Endurance training based on individual physiological events is effective to enhance training responsiveness and maximize cardiorespiratory, neuromuscular, and functional adaptations (40). This training method requires determining individualized intensities corresponding to physiological milestones, such as the maximal oxygen uptake (VO2max) and the lactate/aerobic-anaerobic thresholds (13,16,34,36,37). An accurate identification of these individual milestones will depend on the testing procedures (4,18,19). Therefore, variations in the testing protocol configuration (e.g., warm-up, workload increments, and total test duration) are decisive in the assessment of endurance performance (11,33).
It is well known that graded exercise testing (GXT), using metabolic systems under laboratory conditions, is the most accurate method to assess physiological responses to exercise in endurance sports (3). In particular, ramp protocols, in which the speed increases in a continuous fashion rather than in bouts (e.g., multistage protocols), are especially recommended for maximal cardiovascular testing and predicting the metabolic cost at individual workload (31,32). Numerous studies confirm that short ramp GXT, lasting 10–14 minutes, are the most appropriate assessment to identify individual physiological events in cyclists (14,24,25,28,33) and runners (11,31,32). The main reason for this choice is that longer protocols (lasting 20–30 minutes) or multistage tests (i.e., speed increments every 2–3 minutes) would prevent athletes from achieving their maximal potential because of accumulative fatigue, dehydration, muscle acidosis, and cardiovascular drift (4,18,32). However, the use of laboratory testing procedures is limited by the requirement of sophisticated equipment that most coaches and athletes are not equipped with or cannot afford. Furthermore, treadmill testing with a metabolic cart is impractical for routine athlete assessment and load adjustment compared with field-based and outdoor assessment using portable technologies. Unfortunately, the technology available for quantifying and monitoring running performance in outdoor conditions such as running power output is still limited (2). Thus, indirect estimations from track tests are, to date, the best alternative for determining individual training intensities in running, when laboratory equipment is not available.
There are 2 main running intensities that coaches can analyze using track tests: the peak velocity (Vpeak) and the maximal aerobic speed (MAS). The Vpeak is the highest speed attained during a test, whereas the MAS is the lowest speed that elicits the VO2max (20). The MAS is a reference value to determine training intensity and workload distribution in endurance sports based on the aerobic performance limits (39). Given the similarity between Vpeak and MAS intensities (11), running track tests use the Vpeak to estimate the VO2max and the corresponding MAS (5,6,21), if no metabolic system is available. Hence, bringing athletes to their maximal aerobic performance (i.e., VO2max) is an essential requirement when designing running track tests to measure athletes' endurance performance in the field (18).
The University of Montreal Track Test (UMTT) is the most famous test to estimate essential running parameters in the field (21). The UMTT follows a multistage GXT protocol with 1 km·h−1 increments every 2 minutes (1:2 ratio) to estimate the VO2max and MAS from the Vpeak attained. The MAS is considered as the speed reached at the last wholly completed stage (e.g., if the last speed [Vpeak] is 18.5 km·h−1, the MAS is 18 km·h−1) (20). Using this simple speed-based calculation, a variety of authors have published their own UMTT modification, including slight variations on testing procedures such as distance between pylons, stage duration, initial speed, and warm-up (7,9,10).
Despite the fact that nowadays most coaches are designing their training plans based on the speed-based estimations derived from these track tests, some issues can question its validity. First, these tests follow long multistage protocols that require athletes to run for a notably long time until exhaustion, especially in well-trained and elite runners (e.g., for a given athlete with MAS = 20 km·h−1, the UMTT will take 24 minutes). As previously noted, this long duration might raise some doubts on whether the Vpeak attained during the test reflects the athletes' true maximal physiological potential (8,18,26,28). Likewise, the laboratory protocols used to validate these track tests followed the same long multistage protocol; consequently, the measures obtained from both are equally limited (7,9,10). Finally, in all these tests, the speed increases in long bouts rather than progressively, which could impede the athletes reaching their real maximal cardiorespiratory performance (32). To overcome these issues, short ramp GXT protocols including 1 km·h−1 increments every 1 minute (1:1 ratio) seem to be a better alternative for running performance assessment in the field (11,31,32). Notwithstanding the aforementioned, although this protocol has been proven in the laboratory (Lab(1:1)) (11), to the best of our knowledge, there is no alternative running track test available for runners.
Therefore, the aim of this study was to validate a new short track test with 1 km·h−1 increments every 1 minute (Track(1:1)) to estimate running performance parameters (VO2max and MAS), based on a laboratory treadmill protocol and gas exchange data analysis (Lab(1:1)). In addition, we compared the results with the UMTT, a multistage longer protocol with 1 km·h−1 increments every 2 minutes.
Experimental Approach to the Problem
Participants performed 4 testing protocols: 2 in the laboratory (Lab(1:1)-pre and Lab(1:1)) and 2 in the field (UMTT and Track(1:1)). Evaluations took place in 4 separate days with 48–72 hours rest in between. On the first day, participants performed a preliminary laboratory test using a treadmill and metabolic cart, following a short ramp GXT protocol with 1 km·h−1 increments every 1 minute (Lab(1:1)-pre), as described elsewhere (11). Measures from the Lab(1:1)-pre were used to determine individuals' Vpeak. On the second day, participants visited the laboratory again to perform another GXT test following the same protocol (Lab(1:1)) but setting the initial running speed according to each participant's Vpeak (i.e., starting at 13 km·h−1 less than each athlete's Vpeak), previously determined in the Lab(1:1)-pre. Measures from the Lab(1:1) were considered as the gold standard for the validity analysis. On the third day, participants completed the multistage UMTT (21). On the fourth day, they completed the Track(1:1), a reproduction of the Lab(1:1) protocol in the field, following the same individual workload adjustment. Participants' heart rate (HR) was continuously monitored (V800; Polar, Kempele, Finland) in each test.
Twenty-two trained male athletes (5,000–21,000 m) and triathletes (5,000 and 10,000 m) volunteered to participate in this study (mean ± SD: age 25.7 ± 7.9 years (all subjects 18 years or older), body mass 67.34 ± 6.5 kg, height 175.9 ± 5.0 cm, body fat 11.3 ± 1.8%, VO2max 60.3 ± 5.9 ml·kg−1·min−1, and endurance training experience 7.3 ± 4.0 years). All participants were competing at regional and national level races and following a regular training load of 4–6 days per week, 1–2 hours per day. Measurements were obtained during the precompetitive season. All participants were familiarized with the testing procedures used in this investigation. They underwent a complete medical examination (including ECG) that showed all were in good health. No physical limitations or musculoskeletal injuries that could affect testing procedures were reported. None of the subjects were taking drugs, medications, or dietary supplements known to influence physical performance. The Bioethics Commission of the University of Murcia approved the study, which was conducted according to the Declaration of Helsinki. Subjects were verbally informed about the experimental procedures and possible risks and benefits. Written informed consent was obtained from all subjects.
Laboratory-Individualized Short Ramp Graded Exercise Testing Protocol
This protocol involved 2 laboratory tests using treadmills and metabolic carts (Lab(1:1)-pre and Lab(1:1)), following a short ramp GXT with 1 km·h−1 increments every 1 minute, as described elsewhere (11). Both tests were performed on the same treadmill (HP Cosmos Pulsar; H Cosmos Sports & Medical GMBH, Nussdorf Traunstein, Germany) with an incline of 1.0% (17). Evaluations were performed under the similar environmental conditions (21–24° C and 45–55% relative humidity) at the same time of the day (16:00–19:00 hours) to minimize the circadian rhythm effects (30). Air ventilation was controlled with a fan positioned 1.5 m from the subject's chest at a wind velocity of 2.55 m·s−1. Ventilatory performance (VO2, VO2max, and ventilation) was recorded on a breath-by-breath basis using a metabolic cart (MetaLyzer 3B-R3; Cortex Biophysik GmbH, Leipzig, Germany). A standardized warm-up was performed before each test. The preliminary test (Lab(1:1)-pre) was made under medical supervision to discard cardiovascular diseases and determine the athletes' Vpeak. The second test (Lab(1:1)) was individualized based on the Vpeak previously determined, as follows: the starting velocity was set at 13 km·h−1 slower than each athlete's Vpeak, after which the workload increased 1 km·h−1 per minute until exhaustion. Maximal effort criteria were considered to verify the outcomes (1). If verified, the MAS was determined as the first running velocity where VO2max was reached (20). The metabolic cart was calibrated before each test according to the manufacturer's instructions. The Vpeak was obtained automatically from the treadmill software using the formula proposed by Kuipers et al. (19):in which Vcomplete is the speed at the last completed stage, Inc is the speed increment (i.e., 1 km·h−1), t is the time in seconds sustained during the incomplete stage, and T is the time in seconds required to complete a stage (i.e., 60 seconds).
University of Montreal Track Test
The UMTT (21) was carried in a 400-m outdoor flat track. Running pace was controlled by audio beeps on a prerecorded file. Participants had to reach a pylon on each beep. Pylons were placed every 25 m along the track. The test ended when the athlete could not keep the imposed pace by the beeps and failed to reach the next pylon twice in a row. Initial speed was set at 8 km·h−1 and thereafter increased by 1 km·h−1 every 2 minutes until exhaustion. The Vpeak was obtained using the formula proposed by Kuipers et al. (19) considering 1 km·h−1 increments and the 120 seconds required to complete a stage. The speed at the last completed stage was taken as the MAS (e.g., if the last speed [Vpeak] is 18.5 km·h−1, the MAS is 18 km·h−1) (5,20). VO2max was calculated using the formula proposed by Léger and Mercier (22):
Track-Individualized Short Ramp Graded Exercise Testing
The Track(1:1) was designed to follow the same ramp GXT protocol as the Lab(1:1) in outdoor conditions. The test was performed on a 400-m outdoor flat track on the basis of the UMTT (i.e., pace controlled by audio beeps and pylons each 25 m), but the velocity increased by 1 km·h−1 every 1 minute. After a standardized warm-up, the test started at 13 km·h−1 slower than each athlete's Vpeak (previously determined during the Lab(1:1)) followed by progressive increments of 1 km·h−1 every 1 minute until exhaustion (11). Participants' running pace was individually set-up and controlled using automated sound beeps. The Vpeak was obtained using the same formula than the UMTT proposed by Kuipers et al. (19), considering 1 km·h−1 increments and the 60 seconds required to complete a stage. The MAS was estimated using the equation proposed by Cerezuela-Espejo et al. (11):
Standard statistical methods were used for the calculation of means, SDs, and 95% confidence interval. Comparisons between VO2max, MAS, Vpeak, and HRmax measures obtained from the tests (Lab(1:1), UMTT, and Track(1:1)) were conducted by analysis of variance with post hoc comparisons, intraclass correlation coefficient, and Bland-Altman bias analyses. The magnitude of agreement was examined through mean difference bias, mean ± SD, and 95% limits of agreement (LoA = bias ± 1.96 mean ± SD) calculations. Effect size (ES) was estimated using the Cohen's d index and interpreted as small (0.20), medium (0.50), and large (0.80). Analyses were performed using GraphPad Prism 6.0 (GraphPad Software, Inc., CA, USA) and SPSS software version 19.0 (IBM Corp., Armonk, NY, USA).
A high linear relationship (r = 0.91) was observed between the Vpeak from the Track(1:1) and the VO2max obtained from the Lab(1:1) (Figure 1). Assuming a standard error of 2.97 ml·kg−1·min−1, the resulting equation was:
The total distances achieved at the end of each test were Lab(1:1) = 2,784 ± 281 m, Lab(1:1) = 2,827 ± 321 m, and UMTT = 4,877 ± 984 m. Outcomes from each test are shown in Table 1. Analysis of variance was not able to detect mean difference in VO2max (F = 1.738; p = 0.184), MAS (F = 1.451; p = 0.242), or HRmax (F = 0.296; p = 0.745) among the tests. In turn, Vpeak (F = 3.181; p = 0.048) was significantly lower in the UMTT compared with the Lab(1:1) (mean difference = −0.81 km·h−1; ES = 0.62; p = 0.04) and the Track(1:1) (mean difference = −0.93 km·h−1; ES = 0.67; p = 0.03). Correlation analysis revealed high and significant coefficients (r > 0.889) between both track tests (UMTT and Track(1:1)) and the Lab(1:1). However, Bland-Altman analyses (bias and 95% LoA) confirmed a greater agreement in all the studied performance parameters in the Track(1:1) compared with the UMTT. Conversely, the UMTT overestimated the VO2max ∼2.9 ml·kg−1·min−1, underestimated the MAS ∼0.5 km·h−1, and importantly underestimated the Vpeak ∼0.8 km·h−1 compared with the laboratory outcomes. Figure 2 depicts these results graphically using Bland-Altman plots.
The current Track(1:1) test proposal (individualized, short ramp GXT on track with 1 km·h−1 increments per minute) was a better alternative than the UMTT to estimate maximal running performance parameters such as the VO2max and MAS. The current findings provide further empirical evidence about the advantages of conducting individualized, short ramp protocols to assess maximal physiological parameters in endurance sports, particularly in running. Moreover, in light of the high agreement between the Track(1:1) test and the gas exchange laboratory conditions, this novel proposal emerges as a valid alternative to longer and multistage traditional track tests for a better evaluation of athletes' running performance.
Our findings revealed that the UMTT underestimated the Vpeak by 4.2% (∼0.81 km·h−1) compared with the Lab(1:1), a validated maximal short treadmill GXT using gas exchange systems (11). Consequently, indirect estimations from the Vpeak attained at the end of the UMTT were notably altered and did not reflect the true maximal aerobic performance. These results confirm earlier findings suggesting that long (>20 minutes) and multistage tests avoid the athletes reaching their maximal running speed (8,18,26,28,32). According to this disclosure, the Vpeak attained in the Track(1:1) was higher than the UMTT (±1.0 km·h−1) and almost the same as the Lab(1:1) (bias = <0.1 km·h−1). Thus, the current short ramp Track(1:1) proposal is a more valid option than UMTT to make estimations derived from maximal running intensities such as the Vpeak using the same human and material resources.
A main advantage of the current Track(1:1) is the estimation of the MAS using a practical formula (11) based on the values obtained from the same protocol under laboratory conditions (Lab(1:1)). The use of this formula allowed us to obtain a better estimation for the MAS (0.3% different) than using the UMTT methods (2.6% different) compared with gas exchange. This is critical, given the MAS is a helpful indicator to monitor training loads, assess changes in aerobic endurance, and individualize theoretical submaximal and maximal training intensities in runners (5,12,25,27,29). This high precision in the MAS estimation based on direct gas exchange measurements constitutes a powerful improvement of the Track(1:1) among available running track tests.
It is worth noting that the laboratory protocol used as a gold standard (Lab(1:1)) has been proven as effective to estimate critical workloads, such as ventilatory thresholds (VT1 and VT2) and maximal lactate steady state (11). In this work, the authors provided a personal approach for exercise prescription (training zones). Given the Track(1:1) follows the same testing procedure as the Lab(1:1), these training zones can be determined from the MAS and Vpeak values obtained in this test.
The estimation of the VO2max derived from the Vpeak attained during the Track(1:1) is a novel contribution of this study. This is possible due to the fact that the current short ramp GXT protocol guarantees that the Vpeak attained is likely to be the true individual fastest speed (11). Thus, we were able to yield an equation (Figure 1) to estimate VO2max, assuming an error of 2.97 ml·kg−1·min−1. This error could mainly come from the athletes' running economy (i.e., a different rate of energy consumption at a given speed), which has been shown to explain 7–12% of VO2max variations in elite running athletes (38). Considering the current sample (VO2max = 60 ml·kg−1·min−1), VO2max might vary between 5.4 and 7.2 ml·kg−1·min−1 because of running economy, which could partially explain the error produced in our resulting equation (Figure 1). Future investigations are required to include running economy assessment during the Track(1:1) and Lab(1:1) testing procedures to refine the accuracy of the estimations.
Finally, although it is true that the UMTT showed high correlations, this coefficient may be limited because it indicates that a value changes when another changes but does not identify the presence of a high systematic error difference between measurements (23). Hence, for training and practical applications, the use of Bland-Altman agreement seems to be more relevant than correlations (15). In this sense, our findings revealed that, despite similar correlation, the Track(1:1) test exhibited higher agreement and lower bias with the laboratory than the UMTT. Therefore, the Track(1:1) should be a preferred option for testing athletes.
This investigation has some limitations that should be noted. A familiarization with the sound signals is recommended for optimal running pace adjustment when performing audio-guided running track tests. No women were included in this test, so further investigations should confirm this validation. In addition, despite the near-perfect agreement, very good correlations, and minimum bias between the Track(1:1) and the laboratory, future studies should corroborate these findings by reproducing the Track(1:1) using a portable metabolic cart (35).
This study demonstrates that the Track(1:1) is a valid, noninvasive, and individualized test to assess both external (Vpeak and MAS) and internal (VO2max and HRmax) running performance parameters, while assuring that the athletes reach their maximal aerobic velocity. The use of the Vpeak to obtain high precision ventilatory estimations compared with laboratory exchange measurements constitutes a main practical application for daily training monitoring. The Track(1:1) intensity is individualized according to the athlete's maximum potential (i.e., Vpeak) to allow for completing the test within 14 minutes; in turn, longer or multistage track tests could lead to an early fatigue failure and underestimate the outcomes. Thanks to this short duration, measurements can be obtained from a single training session without compromising the training plan.
1. ACSM. Interpretation of Clinical Exercise Test Results. In: 9th, ed. ACSM's Guidelines for Exercise Testing and Prescription. Philadelphia, PA: Lippincott Williams and Wilkins, 2013. pp. 142–160.
2. Aubry RL, Power GA, Burr JF. An assessment of running power as a training metric for elite and recreational runners. J Strength Cond Res 32: 1, 2018.
3. Beltz NM, Gibson AL, Janot JM, Kravitz L, Mermier CM, Dalleck LC. Graded exercise testing protocols for the determination of vo2
max: Historical perspectives, progress, and future considerations. J Sports Med 2016: 1–12, 2016.
4. Bentley DJ, Newell J, Bishop D. Incremental exercise test design and analysis: Implications for performance diagnostics in endurance athletes. Sport Med 37: 575–586, 2007.
5. Berthoin S, Gerbeaux M, Turpin E, Guerrin F, Lensel‐Corbeil G, Vandendorpe F. Comparison of two field tests to estimate maximum aerobic speed. J Sports Sci 12: 355–362, 1994.
6. Berthon P, Fellmann N, Bedu M, Beaune B, Dabonneville M, Coudert J, et al. A 5-min running field test as a measurement of maximal aerobic velocity. Eur J Appl Physiol Occup Physiol 75: 233–238, 1997.
7. Billat V, Hill DW, Pinoteau J, Petit B, Koralsztein JP. Effect of protocol on determination of velocity at VO2max and on its time to exhaustion. Arch Physiol Biochem 104: 313–321, 1996.
8. Buchfuhrer MJ, Hansen JE, Robinson TE, Sue DY, Wasserman K, Whipp BJ. Optimizing the exercise protocol for cardiopulmonary assessment. J Appl Physiol 55: 1558–1564, 1983.
9. Cazorla G. Field tests to Evaluate Aerobic Capacity and Maximum Aerobic Speed. In: The Evaluation in Physical Activity and in Sport. Cazorla G, Robert G, eds. Cestas, France: AREAPS, 1990. pp. 151–174.
10. Cazorla G, Leger L. How to Evaluate and Develop Your Abilities Aerobics: The Progressive Test of the Shuttle Race, The Progressive Test of VAMEVAL Track Race. Cestas, France: AREAPS Research and Evaluation Association in Physical Activity and Sport, 1993. p. 123.
11. Cerezuela-Espejo V, Courel-Ibáñez J, Morán-Navarro R, Martínez-Cava A, Pallarés JG. The relationship between lactate and ventilatory thresholds in runners: Validity and reliability of exercise test performance parameters. Front Physiol 9: 1320, 2018.
12. Esfarjani F, Laursen PB. Manipulating high-intensity interval training: Effects on VO2max, the lactate threshold and 3000 m running performance in moderately trained males. J Sci Med Sport 10: 27–35, 2007.
13. García-Pallarés J, García-Fernández M, Sánchez-Medina L, Izquierdo M. Performance changes in world-class kayakers following two different training periodization models. Eur J Appl Physiol 110: 99–107, 2010.
14. Gaskill SE, Ruby BC, Walker AJ, Sanchez OA, Serfass RC, Leon AS. Validity and reliability of combining three methods to determine ventilatory threshold. Med Sci Sports Exerc 33: 1841–1848, 2001.
15. Giavarina D. Understanding Bland Altman analysis. Biochem Med 25: 141–151, 2015.
16. Helgerud J, Høydal K, Wang E, Karlsen T, Berg P, Bjerkaas M, et al. Aerobic high-intensity intervals improve VO2max more than moderate training. Med Sci Sport Exerc 39: 665–671, 2007.
17. Jones AM, Doust JH. A 1% treadmill grade most accurately reflects the energetic cost of outdoor running. J Sports Sci 14: 321–327, 1996.
18. Julio UF, Panissa VLG, Shiroma SA, Franchini E. Effect of protocol manipulation for determining maximal aerobic power on a treadmill and cycle ergometer: A brief review. Strength Cond J 5: 58–71, 2017.
19. Kuipers H, Rietjens G, Verstappen F, Schoenmakers H, Hofman G. Effects of stage duration in incremental running tests on physiological variables. Int J Sports Med 24: 486–491, 2003.
20. Lacour JR, Padilla-Magunacelaya S, Chatard JC, Arsac L, Barthélémy JC. Assessment of running velocity at maximal oxygen uptake. Eur J Appl Physiol Occup Physiol 62: 77–82, 1991.
21. Léger L, Boucher R. An indirect continuous running multistage field test: The Universite de Montreal Track Test. Can J Appl Physiol 5: 77–84, 1980.
22. Léger L, Mercier D. Gross energy cost of horizontal treadmill and track running. Sport Med 1: 270–277, 1984.
23. Lin L, Hedayat AS, Sinha B, Yang M. Statistical methods in assessing agreement. J Am Stat Assoc 97: 257–270, 2002.
24. Lucía A, Hoyos J, Pérez M, Chicharro JL. Heart rate and performance parameters in elite cyclists: A longitudinal study. Med Sci Sports Exerc 32: 1777–1782, 2000.
25. Lucía A, Sánchez O, Carvajal A, Chicharro JL. Analysis of the aerobic-anaerobic transition in elite cyclists during incremental exercise with the use of electromyography. Br J Sports Med 33: 178–185, 1999.
26. Lukaski HC, Bolonchuk WW, Klevay LM. Comparison of metabolic responses and oxygen cost during maximal exercise using three treadmill protocols. J Sports Med Phys Fitness 29: 223–229, 1989.
27. McLaughlin JE, Howley ET, Bassett DR, Thompson DL, Fitzhugh EC. Test of the classic model for predicting endurance running performance. Med Sci Sports Exerc 42: 991–997, 2010.
28. Midgley AW, McNaughton LR, Carroll S. Time at VO2max during intermittent treadmill running: Test protocol dependent or methodological artefact? Int J Sports Med 28: 934–939, 2007.
29. Millet GP, Candau R, Fattori P, Bignet F, Varray A. VO2 responses to different intermittent runs at velocity associated with VO2max. Can J Appl Physiol 28: 410–423, 2003.
30. Mora-Rodríguez R, Pallarés JG, López-Gullón JM, López-Samanes Á, Fernández-Elías VE, Ortega JF. Improvements on neuromuscular performance with caffeine ingestion depend on the time-of-day. J Sci Med Sport 18: 338–342, 2015.
31. Myers J, Bellin D. Ramp exercise protocols for clinical and cardiopulmonary exercise testing. Sport Med 30: 23–29, 2000.
32. Myers J, Buchanan N, Walsh D, Kraemer M, McAuley P, Hamilton-Wessler M, et al. Comparison of the ramp versus standard exercise protocols. J Am Coll Cardiol 17: 1334–1342, 1991.
33. Pallarés JG, Morán-Navarro R, Ortega JF, Fernández-Elías VE, Mora-Rodriguez R. Validity and reliability of ventilatory and blood lactate thresholds in well-trained cyclists. PLoS One 11: e0163389, 2016.
34. Pallarés JG, Sánchez-Medina L, Carrasco L, Díaz A, Izquierdo M. Endurance and neuromuscular changes in world-class level kayakers during a periodized training cycle. Eur J Appl Physiol 106: 629–638, 2009.
35. Perez-Suarez I, Martin-Rincon M, Gonzalez-Henriquez JJ, Fezzardi C, Perez-Regalado S, Galvan-Alvarez V, et al. Accuracy and precision of the cosmed k5 portable analyser. Front Physiol 9: 1764, 2018.
36. Scharhag-Rosenberger F, Walitzek S, Kindermann W, Meyer T. Differences in adaptations to 1 year of aerobic endurance training: Individual patterns of nonresponse. Scand J Med Sci Sports 22: 113–118, 2012.
37. Seiler S, Tønnessen E. Intervals, thresholds, and long slow distance: The role of intensity and duration in endurance training. Sportscience 13: 32–53, 2009.
38. Shaw AJ, Ingham SA, Atkinson G, Folland JP. The correlation between running economy and maximal oxygen uptake: Cross-sectional and longitudinal relationships in highly trained distance runners. PLoS One 10: e0123101, 2015.
39. Stöggl T, Sperlich B. The training intensity distribution among well-trained and elite endurance athletes. Front Physiol 6: 295, 2015.
40. Wolpern AE, Burgos DJ, Janot JM, Dalleck LC. Is a threshold-based model a superior method to the relative percent concept for establishing individual exercise intensity? A randomized controlled trial. BMC Sports Sci Med Rehabil 7: 16, 2015.