Heart Rate Variability in Athletes

Abstract

This review examines the influence on heart rate variability (HRV) indices in athletes from training status, different types of exercise training, sex and ageing, presented from both cross-sectional and longitudinal studies. The predictability of HRV in over-training, athletic condition and athletic performance is also included. Finally, some recommendations concerning the application of HRV methods in athletes are made.
The cardiovascular system is mostly controlled by autonomic regulation through the activity of sympathetic and parasympathetic pathways of the autonomic nervous system. Analysis of HRV permits insight in this control mechanism. It can easily be determined from ECG recordings, resulting in time series (RR-intervals) that are usually analysed in time and frequency domains. As a first approach, it can be assumed that power in different frequency bands corresponds to activity of sympathetic (0.04–0.15Hz) and parasympathetic (0.15–0.4Hz) nerves. However, other mechanisms (and feedback loops) are also at work, especially in the low frequency band.
During dynamic exercise, it is generally assumed that heart rate increases due to both a parasympathetic withdrawal and an augmented sympathetic activity. However, because some authors disagree with the former statement and the fact that during exercise there is also a technical problem related to the non-stationary signals, a critical look at interpretation of results is needed.
It is strongly suggested that, when presenting reports on HRV studies related to exercise physiology in general or concerned with athletes, a detailed description should be provided on analysis methods, as well as concerning population, and training schedule, intensity and duration. Most studies concern relatively small numbers of study participants, diminishing the power of statistics. Therefore, multicentre studies would be preferable.
In order to further develop this fascinating research field, we advocate prospective, randomised, controlled, long-term studies using validated measurement methods. Finally, there is a strong need for basic research on the nature of the control and regulating mechanism exerted by the autonomic nervous system on cardiovascular function in athletes, preferably with a multidisciplinary approach between cardiologists, exercise physiologists, pulmonary physiologists, coaches and biomedical engineers.

Full text

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 1
Heart Rate Variability in Athletes
A.E. Aubert, B. Seps and F. Beckers
Laboratory of Experimental Cardiology, School of Medicine, K.U. Leuven, Leuven, Belgium
Address for corresponding author:
André E. Aubert
Laboratory of Experimental Cardiology
University Hospital Gasthuisberg O/N
Herestraat 49
3000 Leuven, Belgium
Tel: 32/16-345840
Fax: 32/16/345844
E-Mail: Andre.aubert@med.kuleuven.ac.be

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 2
Contents
Abstract
1. Introduction
2. Control of heart rate: the autonomic nervous system
3. Methodology and analysis of cardiovascular variability: heart rate variability (HRV), blood
pressure variability (BPV) and baroreflex sensitivity
3.1 Time domain
3.2 Frequency analysis.
3.2.1 FFT approach
3.2.2 Autoregressive modelling (AR)
3.2.3 Wavelet decomposition
3.3 Selection of the most relevant frequency ranges and physiologic significance
3.4 Non-linear methods
4. Exercise physiology related to HRV
4.1 General cardiovascular changes due to exercise
4.2 Exercise and the autonomic nervous system
5. Changes in HRV related to exercise training
5.1 HRV during exercise
5.2 Cross sectional: comparison of athletes and sedentary groups
5.3 Longitudinal: effect on HRV of exercise training of non-athletes
5.4 Differences due to age and gender
5.5 Over-training and the autonomic nervous system
6. Conclusions

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 3
Abstract
The present review examines the influence on heart rate variability (HRV) indices in athletes from
training status, different types of exercise training, gender and ageing, presented from both cross-
sectional and longitudinal studies. Also the predictability of HRV in over-training, athletic
condition and athletic performance is included. Finally some recommendations concerning the
application of HRV methods in athletes are made.
The cardiovascular system is mostly controlled by autonomic regulation through the activity of
sympathetic and parasympathetic pathways of the autonomic nervous system. Analysis of HRV
permits insight in this control mechanism. It can easily be determined from ECG recordings,
resulting in time series (RR-intervals) that are usually analysed in time and frequency domain. As
a first approach it can be assumed that power in different frequency bands corresponds to activity
of sympathetic (0.04-0.15 Hz) and parasympathetic nerves (0.15-0.4 Hz). However also other
mechanisms (and feedback loops) are at work, especially in the low frequency band.
During dynamic exercise it is generally assumed that heart rate increases due to both a
parasympathetic withdrawal and an augmented sympathetic activity. However, because some
authors disagree with the former statement and the fact that during exercise there is also a
technical problem related to the non-stationary signals, a critical look at interpretation of results is
needed.
It is strongly suggested that, when presenting reports on HRV studies related to exercise
physiology in general or concerned with athletes, a detailed description should be provided on
analysis methods, as well as concerning population, training schedule, intensity and duration.
Most studies concern relatively small numbers of subjects, diminishing the power of statistics.
Therefore multicentre studies would be preferable.
In order to further develop this fascinating research field, we advocate prospective, randomised,
controlled, long term studies using validated measurement methods. Finally, there is a strong need
for basic research on the nature of the control and regulating mechanism exerted by the autonomic
nervous system on cardiovascular function in athletes, preferably with a multidisciplinary
approach between cardiologists, exercise physiologists, pulmonary physiologists, coaches and
biomedical engineers.
1. Introduction
The manner in which the intact organism in
general and the cardiovascular system more
specific, responds to the stress of exercise has
intrigued sports physiologist for the past century.
The cardiovascular adjustments, necessary to meet
the extraordinary demands of the working
musculature, which begins even before the onset of
exercise, remain areas of intense investigation and
speculation
(1-2)
. As well anatomical geometry as
cardiovascular function of the heart are altered
after chronic physical activity
(3)
. For example, on
the one hand, persistent volume load such as
elicited after endurance training (or its pathologic
equivalent after aortic or mitral insufficiency) leads
to enlargement of the left ventricular internal
diameter and a proportional increase in wall
thickness
(4-5)
. This type of adaptation is called
eccentric left ventricular hypertrophy. On the other
hand, a pressure load such as elicited after power
training (or its pathologic equivalent of aortic
stenosis or hypertension) leads to a thickening of
the ventricular wall and an unchanged internal
dimension. This type of adaptation is called
concentric left ventricular hypertrophy. An
essential difference between exercise and
pathologic conditions is that the load on the heart
is continuous in the latter case and intermittent in
the former.
Other adjustments take place in almost every organ
system of the body and involve all aspects of
cardiac and peripheral vascular control, including
regulation by the autonomic nervous system
(ANS). Neural mechanisms appear to be of great
importance in mediating the initial response to
exercise, which involves very rapid changes in
heart rate and blood pressure. All these phenomena
involving heart rate and blood pressure are
described as “cardiovascular variability”. Both
phenomena covered in this review, exercise
training and its relation to control and regulation of
the cardiovascular function by the ANS, have also

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 4
important clinical aspects: 1. Can exercise training
be used to retard the advance of coronary and other
heart diseases? 2. Can HRV be used as a predictor
or as a marker of the progression of cardiovascular
disease?
Understanding interactions between cardiovascular
function, activity of the autonomic nervous system
and exercise training, remains a difficult problem.
The disciplines of medicine, exercise and
environmental physiology, physical education and
biomedical engineering are all closely allied to
study the effects of exercise and other stresses on
cardiac structure and function.
The goal of this review is to discuss how some of
the consequences of exercise training on the
cardiovascular system, can be deducted from
measured basic experimental data of heart rate
variability (HRV), aortic blood pressure variability
(BPV) and baroreflex sensitivity (BRS).
More specific, time and frequency analysis of heart
rate will be described as a valuable tool to
investigate the reflex mechanisms of
cardiovascular regulation in active athletes in a
fully non-invasive way.
The parameters of HRV, BPV and BRS can simply
be obtained from the measurement of the ECG
(and heart rate) and (non-invasive) blood pressure
as will be shown further. Indices from HRV and
BPV can be studied in time (statistical studies) and
frequency domain (power spectrum). These indices
can be a valuable non-invasive tool to investigate
the reflex mechanisms of cardiovascular regulation
during and after exercising, for de-training and
over-training, gender differences and the effects of
ageing.
This review will discuss consecutively: 1. control
mechanisms of heart rate and blood pressure and
the role of the ANS; 2. how to measure
experimentally HRV and BPV, starting from the
ECG and (non-invasive) blood pressure signals and
how to analyse; 3. correlation between HRV and
physical and physiological parameters, 4. HRV
data obtained from studies on athletes and related
to training, training overload and age and gender
differences.
2. Control of heart rate: the autonomic nervous
system
The cardiovascular system, the heart and the
circulation, are mostly controlled by higher brain
centres (central command) and cardiovascular
control areas in the brain stem through the activity
of sympathetic and parasympathetic nerves
(6)
.
Control is also affected by baroreceptors,
chemoreceptors, muscle afferent, local tissue
metabolism and circulating hormones
(7)
. Study of
cardiovascular variability allows mainly access to
the activity of the nerves and the baroreceptors.
The autonomic nervous system describes those
nerves that are concerned predominantly with the
regulation of bodily functions. These nerves
generally function without consciousness or
volition. Autonomic nerves comprise sympathetic
nerves and parasympathetic nerves (the latter often
being used as synonym of vagal, because the
parasympathetic supply to the heart runs in the
vagal nerves). Both divisions contain both afferent
and efferent nerves and both myelinated and non-
myelinated fibers. In general the effects of the two
divisions are complementary, with activity in
sympathetic nerves exciting the heart (increasing
heart rate), constricting blood vessels, decreasing
gastrointestinal motility and constricting
sphincters, and parasympathetic nerves inducing
the opposite response. The autonomic system
supplies both afferent and efferent nerves to the
heart, with sympathetic nerve endings all over the
myocardium and parasympathetic on the sino-atrial
node, on the atrial myocardium and the atrio-
ventricular node. These nerves not only control
heart rate and force, but both sympathetic and
parasympathetic nerves supply important
reflexogenic areas in various parts of the heart
which, when excited by either mechanical or
chemical stimuli, give rise to reflexes which
influence both the heart itself and the state of
constriction of blood vessels
(8)
. These neural
pathways are also closely linked to baroreceptor
reflex activity, with changes in blood pressure
playing a key role in either increasing or
decreasing activity of one or the other pathway.
Analysis of cardiovascular variability permitted
insight into the neural control mechanism of the
heart, leading to a new discipline:
“Neurocardiology”
(9-10-11)
. This area combines the
disciplines of neurosciences and cardiovascular
physiology on the research side and of neurology
and cardiology on the clinical side.
The normal heartbeat and blood pressure vary
secondary to respiration (respiratory sinus
arrhythmia), in response to physical,
environmental, mental and multiple other factors
and is characterized by a circadian variation. Both
the basic heart rate and its modulation are
primarily determined by alterations in autonomic
activity. Increased parasympathetic nervous
activity slows the heart rate and increased
sympathetic activity increases the heart rate
(Figure 1a)
(12)
. In reality however the situation is
much more complex and figure 1b depicts a more

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 5
Therefore, cardiovascular control, as expressed by
the time-dependence of haemodynamic variables,
is a direct reflection of autonomic activity. It may
be a useful tool to examine autonomic fluctuations
under different physiological circumstances
(14)
or
to study external influences such as the effect of
training.
evolved working model that, starting from central
cardiovascular control as a black box, identifies the
output of the ANS to blood pressure and heart rate
and describes the feedback loop via the
baroreceptors
(13)
. In a healthy individual, the role
of the autonomic nervous system in the beat-to-
beat adjustment of haemodynamic parameters is
essential to adequate cardiovascular functioning. Autonomic nerves thus have a pivotal role in the
regulation of the cardiovascular system both in
ensuring optimal function during various activities
in health under varying physical conditions, even
during weightlessness
(15)
, and also in mediating
several of the manifestations of cardiac diseases.
a.
+
_
Brain
+
Sympathetic
Parasympathetic
Heart
HR, BP
SV
3. Methodology and analysis of cardiovascular
variability: heart rate variability (HRV), blood
pressure variability (BPV) and baroreflex
sensitivity (BRS).
The first step for the analysis of HRV and BPV
signals are obtaining high quality ECG and
(non)invasive blood pressure tracings under
stationary conditions (Figure 2). As the analysis of
the ECG and blood pressure are very similar, only
the ECG will be discussed further. Duration of
recordings can extend from a minimum of 10 min
to 24 h in Holter recordings. The duration has to be
sufficiently long and stationary during that period,
allowing a good frequency resolution. For
frequency domain measurements it is
recommended that the duration of the recording is
at least two times the wavelength of the lowest
frequency component. Accordingly the minimum
duration for the assessment of the high frequency
component (0.15 Hz) would be 13.3 s and for the
low frequency component (0.04Hz) 50 s. However,
it is generally recommended to have minimum
duration recordings of 5 minutes or even better 10
minutes. For the study of circadian variations
Holter recordings (24 hour) covering a full
day/night cycle are needed. Also, as will be shown
later, many HRV indices depend upon the duration
of the recording. Thus, it is inappropriate to
compare HRV indices obtained from recordings of
different duration with each other.
b.
Figure 1 a. A very simple model illustrating the
influence of the sympathetic (increase heart
rate) and parasympathetic nervous activity
(decrease heart rate) on heart rate, so called
“balance model”. 1. b. Block diagram of a more
elaborated working model of cardiovascular
control mechanisms of heart rate and blood
pressure and feedback mechanism from
baroreflex. The diagram illustrates independent
actions of the vagal, alpha-sympathetic and
beta-sympathetic systems. Their action can be
assessed by measuring HRV, BPV and
baroreflex mechanism. The parasympathetic
activity is responsible for the bradycardia
accompanying baroreceptor stimulation and for
the tachycardia accompanying baroreceptor
deactivation, with the sympathetic nervous
system also playing a minor role. TPR: total
peripheral resistance, CO: cardiac output, SV:
stroke volume.
While laboratory conditions may be closely
controlled, artifacts are present in almost all Holter
recordings or telemetry recordings as obtained in
the field. These signals are analog/digital
converted for computer processing. In order to
have a good time resolution and event definition, a
sampling rate of at least 250 Hz and up to 1000 Hz
(giving a time resolution of 1 ms) is recommended.
The second step is the recognition of the QRS
complex. Peak detection is often performed with
commercially available software included in the

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 6
a
b
c
d
Figure 2. Analysis of HRV: calculation of consecutive RR intervals (a) on the ECG, results in the
tachogram (b) which can be analysed in the frequency domain (c) and the time domain (d). Spectral
analysis and histogram are results from a 24 h Holter recording. Therefore the histogram shows two
peaks: around 1100 ms, corresponding to mean heart rate at night and around 750 ms,
corresponding to mean heart rate during day time.
Holter analysis systems. An algorithm was
developed in house for threshold detection
(16)
. This
algorithm functions as well on the ECG as on the
blood pressure recordings. The result is a discrete,
unevenly spaced time event series: the tachogram,
obtained from the ECG. It is crucial that before
processing, these signals are corrected for ectopic
and missed beats
(11-17)
. This is performed with
filtering (elimination of spurious peaks) and
interpolation algorithms (i.e. replacing beats to be
corrected by the mean of a combination of
preceding and following beats)
(18)
. After this step a
normal-to-normal interval (NN) is obtained.
A final step is needed before spectral analysis can
be performed. Computation of the spectral
components of the tachogram requires a signal
sampled at regular intervals, which is not the case
for the tachogram, sampled by each (variable)
heartbeat. A regular signal is obtained by
modifying the tachogram. An interpolation is
performed and, on this last signal, equidistant
points are sampled every 0.5 s. Different
algorithms have been proposed to achieve
equidistant sampling
(19-20-18)
.
Non-invasive blood pressure can be measured
using finger cuffs
(21-22)
or a pulse displacement
device
(23)
. Both methods allow continuous
recording of blood pressure and can be calibrated
with a conventional arm cuff device. The analysis
of blood pressure signals is very similar.
Therefore, a separate description will not be given.
The only supplementary differences are: 1.
maxima (systolic blood pressure values) and
minima (diastolic blood pressure values) should be
detected as well and 2. on the contrary to the QRS
peak where only the timing of its occurrence has to
be recorded, here both coordinates: amplitude (in
mmHg) and timing (s) have to be recorded. The
variations in systolic blood pressure lead to the
systogram and the variations in diastolic blood
pressure to the diastogram.
Data analysis on all these graphs can be
approached from different viewpoints,
accentuating different underlying physiological
mechanisms. Traditionally time and frequency

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 7
domain have been considered, recently also non-
linear dynamics methods have been added.
3.1 Time domain:
Parameters in the time domain are easily computed
with simple statistical methods, even from short
time windows. Their main limitation is the lack of
discrimination between the activity of the different
autonomic branches.
Recommendations for a standardization of valid
parameters have been published
(24)
. The most
frequently used time domain parameters include
SD and SDANN, which represent global
variability, and rMSSD and pNN50, which are
highly correlated to high frequency power in the
frequency domain and represent markers for vagal
modulation
(25)
, (as will be explained later). The
definition of the different indices is as follows:
SDNN (or SD), (ms): standard deviation of the
normal to normal (NN) interval (result from
corrected signals for ectopic and missed beats by
filtering and interpolation algorithms) over the
recorded time interval. Theoretically heart rate
variance, equal to (SDNN)
2
and total power are
mathematically identical. In practical terms
however, correspondence between SDNN and the
total spectral power depends on data processing:
treatment of ectopic beats, interpolation, definition
of total power, etc
(26)
. It depends largely on the
duration of the recording; therefore, SDNN values
from recordings of different duration should not be
compared.
SDANN (ms): standard deviation of the 5-minute
mean NN interval over the entire recording. As
SDANN values are obtained from successive short
5-minute periods, it can only estimate changes in
heart rate caused by cycles shorter than 5 minutes.
Previous indices can be obtained from statistical
methods such as shown in the histogram in figure
2d. It provides mean values, standard deviation,
coefficient of variation and related parameters.
rMSSD (ms): the square root of the mean squared
successive differences between adjacent RR
intervals over the entire recording.
pNN50 (%): the percentage of successive interval
differences larger than 50 ms computed over the
entire recording.
Some typical values of previously mentioned
parameters are shown in Table 1. It gives values
for Mean NN, SD, rMSSD and pNN50, obtained
from 10 control subjects and 10 aerobic trained
athletes
(143)
, in supine and standing position.
Aerobic trained athletes show a higher NN (lower
heart rate) compared to the control group and
higher rMSSD and pNN50 as well in supine as in
standing position. Also rMSSD and pNN50 are
significantly larger (p<0.05) in supine compared to
standing position. This corresponds to a larger high
frequency modulation in supine position compared
to standing (more vagal modulation) as will be
discussed later (see also Figure 4).
Another possibility to process RR intervals in the
time domain is the use of geometrical methods
(27)
.
The simplest one is the sample histogram (Figure
2d), of which parameters related to the distribution
can be calculated: mode (value that occurs most
often), skewness (a measure of symmetry) and
kurtosis (a measure of peakedness). Lorenz or
Poincaré maps plot the duration of each RR
interval against the duration of the immediately
preceding RR interval. The practical use of the
geometrical methods seems to be rather limited
and up to now, not so often used in the literature.
Table 1. HRV parameters in time domain
obtained from 10 control (sedentary) subjects
and 10 aerobic trained athletes. Values are
mean±SD. *p<0.05 (Modified from
[143]
)
Mean NN (ms) SDNN (ms) rMSSD (ms) pNN50 (%)
Supine
Control
880.7±263.8 69.7±37 45.5±26.8 21.8±19.7
Aerobic
1100.3±158.5* 97.9±15.7* 73.5±23.7* 40.1±16.6*
Standing
Control
749.7±165.6 65.4±38.9 30.6±16.9 10.5±12.4
Aerobic
947.7±108.8* 92.9±30.9 47.2±11.1* 22.4±8.9*
3.2 Frequency analysis
By definition, spectral analysis decomposes any
steady, stationary, fluctuating time dependent
signal into its sinusoidal components. It allows
plotting the power of each such component as a
function of its frequency and the computation of
the power in defined frequency regions. Power
spectral analysis has been performed by Fast
Fourier Transform (FFT)
(28)
, by autoregressive
modelling
(29)
and by wavelet decomposition
(30)
.
3.2.1 FFT approach
The FFT method is an objective method because
no information is lost: the tachogram can be shown
in the frequency domain after FFT and the latter
signal can be backward transformed to retrieve the
original tachogram. Units of the spectral
components are: ms
2
/Hz for HRV and mmHg
2
/Hz
for BPV. The advantage of the classical FFT
approach consists mainly in its computational
efficiency and its simple implementation (Figure
2c). However these advantages are
counterbalanced by some limitations. These are
mainly related to the limited frequency

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 8
resolution
(18)
, which is directly related to the
duration of the recording period (which also
determines the lower limit of the spectrum, the
latter equals the inverse of the recording length)
which is affected by the windowing process as
well. The upper frequency limit (1 Hz in humans)
is imposed by the Nyquist criterion: it equals half
the sampling rate, which in case of resampling the
signal every 0.5 s corresponds to 2 Hz. Therefore
the upper frequency limit is at 1 Hz.
The main reason why FFT analysis is so popular in
the scientific community is that is relatively simple
to apply, gives a nice graphical representation and
is readily available for application on computers, it
is even used for analysis of running velocity
(31)
.
3.2.2. Autoregressive modelling
This approach considers the time series as a
difference equation, such that the signal at every
time step is expressed as a linear function of its
values at J previous time steps. Therefore the
autoregressive model (AR) requires an a priori
choice of the value of J (the order of the parametric
model) to provide the best fit to the data set that is
being processed. Visually the autoregressive
spectrum presents smoother spectral components,
which can be distinguished independently of pre-
selected frequency bands
(29)
. The power content in
these peaks can be calculated without the need for
predefined spectral bands.
The limitations of this method are linked with the
adequacy of the choice of the order J, which may
affect the accuracy of the determination of the time
series and the power spectra. The model order J,
even if selected objectively by information theory
criteria, importantly determines both centre
frequency and the magnitude of the spectral
components
(32)
.
3.2.3 Wavelet decomposition.
Wavelet transform
(30-33)
, a relative recent
development, provides a general signal processing
technique that can be used in numerous biomedical
applications. Its development was originally
motivated by the desire to overcome the drawbacks
of traditional Fourier analysis (e.g. fast Fourier
transform FFT), simultaneously providing time and
frequency information of the signal. The wavelet
transform (WT) indicates which frequencies occur
at what time, showing good time resolution at high
frequencies and good frequency resolution at low
frequencies. This multiresolution joint time-
frequency analysis is therefore suited for the
examination of non-stationary signals. Real
signals, like an electrocardiogram (ECG) or a
tachogram, are mostly non-stationary. The
information obtained by the wavelet
decomposition can, be used to compare differences
in power or standard deviations at each of the
wavelet levels analysed.
a.
b.
Figure 3a. Comparison of spectral analysis
methods: upper panel FFT, lower panel:
autoregressive modelling (order is 24). Peaks
(due to respiration at fixed rate) are at the same
frequency, but the autoregressive signal is
smoother than the FFT signal.
3b. Comparison of power bands as obtained
from FFT (top) and wavelet transform (WT)
(bottom). Control measurements from 10
(sedentary) control subjects and aerobic: from
10 aerobic trained athletes. Recordings were
obtained in supine position (modified from
[30]).

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 9
Wavelet transform offers superior time resolution
and time localisation compared to FFT or
autoregressive models. Wavelet transform analysis
also is not restricted to stationary signals. The
advantage of WT over AR is that no assumptions
have to be made about model parameters. It offers
rapid frequency decomposition with time
resolution, useful when one is interested in a
particular power spectral band over time and a
potential use to assess fractal characteristics.
A limitation of the method consists in the choice of
the basic wavelet function (the mother wavelet),
which has to possess some specific properties.
Furthermore the wavelet transform results in
coefficients, which have to be related to power in
specific frequency bands.
The previously mentioned frequency analysis
methods are compared in Figure 3. Both FFT and
AR methods provide very comparable results, with
AR providing a smoother spectral shape. It also
allows decomposition of the spectrum (division of
the spectrum in its root components) without the
need for predefined spectral bands.
In the same figure (Figure 3b), power bands
obtained from FFT and from WT are compared
between the same two groups (control subjects and
aerobic trained athletes) as described in Table 1.
Two conclusions can be drawn from this figure: 1.
FFT and WT provide very comparable results; 2.
aerobic trained athletes, with a low resting heart
rate, have indications of increased power in all
frequency bands compared to the control
(sedentary) group. This implies an increased
modulation of heart rate by the ANS, especially of
the parasympathetic component.
3.3 Selection of the most relevant frequency
ranges and physiologic significance
The power spectrum of the HRV signal, as
obtained from spectral analysis (FFT,
autoregressive modelling or wavelet transform),
was proposed to be used as a quantitative probe to
assess cardiovascular control mechanisms
(14)
In a typical heart rate power spectral density (PSD)
(which is the integral of the amplitude-frequency
curve and is expressed in ms
2
for HRV and in
mmHg
2
for BPV) three main frequency bands can
be observed: very low frequency (VLF), low
frequency (LF) and high frequency (HF)
components (Figure 2c). Power in the LF and HF
bands can also be expressed in normalised units:
LFnu and HFnu, these are the values of LF and HF
divided by the total power minus VLF and
multiplied by 100 (in %). The distribution of the
power and the central frequency of these
components are not fixed but may vary in relation
to changes in autonomic modulation of heart rate
and blood pressure
(17)
. In man, the spectral
components are usually integrated over two
frequency regions defined as LF (0.04-0.15 Hz,
with a central frequency around 0.1 Hz) and HF
(0.15-0.4 Hz, with a central frequency at the
respiratory rate around 0.25 Hz). The LF and HF
bands are indicated in figure 2c. In other mammals
these regions are differently chosen according to
the heart rate of the specific species
(18)
.
Which neural mechanisms are underlying these
spectral bands fluctuations? Parasympathetic
efferent activity was considered responsible for
HF, i.e. respiration linked oscillation of HRV. This
statement was made in conclusion after
experiments with vagotomy performed in
experiments on decerebrate cats
(34)
, or after
muscarinic receptor blockade in conscious dogs
(35)
and in humans
(36)
. Both parasympathetic and
sympathetic outflows were considered to
determine LF, together with other regulatory
mechanisms such as the renin-angiotensin system
and baroreflex
(37-36)
.
The LF/HF ratio can assess the fractional
distribution of power
(38)
, although like any ratio, it
can emphasize the opposite changes.
Below the LF frequency range (referred to as
VLF), there is often a continuous increase in
power. In part, this is the expression of very slow
frequency oscillations, probably related to
thermoregulation, but also non-harmonic DC noise
and the windowing process. These rhythms cannot
be satisfactorily resolved and quantified by the
traditional spectral analysis methods that are
performed on short recordings (of the order of
minutes). Different techniques and specific
methodologies have to be applied for a correct
understanding and quantification of these complex
and not yet fully clarified mechanisms. Spectral
analysis of 24-h traces provides information down
to 10
-5
Hz and shows a circadian pattern. The long-
term power spectrum of heart rate
(39)
(40-33-41)
seems
to display a 1/f shaped frequency dependence (with
a slope around –1 in humans), raising the question
whether the cardiovascular control mechanism is
of fractal nature.
A simple autonomic provocation consists in an
active change of posture from supine (Figure 4
left) to standing (Figure 4 right) (see also Table 1).
This results in a shift of blood away from the chest
to the venous system below the diaphragm, usually
referred to as venous pooling. Almost invariably in
all normal subjects an increase in heart rate is the

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 10
Figure 4. Tachogram and corresponding power spectral density (PSD) of a standing subject (left)
and a supine subject (right). From supine to standing heart rare rises (RR intervals become shorter)
and high frequency power (parasympathetic) is depressed compared to supine, whereas low
frequency power (partially sympathetic) increases.
result (from mean value of 85 beats/min supine to
120 beats/min standing on figure 4). While
standing, the regulatory system increases heart
rate, cardiac contractility and vascular tone by a
decrease in parasympathetic outflow and an
increase in sympathetic outflow. The latter
increase is reflected in the LF content of the PSD
(Figure 4 left) and the former decrease in the HF
content. While being supine, there is a
parasympathetic predominance, which switches to
sympathetic predominance on standing.
3.4 Non-linear methods
Chaotic behaviour exhibits a number of
characteristics that distinguish it from periodic and
random behaviour
(42-43)
i.e. HRV spectra show a
broad band noise-like variability over a large
frequency span
(44-45)
.
.
This seems to be due to non-
linearity in the cardiovascular control network. The
long-term regulation of heart rate contains both
short-time periodic (e.g. respiratory) modulations
and entirely non-periodic fluctuations. There are
indications that a reduction in complexity comes
along with a decrease in parasympathetic activity,
suggesting that a considerable amount of non-
linear behaviour be provided by this branch of the
ANS. Methods of non-linear dynamics define
parameters that quantify complicated interactions
of independent and interrelated components, which
can be described as ‘complexity measure’
(46-47-48)
.
Non-linear dynamical methods have made their
appearance in the analysis of HRV only recently
and methods have still to be established. Methods
related to chaos theory are used to describe the
non-linear properties of heart rate fluctuations
(attractors, 1/f behaviour of the power spectrum,
fractal-
(49-50)
and correlation dimension
(51)
,
Poincaré- and higher order moment plots,
approximate entropy
(52)
, pointwise correlation
dimension, detrended fluctuation analysis
(47)
,
Lyapunov exponents
(53)
).
The use of the new methods from non-linear
dynamics for HRV analysis may provide a more
sensitive way to characterise function or
dysfunction of the control mechanism of the
cardiovascular system. These tools are promising
with regard to the understanding of the latter
mechanism, but are still under development and
evaluation. Moreover these methods require more
powerful computing and are less visual attractive
compared to frequency analysis.
3.5 Baroreflex sensitivity
Evaluation of RR interval changes corresponding
to aorta blood pressure variations, allow to assess
the activity of the baroreceptive mechanism
(54)
.
Results from combined HRV and BPV signal
analysis lead to different methods that relate to the
baroreflex mechanism. The enormous complexity
of baroreflex interactions has been extensively
reviewed recently
(55)
.
Several methods have been described to study
arterial baroreflex activity. The majority of the
methods depend on pharmacological or
physiologic maneuvers that produce an abrupt
increase or decrease in blood pressure
(56)
.

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 11
Subsequently, quantification of the (linear) relation
between blood pressure and corresponding heart
rate changes is performed by calculation of the
slope of the fitted linear curve
(57)
. With standing or
passive tilt transient hypotension occurs that results
in a reflex increase in heart rate, whereas the post-
Valsalva increase in blood pressure causes reflex
slowing
(58)
. Maneuvers like neck suction or neck
pressure that alter the transmural pressure or
stretch in the carotid sinus also can be used in
humans to activate (load) or deactivate (unload)
arterial baroreceptor reflexes
(59)
. Drugs such as α-
adrenergic agents (phenylephrine or PE) or
angiotensin II that increase blood pressure produce
reflex slowing of the heart rate, whereas drugs like
nitrates or sodium nitroprusside that lower blood
pressure directly by relaxing vascular smooth
muscle augment sympathetic efferent nerve
activity and cause tachycardia and an increase in
cardiac contractility. A high slope of the regression
line is interpreted as indicating the presence of
strong vagal reflexes while a relatively flat slope
indicates the presence of weak vagal reflexes,
possibly associated with high reflex sympathetic
activity
(60)
.
The usefulness and constraints of traditionally used
methods have been reviewed elsewhere
(61)
. Some
investigators have even viewed the traditional
drug-induced baroreflex as misleading
(62)
.
Recently, several methods have been developed to
quantify spontaneous BRS or spBRS. Some are
based on the use of the spectral analysis of both
RR and BPV variabilities (α-index)
(63)
, on the
analysis of sequences of concurrent alterations in
BP and HR (sequence method)
(64)
, or on the
method of statistical dependence
(65-66-67)
. The
spontaneous BRS has a number of important
advantages: is does not require the use of i.v. drugs
or a neck chamber apparatus and it measures BRS
in the normal physiological range over a period of
time rather than brief and extreme perturbations as
induced by other methods. In this respect, it
represents a true steady-state assessment of the
cardiac baroreflex under stationary conditions.
It is out of the scope of this review, but suffice it to
mention that HRV methods have many
physiological and clinical applications studying the
influence of: ageing and gender studies
(68-69)
,
anxiety, stress
(70)
and depression
(71-72)
, smoking
(73-
74)
, caffeine
(75-76)
and alcohol consumption
(77-78)
,
risk assessment after myocardial infarction
(79)
or
predictor of mortality
(80-81)
, hemodialysis
(82)
,
congestive heart failure
(83)
and heart transplant
patients
(84)
, diabetes
(85)
, hypertension
(86)
, drug
testing
(87-88)
, sudden infant death syndrome
(89)
,
influence of gravity
(90-91)
, exercise training in
patients after coronary artery disease
(92)
or heart
transplantation
(93-94)
.
All the HRV and BPV analysis methods as
described above, have been implemented in
appropriate algorithms in our laboratory and
accordingly software was developed in-house. All
programs were implemented in LabVIEW, (which
is a graphical language) and variability parameters
determined according to the standards provided in
the Task Force on HRV
(24)
and extensively tested
and validated
(16-18-26-47-52-67)
.
4. Exercise physiology aspects as related to
HRV
The cardiovascular adjustments in exercise
represent a combination and integration of neural
and local chemical factors. The neural factors
consist of: 1. central command, 2. reflexes
originating in the contracting muscle, and 3. the
baroreflex. Central command is the cerebrocortical
activation of the sympathetic nervous system that
produces cardiac acceleration, increased
myocardial contractile force and peripheral
vasoconstriction. When exercise stops, an abrupt
decrease in heart rate and cardiac output occurs
and the sympathetic drive to the heart is essentially
removed. Blood pressure will be stabilised by the
baroreflex and parasympathetic activity will be
enhanced.
4.1 General cardiovascular changes due to
exercise
Physical activity is associated with hemodynamic
changes and alters the loading conditions of the
heart
(95)
. Cardiovascular responses to physical
activity depend on the type and intensity of
exercise. The main difference, at heart level is the
increased volume load during endurance exercise
in contrast to pressure load during strength
exercises
(96)
. These differences in loading will
cause various cardiovascular responses to physical
activity. After long term athletic training left
ventricular diastolic cavity dimensions, wall
thickness and mass will increase
(4-5)
. These
changes are described as the "athlete's heart".
However in comparison to men, female athletes
show smaller left ventricular mass
(97)
. This gender
difference has been associated with a lower
systolic blood pressure during 24-h Holter
recordings and during exercise in female
athletes
(98)
.

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 12
The volume load during endurance training results
in adaptive changes in many aspects of
cardiovascular function
(2)
. The heart improves its
ability to pump blood, mainly by increasing its
stroke volume, which occurs because of an
increase in end-diastolic volume and a small
increase in left ventricular mass. In contrast,
strength training results in larger increases in left
ventricular mass. There is little or no change in
ventricular volume. Endurance exercise also
decreases the metabolic load on the heart at rest
and at any submaximal exercise intensity. It does
so by increasing stroke volume and decreasing
heart rate. The result is a more efficient pressure-
time relationship.
A short overview of the major cardiovascular
changes will be given.
Heart rate is the predominant mechanism by which
cardiac output rises during exercise under
physiological circumstances
(99)
. Tachycardia can
occur either by neural stimulation or by an
elevation in circulating catecholamines
(100)
.
Increased heart volume and contractility will lead
to higher values of stroke volume, as well during
rest as during submaximal and maximal exercise.
Also, the lower heart rate will increase stroke
volume because of longer periods of diastole. The
heart ejects the extra blood due to the Frank-
Starling mechanism
(2)
. Another factor inducing
higher stroke volume is the larger blood volume in
athletes
(101)
.
Endurance training reduces resting and
submaximal exercise systolic, diastolic and mean
arterial blood pressures
(102)
. The mechanism of
reduced blood pressure at rest is not known.
Endurance training will also influence the release
of catecholamines. Norepinephrine is released by
the sympathetic nerve processes. An endurance
training programme will result in less
catecholamine response to submaximal exercise
but not to maximal exercise
(103)
.
4.2 Exercise and the autonomic nervous system
Heart rate is generally regulated predominantly by
the ANS
(7)
. The two major efferent mechanisms by
which tachycardia occurs are either through a
decrease in parasympathetic or through an increase
in sympathetic stimulation
(6)
. The latter can occur
either by neural stimulation or by an elevation in
circulating catecholamines. The mechanism of the
(exercise induced) tachycardia appears to involve
parasympathetic and spinal sympathetic reflex
circuits (Brainbridge reflex). The latter mechanism
is important to mention, since stimulation of
cardiovascular sympathetic afferent fibres produce
cardiovascular reflexes that operate through a
positive feedback mechanism and thus may be
particularly responsible for the increased
sympatho-adrenal activity of exercise
(2)
. This is
opposed to reflex responses initiated by
baroreceptor or parasympathetic innervated
cardiopulmonary receptors that operate through
negative feedback mechanisms
(7)
.
Thus both the sympathetic and parasympathetic
arms of the ANS play a pivotal role during
exercising. Therefore it can be expected to find
changes in HRV indices according to the degree
and duration of training and/or the kind of
training
(104)
.
Long term physical training influences cardiac
rhythm: sinus bradycardia in resting conditions and
a slower increase in heart rate at any degree of
submaximal oxygen uptake due to a shift of the
sympathovagal balance
(105)
towards
parasympathetic dominance
(106)
. Although the
latter point has been questioned recently
(107-108)
and
a direct involvement of the sinus node was
suggested. This point will be dealt with later on.
Heart rate during exercise is regulated by increased
sympathetic modulation and withdrawal of
parasympathetic activity
(8)
. It varies within an
individual according to heredity (size of the left
ventricle: predisposition for certain sport
activities), fitness level, exercise mode (endurance
or static training) and skill (economy of exercise).
Body posture (supine, sitting, standing
(109)
),
environmental variables (temperature
(110)
,
humidity, altitude
(111)
), state of mood
(112)
and
hormonal status
(113)
also alter heart rate response.
Heart rate and HRV as well are also affected by
drugs, stimulants
(76)
and eating habits.
Reflex adjustments initiated by the stimulation of
afferent nerve fibres from the exercising muscles
are also likely to play a role in the cardiovascular
response to exercise
(114)
. There is evidence that
reflex cardiovascular adjustments originating in the
contracting muscles are not mediated by muscle
spindle afferents but rather by small myelinated
and unmyelinated afferent fibres
(115)
.
Since exercise is accompanied by major
cardiovascular alterations, including marked
tachycardia, increases in cardiac output and in
arterial and atrial pressures and a reduction in total
peripheral resistance, it could be expected that a
cardiovascular regulating mechanism as important
as the arterial baroreceptor reflex would play a
significant role in mediating and modifying the
exercise response
(116)
. Investigations into the role
of the arterial baroreflex in the control of the

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 13
Table 2. HRV during exercise.
Author N Age Sympathetic Parasympathetic Remarks
(years)
Arai
(126)
43 25-65 no change withdrawal
FFT
Maciel
(129)
23 25-35 no change withdrawal pharmacologic
increase at higher blockade
activity level
Brenner
(130)
increase at onset withdrawal FFT
later on attenuated review
increase due to
higher temperature
Perini
(125)
7 23.7±0.5 no change at no change at AR
low intensity low intewnsity
decrease at higher
Kamath
(133)
19 20-32 decrease AR
Shin
(135-136)
5 17-21 decrease decrease AR
8 21-40 decrease decrease Non-athletes
N : Number of subjects, FFT : fast Fourier transform, AR : autoregressive method
cardiovascular system during exercise have yielded
conflicting conclusions as to their importance
(117)
.
At first it was suggested that the baroreflex is just
as active during exercise as at rest. On the other
hand, if the baroreflex was also important during
exercise, than the occurrence of tachycardia
associated with an elevated pressure is opposite to
the predicted response, since the baroreceptor
should act to restrain heart rate in the face of an
elevated pressure
(118)
.
There is now a large body of evidence suggesting
the lack of importance of the baroreflex during
exercise
(119-120)
(also similar response to moderate
exercise in intact dogs and arterial baroreceptor
denervation
(121)
).
In reality, cardiovascular control mechanisms are
much more complex as shown in figure 1 as was
recently shown in a review by Malpas
(122)
. Stroke
volume and end-diastolic volume also contribute
(13)
in an intricate feedback system .
Taking all these considerations together
concerning HRV and its relationship to training,
some questions still remain unanswered: 1. Are
differences in ANS control of the cardiovascular
system between trained athletes and a sedentary
population due to a training effect or are other
factors involved? 2. Can cardiovascular variability
(HRV and BPV) parameters be used as a predictive
factor for athletic achievements, or in other words,
can HRV and BPV be used to predict optimal
training and athletic performance?
5. Changes in HRV related to exercise training
Highly trained athletes have a lower resting heart
rate
(3)
. Anticipation of physical activity inhibits the
vagal nerve impulses to the heart and increases
sympathetic discharge
(123-124)
. The concerted
inhibition of parasympathetic control areas and
activation of sympathetic control areas in the
medulla oblongata elicit an increase in heart rate
and myocardial contractility.
Technically a problem arises for heart rate
measurements during exercise: as it is increasing
according to the intensity of exercising, no steady
state is obtained, which is necessary for spectral
analysis. Two approaches are usually proposed in
the literature to solve this problem: 1. perform
measurements at a fixed intensity level
(125)
, 2.
subtract a background trend to decrease the
contribution of the continuous increase in heart
rate with increasing exercise intensity
(126)
. The
latter method is based on the fact that the linear
trend (first order) represents the largest non-
stationarity of heart rate during and after exercise.
Normally one is also only interested in resolving
spectral components in the range where baroreflex
and respiratory inputs are the dominant effectors of
heart rate fluctuations (higher then 0.03 Hz).
During exercise sometimes an exponential trend is
subtracted.
ECG and/or blood pressure recordings before or
after exercise cause no particular problems. Best
practice is to perform these measurements in a quit
surrounding, at comparable timings for all subjects
in order to avoid circadian variations among
subjects. Usually, ECG is recorded as in clinical
practice, sometimes only RR-intervals are stored
with a wrist watch. For particular studies, requiring
day/night resolution or circadian variations, 24 h
Holter recordings are used.

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 14
As well Warren et al.
(131)
and Cottin et al.
(132)
concluded that HRV is a valid technique for non-
invasive measurement of parasympathetic activity
during exercise, but its validity as a measure of
sympathetic activity during exercise is equivocal.
The former concluded this from measurements
during exercise (progressive cycling tests at 40, 60
anf 80% of each subject’sheart rate reserve) with
infusion of saline, esmolol (beta-1 blocker),
glycopyrrolate (muscarinic blocker), or both drugs.
HF power decreased exponentially with workload
and was attenuated by glycopyrrolate and
combined treatments. The latter group showed
spectral analysis to confirm withdrawal of
parasympathetic control during graded exercise
load (25, 50 and 75% of VO
2max
), as the power
spectral density of the HF band significantly
decreased with exercise loads. However also the
LF power decreased with exercise load, suggesting
that LF and LF/HF is not a good indicator of
cardiovascular modulation during exercise.
5.1 HRV during exercise
It has since long been shown that during dynamic
exercise heart rate increases due to both a
parasympathetic withdrawal and an augmented
sympathetic activity
(127-128)
. The relative role of the
two drives depends on the exercise intensity
(120-125)
.
Arai et al.
(126)
were the first to test this hypothesis
with the aid of Fourier spectrum analysis of heart
rate time series in 43 normal subjects (range 25-69
years of age), who exercised until peak level. Their
data (Table 2) support a progressive withdrawal of
parasympathetic activity during exercise but no
changes in normalized values of LF and HF with
respect to rest and no correlation between LF
power and sympathetic activity have been
observed during muscular exercise.
Maciel et al.
(129)
came to similar conclusions. They
performed bicycle ergometer test in a group of 23
untrained subjects, at 3 levels (25W, 50W and
100W), before and after blockade with atropine or
propranolol. Their results showed that tachycardia
induced by dynamic exercise is mediated by a
biphasic mechanism initially depending on rapid
vagal release and an increased sympathetic
activity, especially at higher levels of exercising.
Kamath et al.
(133)
in his study compared orthostatic
stress (10 min supine followed by 10 min standing)
and exercising on a cycle ergometer (at 50% of
their maximum predicted power output) in a group
of 19 healthy untrained subjects (16 male, 3
females; 20-32 years of age). They found the same
significant decrease in the LF component due to
exercise, but an enhanced during orthostatic stress.
Therefore they concluded that humoral factors,
such as circulating catecholamines, probably play a
more dominant role in maintaining the tachycardia
during exercise instead of neurogenic control
which takes place during orthostatic stress. The
existence of a non-neural mechanism in the
reduction of the HF component was also supported
by a study from Casadei et al.
(134)
.
In a recent review article Brenner et al.
(130)
also
supports this hypothesis: at the onset of exercise,
heart rate is increased by a reduction in
parasympathetic activity and a temporally increase
in sympathetic tone. A continuation of physical
activity is associated with a continued withdrawal
of vagal activity and an attenuation of sympathetic
nervous system tone.
In contrast to Arai et al.
(126)
and other previously
mentioned authors, Perini et al.
(125)
performed
power spectral analysis (with an autoregressive
modelling) during steady-state exercise at different
intensities (3 levels: low at 50W, medium at 100W
and high at 150W) and during the corresponding
recovery periods in seven sedentary young males
(age: 23.7±0.9 years). They found only at low
exercise intensities no changes in the relative
power of the three components with respect to rest.
Above 30% VO
2max
, a marked decrease in LF
normalized power coupled to an increase in VLF%
was found. Their hypothesis was that above this
threshold additional mechanisms were involved in
cardiovascular adjustment and that a not negligible
portion of the power of HRV was in the VLF band
and that this component might reflect, at least in
part, the sympathetic activity. However, they also
mentioned a technical problem with the VLF
detection after trend removal. Therefore
conclusions about this component are maybe not
entirely justified.
Shin et al.
(135-136)
submitted 5 runners (18±2 years)
and 8 sedentary subjects (27±7 years) to a bicycle
ergometer exercise to the point of exhaustion. The
found that as well in athletes as in non-athletes LF
and HF gradually decreased with exercise
intensity. They suggested two possible reasons: a
marked absence of vagal modulation may have led
to reductions in LF accompanied by an influence
on the baroreflex (restored at higher operating
point or turned off), or hormonal factors. Possible
limitations of this study are: 1. the choice of order
for the AR analysis, which influences power
distribution over different bands, 2. the small
number of athletes (N=4), rather young and
compared to 3. an older non-athletes
(137)
population.

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 15
Table 3. Cross-sectional: athletes versus sedentary population
Author N Age Spectral
analysis
Remarks
(years)
Tonkins
(142)
39 21.2±3 No change 24 h Holter
39 time domain
Aubert
(211)
10 18-34 HF (increase) FFT
10 19-31
Verlinde
(30)
10 18-34 HF (increase) Wavelet
10 19-31
Dixon
(146)
10 22-33 HF (increase) AR
14 23-33
Goldsmith
(147)
8 24-38 HF (increase) 24 h Holter, FFT
8 24-38 sleeping and
awake
Furlan
(151)
21 16±0.6 LF (increase) trained
15 16±0.5 HF (increase) detrained
29 16±0.4
Jansen
(152)
18 19-32 LF (decrease) supine
11 23-33
N: number of subjects, LF: low frequency power, HF: high frequency power,
FFT: fast Fourier transform, AR: autoregressive method.
Last row for each author: sedentary comparison group
Yamamoto et al.
(138)
found an increase of LF
component with increasing exercise intensity. In
their study 6 healthy male volunteers performed
incremental exercise test on an electrically braked
cycle ergometer, consisting of a 5 min warm up
period at 50W, followed by work rate increment in
a ramp fashion until exhaustion. But these authors
used 0.0 Hz to 0.15 Hz as limits for the low
frequency bands. Therefore we cannot interpret
these data compared to previous ones, because
their LF component involves also the VLF
component as proposed by the Task Force
(24)
.
Parasympathetic activity of heart rate during
exercise was investigated with a time series
analysis by way of geometrical methods (Poincaré
plot) in a study in 31 subjects by Tulppo et al.
(139)
.
They showed that during recovery parasympathetic
activity decreased progressively until the
ventilatory threshold level was reached, when
sympathetic activation was reflected from changes
in the Poincaré plot. They concluded that poor
physical fitness is associated with an impairment
of cardiac parasympathetic function during
exercise and that their data support the concept that
good aerobic fitness may exert cardioprotective
effects by enhancing the cardiac parasympathetic
function during exercise.
A totally different technique to analyse heart rate
variability during exercise is proposed by Anosov
et al.
(140)
. They examined a group of 22 untrained
subjects (13 females, 9 males; 20-40 years of age)
on a cycle ergometer with a ramp load until
exhaustion. The authors were only interested in the
HF component. Therefore the tachogram was high-
pass filtered with a low frequency cut-off at 0.15
Hz. As ramp loading leads to non-stationary time
series, Fourier analysis was not applicable. To
obtain the instantaneous frequency of the HF
component of the HRV, the analytic signal
approach was used. This method consists in
constructing a complex function (the analytic
function) where the real part is the time series and
the complex part is the Hilbert transform of
previous time series. From this complex function
the amplitude and phase of the time series can be
obtained, and finally the instantaneous frequency is
the derivative of the instantaneous phase. They
concluded that the instantaneous frequency
component of the HF power of HRV and of the
respiratory signal developed in parallel during a
ramp load test. Both signals were closely linked,
showing a strong correlation between respiration
and heart rate. Due to this correlation, the HF
component of HRV was modified during ramp
load and in most cases can be used for the
detection of the ventilatory anaerobic threshold,
because the shift in instantaneous frequency of the
HF component occurred during the transition from
aerobic to anaerobic work. The modulation of
HRV in terms of its frequency is strong, even at

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 16
high physical activity levels, whereas the absolute
power of HRV is clearly reduced at high work
loads.
Gonzalez-Camarena
(141)
compared heart rate and
blood pressure variabilities during static and
dynamic (cycling at 30% and 60% of VO
2max
))
exercise in 10 subjects. They found a
parasympathetic withdrawal and sympathetic
augmentation during dynamic exercise and an
overall increase in HRV indices during static
exercise, suggesting an increased activity of both
autonomic branches.
HRV analysis during exercise remains a problem.
There are not so many studies and almost all of
them mention the technical problem of not dealing
with stationary time series. There is also a problem
related to the interpretation due to the
methodology. There are nearly as many protocols
proposed as there are papers written on this topic.
Methodology differs widely especially concerning
training intensity and/or exercise intensity, even in
some papers they are only vaguely mentioned.
Therefore it is strongly recommended to establish a
protocol, one for performing studies in sedentary
populations and another for athletes, while using
appropriate blockade mechanisms and different
analysis methods: time domain (and geometric
methods), spectral analysis and its variations
(Hilbert transform) and non-linear methods as
well.
Changes in HRV after exercise training
5.2 Cross –sectional studies: comparison of
athletic and sedentary groups
In this section, the differences between a sedentary
group and one or more groups of athletes (Table 3)
will be discussed as described in the literature.
Tonkins
(142)
reports a positive effect of time
domain parameters, as obtained from Holter
recordings in 39 trained athletes but did not find a
difference between aerobic trained and anaerobic
trained athletes. This is in contrast with results of
Aubert et al.
(143)
. They found significant higher
values of rMMSD and pNN50 between aerobic
trained athletes and anaerobic trained athletes or
rugby players, the latter are involved in combined
aerobic and anaerobic training. These differences
were also found in the frequency spectrum: larger
high frequency component in aerobic trained
athletes as well with FFT as with wavelet
analysis
(30)
. In an earlier study
(144)
they came to the
same conclusion: significantly higher rMSSD in 14
middle aged athletes, compared to a sedentary age
matched population (N=14, 35-55 years of age).
Many other studies confirm these findings for
young endurance trained athletes (mean age lower
than 30 year; disciplines: cycling, canoeing,
athletics, roller-skating, volleyball)
(145-146-147-148-149-
150)
. These studies concluded that endurance
training results in the enhanced vagal tone in
athletes, which may contribute in part to the lower
resting heart rate. Goldsmith
(147)
, who performed a
Holter study in 8 endurance trained athletes and
compared with 8 age matched untrained men,
suggests that aerobic exercise training may be a
useful adjunct or alternative to drug therapy in
lessening the derangements of autonomic balance
in many cardiovascular diseases.
In a combined RR-interval blood pressure study,
Macor et al.
(149)
concluded that competitive cycling
causes an enhanced parasympathetic drive to the
sinus node, whereas the neural control of blood
pressure is not affected. Furlan et al.
(151)
examined
two groups of endurance athletes: one group in rest
period (detrained athletes, 15 in total: 6 male, 9
female) and one group during peak season (21
swimmers, 14 male, 7 female). The latter had, in
contrast with the former group, elevated
sympathetic activity and higher parasympathetic
activity compared to a control group. They
concluded that the enhanced athletic performance
resulting from long-term training might depend on
an increase of both parasympathetic and
sympathetic modulation. Janssen et al.
(152)
compared athletes (18 cyclists, 19-32 years of age)
with 11 sedentary subjects (23-33 years of age) in
both supine and standing position. Spectral
analysis was performed with autoregressive
methods. Their measurements would suggest that
in the supine position, the sympathovagal balance
of the athletes differed from the control values,
caused by lowered sympathetic and/or increased
parasympathetic tone. This is mostly due to a
persistent sympathetic activation after exercise,
lasting up to 24 h, which is also studied in the same
work. They concluded that the differences in
autonomic control between the athletes and the
controls, were reflected in the quality (balance
between slow and fast heart rate fluctuations)
rather than in the quantity of heart rate variability.
De Meersman et al.
(153)
performed a cross-sectional
study for all age groups in 72 runners (15-83 years
of age) and in 72 sedentary controls matched for
age, body weight, blood pressure and social status.
HRV however was not determined from spectral
analysis but defined as % change of heart rate with
breathing (imposed breathing at 6/min). Although
no correlations with spectral component were
made, it can be assumed that this parameter is
related to the HF component of HRV. The

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 17
physically active group had significantly higher
levels of % change of heart rate, when compared
with their sedentary counterparts. These authors
concluded that habitual aerobic exercise augments
some parameters of HRV and could be a beneficial
modulator of heart rate variability in an ageing
population. They also suggested that this
augmented HRV in physically active individuals
provided further support for life-long aerobic
exercise as a possible non-pharmacological cardio-
protective therapy. However this statement remains
highly speculative, as it is not entirely supported
by their data.
All previous studies showed an increment in
parasympathetic activity due to an aerobic exercise
program. Some other studies
(154-155-156-157-69-158-108-
105)
did not find this positive effect on the
autonomic nervous system. Migliaro et al.
(69)
found
no differences in HRV (as determined from
spectral analysis: LF and HF) parameters between
sedentary (N=29, 15-24 years of age) and non-
sedentary young subjects (N=29, 15-24 years of
age). They also did not observe training
bradycardia which can probably explain their
observation.
A recent pharmacological blockade study by Stein
et al.
(108)
, with atropine and propranolol, caused
parallel shifts in the sinus automaticity of athletes
(6 runners, 29±4 years of age and 6 non-athletes
28±5 years of age). Increased parasympathetic
activity would cause greater heart rate response
post-atropine and a reduction in sympathetic
activity would cause lesser heart rate response
post-propranolol in athletes compared to non-
athletes. These conclusions were obtained after
electrophysiological studies of the conduction
system. The authors concluded that sinus
automaticity and AV node conduction changes of
endurance athletes were related to intrinsic
electrophysiology and not to autonomic influences.
The same group suggested earlier
(107)
that in
addition to its parasympathetic effects, athletic
training might induce intrinsic adaptations in the
conduction system (mostly by influencing
conduction velocity), which could contribute to the
higher prevalence of atrioventricular abnormalities
observed in athletes.
The latter study was in agreement with the results
of a blockade study of Smith et al.
(159)
who found
greater parasympathetic influence in endurance
trained subjects as well as lower intrinsic heart
rate, but in disagreement with all the studies as
mentioned in the first paragraph of this chapter and
Goldsmith
(160)
who indicated from their results that
physical fitness is strongly associated with vagal
modulation. Most studies
(146-148-150)
mention that the
higher parasympathetic activity is not the only
factor that contributes to the bradycardia in athletes
but that it is only a part of the lower heart rate. All
these studies point to endurance training as an
effector of enhanced parasympathetic activity in
athletes, which may contribute in part to the resting
bradycardia. Katona et al.
(155)
already found in
1982 that lower resting heart rate in endurance
trained athletes (8 world class oarsmen) is solely
due to a reduction in intrinsic cardiac rate, and not
to an increase in parasympathetic tone. They
showed it by using pharmacological blockade
(propranolol and atropine) to suppress either
sympathetic or parasympathetic activity of the
autonomic nervous system. Also, Bonaduce et
al.
(161)
came to the conclusion that other
mechanisms than changes in cardiac autonomic
control could be involved in determining the
profound bradycardia of athletes.
Another possible reason for the controversial
results concerning ANS activity in athletes is due
to a disturbance on the LF power caused by
respiration. This was shown in a study from Strano
et al.
(162)
comparing controlled versus paced
breathing. A slow breathing rate, which is a
common feature in athletes, caused the HF and LF
components to overlap, leading to a
misinterpretation of the LF power. ECG was
recorded in supine position in athletes, while they
were breathing at their spontaneous frequency and
at rates of 15, 12 and 10 to 14 (in random order)
breaths/min (corresponding to 0.25, 0.2, 0.16 and
0.23 Hz). Uncontrolled and random breathing rates
significantly altered spectral sympathetic indices.
On the other hand, 15 and 12 breaths/min
redistributed respiratory related power through the
HF, thus yielding correct LF power estimation.
The authors conclude and recommend to
standardize respiration at 0.25 Hz (15 breaths/min)
in athletes for assessing ANS activity.
A possible hypothesis as to the controversy about
autonomic versus non-autonomic determinants of
electrophysiological adaptations in athletes could
be a fundamental difference between short-term
and long-term physical training programs
(108)
.
Short-term training, as in prospective studies,
could induce autonomic adaptations, with a
reduction in sympathetic activity and an increase in
parasympathetic activity (leading to bradycardia).
On the other hand, long-term aerobic training,
eliciting atrial and ventricular dilation, would
induce intrinsic electrophysiological adaptations
and enhance parasympathetic activity.

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 18
Table 4 Effect of training on a sedentary population.
Author N Age Duration repetions TP LF HF Remarks
(years) (weeks) (/week) (ms
2
) (ms
2
) (ms
2
)
Melanson
(157)
11 25-45 16 3 234 398 before
416 798* after training
5 25-45 No training 173 331 before
169 446 after 16 weeks
Loimaala
(171)
26 35-55 20 863 321 control/before
829 391 after training
26 4 to 6 1212 572 before
1300 659 jogging/walking
55%
28 4 to 6 846 317 before
1054 478 jogging 75%
Catai
(154)
10 19-21 12 3 1821 818 277 Before/awake
2870 1048 429 Jogging/walking
70-85% peak HR
7 50-59 12 3 2601 687 265 before
2942 513 253 after
10 19-21 12 3 4862 1030 2589 Before/asleep
3152 930 1374 after
7 50-59 12 3 1225 357 342 before
1584 502 488 after
*: p<0.05
N: number of subjects, TP: total power, LF: low frequency power, HF: high frequency power
Melanson and Loimaala: values LF and HF transformed from log.
5.3 Longitudinal: effect on HRV of exercise
training of non-athletes.
Beneficial effects of physical training have been
reported in post-myocardial patients
(163-164)
and in
heart transplant patients
(165)
. Therefore it can
hypothesised that exercise training would be
effective in improving the autonomic balance in a
general public while developing physical fitness as
well.
Melanson and Freedson showed influence of
exercise training on HRV parameters on a young
(11 subjects, 25-40 years of age) male
population
(166)
. The subjects performed moderate to
vigorous intensity stationary cycling on 3 days
each week for 30 min per session. In their study
they showed that a moderate-to-vigorous-intensity
endurance training program in adult, previously
sedentary men increased markers of cardiac
parasympathetic activity after 12 weeks. This was
proven by a significant increase in HF power after
training (Table 4) and a significant increase in time
domain parameters related to parasympathetic
activity (pNN50 and rMSSD) as well.
Boutcher and Stein
(167)
found no change in HRV in
a group of 19 middle aged men (46.2±1.6 years of
age) compared to an age-matched control group
(N=15). HRV was assessed after 24 exercise
sessions of moderate intensity exercise training
(during 8 weeks). The subjects exercised 3 times
each week at an intensity of 60% of heart rate,
determined through baseline at maximal exercise
heart rate. The exercise session consisted of a 0.25
mile walking warm up, a series of stretches, an
aerobic exercise period (20 min for the first 3
sessions, 15 for the next 3, and 30 min for 7 to 24),
a 0.25 mile cool down walk and a repeat of the
stretching. LF and HF components were obtained
after band pass filtering of the tachogram and
variance was determined in these bands. In the
exercise group VO
2max
increased (12% absolute
value) after the training period, but without altering
HRV. These results show that a short duration and
moderate intensity aerobic training in a middle-
aged population, is insufficient to alter HRV
parameters in that age group.
The same conclusion was reached by Perini et
al.
(168)
in a training program in an elderly

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 19
population of 7 men and 8 females (73.9±3.5 years
of age). They reported no changes in HRV
parameters after an intense 8-week aerobic training
program. However after a training program of 6
months (3 aerobic training sessions a week lasting
45 min) in an elderly population (51 older men and
women, 67±5.1 years of age), Schuit et al.
(169)
found a general increase in HRV, after a training
program of 6 months (3 times a week aerobic
supervised training) in an elderly population (51
men and women, 67.0±5.1 years of age).
Theyshowed specifically that the very low and LF
power bands, were significantly increased
compared with a control group. Their conclusion:
“In older subjects physical training may be an
effective means to modify positively a factor that is
associated with increased incidence of cardiac
events.”, however is questionable as LF power is
associated with arrhythmogenic activity and low
LF in pre-menopausal women is
cardioprotective
(170)
.
Again in a 5 month duration aerobic training
program in 83 middle aged (35-55 years of age)
men, Loimaala et al.
(171)
found no changes in HRV
parameters in both time and frequency domains
(Table 4). Subjects were trained 4-6 times a week
during 30 min in two different groups: 1. jogging
at a heart rate level corresponding to 55% of the
VO
2max
measured at baseline; 2. jogging at an heart
rate level corresponding to 75% of the VO
2max
measured at baseline. Indices reflecting tonic
parasympathetic outflow (SDNN, pNN50 and HF
power) did not change significantly during the
intervention. They concluded: “exercise training
was not able to modify the cardiac parasympathetic
activity in sedentary, middle-aged persons.”
No consistent changes were observed in BRS,
although a significant reduction in heart rate was
found. The authors blame the short duration of the
training program and suggest that in order to obtain
any effect on HRV it should last for a period of a
year at least
(171)
.
Many factors affect the physiological significance
of these studies. One of the most important is the
age factor, which contributes to the discrepant
findings in the literature. It is well known that
HRV parameters are decreasing with ageing
(170)
(and a function of gender as well). Exercise
training studies in young adults
(172)
generally report
increases in measures of HRV, whereas studies in
middle-aged
(167)
and older adults
(173)
show no
changes in cardiac autonomic function, as
determined from HRV.
Duration and intensity of training, the accent of the
program even gender distribution, also vary widely
among different studies. In one study it was even
suggested that endurance should be practiced for a
prolonged period, even extending over many
years
(171)
in a middle-aged population. On the other
hand in a young population (20-22 years of age)
we have seen (Figure 5) some influence on HF of
HRV after only 6 weeks of training (unpublished
data). In most of the studies dynamic exercise is
performed, however in some studies also static
training is used
(141)
. This again, if not taken into
account, can lead to differing conclusions. A last
factor is usually the small number of subjects in
the training program. The effect working with
small numbers is to reduce the statistical power,
making it more difficult to detect differences due
to the training. Therefore, whether age or other
factors would modulate the effects of training on
HRV parameters is still unclear and is an area for
further investigation.
Figure 5. Top: Tachogram and power spectral
density of a recording in a young sedentary
subject before training (HF=812,3 ms
2
) and
bottom: the same subject after a 6 months
aerobic training program (HF=1878.4 ms
2
).
Levy et al.
(174)
submitted an elderly (N=13, 60-82
years of age) and a younger population (N=11, 24-
32 years of age) to a 6 months aerobic training
program (walking, jogging and bicycling). The
subjects trained as follows: 10 min warm up, 45
min exercise and 10 min cool down. Training
began at 50 to 60% of heart rate reserve and
increased to 80 to 85% by the 10
th
month. However
HRV was only measured as SD (ms) of all normal

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 20
5.4 Differences due to age and gender
RR intervals during a 2 min acquisition. They
found an increase of this parameter of 68% in the
elderly and of 17% in the younger population.
Their conclusion was that exercise training
increases parasympathetic tone in both the healthy
older and young men. However it has been
proven
(18)
, and it is a mathematical law (Parseval’s
theorem), that SD corresponds to total power and
as such is a combination of sympathetic and of
parasympathetic activity and as such, their
conclusion is wrong, or at least an overstatement.
A contribution by each division of autonomic
modulation to HRV is only possible when this
variable is represented in the frequency domain as
a power spectral density graph.
Many studies with a large number of subjects
groups, have focused on the influence of age and
gender on cardiac autonomic tone (HRV
parameters)
(176-177-68-178-179-180)
. The general
conclusion of these studies was that: 1. ageing
reduced the global measure of HRV, at rest, in
general and of both its spectral components (LF
and HF) as well. Therefore this decline might
reflect reduced responsiveness of autonomic
activity with age. 2. All HRV parameters, except
for HF power were higher in men and this gender
difference was confined to the age categories less
than 40-50 years. The lower sympathetic tone (LF)
in women might provide protection against
arrhythmias and the development of coronary heart
disease.
Catai et al.
(154)
also trained an elderly (50-59 years
of age) and a younger (19-24 years of age)
population and reported HRV values obtained in
the frequency domain, awake and during sleep
(Table 4). The training programs were conducted
for 3 months on a field track and included
stretching for 10 min followed by walking and/or
jogging for 40 min; 3 times a week at a prescribed
heart rate corresponding to 70-80% of peak heart
rate. The authors found no significant changes in
HRV associated with an increase in aerobic
capacity induced by aerobic training. They
concluded that resting bradycardia induced by
short-term aerobic training in both young and
middle-aged men is more related to intrinsic
alterations in the sinus node than to changes in
efferent parasympathetic-sympathetic modulation.
As they mentioned, however, the primary goal of
the experimental design was directed to evaluate
the cardiorespiratory adaptation in short-term
training: they only used two 1000 s epochs out of a
24 h Holter recording (awake and asleep). The
training period was very short (12 weeks) with a
small number of subjects.
A potential confounding effect of the menstrual
cycle can arise in studies that address gender
differences in HRV parameters. Effects of the
menstrual cycle have been shown on cardiac
autonomic function as assessed by HRV
methods
(181-182-183)
and even of hormonal
replacement therapy in post-menopausal
women
(183)
. All studies agreed that regulation of
the autonomic tone is modified during menstrual
cycle. The alteration in the balance of ovarian
hormones might be responsible for these changes
in the cardiac autonomic activity. Results suggest
that parasympathetic nerve activity is predominant
in the follicular phase. Unfortunately in the few
gender studies concerning young female athletes,
no mention is made of timing within the menstrual
cycle.
This view is also supported in a study from
Boutcher et al.
(184)
and confirmed by Davy et al.
(185)
and McCole et al.
(113)
who found that older women
athletes (postmenopausal women), who had
habitually performed vigorous endurance training,
had higher stroke volume and cardiac outputs
during maximal exercise, than their sedentary
postmenopausal peers. On the other hand in young
female athletes, similar results are found compared
to their male counterparts. Pigozi et al.
(186)
performed a 24-h Holter study (spectral analysis
with AR) in 26 highly trained female athletes
(24.5±1.9 years). They were assigned to a 5-week
aerobic training program during a yearly rest
period. They concluded that from the relative
night-time increase in LF and the decrease in the
day-night difference in time domain indexes,
exercise training is able to induce an increase in
the sympathetic modulation of the sinus node,
coexisting with signs of reduced or unaffected
To conclude this paragraph it can be stated that
there are conflicting reports in the literature
concerning the effects of aerobic training in a
general population on HRV parameters under
resting conditions. While some studies have
reported an increase in the magnitude of HRV in
the time domain
(175)
, in the frequency domain
others have reported absence of modifications
(167)
,
and an increase
(151)
or decrease
(150)
of
sympathovagal balance in the sinus node.
Therefore studies of aerobic training effects on
HRV parameters on a previously not-trained
(young and/or elderly) population still remain
necessary, preferably under well controlled
conditions.

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 21
Table 5. HRV parameters in elderly athletes and age matched sedentary population
Author N Age LF HF Remarks
(years) (ms
2
) (ms
2
)
Holter 24 H
Yataco
(191)
15 69±7 891* 575* athletes
14 537 102 sedentary
Banach
(192)
9 52.9±7.2 1088* 920* athletes
9 52.9±7.2 220 294 sedentary
Jensen-
Urstad
(148)
11 73.2±2.8 673±244 353±349 athletes/24 h Holter
12 74.5±2.7 492±290 209±172 sedentary
764±327 475±654 athletes/night
728±485 328±48 sedentary
587±250* 267±163* athletes/day
346±177 127±41 sedentary
*: p<0.05
N: number of subjects, LF: low frequency power, HF: high frequency power
Yataco: values LF and HF transformed from log
vagal modulation in this group of young female
athletes.
A gender difference was obtained by Hedelin et
al.
(187)
in junior athletes. They compared short-term
HRV recordings (AR power spectrum) in 17 cross-
country skiers (9 females, 8 males, 16-19 years of
age) before and after the competitive season. After
the intensive training/competition season there was
a general increase in HRV. No difference in resting
heart rate was found, pre- and post season.
However in females they found a higher level of
parasympathetic activity than in males, reflected
by a consistently higher HF and total variability.
A difficulty comparing previous data is that 1.
training level is different, 2. training duration is
different: short-term or long term effects, 3.
duration of ECG recording is different: 24 h Holter
recordings versus short duration ECG recordings.
In general the literature proposes three conclusions
concerning ageing: 1.cardiovascular and
cardiorespiratory function are higher in elderly
athletes than in age comparable sedentary
groups
(188)
, 2. the capacity for significant function
in endurance and power persists throughout life in
trained individuals, 3. strength decreases more
rapidly than endurance
(189)
.
How do these findings in a general population
relate to the ageing athlete?
There are many physiological, structural and
psychological differences, which distinguish
elderly athletes from younger ones and from a
similar aged sedentary group, especially if still
active. Regular exercise may be able to retard the
physiological decline
(190)
.
This is supported by the very few HRV studies
performed in senior athletes so far (Table 5).
Yataco et al.
(191)
determined the age-associated
decline in HRV with decreases in HRV by
comparing HRV parameters in older athletes
(N=15, 69±7 years of age) with age-matched
sedentary persons (N=14, 69±4 years of age). They
showed positive correlations between HRV
parameters and aerobic fitness (as determined from
maximal treadmill exercise). Frequency analysis
was performed after Holter monitoring. Senior
competitive athletes had increased HRV and
parasympathetic heart rate activity (Table 5) when
compared with their sedentary counterparts. This
work support the hypothesis that the age associated
decline in HRV parameters is due in part to
lifestyle and not solely to ageing. Similar results
were shown in the study by Banach et al.
(192)
:
higher HRV parameters in middle-aged athletes
compared to a sedentary population (Table 4),
indicating that the autonomic activity in sportsmen
is not affected by ageing to the sixth decade of life.
Jensen-Urstad et al.
(193)
on the other hand showed
that elderly athletes (N=11, 73.2±2.8 years of age)
with a lifelong training history seem to have more
complex arrhythmias and profound brady-
arrhythmias than do healthy elderly controls,
which may increase the risk of sudden cardiac
death. In contrast however, the age-related
decrease in HRV seems also retarded as in
previous studies (Table 5). The latter has a positive
prognostic value and may decrease the risk of life
threatening ventricular arrhythmias.
Results from the few HRV studies in elderly
athletes all point in the same direction: the decline

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 22
in HRV parameters associated with age is
overcome to some extend by sustained endurance
training into high age. However, more studies are
needed, especially to show any beneficial effect of
lifelong regular training on quality of life and on
life expectancy.
5.5 Over-training and the autonomic nervous
system
In athletic training, workloads are gradually
increased, thereby exceeding the previously
employed workload. This ‘overload’ principle is an
important component of modern training
(194)
and is
a positive stressor that can be quantified according
to load, repetition, rest and frequency
(2)
.
Application of too great training stress and too
frequent training sessions can result in exhaustion
of a physiological system. ‘Over-training
syndrome’ or ‘staleness’ in athletes results from
long-term stress or exhaustion due to prolonged
imbalance between training and other external and
internal stressors and recovery
(195-196-197)
.
It is well known that over-training causes
hormonal imbalance
(198-199)
in athletes. Due to these
hormonal changes, over-training will lead to an
autonomic imbalance
(200-201-202)
. In which way the
autonomic nervous system changes
(parasympathetic versus sympathetic) is still
controversial. From a clinical standpoint, Israel
(201)
distinguished between a parasympathetic type or
vagal type (Addison type) over-training syndrome
and a sympathetic type (Basedow type). The two
types of over-training were the consequence of an
imbalance between training and rest periods, but it
was expected that a sympathetic type over-training
syndrome might rather be the consequence of too
much accompanying psycho-emotional stress, such
as too many competitions and too many non-
training stress factors (social, educational,
nutritional,…)
(200)
. Kuipers
(198)
hypothesized that
during the early stage of the over-training
syndrome, the sympathetic system was
continuously altered, whereas during advanced
over-training the activity of the sympathetic system
was inhibited, resulting in a marked dominance of
the parasympathetic system.
Hedelin et al.
(203)
investigated nine canoeists (6
men and 3 women, 18-23 years of age) before and
after a training regimen corresponding to a 50%
increase in normal training load applied during 6
days. Heart rates reduced (-5 to –8 bpm) both at
sub-maximal and maximal levels which could be
due to hypervolemia leading to increased stroke
volume and maintenance of cardiac output with
lower heart rates. Unlike these changes in heart
rate, no significant differences were found in HRV
parameters, neither when stressing the
parasympathetic system (controlled breathing) nor
when stressing the sympathetic system (tilt test: the
subject starts in a supine position on a special bed
that and the subject is raised passively to an angle
of 60°). So, they concluded that these HRV data
did not support an altered autonomic balance in
these athletes. A possible explanation could be that
a 6 day training period has only a small effect on
individual HRV parameters and also that group
differences will be difficult to determine in small
groups. A case study of the same authors
(204)
in a
cross-country skier, showed a relative
parasympathetic dominance when the athlete was
over-trained.
We suggest that it is impossible to find group
changes in HRV because of the two types of over-
training. Individual HRV, however, can change
due to over-training. These hypotheses were
confirmed by Uusitalo et al.
(205)
who investigated
HRV and BPV of young female athletes during 6-9
weeks of training period. They compared a high
intensity training group (4 long distance runners, 1
cross country skier, 2 triathletes, 1 orienteer), with
a low intensity training group (1 long distance
runner, 3 cross country skiers, 1 triathlete, 1
orienteer). The purpose of the experimental
training period was to over-train this group after a
period of 6 to 9 weeks. Heavy endurance training
seemed to induce a significant (Table 6) increase in
LF of HRV during supine position, but not in the
low intensity training group. In many subjects the
changes in supine and standing heart rate
variability seemed to be rather contrary. Since
there were no large uniform findings in the over-
trained athletes the authors looked also at
individual results during a tilt-test. Increased as
well as decreased changes due to upright tilt were
found in the over-trained athletes compared to their
values in the normal training state. This is a sign of
either increased or decreased ability to increase
sympathetic discharge during standing and
corresponds to the two over-training types.
However, the changes were not specific to over-
The cardiac autonomic imbalance observed in
over-trained athletes implies changes in HRV and
therefore would suggest that heart rate variability
could provide useful parameters to detect over-
training in athletes. Despite these expectations,
little is known about changes in heart rate
variability due to over-training and only a few
studies are available (Table 6).

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 23
Table 6. Effect of over-training on HRV parameters
Author N Age TP LF HF Remarks
(years) (ms
2
) (ms
2
) (ms
2
)
Hedelin
(204)
9 18-23 3.71±0.29 2.9±0.57 3.4±0.27 control
3.66±0.26 2.77±0.29 3.36±0.43 Over-trained
Uusitalo
(205-212)
6 19-27 5100±900 800±200 2800±700 control
8600±3700 700±200 5600±3200 light training
9 20-27 5500±100 600±100 2700±600 control
5500±1200 900±200* 2900±700 Over-training
N: number of subjects, TP: total power, LF: low frequency power, HF: high frequency power
Values from Hedelin have been log transformed, mean ±SD, *: p<0.05
Values from Uusitalo: absolute values mean ±SEM
training because there also were similar changes in
the not over-trained athletes.
Portier et al.
(206)
tested 8 runners twice : after a
relative rest period of 3 weeks and after a 12 week
intense training period for endurance and each time
determined HRV parameters. Although the athletes
were not trained until over-training, they
concluded that spectral analysis could be a means
of demonstrating impairment of autonomic balance
for the purpose of detecting a state of fatigue that
could result in over-training. Pichot et al.
(207)
came
to similar conclusions. They assessed ANS activity
in 7 middle distance runners (24.6±4.8 years)
during their training cycle: 3 weeks heavy training,
followed by a relative resting week. HRV was
analysed using FFT and wavelet transform. Their
results confirmed that heavy training shifted the
cardiac autonomic balance of the sympathetic over
the parasympathetic drive. Night–time results of
HRV parameters proved a good tool to estimate
cumulated physical fatigue. Therefore they
concluded that HRV could be valuable for
optimizing individual training profiles.
Concerning the use of HRV methods during over-
training in athletes, no definite conclusions can be
reached as only very few studies are available,
even so with conflicting results. It remains to be
proven that the autonomic imbalance observed in
over-trained athletes, manifests itself from HRV
indices.
6. Conclusions
Innumerable studies have been published
concerning training in general (computer search on
the keyword “training” results in 409395 hits)
concerning physical and physiological condition of
athletes. However, only very few papers are
dealing with studies of HRV regarding applications
in athletics (117 hits). Therefore, cardiovascular
variability studies in athletes are still an almost
unexplored domain. Much work still needs to be
done to advance in understanding of the action of
the autonomic nervous system in athletes as a
function of athletic discipline, age, gender,
intensity and duration of training, detraining and
over-training effects, comparison with sedentary
population, and so on.
Another key issue is that almost no studies are
available as a longitudinal section for the follow-
up of athletes during ageing, as well as very few
studies about active elderly athletes.
For further studies it is recommended to apply
standardized conditions: 1. selection of subjects:
age, gender, training or physical fitness level,
athletic discipline and accent on aerobic or
anaerobic training; 2. measurements: minimal
number of parameters proposed: ECG, (non-
invasive) blood pressure, eventually respiration; 3.
measurements at rest with a minimum of 10
minutes supine and 10 minutes standing, to
activate the sympatho-vagal balance, eventually
breathing at different fixed frequencies, to activate
primarily the parasympathetic system, 24 h Holter
monitoring when day to night separation is needed
for circadian pattern detection; 4. measurements
during exercise: either with adapted trend removal
or else at constant work levels in order to have
stationary signals.
For interpretation of the data in as well time as
frequency domain the use of the guidelines
(24)
are
recommended in order to be able to compare
different studies.
It is strongly suggested that, when presenting
reports on HRV studies related to exercise
physiology in general or concerned with athletes, a
detailed description should be provided on analysis
methods, as well as concerning population, training
schedule, intensity and duration. Only with such
information will it be possible to understand and

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 24
evaluate conclusions drawn and compare results
with other studies. As until now this is not the case
in most studies on HRV in athletes as found in the
literature, it is only possible to make general
comparisons.
Most studies concern relatively small numbers of
subjects, diminishing the power of statistics. It is of
course not so easy to find and motivate large
numbers of athletes to participate in scientific
studies: the usual answer (especially from coaches)
is that the athletes should train and refrain from
loosing time on other topics such as specific
physiologic measurements. Therefore, multicentre
studies would be preferable: 1. to enhance the
value of the study and motivate the subjects, 2. to
increase significantly the number of participants.
This would also facilitate a multidisciplinary
approach between cardiologists, exercise
physiologists, pulmonary physiologists, coaches
and biomedical engineers needed to evaluate the
many different and interrelated aspects of
cardiovascular variability in the athlete.
In order to further develop this fascinating research
field, we advocate prospective, randomised,
controlled, long term studies using validated
measurements methods. However, there is a strong
need for basic research on the nature of the control
and regulating mechanism exerted by the ANS on
cardiovascular function in athletes, needed for
better understanding of this phenomenon.
It remains a question whether aerobic exercising
helps in maintaining and developing
cardiovascular fitness in a general population
(208)
,
or whether it can be used as a predictor of
mortality
(80)
, or whether physical activity has
beneficial effects on the cardiovascular risk
profile
(209)
, or in other words: is life-long
exercising cardioprotective? Fact is that physical
activity has a great impact on quality of live
improvement of all those involved in athletic
activities.
Finally: study of cardiovascular variability (HRV
and BPV) is a potentially powerful method as a
basic scientific tool for better understanding the
regulation and control of the cardiovascular
system. From the practical point of view it remains
to be determined if it can also be used as a
predictor of athletic condition
(210)
and of athletic
achievements.
Acknowledgement
We thank all the athletes and subjects who
participated to some of the research projects from
our laboratory.
We thank Bart Verheyden for his suggestions and
for carefully reading the manuscript.
References
1. Vatner, S. F. and Pagani, M. Cardiovascular
adjustments to exercise: hemodynamics and
mechanisms. Prog.Cardiovasc.Dis.1976; 19
(2): 91-108
2. Brooks, G. A., Fahey, T. D., and White, T. P.
Exercise physiology. Human bioenergetics and
its applications. ; 2nd Ed, Mayfield Publ
Comp, Mountain View CQA, USA, 1995:
3. Fagard, R. H. Impact of different sports and
training on cardiac structure and function.
Cardiol.Clin.1992; 10 (2): 241-56
4. Fagard, R., Aubert, A. E., Lysens, R.,
Staessen, J., Vanhees, L., and Amery, A.
Noninvasive assessment of seasonal variations
in cardiac structure and function in cyclists.
Circulation 1983; 67 (4): 896-901
5. Fagard, R., Aubert, A. E., Staessen, J., Eynde,
E. V., Vanhees, L., and Amery, A. Cardiac
structure and function in cyclists and runners.
Comparative echocardiographic study.
Br.Heart J. 1984; 52 (2): 124-9
6. Hainsworth, R. Physiology of the cardiac
autonomic system.Ch 1, In: Clinical guide to
cardiac autonomic tests. Ed: M Malik, Kluwer
Ac Publ, 1998, Dordrecht, The Netherlands
7. Levy M.N. and Martin P.J. Neural control of
the heart. In: Handbook of physiology. Ed
Berne RM pp. 581-620, American
Physiological Society, 1979
8. Persson, P. B. Modulation of cardiovascular
control mechanism and their interaction.
Phys.Rev.A 1996; 76: 193-244
9. Natelson B.H. Neurocardiology, an
interdisciplinary area for the 80s. Arch Neurol
1985; 42: 178-84
10. Malik, M. and Camm, A. J. Heart rate
variability. Clin.Cardiol.1990; 13 (8): 570-6
11. Aubert, A. E. and Ramaekers, D.
Neurocardiology: the benefits of irregularity.
The basics of methodology, physiology and
current clinical applications. Acta
Cardiol.1999; 54 (3): 107-20
12. Pumprla J., Howorka K., Groves D., Cheste
M., and Nolan J. Functional assessment of
heart rate variability: physiological basis and
practical applications. Int J Cardiol 2002; 84:
1-14
13. Ramaekers, D. Effects of melanocortins and
N-terminal proopiomelanocortin on
cardiovascular function and autonomic

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 25
dynamics.Acta Biomedica Lovaniensia, pp.
130, Leuven University Press, 1999, Leuven,
Belgium.
14. Akselrod, S., Gordon, D., Ubel, F. A.,
Shannon, D. C., Berger, A. C., and Cohen, R.
J. Power spectrum analysis of heart rate
fluctuation: a quantitative probe of beat-to-beat
cardiovascular control. Science 1981:10; 213
(4504): 220-2
15. Eckberg, D. L. and Fritsch, J. M. Human
autonomic responses to actual and simulated
weightlessness. J.Clin.Pharmacol.1991; 31
(10): 951-5
16. Beckers, F., Ramaekers, D., and Aubert, A. E.
ACTS: automated calculation of tachograms
and systograms. Prog.Biomed.Res.1999; 4:
160-5
17. Kamath, M. V. and Fallen, E. L. Power
spectral analysis of heart rate variability: a
noninvasive signature of cardiac autonomic
function. Crit Rev.Biomed.Eng 1993; 21 (3):
245-311
18. Aubert, A. E., Ramaekers, D., Beckers, F.,
Breem, R., Denef, C., Van de, Werf F., and
Ector, H. The analysis of heart rate variability
in unrestrained rats. Validation of method and
results. Comput.Methods Programs
Biomed.1999; 60 (3): 197-213
19. de Boer, R. W., Karemaker, J. M., and
Strackee, J. Comparing spectra of a series of
point events particularly for heart rate
variability data. IEEE Trans.Biomed.Eng
1984; 31: 384-7
20. de Boer, R. W., Karemaker, J. M., and
Strackee, J. Spectrum of a series of point
events, generated by the integral pulse
frequency modulation model. Med.Biol.Eng
Comput.1985; 23 (2): 138-42
21. Parati, G., Casadei, R., Groppelli, A., Di
Rienzo, M., and Mancia, G. Comparison of
finger and intra-arterial blood pressure
monitoring at rest and during laboratory
testing. Hypertension 1989; 13 (6 Pt 1): 647-55
22. Mulder L.J.M., Veldman J.B.P., Ruddel H.,
Robbe H.W.J., and Mulder G. On the
usefulness of finger-pressure measurments for
studies on mental workload. Homeostasis
1991; 33: 47-60
23. Sato T., Nishinaga M., Kawamoto A., Ozawa
T., and Takatsuji H. Accuracy of a continuous
blood pressure monitor based on arterial
tonometry. Hypertension 1993; 21: 866-74
24. Heart rate variability: standards of
measurement, physiological interpretation and
clinical use. Task Force of the European
Society of Cardiology and the North American
Society of Pacing and Electrophysiology.
Circulation 1996: 93 (5): 1043-65
25. Kleiger, R. E., Stein, P. K., Bosner, M. S., and
Rottman, J. N. Time domain measurements of
heart rate variability. Cardiol.Clin.1992; 10
(3): 487-98
26. Aubert, A. E., Ramaekers, D., Beckers, F.,
Breem, R., Ector, H., and Van de Werf, F.
TIFAHR: Time- and Frequency Analysis of
Heart Rate Variability. Pitfalls and
misinterpretations.1998; 323-7
27. Malik, M. and Camm, A. Heart rate
variability.Futura Publ Co, Armonk, New
York, 1995;
28. Akselrod, S. Spectral analysis of fluctuations
in cardiovascular parameters: a quantitative
tool for the investigation of autonomic control.
Trends Pharmacol.Sci.1988; 9 (1): 6-9
29. Bartoli, F., Baselli, G., and Cerutti, S. AR
identification and spectral estimate applied to
the R-R interval measurements.
Int.J.Biomed.Comput.1985; 16 (3-4): 201-15
30. Verlinde, D., Beckers, F., Ramaekers, D., and
Aubert, A. E. Wavelet decomposition analysis
of heart rate variability in aerobic athletes.
Auton.Neurosci.2001: 20; 90 (1-2): 138-41
31. Cottin, F., Billat, L., Durbin, F., Koralsztein
J.P., and Papelier Y. Effects of fatigue on
spontaneous velocity variations in human
middle-distance running: use of short-term
Fourier transform. Eur J Appl Physiol 2002;
87: 17-27:
32. Fagard, R. H., Pardaens, K., Staessen, J. A.,
and Thijs, L. Power spectral analysis of heart
rate variability by autoregressive modelling
and fast Fourier transform: a comparative
study. Acta Cardiologica 1998; 53 (4): 211-8
33. Mansier, P., Clairambault, J., Charlotte, N.,
Medigue, C., Vermeiren, C., LePape, G.,
Carre, F., Gounaropoulou, A., and
Swynghedauw, B. Linear and non-linear
analyses of heart rate variability: a minireview.
Cardiovasc.Res.1996; 31 (3): 371-9
34. Chess GF., Tam RMK., and Calarescu FR.
Influence of cardiac neural inputs on rhythmic
variations of heart period in the cat. Am J
Physiol 1975; 228: 775-80
35. Convertino, V. A. G-factor as a tool in basic
research: mechanisms of orthostatic tolerance.
J.Gravit.Physiol 1999 Jul; 6 (1): 73-6
36. Selman A., McDonald A., Kitney R., and
Linkens D. The interaction between heart rate
and respiration: part I - experimental studies in
man. Automedica 1982; 4: 131-9

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 26
37. Eckberg, D. L. Sympathovagal balance: a
critical appraisal. Circulation 1997; 96 (9):
3224-32
38. Malliani, A., Pagani, M., Lombardi, F., and
Cerutti, S. Cardiovascular neural regulation
explored in the frequency domain. Circulation
1991; 84 (2): 482-92
39. Bigger, J. T., Jr., Steinman, R. C., Rolnitzky,
L. M., Fleiss, J. L., Albrecht, P., and Cohen, R.
J. Power law behavior of RR-interval
variability in healthy middle-aged persons,
patients with recent acute myocardial
infarction, and patients with heart transplants.
Circulation 1996; 93 (12): 2142-51
40. Kobayashi, M. and Musha, T. 1/f fluctuation of
heartbeat period. IEEE Trans.Biomed.Eng
1982; 29(6): 456-7
41. Saul, J. P., Arai, Y., Berger, R. D., Lilly, L. S.,
Colucci, W. S., and Cohen, R. J. Assessment
of autonomic regulation in chronic congestive
heart failure by heart rate spectral analysis.
Am.J.Cardiol.1988; 61 (15): 1292-9
42. Wagner, C. D. and Persson, P. B. Chaos in the
cardiovascular system: an update.
Cardiovasc.Res.1998; 40 (2): 257-64
43. Pagani, M., Furlan, R., Dell'Orto, S., Pizzinelli,
P., Lanzi, G., Baselli, G., Santoli, C., Cerutti,
S., Lombardi, F., and Malliani, A. Continuous
recording of direct high fidelity arterial
pressure and electrocardiogram in ambulant
patients. Cardiovasc.Res.1986; 20 (5): 384-8
44. Beckers, F., Ramaekers, D., and Aubert, A. E.
Nonlinear dynamics in heart rate variability.
IEEE Comp.Cardiol. 2000; 27: 131-4
45. Aubert, A. E. and Ramaekers, D. Heart rate
variability: is there more information in
spectral analysis or in methods of nonlinear
dynamics? Med.Biol.Eng Comput.1999; 37
(sup 2): 378-9
46. Lombardi, F., Sandrone, G., Pernpruner, S.,
Sala, R., Garimoldi, M., Cerutti, S., Baselli, G.,
Pagani, M., and Malliani, A. Heart rate
variability as an index of sympathovagal
interaction after acute myocardial infarction.
Am.J.Cardiol.1987; 60 (16): 1239-45
47. Beckers, F. Non-linear dynamics of
cardiovascular variabilty. Validation and
clinical applications. Acta Biomedica
Lovaniensia pp. 155, Leuven University Press
2002, Leuven, Belgium
48. Schmidt, G. and Morfill, G. E. Complexity
diagnostics in cardiology: fundamental
considerations. Pacing
Clin.Electrophysiol.1994; 17 (6): 1174-7
49. Nakamura, Y., Yamamoto, Y., and Muraoka, I.
Autonomic control of heart rate during
physical exercise and fractal dimension of
heart rate variability. J.Appl.Physiol 1993; 74
(2): 875-81
50. Tulppo, M. P., Hughson, R. L., Makikallio, T.
H., Airaksinen, K. E., Seppanen, T., and
Huikuri, H. V. Effects of exercise and passive
head-up tilt on fractal and complexity
properties of heart rate dynamics.
Am.J.Physiol Heart Circ.Physiol 2001; 280
(3): H1081-H1087
51. Bogaert, C., Beckers, F., Ramaekers, D., and
Aubert, A. E. Analysis of heart rate variability
with correlation dimension method in a normal
population and in heart transplant patients.
Auton.Neurosci. 2001; 90 (1-2): 142-7
52. Beckers, F., Ramaekers, D., and Aubert, A. E.
Approximate entropy of heart rate variability:
validation of methods and application in heart
failure. Cardiovasc.Eng. 2001; 1: 177-182:
53. Casaleggio, A., Cerutti, S., and Signorini, M.
G. Study of the Lyapunov exponents in heart
rate variability signals. Methods Inf.Med.1997;
36 (4-5): 274-7
54. Malliani, A. Principles of cardiovascular
neural regulation in health and disease. Kluwer
Academic Publ, Dordrecht, The Netherlands,
2000;
55. Eckberg DL and Sleight P. Human
baroreflexes in health and disease. Clarendon
Press, Oxford, UK1992;
56. Smith S.A., Stallard T.J., and Salih M.M. Can
sinoaortic baroreceptor heart rate reflex be
determined from phase 4 of the valslava
manoeuver. Cardiovasc Res 1987; 21: 422-7
57. Smyth H.S., Sleight P., and Pickering G.W.
Reflex regulation of arterial pressure during
sleep in man: a quantitative method of
assessing baroreflex sensitivity. Circ Res 1969;
24: 109-21
58. Bernardi L., Savioli R., and Spodick D.H. Do
hemodynamic responses to the valsalva
manoeuvre reflect myocardial dysfunction?
Chest 1989; 95: 986-91
59. Eckberg D.L., Cavanaugh M.S., and Mark
A.L. A simplified neck suction device for
activation of carotid baroreceptors. J Lab Clin
Med 1975; 85: 167-73
60. Smith, S. A., Querry, R. G., Fadel, P. J.,
Welch-O'Connor, R. M., Olivencia-Yurvati,
A., Shi, X., and Raven, P. B. Differential
baroreflex control of heart rate in sedentary
and aerobically fit individuals. Med.Sci.Sports
Exerc. 2000; 32 (8): 1419-30

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 27
61. Eckberg, D. L. Parasympathetic cardiovascular
control in human disease: a critical review of
methods and results. Am J Physiol 1980; 239:
H581-H593
62. Ludbrook, J. Concern about gain: is this the
best measure of performance of cardiovascular
reflexes? Clin.Exp.Pharmacol.Physiol 1984;
11: 385-90
63. Robbe, H. W., Mulder, L. J., Ruddel, H.,
Langewitz, W. A., Veldman, J. B., and
Mulder, G Assessment of baroreceptor reflex
sensitivity by means of spectral analysis.
Hypertension 1987; 10: 538-43
64. Bertinieri, G., Di Rienzo, M., Cavallazzi, A.,
Ferrari, A. U., Pedotti, A., and Mancia, G A
new approach to analysis of the arterial
baroreflex. J Hypert 1985; S79-S81
65. Ducher, M., Cerutti, C., Gustin, M. P., and
Paultre, C. Z. Statistical relationships between
systolic blood pressure and heart rate and their
functional significance in conscious rats.
Med.Biol.Eng Comput.1994; 32 (6): 649-55
66. Ramaekers, D., Beckers, F., Demeulemeester,
H., Bert, C., Denef, C., and Aubert, A. E.
Effects of melanocortins on cardiovascular
regulation in rats. Clin.Exp.Pharmacol.Physiol
2002; 29: 549-58
67. Ramaekers, D., Beckers, F., Demeulemeester,
H., and Aubert, A. E. Cardiovascular
autonomic function in conscious rats. A novel
approach to facilitate stationary conditions.
Annals of Nonivasive Electrocardiology 2002,
7: 307-318
68. Ramaekers, D., Ector, H., Aubert, A. E.,
Rubens, A., and Van de, Werf F. Heart rate
variability and heart rate in healthy volunteers.
Is the female autonomic nervous system
cardioprotective? Eur.Heart J.1998; 19 (9):
1334-41
69. Migliaro, E. R., Contreras, P., Bech, S.,
Etxagibel, A., Castro, M., Ricca, R., and
Vicente, K. Relative influence of age, resting
heart rate and sedentary life style in short-term
analysis of heart rate variability.
Braz.J.Med.Biol.Res.2001; 34 (4): 493-500
70. Dishman, R. K., Nakamura, Y., Garcia, M. E.,
Thompson, R. W., Dunn, A. L., and Blair, S.
N. Heart rate variability, trait anxiety, and
perceived stress among physically fit men and
women. Int.J.Psychophysiol.2000; 37 (2): 121-
33
71. Ramaekers, D., Ector, H., Demyttenaere, K.,
Rubens, A., and Van de, Werf F. Association
between cardiac autonomic function and
coping style in healthy subjects. Pacing
Clin.Electrophysiol.1998; 21 (8): 1546-52
72. Merz, C. N. and Pardo, Y. Mental versus
physical stress, QT prolongation, and the
autonomic nervous system. Circulation 2000;
101 (22): E213-E214
73. Hayano J., Yamada M., Sakakibara Y.,
Fujinami T., and Takata K. Short and long
term effects of cigarette smoking on HRV. Am
J Cardiol 1990; 65: 84-8
74. Niedermaier O., Smith M., Beigthol M.,
Zukowska-Grocec Z., and Goldstein G.
Influence of cigarette smoking on human
autonomic nervous system. Circulation 1993;
88: 562-71
75. Kolodiichik E. and Arushanian E.B. The effect
of caffeine on the cardiac intervalogram
indices depending on the ovarian cycle phase
in women. Farmal Toksil 1991; 54: 28-30
76. Nishijima Y., Ikeda T., Takamatsu M., Kiso
Y., Shibata H., Fushiki T., and Moritani T.
Influence of caffeine ingestion on autonomic
nervous activity during exercise in humans.
Eur.J.Appl.Physiol 2002; 87: 475-480:
77. Weise, F., Muller, D., Krell, D., Kielstein, V.,
and Koch, R. D. Heart rate variability in
chronic alcoholics: a follow-up study. Drug
Alcohol Depend.1986; 17 (4): 365-8
78. Hirsch J.A., Bishop B., and Yook J.L.
Recovery of respiratory sinus arrhythmia in
detoxified alcohol subjects. J Appl Physiol
1993; 74: 1816-23
79. La Rovere, M. T., Bigger, J. T., Jr., Marcus, F.
I., Mortara, A., and Schwartz, P. J. Baroreflex
sensitivity and heart-rate variability in
prediction of total cardiac mortality after
myocardial infarction. ATRAMI (Autonomic
Tone and Reflexes After Myocardial
Infarction) Investigators. Lancet 1998; 351
(9101): 478-84
80. Cole, C. R., Blackstone, E. H., Pashkow, F. J.,
Snader, C. E., and Lauer, M. S. Heart-rate
recovery immediately after exercise as a
predictor of mortality. N.Engl.J.Med.1999; 341
(18): 1351-7
81. Desai, M. Y., Pena-Almaguer, E., and
Mannting, F. Abnormal heart rate recovery
after exercise: a comparison with known
indicators of increased mortality. Cardiology
2001; 96 (1): 38-44
82. Deligiannis, A., Kouidi, E., and Tourkantonis,
A. Effects of physical training on heart rate
variability in patients on hemodialysis.
Am.J.Cardiol.1999; 84 (2): 197-202

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 28
83. Ponikowski, P., Chua, T. P., Piepoli, M.,
Banasiak, W., Anker, S. D., Szelemej, R.,
Molenda, W., Wrabec, K., Capucci, A., and
Coats, A. J. Ventilatory response to exercise
correlates with impaired heart rate variability
in patients with chronic congestive heart
failure. Am.J.Cardiol.1998; 82 (3): 338-44
84. Bernardi, L., Salvucci, F., Suardi, R., Solda, P.
L., Calciati, A., Perlini, S., Falcone, C., and
Ricciardi, L. Evidence for an intrinsic
mechanism regulating heart rate variability in
the transplanted and the intact heart during
submaximal dynamic exercise?
Cardiovasc.Res.1990; 24 (12): 969-81
85. Bernardi, L. Clinical evaluation of arterial
baroreflex activity in diabetes. Diabetes
Nutr.Metab 2000; 13 (6): 331-40
86. Narkiewicz, K. and Somers, V. K. Endurance
training in mild hypertension - effects on
ambulatory blood pressure and neural
circulatory control. Blood Press Monit.1997; 2
(5): 229-35
87. Kontopoulos, A. G., Athyros, V. G.,
Papageorgiou, A. A., Skeberis, V. M.,
Basayiannis, E. C., and Boudoulas, H. Effect
of angiotensin-converting enzyme inhibitors
on the power spectrum of heart rate variability
in post-myocardial infarction patients.
Coron.Artery Dis. 1997; 8 (8-9): 517-24
88. Kontopoulos, A. G., Athyros, V. G.,
Didangelos, T. P., Papageorgiou, A. A.,
Avramidis, M. J., Mayroudi, M. C., and
Karamitsos, D. T. Effect of chronic quinapril
administration on heart rate variability in
patients with diabetic autonomic neuropathy.
Diabetes Care 1997; 20 (3): 355-61
89. Perticone, F., Ceravolo, R., Maio, R., Cosco,
C., and Mattioli, P. L. Heart rate variability
and sudden infant death syndrome. Pacing
Clin.Electrophysiol. 1990; 13 (12 Pt 2): 2096-9
90. Harrison, M. H. Athletes, astronauts and
orthostatic tolerance. Sports Med. 1986; 3 (6):
428-35
91. Seps, B., Beckers, F., Ramaekers, D., and
Aubert, A. E. Acute heart rate response to
weightlessness during parabolic flight.2002,
pp. 151-158, Federal Office for Scientific,
Technical and Cultural Affairs, Brussels.
Belgium
92. Iellamo, F., Legramante, J. M., Massaro, M.,
Raimondi, G., and Galante, A. Effects of a
residential exercise training on baroreflex
sensitivity and heart rate variability in patients
with coronary artery disease: A randomized,
controlled study. Circulation 2000; 102 (21):
2588-92
93. Beckers, F., Ramaekers, D., Aubert, A. E., van
Cleemput, J., Droogne, W., Vanhaecke, J.,
Ector, H., and Van de Werf, F. Intracardiac
heart rate variability of the native sinus node in
heart transplant patients. Med.Biol.Eng
Comput.1999; 37 (sup 1): 232-3
94. Beckers, F., Ramaekers, D., van Cleemput, J.,
Droogne, W., Vanhaecke, J., Van de Werf, F.,
and Aubert, A. E. Association between
restoration of autonomic modulation in the
native sinus node and haemodynamic
improvement after cardiac transplantation.
Transplantation 2002; 73: 1614-20
95. Rost R. and Hollman W. Athlete's heart-A
review of its historical assessment and new
aspects. Int J Sports Med 1983; 4: 147
96. Roy A., Doyon M., and Dumesnil J.G.
Endurance vs strength training: comparison of
cardiac structures using normal predicted
values. J Appl Physiol 1988; 64: 2552
97. Pelliccia A., Maron B.J., Culasso F., Spataro
A., and Caselli G. Athlete's heart in women. J
Am Med Assoc 1996; 276: 1048-54
98. Zemva, A. and Rogel, P. Gender differences in
athlete's heart: association with 24-h blood
pressure. A study of pairs in sport dancing.
Int.J.Cardiol. 2001; 77 (1): 49-54
99. Janicki J.S., Sheriff D.D., Robotham J.L., and
Wise R.A. Cardiac output during exercise:
contributions of the cardiac, circulatory, and
respiratory systems.1996; (12): 649-704
100. Mazzeo R.S. and Marshall P. Influence of
plasma catecholamine on the lactate threshold
during graded exercise. J Appl Physiol 1989;
67: 1319-22
101. Roeske W.R., O'Rourke R.A., and Klein
A. Noninvasive determination of ventricular
hypertrophy in professional athletes.
Circulation 1976; 53: 286-91
102. Huston T.P., Puffer J.C., and Rodney
W.M. The athletic heart syndrome.
N.Engl.J.Med.1985; 313: 24-32
103. Orizio C., Perini R., Comande A.,
Castellano M., Beschi M., and Veicsteinas A.
Plasma cathecolamines and heart rate at the
beginning of muscular exercise inman.
Eur.J.Appl.Physiol 1988; (57): 644-51
104. O'Sullivan, S. E. and Bell, C. The effects
of exercise and training on human
cardiovascular reflex control.
J.Auton.Nerv.Syst. 2000; 81 (1-3): 16-24
105. Uusitalo, A. L., Tahvanainen, K. U.,
Uusitalo, A. J., and Rusko, H. K. Non-invasive

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 29
evaluation of sympathovagal balance in
athletes by time and frequency domain
analyses of heart rate and blood pressure
variability. Clin.Physiol 1996; 16 (6): 575-88
106. Seals D.R. and Chase P.B. Influence of
physical training on heart rate variability and
baroreflex circulatory control. J Appl Physiol
1989; 66: 1886-95
107. Stein, R., Moraes, R. S., Cavalcanti, A. V.,
Ferlin, E. L., Zimerman, L. I., and Ribeiro, J.
P. Atrial automaticity and atrioventricular
conduction in athletes: contribution of
autonomic regulation. Eur.J.Appl.Physiol
2000; 82 (1-2): 155-7
108. Stein, R., Medeiros, C. M., Rosito, G. A.,
Zimerman, L. I., and Ribeiro, J. P. Intrinsic
sinus and atrioventricular node
electrophysiologic adaptations in endurance
athletes. J.Am.Coll.Cardiol. 2002; 39 (6):
1033-8
109. Mangin, L., Kobeissi, A., Lelouche, D.,
Dherouville, T. Y., Mansier, P.,
Swynghedauw, B., and Macquin-Mavier, I.
Simultaneous analysis of heart rate variability
and myocardial contractility during head-up tilt
in patients with vasovagal syncope.
J.Cardiovasc.Electrophysiol. 2001; 12 (6):
639-44
110. Matsumoto, T., Miyawaki, T., Ue, H.,
Kanda, T., Zenji, C., and Moritani, T.
Autonomic responsiveness to acute cold
exposure in obese and non-obese young
women. Int.J.Obes.Relat Metab Disord. 1999;
23 (8): 793-800
111. Bernardi, L., Passino, C., Spadacini, G.,
Calciati, A., Robergs, R., Greene, R.,
Martignoni, E., Anand, I., and Appenzeller, O.
Cardiovascular autonomic modulation and
activity of carotid baroreceptors at altitude.
Clin.Sci.(Lond) 1998; 95 (5): 565-73
112. McCraty, R., Atkinson, M., Tiller, W. A.,
Rein, G., and Watkins, A. D. The effects of
emotions on short-term power spectrum
analysis of heart rate variability. Am.J.Cardiol.
1995; 76 (14): 1089-93
113. McCole SD., Brown MD., Moore GE.,
Zmuda JM., Cwynar JD., and Hagberg J.M.
Enhanced cardiovascular hemodynamics in
endurance-trained postmenopausal women
athletes. Med Sci Sports Exerc 2000; 32: 1073-
9
114. DiCarlo, S. E. and Bishop, V. S. Exercise
training enhances cardiac afferent inhibition of
baroreflex function. Am J Physiol 1990; 258:
H212-H220
115. Iellamo, F. Neural control of the
cardiovascular system during exercise.
Ital.Heart J. 2001; 2 (3): 200-12
116. Bernardi, L., Passino, C., Robergs, R., and
Appenzeller, O. Acute and persistent effects of
a 46-kilometer wilderness trail run at altitude:
cardiovascular autonomic modulation and
baroreflexes. Cardiovasc.Res. 1997; 34 (2):
273-80
117. Halliwill, J. R., Taylor, J. A., Hartwig, T.
D., and Eckberg, D. L. Augmented baroreflex
heart rate gain after moderate-intensity,
dynamic exercise. Am.J.Physiol 1996; 270 (2
Pt 2): R420-R426
118. Grassi G., Seravalle G., Calhoun D.A., and
Mancia G. Physical training and baroreceptor
control of sympathetic nerve activity in
humans. Hypertension 1994; 23: 294-301
119. Negrao C.E., Irigoyen M.C., Moreira E.D.,
Brum P.C., Freire C.E., and Krieger E.M.
Effect of exercise training on RSNA,
baroreflex control and blood pressure
responsiveness. Am J Physiol 1993; 265:
R365-R370
120. Robinson, B. F., Epstein, S. E., Beiser, G.
D., and Braunwald, E. Control of heart rate by
the autonomic nervous system. Studies in man
on the interrelation between baroreceptor
mechanisms and exercise. Circ.Res.1966; 19
(2): 400-11
121. Schwartz, P. J., Verrier, R. L., and Lown B
Effect of autonomic blockade and vagotomy
on ventricular refractoriness in dogs. Circ Res
1977; 44 (637): 645
122. Malpas, S. C. Neural influences on
cardiovascular variability: possibilities and
pitfalls. Am J Physiol.Heart Circ Phys 2002;
282: H6-H20
123. Roure, R., Collet, C., Deschaumes-
Molinaro, C., Dittmar, A., Rada, H.,
Delhomme, G., and Vernet-Maury, E.
Autonomic nervous system responses correlate
with mental rehearsal in volleyball training.
Eur.J.Appl.Physiol Occup.Physiol 1998; 78
(2): 99-108
124. Collet, C., Roure, R., Delhomme, G.,
Dittmar, A., Rada, H., and Vernet-Maury, E.
Autonomic nervous system responses as
performance indicators among volleyball
players. Eur J Appl Physiol Occup Physiol
1999; 80: 41-51
125. Perini, R., Orizio, C., Baselli, G., Cerutti,
S., and Veicsteinas, A. The influence of
exercise intensity on the power spectrum of

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 30
heart rate variability. Eur.J.Appl.Physiol
Occup.Physiol 1990; 61 (1-2): 143-8
126. Arai, Y., Saul, J. P., Albrecht, P., Hartley,
L. H., Lilly, L. S., Cohen, R. J., and Colucci,
W. S. Modulation of cardiac autonomic
activity during and immediately after exercise.
Am.J.Physiol 1989; 256(1Pt 2): H132-H141
127. Iellamo, F. Neural mechanisms of
cardiovascular regulation during exercise.
Auton.Neurosci. 2001; (20): 66-75
128. Bernardi, L. and Piepoli, M. Autonomic
nervous system adaptatiuon during physical
exercise. Ital Heart J 2001; (2): 831-9
129. Maciel, B. C., Gallo, L., Jr., Marin Neto, J.
A., Lima Filho, E. C., and Martins, L. E.
Autonomic nervous control of the heart rate
during dynamic exercise in normal man.
Clin.Sci.(Colch.) 1986; 71 (4): 457-60
130. Brenner, I. K., Thomas, S., and Shephard,
R. J. Autonomic regulation of the circulation
during exercise and heat exposure. Inferences
from heart rate variability. Sports Med. 1998;
26 (2): 85-99
131. Warren, J. H., Jaffe, R. S., Wraa, C. E.,
and Stebbins, C. L. Effect of autonomic
blockade on power spectrum of heart rate
variability during exercise. Am.J.Physiol 1997;
273 (2 Pt 2): R495-R502
132. Cottin F, Papelier Y, and Escourrou P
Effects of exercise load and breathing
frequency on heart rate and blood pressure
variability during dynamic exercise. Int J
Sports Med 1999; 20 (232): 238
133. Kamath, M. V., Fallen, E. L., and
McKelvie, R. Effects of steady state exercise
on the power spectrum of heart rate variability.
Med.Sci.Sports Exerc. 1991; 23 (4): 428-34
134. Casadei B., Moon J., Johnston J, Caiazza
A, and Sleigh J.W. Is respiratory sinus
arrhthymia a good index of cardiac vagal tone
in exercise? J Appl Physiol 1996; 81: 556-64
135. Shin, K., Minamitani, H., Onishi, S.,
Yamazaki, H., and Lee, M The power spectral
analysis of heart rate variability in athletes
during dynamic exercise-Part II. Clin Cardiol
1995; 18: 664-8
136. Shin, K., Minamitani, H., Onishi, S.,
Yamazaki, H., and Lee, M. The power spectral
analysis of heart rate variability in athletes
during dynamic exercise--Part I. Clin.Cardiol.
1995; 18 (10): 583-6
137. Casadei, B., Cochrane, J. E., Johnston J,
Conway, J., and Sleigh, J. W. Pitfalls in the
interpretation of spectral analysis of the heart
rate variability during exercise in humans.
Acta Physiol Scand 1995; 153: 125-31
138. Yamamoto, Y., Hughson, R. L., and
Nakamura, Y. Autonomic nervous system
responses to exercise in relation to ventilatory
threshold. Chest 1992; 101 (5 Suppl): 206S-
10S
139. Tulppo, M. P., Makikallio, T. H., Takala,
T. E., Seppanen, T., and Huikuri, H. V.
Quantitative beat-to-beat analysis of heart rate
dynamics during exercise. Am.J.Physiol 1996;
271 (1 Pt 2): H244-H252
140. Anosov, O., Patzak, A., Kononovich, Y.,
and Persson, P. B. High-frequency oscillations
of the heart rate during ramp load reflect the
human anaerobic threshold. Eur.J.Appl.Physiol
2000; 83 (4 -5): 388-94
141. Gonzalez-Camarena, R., Carrasco-Sosa,
S., Roman-Ramos, R., Gaitan-Gonzalez, M. J.,
Medina-Banuelos, V., and Azpiroz-Leehan, J.
Effect of static and dynamic exercise on heart
rate and blood pressure variabilities.
Med.Sci.Sports Exerc. 2000; 32 (10): 1719-28
142. Tonkins W.P. Analysis of the relationship
between exercise capacity and heart rate
variability in trained and untrained individuals.
University of Oregon, microform publications
1999;
143. Aubert, A. E., Beckers, F., and Ramaekers,
D. Short-term heart rate variability in young
athletes. J.Cardiol. 2001; 37: 85-8
144. Aubert, A. E., Ramaekers, D., Cuche, Y.,
Leysens, R., Ector, H., and Van de Werf, F.
Effect of long-term physical training on heart
rate variability. IEEE Comp.Cardiol. 1996; 16:
17-20
145. Puig, J., Freitas, J., Carvalho, M. J., Puga,
N., Ramos, J., Fernandes, P., Costa, O., and de
Freitas, A. F. Spectral analysis of heart rate
variability in athletes. J.Sports
Med.Phys.Fitness 1993; 33 (1): 44-8
146. Dixon, E. M., Kamath, M. V., McCartney,
N., and Fallen, E. L. Neural regulation of heart
rate variability in endurance athletes and
sedentary controls. Cardiovasc.Res. 1992; 26
(7): 713-9
147. Goldsmith, R. L., Bigger, J. T., Jr.,
Steinman, R. C., and Fleiss, J. L. Comparison
of 24-hour parasympathetic activity in
endurance-trained and untrained young men.
J.Am.Coll.Cardiol. 1992; 20 (3): 552-8
148. Jensen-Urstad, K., Saltin, B., Ericson, M.,
Storck, N., and Jensen-Urstad, M. Pronounced
resting bradycardia in male elite runners is

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 31
associated with high heart rate variability.
Scand.J.Med.Sci.Sports 1997; 7 (5): 274-8
149. Macor, F., Fagard, R., and Amery, A.
Power spectral analysis of RR interval and
blood pressure short-term variability at rest
and during dynamic exercise: comparison
between cyclists and controls. Int.J.Sports
Med. 1996; 17 (3): 175-81
150. Shin, K., Minamitani, H., Onishi, S.,
Yamazaki, H., and Lee, M. Autonomic
differences between athletes and nonathletes:
spectral analysis approach. Med.Sci.Sports
Exerc. 1997; 29 (11): 1482-90
151. Furlan, R., Piazza, S., Dell'Orto, S.,
Gentile, E., Cerutti, S., Pagani, M., and
Malliani, A. Early and late effects of exercise
and athletic training on neural mechanisms
controlling heart rate. Cardiovasc.Res. 1993;
27 (3): 482-8
152. Janssen, M. J., de Bie, J., Swenne, C. A.,
and Oudhof, J. Supine and standing
sympathovagal balance in athletes and
controls. Eur.J.Appl.Physiol Occup.Physiol
1993; 67 (2): 164-7
153. De Meersman, R. E. Heart rate variability
and aerobic fitness. Am.Heart J. 1993; 125 (3):
726-31
154. Catai, A. M., Chacon-Mikahil, M. P.,
Martinelli, F. S., Forti, V. A., Silva, E.,
Golfetti, R., Martins, L. E., Szrajer, J. S.,
Wanderley, J. S., Lima-Filho, E. C., Milan, L.
A., Marin-Neto, J. A., Maciel, B. C., and
Gallo, L., Jr. Effects of aerobic exercise
training on heart rate variability during
wakefulness and sleep and cardiorespiratory
responses of young and middle-aged healthy
men. Braz.J.Med.Biol.Res. 2002; 35 (6): 741-
52
155. Katona, P. G., McLean, M., Dighton, D.
H., and Guz, A. Sympathetic and
parasympathetic cardiac control in athletes and
nonathletes at rest. J.Appl.Physiol 1982; 52
(6): 1652-7
156. Lazoglu, A. H., Glace, B., Gleim, G. W.,
and Coplan, N. L. Exercise and heart rate
variability. Am.Heart J. 1996; 131 (4): 825-6
157. Melanson, E. L. Resting heart rate
variability in men varying in habitual physical
activity. Med.Sci.Sports Exerc. 2000; 32 (11):
1894-901
158. Reiling, M. J. and Seals, D. R. Respiratory
sinus arrhythmia and carotid baroreflex control
of heart rate in endurance athletes and
untrained controls. Clin.Physiol 1988; 8 (5):
511-9
159. Smith, M. L., Hudson, D. L., Graitzer, H.
M., and Raven, P. B. Exercise training
bradycardia: the role of autonomic balance.
Med.Sci.Sports Exerc. 1989; 21 (1): 40-4
160. Goldsmith, R. L., Bigger, J. T., Jr.,
Bloomfield, D. M., and Steinman, R. C.
Physical fitness as a determinant of vagal
modulation. Med.Sci.Sports Exerc. 1997; 29
(6): 812-7
161. Bonaduce, D., Petretta, M., Cavallaro, V.,
Apicella, C., Ianniciello, A., Romano, M.,
Breglio, R., and Marciano, F. Intensive
training and cardiac autonomic control in high
level athletes. Med.Sci.Sports Exerc. 1998; 30
(5): 691-6
162. Strano, S., Lino, S., Calcagnini, G., di,
Virgilio, V, Ciardo, R., Cerutti, S., Calcagnini,
G., and Caselli, G. Respiratory sinus
arrhythmia and cardiovascular neural
regulation in athletes. Med.Sci.Sports Exerc.
1998; 30 (2): 215-9
163. Malfatto, G., Facchini, M., Bragato, R.,
Branzi, G., Sala, L., and Leonetti, G. Short and
long term effects of exercise training on the
tonic autonomic modulation of heart rate
variability after myocardial infarction.
Eur.Heart J. 1996; 17 (4): 532-8
164. Malfatto, G., Facchini, M., Sala, L.,
Branzi, G., Bragato, R., and Leonetti, G.
Effects of cardiac rehabilitation and beta-
blocker therapy on heart rate variability after
first acute myocardial infarction.
Am.J.Cardiol. 1998; 81 (7): 834-40
165. Meyer, M., Marconi, C., Ferretti, G.,
Fiocchi, R., Cerretelli, P., and Skinner, J. E.
Heart rate variability in the human transplanted
heart: nonlinear dynamics and QT vs RR-QT
alterations during exercise suggest a return of
neurocardiac regulation in long-term recovery.
Integr.Physiol Behav.Sci. 1996; 31 (4): 289-
305
166. Melanson, E. L. and Freedson, P. S. The
effect of endurance training on resting heart
rate variability in sedentary adult males.
Eur.J.Appl.Physiol 2001; 85 (5): 442-9
167. Boutcher, S. H. and Stein, P. Association
between heart rate variability and training
response in sedentary middle-aged men.
Eur.J.Appl.Physiol Occup.Physiol 1995; 70
(1): 75-80
168. Perini R., Fisher N., Veicsteinas A., and
Pendergast D.R. Aerobic training and
cardiovascular responses at rest and during
exercise in older men and women. Medicine

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 32
and Science in Sports and Exercise 2002; 34:
700-8
169. Schuit, A. J., van Amelsvoort, L. G.,
Verheij, T. C., Rijneke, R. D., Maan, A. C.,
Swenne, C. A., and Schouten, E. G. Exercise
training and heart rate variability in older
people. Med.Sci.Sports Exerc. 1999; 31 (6):
816-21
170. Ramaekers, D., Ector, H., Aubert, A. E.,
Rubens, A., and VandeWerf, F. Heart rate
variability and heart rate in healthy volunteers
- Is the female autonomic nervous system
cardioprotective? European Heart Journal
1998; 19 (9): 1334-41
171. Loimaala, A., Huikuri, H., Oja, P.,
Pasanen, M., and Vuori, I. Controlled 5-mo
aerobic training improves heart rate but not
heart rate variability or baroreflex sensitivity.
J.Appl.Physiol 2000; 89 (5): 1825-9
172. Al Ani, M., Munir, S. M., White, M.,
Townend, J., and Coote, J. H. Changes in R-R
variability before and after endurance training
measured by power spectral analysis and by
the effect of isometric muscle contraction.
Eur.J.Appl.Physiol Occup.Physiol 1996; 74
(5): 397-403
173. Davy, K. P., Willis, W. L., and Seals, D. R.
Influence of exercise training on heart rate
variability in post- menopausal women with
elevated arterial blood pressure. Clin.Physiol
1997; 17 (1): 31-40
174. Levy, W. C., Cerqueira, M. D., Harp, G.
D., Johannessen, K. A., Abrass, I. B.,
Schwartz, R. S., and Stratton, J. R. Effect of
endurance exercise training on heart rate
variability at rest in healthy young and older
men. Am.J.Cardiol. 1998; 82 (10): 1236-41
175. Sacknoff, D. M., Gleim, G. W.,
Stachenfeld, N., and Coplan, N. L. Effect of
athletic training on heart rate variability.
Am.Heart J. 1994; 127 (5): 1275-8
176. Ryan, S. M., Goldberger, A. L., Pincus, S.
M., Mietus, J., and Lipsitz, L. A. Gender- and
age-related differences in heart rate dynamics:
are women more complex than men?
J.Am.Coll.Cardiol. 1994; 24 (7): 1700-7
177. Reardon, M. and Malik, M. Changes in
heart rate variability with age. Pacing
Clin.Electrophysiol. 1996; 19 (11 Pt 2): 1863-6
178. Ramaekers, D., Ector, H., and Aubert, A.
E. The influence of age and gender on heart
rate variability (HRV). J.Am.Coll.Cardiol.
1999; 33 (3): 900-2
179. Kuo, T. B., Lin, T., Yang, C. C., Li, C. L.,
Chen, C. F., and Chou, P. Effect of aging on
gender differences in neural control of heart
rate. Am.J.Physiol 1999; 277 (6 Pt 2): H2233-
H2239
180. Tasaki, H., Serita, T., Irita, A., Hano, O.,
Iliev, I., Ueyama, C., Kitano, K., Seto, S.,
Hayano, M., and Yano, K. A 15-year
longitudinal follow-up study of heart rate and
heart rate variability in healthy elderly persons.
J.Gerontol.A Biol.Sci.Med.Sci. 2000; 55 (12):
M744-M749
181. Sato, N., Miyake, S., Akatsu, J., and
Kumashiro, M. Power spectral analysis of
heart rate variability in healthy young women
during the normal menstrual cycle.
Psychosom.Med. 1995; 57 (4): 331-5
182. Saeki Y, Atogami F Takahashi K
Yoshikawa T Reflex control of autonomic
function induced by posture change during the
menstrual cycle. J Auton Nerv Syst1 1997; 10:
69-74
183. Yildirir, A., Kabakci, G., Yarali, H.,
Aybar, F., Akgul, E., Bukulmez, O.,
Tokgozoglu, L., Gurgan, T., and Oto, A.
Effects of hormone replacement therapy on
heart rate variability in postmenopausal
women. Ann.Noninvasive.Electrocardiol.
2001; 6 (4): 280-4
184. Boutcher, S. H. and Meyer, B. J. Craig G.
A. Astheimer L. Resting autonomic function in
aerobically trained and untrained post-
menopausal women. J Aging Phys Activity
1998; 6: 310-6
185. Davy, K. P., Miniclier, N. L., Taylor, J. A.,
Stevenson, E. T., and Seals, D. R. Elevated
heart rate variability in physically active
postmenopausal women: a cardioprotective
effect? Am.J.Physiol 1996; 271 (2 Pt 2): H455-
H460
186. Pigozzi, F., Alabiso, A., Parisi, A., Di,
Salvo, V, Di Luigi, L., Spataro, A., and
Iellamo, F. Effects of aerobic exercise training
on 24 hr profile of heart rate variability in
female athletes. J.Sports Med.Phys.Fitness
2001; 41 (1): 101-7
187. Hedelin, R., Wiklund, U., Bjerle, P., and
Henriksson-Larsen, K. Pre- and post-season
heart rate variability in adolescent cross-
country skiers. Scand.J.Med.Sci.Sports 2000;
10 (5): 298-303
188. Muster A.J., Jungblutt P.R., Quigg R.J.,
McNeal M.A., Clauser J., and McPherson D.D.
Cardiopulmonary and cardiovascular
physiology in elderly competitive endurance
athletes. Am J Geriatr Cardiol 1999; 8: 162-8

Heart rate variability in athletes. AE Aubert, B Seps and F Beckers. Sports Medicine 33(12):889-919, 2003 33
189. Galloway M.T., Kadoko R., and Joki P.
Effect of aging on male and female master
athlete's performance in strength versus
endurance activities. Am J Orthop 2002; 31:
93-8
190. Menard D. and Stanish W.D. The aging
athlete. Am J Sports Med 1989; 17: 187-96
191. Yataco, A. R., Fleisher, L. A., and Katzel,
L. I. Heart rate variability and cardiovascular
fitness in senior athletes. Am.J.Cardiol. 1997;
80 (10): 1389-91
192. Banach, T., Zoladz, J. A., Kolasinska-
Kloch, W., Szygula, Z., and Thor, P. J. [The
effect of aging on the activity of the autonomic
nervous system in long distance runners]. Folia
Med.Cracov. 2000; 41 (3-4): 113-20
193. Jensen-Urstad, K., Bouvier, F., Saltin, B.,
and Jensen-Urstad, M. High prevalence of
arrhythmias in elderly male athletes with a
lifelong history of regular strenuous exercise.
Heart 1998; 79 (2): 161-4
194. Fry R.W. and Kraemer W.J. Resistance
exercise overtraining and overreaching. Sports
Med. 1997; 23: 106-29
195. Kuipers, H. and Keizer, H. A. Overtraining
in elite athletes. Review and directions for the
future. Sports Med. 1988; 6 (2): 79-92
196. Fry R.W., Morton A.R., and Keast D.
Overtraining in athletes. Sports Med 1991; 12:
32-65
197. Lehmann M., Foster C., and Keul J.
Overtraining in endurance athletes: a brief
overview. Med Sc Sports Exerc 1993; 25: 854-
62
198. Kuipers, H. Training and overtraining: an
introduction. Med.Sci.Sports Exerc. 1998; 30
(7): 1137-9
199. Gastmann, U. A. and Lehmann, M. J.
Overtraining and the BCAA hypothesis.
Med.Sci.Sports Exerc. 1998; 30 (7): 1173-8
200. Lehmann, M., Foster, C., Dickhuth, H. H.,
and Gastmann, U. Autonomic imbalance
hypothesis and overtraining syndrome.
Med.Sci.Sports Exerc. 1998; 30 (7): 1140-5
201. Israel S. Problems of overtraining from an
internal medical and performance
physiological standpoint. Medizin Und Sport
1976; 16: 1-12
202. Kindermann W. Overtraining: expression
of a disturbed autonomic regulation (Das
Übertraining: Ausdruck ainer vegetativen
Fehlsteuerung). Dtsch.Z.Sportmed. 1986; 8:
238-45
203. Hedelin, R., Kentta, G., Wiklund, U.,
Bjerle, P., and Henriksson-Larsen, K. Short-
term overtraining: effects on performance,
circulatory responses, and heart rate
variability. Med.Sci.Sports Exerc. 2000; 32
(8): 1480-4
204. Hedelin, R., Wiklund, U., Bjerle, P., and
Henriksson-Larsen, K. Cardiac autonomic
imbalance in an overtrained athlete.
Med.Sci.Sports Exerc. 2000; 32 (9): 1531-3
205. Uusitalo, A. L., Uusitalo, A. J., and Rusko,
H. K. Heart rate and blood pressure variability
during heavy training and overtraining in the
female athlete. Int.J.Sports Med. 2000; 21 (1):
45-53
206. Portier, H., Louisy, F., Laude, D.,
Berthelot, M., and Guezennec, C. Y. Intense
endurance training on heart rate and blood
pressure variability in runners. Med.Sci.Sports
Exerc. 2001; 33 (7): 1120-5
207. Pichot, V., Roche, F., Gaspoz, J. M.,
Enjolras, F., Antoniadis, A., Minini, P., Costes,
F., Busso, T., Lacour, J. R., and Barthelemy, J.
C. Relation between heart rate variability and
training load in middle- distance runners.
Med.Sci.Sports Exerc. 2000; 32 (10): 1729-36
208. Monahan K.D., Dinenno F.A., Tanaka H.,
Clevenger C.M., DeSouza C.A., and Seals
D.R. Regular aerobic exercise modulates age-
associated declines in cardiovagal baroreflex
sensistivity in healthy men. J.Physiol 2000;
(529): 263-71
209. Goldsmith, R. L., Bloomfield, D. M., and
Rosenwinkel, E. T. Exercise and autonomic
function. Coron.Artery Dis. 2000; 11 (2): 129-
35
210. Hedelin, R., Bjerle, P., and Henriksson-
Larsen, K. Heart rate variability in athletes:
relationship with central and peripheral
performance. Med.Sci.Sports Exerc. 2001; 33
(8): 1394-8
211. Aubert, A. E., Ramaekers, D., Collier, B.,
Beckers, F., Van de Werf, F., and Ector, H.
Comparison of the effect of different types of
exercise on short-term heart rate variability.
Med.Biol.Eng Comput. 1999; 37: 568-569
212. Uusitalo, A. L., Uusitalo, A. J., and Rusko,
H. K. Endurance training, overtraining and
baroreflex sensitivity in female athletes.
Clin.Physiol 1998; 18 (6): 510-20