Acute Physiological Responses to Four Running Sessions Performed at Different Intensity Zones

Abstract This study investigated acute responses and post 24-h recovery to four running sessions performed at different intensity zones by supine heart rate variability, countermovement jump, and a submaximal running test. A total of 24 recreationally endurance-trained male subjects performed 90 min low-intensity (LIT), 30 min moderate-intensity (MOD), 6×3 min high-intensity interval (HIIT) and 10×30 s supramaximal-intensity interval (SMIT) exercises on a treadmill. Heart rate variability decreased acutely after all sessions, and the decrease was greater after MOD compared to LIT and SMIT (p<0.001; p<0.01) and HIIT compared to LIT (p<0.01). Countermovement jump decreased only after LIT (p<0.01) and SMIT (p<0.001), and the relative changes were different compared to MOD (p<0.01) and HIIT (p<0.001). Countermovement jump remained decreased at 24 h after SMIT (p<0.05). Heart rate during the submaximal running test rebounded below the baseline 24 h after all sessions (p<0.05), while the rating of perceived exertion during the running test remained elevated after HIIT (p<0.05) and SMIT (p<0.01). The current results highlight differences in the physiological demands of the running sessions, and distinct recovery patterns of the measured aspects of performance. Based on these results, assessments of performance and recovery from multiple perspectives may provide valuable information for endurance athletes, and help to improve the quality of training monitoring.


Introduction
Endurance training typically consists of various training modes differing in duration and intensity. Traditional intensity zones can be set based on individual ventilatory or lactate thresholds in lowintensity training below the first lactate threshold, moderate-intensity training between first and second lactate thresholds, and high-intensity training between second lactate threshold and maximal oxygen consumption [1]. In addition to endurance intensity zones, supramaximal intensity training above the intensity of VO 2max may improve maximal endurance performance [2,3], and induce similar adaptations in the skeletal muscle oxidative capacity than traditional endurance training [3]. Training intensity has effects on the cardiovascular workload, substrate utilization in energy metabolism, as well as the number and type of motor units recruited during the exercise [1], all of which may influence the type of fatigue induced and responses observed followed by the session.
Fatigue during endurance exercise can be stated as perceived tiredness with concurrent decrements in muscular performance and function [4]. Typically, the body needs to adjust to the growing demand of the activity performed by increasing heart rate [5][6][7], oxygen consumption, [5][6][7] and perceived effort [8] at a given workload. The autonomic nervous system responds to exercise by increasing sympathetic drive and catecholamine secretion [9], while parasympathetic activity diminishes [10]. The origin of the fatigue and time frame to restore the normal function in the neuromuscular system seems to depend on the duration and the intensity of the preceding exercise [11].

Abstr Ac t
This study investigated acute responses and post 24-h recovery to four running sessions performed at different intensity zones by supine heart rate variability, countermovement jump, and a submaximal running test. A total of 24 recreationally endurance-trained male subjects performed 90 min low-intensity (LIT), 30 min moderate-intensity (MOD), 6 × 3 min high-intensity interval (HIIT) and 10 × 30 s supramaximal-intensity interval (SMIT) exercises on a treadmill. Heart rate variability decreased acutely after all sessions, and the decrease was greater after MOD compared to LIT and SMIT (p < 0.001; p < 0.01) and HIIT compared to LIT (p < 0.01). Countermovement jump decreased only after LIT (p < 0.01) and SMIT (p < 0.001), and the relative changes were different compared to MOD (p < 0.01) and HIIT (p < 0.001). Countermovement jump remained decreased at 24 h after SMIT (p < 0.05). Heart rate during the submaximal running test rebounded below the baseline 24 h after all sessions (p < 0.05), while the rating of perceived exertion during the running test remained elevated after HIIT (p < 0.05) and SMIT (p < 0.01). The current results highlight differences in the physiological demands of the running sessions, and distinct recovery patterns of the measured aspects of performance. Based on these results, assessments of performance and recovery from multiple perspectives may provide valuable information for endurance athletes, and help to improve the quality of training monitoring.
In addition to appropriate training load, sufficient recovery is required to induce training adaptations. Resting heart rate variability (HRV) is a noninvasive measurement of the autonomic nervous system function and is suggested to provide comprehensive information about the recovery status [12]. Previous research has shown that the reactivation of the parasympathetic nervous system measured as HRV after training appears to be affected most by the intensity of the session [10,13,14]. Full restoration after exercise at or above the first ventilatory threshold intensity can take up to 24 − 48 h [15]. Furthermore, individually adjusted endurance training based on the fluctuations of resting HRV has provided superior results compared to predefined training [16,17]. A similar approach has also been examined with a heart rate-based submaximal cycling test [18]. The general assumption in submaximal tests is that increase in the power or speed at the same relative heart rate and perceived effort reflects positive training adaptation [19] and readiness to train [18].
Despite the potential of the resting HRV and heart rate based submaximal tests, it is unclear how well these tests reflect all aspects of recovery such as the subjective recovery or readiness of the neuromuscular system. Recently, it has been observed that the recovery timeframe of neuromuscular performance and exerciseinduced muscle damage may differ from that of HRV after strength training exercise [20]. Acute responses following endurance exercise also seem to differ as countermovement jump (CMJ) performance may even improve in endurance-trained athletes after high-intensity sessions [21,22]. Additional monitoring variables may also help to contextualize whether changes in resting HRV or submaximal heart rate are due to fatigue or positive training adaptation, as similar responses may be observed in both situations [12,23].
Responses to training are likely related to the intensity and duration of the preceding session and subsequently, the timeframe to recover may vary depending on the viewpoint taken. In endurance sports, heart rate assessments during rest and exercise have been studied and also utilized in training monitoring widely, while less is known about the other aspects of recovery and resemblance of different markers. The purpose of this study was to compare acute responses and post 24-h recovery in the function of the autonomic nervous system, neuromuscular performance and submaximal running test. It was hypothesized that the acute HRV decrease is related to the intensity of the training, while CMJ performance would improve after moderate and high-intensity training sessions. In addition, it was anticipated that CMJ, metabolic and cardiorespiratory recovery would occur in the 24 h after all sessions, but parasympathetic nervous system activity would only be fully recovered after LIT.

Subjects
Twenty-five recreationally endurance-trained men, ages 20 − 45 years, were recruited for the study. Basic characteristics of the subjects are presented in ▶ table 1. After being informed about the study design and possible risks and benefits of participation, subjects with an appropriate training background and health status signed a written informed consent form. One subject could not fin-ish all the training sessions due to an injury, and therefore the total number of subjects was 24. After low-intensity session and highintensity interval session, one subject did not perform the post-24-h measurements. The study was approved by the Ethical Committee of the University of Jyväskylä, and it was conducted according to the provisions of the Declaration of Helsinki and recommendations of Harriss et al. [24].

Study design
The study compared acute responses to and post 24-h recovery following four different training sessions performed on a treadmill. The order of the training session was randomized by drawing the sequence for each subject. After a preliminary performance testing week, training sessions were performed during the one-month study period. Before (pre), immediately after (post) and 24 h after (post24) each session, supine heart rate variability, countermovement jumps, and a submaximal running test were performed. Additionally, perceived recovery and muscle soreness was measured at pre and post24. Subjects could continue their regular training during the study period. However, on the day before each training session, no exercise was performed and two days before only light exercise was permitted. During the recovery phase before post24 measurements, exercising was not allowed. Subjects were advised to avoid heavy meals and caffeine 3 − 4 h preceding each measurement to avoid any gastrointestinal symptoms or any other possible effects on measured variables. The structure of one training session and the measurements performed are presented in ▶Fig. 1.

Incremental treadmill test
The incremental treadmill test was performed on a treadmill (Telineyhtymä Oy, Kotka, Finland). Starting speed (8.2 ± 1.1 km/h − 1 ) was individually set based on information obtained from the previous performance and training background of the subjects, in order to have at least two stages before the velocity of the first lactate threshold and thus allow a reliable estimation of lactate thresholds. Three-minute stages were used, and speed increased by ▶table 1 Basic characteristics of the subjects (n = 24).

Mean ± sD
Age (yrs.) 35  and running speed at the second lactate threshold (vLT2) were determined based on the change in the inclination of the blood lactate curve during the test [19]. The first lactate threshold was set at 0.3 mmol · l − 1 above the lowest lactate value and the second lactate threshold at the intersection point between 1) a linear model between first lactate threshold and the next lactate point and 2) a linear model for the lactate points with the La increase of at least 0.8 mmol · l − 1 similar to Vesterinen et al. [19] 20-m flying sprint test 20-m flying sprint test was performed on the indoor track. Warmup before the test included a 10-min low-intensity run, dynamic stretching for the lower limbs and three submaximal 50-m accelerations. Maximal running speed (v20m) was measured with two photocell gates after 30-m acceleration. Three attempts were performed with three-minute recovery, if no more than 5 % improvement were found between the last two attempts. The best result was used in further analysis.

Anthropometrics
Fat percentage was analyzed as a sum of four skinfolds [25]. Subjects were weighed before each measurement session and the current body mass was used in VO 2 (ml · kg − 1 · min − 1 ) calculations.

Training sessions
The duration and intensity of the training sessions were pre-determined in order that they represented typical training of each intensity zone (low, moderate, high and supramaximal intensity) and to ensure it would be possible for each subject to perform the ses-sions. Previous studies [2,14,26] have also utilized similar types of training. Running speeds of the sessions were set individually based on their lactate thresholds and maximal running speed during the incremental treadmill test and the 20-m flying sprint test. A lowintensity (LIT) session was a 90-min run performed at 80 % of the speed of the first lactate threshold (vLT1). A moderate-intensity (MOD) session was a 30-min run performed at the average speed of the first and second lactate thresholds ((vLT1 + vLT2)/2). A highintensity interval (HIIT) session was 6 × 3 min with 2-min recovery performed between the second lactate threshold and maximal incremental treadmill test speed (vLT2 + (vMax-vLT2)/3). A supramaximal intensity interval (SMIIT) session was 10 × 30 s with 2.5-min recovery performed at 75 % of the speed from the 20-m flying sprint (v20m · 0.75). During the recovery, treadmill speed was set at 5 km/h in both interval sessions. The submaximal running test acted as a warm-up and cool-down for the sessions. Before SMIT, one short acceleration (15 s) to the speed of the upcoming session was performed to familiarize subjects with the treadmill velocity, and the actual session started after 2.5-min recovery. All training sessions were performed within-subject at the same time of the day ( ± 1 h) on the treadmill (Telineyhtymä Oy, Kotka, Finland). Heart rate was measured throughout the sessions with a Garmin Forerunner 920 XT -monitor (Garmin Ltd, Schaffhausen, Switzerland). Average and peak heart rates as well as training impulse (TRIMP), based on the Edwards [27] model, were analyzed from the training sessions. In addition, at the end of each session rating of perceived exertion (RPE) was asked using the 6 − 20 Borg scale [28], and blood samples were drawn from the fingertip. Blood lactate was analyzed using Biosen S_line Lab + lactate analyzer (EKF Diagnostic, Magdeburg, Germany). After each training session, subjects were given the same recovery drink (Fast Reco2), including 41 g of carbohydrates and 20 g of proteins mixed in a 500 ml of water. The recovery drink was served to ensure similar immediate nutrition for all subjects.

Recovery measurements
Heart rate variability Heart rate variability (HRV) was measured in a supine position using a Garmin Forerunner 920XT monitor. Before starting three-minute data collection [29], a one-minute stabilization period was performed [30]. Subjects were able to breathe at their natural rhythm. The average natural logarithm of the square root of the mean sum of the squared differences (lnRMSSD) was calculated from the three-minute measurement period. Because the measurements were performed in the lab and not right after awakening, baseline values in each athlete were derived from pooled pre-exercise data for the four test sessions comparable to Seiler et al. [14].

Countermovement jump
Countermovement jumps were performed on a contact mat. Jump height (h) was calculated with a formula: h = g · t 2 · 8 − 1 , where t is the recorded flight time in seconds and g is the acceleration due to gravity (9.81 m · s − 2 ) [31]. Subjects were advised to keep their hands on their hips and jump as high as possible. The lowest knee angle for the jump was about 90 degrees. Three attempts with 30 s recovery were performed unless an improvement of 5 % or more was found between two last jumps. The best jump of three was used in further analysis. Subjects were familiarized with the jumping technique during the preliminary tests.

Submaximal running test
The submaximal running test (SRT) was modified from the Vesterinen et al. [19] submaximal running test, and it acted as a warm-up and cool-down for each training session. SRT in the current study consisted of two 5-min stages, which were performed at the speeds corresponding individually to 70 % (1. stage) and 80 % (2. stage) of HR max during the incremental treadmill test. The same individually set speeds, which were calculated from the incremental treadmill test, were used in all measurements, despite possible changes in heart rate, to allow fair comparison between sessions and conditions. During SRT, heart rate (HR) was recorded (Garmin Forerunner 920 XT) and oxygen consumption (VO 2 ) and respiratory exchange ratio (RER) were measured (OxygonPro, Jaeger, Hoechberg, Germany). The average of the last two minutes during the 80 % running speed was used in further analysis, as higher intensities reflect better changes in maximal performance [19]. After SRT, RPE was asked using the 6 − 20 Borg scale [28] and blood lactate values were analyzed from the fingertip sample.

Subjective markers
Perceived recovery was estimated on the 0 − 10 scale [32]. Perceived muscle soreness of the lower limbs was estimated on the 10 cm visual analogy scale where 0 represented no soreness at all and 10 represented the highest possible soreness [33].

Statistical analyses
All values are expressed as mean and standard deviation (SD). Normal distribution of the data was checked with the Shapiro-Wilk test and homogeneity of the variances by Levene's test. A one-way repeated measures ANOVA was used to compare training load variables measured during training sessions. A two-way repeated measures ANOVA was performed to examine main effects (training mode, time) and interaction (training mode × time) across measured variables. When appropriate, a Bonferroni post hoc test was used. Furthermore, in case of significant training mode × time interaction, relative changes from pre-values to post and post 24 were compared between training modes using paired samples t-test with Bonferroni correction. Muscle soreness was not normally distributed even after log-transformation, so Wilcoxon signed-rank test was used for within-group comparisons and Mann-Whitney U-test for between-group comparisons. To further analyze the magnitude of observed changes, the effect size was assessed by Cohen's D (difference of the means divided by the pooled standard deviation) [34], and after nonparametric tests by a formula: ES = Z/√n, where Z is the z-score, and n are the number of observations on which Z is based. An effect size of < 0.20 was considered trivial, ≥ 0.20 small, ≥ 0.50 medium, and ≥ 0.80 large [34]. The statistical significance level was set to p < 0.05. Analysis was performed with IBM SPSS Statistics v. 26

Training sessions
Results of the training sessions performed are presented in ▶table 2.  Values are presented as means ± SD. AvgHR, average heart rate; PeakHR, peak heart rate; TRIMP, training impulse; RPE, rating of perceived exertion; LIT, low-intensity session; MOD, moderate-intensity session; HIIT, high-intensity interval session; SMIT, supramaximal intensity interval session.

Submaximal running test
A significant main effect for time (p < 0.01) was observed in all variables measured during the submaximal running test. In addition, significant main effect for training mode as well as training mode × time interaction was found in blood lactate (p < 0.001). Heart rate during submaximal running test increased after all sessions (p < 0.001) followed by a decrease below the baseline at 24 h after LIT (p = 0.001), MOD (p = 0.023), HIIT (p = 0.016) and SMIT (p = 0.011). RPE during submaximal running test increased after LIT (p < 0.001), MOD (p = 0.004), HIIT (p = 0.001) and SMIT (p = 0.002), and it returned to baseline at 24 h after LIT and MOD but remained increased after HIIT (p = 0.048) and SMIT (p = 0.007). Oxygen consumption during submaximal running test increased after LIT (p = 0.017) and MOD (p = 0.002), while no significant difference was observed after SMIT or HIIT. Oxygen consumption returned to baseline after all sessions at post24 measurements. The only betweengroup difference observed during the submaximal running test was in blood lactate, which was higher after SMIT than any other session (p < 0.001) The absolute values and effect sizes measured during the submaximal running test are presented in ▶table 3.

Discussion
The aim of this study was to compare acute responses and post 24-h recovery after training sessions performed at different intensity zones. The main findings of the study were that parasympathetic reactivation measured as HRV was diminished the most after MOD and HIIT compared to LIT and SMIT. Contradictory, CMJ performance did not decrease after MOD or HIIT but acutely decreased combined with increased muscle soreness at post24 after LIT and SMIT. The main result of the submaximal running test was that all measured metabolic parameters recovered and heart rate decreased significantly at 24 h after all sessions, despite perceived exertion being elevated after HIIT and SMIT at the respective time points. The current results highlight how physiological demands differ between training modes. Different measures of performance and recovery may induce even contradictory results illustrating the usefulness of a broad approach to endurance training monitoring.

Training sessions
Blood lactate, heart rate, and RPE responses to running sessions confirmed that they could be regarded as representative measures of each intensity zone. Peak and average heart rate values observed during MOD and HIIT indicated a high cardiovascular demand dur-ing these sessions. Perceived effort during SMIT was estimated by the subjects similarly as after MOD and HIIT, and despite lower heart rate values, blood lactate increased the most suggesting higher anaerobic contribution during the session. During LIT, more than double the distance was covered compared to MOD and HIIT, while there was almost a fourfold increase in distance covered compared to SMIT. Although the duration or distance was not the same between the sessions, they were likely similar to the ones typically utilized by athletes [1]. The only unexpected response was increased blood lactate value after LIT despite the low relative intensity (52 %/vMax, 80 %/vLT1), low heart rate (avg: 70 %/HR max , peak: 76 %/HR max ) and RPE (12 on a 6 − 20 scale). In the study of Seiler et al. [14], no changes in blood lactate were observed after 1-or 2-h exercises performed below the ventilatory threshold. Differences compared to the present protocol were that Seiler et al. [14] had higher caliber athletes, the 1-h session was performed on the 2 − 5 % incline and the longer 2-h session was performed outside. All of these methodological differences may at least slightly influence the physiological responses compared to the present protocol. In line with the present results, Kaikkonen et al. [26] found that after a 14 km run on a treadmill at 60 % vVO 2max blood lactate elevated significantly compared to the con-
trol session performed at the same intensity but 3 km in distance (1.4 vs. 2.6 mmol/l) in recreational athletes. The athlete level could, therefore, be a major factor in the observed response. Because RPE and heart rate remained in target values despite elevated blood lactate levels during LIT, the training session of the present study likely illustrates a typical LIT session of a recreational endurance athlete.

Acute responses
Acute responses in supine HRV differed significantly between the sessions. HRV decreased less after LIT compared to MOD and HIIT despite higher TRIMP and more than the double distance covered during the session. The results are in line with previous studies [10,14] indicating that intensity of the sessions when performed below vVO 2max seems to influence the parasympathetic reactivation more than the duration of the session. Furthermore, MOD and HIIT produced quite similar acute responses in HRV supporting the theory that lactate or the ventilatory threshold may act as a lower bound for the intensity related delay in parasympathetic reactivation [14]. An unexpected finding was that a smaller decrease in HRV was also observed after SMIT than MOD, and no difference was observed in the HRV responses between LIT and SMIT despite significantly higher lactate values and RPE measured in SMIT. This also contradicts the results of Niewiadomski et al. [13] who found a greater decrease in HRV one hour after a supramaximal session (2 × 30 s Wingate) compared to a moderate intensity session (30 min 85 %/HRmax). In addition, Buchheit et al. [35] have suggested that the delay in parasympathetic reactivation is mainly related to the contribution of anaerobic processes. However, maximum heart rate [13], as well as mean heart rate and blood lactate levels [35] during supramaximal exercises were substantially higher compared to the present study, which may relate to a lower cardiovascular load and the sympathetic nervous system activity during the session. It should be also acknowledged that there is a wide range of anaerobic interval sessions differing in the intensity, work:relief-ratio and anaerobic glycolytic energy contribution [36].
Further studies are needed to understand how manipulation of these variables would affect the parasympathetic reactivation following anaerobic exercise. Acute changes in CMJ occurred oppositely when compared to HRV responses. While no change was observed after HIIT or MOD, jump height decreased after LIT and SMIT. Previously, Boullosa et al. [21] have found improved CMJ performance after intensive running sessions in endurance-trained athletes. Because both the intensity and the work:relief-ratio of the interval exercise significantly affect neuromuscular demands of the session [36], it is somewhat expected that supramaximal intervals may induce different types of response compared to the intervals of lower intensity. Wiewelhove et al. [37] reported decreased CMJ performance after sprint interval training, while aerobic high-intensity training did not induce such an effect. Blood lactate is probably not the main contributor behind this difference as improved CMJ performance has been observed in the absence of higher blood lactate values [21] compared to the present study. In general, it is thought that neuromuscular fatigue after high-intensity exercise is mainly peripheral and caused by contractile mechanisms disturbances [11,38], while fatigue induced by longer duration activities are mainly of central origin observed as decreased voluntary activation [11] or changes in stretch-reflex sensitivity and muscle stiffness [39]. It is plausible that different mechanisms are behind the CMJ decrease observed after LIT and SMIT, and the time needed to recover is longer after SMIT, at least in recreationally trained athletes.
All cardiorespiratory, metabolic and perceptual measures during the submaximal running test were quite similar between the sessions, and the responses were mainly in line with previous studies using similar types of running protocols [5][6][7]. Heart rate and RPE during submaximal running test increased after all sessions, with a concurrent decrease in RER indicating higher reliance on fat as a substrate despite the nature of the preceding exercise. Oxygen consumption during submaximal running test increased slightly, but still significantly only after continuous sessions. This was somewhat surprising and in contrast to the results reported by previous studies [5,7]. However, effect sizes in acute responses remained trivial after all sessions, so any major difference between session types cannot be stated. Blood lactate during the submaximal running test remained elevated after SMIT, which was probably mainly the outcome of a higher absolute value after the exercise itself. A longer recovery period seems to be necessary for lactate clearance after such a session. Increases in body core temperature and sympathetic nervous system activity along with dehydration and a decrease in blood volume are likely the main contributors to the cardiovascular responses observed in the present study [40]. Increased heart rate may also contribute to the impaired running economy [41]. Muscle glycogen content has been shown to decrease after prolonged as well as high-intensity exercises [42], and its depletion further influences substrate utilization [43] and oxygen cost during running [44]. It is plausible that running sessions of the present study did not induce significant differences in the aforementioned factors. The lack of major differences between the sessions in the submaximal running test could also be related to the time frame of the measurement, as some session-related effects may have already disappeared at the time point used in the current study (post 18 − 20 min).

24-h recovery
Supine HRV returned to baseline in 24 h after all sessions. Stanley et al. [15] concluded in their review, that cardiac autonomic recovery after strenuous exercise may take up to 24-48 h. The lack of any differences within or between the sessions at post24 could possibly be related to methodological differences between nocturnal and daytime recordings. It is well known that HRV is affected by many external factors especially during the daytime [12], which makes it challenging to find significant changes in the autonomic modulation caused by a single training session. While nocturnal HRV has remained suppressed after a moderate and heavy endurance exercise [45], Niewiadomski et al. [13] measured HRV in the laboratory conditions, and found no change compared to baseline 24 h after moderate or high-intensity training sessions.
Recently, it has been observed that the recovery of the neuromuscular performance and markers of muscle damage follow different patterns than HRV after strength training exercise [20]. Similar observations were found in the present study, as HRV was fully recovered at 24 h after all sessions, but CMJ performance after SMIT remained decreased and muscle soreness was apparent after LIT and SMIT. Although it should be acknowledged that effect size of the pre-post24 change in CMJ was trivial after all sessions, it is likely that cardiac parasympathetic reactivation after exercise does not reflect all aspects crucial to recovery in endurance training, such as repletion of the energy stores and neuromuscular performance [12,15].
All physiological variables measured during the submaximal running test recovered at least to baseline levels at 24 h after all exercises with no significant differences between the sessions. However, increased perceived exertion during the submaximal running test was still apparent after both interval exercises and muscle soreness were increased after LIT and SMIT. All these changes took place despite the significantly decreased heart rate in the submaximal running test. In addition to increased muscle soreness and perceived exertion during the submaximal running test, also CMJ remained decreased at 24 h after SMIT, which emphasizes the high neuromuscular demand of these types of sessions, which is also supported by Wiewelhove et al. [37]. It is possible that in addition to peripheral factors [11], high mechanical load of the running speed that recreational endurance athletes may be unaccustomed to further amplified the neuromuscular fatigue and time needed to recover after SMIT.
Similar to resting HRV, heart rate recordings during exercise could also be affected by multiple external factors, and a natural day-to-day variation during submaximal exercise can be up to 3 − 8 bpm [46]. Taking this into account, it was interesting that the heart rate during submaximal running test decreased significantly at 24 h after all sessions by 3 − 4 bpm. Previously, submaximal running tests have mainly been studied after intensive training or the competition period, while effects of a single session have been less examined. Siegl et al. [47] found heart rate decrements of 3 bpm (70 %/vVO 2max ) and 2 bpm (85 %/vVO 2max ) with a concurrent increase of RPE two days after an ultramarathon event. Other studies have failed to show any differences in submaximal heart rate 1 − 4 days after a 30-min high-intensity run [48] or 1 − 3 days after a 26 km run at an intensity of 80 %/HR max [49]. One possible reason for heart rate decrease during submaximal exercise could relate to an increase in plasma volume [12], but this variable was not measured in the present study and the actual reason remains inconclusive.
Running economy is an important determinant of endurance performance [41], and a recovery pattern of the economy would therefore be an important aspect to examine. In the present study, oxygen consumption as well as RER during the submaximal running test returned to the baseline level at 24 h after all sessions. This was not surprising, though already acute responses in oxygen consumption could be stated as trivial after all sessions. In previous studies, running economy has recovered 24 h after a 30-min high-intensity run [48], 26 km run at the intensity of 80 %/HR max [49] as well as 48 h after a marathon [50]. Despite exercise-induced fatigue potentially declining running economy via multiple mechanisms [41], it seems that metabolic recovery occurs quite rapidly, and running economy will not likely be impaired in the following day of a single intensive or prolonged low-intensity training session.
Despite clear relationships between maximal and submaximal endurance performance [19] results of the submaximal tests are sometimes complicated to interpret. Similar changes in submaximal heart rate can be a sign of a positive training adaptation [19] or negative adaptation indicating fatigue/overreaching [51]. Similarly, it may be difficult to make assertions related to blood lactate values during submaximal exercise [51]. In the present study, RPE during the submaximal running test remained elevated after both interval sessions despite the decreased heart rate and the baseline level blood lactate values. Because maximal performance was not measured in the present study, it is difficult to ascertain whether the changes in perceptual markers would be a sign of decreased maximal performance. However, Marcora and Bosio [52] found that exercise-induced muscle damage and muscle soreness may impair 30-min maximal time-trial running performance. The authors suggested that the performance impairing effect might be mediated by the increased sense of effort during the time trial caused by muscle soreness [52]. Therefore, it seems reasonable to assume that higher perceptual effort at the submaximal level and increased muscle soreness may also have influenced maximal performance.

Study limitations
The study population consisted of recreationally endurance-trained men, and the results cannot be straightforwardly transferred to well-trained or elite-athletes. Although the physiological characteristics of each intensity zone are likely quite universal, more studies are needed to confirm the findings among high-level athlete populations and in both sexes. Because the follow-up measurements were not performed later than 24 h after each session, all variables did not reach the baseline in all subjects. It could not, therefore, be concluded, what would have been the time frame to recover for each of the variables. Training sessions of the present study were not matched for training load but were instead chosen as one representative session type of each intensity zone. Changing session intensity or duration would possibly influence the results, so further studies are needed to grow the understanding of how manipulation of these variables would affect responses and recovery. Lastly, although trying to standardize the testing protocol and days surrounding it, there remain some aspects that may potentially influence recovery within and between individuals such as nutrition, sleep or leisure-time activity.

Conclusions
In conclusion, the results of the present study highlight differences in the physiological demands of the running sessions performed and distinct recovery patterns following these sessions in the measured variables of performance and training state. The delay of the parasympathetic reactivation after endurance exercise seems to relate to the intensity and cardiovascular load of the preceding session. Running sessions of a long-duration or a supramaximal intensity have high neuromuscular demands, observed as acute decreases in neuromuscular performance, and increased muscle soreness 24 h afterwards. Cardiovascular and metabolic recovery occurs rapidly, and these components of physical performance are not likely compromised 24 h after a single intensive or a low-intensity prolonged session. Because subjective markers may give even contradictory results when compared to the objective measurements, it would be recommended to combine information from different sources, when estimating the actual recovery state and readiness to train.