Can HRV be used to evaluate training load in constant ... - Springer Link

Eur J Appl Physiol (2010) 108:435–442 DOI 10.1007/s00421-009-1240-1

O R I G I N A L A R T I CL E

Can HRV be used to evaluate training load in constant load exercises? Piia Kaikkonen · Esa Hynynen · Theresa Mann · Heikki Rusko · Ari Nummela

Accepted: 29 September 2009 / Published online: 14 October 2009 © Springer-Verlag 2009

Abstract The overload principle of training states that training load (TL) must be suYcient to threaten the homeostasis of cells, tissues, organs, and/or body. However, there is no “golden standard” for TL measurement. The aim of this study was to examine if any post-exercise heart rate variability (HRV) indices could be used to evaluate TL in exercises with diVerent intensities and durations. Thirteen endurance-trained males (35 § 5 year) performed MODE (moderate intensity, 3 km at 60% of the maximal velocity of the graded maximal test (vVO2max)), HI (high intensity, 3 km at 85% vVO2max), and PRO (prolonged, 14 km at 60% vVO2max) exercises on a treadmill. HRV was analyzed with short-time Fourier-transform method during rest, exercise, and 15-min recovery. Rating of perceived exertion (RPE), blood lactate (BLa), and HFP120 (mean of 0–120 s postexercise) described TL of these exercises similarly, being

diVerent for HI (P < 0.05) and PRO (P < 0.05) when compared with MODE. RPE and BLa also correlated negatively with HFP120 (r = ¡0.604, ¡0.401), LFP120 (¡0.634, ¡0.601), and TP120 (¡0.691, ¡0.569). HRV recovery dynamics were similar after each exercise, but the level of HRV was lower after HI than MODE. Increased intensity or duration of exercise decreased immediate HRV recovery, suggesting that post-exercise HRV may enable an objective evaluation of TL in Weld conditions. The Wrst 2-min recovery seems to give enough information on HRV recovery for evaluating TL. Keywords Autonomic nervous system · Running exercise · Recovery · Short-time Fourier transform

Introduction Communicated by Susan Ward. P. Kaikkonen Tampere Research Center of Sports Medicine, Tampere, Finland E. Hynynen · H. Rusko · A. Nummela KIHU, Research Institute for Olympic Sports, Jyväskylä, Finland T. Mann MRC/UCT Research Unit for Exercise Science and Sports Medicine, University of Cape Town, Cape Town, South Africa H. Rusko Department of Biology of Physical Activity, University of Jyväskylä, Jyväskylä, Finland P. Kaikkonen (&) Kaupinpuistonkatu 1, PO Box 30, 33501 Tampere, Finland e-mail: [email protected]

The overload principle of training states that training load (TL) must be suYcient to threaten the homeostasis of cells, tissues, organs and/or body. TL of aerobic exercise depends on both exercise intensity and duration. There is no “golden standard” for TL but it has generally been measured with heart rate (HR), oxygen uptake (VO2), blood lactate concentration (BLa), and rating of perceived exertion (RPE) (Borg 1982). These parameters reXect mainly the intensity of exercise. Because the duration of exercise also aVects TL, indices like session RPE (RPEs) (Foster 1998) and Training Impulse (TRIMP) (Banister 1991) have been introduced. These measures aim to account for exercise duration as well as intensity, nevertheless their validity as a comprehensive measure of TL has not been proven. The autonomic nervous system (ANS) regulates homeostatic function of the body (Porges 1992), including cardiovascular function during exercise and recovery, executing a

123

436

rapid shift in autonomic output during the transition from one state to the other. SpeciWcally, the termination of exercise is known to trigger an increase in vagal activity with a simultaneous reduction in sympathetic drive (Savin et al. 1982). Possible mechanisms inducing the immediate changes in cardiac function include also, e.g., fast changes in cardiac pre-load, after-load and contractility of the heart (Miles et al. 1984; Plotnick et al. 1986). Simultaneously, there is also a loss of central command and greater activation level of the arterial baroreXex, resulting a decrease in HR (Arai et al. 1989; Oida et al. 1997; O’Leary 1993). An increase in vagal activity is generally agreed to play a major role in decreasing HR during the Wrst minute of recovery and further decrease in HR is mediated by both sympathetic and vagal systems (Perini et al. 1989). To quantify vagal reactivation after exercise, the time course of heart rate recovery and several heart rate variability (HRV) indices have been studied (Cole et al. 1999; Goldberger et al. 2006; Pierpont et al. 2000; Savin et al. 1982). HRV has been widely used as a noninvasive method to estimate function of ANS during rest, exercise, and recovery, and it is well established that high-frequency power (HFP) is modulated essentially by Xuctuations in the vagal branch of the ANS (Berntson et al. 1993). A consensus has emerged regarding a decrease in HRV, especially the vagal-related HFP, during exercise followed by a gradual increase during the Wrst minutes of recovery (Casties et al. 2006; Goldberger et al. 2006; Kaikkonen et al. 2008; Kaikkonen et al. 2007). Moreover, a signiWcant delay in post-exercise HRV recovery has been observed following moderate- to high-intensity exercise when compared with exercise at a low to moderate intensity (Buchheit et al. 2007; Kaikkonen et al. 2008; Kaikkonen et al. 2007; Martinmäki and Rusko 2008; Seiler et al. 2007). The inXuence of exercise duration on the HRV recovery dynamics has been examined in only a few studies. Doubled exercise duration when compared with a “default” exercise of 60 min (Seiler et al. 2007) or 3,500 m (Kaikkonen et al. 2007) was not found to aVect HRV recovery in highly trained athletes (Seiler et al. 2007) or in sedentary women (Kaikkonen et al. 2007). However, the exercise intensity in these studies was less than 65% of VO2max and it may be proposed that the increased disturbance of homeostasis at higher exercise intensities would produce diVerent Wndings. The aim of this study was to examine if any post-exercise HRV indices could be used to evaluate TL in exercises with diVerent intensities and durations. More speciWcally we wanted to Wnd out if HRV was able to diVerentiate exercises of diVerent intensity and duration, and how HRV was related to the present TL parameters. We hypothesized that HRV may reXect TL-induced changes in body homeostasis.

123

Eur J Appl Physiol (2010) 108:435–442

Methods Thirteen recreational level distance runners volunteered to participate in this study from a larger group of subjects recruited e.g., from marathon clubs. They all had completed 4 (§1) training sessions per week during the last 2 months, and all except one had run at least one marathon during the last few years. The subjects were selected for the study if they were low-risk participants according to the ACSM (American college of sports medicine) guidelines. To optimize safety during exercise testing, it is important to screen potential participants for risk factors and symptoms of various cardiovascular, pulmonary or metabolic diseases that may be aggravated by exercise. In general, there is no risk of cardiovascular events provoked by low to moderate intensity exercise, but by selecting the subjects classiWed as low risk, also higher than moderate intensities may be pursued safely. The descriptive and performance characteristics of the subjects are presented in the Table 1. Subjects gave a written informed consent to participate and had the right to withdraw from the study at any time. The study was approved by the ethics committee of the University of Jyväskylä. Graded maximal treadmill test Maximal aerobic power of each subject was assessed by a graded maximal treadmill test. This test consisted of an initial speed of 8 km h¡1 (gradient 0.5°) followed by increments of 1 km h¡1 every 3 min until exhaustion. Breathby-breath respiratory data (Oxycon Mobile, Viasys Healthcare GmbH, Hoechberg, Germany) and R-R intervals (RRI) (Suunto t6 wristop computer, Suunto Oy, Vantaa, Finland) were collected continuously during the test. Fingertip blood Table 1 Mean (§SD) descriptive data and graded maximal test results of the subjects (N = 13) Age (years)

35 § 5

Height (cm)

179 § 6

Body mass (kg)

76.6 § 5.6

BMI (kg m¡2)

23.9 § 1.6

VO2max (l min¡1) VO2max (ml kg¡1 min¡1)

4.1 § 0.4 54.1 § 3.6

AerT (% of VO2max)

57.9 § 6.2

AnT (% of VO2max)

80.6 § 4.7

HRmax (bpm)

183 § 9

vVO2max (km h¡1)

15.9 § 0.8

BLamax (mmol l¡1)

11.8 § 2.5

BMI body mass index, VO2max maximal oxygen uptake, AerT aerobic threshold according to Aunola and Rusko (1984), AnT anaerobic threshold according to Aunola and Rusko (1984), vVO2max maximal speed at the graded maximal treadmill test, HRmax maximal heart rate, BLamax maximal blood lactate level at the graded maximal treadmill test


437

This study consisted of three constant-load endurance exercises: a short moderate-intensity exercise (MODE, 3 km at 60% vVO2max), a short high-intensity exercise (HI, 3 km at 85% vVO2max), and a prolonged moderate-intensity exercise (PRO, 14 km at 60% vVO2max). Exercises were performed in random order on three diVerent days, separated by at least 2 days. The subjects were asked to refrain from intense physical exertion for 2 days, and from alcohol consumption for 1 day prior to controlled exercises, and to avoid caVeine during the test day. The controlled exercises were carried out on a treadmill at the speed individually calculated for each subject (gradient 0.5°). Each session consisted of a 5-min baseline measurement with the subject seated, 1-km controlled warm up at 60% vVO2max, the exercise session and a 15-min controlled recovery during which the subject was once again seated. Moving and talking was prohibited during baseline and recovery phases. MODE was considered as a “basic” exercise, to which the eVects of increased intensity or prolonged duration were compared.

Since the RRI time series were not stationary during the recovery, conventional spectral analysis could not be used. Short-time Fourier-transform (STFT) method was used in HRV analysis, since it provides instantaneous time-frequency information of RRI and can be used during stationary as well as transient phases of signal. STFT has been previously used to measure HRV during exercise (Pichon et al. 2004) and recovery (Kaikkonen et al. 2008; Kaikkonen et al. 2007; Martinmäki and Rusko 2008), and has been proven to detect fast changes in vagal responses (Keselbrener and Akselrod 1996; Martinmäki et al. 2006). It provides time-frequency decompositions of the consecutive RRI time series by calculating consecutive power spectra of short sections of the signal. A section of 512 samples was multiplied by the Hanning window function and the fast Fourier transform of their product was taken. The window was then shifted from one sample to another and the same calculations were performed until the whole RRI time series, baseline, exercise, and recovery, was covered. Lowfrequency power (LFP, 0.04–0.15 Hz), high-frequency power (HFP, 0.15–1.0 Hz), and total power (TP, 0.04–1.0 Hz) were calculated as integrals of the respective power spectral density curve. The frequency limit of 1.0 Hz was used to include the respiratory frequency during exercise and recovery to the analysis. In order to meet the assumptions of parametric statistical analysis, spectral powers were expressed in natural log-transformed values (LFPln, HFPln, TPln). A 4-min average was calculated during the pre-exercise sitting. Averages of 60 s were used in statistical analysis during the recovery minutes 1–5 and 15. HFPdiV was calculated as diVerence between HFP at the pre-exercise sitting (HFPbaseline) and the Wrst minute of the recovery (HFPrec1). Because the results of the previous studies have shown signiWcant diVerences in HRV during the Wrst recovery minutes, averages of 120 s (HFP120, LFP120, TP120) and 180 s (HFP180, LFP180, TP180) were also calculated during the Wrst minutes of the recovery.

HRV measurements

Training load measurements

RRIs were recorded (Suunto t6 wristop computer, Suunto Oy, Vantaa, Finland) continuously during the sessions with a sampling frequency of 1,000 Hz. RRI data was transferred to computer with Suunto Training Manager-software (Suunto Oy, Vantaa, Finland). Further signal processing and HRV calculations were performed with Matlab-software (Matlab 7, MathWorks, Inc., Natick, USA). RRIs were checked and edited for artifacts using a detecting algorithm and subsequently veriWed by visual inspection. The RRI series were then resampled at a rate of 5 Hz using linear interpolation to obtain equidistantly sampled time series. In order to remove low-frequency trends and variances below and above interest, a polynomial Wlter and a digital band-pass Wlter were used.

Breath-by-breath respiratory data (Oxycon Mobile, Viasys Healthcare GmbH, Hoechberg, Germany) was recorded continuously during rest, exercise, and recovery, except in PRO, in which the data were recorded at rest, during the last three kilometers of the exercise, and during the recovery. Averages of 60 s were used in statistical analysis. EPOC was calculated for the entire 15-min recovery. Rating of perceived exertion (RPE, scale 0–10) (Borg 1982) and Wngertip blood samples for blood lactate analysis (Biosen S_line Lab+, EKF-diacnostic GmbH, Barleben, Germany) were obtained immediately after the exercise. Also Training Impulse (exercise duration (min) £ (HRexercise ¡ HR rest)/(HRmax ¡ HR rest) £ 0.64 £ e[1.92 £ (HRexercise¡HRrest)/(HRmax¡HRrest)]), where e = Naperian

samples for blood lactate analysis (Biosen S_line Lab+, EKF-diacnostic GmbH, Barleben, Germany) were taken at the end of each work load. The highest 60-s VO2 value was considered as a maximal oxygen uptake (VO2max). Aerobic and anaerobic thresholds were determined using blood lactate, ventilation, VO2, and VCO2 (production of carbon dioxide) according to Aunola and Rusko (1984). Maximal velocity (vVO2max) was determined as the highest velocity of the test. If the subject could not complete the 3-min of the last velocity, the vVO2max was calculated using the last completed velocity (vlast) and the relative duration of the last uncompleted velocity (frac) as follows: vVO2max = vlast + frac. Constant load exercises

123

438

logarithm having a value of 2.712) and RPEs (Session RPE, exercise duration £ RPE) were calculated (Banister 1991). The term “training load parameters” is used in the present study to describe the parameters mentioned in this section. Statistical analysis

Eur J Appl Physiol (2010) 108:435–442 Table 2 Physiological and training load parameters in the exercises

HR (bpm) ¡1

VO2 (ml kg

Results Baseline HR and HRV The baseline values of HR and HRV did not diVer signiWcantly between the exercises. HR values prior to MODE, HI, and PRO were 59 (§12) bpm, 57 (§9) bpm and 56 (§7) bpm, respectively. HFPbaseline values prior to MODE, HI and PRO were 7.8 (§1.0), 7.9 (§0.9) and 7.8 (§0.9) ln(ms2), LFPbaseline 8.2 (§1.5), 8.2 (§1.4) and 8.2 (§1.3) ln(ms2) and TPbaseline values 8.9 (§1.2), 9.0 (§1.0) and 9.0 (§1.1) ln(ms2), respectively. Physiological and training load parameters during exercises HR reached the levels of 73 (§4) %, 92 (§3) % and 76 (§3) % of HRmax in MODE, HI and PRO, respectively. VO2 at the end of each exercise was 66 (§5) %, 95 (§4) % and 66 (§5) % of VO2max, respectively. The results of physiological and TL parameters are presented in Table 2. RPE and BLa values during PRO were not even doubled when compared with MODE, whereas session RPE and TRIMP values were over six and over four times greater in PRO when compared with MODE, respectively.

123

HI

134 § 11

167 § 9***

¡1

min ) 35.7 § 3.5 51.5 § 3.7***

PRO 138 § 10 35.7 § 2.9

RPE (0–10)

2.7 § 1.6

6.4 § 2.2***

3.6 § 2.0**

BLapeak (mmol l¡1)

1.4 § 0.6

7.6 § 2.9***

2.6 § 2.0*

52 § 31

RPEs

All values were expressed as means (§SD). Repeated measures analysis of variance (ANOVA) was used to compare the main eVects of exercise mode, recovery time, and their interaction during the recovery minutes 1–5. Repeated measures ANOVA with contrasts were used (1) to compare MODE and HI as well as MODE and PRO at each recovery minute and (2) to compare diVerences between the successive recovery minutes determined separately for each exercise. In addition, Student’s t-test was used to Wnd out the eVects of increased intensity or duration of MODE on TL parameters as well as on HFP120, LFP120, and TP120. Pearson correlation coeYcient was used to study relationships between TL parameters and HRV. The results of HRV during the last minute of exercise were excluded from the statistical analysis because of the noisy RRI data of several subjects. DiVerences between means were considered signiWcant when P < 0.05.

MODE

86 § 30***

326 § 184***

TRIMP

26.0 § 4.5 34.6 § 4.1***

116.6 § 20.7***

EPOC (ml kg¡1)

28.5 § 6.8 77.0 § 25.0***

33.6 § 7.9

MODE (moderate intensity) = 3 km at 60% vVO2max, HI (high intensity) = 3 km at 85% vVO2max, PRO (prolonged) = 14 km at 60% vVO2max VO2 oxygen uptake during the last 60-s of the exercise, BLapeak blood lactate at the end of the exercise, RPEs = session RPE = exercise duration £ RPE, TRIMP Training Impulse, EPOC excess post-exercise oxygen consumption during the 15-min recovery SigniWcantly diVerent from MODE *P < 0.05, **P < 0.01, ***P < 0.001

HRV recovery The results of HR, HFPln, LFPln, and TPln during the immediate 5-min recovery after exercise are presented in Fig. 1. In HR, HFPln, and TP the main eVects of exercise, recovery time and their interaction were all signiWcant (P < 0.05) between these three exercises. The successive HR values decreased (P < 0.001) during the 5-min recovery after each exercise, with the exception of the non-signiWcant diVerence between the third and fourth recovery minutes after MODE. The successive HFPln values increased during the Wrst two recovery minutes after MODE (P < 0.05), HI (P < 0.001) and PRO (P < 0.05). In LFPln, the main eVect of exercise (P < 0.001) and recovery time (P < 0.001) were signiWcant between these three exercises, but not their interaction. The successive LFPln values increased during the Wrst two recovery minutes in MODE (P < 0.01) and HI (P < 0.001), and also during the third minute after PRO (P < 0.05). As presented in Fig. 2, HFP120, LFP120 and TP120 were lower (P < 0.05) in HI and in PRO when compared with MODE. Similar results were also found in HFP180, LFP180, andTP180 (Fig. 2), HFPdiV (HFPbaseline ¡ HFPrec1), an additional HRV parameter indicating TL, was higher both in HI (P < 0.05) and PRO (P < 0.05) when compared with MODE. At the end of the 15-min recovery, HR was still (P < 0.001) higher and HFPln, and TPln lower (P < 0.05) after each exercise when compared with the baseline. LFPln was lower than baseline in the end of the recovery after HI (P = 0.05), but did not diVer from the baseline after MODE and PRO. The diVerences between exercises at the end of the recovery were similar to the fast recovery; HR was higher (P < 0.001) and HFPln, LFPln, and TPln lower (P < 0.001)


439

Fig. 1 The recovery dynamics of HR (a), LFP (b), HFP (c) and TP (d). MODE (moderate intensity) = 3 km at 60% vVO2max, HI (high intensity) = 3 km at 85% vVO2max, PRO (prolonged) = 14 km at 60%

vVO2max. DiVerences between MODE and HI *P < 0.05, **P < 0.01, ***P < 0.001, diVerences between MODE and PRO P < 0.05

Fig. 2 DiVerences in HFP, LFP and TP between the exercises during the acute recovery. Two-minute means in a and 3-min means in b. MODE (moderate intensity) = 3 km at 60% vVO2max, HI (high

intensity) = 3 km at 85% vVO2max, PRO (prolonged) = 14 km at 60% vVO2max. SigniWcantly diVerent from MODE *P < 0.05, **P < 0.01, ***P < 0.001

after HI when compared with MODE. HR was also higher after PRO when compared with MODE, but no diVerences in HFPln, LFPln, or in TPln was found between them (Fig. 1).

Discussion

Relationships between training load parameters and HRV When all three exercises were included, a negative (P < 0.05) correlation between EPOC and LFP120 and also between EPOC and LFP180 was found. RPE (P < 0.001) and BLa (P < 0.05) correlated negatively with HFP120, LFP120, and TP120 and also with HFP180, LFP180, and TP180. Results of the correlations are presented in Table 3.

Although the physiological eVects of exercise intensity and duration have conventionally been estimated with measures such as HR, BLa, RPE, EPOC, and TRIMP, there is no “golden standard” method for TL analysis. The aim of this study was to examine if any post-exercise HRV indices could be used to evaluate TL in exercises when one of the two major determinants of TL, intensity or duration, was increased. The results of the present study can be summarized with two main Wndings. First, over fourfold increase in the distance of running at 60% vVO2max as well as the

123

440


Table 3 Correlations (r) between training load parameters and HRV RPE

BLa

EPOC

TRIMP

RPEs

HFP120

¡0.604***

¡0.401*

ns.

ns.

¡0.414*

LFP120

¡0.634***

¡0.601***

¡0.399*

ns.

ns.

TP120

¡0.691***

¡0.569**

ns.

ns.

ns.

HFP180

¡0.586***

¡0.409*

ns.

ns.

¡0.367*

LFP180

¡0.638***

¡0.622***

¡0.406*

ns.

ns.

TP180

¡0.674***

¡0.580**

ns.

ns.

ns.

BLa blood lactate at the end of the exercise, RPEs (session RPE) = exercise duration £ RPE, TRIMP Training Impulse, EPOC excess post-exercise oxygen consumption during the 15-min recovery *P < 0.05, **P < 0.01, ***P < 0.001. All three exercises are included

increased intensity from 60 to 85% vVO2max decreased HFPln, LFPln, and TPln during the Wrst 2 min of recovery. Second, the decreased post-exercise HRV was strongly related to the increased RPE and BLa. EVects of exercise duration and intensity on HRV The eVect of exercise duration on immediate post-exercise HRV has remained largely unexplored with the exception of a previous study of Kaikkonen et al. (2007), who found no eVect of doubled exercise duration, from 3,500 to 7,000 m at 50% and 64% of vVO2max, in acute post-exercise HRV in sedentary women. The diVerences compared with the present results may be partly due to the magnitude of the increase in exercise duration; prolonged exercise in the study of Kaikkonen et al. (2007) was “only” doubled, while the PRO exercise in the present study was almost Wve times longer than MODE. Also Seiler et al. (2007) recently investigated the eVect of exercise duration on post-exercise HRV in highly trained athletes, but they did not analyze the immediate 5-min HRV recovery. They found that HRV recovery during 4-h period was not aVected at all by extending the exercise from 60 min to 120 min at the intensity below the Wrst ventilatory threshold (61% VO2max). The discrepancy in Seiler et al.’s Wndings and those of the present study may have a number of explanations. In addition to the diVerence in the timing of HRV analysis, Seiler et al. examined the recovery of highly trained athletes, which is presumably faster than that of the moderately trained participants of the present study. Furthermore, in the study of Seiler et al., the long exercise was, similar to Kaikkonen et al. (2007), twice the duration of the “baseline” exercise, while the duration of PRO in the present study was almost Wve times longer than the MODE. It is also important to note that the intensity of the exercise may be very signiWcant when assessing the eVects of duration on HRV recovery. At intensities exceeding anaerobic threshold the metabolic load of the exercise

123

accumulates, which might be observed as decreased HRV. In comparison, low- or moderate-intensity exercises, during which the sympathetic nervous system is not greatly activated, can be supposed to have smaller eVects on vagal activity during the exercise and reactivation during the recovery. As expected, based on the previous studies (Buchheit et al. 2007; Casties et al. 2006; Goldberger et al. 2006; Kaikkonen et al. 2008; Kaikkonen et al. 2007; Martinmäki and Rusko 2008; Seiler et al. 2007), the recovery of HRV was signiWcantly delayed after the high-intensity exercise. For example, Seiler et al. (2007) found similar HRV recovery during a 4-h period after 60 min and 120 min exercises below the Wrst ventilatory threshold in highly trained athletes, but signiWcantly delayed recovery after exercise at the second ventilatory threshold. No further delay was found despite the increase of exercise intensity up to 95% of VO2max. The authors concluded that the Wrst ventilatory threshold appeared to be a clear threshold for ANS perturbation in highly trained athletes. The results of the present study conWrm the signiWcantly delayed HRV recovery after exercise near anaerobic threshold. The high-intensity exercise (85% vVO2max) was performed above anaerobic threshold in all but three subjects whereas MODE and PRO were performed near the aerobic threshold in most of the subjects. HR and HRV recovery dynamics In the present study, a signiWcant decrease in HR was found during the Wrst Wve recovery minutes. As previous studies have proved, HR decreases with an exponential pattern during the Wrst recovery minutes (Savin et al. 1982), and the decrease of the Wrst minute is suggested to be due to increase in vagal activity. ConWrming that, we also found signiWcant increase in HFPln during the Wrst two recovery minutes after each exercise. After 2 min, HFPln increased at a more gradual rate toward the pre-exercise baseline values. This Wnding was in contrast to the previous studies of Kaikkonen et al. (2007, 2008) in which barely any recovery of HFPln was detected during the Wrst Wve recovery minutes after the high-intensity exercises, e.g., 85–93% of vVO2max in athletes (2008) and 63–74% of vVO2max in sedentary women (2007). These diVerent Wndings are probably a result of diVerences in training status as well as in exercises; the recovery was slower in sedentary women when compared with athletes, probably because they were unaccustomed to such exercises. This contributing factor includes the lower anaerobic threshold of the subjects of Kaikkonen et al. (2008) (74% vs. 81% of vVO2max) resulting in higher perceived exertion (9 § 1 in the study of Kaikkonen et al. vs. 6 § 2 in the present study).


Training load and HRV The prolonged exercise at the present intensity aVected signiWcantly RPE and blood lactate, in addition to HRV. RPE and blood lactate also correlated negatively with post-exercise HRV. The greater was RPE or blood lactate during the exercise, the lower was HRV during the Wrst 2 min recovery. In previous studies, RPE has been found to correlate with the level of cardio-respiratory and metabolic demand (Noble 1982; Skinner et al. 1973) and to integrate various informations, e.g., signals from working muscles and joints, and also from the central nervous system (Borg 1982). Although RPE was invented to describe the intensity of exercise, its subjective nature takes also the duration of exercise into account. Since HRV has been used as a noninvasive method to estimate function of ANS, which regulates body homeostasis (Porges 1992), during exercise and recovery, the relationship between HRV and RPE supports our assumption of RPE being a “basic” subjective variable giving us quite good knowledge of the exercise-induced TL. However, HRV as a physiological variable can be seen as a more reliable measure of changes in body homeostasis than RPE. We found no eVect of prolonged exercise duration on EPOC, despite the moderate (60% vVO2max) intensity of the exercises. The correlation between EPOC and HRV was also weaker than that between HRV and RPE, and correlations were found only between LFPln and EPOC. Exercise intensity has been shown to have a curvilinear relationship to EPOC, whereas the eVect of exercise duration has been found more linear (Børsheim and Bahr 2003). Gore and Withers (1990) reported that both the intensity and the duration of exercise have a threshold point from which the increase of EPOC is fastened. They found that duration aVects the magnitude of 8-h EPOC only when exercise intensity exceeded 50% VO2max. In this study, the intensity of PRO was 60% of vVO2max, yet increased exercise duration failed to have an eVect on EPOC. One explanation may be the duration of EPOC measurement. We measured EPOC only for 15 min post-exercise, and the result could have been diVerent if the duration had been closer to the recommended 60 min–24 h (Børsheim and Bahr 2003). When compared with all other TL parameters, the results of TRIMP and RPEs were noticeably diVerent and seemed to overestimate the eVect of exercise duration. In the present study, TRIMP was over four times and session RPE over six times greater in PRO compared with MODE, despite similar HR and EPOC values between exercises. In addition, TRIMP did not correlate with any HRV parameter, so the results of the present study do not encourage their use as TL indicators. In summary, the aim of this study was to examine if any post-exercise HRV indices could be used to evaluate TL in

441

exercises with diVerent intensities and durations. The evaluation of HRV as a method for estimation of TL is complex since there is at present no “golden standard” to which HRV parameters may be compared. We hypothesized that since HRV has been used as an indicator of ANS function, which regulates body homeostasis, post-exercise HRV might reXect TL-induced changes in body homeostasis. In this study, we found (a) signiWcantly delayed HRV recovery after increased intensity or prolonged duration of exercise and (b) a signiWcant relationship between post-exercise HRV and RPE. Furthermore, because both increased intensity as well as duration of exercise was shown to have an eVect on post-exercise HRV, it seems to be one objective tool for estimation of TL. However, the Wndings of this study are preliminary and further research is required when developing HRV-based TL measurements. Acknowledgments This study was funded by grants from TEKESNational Technology Agency of Finland, Emil Aaltonen Foundation and Foundation of Sports Institute. The results of the present study do not constitute endorsement by ACSM.

References American College of Sports Medicine (2006) ACSM’s guidelines for exercise testing and prescription, 7th edn. Lippincott Williams & Wilkins, Philadelphia, pp 19–35 Arai Y, Saul J, Albrecht P, Hartley L, Lilly L, Cohen R, Colucci W (1989) Modulation of cardiac autonomic activity during and immediately after exercise. Am J Physiol 256:H132–H141 Aunola S, Rusko H (1984) Reproducibility of aerobic and anaerobic thresholds on 20–50 year old men. Eur J Appl Physiol 53:260– 266 Banister E (1991) Modelling elite athletic performance. In: MacDougall JD, Wenger HA, Green HJ (eds) Physiological testing of the high-performance athlete, 2nd edn. Human Kinetics Publishers Ltd, Champaign, pp 403–424 Berntson G, Cacioppo J, Quigley K (1993) Respiratory sinus arrhythmia: autonomic origins, physiological mechanisms, and psychophysiological implications. Psychophysiology 30:183– 196 Borg G (1982) Psychophysical bases of perceived exertion. Med Sci Sports Exerc 14:377–381 Børsheim E, Bahr R (2003) EVect of exercise intensity, duration and mode on post-exercise oxygen consumption. Sports Med 33:1037–1060 Buchheit M, Laursen P, Ahmaidi S (2007) Parasympathetic reactivation after repeated sprint exercise. Am J Physiol Heart Circ Physiol 293:H133–H141 Casties J-F, Mottet D, Le Gallais D (2006) Non-linear analysis of heart rate variability during heavy exercise and recovery in cyclists. Int J Sports Med 27:780–785 Cole C, Blackstone E, Pashkow F, Snader C, Lauer M (1999) Heartrate recovery immediately after exercise is a predictor of mortality. N Engl J Med 341:1351–1357 Foster C (1998) Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc 30:1164–1168 Goldberger J, Le F, Lahiri M, Kannankeril P, Ng J, Kadish A (2006) Assessment of parasympathetic reactivation after exercise. Am J Physiol 290:2446–2452

123

442 Gore C, Withers R (1990) The eVect of exercise intensity and duration on the oxygen deWcit and excess post-exercise oxygen consumption. Eur J Appl Physiol 60:169–174 Kaikkonen P, Nummela A, Rusko H (2007) Heart rate variability dynamics during early recovery after diVerent endurance exercises. Eur J Appl Physiol 102:79–86 Kaikkonen P, Rusko H, Martinmäki K (2008) Post-exercise heart rate variability of endurance athletes after diVerent high-intensity exercises. Scand J Med Sci Sports 18:511–519 Keselbrener L, Akselrod S (1996) Selective discrete Fourier transform algorithm for time-frequency analysis: method and application on simulated and cardiovascular signals. IEEE Trans Biomed Eng 43:789–802 Martinmäki K, Rusko H (2008) Time-frequency analysis of heart rate variability during immediate recovery from low and high intensity exercise. Eur J Appl Physiol 102:353–360 Martinmäki K, Rusko H, Saalasti S, Kettunen J (2006) Ability of short-time Fourier Transform to detect transient changes in vagal eVects on hearts: a pharmacological blocking study. Am J Physiol 290:2582–2589 Miles D, Sawka M, Hanpeter D, Foster J, Doerr B, Basset Frey M (1984) Central hemodynamics during progressive upper and lower body exercise and recovery. J Appl Physiol 57:366–370 Noble B (1982) Clinical applications of perceived exertion. Med Sci Sports Exerc 14:406–411 O’Leary D (1993) Autonomic mechanisms of muscle metaboreXex control of heart rate. J Appl Physiol 74:1748–1754

123

Eur J Appl Physiol (2010) 108:435–442 Oida E, Moritani T, Yamori Y (1997) Tone-entropy analysis on cardiac recovery after dynamic exercise. J Appl Physiol 82:1794–1801 Perini R, Orizio C, Comandè A, Castellano M, Beschi M, Veicsteinas A (1989) Plasma norepinephrine and heart rate dynamics during recovery from submaximal exercise in man. Eur J Appl Physiol 58:879–883 Pichon A, De Bisschop C, Roulaud M, Denjean A, Papelier Y (2004) Spectral analysis of heart rate variability during exercise in trained subjects. Med Sci Sports Exerc 36:1702–1708 Pierpont G, Stolpman D, Gornick C (2000) Heart rate recovery postexercise as an index of parasympathetic activity. J Auton Nerv Syst 80:169–174 Plotnick C, Becker L, Fisher M (1986) Changes in left ventricular function during recovery from upright bicycle exercise in normal persons and patients with coronary artery disease. Am J Cardiol 58:247–251 Porges S (1992) Vagal tone: a physiologic marker of stress vulnerability. Pediatrics 90:498–504 Savin W, Davidson D, Haskell W (1982) Autonomic contribution to heart rate recovery from exercise in humans. J Appl Physiol 53:1572–1575 Seiler S, Haugen O, KuVel E (2007) Autonomic recovery after exercise in trained athletes: intensity and duration eVects. Med Sci Sports Exerc 39:1366–1373 Skinner J, Hustler R, Bergsteinova V, Buskirk E (1973) The validity and reliability of a rating scale of perceived exertion. Med Sci Sports 5:97–103