Overestimation of Critical Power and Underestimation of W’ in a 3-minute all-out Exercise Test ABSTRACT:The purpose of this study was to investigate if a 3-minute all-out exercise test was an accurate method of measuring CP and W’. A multi-step exercise test was used as a baseline in order confirm these values to be valid. It was hypothesized that both the CP and W’ values will remain the same between the 3MT and the MST, with no significant difference seen. 13 total subjects (7 females and 6 males) were tested on a cycle ergometer through 2 testing protocols; the 3-minute all out-test (3MT) and the multi-step exercise test (MST). CP estimated from the 3MT was significantly higher than the MST, as well W’ was significantly lower, displaying an overestimation of CP and an underestimation of W’. These values were recorded as a 12% decrease in CP and a 98% increase in W’ from the 3MT to the MST. Thus proving that the 3MT was an inaccurate method of estimating CP and W’. The hypotheses were rejected and proven opposite.INTRODUCTION:Important information regarding metabolic work capacities can be extrapolated from the study of 2 parameters, CP and W’. The concept of critical power (CP) can been defined in a multitude of capacities. Firstly, CP was originally estimated through a series of graphs by Hill (1927). The first relationship was a hyperbolic equation modeling power versus time. When graphing power output (measured in Watts) on the y-axis and time until exhaustion (tLIM) (measured in seconds) on the x-axis, CP (measured in Watts) appeared as a horizontal asymptote at a specific power output. The variable of W’ was also studied through the power-time curve, displayed as the “curvature constant” (Miura, Sato, Sato, Whipp, & Fukuba, 2000). W’ was related to a distinct anaerobic energy store that did not fluctuate although ranging power outputs according to Miura et al. (2000). Sustainable power outputs fell with increasing exercise duration, leveling off at the critical power asymptote when graphed as the hyperbolic equation (Jones & Vanhatalo, 2017). Furthermore, CP was also fitted to a linear form. For this graph, the y-axis was still maintained as power output, while the x-axis was changed to the reciprocal of TLIM, which became 1/TLIM (Barker, Poole, Noble, & Barstow, 2006). For this equation, CP was calculated as the y-intercept, while the slope of the line was equated to be W’ (Barker et al., 2006).CP was further defined by its location on these graphs within exercise domains and boundaries. Exercise physiologists have defined three domains of exercise intensity; moderate, heavy and severe. The boundary between heavy and severe exercise domains has been established as CP (Burnley, Doust & Vanhatalo, 2006). CP occurred at a work output between the lactate threshold and VO2 max (Monod & Scherrer, 1965). Under the boundary of CP, a subject could theoretically exercise for an infinite amount of time, however upon passing from heavy to severe exercise, steady state is no longer maintained and VO2 peak will eventually be achieved alongside fatigue.Threfore, the appearance of fatigue can be further understood by the disappearance of steady state. CP was also demonstrated as the highest work rate at which a steady state can be attained (Vanhatalo, Doust & Burnley, 2007). At this power output variables such as phosphocreatine, blood lactate and pulmonary oxygen uptake were maintained at stable values (Jones & Vanhatalo, 2017). Above the CP output, these fatigue inducing metabolites began to accumulate leading to exercise intolerance, therefore CP was further defined as the commencement of fatigue inducing metabolites (Chidnok et al., 2013). Burnley and Jones (2007) stated that W’ was related to the accumulation of fatigue inducing metabolites. Additionally, Fukuba et al. (2003) declared that encountering a limit of tolerance was more likely to occur due to an increasing amount of fatigue inducing metabolites versus from a lack of energy stores. This accumualtion of metabolite build-up was proportional and interrelated to the utilization rate of W’ (Fukuba et al., 2003).  Moritani, Nagata, Devries, & Muro’s (1981) work, displayed that W’ stores were not altered although subjects were introduced to a hypoxic environment, thus proving that W’ is solely reliant on anaerobic stores within the body and is not aerobically dependent. As W’ stores were proposed to be stored energy, W’ was said to remain constant independent of power output and trial sequence (3MT versus MST).Traditionally CP and W’ were assessed through a multi-step “constant load” exercise test (MST). This conclusion came from the basis of  Moritani et al.’s (1981) work which determined the MST to accurately predict CP and W’ making it the “gold standard”. However, due to its long nature and multiple tests traditionally requiring 4 or 5 exhaustive trials, a new test was created in order to decrease the amount of time under maximal conditions (Bartram, Thewlis, Martin, and Norton, 2017). Therefore testing protocols were altered by decreasing the time to a 3 minute “all-out” sprint-exercise test (3MT) firstly done by Burnley et al. (2006). A 3MT was further proven by Vanhatalo et al. (2007) to accurately predict CP in a short window of time, however other studies such as Bartram et al. (2017) have proven the opposite. These theories were tested within this lab, comparing the 3MT values to the confirmed MST values to analyze if this testing protocol was reliable. Since the literature provided a valid argument for both sides of this discussion, a general consensus has yet to be met/made. Thus, establishing the necessity for a comparison.The principal purpose of this study was to investigate if the 3MT was a valid method for predicting CP and W’. Comparing these values to the trusted MST values established if they were consistent and valid predictions.It was hypothesized that:i)                    Critical power was similar across both the 3MT and the MST for all subjects.ii)                 W’ would remain constant across both the 3MT and the MST for all subjects.RESULTS: The mean CP during the 3MT was 178 ± 78 W, which was significantly different (P <0.05) then the mean CP during the MST which was 151 ± 52 W. The mean % change in CP from the 3MT to the MST was a decrease by 12% ± 12%. The correlation coefficient value of the trendline was 0.887, as seen in figure 1.The mean W’ value for the 3MT was 11 ± 6 kJ, which was significantly different (P<0.05) then the mean W’ value for the MST which was 19 ± 7 kJ. The mean % chage in W’ from the 3MT to the MST was an increase by 98% ± 121%. The correlation coefficient value of the trendline was 0.225, as seen in figure 2.DISCUSSION: The CP and W’ values measured across both testing methods (3MT and MST) were proven to be statistically different, thus disproving the individual hypotheses about CP and W’. The general tendency was for CP to decrease from the 3MT to the MST, meaning an overestimation of CP occurred during the 3MT. While, the W’ values were seen to increase from the 3MT to the MST, characterizing an underestimation of W’ from the 3MT. These trends prove that the 3MT was an inaccurate method of predicting both the CP and W’ values and this testing protocol should be rejected as reliable.CP was overestimated in the 3MT when compared to the “gold standard” MST. The values across test protocols were proven to be statistically significant with a mean percentage decrease of 12% seen from 3MT to MST. This percentage overestimation was seen in other studies, ranging from 3% (Black, Jones, Kelly, Bailey, & Vanhatalo, 2016) to 17% (Housh, Housh, & Bauge, 1989). As this finding was consistent with other studies, together these works provide validity for the argument that a 3MT was improper in fully predicting one’s CP as the data were not replicable.This being said, our data was also in direct contradiction to scientific research, thus there was equal evidence proving that the 3MT was valid as well as invalid. Inconsistencies with our findings, arose from literature that stated that CP and W’ values from the 3MT were not significantly differentfrom the values obtained from the predicting trials (Burnley et al., 2006). Evidently, this was not the case in our group data. One basis for this explanation came from a theory developed by Faria (1992) named “teleoanticipation”. Ulmer (1996) defined teleoanticipation as a subconscious pacing strategy which allows exercise performance to be regulated and completed within the biomechanical and metabolic limits. Further, this feedback mechanism was instated in Bartram et al.’s (2017) work, where it was proven that pacing occurred during a 3-minute all-out approach where similar results for CP were seen. Bartram et al. (2017) stated that although subjects were motivated, naturally teleoanticipation deceived the participant, making true attainment of CP as well as full depletion of W’ stores not possible. Therefore the main reason for inaccuracy between the two tests types was explained by the failure to attend to the all-out pacing technique. For this reason, the conclusion was drawn that the 3MT provided an overestimation of CP, which mimics the data seen in our lab. W’ was underestimated in the 3MT when compared to the MST. This underestimation was proven by the low correlation coefficient observed in figure 2 as well as the mean values across testing techniques were proven to be significantly different. This mimics the results seen by Bartram et al. (2017), where a mean increase of 8.8 kJ was seen in the W’ when switching from the 3MT to the MST, whereas in our lab an 8 kJ increase was seen. Despite this research, studies have also proven the opposing argument, which stated that W’ remained constant between trial types (Burnley et al., 2006). A physiological justification for the underestimation of W’ was that participants did not use their full W’ stores during the 3MT, due to the teleoanticipation concept as previously stated (Bartram, 2017). If there was a surplus of anaerobic energy uncounted for such as ATP, PCr, or stored muscle glycogen, this would lead to a wrongful estimate, and thus when compared to the MST would appear as an underestimation. Another rationale for the mismatch of W’, was due to exercise intensity and duration. The rate of W’ utilization was dependent on the power output relative to CP (Jones & Vanhatalo, 2017), therefore it has differing rates of depletion. Although W’ was originally thought to be constant, Bishop, Jenkins, and Howard (1998) stated that work rate and duration of exercise had a significant effect on W’ usage. Bishop et al. (1998), went on to say that during higher intensity bouts a lower anaerobic energy yield was produced as it is not the predominant energy source for this intensity. W’ was best predicted by longer duration exercises with lower power outputs, in order to maximize the anaerobic capacity output (Bishop et al., 1998).A low estimation of W’ was seen when performing higher power outputs similar to the 3MT, as the anaerobic capacity cannot keep up with the energy demands at this extreme or maximal rate.In addition, it was found that W’ underestimation occurred when test protocols lasted under 7 minutes in length (Maturana, Fontana, Pogliaghi, Passfield, & Murias, 2018). The trial time of 3 minutes for the 3MT falls within this criteria, thus this was one method of explaining the inconsistencies between the 3MT and MST, further disproving the validity in the 3MT. Taking our data into account as well as the previous research, the conclusion that the 3MT protocol cannot accurately predict W’ when compared to a MST was proven.The main assumptions of this study fell upon the participants. There was no method to confirm in real time if the subject had obtained their true critical power, it was solely left up to the individual when to decide to quit. It was also assumed that individuals were all healthy and nobody was sick across the 3 trial weeks. The cycle ergometer was assumed to be properly calibrated and it was assumed that the subject maintained the proper cadence for the entirety of the trial. Furthermore, it was assumed that the subjects’ anaerobic energy stores did not change between the 3 trial weeks. It was assumed the subjects maintained the same relative diet as well as the same level of fitness between the 3 weeks to allow for similar testing. Lastly, it was assumed that all subjects were highly motivated and did not pace themselves for the 3MT.The collected data could have provided more accurate results had certain limitations not been in place. The first limitation being that the calculation of CP for the 3MT was dependent on an accurate interpretation of a video recording of the subject during the 3 minutes. Analyzing these videos on cell phones made it difficult to scroll through the video accurately instead of being able to analyze the video frame by frame. Therefore inaccurate counting of revolutions during the 3MT is a limitation. Secondly, tLIM was not provided within the class data set thus the graphs to calculate CP and W’ could not be replicated for this lab. One further limitation was the sample size (n=13) as well as the demographic. Since all subjects were similar in age and physical fitness levels, these results could be varied given a different sample of participants, displaying subject-group bias.CONCLUSION:The 3-minute all-out exercise test was deemed invalid in estimating CP and W’, as the predicted/projected values were inaccurate and not reproducible when compared to the MST. The 3-minute all-out exercise test significantly overestimated CP and underestimated W’ when compared to the baseline multi-step test. These conclusions were in direct disagreement with the original hypotheses, thus the hypotheses was rejected. The physiological basis behind these observations comprised the concept of teleoanticipation as well as the exercise intensity and duration relationship. Future research possibilities could include the manipulation of W’ stores through different nutritional practices i.e. carboloading in order to maximize anaerobic fuel sources. As well, analyzing speeds of movement in order to maintain a power output just under the CP to avoid the negative consequences of CP. For example, analyzing running speed in marathon runners in order to reducing triggering of fatigue inducing metabolites. Upon analyzing our current findings it was proposed that the 3MT was invalid in accurately predicting CP and W’ due to numerical inconsistencies when compared to the “gold standard” MST.REFRENCES:Barker, T., Poole, D. C., Noble, M. L., & Barstow, T. J. (2006). Human critical power-oxygen uptake relationship at different pedalling frequencies. Experimental Physiology,91(3), 621-632. doi:10.1113/expphysiol.2005.032789Bartram, J. C., Thewlis, D., Martin, D. T., & Norton, K. I. (2017). Predicting Critical Power in Elite Cyclists: Questioning the Validity of the 3-Minute All-Out Test. International Journal of Sports Physiology and Performance,12(6), 783-787. doi:10.1123/ijspp.2016-0376Black, M. I., Jones, A. M., Kelly, J. A., Bailey, S. J., & Vanhatalo, A. (2016). The constant work rate critical power protocol overestimates ramp incremental exercise performance. European Journal of Applied Physiology,116(11-12), 2415-2422. doi:10.1007/s00421-016-3491-yBishop, D., Jenkins, D., & Howard, A. (1998). The Critical Power Function is Dependent on the Duration of the Predictive Exercise Tests Chosen. International Journal of Sports Medicine,19(02), 125-129. doi:10.1055/s-2007-971894Burnley, M., Doust, J. H., & Vanhatalo, A. (2006). A 3-min All-Out Test to Determine Peak Oxygen Uptake and the Maximal Steady State. Medicine & Science in Sports & Exercise,38(11), 1995-2003. doi:10.1249/01.mss.0000232024.06114.a6Burnley, M., & Jones, A. M. (2007). Oxygen uptake kinetics as a determinant of sports performance. European Journal of Sport Science,7(2), 63-79. doi:10.1080/17461390701456148Chidnok, W., Fulford, J., Bailey, S. J., DiMenna, F. J., Skiba, P. F., Vanhatalo, A., & Jones, A. M. (2013). Muscle metabolic determinants of exercise tolerance following exhaustion: Relationship to the “critical power”. Journal of Applied Physiology,115(2), 243-250. doi:10.1152/japplphysiol.00334.2013Faria, I. E. (1992). Energy Expenditure, Aerodynamics and Medical Problems in Cycling. Sports Medicine,14(1), 43-63. doi:10.2165/00007256-199214010-00004Fukuba, Y., Miura, A., Endo, M., Kan, A., Yanagawa, K., & Whipp, B. J. (2003). The Curvature Constant Parameter of the Power-Duration Curve for Varied-Power Exercise. Medicine & Science in Sports & Exercise,35(8), 1413-1418. doi:10.1249/01.mss.0000079047.84364.70Hill, A. V. (1927). Muscular Movement in Man: the Factors Governing Speed & Recovery from Fatigue. McGraw-Hill.Housh, D. J., Housh, T. J., & Bauge, S. M. (1989). The accuracy of the critical power test for predicting time to exhaustion during cycle ergometry. Ergonomics,32(8), 997-1004. doi:10.1080/00140138908966860Jones, A. M., Vanhatalo, A., Burnley, M., Morton, R. H., & Poole, D. C. (2010). Critical Power: Implications for Determination of V?O2max and Exercise Tolerance. Medicine & Science in Sports & Exercise,42(10), 1876-1890. doi:10.1249/mss.0b013e3181d9cf7fJones, A. M., & Vanhatalo, A. (2017). The ‘Critical Power’ Concept: Applications to Sports Performance with a Focus on Intermittent High-Intensity Exercise. Sports Medicine,47(S1), 65-78. doi:10.1007/s40279-017-0688-0Maturana, F. M., Fontana, F. Y., Pogliaghi, S., Passfield, L., & Murias, J. M. (2018). Critical power: How different protocols and models affect its determination. Journal of Science and Medicine in Sport,21(7), 742-747. doi:10.1016/j.jsams.2017.11.015Miura, A., Sato, H., Sato, H., Whipp, B. J., & Fukuba, Y. (2000). The effect of glycogen depletion on the curvature constant parameter of the power-duration curve for cycle ergometry. Ergonomics,43(1), 133-141. doi:10.1080/001401300184693Monod, H., & Scherrer, J. (1965). The work capacity of a synergic muscular group. Ergonomics, 8(3), 329-338.Moritani, T., Nagata, A., Devries, H. A., & Muro, M. (1981). Critical power as a measure of physical work capacity and anaerobic threshold. Ergonomics,24(5), 339-350. doi:10.1080/00140138108924856Ulmer, H. (1996). Concept of an extracellular regulation of muscular metabolic rate during heavy exercise in humans by psychophysiological feedback. Experientia,52(5), 416-420. doi:10.1007/bf01919309Vanhatalo, A., Doust, J. H., & Burnley, M. (2007). Determination of Critical Power Using a 3-min All-out Cycling Test. American College of Sports Medicine,2007(39), 548-555. doi:10.1249/mss.0b013e31802dd3e6APPENDIX A:Calculation 1. %  Change in CP between 3 minute “all-out” and multi-stepSubject 2: CP 3 minute “all-out” = 141 WSubject 1: CP multi-step = 133 W(CPmulti–step – CP3 minute “all–out”)CP3 minute “all–out”x 100%= (133 W–141W)141 Wx 100%=–6%Therefore a 6% decrease is seen in subject 2 when comparing CP from the 3MT to the MST. Figure 1.Relationship between the CP values predicted from the 3MT and the MST. The r value for the trendline = 0.887.Figure 2.Relationship between the W’ values predicted from the 3MT and the MST. The r value for the trendline = 0.225.