JapanSociety

Japan Society of Physiological

of Physiological Anthropology
Anthropology

,JollTTialof
PHYSIOLOGICAL
ANTHROPOLOGY

CardiacAutonomic ControlImmediately after Exercise

inFemale Distance Runners

Suzukii),Katsuhiro
Shigehiro Sumi2) and Michitaka Matsubara3)

ijepsp
RieultyofScience
& 7lechnologrl
Meijo University
SchoolofJnternatienalLibernl
Studies,
Chutvo Uhivensity
Department ofNkurotagnl
Mrgo.va & SPorts
CityRehabilitation Center

Abstract The purposeof this study was to examine the In contrast, HR

system. recovery after exercise islinkedto the
timing of recovcry of the high-frequency (HF)component of inhjbition
of sympathetic nerve activity and parasympathetic
heartrate variability (HRV)after exercise in fernale distance reactivation (Arai et al., 1989;Imaiet al., 1994;Javorka et al.,

runners using frequency analysis of HRM ]beenty-two female 2003;Kannankeril et al.,2004;Pierpont et al.,2000).Imaiet

young (n;1I; YG) and middle-aged (n=11; MG) distance al, (1994) reported that the HR recovery 30 seconds
runners participated in thisstudy The two groupsperformed irnmediately after exercise was strongly influenced by the
incremental cycle exercise with progressive intensity until vagus nerve activitM not by thedecrease in sympathetic nerye
exhaustion. The RrR intervals were processed by the activity. Kannankeril et al. (2004)
also reported that cardiac
maximum entropy methed fordetermination of HF power on parasympathetic nerve activity increasedone minute after
successive 7-secondsegments of 70 seconds of the recovery exercise. The results of these previous studies (Imai et al.,

period.In the YG, theHF power of the second 7-secsegment 1994;Kannankerilet al., 2004) suggest that reactivation of
showed significantly higher values than the 7sec before parasympathetic nervous activity after exercise occurs to a
cessation of exercise ip<O,O05), whereas the MG exhibited considerable degree during the early recovery period
significantly highervalues inthe thirdsegment (p<O.O05), The immediately after exercise, However, itdoesnot seem thatthe
YG indicated significantly higherHF power than the MG in timing hasbeenfuIIy examined.
the fifth segment (p<O.O045), These findings suggested the Heartrate variability (HR]V) iswidely used as an indexto
occurrence efparasympathetic reactivation at an car]ier period assess cardiac autonomic nervous activity noninvasively Many
compared to the previous findings. Multipleinfiuences of researchers havc exarnincd the autonomic nervous activity

various factors including the subjects' characteristics to HF duringand after

exercise using frequencyanalysis of HRV
recovery were suggested. However,the detection ofthe timing (Arai 1989;
et al., Perini et al., 1990; Yamamoto et al., 2001).
of the duration
HF recovery despite of sharp change in HR Sinceconventional methods, such as thefast Fouriertransfbrm
indicatedthatHRV was an effective evaluation technique for (FFT)and autoregressive (AR) techniques, need stationary
determination
of autonomic control imrnedi
ately after exercise. data,thcy cannot be applied to the duration of sharp change in
J P]rysiolAnthmpol 27(kE):325-332, 2008 http:Nwwvvjstage. HR immediatelyafter exercise (Goldberger et al,, 2006).
as the rnaximum
jst,gojpfbrowseljpa2 Recently, newer methods, such entropy

[DOI:
10.2114ijpa2.27.325] method (Macor et al., 1996; Murasato et al., 1998; Sumi et al.,

2006)and the complex demodulationmethocl (Hayano et al.,

Keywords: heartrate recovery, autonomic nervous system, 1994),which do not depend on stationary data,have been
maximum entropy method developed. These newer methods make itpossible to assess
autonomic nervous activity based on a short-term R-R
intervat, Sincethe lowerfrequencylimitofthe high-frequency
Introduction (HF) component of HRM which is the index of cardiac
parasympatheticnervous activity (Task Force of the European
The heartrate (HR)is antagonistically controlled by the Society of Cardiologyand the North American Society of
sympathetic the autonomic nervous
activity of system and the Pacing and Electrophysiology,1996),is O.15Hz, the required
parasympathcticbranchof the vagus nerve. The increase in shortest time-series data length to assess HF power is
HR duringexercise isbroughtabout by the inhibitionof the appreximately 7 seconds (1seclO.15), By analyzing a 7-second
vagus nerve and the activation of thesympathetic nervous segment using the newer methods, itispossibleto evaluate HF

NII-Electronic
NII-Electronic Library Service
JapanSociety
Japan Society of Physiological
of Physiological Anthropology
Anthropology

326 Timingof H RV Recovery after Exercise

powerduringsharp change inHR. 2. ExeTzriseprotoeol

Exercisetrainingincreases parasympathetic tone (Levy et An incremental
cycle exercise was carried out in a quiet
al., 1998), and highaerobic capacity isassociated with fast HR room maintainedat constant temperature (22-240C). The
recovery after exereise (Darret al., 1988).Distance running is subjects performed the test in upright positionon an
one of the typicalendurance exercises thatirnproye aerobic electronically brakedergometer (Well Bike BE-360, Fukuda
capacity. Nevertheless, HI(Svi immediatelyafter exercise in Denshi,Tbkyo,Japan), At first,
the subjects rested in a supine
distance runners isnot well understood. In addition, very littlc positionfor20min, Thcn,the subjects rode on the ergometer,
is known about the autonomic control of post-exercise heart Seatand handlebarheights were set fbreach subject and kept

rate intrainedfemale subjects (Duet al,,2005), The possible constant during Followinga 3-minrest, the subjects
the test,
factors
related to post-exerciseHR recovery are age and perforrncd a 2-min warm-up pedalingat OXM and then
athletic level(Buchheit and Gindre, 2006; Darr et al., 1988; exercised with progressiveintensityuntil a subject could no

Lipsitz et al., 1990; Shannon et al., 1987; Yamasaki et al., longermaintain the pedaling rate (volitional exhaustion). The
1996). Darr et al. (1988) reported that trajned subjects, work load was increased by 20W･min-i. The pedaLing
irrespectiveof age, demonstrated a significantly fasterHR frequency was set at 50 rev･ min-T. Irnrnediately
after cessation

recovery than untrained subjects, Increasingage rcsulted in a of exercise, the subjects were instructed
to stop pedalingand to
slower decline of heartrate after exercise (KDstis et al,, 1982), stay on for2min, No
the ergometer attempt was made to
Therefore,post-exerciseHR recovery of a young group with control breathing
frequency
duringand after the test,
Alltests
high athlctic levelsis infiuenced by the status of age and were conducted in a centrolled environment with proper
athletic leve]. It is expected thatthe post-exercise HR wM experimental design.
During the tests,all subjects were
recover earlier than thatof older runners. The significance of mentally andphysically fit.
this study is to findthe timing of parasympathetic reactivation Oxygen uptake was obtained using an Oxycon-a automatic

after exercise using a newer method and to obtain newer online breath-by-breath system (Jaeger, Mljnhard by
findings on the cardiac autonomic modulation in female Netherlands). The system isadapted forexpired and inspired
distancerunners, volumes, Breath-by-breath datawere averaged to provideone
This study aims to assess the cardiac autonomic nervous datapointfbreach 30-secperiod.This was subjected to a
activity immediately after exercise in femaledistance runners three-way calibration process,involving
a flowvolume sensor,
by using the newer method of frequencyanalysis of HR]l Itis gas analyzer, and de]aytime calibration. The flowvolume
hypothesized thatthe timing of recovery of HF power can be sensor calibration ensures that a measuring system of thc
distinguishedby analyzing a shorter time segment than was Oxycon (consisting the amplifier, TripleM and pressure
of
done in previous studies (Imai et al., I994; Kannankeril et a],, transducer)is functioning correctly. A calibrated 3-L syringe
2004), even in the change in HR with time immediately after connected to the [[tiple V assembly was used, A series of six
exercise. Funherrnore, the timing of recovery of HF power in complete pumps of the syringe was repeated until thepercent
highlytrained young femalerunners isfaster than in well- difference betweenthecurrent and previous volume calibration
trained middle-aged femalenmners. was less than 1%. The gasanalyzer and delaytime calibration
involvedan automated calibration procedure (Carter and

Metheds Jeukendrup, 2002).The eriteria fOrachicvement of maximal

oxygen uptake (V02..) includedall thefo11owing measures:

1, Subjects 1)Leveling off of oxygen uptake despitc a further progression

twenty-twohealthy young (n=:11) and middle-aged (n=11) ofthe cxcrcise toad(<150mVmin increasein oxygen uptake),
femalesubjects participated in this study, The young group 2) finalrespiratory exchange ratio of >1.1, and 3) visible
(YG)was composed ofhighly trainedcollegiate longdistance exhaustion. In this studM considering the diMculty of
runners. The middle-aged group (MG)ineludedwell-trained estimating maximal HR (Gellish
et al,, 2007), the age-
regionalrurmers. predicted maximal HR was not adopted as a criterion for
No subjects were en any regular medication, including V02mn..Breathing rate (BR) was measurcd continuously,
hormone replacement therapy. No one was diagnosedwith Throughout the test, determination
l2-lead of
diabetesand hypertension, and none were smokers, Thc electrocardiogram (ECG)
was digitized
at 4kHz
continuously
subjects were asked to refrain from strenuous exercise and (ML-5000, Fukuda Denshi,Tbkyo,Japan). HR was measured
caffeine consumption forat least48 hoursand 24 hoursbefore and r ¢ corded with ECG monitoring. Cuffbloodpressureswere

the test,respcctively. Thc subjects wcre informcdabout the obtained every indirectautomatic manometer
minute with an
aim of the study and itspossible risks, and all gave their (STBF-680E
CollinDenshi, Aichi,Japan).
written infbrrned consent beforeentering thc study. This study

protocol was approved by the Human Subjects Committee at 3. Measunementqfeareliacautonomicactivity

Nagoya CityRehabilitation & SportsCenter. In order to obtain theRrR intervalof each subject, ECG
(CM5)was recorded (LRR-03, GMS, Tokyo, Japan) over a
period extending from the supine rest up to the end of

NII-Electronic
NII-Electronic Library Service
JapanSociety
Japan Society of Physiological
of Physiological Anthropology
Anthropology

Suzuki, S et al.J f'hysiolAnth,opoL 27.'325-332, 200S 327

the recovery. The measured R-R interval time series data considered thattotalpower immediately after exercise was as
were then transferred by an AD converter (AD12-8(PM),lowas thatduringcxercisc and confirrned thecentral frequency
CONTEC, Osaka, Japan) from the ECG to a personal in the power spectrum. Furthermore, BR during the testwas
computer (XexIO VGN-T91PSY SONYI Ibkyo, Japan) with a measured by automatic an online breath-by-breathsystem. The
sampling rate at 1 kHz. Based on theR-R interval
timeseries peak frequency
range of HF power was determined
basedon
data,a time series analysis system (MemCalc,
Suwa TTust, the results ofthese caleulations.

Tbkyo, Japan) was then used to obtain the power spectral

distribution
ofHRM 4, Statisticalanalysis
MemCalc is a computer programused to calculate spectrum No subjects were eliminated from the analysis by virtue of
based on the maximum entropy method. The procedure thc irregularitiesin the ECG during the recovery period. HR
overcomes the disadvantage of the eonventional spectral was calculated using the
average R-R intervalsin each
analysisin the frequency domain,such as poor rcsolution and segment. Characteristics, HR, and BR were expressed as
insuMciency for estimating short time-series datain FFT and mean ± SD, HF was expressed as median (IS`-3rd
quartile),A
AR methods (Ohtomoet al,, t994; Radoski et al,, 1976), Student's
unpaired t-test was used for the corrrparison of
MemCalc can also detect
properfrequencyeven from dataof characteristics of subjects, Serialchanges inHR and BR werc
the lengthcorresponding to only period (Ohtomo
one et al., evaluated × 11 (time
by 2 (groups) points)repeated measures
1994; Ohtomo et al.,1995; Tsuchida et aT., 1998). ANOVA. Analyses that did not meet Mauchley's sphericity
In order to filter
ectopic beatsin thissystem, we extracted criteria were interpreted using the Greenhouse-Geisser
the R waves, First,we recognized thc wave patterns,
and then correction fbrthe inflated
risk of a type I error (Ludbrook,

removed thevalues, which were not adequate as human R-R 1994),Post-hoctestingusing a Bonferroniadjustment was
intervals (less than 272msec, more than 3000msec).If the used to assess specific differences between 7sec before
ectopic beatswere mixed inbetweennormal R waves, they cessation of exercise and 70see of the recovery period and
were removed and made primaryR-R intervals, In thisstudy, between-groupcomparisons at the same time points,The
the total numbers of R-R intervals removed from each results ofLilliefbrs testrevealed thatno normality concerning
subjeet's original data were 22,5± 18,82 in the YG and HRV parameterswas obtainea so that non-parametric tests
6.5± 7.61in theMG. The outliers of theoriginal datawere were used. As regards resting value and each segment ofHF
O.9± O.72% in the YG and O,2± O.24% in theMG. Sincethe powerbetween7 sec befbrecessation of exercise and 70 sec of
R-R interval time series datawere unequally spaced they were the recovery periodthe comparisen betweenbothgroupswas
made even. The function of MemCalc isto interpolate datato performedusing a Mann-WhitneyU test.A Wilcoxon signed-
even it out. That is,it connects anteroposterior data dots raTik testwith Bonferroni correction was employed in the pair-
linearly and sets the height of the dotto where the straight line wise comparisons between7sec beforecessation of exercise
crosses the vertical line. The latter was taken at the observed and each successive time pointin each group.The significant
tirne and observed value. We resampled the data at even differencebetween the proponionsof the tenth 7 sec segrnent
intervals fromthe R-R interval time series data,which became of HF (HF6].7e) to the resting value in both groups was
a continuous function. The resampling frequencyat rest was 1 assessed using a chi-square test for independence.The ]evelof
Hz,and thedataduringexercise and recovery was 3.33Hz, sigriificance was set at p<O,05, Data were analyzed using
The resting value was calculated as theaverage value taken StatVicwJ5.0and SPSS 15.0JforWindows.
oyer a 3-minperiodin sitting position on the cycle ergometer

prior to exercise. Tb elucidate the detailed structure of HRV Results

duringrecovery, segment time series analysis was carried out
Previousstudies describedthe increasein parasympathetic The characteristics of the subjects are presented in Rible1.
nervous activity between30sec and one minute immediately The subjects of the YG were significantly younger than the
'
after exercise (Goldberger
et al., 2006; Imai et al., 1994; MG subjects Cp<O.OOOI). V02.. was significantly higherin
Kannankeril et al., 2004), Therefbre,we divided the original the YG (p<O.OOI). HR at rest was significantly lower in the
time series from 7 sec before cessation of exercise to 70 sec of YG (p<O,Ol). HF power at rest of the YG was sigriificantly
the recovery periodintoa sub-series of ten 7 see segments. highcr than for the MG Cp<O,Ol), Training experience was

Thc HF power spectrum (ms2) was set as the sum of the significantly longer in the MG ip<O,OOOI), but current
powerfromO,15to O.4Hz at rest.The HF power fromthe peak training status was significantly higherintheYG (P<O.OOO1).
of exercise to the end of the recovery period was set between Therewas a significant groupsXtime points interaction fbr
O.IS and 1.0Hz,dependingon the subjects. The upper limit of HR [F(2.390, 47.794)=26.100;p<O.OOI; Greenhouse-Geisser
the frequency range was altered by respiratory frequency correction applied; Fig.1],The HR in bothgroups decreased
becauseHF power was infiuenced by respiratory activity immediately after exercise, Inthe YG, HR of the second 7-sec

(Hirsh and Bishop, 1981). During high-intensity exercise, the segment (HR7-m) indicated a significantly lowerlevelthan
central frequency responded to respiratory frequcncy even if HRp,ak (p<O,O05), whereas, inthe MG, HR,4n, demonstrated a
tota1power was low (Sumi et al., 2006), In this studM we significantly lower level than HRp,,k (p<O･O05)･ HR,,,,kWaS

NII-Electronic
NII-Electronic Library Service
JapanSociety
Japan Society of Physiological
of Physiological Anthropology
Anthropology

328 TimingofHIW Recovery after Excrcise

Table 1. Churacteristicsefsubjects

YG(n=:11) MG {n 11)
==

Age (yrs)Mass(kg)Height 19.g± O.87**** 57.1± Z17

47.3± 624 47.0± 4,07
(cm) 160.9± 5.18** 153.0± 528
V02.,.(m!lmirVkg) 57.6± 4.91*** 42.0± 11.20
HR at rest (bpm) S6.5± 7.31** 66,5± 9.00
HF pewerat sitting rcst (ms2) 650,8(300.92-750.8D** ISO.6(100.83-250.50)
Trainingexperience (yrs) 7.8± O.S7 23.3± 4.10****
Traming frequency Csessions/week) 6.9± O.3o**s* 4.0± 1.10
Trainingtime (minfsession) 15S.S± 32,36**** 58,2± 26,77

(1 YG, young group;MG, greup;VO!..., HR, heaTtrate;

St
Milues are means ± SD or median quartilc-3nd
quartile); middle-aged maximal oxygell uptake;

IIF,high-frequency fp<O.05, significant differences betweentwo groups,

*fp<O.Ol, "*p<O,OOI, ****p<O,OOOI
eomponent ofheart ratc variability.

lgo (496.5f650.8)
inthe YG 5,8% (8.71150.6)
and intheMG. This
differenccwas statistically significant (x2=84,OO1, p<O.OO1).
160
In the comparison betweenboth groups,the YG indicated
significantly highervalues after HF2s 3s ip<O.O045).
A 140ge
Therewas a significant gToupsXtime points interaction fbr
12ooEE BR [F(2.949,58.979) 29,244;p<O,OO1; Greenhouse-Geisser
==

correction applied; Table3],BR after exercise in bothgroups

loodi demonstratedsimilar values to HF, In the YG, BR of the
go
second 7-secsegrnent (BR7i4) showed a significantly lower
value than (p<O.O05).
BRp,,k SubsequentlM the reduction
60 was (p<O.O05).the MG, BRT4.2i showed
continued In
a significantly lower yalue than BRp,,k (p<O･O05)･
40
SubsequentlM the reduction was continued as well (P<O.O05),

g ., ,s ptA>sg y'wO SS crmp In the bctween both groups, there was no

comparison
significant difference BRp,ak,
The in
YG indicated
significantly
(sec)
Time inRecover", lowervalues in BR7.i4(p<O.O045). Thereafter,
lowervalues
Fig,1. Mean and standard deviationof hcart ratc at peak exeTcise and were until BR63"o(p<O,O045).
similarly demonstrated
first70 seconds ef recoveryforthe young group <n=11) and middle-
aged group(n=1 1),The results from the repeated-mcasures two-way Discussion
ANOVPL for the heartrate data demonstratethat there was a
significant groupXtime interaction (F (2,390, 47.794)=26.100;
The major findings ofthis study are as fbllows]1) We were
p<O.OO1;Greenhouse-Geisser correction applied). Alpha lcvcls were
adjusted by the Bonferroni technique for comparisons at ten time able to comprehend clear)y the change inHF power duringthe
points forspecific diffCrcnee betweenpeak and 7esec of the recovery duration ofthe sharp change in HR immediately after exercise
period(i.e., *p<O.05/1O=O.O05) and eleven timepoints forbetween- using the newer frequencyanalysis of HRV based on the
groupdifferences(i.c.,'p<O.05,'11=O.O045). maximum entropy method and 2) the timing ofthe recovery of
HF power immediately after exercise inthe YG was faster than
higher in YG
sigriificantly (p<O.O045). A significant in the MG. As fhras we know; this study reveals for the first
differencebetween groups was revealed in HR2B-3s timc the detection ofHRV recovery immediately after exercise

(p<O.O045). Thereafter, the YG similarly demonstrated by anaLyzing theshortest segment oftime.

significantly lower values Cp<O,O045), HR63.7o demonstrated
91.5± 15.34 (bpm) in the YG and 119,5± 20,83 (bpm)in the HRVindices immediately qfier exercise
MG, The ratio of HR6ye to HRp,ak was 55.1± 8.48% in the In this studM HF power inthe YG started to rise between 7
YG and 76,6± 9,74% inthe MG. and 14 seconds after exercise. On the other hana HF power in
The change with time in HF power isshown inTable 2. In the MG increased between14 and 21 seconds. These results
the YG, the HF power of the second 7-secsegment (HF7.i4) suggested the occurrence of parasympatheticreactivation at an
showed a significantly highervalue than 7 sec befbrecessation earlier periodcempared to the 30 seconds or one minute after
of cxercise Cp<O.O05). Thereafter,significantly highervalues exercise described in previous studies (Goldberger et al.,2006;
were similarly demonstrated (p<O.O05). The MG exhibited Imai et al., 1994; Kannankeril et al., 2004), Immediately after
significantly higher values after HF,4-2iCp<O.O05). The exercise, the autonomic nervous activity beginsto change

proponionsof HFG]"o to the resting value were 76.3% (Kannankeril and Goldbcrger, 2002).Thatis,duringexercise

NII-Electronic
NII-Electronic Library Service
JapanSociety
Japan Society of Physiological Anthropology
ofPhysiologicalAnthropology

Suzuki,S et al.JPh.vsiolAnthiepol,
27:325-332,2008 329

Ta'ble2.High-frcquencycomponentefheartratevariabilityimmediatelyafterexercise

Timesegment
HF power (ins2) pvalue

Csec) MG
YG(n=11) Ma(n=11) vs. Peak {YG}vs. Pcak (MG)YG vs.

Peak 1.8(1,354,44) 1.5(O,70-6,OO) O,467

O-7 4.1(2.47-5,06) 1.8(O,876.0S) O.269O.OOI*O.OOI*O.O02*O.OOI*O.OOI*O.OOI*O.OOI*O,OOI*O
O,080O,189O,O02fio.oolgO.OOIfio,oolgO.
O,166
7-1414-2121m2828-3S3542424949-5656-6363-70
6.4(2,649,S6) 1.7(1,35-6,7S) e,os3
Z4(3.81-20,37) 2.3(1,54-6,Ol) o,olg
21.9(5.IS-93,64) 2.8(1,S4-6,77) O.027
78.S(8,64]46.8S) 4.9(1,93-8,60) O.O03'1'
272.3(12.09-340.10) 5.4(1.90･-15.03) o.oe2t<O.OOIt<O.OOIt<O
271.4(20.9g-666.84) 6.4(1.5g-10.05)
264.7(46.10-664.70) 6.9(4.87-14.90)
352.3(12121-671,84) 7,O(2,49-1629)
496,5(172.98-80e,68) 8,7(2,30-13.95)

Nh]ues are median (1SLquartile-3rd

quartile);HE high-frequcncy component of hcart rate variability; YG, young group;MG, middle-aged group,p values
are from Wilcoxen signed-raiik testwith Bonferronicorrection forthe pairwise comparisons betwccn7 scc beforecessation of exercise and each (Peak)
suceessive time pointand Mann-WhitneyU testwith Bonferroni correction forthe cornparison bctweenbothgroups.* Significant difference fromPeakin
the young group (p<O.05110=O,OOS), g Significant
differenee from Peak in the midd[e-aged group O)<O.05IIO=O.O05). t Significantdifferencebetween
two gToups (p<O.05fl l=O.O045).

Table3. Breathing rate immediately after exercise

Time segment
Breathing rate (tirneslmin) p value

(sec) YG(n=11} MG(n=11) vs. PeakCYG)vs. Peak(MG)YG vs. MG

Peak 52.6± 5.95 47.6± 8.14 O.lt8

O-7 48,g± 5,04 46,9± 6.03 O,602<O,OOI*<O,OOI*<O.OOI*<O.OOI*<o.oel*<O.OOI*<O.OOI*<O.OOI*<O.
1,OOO O.417<O.OOIt<O.OOIt<O.OOI"
7-14I4r2121-2828-3535-42424949-5656-6363-70
29.1± 3,13 46,5± 7,61 O,744<o,oeli<O.OOI"<o.oeli<o,oolg<O.o
27,7± 2,82 3Z9 ± 3,75
26,3± 4,07 35.8± 3.37
24.9± 3,13 33.8± 3.37
24.6± 3.45 32.2± 3.40
24.1 ± 2.g2 31,4± 3.37
23.8± 2.51 30.5± 324
23,2± 188 29,7± 3.31
23,O± 1,68 28,7± 3.13

Valuesare means ± SD; YG, youmg group; MG, middle-aged group. The results frem the repeated-measures two-way ANOVA for the bTeathingrate data
demonstrated thatthere was a signthcant groupXtime interaction(F (2.94Y, 58.979)=29.244; p<O,OOI; Greenhouse-Geisser correction applicd). Alpha
levelswere adjusted by the Bonferreni tcchnique forcomparisons at ten time points between7sec beforecessation ofexeTcise (Peak)
forspeeific differenee
in young group(i,e,, group(i.c.,
gp<O.OS/10=O.O05), and eleven time points
'p<O.05AO=:O.O05)
and 70secof thc rccovery period the and rn{ddle-aged
forbetween-groupdifTerences(i.c.,tp<O.OSIII=O.O045).

the decreascin parasympathetic nervous activity and the nervous system reactivation appeared after HR reached lower
increase
in sympathetic nerve activity result in therise ofHR. levels,
In contrast, after exercise, HR is decreasedby interaction We think that the methods and segment of analysis are
betweenthe inhibition of the sympathetic nerve activity and different from the findingsof these studies. The previous
parasympathetic reactivation (Goldberger et al., 2006; Pierpont studies examined a segment datasetting after exercise between
and Xloth,2004), However, there isno agreernent of findings 15 and 60 seconds (Goldberger et al.,2006;]mai et al., 1994).
concerning the timing of change of the autonomic nervous Tn the current studM HR decreasedup to 55.1% of HR.,,.in
system. The previous stud{es stated that vagus nerve activity only 70 seconds after exercise inthc YG and 76.6% intheMG.
was reactivated immediately after exercise (Arai et al., 1989; Therefbre, inorder to assess the cardiac autonomic modulation
Imai et al., 1994) and was associated with the exponential mechanism, which changes from moment to moment, data
decline ofcardiac activity (Nishime et al,, 2000;Perini et al., lengthfbranalysis should be as short as possible. The 7-sec
1989). On the other hana Savin et al. (1982) reportcd thatthe segment used in our study isthe shortest datalengthto detect
reduction in sympathetic nerve activjty contributed greatly to HI7 using frequencyanalysis. Heretofbre, HR recovery has
HR recovery immediately after exercise and parasympathetic been a clinically important indexto evaluate possible mortality

NII-Electronic
NII-Electronic Library Service
JapanSociety
Japan Society of Physiological
of Physiological Anthropology
Anthropology

33U Timing ofHRY Recovery after Excrcise

causcd by rnyocardial ischemia(Cole et al,,2000;Nishime et Flactony injlueneingthe timing of'HF recoveny)?

al.,2000).However,HR recovery cannot bc used as an index In thisstudM HF power inYG started to recover ear!ier A,han
ofthe assessment of parasympatheticreactivation becauseitis in MG. The result suggests thatfactors such as age, physical

influencedby sympathetic fitness, and training status influence the timing ofHF recovery
and parasympathetic nervous
activity (Kannankeril et al.,2004),On the other hand since the in female distancerunners, Previous studies showed the
conventional techniques such as FFT and AR need stationary influence of age on parasympathetic modulation of HR by
data,thcy could not be used duringrccovery after exercise, estimating HF power (Stratton et al.,2003). Darr et al. (1988)
Thcrefbre, validity is doubtfu1in the assessment of cardiac indicatedthattrained subjects demonstrateda significantly
autonomic modulation mechanisms using these techniques. fasterHR recoyery than untrained subjects, which was
RecentlMKaiklconen et al. (2007), and Martinmtiki and R/usko particularly marked duringfast-phase recoyery, Buchheitand

(2008) jnvestigated the acute rccovery of HRV by analyzing Gindre (2006) reported a significant relationship betweenHR
minute-by-minute values immediate]y after different jntensity recovery and trainingloadcstimated by the Baeckesport score.
exercises using the Short-time Fourier transfbrm,which jsan Therefbre, multiple infiuencesof thesefactors to HF recovery

extension of theFFT,However,spectral powerswere averaged afier exercise are suggested, A future examination isrequired
for each successive 60-sec periodin thc methods of the to examine the factor which reflects most strongly the timing
aforementioned study, We stM believethat a preeisc ofHF recovery or the presenceof other factors.
descriptionof HI(V dynamics immediately after exercjse The validity of themethod used in our study was verified by
caniiot be provided through such a long time segment (i.e., 1 our investigation ofpossible djfl}]rentiation of HF recovery in
min). Goldberger et al, (2006) recently examined HRV using the subject groups with different properties. With respect to the
the time-domain method and revealed thatthe effect of the occurrence of parasympathetie reactiyatjon after exercise, it
parasympathetic nervous system appeared 30 seconds after remains unclear becausethe time varies wildly from 30
exercisc, We provedthatHF powcr increases within 30 seconds seconds (Imai et al., 1994)to one minute (Kannankeril et al.,

after excrcise in both groups.It indicates that frcquency 2004),In this studM we were able to detectthe timingof
analysis ofHRV using themaximum entropy method can also reactivation with more detailedscales by measuring HF
detectaspects of change in autonomic nervous activity after recovery as one segment fbr 7 seconds compared to previous
cxerclse. studies.In spite of the differeneesin age and athletic level
In this studM HRV was observed for 70 seconds after among the subjects, the differencein timing of HF recovery
exercise, which isa shorter time compared to previous studies, was just one segment (YG:HF7.i4, MG: HFi4.2i).However,
e.g., Martinrnaki and Rusko (2008), which observed HRV for thisis not a small difference becauseHF is infiuenced by
10 minutes after exercise and Kaikkonen et al, (2007) and respiration (Hirsh
and Bishop,1981).Thc BR of YG was
Goldberger et al. (2006), which observed HRV for5 minutes approximately 3.4times (29,1/60'7) forthe 7-secsegment of
after exercise. Thisiswhy thisstudy aims to findthe mutation HF7.i4,while thatof MG was approximately 4.4 timcs
point when HF power begins to increase,whjch is (37.9/60*7) fbr the 7-sec segment of HF,4 ,,. Considering the
demonstrated at a near-zero level duringexercise, HF recovery results, a differencc up to 14 second in time pointwhen the
ofboth groupsfbr70 seconds shows a small value, HF after 70 eardiac vagus nerve activity ofboth groupsstarted te reactivate
seconds dcmonstrates 76.3% at rest in the YG and only 5.8% is possible.Therefore,the detectionof fUrtherdetailcd
in the MG, However,the discovery of the mutation pointsof differences remains to be solved, However,the method used in
bothgroupswill providea basisto predict modulation of HR this study showed validity in evaluating eardiac autonomic
recovery after exercise. Earlyrecovery ofHF immediately after nervous activity becausewe could detect more detailed timing
exercise could be observed becausewell-trained subjects were ofthe reactivation and the difference betweentwo groupswith
examined in thisstudy. The characteristics of trainingand age differentcharacteristics.
of the subjects require a longer analysis time,
The time pointthat jndicatcda significant increaseor Mlethodologicallimitations
decreasein each peak of HR and HF after exercise in both This study hasdiscovered threemethodological limitations.
groupswas thc same, (Reference to Fig. 1 and Table 2), The First,HRY indicates yariability of the R-R interval, not an

strong association of parasympathetic nervo"s activity with indexfordetermination of cardiac autonomic nervous activity

HR Tecovery immediatelyafter exercise is suggested, [Irhis Parasyrnpathetic

quantitatively. blockadeby using atropine isa
result supports the previousstudy (Imai et al., 1994; standard method to evaluate parasympathetic activity
Kannankeri1et al., 2004).In addition, the time point that (Berntson
et al., 1997;Vukajlovic
et al,, 2006).
However,this
indicated significant increase or decrease in each peak of HF study does not clarify the infiuenceof medical agents.
and BR in bothgroups was also thesame. This result supports Therefore,as thiscontext, we haveto wait fOrfuture
regards

theideathatHF issynchronous with a change intidalvolume physiological verification fbr the HI(V index in autonomic
or breathingfrequency (Hirshand Bishop, 1981). It is nervous activity immediately after exercise, including the
confirmed that the main driverofHRV even immediately after influenceofmedical agents,

exercise is respiration. Second refiex cardiac rate regulation irnmediately

after

NII-Electronic
NII-Electronic Library Service
JapanSociety
Japan Society of Physiological
of Physiological Anthropology
Anthropology

Suzuki,S et al.JPbj,siolAnthmpoL 27.'

325-332,2008 331

(2005)
indicated Goldberger
JJ,Le FK LahiriM, Kannankeri] PJ,Ng J,Kadish
rlahara
exercise was not examined. et al. a

reperfusion refleK immediatelyafter exercise. It is also AH (2006)Assessmentofparasympathetic reactivation after

indicated that direct stirnulation of the central chemoreceptor exercise. Am J Physiol Heart Circ Physiol 290:
by C02 increases the ampljtude of respiratory arrhythrnia H2446-H2452
and Aguilaniu,2006). Therefbre,this study indicates
<Peronnet Hayano J,[[ajrlor
JA, Mukai S, Okada A, Watanabe Yl Takata
that HF power is influenced not only by vagus nerve activ{ty K, FiijinamiT (1994) Assessment of frequencyshifts in
butalso by reflex enhancement. R.R interval variability and respiration with complex
Thirathc LF (sumof the power from 0.04to O.15Hz)fHF demodulation.
J Appl Physiol77i2879-2888
ratio refiects the balance of sympathetic and parasympathetic HirschJA, BishopB (1981) Respiratory sinus arrhythmia in
nervous activity (fask Force of the European Societyof humans:how breathing patternmodulates heartrate, Am J
Cardiologyand the North American Societyof Pacing and Physiol241: H620-H629
ElectrophysioTogyl 1996).Inthe presentstudy, a 7-secsegment ImaiK, SatoH, HoriM, Kusuoka H, OzakiH, Ybkoyama H,
isanalyzed in order to examine parasympathetic reactivation. Takeda H, Inoue M, Kamada T (1994) Vagal]ymediated
The segment data of 25 sec (1secfO.04) at the earliest is heartrate recovery after exercise is accelerated in athletes
needed to estimate LF power.Therefbre, the influences of butbluntedinpatientswith chronic heartfailure. j Am Coll
sympathetic nerve activity and sympathovagal balance Cardiol24: 1529-1535
immediately after exercise are net clear. JavorkaM, ZilaI,BalharekT, JavorkaK (2003) On- and off-
responses of heartrate to exercise-relations to heartrate
Acknewledgements Thisstudy was partly supported by a variability,ClinPhysiolFunctImagjng23:1-8
Grant-jn-Aid Research(C)of the Japan Society KaiklconenR Nummela A, Rusko H (2007)
fbrScientific Heart rate
fbrthe Promotionof Science(No17500486), variability dynamicsduring ear]y recovery after d{fTerent
endurance exercises. Eur J App] Physio] 102: 79-86
References Kannankeril PJ, GoldbergerJJ (2002) Parasympathetic effects
on cardiac electrophysiology during exereise and recovery.
AraiY] SaulJR AIbrechtR HartleyLH, LillyLS, Cohen Rg Am J PhysiolHe art CircPhysiol 282: H2091-H2098
ColucciWS (1989)Modulationof cardiac autonomic Kannankeril PJ, Le FK, Kadish AH, GoldbergerJj (2004)
activity duringand immediately
after exercise. Am J Physiol effects on heart
Parasympathetic rate recovery after exercise.
256:H132-H141 J InvestigMed 52: 394401
BerntsonGG, Bigger JT Jr, Eckberg DL, Grossman e KostisJB, Moreyra AE, Amendo MT, Di PietroJ,CosgroveN,
KaufinannPG, MalikM, NagarajaHN, PorgesS"1 SaulJe Kuo PT (1982) The effect of age on heartrate in subjects
StonePH, van derMolen MW (1 997)Heart rate variability: free of heart disease. Studies by ambulatory
origins, methods, and interpretive caveats. electrocardiography and maximal exercise stress test,
Psychophysio]ogy 34:623-648 Circulation 65:141-145
Buchheit M, Gindre C (2006) Cardiac parasympathetic Levy WC, Cerqucira MD, Harp GD, Johannessen KA, Abrass
regulation: respective associations with cardiorespiratory IB,SchwartzRS, Stratton JR (1998) Etfectof endurance
fitness and trainingload.Am J PhysiolHeartCircPhysiol exercise trainingon heartrate variability at rest in healthy
291:H451-H458 young and older men. Arn JCardiol82:1236-1241
Carter J,JeukendrupAE (2002) Vhlidityand reliability of three LipsitzLA, Mietus J, Moody GB, GoldbergerAL (1990)
commercially available breath-by-breathrespiratory Spectral characteristics of heartrate variability beforeand
systems. EurJ Appl Physiol 86: 43S-441 during posturaL tilt.Relationsto aging and riskof syncope.

CoieCR, Foody JM, Blackstone EH, Lauer MS (2000) Heart Circulation 81:1803-1810
rate rec[wery after submaximal exercise testing as a Ludbrook J (1994)
Repeated measurements and rnultiple

predictor of mortality in a cardioyascularly healthycohQrt. comparisons in cardiovascular research. Cardiovasc

Res 28:
Ann IntcrnMed 132: 552-555 303-311
Darr KC, BassettDR, Morgan BJ, Thomas DP (1988) Effects Macor E FagardR, Amery A (1996) Power spectral analysis of
of age and training status on heart rate recovery after peak RR interval and bloodpressure short-tcrm variability at rest
exercise. Am J Physiol254:H340-H343 and duringdynamic exercise: comparison betweencyclists
Du N, Bai S, Oguri K, Kato Y Matsumoto I,Kawase H, and controls, IntJ SportsMed 17:175-181
Matsuoka T (2005) Heart rate recovery after exercise and MartinmakiK, Rusko H (2008) Time-frequency analysis of
neural regulation of heartrate variability in 30-40 yearold heartrate variability duringimmediate recovery from 1ffw
femalemarathon runners, J SportsSciMed 4:9-17 and high intensity exercise. Eur J Appl Physiol 102:
GellishRL, GoslinBR, Olson RE, McDonald A, Russi GD, 353-360
Moudgil VK (2007) Longitudinalmodeljng of the Murasato Yl Hirakawa H, Harada YlNakamura T,HayashidaY
relationship betweenage and maximal heart rate. Med Sci (1998) Effects of systemic hypoxiaon R-R interval and
SportsExerc39:822-829 bloodpressurevariabilities in conscious rats, Am J Physiol

NII-Electronic
NII-Electronic Library Service
JapanSociety
Japan Society of Physiological
of Physiological Anthropology
Anthropology

332 Timingof HRY Recoveryafter Exercisc

275:Ff797-H804 cardiovascular responses withdrawal. J

to parasympathetic
NishimeEO, ColeCR, Blackstone EH, PashkowFJ, Lauer MS Am CollCardiol 41:2077-2083
(2000) Heart rate recovery and treadmillexercise score as Sumi K, SuzukiS, MatsubaraM, Ando YlKobayashiF (2006)
predictors ofmortality in patients
referred forexercise ECG, Heartrate variability duringhigh-intensity
fieldexercise in
JAMA 284:1392-1398 female distancerunners. Scand J Med Sci Sports 16:
Ohtomo N, IerachiS, [lanakaYl [[bkiwanoK, Kaneko N 314-320
(1994)New method of time series analysis and its IlaharaN, Thkaki H, Thguchi A, Suyama K, KuritaT, Shimizu
application to WOIf'ssunspot number data, Jpn J Appl Phys IM Miyazakj S,Kawada Z SunagawaK (2005) Pronounced
33:2821-2831 HR variability after exercise in inferior ischemia:
evidence

Ohtomo N, TfbkiwanoK, Tlanaka Yl Sumi A, TbrachiS,Konno thatthe cardioinhibitory vagal reflex isinvoked
by exercise-
H (1995) Exponentialcharacteristics of power spectral induced inferior ischemia. Am J Physioi HeartCircPhysiol
densities caused by chaotic phenomena.J Phys Soc Jpn 64: 288:Hl179-Hll85
1104-1113 Tbsk Force of the European Society of Cardiology and the
Perini R, OrizioC, Comande A, Castellano M, Beschi M, North American Society of Pacing and Electrophysiology
Veicsteinas A (1989) Plasmanorepinephrine and heartrate (1996) Heart rate variability: standards of measurement,

dynamics duringrecovery from submaximal cxercise in physjologicalinterpretation,

and clinical use. Circulation 93i
man, Eur J Appl Physjol Occup Physio] 58:879-883 1043-1065
PeriniR, OrizioC, BaselliG, Cerutti S, VeicsteinasA (1990) TsuchidaM, Tokiwano K, Hosoi H, Sueoka K, Ohtomo N,
The infiuence ofexercise intensity
on thepower spectrum of TimakaY Mukasa K (1998) Restoration of imagesobtaincd
heartratc variability. Eur J Appl PhysiolOccup Physiol61: from the scanning tunnelingmicroscope, Jpn J Appl Phys
143-148 37:3500-3505
Peronnet E AguilaniuB (2006) Lactic acid bufTering, Vukaj1ovicDD, GuettlerN, Miric M, Pitschner HF (2006)
nonmetabolic C02 and exercise byperventilation] a critical Effectsof atropine and pirenzepine on heartrate turbulence,

reappraisal. Respir PhysiolNeurobiol150:4-18 Ann Noninvasive Electrocardiol11:34-37

Pierpont GL, StolpmanDR, GornickCC (2000) Heart rate Yamamoto K, MiyachiM, SaitohT, YbshiokaA, OnoderaS
recovery post-exercise as an index of parasympathetic (2001) Effectsof endurancc trainingon resting and post-
activity. J Auton Nerv Syst80:169-174 exercise cardiac autonomic eontrol. Med Sci SportsExerc
Pierpont GL, NlothEJ (2004) Assessingautonomic function by 33:1496-1502
analysis of heartrate recovery from exercise in healthy Yitmasaki Y Kodama M, MatsuhisaM, KishimotoM, Ozaki
subjects. Am J Cardio] 94:64-68 H, faniA, Ueda N, lshidaYl Kamada T (1996) Diurnal
Radoski HR, ZawalickEJ,FougerePF (1976) The superiority heartrate variability in healthysubjects: effects aging and
of maximum entropy power spectrum techniques to
applied sex diffbrence. Am J Physiol 27 1:H3 03-H31O
geomagneticmicropulsations, Phys Earth Planet Interiors
12:208-216 Received: February
29,2008
Savin WM, Davidson DM, Haskell WL (I982)Autonomic Accepted:September30,2008
contribution to heartrate recovery from exercise inhumans. Correspondence to: Shigehiro
Suzuki,Facultyof Science&
J ApplPhysiol53:1572-1575 Technology,Meijo University,1-501 Shiogarnaguchi,
Shannon DC, Carley DVIC Benson H (1987) Aging of [Ibmpaku-ku,Nagoya468-8502,Japan
modulatien ofheart rate. Am J Physiol
253:H874H877 Phonei+81-52-838-2402
Stratton JR, Levy WC, CaldwellJH, Jaeobson A, May J, Fax:+81-52-832-1170
Matsuoka D, Madden K (2003)
Effects of aging on e-mail:ssuzuki@ccmfs.meijo-u.ac.jp

NII-Electronic
NII-Electronic Library Service