1 INTRODUCTION

Several physiological measures might be used to detect positive and negative training effects in athletes, helping coaches to better design and adjust their training programs.1 Among these measures, heart rate variability (HRV) has been attracting attention2; its use is boosted by the non-invasive nature and technical simplicity of the data collection. HRV indices are commonly derived from ultra-short (1-min)3 or short-term (5–10-min)4 R–R interval recordings in the seated or supine position.2 It was previously shown that ultra-short-term and short-term HRV indices are similar,5 and both reflect the resting cardiac autonomic modulation under standardized conditions.6 In fact, short-term vagal indices are reduced by acutely administering muscarinic cholinergic blockade.7 Although these measures are practical and easy to perform, more laborious analysis of HRV throughout 24 h of free-living ambulatory conditions could reveal other facets of cardiac autonomic regulation.

The relationship between R–R intervals and the high-frequency oscillations of heart rate (HR) in 5-min segments over 24-h during a non-exercising day could reveal the occurrence or not of the saturation effect of vagal activity.8, 9 In individuals with non-saturated dynamics, the relationship between R–R intervals and HF power collected over 24-h is highly linear, indicating that bradycardia is modulated by an increased vagal outflow to the sinus node. Bradycardia is more pronounced during overnight sleep when vagal activity reaches its peak due to the prolonged time without stressful stimuli or events.10 On the other hand, an increase in HR during an ambulatory activity during awake hours is determined by the parasympathetic withdrawal, associated with sympathetic activation.11

Though, in subjects with saturation of HF oscillations, the relationship between R–R intervals and HF is better fitted to a quadratic function, displaying a plateau or even decay in HF power at high R–R intervals, especially during night sleep.8 It is interesting that in a study8 conducted with 76 healthy individuals (non-athletes) and 82 patients with post-acute myocardial infarction (AMI), the relationship between R–R intervals and HF power was linear in 50% of the healthy participants, while it was saturated in 46% and non-significantly correlated in only 3 healthy participants (4%). This latter case is suggestive of non-autonomic mechanisms driving HR control.8, 9 However, in patients with post-AMI, only 11% of the participants presented saturated dynamics, while most participants presented linear dynamics (54%) or low-correlated dynamics (35%). In both healthy and post-AMI groups, the participants that displayed saturation of the vagal outflow tended to display greater aerobic power as assessed by maximal oxygen consumption (VO2max) when compared to participants with linear and, especially, low-correlated dynamics. Therefore, it seems that aerobic fitness is a determinant of the saturation effect of vagal activity.8, 9 Furthermore, it has also been shown that resting HRV and the aerobic fitness performance are positively associated with novice male soccer players (r = 0.72).12 It is important to emphasize that other factors such as age, sex, and body composition13, 14 can also influence HRV and thus the occurrence or not of the saturation effect; however, this topic deserves future investigations.

To the best of our knowledge, no studies so far reported the prevalence of saturated and linear dynamics of the R–R intervals and HF power oscillations in team sports athletes, whereas only one study has been conducted with endurance athletes.15 In addition, the prevalence of vagal saturation has not been explored in female athletes to date. Generally, it is expected that athletes enhance cardiorespiratory fitness during the preseason period,16 which might increase the prevalence of players displaying saturation of vagal activity and vagal-mediated HRV indices determined during 24 h of free-living R–R interval data collection. In a previous study involving healthy male individuals, a pre-training analysis showed that 7 out of 17 subjects displayed saturation of vagal activity and post-training five new cases of saturation of vagal activity were observed.9 Therefore, one can expect that a soccer preseason could increase the prevalence of the saturation of the HF oscillations of the R–R interval variability in soccer players.

Due to the lack of evidence in team sports athletes and in women, the aim of the current investigation was to analyze the changes in the relationship between R–R intervals and HF over 24-h (and its saturation) in female soccer players during a typical 4-week preseason period. Additionally, it was investigated the association between aerobic fitness and vagal-related HRV indices (ie, 24-h HRV), and the respective changes in response to training.

2 METHODS 2.1 Participants

Sixteen female soccer players (age: 21.8 ± 2.6 years; height: 159.6 ± 5.8 cm; body mass: 56.8 ± 5.5 kg) competing in the 1st division of the Portuguese soccer championship and involved in the 2018–19 preseason were included.

The study design was carefully explained to the participants and written informed consent was obtained from each player. Inclusion criteria were: age ≥18 years, playing soccer for at least 5 years, participating in all training sessions of the preseason, and classified as good sleepers17 (ie, scores <5; according to Pittsburgh Sleep Quality Index [PSQI] questionnaire). Exclusion criteria were being a goalkeeper (n = 2), tobacco use (n = 2), and medical conditions contraindicating physical exercise (n = 3). The study followed the Declaration of Helsinki and was approved by the Ethics Committee of the Faculty of Sports, University of Porto (CEFADE 03.2017).

2.2 Study design

All training sessions took place at the same period of the day, starting at 18:12 ± 00:08 h. The training sessions planned by the coach consisted mostly of warm-up and activation; technical training; small-sided games; and scrimmages. Training load was measured during each training session by the session-rating of perceived exertion (s-RPE) to quantify the internal load of the practices during the period of observation.

After preseason all players filled out the PSQI questionnaire, which has been previously validated to Portuguese.17 The sleep habits of participants were not constrained by the study procedures in order to grant high ecological validity to the investigation. All players were classified as good sleepers during the period of observation (PSQI score: 4 ± 1 [mean ± SD; individual range: 3–5]). During the study period, the players went to bed at 22:00 ± 0:50 h (individual range: 21:30–23:30 h) with a sleeping duration of 8:30 ± 0:30 h (individual range: 8:00–9:00 h).

Ambulatory R–R intervals (24-h duration) were recorded during a non-exercise day, 48-h before the start (PRE-preseason) and 48-h after the end (POST-preseason) of the preseason period. Each player self-performed the R–R interval recordings at home. The players were instructed not to practice any type of exercise 48-h before the R–R interval recordings performed in PRE- and POST-preseason. In addition, the players completed a YYIR1 during the first and last week of the preseason to assess aerobic fitness. A chronological scheme of the study is illustrated in Figure 1.

Scheme of the study design

2.3 Yo-Yo recovery test—level 1

The YYIR1 is an intense intermittent exercise performance test, specially developed to evaluate the aerobic fitness to perform high-intensity intermittent exercise of soccer players. Moreover, the YYIR1 test can be used as an indicator of aerobic fitness for female soccer players.18 The aim of the test was to perform as many shuttles (2 × 20-m runs with 5 m of active recovery) as possible. When the player failed twice to reach the finish line in time, the final distance covered was recorded.

2.4 Session-rating of perceived exertion

Training load during the preseason was quantified using the s-RPE method, which has been validated for female soccer players.19 The players reported individual ratings of perceived exertion (RPE) using Borg’s category ratio scale (CR10) after each training session. The CR10 score was subsequently multiplied by individual exposure time, thus providing an overall load quantification of the session. The daily average training load across the preseason was 604 ± 70 AU (mean ± SD; individual range: 510–900 a.u.), with each training session lasting 90 ± 5 min (individual training time range: 85–95 min). The accumulated s-RPE for each week were: week 1: 2453 ± 534 a.u. [1730–2860 a.u.]; week 2: 2454 ± 426 a.u. [1710–2840 a.u.]; week 3: 2371 ± 315 a.u. [1610–2720 a.u.]; and week 4: 2379 ± 321 a.u. [1640–2710 a.u.]. No significant differences were found in accumulated s-RPE across the weeks (p > 0.05).

2.5 Twenty-four-hour analysis of heart rate variability

During the previous season, all players were familiarized with the portable HR monitor used during the current investigation. Ambulatory R–R intervals (R-R24h) were recorded using HR monitors (Firstbeat Bodyguard2®, Firstbeat Technologies). The wearable R–R interval recording has been validated against standard ECG equipment to detect heartbeats.20 The R-R intervals were analyzed using Kubios HRV, 3.0.0® software (Biosignal Analysis and Medical Imaging Group at the Department of Applied Physics, University of Kuopio, Kuopio, Finland).

The mean R-R24h (ie, intervals between consecutive heartbeats [beat-to-beat]) lengths and the corresponding HF power values (HF24h) were analyzed in 5-min segments over the whole 24-h recording.8 HF24h values are expressed in ms2.

All 5-min values of HF24h (y-axis) were plotted as a function of the corresponding R-R24h (x-axis). A quadratic regression model (HF, ln = a(R–R)2 + b(R–R) + c) was used to model the relationship between the R-R24h length and the magnitude of beat-to-beat HRV.8 If the quadratic correlation coefficient was >0.50, it was established whether the relationship between R-R24h length and HF24h was saturated or linear.8 Otherwise, if the quadratic correlation coefficient was <0.50, indicating a low correlation, the participants were excluded from the following analyses.8 The value of R-R24h at which the derivative of the quadratic regression model reached a zero value was defined as the deflection point (R–R0).8 If the R–R0 occurred before the maximum R-R24h value, indicating the plateau of HF24h (HF24h reached its maximum value), the relationship between R-R24h length and HF24h was defined as saturated (Figure 2).8 Otherwise, the relationship was considered linear. The HF index was intended to measure cardiac vagal outflow without saturation disturbing the analysis.8 In this case, HF24h was calculated separately using the linear portion of the R-R24h versus the HF24h regression curve. Hence, HF24h was analyzed from the window of R-R24h lengths shorter than R–R0. The lower 95% confidence interval of R–R0 was defined as the upper limit for the analysis of the HF index.8 All 5-min HF24h values at R-R24h lengths of less than 95% confidence interval of R–R0 (saturated cases) or less than 95% confidence interval of the maximum R-R24hvalue (linear cases) were averaged for each case, and the mean value was defined as the HF index.8 The average of the corresponding R-R24h values was defined as the R–R index.

Representative example of cases in which the relationship between HF24h and R-R24h were as follows: (A) linear PRE-preseason and became saturated POST-preseason (LiSa); (B) maintained saturated PRE- and POST-preseason (SaSa); and (C) maintained linear PRE- and POST-preseason (LiLi). R–R, beat-beat interval; R–R0, R–R interval length at deflection point; HF, high-frequency power of R–R interval variability; Δ, % change; R, quadratic correlation coefficient

In the current study, different groups of players were defined to cluster the patterns of cardiac autonomic adaptation: linear PRE-preseason that became saturated POST-preseason (LiSa); remained saturated PRE- and POST-preseason (SaSa); and linear PRE- and POST-preseason (LiLi).

2.6 Statistical analysis

To determine the number of athletes required for the study we computed a priori statistical power analysis (G*Power 3, University Dusseldorf, Germany). The sample size was calculated by a priory power analysis based on the results of Kiviniemi et al studies.9, 15 The sample size calculation to detect changes in the main outcomes revealed that 12 subjects would be necessary to detect differences with a power of 80% and α = 0.05.

The results for the groups are expressed as mean ± standard deviation. All data sets were assessed for normal distribution using the Shapiro–Wilk test. The within-subject differences (n = 16) for the YYIR1, HR-derived indices (ie, R-R24h, HF24h, R-R index, HF index), and within each group (ie, LiSa, n = 5; SaSa, n = 7; and LiLi, n = 4) between PRE- versus POST-preseason comparisons were analyzed using paired-sample t-test. The differences for the linear versus saturated participants’ results, during PRE- and POST-preseason, were analyzed using independent-sample t-test. Finally, the differences in accumulated s-RPE between-participants for each week (ie, 4 weeks) of the preseason were analyzed using one-way ANOVA with Bonferroni-corrected paired-sample t-tests. The magnitude of the differences was examined using the standardized differences based on Cohen’s d units by means of the effect sizes (ES) analysis, with corresponding 95% confidence intervals (CI).21 The ES was qualitatively interpreted using the following thresholds: <0.2, trivial; 0.2 to 0.6, small; 0.6 to 1.2, moderate; 1.2 to 2.0, large; 2.0 to 4.0, very large; and >4.0, nearly perfect.22 Pearson's product-moment correlation analysis was used to compute associations between YYIR1 performance and vagal-related HRV indices (ie, HF24h and HF index), PRE- and POST-preseason. Additionally, associations between changes in aerobic fitness and vagal-related HRV indices (absolute and % changes) were also calculated. The threshold used to qualitatively assess the correlations was based on the following criteria: <0.1, trivial; 0.1–0.3, small; 0.3–0.5, moderate; 0.5–0.7, large; 0.7–0.9, very large; >0.9 nearly perfect.23 For all comparisons, a p-value of <0.05 was considered statistically significant.

3 RESULTS

Significant differences were found in YYIR1 PRE-preseason compared to POST-preseason: 930 ± 286 m (mean ± SD; individual range: 400–1240 m) versus 1265 ± 252 m (640–1640 m), respectively; ES = 1.25 [0.50; 2.0] (ES, effect size [confidence intervals]); large effect; p < 0.001.

The whole group R-R24h (moderate effect), HF24h (moderate effect), R-R index (moderate effect), and HF index (moderate effect) significantly increased POST-preseason (p < 0.05; Table 1) compared to PRE-preseason (n = 16).

TABLE 1. Mean, SD, ES, and 95% CI values of selected demographic parameters and related HR indices over 24-h recordings PRE- and POST-preseason in female soccer players All players (n = 16) PRE-preseason POST-preseason PRE-preseason POST-preseason ES (95% CI) Linear (n = 9) Saturated (n = 7) ES (95% CI) Linear (n = 4) Saturated (n = 12) ES (95% CI) PRE-preseason POST-preseason R–R24h (ms) 973 ± 207 1045 ± 215* 0.64 (0.60–0.67) 935 ± 183 1024 ± 224** 0.64 (0.58–0.70) 972 ± 189 1070 ± 217** 0.67 (0.60–0.73) HF24h (ln ms) 7.53 ± 1.17 7.81 ± 1.01* 0.65 (0.62–0.69) 7.44 ± 1.12 7.63 ± 1.21 ** 0.67 (0.61–0.72) 7.52 ± 1.01 7.97 ± 0.95** 0.66 (0.59–0.72) R–R index (ms) 763 ± 89 829 ± 110* 0.75 (0.69–0.82) 750 ± 81 770 ± 95** 0.62 (0.42–0.72) 817 ± 103 834 ± 115** 0.63 (0.55–0.76) HF index (ln ms) 6.56 ± 1.01 7.01 ± 1.01* 0.68 (0.57–0.71) 6.62 ± 1.04 6.74 ± 0.98 0.40 (0.31–0.52) 6.68 ± 0.84 7.20 ± 1.04** 0.61 (0.54–0.71) Abbreviations: CI, confidence interval; ES, effect size; HF, high-frequency power of R–R interval variability; R–R, beat-beat interval.

In the PRE-preseason, the relationship between R-R24h length and the corresponding HF24h was saturated in 7 players and linear in 9 players, while in the POST-preseason, 5 new cases of saturated HF24h were observed and the remaining 4 players maintained linear relationship. Moderate effects with increased values were found for the mean R-R24h, HF24h, R-R index and HF index, when compared the saturated and linear players for both PRE and POST-preseason (p < 0.05; Table 1).

In all groups of players (LiSa, SaSa and LiLi), moderate effects for the R-R24h, HF24h, R-R index, and HF index were found between PRE and POST-preseason (p < 0.05; Table 2).

TABLE 2. Mean, SD, ES, and 95% CI of selected demographic parameters (ie, (A) linear PRE-preseason and became saturated POST-preseason [LiSa]; (B) maintained saturated PRE- and POST-preseason [SaSa]; and (C) maintained linear PRE- and POST-preseason [LiLi]) and related HR indices over 24-h recordings PRE- and POST-preseason in female soccer players LiSa (n = 5) SaSa (n = 7) LiLi (n = 4) PRE-preseason POST-preseason ES (95% CI) PRE-preseason POST-preseason ES (95% CI) PRE-preseason POST-preseason ES (95% CI) R–R24h (ms) 957 ± 175 1066 ± 217 * 0.65 (0.59−0.71) 1023 ± 225 1072 ± 218 * 0.62 (0.56–0.71) 905 ± 188 972 ± 189 * 0.68 (0.60–0.72) HF24h (ln ms) 7.64 ± 1.01 7.82 ± 0.86 * 0.62 (0.52–0.75) 7.76 ± 1.21 8.10 ± 1.01* 0.64 (0.54–0.70) 7.20 ± 1.22 7.32 ± 1.01 * 0.61 (0.54–0.72) R–R index (ms) 782 ± 97 838 ± 127 * 0.69 (0.58–0.81) 751 ± 84 835 ± 103 * 0.89 (0.77–0.91) 760 ± 86 816 ± 103 * 0.61 (0.52–0.73) HF index (ln ms) 6.76 ± 0.99 7.20 ± 0.95 * 0.65 (0.53–0.77) 6.78 ± 1.00 7.31 ± 1.08 * 0.61 (0.49–0.72) 6.33 ± 0.98 6.59 ± 0.84 * 0.68 (0.56–0.80) Abbreviations: CI, confidence interval; ES, effect size; HF, high-frequency power of R–R interval variability; R–R, beat-beat interval.

Figure 3 displays the correlation between the distance covered in the YYIR1 (335 ± 202 m [individual range: 20–720 m]) and vagal-related HRV indices in both PRE- and POST-preseason. HF24h and HF index were largely correlated with the distance covered in the YYIR1 (PRE-preseason: r = 0.58 [0.20; 0.84], and r = 0.69 [0.37; 0.83]; POST-preseason: r = 0.64 [0.38; 0.87], and r = 0.65 [0.34; 0.82]; respectively; p < 0.05).

Correlation coefficients (95% confidence limits) between intermittent recovery teste level 1 distance (YYIR1) and vagal-related HRV indices (ie, HF24h and HF index), PRE- and POST-preseason. Associations between aerobic capacity changes (absolute and % YYIR1 changes) and vagal-related HRV indices changes (absolute and % changes for both HF24h and HF index) were also calculated, in female soccer players (n = 16). p < 0.05 for all correlation coefficients

Significant and large correlations were found between absolute and % changes in aerobic fitness and vagal-related HRV indices (absolute changes HF24h: r = 0.65 [0.32; 0.86], and % changes HF24h: r = 0.68 [0.30; 0.88]; absolute changes HF index: r = 0.53 [0.28; 0.81], and % changes HF index: r = 0.56 [0.29; 0.83]); (p < 0.05) (Figure 3).

4 DISCUSSION

In the current investigation, the proportion of female soccer players displaying saturated vagal modulation of 24-h R-R interval recordings versus HF24h prior to the start of preseason preparation was substantially lower than after only 4 weeks of preseason training. In addition, the whole group displayed a significant increase in the R-R intervals and HF24h in the POST-preseason compared to PRE-preseason, which indicates enhanced vagal modulation during the 24-h recording. In addition to the large correlations found between HF24h and HF index with the distance covered in the YYIR1, in both PRE- and POST-preseason, changes in HF24h and HF index were also largely correlated with changes in aerobic fitness.

The increased cardiac vagal activity in response to preseason training in team sports has been described,24-26 and there are reports of changes during this period of preparation (ie, positive adaptations).27, 28 It is important to note that all the cited studies used overnight or short-term (5–15 min in the seated or supine position) awake recordings, while we performed, for the first time, 24-h recordings in team sports players. It is clear that the former options are easier to implement in sports settings, but the latter can provide a more thorough analysis of the autonomic nervous system function across different daily behaviors (eg, walking and sleeping).6, 8, 9, 15 The increase in vagal activity might be an adaptive advantage to the players, as HF has been reported to be highly correlated with aerobic fitness.12, 29, 30 Therefore, the increment in YYIR1 performance (34%) observed in the current investigation is consistent with the improvement in the HF24h and HF index. In fact, large and positive correlations were found between HF24h and HF index with YYIR1 (PRE- and POST-preseason), and between changes in 24-h HRV and changes in YYIR1 performance. Accordingly, it appears that improvement in vagally mediated HRV and aerobic fitness are associated, as previously suggested while using short-term exercise and resting HRV measures (3 to 10-min).29, 31 However, the mechanisms behind this association need to be clarified.

Using 24-h recording data, the results presented here show that players displayed greater cardiac vagal activity in the POST-preseason than in the PRE-preseason. Accordingly, Scharf, Brem, Wilhelm, Schoepf, Uder, Lell32 found cardiac morphological and functional adaptations (ie, left and right ventricular dilation accompanied by increased myocardial mass and wall thickness) in a group of male professional soccer players but not in a control group of healthy men. In fact, aerobic exercise training leads to electrical, functional, and morphologic cardiovascular adaptations, including resting bradycardia, increased maximal oxygen uptake, enhanced diastolic function, and atrial and ventricular remodeling.33 These adaptations are mainly due to increased volume load on the ventricle and episodes of high cardiac output,34 and can partly explain the reduction in 24-h HR. Furthermore, this bradycardic adaptation at rest has been explained by the increased vagal tone induced by dynamic exercise35 and agrees with the results of other studies in soccer players.32

Interestingly, the magnitude of these cardiac physiological adaptations might be sex-dependent, being more marked in male athletes. In a study performed by Huten et al,36 electrocardiographic training-induced changes were common among elite male soccer players, especially in electrical criteria for left ventricular hypertrophy (37%) and sinus bradycardia (17%). Highly trained female athletes might also have display structural cardiac adaptations, but those adaptations are less pronounced than in male athletes.37 Nonetheless, due to differences in circulating hormones leading to a higher sympathetic activity and a lower parasympathetic modulation in men compared to women,13 differences in training-related adaptations may arise, even when sexes are matched for training volume.38

It is likely that the increased R-R24h and R-R index can also be modulated by the higher vagal outflow throughout the day. Previously, Kiviniemi, Hautala, Makikallio, Seppanen, Huikuri, Tulppo9 found that HF power analyzed over 24 h did not change in male healthy subjects who already had a saturated HF profile, despite the lengthening of R-R intervals. Longer R-R intervals after training could be mostly explained by slower HR during sleep due to enhanced vagal outflow. However, when HF power was analyzed only from the linear portion between HF power and R-R interval (ie, assessing the cardiac vagal outflow without saturation disturbing the analysis), HF power significantly increased. In this regard, it has been suggested that the HF index would be more sensitive to training than HF power.9 In contrast, we found that both HF24h and HF index were positively changed during the preseason period, indicating that both variables can be sensitive to training in team sports athletes.

Improvement in vagal-related HRV indices could be advantageous to soccer players. It has been shown that female soccer players with increased parasympathetic HRV indices in response to the first 3 weeks of preseason training are more prone to present increased VO2max after 10 weeks of preparation.39 Furthermore, it has been demonstrated that players showing higher natural logarithm of the square root of the mean of the sum of the squares of differences between adjacent NN intervals (lnRMSSD) are more resistant to suffer homeostatic perturbation of the autonomic nervous system during a training week.27 Ultimately, severe perturbation of the autonomic nervous system regulation due to excessive training loads (and other sources of stress) can lead to overreaching, with reduction in vagal activity.40 In this case, one can speculate that saturated players might be more protected from overreaching than non-saturated players.41

It is important to highlight that none of the players in the current study transitioned from the saturated to linear 24-h R-R interval recordings versus HF24h (ie, HF power). On the other hand, 5 players who had a linear relationship between R-R interval and vagal outflow transitioned to saturated after the preseason. This change, along with the increase in the HF24h and HF index of all players, was probably caused by adequate training loads sustained during the observation period.

It is imperative to note that the physiological mechanism of saturated HF power lies in the lack of increase in respiratory sinus arrhythmia in consonance with increasing cardiac vagal tone.9 The dose-response of the heart to the acetylcholine secreted by the vagal nerve ending may explain the saturation of beat-to-beat HRV.42 The cardiac activity dose-response to acetylcholine has been considered to be linear until its concentration reaches the level at which a further increase in acetylcholine concentration does not produce a change in the HRV response.42 Rapid discharge of the vagal nerve during expiration may produce high acetylcholine secretion to maintain a high parasympathetic effect even during inspiration, which results in the saturation of respiratory sinus arrhythmia and HF power at rest.43 However, saturation of respiratory sinus arrhythmia was not measured during the 24-h free-living ambulatory condition, and we recognize this as an important limitation of the current investigation. In addition, one of the main common limitations of 24-h ambulatory R-R interval recordings is the uncontrolled daily activity. Despite the good reproducibility of 24-h recordings, the individual differences in daily activity and respiration patterns may limit the analysis of ambulatory HRV in comparison to more controlled short-term recordings. Similarly, we did not measure ambulatory breathing frequency. It is well known that this factor has marked effects on HR variability. It is possible that the saturated and linear groups differ in terms of respiratory pattern. Finally, our study design did not allow identifying mechanisms behind the associations found between HRV and endurance performance and their respective changes caused by training.

In conclusion, high-level female soccer players undertaking high training loads during a 4-week preseason period demonstrated increased aerobic fitness as assessed by a field-based intermittent running test, along with increased 24-h cardiac vagal activity. This enhanced vagal modulation led some players to transition from non-saturated to saturated vagal outflow throughout the day. Additionally, changes in 24-h HRV were associated with changes in aerobic fitness, suggesting that monitoring cardiac autonomic can aid in the optimization of training responses.

5 PERSPECTIVE

The 24-h HRV techniques used in the present study that can be easily assessed at the training facilities and/or the players’ own homes, seem to provide specific information about the occurrence of vagal outflow saturation in female soccer players during preseason training. This information might be of interest to practitioners’ intent on monitoring the response to specific training prescriptions, supporting the imp