The influence of type of recovery on subsequent performances has not received consensus in the literature. Several authors (1,4,8,20,24) have reported better performances after active recovery compared with passive recovery. For example, when two Wingate tests (WAnT) were separated by 4 min of recovery, Spierer et al. (24) found that the performance of the second WAnT was better after an active recovery performed at 28% of maximal oxygen uptake (V˙O2max). The positive effects associated with active recovery involve to a greater contribution of aerobic metabolism to energy yield during the subsequent high-intensity exercises (8) and to a faster intracellular pH recovery (22). It also has been suggested that active recovery would enable the blood flow to be higher from the previously exercised muscle, which would increase oxygen delivery and, therefore, phosphorylcreatine (PCr) resynthesis (4,8).
However, active recovery does not always improve performance when compared with passive recovery (5,15,16) and several authors found that performance was improved when passive recovery was introduced between high-intensity exercise rather than active recovery (9,10,25). For example, when short intermittent exercises (15 s) were alternated with short recovery periods (15 s), time to exhaustion (TTE) was significantly longer when passive recovery was introduced rather than active recovery at 50% of maximal aerobic speed (9) or at 40% of V˙O2max (10). The positive effects of passive recovery on performance were attributed to a slower decline in the oxyhemoglobin (10), suggesting that passive recovery allows a higher muscular reoxygenation than active recovery and, therefore, a higher PCr resynthesis. However, in these studies, exercise intensity was set at a predetermined load: 120% of maximal aerobic speed (9) and 125% of maximal aerobic power (10). During repeated sprint tests (six 4-s sprints, every 25 s), Spencer et al. (23) found that active recovery resulted in a significantly lower peak power output for the final sprint, a greater peak power decrement, but no significant difference in work decrement or total work. These results were associated with a strong trend (P = 0.06) for lower PCr concentrations immediately after active recovery compared with passive recovery.
The effects of recovery type on performance may be linked to exercise intensity (sprints vs predetermined intensity), performance criteria (peak power and mean power vs TTE), recovery duration, and the intensity of active recovery. In previous studies where performance was improved after passive recovery, recovery duration was short (15-120 s) and intensity was relatively high (9,10,25). According to Dorado et al. (8), when the intensity of the active recovery is relatively high (i.e., 50% of maximal aerobic speed), active recovery would fail to show beneficial effects on subsequent performance. Consequently, it would be interesting to compare the effects of several recovery intensities on performance when two all-out sprints are alternated with a short recovery period (15 s). Indeed, all-out sprints induce greater PCr depletions and higher deoxygenations than exercises at predetermined intensity. To our knowledge, no study has attempted to investigate the effects of short recovery intensities on all-out performance and deoxyhemoglobin variations. The purpose of this study was to investigate the effects of recovery intensities on WAnT performance and on deoxyhemoglobin variations. We hypothesized that when two WAnT were interspersed with a 15-s recovery period, the performance of the second WAnT would be higher after the passive recovery compared with the active one.
Subjects were 12 males who specialized in soccer, trained from three to five times per week, and performed one match per week. Age, height, and body mass of the participants were 22.8 ± 4.5 yr (range: 20-32 yr), 178.2 ± 5.4 cm (range: 172-187 cm), and 71.3 ± 5.5 kg (range: 65.7-81.8 kg), respectively. Their percent body fat, estimated from a calibrated bioelectrical impedance balance (Tanita TBF 543, Tokyo, Japan), was 13.0 ± 4.0% (range: 9.2-19.4%). All subjects were fully informed of the protocol, benefits, and risks before giving their written informed consent. This investigation was made with the agreement of the local ethics committee in biomedical research.
All subjects performed four tests on a cycle ergometer (Monark 894E, Varberg, Sweden): a graded test and then three sessions of repeated WAnT with different recovery natures, randomly performed. Individual tests were carried out at the same time of day, separated by at least 48 h of rest and completed within 1 month. Subjects were required to have their last meal at least 3 h before the test, to not drink beverages containing alcohol or caffeine, and to consume the same meals and the same water quantity before each test. These recommendations were checked by questionnaire before each test. All subjects were fully familiarized with laboratory testing, having already performed similar tests: graded tests and sprints exercises on a cycle ergometer with the same gas-exchange analyzer. The subjects were instructed to grip the handlebars and to not lift off the saddle while cycling; seat and handlebar height were kept constant during the sessions for each participant. The subject's feet were firmly strapped to the pedals, and the saddle height was adjusted to allow slight flexion of the knee at the lowest level of the pedal cycle. Strong verbal encouragements were provided to all subjects during each test.
The graded test aimed to determine V˙O2max and maximal aerobic power (MAP). During this test, the subjects were asked to maintain a cycling cadence of 60 rpm. Subjects performed a 2-min warm-up at 60 W, and then the power output was increased by 30 W every minute until exhaustion. For this test, the subjects were asked to exercise for as long as possible. They were considered to be exhausted when they could not maintain the required frequency despite vigorous encouragement. The power at the last completed stage was recorded as the MAP.
Before beginning the repeated sprints, each subject completed a standardized warm-up of 8 min of pedaling at 40% of MAP, followed by 1 min of rest before two sets of all-out sprints. These two sets of all-out sprints of 4-s duration were performed with the same braking force as during the warm-up of 8 min at 40% of MAP, and they were interspersed with 30 s of rest. They required the subject to pedal maximally. A 4-min rest period followed this warm-up. The repeated sprints included two WAnT: a 15-s Wingate (WAnT15) and a 30-s Wingate (WAnT30), separated by a 15-s recovery period of a different nature, randomly assigned. The 15- and 30-s Wingate tests consisted of pedaling as fast as possible against a braking force corresponding to 75 g·kg−1 of the subject's body mass (2). The 15-s recovery was either active at 20 or 40% of MAP, or passive. These recovery intensities were chosen because authors have often used these intensities to compare active versus passive recovery (4,8). In these studies, performance was higher after active recovery than passive recovery. Six magnetic interrupters, placed on the wheel every 60°, were activated by a magnet attached to the cycle ergometer. The signal thus generated was transmitted to a computer to calculate cranking velocity. Power was computed using the Monark Anaerobic Test software (Monark, Varberg, Sweden), which accounted for both the load on the flywheel and crank kinematics. Peak power was defined as the highest power achieved at any given 5-s period, and mean power was defined as the average power throughout the test.
During the graded test only, respiratory gas-exchange values were measured breath-by-breath using an automated gas-analysis system (Oxycon Pro, Jaeger, Hoechberg, Germany) to determine ventilation (VE), oxygen uptake (V˙O2), and carbon dioxide production (V˙CO2). Respiratory gas-exchange and heart rate (HR; Polar Electro, Kempele, Finland) values were averaged every 15 s. Before each test, the O2- and CO2-analysis systems were calibrated using ambient air and with a gas mix of known O2 and CO2 concentrations (16% O2 and 5% CO2). The volume transducer was calibrated before each test with a 3-L calibration syringe (Jaeger, Hoechberg, Germany). The V˙O2max corresponded to the highest mean V˙O2 attained in two successive 15-s periods. It was judged that subjects had reached their V˙O2max when three or more of the following criteria were met: 1) a plateau in V˙O2 despite increasing power; 2) a final respiratory exchange ratio (RER) higher than 1.1; 3) an inability to maintain the required power; and 4) a lactate concentration higher than 9 mM.
During the repeated sprints, changes in tissue oxyhemoglobin (O2Hb) and deoxyhemoglobin (HHb) were assessed using a near-infrared spectroscopy device (NIRS; Oxymon, Artinis Medical System, Nijmegen, The Netherlands). The procedure used to collect data was the same as those used by Van Beekvelt et al. (26) with a similar device. The unit consists of a continuous-wave near-infrared spectrophotometer that generates a light source using three wavelengths of 770, 850, and 905 nm. NIRS is based on the relative tissue transparency for light in the near-infrared region and on the oxygen-dependent absorption changes of hemoglobin (Hb) and myoglobin (Mb). The absorption changes at the discrete wavelength are converted into concentration changes of O2Hb and HHb using a modified Lambert-Beer law, in which a path-length factor is incorporated to correct for scattering of photons in the tissue (17). A fixed differential path-length factor of 4.0 was used for the calculation of absolute concentration changes (11). NIRS measurements were made on the right vastus lateralis muscle, approximately 14-20 cm from the knee joint along the vertical axis of the thigh, with an interoptode distance of 40 mm. To prevent variations in placement of the optodes and to avoid operator errors, the angle and place of the optodes were kept constant during the sprint tests, using a special support that was attached to the skin with adhesive stickers. Waterproof markers were used to identify the location on the skin where the probe was placed, to put it on the same area during the subsequent exercise. Three consecutive skinfold-thickness measurements were performed, with the NIRS optodes placed using a Harpenden skinfold caliper. Adipose tissue thickness (fat + skin layer) corresponded to the half of the mean of the three consecutive measurements. Data were sampled at 10 Hz, displayed in real time, and stored for offline analysis. In the present study, only ΔHHb values are presented. Indeed, the interpretation of the ΔO2Hb signal is complicated by its dependence on changes in perfusion of the field of NIR interrogation, whereas the ΔHHb signal is dependent on changes in O2-extraction values and normally is essentially unaffected by perfusion or by changes in arterial Hb volume (12). For the three recovery conditions, the signals at rest values, at the end of the WAnT15, at the end of the recovery period, and at the end of the WAnT30 were analyzed. Rest values were recorded before the warm-up for 5 min in a standard position: subjects were seated on the cycle ergometer with their leg in a relaxed position at the lowest point of the pedal. Rest values corresponded to the averaged value of the last 2 min. The end values corresponded to the averaged values of the last 2 s. ΔHHb values were calculated from the rest values, the end values for the WAnT15, and the end values for the WAnT30. ΔHHbWAnT15 corresponded to the differences between resting values and end values for the WAnT15, ΔHHbrec corresponded to the differences between end values for the WAnT15 and end values for the recovery, and ΔHHbWAnT30 corresponded to the differences between resting values and end values for the WAnT30.
Before beginning the graded test, a venous catheter was inserted into an antecubital vein. Venous blood samples were taken at rest, at the completion of the test, and at 2 and 5 min of recovery. Blood samples were collected in heparinized syringes and were analyzed by Boehringer's bienzyme method (21) for lactate concentrations ([La]b). Peak [La]b values corresponded to the highest value collected during recovery period. Hemoglobin concentration ([Hb]) was measured at rest using a hemoximeter (Radiometer model OSM3, Copenhagen, Denmark). Hematocrit (Hct) determination was made at rest using heparinized microcapillary tubes (100 μL) centrifuged for 15 min at 15,000g (IEC MB microhematocrit centrifuge).
Results are expressed as means ± standard deviations (mean ± SD). The normality distribution of the data was checked with the Kolmogorov-Smirnov test completed by the Lilliefors method. A repeated-measures analysis of variance (ANOVA) was used to determine the effect of recovery condition on peak power and mean power of the two sprints, and on HHb, tHb, and calculated Δ values. Where appropriate, post hoc comparisons were made with the Student-Newman-Keuls test. The level of significance was set at P < 0.05.
Before beginning the graded test, resting blood values for [La]b, [Hb], and Hct were 1.5 ± 0.6 mM, 14.4 ± 1.2 g·100 mL−1, and 44.3 ± 3.33%, respectively. The MAP corresponded to 280.0 ± 32.3 W. The V˙O2max, HRmax, RER, and VEmax obtained for the graded test were 58.7 ± 6.0 mL·kg−1·min−1, 188 ± 11 bpm, 1.22 ± 0.03, and 151.2 ± 17.9 L·min−1, respectively. The peak [La]b value obtained within 5 min after the end of the graded tests was 14.5 ± 2.8 mM.
Performance on the two repeated sprints.
Table 1 presents the values of peak powers and mean powers from the two Wingate tests: the WAnT15 performed before the three recovery conditions, and the WAnT30 performed after the three recovery conditions. Although the peak power values were significantly higher (P < 0.05) after passive recovery (about +11.5% higher than active recovery at 20% of MAP, and about +16.9% higher than active recovery at 40% of MAP) and the mean powers of the WAnT30 (about +6.9% higher than active recovery at 20% of MAP, and about +5.1% higher than active recovery at 40% of MAP), no significant difference was found between the two active recovery conditions.
NIRS data during the two repeated sprints.
Adipose tissue thickness on the right vastus lateralis muscle was 3.6 ± 1.1 mm. Figure 1 presents an example of the HHb kinetics-time relationship during repeated sprints for one subject and under the three recovery conditions: passive, active recovery at 20% of MAP, and active recovery at 40% of MAP. In Table 2, the HHb values are presented at rest, at the end of the WAnT15, at the end of the recovery period, and at the end of the WAnT30. HHb values at the end of passive recovery (3.5 ± 1.7 μM) were significantly lower (more than 3.5 times the active recovery values; P < 0.05) than HHb values at the end of active recovery at 20% (12.4 ± 5.3 μM) and at 40% of MAP (12.9 ± 4.7 μM). HHb values at the end of the WAnT15 were not significantly different from HHb values at the end of the WAnT30. In Table 2, ΔHHb values are also presented values according to the three recovery conditions. ΔHHb values at the end of passive recovery (12.8 ± 5.3 μM) were significantly higher (about three times the active recovery values; P < 0.05) than ΔHHb values at the end of active recovery at 20% (4.3 ± 2.6 μM) and at 40% of MAP (3.9 ± 2.6 μM). Similarly, ΔHHb values at the end of the WAnT15 were not significantly different from ΔHHb values at the end of the WAnT30. For all parameters studied, no significant results were found between the two active recovery conditions.
The purpose of this study was to assess the effects of recovery intensities on WAnT performances and on the deoxyhemoglobin variations. It was hypothesized that when two WAnT were interspersed with a 15-s recovery period, the performance of the second test would be higher after passive recovery compared with active recovery, and that the deoxyhemoglobin variations would be higher during passive recovery in comparison with active recovery. The present results confirm these hypotheses: the mean power and peak power values of the second WAnT were significantly higher when the test was followed by passive recovery rather than active recovery performed at 20 and 40% of MAP; and ΔHHbREC values were significantly higher during passive recovery than during active recovery at 20% and 40% of MAP. Conversely, neither performance nor muscle deoxygenation were influenced by the intensity of the active recovery: peak power, mean power of the WAnT30, HHb during recovery, and ΔHHbREC values were not significantly different between the two active recovery conditions (20 vs 40% of MAP).
In most of the studies investigating the effects of recovery intensities, subjects performed two WAnT of similar 30-s duration (4,24). In the present study, the WAnT15 aimed at inducing a prior fatigue to analyze the effects of different types of recovery on a subsequent exercise performance. The choice of a first sprint of only 15-s duration was based on preliminary experiments having demonstrated subjects' inability to repeat two WAnT of 30 s when intercepted by only 15 s of recovery. Values measured during this prior WAnT15 (peak power, mean power) and at its end (HHb and tHb) were not significantly different between the three recovery conditions, showing that this prior all-out test had been equally performed. The choice of the second WAnT30 was intended to analyze the effects of recovery type on anaerobic capacity and to then compare our results with those obtained by other authors (4,24). Peak power values of the WAnT15 obtained in the present study (1257 ± 139 W before passive recovery, 1234 ± 160 W before active recovery at 20% of MAP, and 1242 ± 155 W before active recovery at 40% of MAP) were close to those obtained by Bogdanis et al. (4) during the first WAnT of 30 s (1291 W before passive recovery and 1297 W before active recovery). However, mean power values of the second WAnT30 were lower in the present study (517 ± 26 W after passive recovery, 484 ± 30 W after active recovery at 20% of MAP, and 492 ± 35 W after active recovery at 40% of MAP) than those obtained by Bogdanis et al. (4) (589 W after passive recovery and 603 W after active recovery) or by Spierer et al. (24) (563 ± 26 W after passive recovery and 589 ± 22 W after active recovery). These results are probably linked to the recovery duration, which was shorter in the present study than in the other studies (4,24). Bogdanis et al. (4) found that 4 min of rest was not enough for complete restoration of power output: peak power during the second WAnT30 represented 87 and 90% of that measured before passive recovery and active recovery, respectively. In the present study, peak power during the second WAnT30 represented 86.4% of the peak power during the first WAnT15 after passive recovery, 78.8% after active recovery at 20% of MAP, and 74.8% after active recovery at 40% of MAP. These results suggest that complete restoration of peak power requires a duration longer than 4 min, but partial restoration is performed within 15 s of passive recovery.
Passive recovery was associated with significantly higher deoxygenation variations (characterized by ΔHHbREC; 12.8 ± 5.3 μM, P < 0.05) compared with the two active recovery conditions (4.3 ± 2.6 μM at 20% of MAP and 3.9 ± 2.6 μM at 40% of MAP; Table 2) and significantly higher performance during the subsequent exercise, characterized by peak power of the WAnT30 (1085 ± 153 W, compared with 973 ± 112 W for active recovery 20% of MAP and 928 ± 116 W for active recovery at 40% of MAP; P < 0.05) and mean power of the WAnT30 (517 ± 26 W, compared with 484 ± 30 W for active recovery at 20% of MAP and 492 ± 35 W for active recovery at 40% of MAP; P < 0.05; Table 1). From this result, it seems that a short, passive recovery after an all-out sprint leads to a higher muscular reoxygenation than during active recovery. Despite the exercise nature proposed (WAnT) and the different active recovery intensities, these results confirm those obtained in previous studies (9,10) for intermittent exercise performed at intensities of 120% of maximal aerobic speed and 125% of MAP, where passive recovery was found to be more appropriate to restore the performance level when the recovery period was short (15 s). These results are also in agreement with those obtained by Toubekis et al. (25), who reported, in swimming, significantly higher performance in a set of eight 25-m sprints when resting periods (45 and 120 s) rather than active recovery periods were introduced between the sprints. In these latter studies and the present study showing that passive recovery is more appropriate than active recovery to restore performance level, the recovery duration ranged from 15 s (9,10) to 120 s (25). Interestingly, when the recovery was longer than 120 s, performance on the subsequent exercise improved after active recovery compared with passive recovery (1,4,8,20,24). In these studies, the recovery durations between the high-intensity exercises ranged between 4 min (4,24) and 20 min (20), which would suggest that the recovery duration is a decisive factor in the choice of recovery mode and intensity.
In the study conducted by Dupont et al. (10), the beneficial effects of passive recovery were associated with a higher muscular reoxygenation. In the present study, the 15-s passive recovery was also characterized by a higher muscular reoxygenation when compared with the active recoveries, whatever the intensities (20 or 40% of MAP). The NIRS was used to analyze changes in tissue HHb and tHb according to the recovery conditions. The measurement depth of the NIRS corresponds to half the distance between the two probes: light in and light out. For an interoptode distance of 40 mm, the light penetrates the tissue approximately 20 mm. In the present study, adipose tissue thickness was 3.6 ± 1.1 mm (range: 1.6-6.5 mm), meaning that the amount of light penetrating the muscle tissue represents 82% of the measurement depth (from 92% for the subject with the least adipose tissue thickness to 67.5% for the subject with the largest adipose tissue thickness). From this result, it seems that the majority of the signal penetrated the muscle tissue for the measurement of muscle oxygenation. During the first WAnT15, deoxygenation, characterized by HHb changes, was similar in the three recovery conditions: HHb increased until the end of the sprint. After this first WAnT15, HHb decreased (Fig. 1). However, passive recovery (3.5 ± 1.7 μM) was characterized by lower HHb values (P < 0.05) compared with the recoveries performed at 20% (12.4 ± 5.3 μM) and 40% of MAP (12.9 ± 4.7 μM). This result suggests that muscular oxygenation was higher during the passive recovery. As phosphorylcreatine (PCr) resynthesis depends on oxygen availability (13,14), it could be suggested that short, passive recovery contributes to a faster PCr resynthesis than active recovery. This hypothesis has been confirmed in studies showing an impairment of PCr resynthesis with short, active recovery (30,18,23). The reasons why short active recovery impairs muscular oxygenation and PCr resynthesis could be linked to blood flow. Van Beekvelt et al. (27) studied forearm blood flow and muscle V˙O2 during moderate and heavy dynamic exercises. They found that, in the case of moderate exercise, forearm blood flow and muscle V˙O2 were not significantly different between the end-exercise values and those observed in the first sample in passive recovery, whereas in the case of heavy exercise, forearm blood flow and muscle V˙O2 increased significantly above the end-exercise values. Consequently, after the first WAnT15 performed in the present study, it could be assumed that hyperemia and muscle V˙O2 were higher during passive recovery conditions than when light exercise was performed (active recovery conditions). Muscular contractions would impair the hyperemia and explain why performance of subsequent exercise and ΔHHbREC values were not significantly different between the two intensities of active recovery.
An additional explanation for the better performance when a passive recovery was imposed could be linked to muscle temperature. Several studies have reported a beneficial effect on performance when the muscle was warmer (3,7,19). Mohr et al. (19) studied muscle temperature and sprint performance during a soccer match. They report that muscle temperature decreased markedly during the half-time period when players recovered passively; when moderate exercise was proposed, muscle temperature remained unchanged (as before and at the end of the first half). Furthermore, these authors found that the decrease in muscle temperature was significantly correlated to the decrease in sprint performance. During a short, passive recovery period, such as the one used in the present study (15 s), muscle temperature probably has no time to drop. Conversely, for passive recovery that is longer in duration, muscle temperature decreases until rest values (19).
Positive effects of active recovery have often been attributed to a faster lactate removal (1). Although lactate removal is faster during a long, active recovery period (6), it cannot be the criterion used to test the quality of recovery. Faster lactate removal does not necessarily involve better performance during the subsequent exercises. In several studies aimed at comparing active and passive recoveries, exercise performance after active recovery did not improve, despite lower lactate concentrations (5,28,29). In other studies, performance after active recovery did improve, but the lactate concentrations determined after the two recovery types (active vs passive) were not significantly different (4,8). Consequently, lactate removal does not seem to be a decisive criterion in performance restoration.
In summary, the findings of the present study demonstrate that when two Wingate tests are performed with 15 s of recovery in between, the performance of the second test is higher when the recovery is passive rather than active. Passive recovery is also associated with higher ΔHHbREC values compared with active recovery at 20 and 40% of MAP. Further studies should investigate the extent to which the type of recovery would affect muscular temperature, blood flow, and PCr resynthesis. A threshold in the duration of the recovery, from which active recovery would be more appropriate to restore the level of performance, could be detected.
The authors gratefully acknowledge the subjects for their cooperation, Willy Colier for the advice and service with the NIRS, Jeanne Dekerle for the translation assistance, and Sophie Robin, Fabrice Vandendorpe, and Anne-Marie Leclercq for medical and technical assistance.
1. Ahmaidi, S., P. Granier, Z. Taoutaou, J. Mercier, H. Dubouchaud, and C. Prefaut. Effects of active recovery on plasma lactate and anaerobic power following repeated intensive exercise. Med. Sci. Sports Exerc.
2. Bar-Or, O. The Wingate anaerobic test. An update on methodology, reliability and validity. Sports Med.
3. Bergh, U., and B. Ekblom. Influence of muscle temperature on maximal muscle strength and power output in human skeletal muscles. Acta Physiol. Scand.
4. Bogdanis, G. C., M. E. Nevill, H. K. A. Lakomy, C. M. Graham, and G. Louis. Effects of active recovery on power output during repeated maximal sprint cycling. Eur. J. Appl. Physiol.
74: 461-469, 1996.
5. Bond, V., R. G. Adams, R. J. Tearney, K. Gresham, and W. Ruff. Effects of active and passive recovery on lactate removal and subsequent isokinetic muscle function. J. Sports Med. Phys. Fitness
6. Bonen, A., and A. N. Belcastro. Comparison of self-selected recovery methods on lactic-acid removal rates. Med. Sci. Sports Exerc.
7. Crowley, G. C., A. Garg, M. S. Lohn, N. Van Someren, and A. J. Wade. Effects of cooling the legs on performance in a standard Wingate anaerobic power test. Br. J. Sports Med.
8. Dorado, C., J. Sanchis-Moysi, and J. A. Calbet. Effects of recovery mode on performance, O2
uptake, and O2
deficit during high-intensity intermittent exercise. Can. J. Appl. Physiol.
9. Dupont, G., N. Blondel, and S. Berthoin. Performance for short intermittent runs: active recovery vs. passive recovery. Eur. J. Appl. Physiol.
10. Dupont, G., W. Moalla, C. Guinhouya, S. Ahmaidi, and S. Berthoin. Passive versus active recovery during high-intensity intermittent exercises. Med. Sci. Sports Exerc.
11. Ferrari, M., Q. Wei, L. Carraresi, R. A. De Blasi, and G. Zaccanti. Time-resolved spectroscopy of the human forearm. J. Photochem. Photobiol. B
12. Grassi, B., S. Pogliaghi, S. Rampichini, et al. Muscle oxygenation and pulmonary gas exchange kinetics during cycling exercise on-transitions in humans. J. Appl. Physiol.
13. Harris, R. C., R. H. T. Edwards, E. Hultman, L. O. Nordesjo, B. Nylind, and K. Sahlin. The time course of phosphorylcreatine resynthesis during recovery of the quadriceps muscle in man. Pflugers Arch.
14. Haseler, L. J., M. C. Hogan, and R. C. Richardson. Skeletal muscle phosphocreatine recovery in exercise-trained humans is dependent on O2
availability. J. Appl. Physiol.
15. Katch, V. L., T. Gilliam, and A. Weltman. Active vs. passive recovery from short-term supramaximal exercise. Res. Q.
16. Lau, S., K. Berg, R. W. Latin, and J. Noble. Comparison of active and passive recovery of blood lactate and subsequent performance of repeated work bouts in ice hockey players. J. Strength Cond. Res.
17. Livera, L. N., S. A. Spencer, M. S. Thorniley, Y. A. Wickramasinghe, and P. Rolfe. Effects of hypoxaemia and bradycardia on neonatal cerebral haemodynamics. Arch. Dis. Child.
18. McAinch, A. J., M. A. Febbraio, J. M. Parkin, et al. Effect of active versus passive recovery on metabolism and performance during subsequent exercise. Int. J. Sport Nutr. Exerc. Metab.
19. Mohr, M., P. Krustrup, L. Nybo, J. J. Nielsen, and J. Bangsbo. Muscle temperature and sprint performance during soccer matches-beneficial effect of re-warm-up at half-time. Scand. J. Med. Sci. Sports
20. Monedero, J., and B. Donne. Effect of recovery interventions on lactate removal and subsequent performance. Int. J. Sports Med.
21. NOLL, A. Enzymatic method analyses, 3rd Ed. In: H. U. Bergmeyer (Ed.). Verlag Chemistry Weinheim, 1974, p. 1521.
22. Sairyo, K., K. Iwanaga, N. Yoshida, et al. Effects of active recovery under a decreasing work load following intense muscular exercise on intramuscular energy metabolism. Int. J. Sports Med.
23. Spencer, M., D. Bishop, B. Dawson, C. Goodman, and R. Duffield. Metabolism and performance in repeated cycle sprints:active versus passive recovery. Med. Sci. Sports Exerc.
24. Spierer, D. K., R. Goldsmith, D. A. Baran, K. Hryniewicz, and S.D. Katz. Effects of active vs. passive recovery on work performed during serial supramaximal exercise tests. Int. J. Sports Med.
25. Toubekis, A. G., H. T. Douda, and S. P. Tokmakidis. Influence of different rest intervals during active or passive recovery on repeated sprint swimming performance. Eur. J. Appl. Physiol.
26. Van Beekvelt, M. C., W. N. Colier, R. A. Wevers, and B. G. Van Engelen. Performance of near-infrared spectroscopy in measuring local O2
consumption and blood flow in skeletal muscle. J. Appl. Physiol.
27. Van Beekvelt, M. C., J. K. Shoemaker, M. E. Tschakovsky, M. T. Hopman, and R. L. Hughson. Blood flow and muscle oxygen uptake at the onset and end of moderate and heavy dynamic forearm exercise. Am. J. Physiol. Regul. Integr. Comp. Physiol.
28. Weltman, A., B. A. Stamford, and C. Fulco. Recovery from maximal effort exercise: lactate disappearance and subsequent performance. J. Appl. Physiol.
29. Weltman, A., and J. D. Regan. Prior exhaustive exercise and subsequent, maximal constant load exercise performance. Int. J. Sports Med.
30. Yoshida, T., H. Watari, and K. Tagawa. Effect active of and passive recoveries on splitting of the inorganic phosphate peak determined by 31
P-nuclear magnetic resonance spectroscopy. NMR Biomed.