inputs.
5.4
TetrapIegia
Patientswithc栅omc
completehigh
cervical
spinal
cordlesionshave
intactefferent
vagal
andsympatheticneural
pathways
directedtothesinusnode.However,spinal
sympatheticneuronsaledeprived
of
modulatory
controland
in
particular
ofbaroreflex
一20一
大连理工大学硕士学位论文
supraspinalinhibitoryinputs.For
this
reason,these
patients
representaunique
clinicalmodel
withwhichtoevaluatethecontributionof
supraspinal
mechanismsindetermining
the
sympatheticaetivity
responsible
forlowfrequencyoscillationsofHRV.Ithasbeenreported
thatnoLFcouldbedetected
in
tetraplegicpatients.thussuggesting
tlle
critical
roleof
supraspinal
mechanismsin
determining
the0.1Hz
rhythm.Tworecentstudies,however,have
indicatedthatan_LFcomponentcanalsobedetectedinHRVandarterial
pressure
variabilities
ofsome
tetraplegiepatients.WhileKoheta1.attributedtheLFcomponentof
HRVtovagal
modulations,Guzzettiet
al,attributedthesamecomponenttosympatheticactivity
beM:auseof
the
delay
withwhich
the
LFcomponent
appearedafter
spinalsection,suggestingallmerging
spinalrhythmicitycapableofmodulatingsympatheticdischarge.
一2l一
坚墅坐!些竺型堕!型塑驾!墅竺塑哑
6godificationsofHRV
byspecifiC
interventions
日壕rationalefor
tryingtomodify
HRVafterMIstemsfrom
multiple
oh∞rvatiom
indicating
thatcardiac
mortality
is
highcramongpost—MIpatientswhohaveamoredepressed
HRV.111einferenceisthatinterventionsthataugmentHRV
may
be
protectiveagainst
cardi∽
mortality
andsuddencardiac
death.Although
the
rationaleforchangingHRVis
sound,it
containsalsotheinhorent
danger
of
leading
totheunwarranted
assumption
thatmodification
ofHRVtranslates
directly
into
cardiac
protection,which
maynotbethecase.
Tk
target
iStheimprovementofcardiacelectrical
stability,and
HRViSinstamarkerof
autonomic
activity.Despite
the
growing
collSCIISllbthat
increases
in
vagalactivitycarlbe
beneficial,itisnot
yet
knownhowmuch
vagalactivity(oritsmarkers)hastoincreasein
ordertowo“deadequateprotcctioa.
6.1
Beta—adrenerg
iCbIockadeandHRV
T量ledataontheeffectofa-biockorsonHRVinpost-MIpatientsa糟surprisingly
scanty.
Despite
theobservationof
statisticallysignificantincreases,the
actual
changesagevery
modest.However,it
isofnotethata-blockade
prevents
therisein
theLFcomponent
observed
inthe
morning
hours.Inconscious
post-MIdogs,a-blockors
donotmodi移HRV.11le
unexpectedobservation
that,prior
toMI,a-blockade
increasesHRV
only
intheanimus
destinedtobeatlowriskforlethalarrhythmiaspost-MImaysuggest
novel
approachesto
post-MIriskstratifieation.
6.2
AntiarrhythmiCdrugs
andHRV
Dataexistforseveral
antiagrhythmicdrugs.Reeainide
and
propafenone,butnot
amiodarone,worereportedtodecrcasetime-domainmeasuresofHRV
inpatientswith
chomcventricular
arrhythmia.In
another
studypropafenone
reducedHRVanddecreasedLF
muchmorethan
HF.resulting
in
asignificantly
smaller
LFmFratio.A
largor
study
confirmedthat
flecainide,and
alSOencainideandmoricizine,decreasedHRV
inpost.MI
patientsbut
found
nocorrelationbetweenthechangein
HRVand
mortalityduring
follow-up.
Thus,¥omeantiarrhythmicdn{缪associatedwithincreased
mortalitycanreduce
HRV.
However,it
isnotknownwhetherthese
changes
inHRVhave
any
direct
prognostic
significance.
6.3
ScopolamiRe
andHRV
Lowdosemuscaginic
receptorblockers,such
as
atropine
andscopolamine,mayproduce
aparadoxical
increase
in
vagal
efferent
activity,as
suggested
byadecrease
inheartrate.
一22—
大连理工大学硕士学位论文
Differentstudiesexamined
the
effectsoftramdermal
scopolamineOnindices
ofvagal
activitv
in
patients
witharecentMIand
witll
congestive
heartfailure.Scopolaminemarkedly
increases
HRV.whichindicates
that
pharmacological
modulationofneural
aetivity
with
scopolaminemayeffectively
increase
vagalactivity.However,efficacyduringlong-term
treatment
hasnotbeenassessed.Furthermore.10Wdosescopolaminedoesnot
prevent
yentricular
fibrillationduetoacute
myocardial
ischaemiainpost—MIdo举.
6.4
lhromboIysiS
andHRV
Theeffect
of
thrombolysisO/IHRV(assessedbypNN50),Wasreportedin95patients
withacuteMI.HRVWas
higher
90minafter
thrombolysis
in
the
patients
with
patency
ofthe
infarct—related
artery.However,this
difference
wasnolonger
evidentwhentheentire24h
were
analysed.
6.5Exeroise
training
andHRV
Exercise
trainingmay
decreasecardiovascular
mortality
andsuddencardiacdeath.
Regular
exercise
training
isalso
thoughtcapable
of
modifying
theautonomic
balance.A
recentexperimentalstudy,designed
toassesstheeffectsofexercisetrainingonmarkersof
vagalactivity,has
simultaneouslyprovided
informationonchanges
incardiacelectrical
stability.Consciousdogs,documentedtobeat
high
risk
by
the
previous
OP.℃:UITanceof
veotricularfibrillationduringacutemyocardial
ischaemia,wererandomlyassignedto6
weeksofeither
daily
exercise
trainingorcage
restfollowed
by
exercise
training.After
training,HRV(SDNr0
inewased
by
74%andallanimalssurvivedanewischaemictest.
ExercisetrainingCallalsoaccelerate
recovery
ofthe
physiologicalsympatho-vagal
interaction,
asshowninpost-MIpatiants.
23—
堕些堕塑垫型墅堕咝燮塑鲤
70IinicaIuse
of
heartratevariabiI
ity
AlthoughHRVhasbeen
thesubjectofnun'lel"OU8clinicalstudies
mv璐figatmg
a州de
spectrum
of
cardiological
and
non-cardiological
diseasesandclinical
conditions,a
general
oonsalsusofthe
practicaluseofHRVinadultmedicinehasbeenreached
only
intwoclinical
scenarios.D印硝sedHRVcanbeused雒a
predictorofriskafteracuteMI
andasallearly
warningsignofdiabetic·neuropathy.Assessment
ofriskafteracutemyocardialinfarction
Fig.7-1
Cumulativesurvivalofpa6entsalk'rMI.(a)Showssurvivalofpa妇bstratified
according幻
24-hSDNNvaluesinthree
groups
wilh
ont-offpointsof50and100ms,(b)Showssimilarsurvivalcurves
0fp撕∞臼stratified
accordingto24-hHRVumgllJarindexvalueswith
cut-offpoints
of
15
and20
tlni扭.
respectively.(DataofSL
George’s
Post—infarctionResearch
SurveyProgrammed.)
Theobservationthatin
patients
withanacuteMItheabsenceof
respiratory
sinus
arrhythraias
is
associatcalwith锄increase
in‘in-hospital’mortalityrepresents
the
firstofa
large
numbcrof
reports
whichhavedemonstrated
the
prognostic
valueof
assessing
HRVto
idctl呖母hi曲riskpatients.DepressedHRVisapowerfulpredictor
of
mortality
andof
矾hytIlmic
cumplicafiom(e.g.symptomatic
sustainedvcntriculartachycardia)inpatients
一24一
大连理工大学硕士学位论文
following
aente
MI(Fig.7.11.Thepredictive
valueofHRVis
independent
ofother
factors
establishedfor
post-infarction
risk
stratification,suchasdepressed
leftventricularejection
fraction,increasedventrieular-eetopic
activity,endpresence
oflate
potentials.For
prediction
ofa11.causemortality,thevalueofHRVis
similartothatofleftventricularejectionfrac矗on,
butHRVis
superior
toleftventricular
ejection
fractionin
predictingarrhyttmaicevents
(suddencardiacdeathendventrieular
tachycardia).Thispermits
speculation
thatHRVisa
strongerpredictor
of
arrhythmicmortality
ratherthan
non-arrhythmicmortality.However,
cleardifferencesbetweenHRVinpatients
suffering
fromsuddenendnon·sudden
cardiac
deathafteracuteMIhavenot
beenobserved.Nevertheless,thismight
also
berelatedtothe
natul'eofthepresentlyuseddefinition
ofsuddencardiac
death,whiehisboundtoincludenot
onlypatients
suffering
fromarrhythmia-related
deathbutalsofatalfeinfhrdionsendother
cardiovascularevents.nevalueofbothconventionaltime-domainend
frequency-domain
parameters
have
been
fillly
assessed
in
severalindependentprospective
studies,butbecause
of
usingoptimized
cut-offvalues
defining
normaland
depressed
HRV.thesestudies
may
slightly
SensitivityI蛳
Fig.7.2Comparison
ofpositive
predictive
characteristics
ofHRV(solidlines)and
ofcombinations
of
HRVwithleftventrieular白瞅ionfraction(dashedlines)andofmⅣ谢Ih1eftv%tricular
ejecdonfraction
andentopic
countsoil“.hECGs(dottedHnes)usedforidenffication
ofpatientsatrisk
ofl-yearcardiac
mortality(a)and1-year
arrhytlunicevents(suddendeathand/or
symptomatic
sustainedventricular
tachycardia(b)afteracute
myocardial
infarction,(DataofSt.George’SPostinfarction
Research
Survey
Pm乎孤nmed)
龉一
塑堕堂垫塑型塑竺咝塑熊鱼型
Useofnominal24-h
recordingsmay
berecommendedforriskstratificationstudiesafter
MI.Ontheotherhand.theass锱smcntofHRVfrom
short-termrecordingscanbe
usedfor
initial
screening
ofsurvivorsofacuteMI.Suehallassessmenthassimilarsensitivitybut
lowerspodfieityfor
predictingpatients
athi曲riskcomparedto24.h
HRV.Spectral
analysis
ofHRV
ins眦-vivorsofMI
suggested
thattheULFandVLFcomponentscarry
thehi曲est
predictive
value.As
the
physiological
correlateofthese
componentsisunknownandasthese
compon伽【忸correspondto
up
tO95%ofthetotal
power
whichCallbe
easily
assessedinthe
time-domain.theBseofindividual
spectralcomponentsofHRVforriskstratificationafterMI
iSnotmore
powerful
than
theuseofthose
time
domainmethodswhiehassessoverallHRV.
7.1
DeveIopment
ofHfivaftoracutemyocardieIinfarctiorl
The
timeaideracuteMIatwhichthe
depressed
HRVreachesthehigh戚predictivevalue
hasnotbeen
investigatedcomprehensively.Nevertheless,thegenial
collsellsu8isthatHRV
shouldbeassessed
shortlyprior
to
hospitaldischarge,i.e.approximately
1weekafterindex
infarction.Sucharecommendationalsofitswellinto
the
common
practiceof
hospital
managementofsurvivors.
Heartrate
variability
isdecreased
early
aiteracuteMIand
begins
toroT,overwithinafew
weeks;itis
maximally
butnotfully
recovered
by
6to12
monthsafterMI[45,46].
Assessmentofheart
ratevariability
atboththe
earlystage
ofMI(2to3
days
afteracuteMI)
【47】andpre-discharge
fromhospital(1to3weeksafteracuteMI)offersimportantprognostic
information.Heartrate
variability
measured
late(1ye哪afteracuteMIalsopredictsfurther
mortalityf48].Data
fromanimalmodels
suggest
thatthe咧ofHRV
recovery
afterMI
correlateswith
subsequcut
risk【491.
7.2HRVusedformuItivariateriskstratificatiOn
11lepredictivevalueofheartratevariabilityalone
ismodest.butcombination
withother
techniquessubstantiallyimproves
itspositivepredictiveaccuracy
OVeraclinicallyimportant
rangeofsensitivity(25%to75%)forcardiac
mortality
and
arrhythmicevents(Fig.7.2).
improvementsin
the
positive
predietiveaccuracyoverthe
range
ofsensitivitieshavebeen
reportedforcombinationsofHRVwithmeanheart
rate,left
ventricular
ejectionfraction,
frequencyofventricular
entopieactivity,parameters
of
high
resolutionelectrocardiograms(c.
吕p懈∞∞orabsenceoflatepotentials),andclinicalassessment[50].However,itisnot
knownwhichotherstratificationfactorsarethemostpractical
andmostfeasibletobe
combinedwithHRVformultifactorriskstratification.Systematicmultivariatestudies
ofpost
MIfiskstratificationareneededbeforeaconscnsBscanbe
reachedandbeforeacombination
ofHRVwithothervariables
ofprovenprognosticimportanceCanberecommended.
一加一
大连理工大学硕士学位论文
Manyaspects
thatarenotrdevantforunivariateriskstratificationneedtobeexamined:
it
isnotobviouswhetherthe
optimmn
cut-offvaluesofindividual
riskfactorsknownfi,om
univariatestudi,'sare印弘邵d如inamultivariate
setting,Different
multivariate
combinationsarcprobably
neededfor
optimizingpredictiveaccuracy
atdifferent
rangesofsensifi哪.Stepwisestrategies
shouldbeexaminedto
i,k-ntifyoptimumsequences
of
performing
individualtestsusedin
multivariatestratification.
Predictivevalue
ofdepressed
HRV
afteracute
myocardial
infarctionThe
following
facts
shouldbenotedwhen
exploiting
HRVassessmentinclinicalstudiesand/ortrials
involving
survivorsofaci.Romyocardialinfarction.DepressedHRV
iSapredictorof
mortality
and
arrhythmiccomplicationsindependentofotherrecognizedriskfactors.Thereis
ageneral
colrL.qffBSUSthatHRV
should
bemeasured
approximately
1
weekafterindexinfarction.
Although
HRVassessedfromshort-term
reeordingsprovides
prognostic
information,
HRVmeasuredinnominal24-h
recordings
isastronger
riskpredictor.HRV
assessedfi'om
short-term
recoMmgs
may
beusedforinitialscreeningofallsurvivors
ofanacuteMI.
NocurrentlyrecognizedHRV
measure
provides
better
prognostic
information
thanthe
time-domainHRVmeasures
assessing
overallHRV
fe.g.SDNNorHRVtriangular
index).
Other
measures,e导ULF
ofentire24.h
spectralanalysis,performequally
well.Ahi曲risk
groupmay
be
selected
by
the
dichotomy
limitsofSDNN<50msorHRV
triangular
index<
15.For
clinicallymeaningfulranges
of
sensitivity,thepredictive
value
ofHRV
aloneis
modest,althoughitishigher
thanthatof
any
otherSO
farrecognizedriskfactor.To
improve
thepredictiVevalue,HRV
may
becombinedwithotherfactors,However,optimumsetofrisk
factorsandcorrespondingdichotomy
limitshavenot
yet
beenestablished.
7.3AssessmentofdiabetiCneuropathy
Asacomplicationofdiabetesmellitus.autonomic
neuropamy
ischaracterized
byearly
and
widespreadneuronal
degeneration
ofsmallIlCa"Vefibers
of
both
sympathetic
and
parasympathetictractsf511.Its
clinicaImanifestationsareubiquitouswith
functional
impairmentand
include
posturalhypotension,persistenttachycardia,gustatorysweating,
gastroparesis,bladderatony
andnocturnaldiarrhea.Onceclinicalmanifestationsofdiabetic
autonomicnmo咖y(DAN)supervene,theestimated
5-yearmortality
is
approximately
50%【s2].Thus,earlysubclinicaldetectionofautonomic
dysfunction
is
important
forrisk
stratificationand
subsequent
management.Analyses
ofshort-termand/or
long-tcrm
HRV
have
proven
nsefulin
detectingDAN『s3].
Forthepatientpresenting
witharealor
suspect
DANthereagethroeHRVmethodsfrom
whichtochoose:①simplebedsideRRintervalmethods,②long-term
time-domain
measBres,whicharcmoresensitiveand
more
reproducible
thantheshort-termtests,and③
一27—-
fiequency..domaln
analysisperformed
undershort-term
steadystateconditionsandwhich眦
usefulin
separatingsympathetic
from
parasympathetic
abnomaalities.
7.4
Long—term
timedomainmeasures
HRVcomputed
from
24.h
Holterrecordsaremoresensitivethan
simple
bedsidetests(e.
吕Valsavaiilanouvel-,orthostatie
test,and
deepbreathing【54])fordetecting
DAN.Most
expedeliCe
hasbeen
obtained
withtheNN50and
SDSD【55】methods.UsiIlg
theNNS0
count,
wheretheIOWer95%confidenceintervalfortotalcounts
rangefrom500to2000
depending
Oilthe
age,about
halfofdiaberic
patientswilldemonstrate
abnormally
lowcounts
per
24h.
Morcowr.thereisaslTong
correlationbetweenthe
percentage
of
patients
witllabnormal
countsandtheextentofautonomic
neuropathy
determinedfromconventionalmeasures.
Besidestheirincreased
sensitivity,these24-htimedomainmethodsare
strongly
correlatedwithotherestablishedHRVmeasurementsandhavebeenfoundtobe
reproducible
andstableOVertime.Similarto
survivorsofMI.patients
withDANarealSOpredisposedtO
13001"outcomessuchas翱也i鼬deathhutitremai整tobedeterminedwhetherthe
HRV
measuresconfer
prognostic
information
among
di西eties.
7.5
Frequenoy
domain
measures
The
following
abnormalitiesinfrequencyHRV
analysisareassociatedwithDAN:①
reduced
power
in
all
spcctral
bands
which
isthemostcommonfinding,②failuretoincrease
LFonstanding,which
isareflectionofimpairedsympatheticresponseordepressed
baroreeeptorsensitivity;,③abnormallyreducedtotalpowerwith
unchangedLF/HF
ratio,
and④aleftwardshiftintheLFcentral
frequency,thephysiologicalmeaning
ofwhich
needsfurtherelucidation[561.
Inadvanced
naturopathie
smtes,the
restingsupine
powerspectrum
oftenreveals
extremely
lowamplitudesofall
spectral
components
making
itdifficultto
separatesignal
fromnoise.Itis
thereforerecommendedthatanmterventionsuchasstandingortiltbe
included.Anothermethodtoovercomethelow
signal
to
noise
ratioistointroducea
coherencefunctionwhichutilizesthetotal
power
coherentwithoneortheother
frequency
band【57].
一28—
大连理工大学硕士学位论文
8MateriaIsandMethod
In
this
paper
ECGsalerecorded丘砌10
healthysubjectsf5maleand5femalelwith
ageranging丘om
23to89.mECG
datatakenisdividedintofour
groups.20
to30,30to40,
41to701and71to90.111edatahas
beencollectedfromMrr.Bm
Arrhythmia
Database.1k
recordings
weredigitizedat360
s=nples
per
second
per
channel、“ml1-bitresolutionoveTa
10mV
range.Twoormore
cardiologists
indqgendently
annotatedeach
record;disagreements
wereresolvedtoobtainthecompeer.readable
referenceannotationsforeachbeat
(approximatelyl
10.000annotationsina11)includedwith
thedatabase.11”numberof
subjectsin
eachgroupisshowntable8.1.
Tab.8.1Numberofsubjectsinvariousgroups
AgerangeNumberofsubjects
20.30
31.40
41.70
71.90
Theheartrateisanalyzedin
byusing
nonlinear
parameters.
8.1
LargestLyapunovExponent(LLE)
LyapunovExponent丑1Saquantitative
measureofthesensitivedependence011the
initialconditions.Itdefines
the
average
rateofdivergenceoftwo
neighboringtrajoratories.An
exponentialdivergenceof
initiallynearbytrajectories证phasespacecoupledwith
folding
of
trajectories,toensllrethatthesolutionswillremain
finite,is
thegeneralmechanismfor
generating
deterministicrandomnessandunpredictability.Therefore,theexistenceofa
positive五for
almost
allinitialconditionsinabounded
dynamicalsystem
is
widely
used
definition
ofdeterministic
chaos.Todiscriminatebetweall
chaotic
dynamics
and
periodic
signals
Lyapunov
Exponent五ale
oftenused.Itisameasureoftherateatwhichthe
trajectoriesseparateoile
fTorn
other.The
trajectoriesofchaotic
signals
in
phasespacefollow
typical
patterns.Closelyspacedtrajectoriescotlverge
and
diverge
exponentially,relativeto
each
other.Fordyrm捌systems,sensitivitytoinitialconditionsisquantifiedby
the
LyapunovExponent五Theycharacterize
the
average
rateof
divergence
ofthese
neighboring旬喇ed鲥鹤.Anegativeexponentimplies
thattheorbits
approachacommonfixed
point.AZeI'Oexponentmeanstheorbitsmaintaintheirrelativepositions;theyare
on
astable
attractor.Finally,apositiveexponent
implies
theorbits矾Oilachaoticattractor.
29—
旦塑堕!堂垫型型些塑婴!堂塑唑
The
algorithmproposedby
Wolfetalisusedto
Largestlyapunov
exponent(LLE)flora
EEGdata.Foragiven
thetimeseries
gt)for
m
dimensional
phasespace
with
delay
coordinate
t'thatisa
pointon
theattractorisgivenby
{工(f),x(t+f),x(t+(肌一l弘))
Welocate
nearest
neighbor
toinitialpoint
{X(to),X(to+f),…x(to+(m—I)0}.
Anddenotethedistancebgtwoenthesetwopoints∞诋).Atalater
timetI,initial
length
willevolvetolength三7(^).Themeanexponentialrateof
divergence
oftwo
initially
close
orbitsischar∞terized
by2=再1渺M器(tk10一fo智92三’一)
Inimplementation
ofthis
program,thefollowing
setofnummcal
parameters
hasto
be
chosen:
P;{m,f,正Smax,Smin,thmax}wherem
is
the
embeddingdimension,t
is
delay,T
beingevaluationtimeandSmax,Sminarethemaximumand
minimum
separations
of
replacementpointrespectively
andthllla,x.isthemaximumorientation
eTfor.According
toDas
etalanembeddingdimensionbetween5to20andadelay
of1shouldbeChOSenwhen
calculating
LEfor
EEG
data.Inouranalysis
wehavechosenallembeddingdimensionof10
anddelayofl.
Results:Theresultsofthenonlinear
param毗ers
for
four
different
ageyoupare
presented
in
table8.2.
Tab.8.2Varia60n
ofuonliaear
parameters
forvarious
agegroup
normalsubjcctB
Agerange
LLB一—————丽j万————————————西蕊丽.0545
31.400.0760-0.0732
41.700.0928-0.0668
71-90O.1122旬,0926
8.2ResuItforfirst
group
The
fi嘲group
containstwosubjects:
(1)Age23Female
一30一
大连理工大学硕士学位论文
Fig.8.1
ECG
ofsubjectI:age
23
Fig.8.2Largest
Lyapunov
graph
forsubject1:age
23
Fig.8.3
ECGofsubjvct2:age
24
Fig.8.4
Largest
Lyapunov
for
subject2:age
24
111:!::!垒!g堡磐些竺P粤!!壁翌!j!垒垡罂!竖
Age
LLE
22
24
0.2354
0.0545
8.3ResuItforsecond
group
Thesecond
group
contains
twosubjects:
(1)Age32Female
(2)Age39Female
一3l一
坐坚坠!型型竺鱼塑竺型墅堕塑型!塑竺壁唑
Fig.8.5
EC(3
ofsubject3:age
32
Fig.8.6LargestLyapunovofsubject3:age
32
Fig.8.7
ECG
ofsubject4:age
39
Fig.8.8LargestLyapunovofsubject4:age
39
Tab.8.4Variationofnonlinear
parameters
forsecond
group
32
39
O.0763
0.0732
8.4ResuItforthird
group
Thet11ird
group
containsthreesubjects:
(1)Age56Female
(2)Age68Male
(3)Age69Male
一32—
大连理工大学硕士学位论文
Fig.8.9
ECG
ofsubject5:age
56
Fig.8.10LargestLyapunov
ofsubject5:age
56
Fig.8.1
1ECGofsubject6:age68
Fig.8.12LargestLyapunovofsubject6:age
68
Fig.8.13
ECGofsubject
7:age
69
Fig.8.14LargestLyapunovofsubject7:age
69
堡竺坠!堂虫型竺型塑竺驾!塑竺熊堕
:!:21玺苎垒竺!!璺竺些竺鬯!堡翌堑!型墅!坚
Age
LLE
56
68
69
O.0928
0.0927
0.075l
8.5ResuItforfourth
group
Thefourth
group
containsthreesubjects:
(1)Age
73Male
(2)Age81Male
(3)Age
89Male
Fig.8.15ECGofsubject8:age
73
Fig.8.16LargestLyapunovofsubject8:age
73
Fig.8.17
ECG
ofsubject9:age
8l
Fig.8.18LargestLyapunovofsubject9:age
81
—34—
大连理工大学硕士学位论文
Fig.8.19
ECGofsu巧ect10:age
89
Fig.8.20Lm'g§tLyapuaovofsubj∞t10:age
89
Tab.8.6Variationofnonlinear
parameters
forfourth
group
LLE
73
8l
89
O.0926
O.1122
0.1122
Nikhil
Iyengar
et
al,haveapplicd
Recently,Echcverria
et
al,haveapplicd
Congestive
HeartFfilure(CHF)subjects.
DFAtothenormalheartrate
variability
signal.
dctrend既lfluctuation
analysis
tothenormaland
L盯gestLyapunovexponent's(LLE)quantifysensitivity
ofthe
system
toimfifl
conditionsandgivesameaslll'e
of
predictability.This
valuede唧ses、Ⅳitllthe
aging
indicating
thattheheartrate
variability
becomeslesschaoticasthe
healthysubjectgrows
old.
一35
塑坠型墅垫型些塑型些生唑
Heartrate
variability
has
considerable
potential
toa88锶¥theroleofautonomicna'voofl
system
fluctuationsinnormal
healthy
individualsandin
patients
with
variouscardiovascular
andnon-cardiovasculardisorders.HRVstudiesshouldenhanceourunderstanding
of
physiologicalphenomena,theactionsofmedications,anddisease
mechanisms.Large
prospectivelongitudinal
studiesareneededtodeterminethesensitivity,specificity,and
predictive
valueofHRVintheidenfificationofindividualsat
riskforsubsequentmorbidand
mortalevents.AfertileareaofresearchiStouseHRV
techniquesto
explore
theroleof
autonomicnervous
system
alterationsindiseasemechanisms,especiallythoseconditionsin
which
sympathovagal
factorsarethoughtto
play
an
important
role.Recentwork
suggests
that
alterationsinautonomicinnervationtothedevelopingheartmightbe
responsibleforsome
formsofthe
longQTsyndrome.FetalHRVstudiesin
pregnant
motherswiththisdisorderis
certainly
feasibleand
might
be
very
informative.Itwouldbeof
interesttoevaluateHRV
in
variouscircadian
patterns
suchasnormal
day-nightcycles,sustained
reversed
day-night
cycles(evening-night
shift
work),andtransiently
altered
day-nightcycles,suchasmight
occurwith
intemationaltravel.nmautonomicfluctuations
occurringduring
various
stages
of
sleepincludingrapideyemovement(REM)sleephavebeenstudiedin
onlyafewsubjects.In
normal
subjects,theHF
vagalcomponentof
thepowerspectrmn
isaugmentedonlyduring
non-REM
sleep.whereas
in
post.MIpatientswiththisincreaseinHFiSabsent.Heartrate
variability(HRⅥsignalcanbeUSedasareliable
indicaterofstateoftheheart.Itbecomes
lessrandomwiththe
aging(1esschaotic).T11isisevaluated
byusing
nonlinear
parameter
LargestLyapunovExponent.Asadiagnostictool:the
Lyapunovspectrumof
ECG
signals
couldbeconsideredasaf脚independenttooltoimprove
automatic
diagnosis
ofheart
disease.Inordertoallow
theiruseitisnecessaryto
repeat
the
analysisonalarge
ECGdata
set.
一36
大连理工大学硕士学位论文
References
【l】Hon
EH,LenST.Electronic嘲fionsofthefetalheartrate
patteruspreceding
fetaldenth,fiather
observations.AmJObstet
Gynee,1965,87(8):814-826.
[2】Sayers
BM.Analysisofheartratevariability.Ergonomics,1973,I“1):17_32.
【3】3
Penaz
J,RoukenzJ,Vendecw枷HJ.Spectralanalysis
ofsome
spontaneousrhytl衄inthecir圳oIL
In:Drischel
H,TiedtN,eds.Leipzig:Biokybernetik,Karl
Marx
Univ,1968:233_241.
t4】LuczakH,Laming
WJ.An
analysis
ofheartratevariability.Ergonomics,1973,16(6):85--97.
f5】5Hirsh
JA,BishopB.Respiratory
sinus
an'hythmia
in
humans;howbreathingpattern
modulall*heart
rate.Am
JPhysiol,1981,241(10):H620-629.
【6】EwingDJ,MartinCN’YonngILleta1.Thevalueofcardiovascularautonomicf11.1lcliolat砷:10
years
cxpcriellce
indiabetes.DiabeticCare,1985,8(5):491-498.
忉woIfMM,Varigos
GA,HontDeta1.Sinusarrhythnminaonte
myocardial
infarction.MealJ
Australia,
1978,2(2):52-53.
f8】8Akselrod
s,GordonD,UbelFAeta1.Power
spectrum
analysis
ofheartratefluctuation:a
quantitative
probe
ofbeattobeatcardiovascularcontr01.Science,1981,213(4504):220-222.
【9】PomeranzM,Macaulay
RJB,Caudill
MA.Assessmentofautonomicfunctionin
huma∞by
heartrate
spectral∞a】ysis.AmJPhysiol,1985,248(1):151—3.
【10]PaganiM,LombardiF,GuzzettiSeta1.Power
spectral
analysis
ofheartrateandarterial
presgar'e
variabilitiesasaInm'korof
sympatho—vagal
interactioninman
and
COnSCious
dog.Circ
Res,1986.
59r2):178--193.
【1l】KleigerRE,MillerJl',Bigger.ITeta1.DecreasedheartratevariabilityanditsaSsociation
witlIincrensed
mortality
afteracute
myocardial
infarction.AmJCardiol,1987,59(4):256-262.
【12】BiggerJT,FleissJL,Steinman
RC,Rolnitzky
LMcta1.Frequency
domain
measuresofheartStandards
ofheartratevariability
375
periodvariability
and
mortality
atk,r
myocardial
infarction.
Circulation,1992,85(1):164-171.
【13】CysarzD,BettermaonH,vail
leenwen
P.Enlropies
ofshort
binarysequences
inheartperiod
dynamics.AmericanJournal
ofPhysiology-Heart
andCirculatoryPhysiology,2000,278(6):2163—2172.
[141Grassberger
P,ProcassiaLMensuringthe
strangeness
ofastrangeattractor.Physica
D
1983,
9(1-2):189-208.
【15】KantzH,KurthsJ,Mayer-Kreas
G.Nanlinear
Analysis
of
Physiological
Data.1stedition.Berlin:
Springer,1998.
【16]Malik
M.Heartrate
variability.CurrentOpinion
in
Cardiology,1998,13(1):36-44.
【17】RosenstienM,ColinsJJ,De
LucaCJ.A
practical
methodfor
calculatinglargestLyapunovexponents
from
small
da诅sets.PhysicaD,1993。65(1):1
17-134.
【18】KobayashiM,Musha
T.1/ffluctuationsofheartbeat
period.mEE
fa'arlsa商OllS011Biomedical
Engineering,1982,29(6):456-457.
【19】PineusSM.ApproximateellR'opyasameasure
of
systemcomplexity.Proc
NatAcad
Sci
USA,
1991-88(6):2297.2301.
【20】PengCKHavlins,Hausdorf
JMeta1.Fraetalmechanismsandheartrate由mmi蕊.J∞mal011
Elecla'ocardiology,1996.28(1):59-.64.
一37—
坚墅堕!型塑垫婪竺咝塾堕壁墼!堑竺塑坚
【2l】WilsonPW,EvansJC.Coronmyarteryprediction.AmericanJournalofHypertens.1993,6(2):309.313.
【22】DalIlbergST.Genderdifferenceintherisk
factors
forsudden
cardiac
death.Cardiology,1990.
77(2):31-40.
【23】RamaekersD,F.etorH,Aubert
AEeta1.Heartrate
variability
and
heartratein
healthy
volunteem:ls
the
femaleautonomousll日VOU8
system
eardioprotective?EurHearlJ,1998,19(9):13捭1341.
【24】Ryan
SM,Golaberger
AL,PincusSMeta1.Genderand
age-related
differonecsin
heartrate:arcwomen
more
complex
thanmen?JournalofAmerican
CollegeofCardiology,1994,24(7):1700-1707.
【25】Yamasaki
Y,KedamaM,Matsuhisa
M.Diurnalheartratevariability
in
healthysubjects:effects
ofaging
andsexdifferences,AmJ
Physiol,1996。271(I):303.310.
[26】TearyKuoBJ,L缸Tsann,CherylYangCHeta1.Effect
ofagingOlagender
differenc∞in
nCl.1rdtlcontrol
ofheartrate.Am
JPhysiol,1999,46(6):2233·2239.
【27】KevinDavyP,Christoper
I)esollza
A,Pamela
JonesPeta1.Elevatedheartrate
variability
in
physically
active
young
and
olderadult
women.ClinicalScience,1998,94(6):579-584.
【28】NagyEmese,OrvosHajnalka,Bardos
C,yorgy矗a1.Gender-related
heartratedifferences
inhuman
neonates.Pediatric
Research,2000,47(6):778.780.
【29]Galeev
A凡IgishevaL
N,Kazin
EM.Heartratevariability
in
healthysix-to-sixteen恻.old
children.FiziolCheloveka,2∞2,28(4):54-58.
【30】SaulJP,AlbrechtP,Berger
RD
et
a1.Analysisoflongtermheart
rate
variability:methods.I/fscaling
and
implications,Compute,rs
in
Cardiology
1987.IEEEComputerSocietypress,Washingtmg
1988:41皿_422.
[31】MalikM,FarrellT,Cripps
Teta1.Heart
ratevariabilityinrelationto
prognosis
afIer
myocardial
infarction:selection
ofoptimalprocessingtechniques.EurHeartJ,1989,10(1):1060-1074.
【32]MalikM')(ia
R,Odemuyiwa
Oet
a1.Influenceoftherecognitionartefactintheautomatic
alialysis
of
long-term
electrocardiogramsontimedomainmeasurementofheart
ratevariability.MedB湖逝
Comput,1993,31(1】:539-544.
【33】SchwartzPJ'Priori
SG.Sympathetic
llervous
system
andcardiac
arrhythmins.In:ZipesDP,JailfeJ。eds.
Cardiac
Electrophysiology.From
CelltoBedside.Philadelphia:W.B.Satmders.1990:33m-343.
[34】WolfMM,varigosGA,Hunt
Deta1.Sinus
antaytlamia
inaeuto
myocardial
infarction.MedJ
Australia,1978,2(2):5}3.
【35】KleigerRE,MillerJP,BiggerITcta1.Decreasedheartrate
variability
anditsassociationwith
increasedmortalityafteracutemyocardialinfarction.AmJCardiol,1987,59(4):256--262.
【36]OdcmuyiwaO,MalikM,FarrellTet
a1.Comparison
ofthepredictivecharacteristicsofheartrate
variability
indexandIeRveiltrieularejectionfractionforall-causemortality,arrhythmiceventsand
suddondeathafteracutemyocardiM
infarction.AmJ
Cardi01.1991,68(5):434_439.
【37】GreeneHL,RichardsonDW,BarkerAHeta1.Classificationof
deathsafter
myocardial
infaretion龃
arrhythIIIicornonarrhythⅡlic(theCardiacarrhyomiaPilotStud”.AmJCardiol,1989,63(I):i-6.
[39】FeiL,Malik
M。Shortand
long-termassessmentofheartrate
variability
for
postinfarcfion
risk
stratification.In:MalikM,CalllHlAJ,eds.Heartratevariability.Armonk:Futura,1995:34I—_346,
【391Bigger
JTJr,F]eiss几,SteinmanRC
et
a1.Frequency
domainmeasuresofheartpcfjod
variability
and
morlalitya脑myocardialinfal谢oB.Circulation,1992,85(1):16扯171.
一鹞一
大连理工大学硕士学位论文
1401LombardiF,SandroneG,MortaraA既a1.Circadlanvailation
ofspoc仃al
indi∞ofheartrate
veriabilityaftermyocardialinfarction.AmHcartJ.1992.123(9):152I-1529.
【4l】BiggerJT’KleigerRE,Fleiss
JLet
a1.Components
ofheartratevariabilitym目su同duringhealing
of
acBte
myocardial
infar℃60n.Am
JCardiol,1988,61(8):208-215.
[42】wolfA,SwifiJB,Swinney
LHet
a1.DeterminingLyapunovexponent
from
atime
sedee.Physica
D.1985.16(3):285.317.
[43】BiggerJT,Fleiss几,RolnitzkyLMet
a1.Frequency
domainmcaslll"esofheart
period
variability
to
蠲s%sriskIateafter
myocardial
infarction,JAmCollardi01.19933l(1):729_.736.
【44】LevyMN,SchwartzPJeds.Vagalconlroloftheheart:Experimentalbasisand
clinical
implications.Armonk:Future,1994.
【45】LombardiF,SandroneG,Mortara
Aeta1.Circadianvariation
ofspectral
indicesofheartratevllriability
aRer
myocardialinfarction.AmHcart
J,1992.123(7):1521—1529.
【蛔BiggerJT,KleigerRE,Fleiss几eta1.Componentsofheartrate
variability
m∞∞red
duringhealing
of
acutemyocardial
infarction.AmJCardiol,1988,61(3):208__215.
【47】CasoloGC,StroderP,Signorini
Ccta1.Heartrate
variabilityduring
theacute
phase
ofmyocardial
infarcfion.Circulation。1992.85(8):2073_-2079.
【48】BiggerJT,Fieiss
JL,Rolnitzky
LMetal。Frequencydomainine∞JA'es
ofheart
periodvariabilityto
assessrisk1ate
aftermyocardial
infarction.JAm
Collardiol,1993。2I(5):729-736.
【49】AdamsonP
B,Huang
M
H,Vanoli
Eeta1.Unexpectedinteractionbetween
a-athenergic
blockadeand
heartratevariability
beforeandafter
myocardial
infarction:a
longitudinal
study
in
dogs
athi曲and
lowrisk
forsudden
death.Circulation,1994.90(3):976_982.
【50]CammA3,FeiL.Riskstratification
followingmyocardial
infar.ion:Heartrate
variability
andothlT
riskfactors.In:Malik
M,CammAJ。eds.Hearlratevariability.Allllonk:Futufa,1995.369-392.
【5l】Bannister&AutonomicF,AtextbookofclinicaIdisordersof血eautonomicne/'voussystenl.Oxford,
NewYork:Oxford
University
Press。1988.
【52]EwlngDJ,CampbellIW,ClarkeBF.mnatul'al
history
ofdiabeticautonomic
neuropathy.QJ
Mcd.1980.193(49):95-108.
【53】IyengarN,PengCKet
a1.Age-related
alterationsinthe
fi'actal
scaling
ofcardiacinterbeatinterval
dynamics.AmJ
Physi01,l996,271(4):1078.1084.
【54]BonnemeierH’Uwe
W
KBrandesAeta1.Circadian
profile
of咖dlacautonomicrlerVOnsmodulation
in
healthysubjects:DifferingEffectsof
aging
andGenderonheartratevariabili哆Jonmalof
Cardiovascular
Electrophysiology,2003,14(8):791-799.
【55】MalpasSC,MalingTJB.Heartrale
variability
and
cardiacautonomicfunctionin
diabetes.Diabetes.1990,39(8):117卜l181.
【56】BellavereF,Balzanil,De
MasiGeta1.Power
spectralanalysisofheartratevariation
improves
assessmentofdiabeticcardiacautonomicneuropathy.Diabetes,1992。41(5):633.枷.
[571BianchiA,BonternpiB,Certuti
Scta1.Spectral∞alysisofhcanrate
variabilitysignal
andrespiration
indiabeticsubjects.MedBinl
EngComimt.1990,28(1):205-211.
一39—
攻读硕士学位期间发表学术论文情况
【1】王兴元,安妮,海登.Heartratevariationinnormalsubjectsamong
various
age
刊物类型:网刊
主办单位:大连理工大学
论文章节:第一章到第八章
【2】马洪连,海登,安妮.IntelIPPforMPEG-4media
player
developmentinembedded
刊物类型:网刊
主办单位:大连理工大学
大连理工大学硕士学位论文
AcknowI
edgement
Firstoffll1want
tothank
mysuper
visorPmf器sorWangXing
Yuanf.orKs
g他athelp
in
allt11e∞yen.Rwouldbe嘶ssibleformet0finish
mythcsis
witlloutKs
hdp.hl
all
也器eyenhehas
hclped
metowalkⅡⅡou曲thepalllofk∞wledge'research.msbroadand
deep虹owledgci11
every
fidd
hasmade
my
r韶earcheasi盯.
Iflsow龇ttothanka11thed船smat鼯of
my
labspecify
Gulma.Shehas
helped
me
throughomtlle
year
to∞lVeV商。璐problcIIlsofmy
rese盯ch.Witllomhcr
hdp
itwofl血'tbe
posfible
formet0f.o皿at
my
tll器is.She
h勰al∞hclped
me、ⅣithChin锶e
l龃guage.
1wantt0thanktllisuniv呻forpro“ding
mechanccto
study
irI驰chaIli∞
aI“∞伽∞‘百Veme
oppommitytoenrich
mylmowlodge.
At
1ast1wanttomank
myfhther.mother,my
sistersand
b∞merfbflheir
con吐册d璐
support
throughout
mystudy.Whenev盯Igct
intomy仃oⅡbletheywe∞flways
thereto
help
me觚dextelldtlleir
loving
handstowards
me.Today
I'm
IIlissmgmy
mothersomuch
tllatMe
leftttlis
worldbefore
seeing
herli砌edaught盯being鲫uated.Thisw嬲h盯dream.Sheis
the∞urcc
ofmyinspimtionforstIldyin孚
1w锄ttomankalI
myt髓chers,cl勰蛐at鼯,伍觚dsandfcla:civ铭fortlleirg僦helpin
a11
日1雠year.Withouttlleir
help
IcouliIn’tbeherc.
一4l一