You are on page 1of 5

FOOTSTEP ANALYSIS AT DIFFERENT SLOPES SPEEDS IN ELITE RACE WALKING

AU1 AU2
1

AND

JOHNNY PADULO,1,2 GIUSEPPE ANNINO,1 STEFANO DOTTAVIO,1 GIANLUCA VERNILLO,3 LAURA SMITH,4 ZSEPH TIHANYI5 GIAN MARIO MIGLIACCIO,2 AND JO
Faculty of Medicine and Surgery Motor Sciences, University Tor Vergata Rome, Italy; 2Coni-Italian Regional Olympic Committee, Sardinia, Italy; 3Department of Sports, Nutrition and Health Sciences, Faculty of Exercise Sciences, Universita` degli Studi di Milano, Milan, Italy; 4Directorate of Sport, Exercise and Physiotherapy, School of Health Sciences, University of Salford, United Kingdom; and 5Department of Biomechanics, Faculty of Physical Education and Sport Sciences, Semmelweis University, Budapest, Hungary

AU3 AU4

ABSTRACT
Padulo, J, Annino, G, DOttavio, S, Vernillo, G, Smith, L, Migliaccio, GM, and Tihanyi, J. Footstep analysis at different slopes and speeds in elite race walking. J Strength Cond Res XX(X): 000000, 2012To investigate the effects of speed and slope on kinematic parameters, we studied the step parameters of 12 elite race walkers on a treadmill at different speeds (3.61, 3.89, and 4.17 ms21) and slopes (0, 2, and 7%). A high-speed digital camera (210 Hz) was used to record motion, and 2D data were analyzed with Dartsh 5.5Pro. The parameters studied were step length (SL), step frequency (SF), and contact time (CT). The results showed that the increases in SL were linearly related to increases in speed: r = 0.37 with p , 0.01, whereas decreases in SL were elicited with an increase in slope r = 0.56, p , 0.0001 ([02% = 3.5%, p , 0.02], [07% = 7.5%, p , 0.01]). The SF was positively correlated with increases in speed, r = 0.56, p = 0.0001, and slopes, r = 0.50, p , 0.0001 (02% = 3.6% n.s.; 07% = 8.5%, p , 0.01). Conversely, CT was negatively correlated with increases of both speed; r = 0.57, p , 0.0001 and slope r = 0.50, p , 0.0001 (02% = 3.4%, n.s.; 07% = 7.7%, p , 0.01). These results suggest that using slopes ,7% could considerably alter the neuromechanical behavior of athletes, whereas slopes around 2% could positively inuence the performance of the elite race walkers without altering the race-walking technique.
AU5

INTRODUCTION

KEY WORDS race walking, uphill locomotion, kinematics

Address correspondence to Johnny Padulo, sportcinetic@yahoo.it. 0(0)/15 Journal of Strength and Conditioning Research 2012 National Strength and Conditioning Association

alking and running have been frequently researched from both a biomechanical and a physiological point of view. Race-walking (RW) research is increasing in popularity because it is one of the special types of gait included in track and eld events (22). It is well documented in the literature that RW is an intermediate mechanical type of locomotion, between walking and running (6,11,12). Consequently, there are signicant differences in kinematic and kinetic variables between the different types of gait; walking, RW, and running. The RW gait, consequently speed, are governed by the strict rules imposed: the International Association of Athletics Federation stipulate the knee must be straightened from initial contact through to the vertical upright position, also that the heel of the advancing foot contacts the ground before the rear foot leaving the ground (double support phase). There are however few studies investigating the effects of speed on RW gait. Murray et al. compared normal fast walking to RW and found that RW elicited greater speed because of increased step length (SL) and cadence, consequently increased muscle activation (18). Similarly, Cairns et al. compared normal walking, slow and fast RW and race walkers jogging at slow velocity. Increases in RW velocity elicited an increase in both SL and step frequency (SF): Increases in SF resulted because of decreased contact time (CT) (4). They also reported that changes in speed altered gait characteristics, including joint kinematics and ground reaction forces, indicating elevated muscle force generation and mechanical work (4). Studies during high-level RW competitions have produced greater speeds than those reported during laboratory-based studies. Average RW speed of the world record during 10- and 50-km races are 4.36 ms21 (3) and a 3.89 ms21 (19), respectively. Therefore, the biomechanical aspect of gait in RW seems to play a crucial role in the energy expenditure and the movement mechanics (5,14,16). It is pertinent to understand the kinematic adaptations of changes in RW speed. Normally, within RW competitions, a slope variability ranging from level to 2% is present. Although taking kinematic
VOLUME 0 | NUMBER 0 | MONTH 2012 |

Uphill Race Walking on Treadmill


parameters into consideration can shed a light on the athletes efciency, no previous study has investigated the kinematic characteristics of RW uphill and at different speeds. Furthermore, within RW, uphill walking is used as a training method to improve cardiovascular and muscular strength such as aerobic power and shifting lactate threshold (13). However, alterations in biomechanics have not been studied with regard to this training tool. Graded walking and running were intensively studied with regard to energy expenditure decades ago (2,9). Recently, Minetti et al. (17) compared energetic cost during normal walking and running in both positive and negative slopes. Walking speed was higher (3.14 ms21) on at ground and reduced with increases in slope (17). The aim of this study was to investigate the effects of both speed and slope on kinematic parameters during RW. We hypothesized that increases in slope would alter RW footstep kinematic parameters, compared with level RW. If this hypothesis were conrmed, it would provide information on the effectiveness of slope training, with regard to altering the mechanical demand in elite-level athletes. velocity within this study, the tests were performed on a motorized treadmill (Run Race Technogym Run 500 Gambettola, Italy). The treadmill was validated and certied at 0, 2, and 7%. The treadmill was calibrated before each test according to the instructions of the manufacturer and regularly checked after the tests. As aforementioned, all the participants were in good general health conditions at the time of the study; they carried out the test during the same period wearing running shoes. The relevant data were acquired within a 7-day period, starting at 4:00 PM up to 6:00 PM under the following environmental condition: average temperature 23 C (minimum 20, maximum 26 C). All the subjects performed a standardized 10-minute warm-up, consisting of RW at 2.50 ms21 on the treadmill (15). After 5 minutes of active muscular stretching, the subjects performed the tests on the treadmill. To exclude the effect of fatigue, the experiment started at different speeds (3.61, 3.89, and 4.17 ms21) and slopes (0, 2, and 7%) with a randomized protocol (Latin Square). Five minutes of passive recovery followed after each session, in accordance with other studies (8,21). The subjects RW on the treadmill was recorded using a high-speed camera (Casio Exilim FH20) sampled at 210 frames per second. The camera was located on a 1.5-m-high tripod, 6 m from and perpendicular to the acquisition space, to be perpendicular to the subjects sagittal plane (1). The lm sequences were analyzed off-line using Dartsh 5.5Pro motion analyzing software (Dartsh, Fribourg, Switzerland). The following kinematic variables were studied, SL (meters), SF (steps per second) and CT (milliseconds): for each velocity 400 steps were sampled (20). Because the velocity of the treadmill was known, both SL and SF could be calculated. Step length was calculated as the ratio between the CT of the right and left foot, and SF calculated as the total number of right and left foot ground contacts per second (7,21). The test-retest reliability of this testing procedure was demonstrated through an intraclass correlation coefcient (ICC) and standard error of measurements (SEM) variables: SL (ICC: 0.970.98, SEM: 0.030.07 m), CT (ICC: 0.960.98, SEM : 911 milliseconds), and SF (ICC: 0.950.98, SEM : 0.80.11 steps per second).
Statistical Analyses

METHODS
Subjects

AU6

Elite race walkers (n = 12), age 23.42 6 5.62 years, height 1.76 6 0.09 m, mass 60.08 6 4.42 kg, body mass index 19.40 6 1.10 kgm22, with 9.92 6 0.72 years of training experience participated in this study. The inclusion criteria were high skill with .7 years training (all were ranked at the national level in their category, one was of an international level). None of the athletes suffered nutritional disturbances, none had any musculoskeletal injuries, and no medications or drugs expected to affect their physical performance were taken during the course of this investigation. To make the group homogeneous with regard to training conditions, none of the subjects performed any endurance strenuous activity and resistance training outside their normal endurance training protocol. A diet control was designed to eliminate the risk of any difference in the total assumption of proteins, carbohydrates and saturated and unsaturated fats; all the athletes lived together and followed a diet provided by the sports nutritionist. The macronutrient composition of the each subject diet was 60% carbohydrate (7.4 gkg21d21), 26% fat, and 14% protein (1.8 gkg21d21). The TCU Institutional Review Board for the use of Human Subjects approved the details of this study and all related informational and consent documentation before any data collection. In accordance with the Institutional Review Boards policies for use of human subjects in research, the investigator informed all the subjects as to the benets and possible risks associated with the participation in the study, and all the subjects signed a written informed consent document indicating their voluntary participation.
Procedures

Starting tests were carried out in May 2010, in the Human Performance Laboratory. To better standardize slope and

Data are presented as mean 6 SD. Linear regression analysis, using Pearsons correlation coefcients (r) was used to indicate the relationship between level and slopes and (a) SL, (b) SF, and (c) CT at each speed (3.61, 3.89, and 4.17 ms21). After the assumption of normality, veried using the Kolmogorov-Smirnov test, a 2-way ANOVA was used to determine any signicant difference in SL, SF, and CTat each speed and slope. In the case of signicant F-value, a Fishers least signicant difference post hoc test was applied. The level set for signicance was p # 0.05. The statistical analyses were performed using SPSS software (version 15, SPSS Inc., Chicago, IL, USA).

Journal of Strength and Conditioning Research

the

TM

Journal of Strength and Conditioning Research

the

TM

| www.nsca-jscr.org

TABLE 1. Effects of slope on race walking at different speeds with increasing slope (0, 2, and 7%, mean 6 SD of the kinematic variables studied). Slope (%) Level +2 Up +7 Up Speed (ms21) 3.61 3.89 4.16 3.61 3.89 4.16 3.61 3.89 4.16 Step length (m) 1.09 1.12 1.15 1.05 1.08 1.11 1.01 1.03 1.06 6 0.06* 6 0.04* 6 0.05* 6 0.07* 6 0.06* 6 0.06* 6 0.08$ 6 0.09$ 6 0.07$ Step frequency (step per s) 3.32 6 0.18 3.48 6 0.14 3.63 6 0.16 3.44 6 0.19 3.60 6 0.15 3.76 6 0.17 3.59 6 0.20* 3.77 6 0.17* 3.95 6 0.18* Contact time (ms) 302 6 288 6 276 6 292 6 278 6 266 6 279 6 266 6 254 6 15.40 11.57 11.53 15.42 11.58 11.57 15.23$ 11.44* 11.20*

AU8

Conditions that differ from level race walking and velocity are denoted in level with an *p , 0.05 or $p , 0.01 signicant differences between 0 and 2%, p , 0.05 or p , 0.01 signicant differences between 0 and 7%. At the 2%: with an *p , 0.05 or $p , 0.01 signicant differences between 2 and 0%, p , 0.05 or p , 0.01 signicant differences between 2 and 7%. At the 7%: with an *p , 0.05 or $p , 0.01 signicant differences between 7 and 2%, p , 0.05 or p , 0.01 signicant differences between 7 and 0%.

RESULTS
Step Length
T1

A signicant decrease (Table 1) was observed in SL with relation to increased treadmill slopes (r = 0.56, 95% condence interval: 20.86 to 0.02, p , 0.0001), whereas a signicant increase resulted because of the increasing velocities (r = 0.37, 0.26 to 0.78, p , 0.01). As shown in Table 1, with a 2% slope the difference compared with level was 4.01% (p , 0.05) while between level and a 7% slope, the difference increased to 9.10% (p , 0.01). Compared with 3.61 ms21, the SL was 2.9 and 5.67% longer at 3.89 and 4.17 ms21, respectively. The SL was greater during level walking at 4.17-ms21 velocity (1.15 6 0.05 m) and shorter at 7% slope at 3.61-ms21 velocity (1.01 6 0.06 m). No signicant correlation was found between the subjects height and their SL with regard to the improvement of velocity (r = 0.20, n.s.) and slope (r = 0.29, n.s.).
Step Frequency

linear with relation to the increasing speeds (3.61 , 3.89 ms21 24.44%; 3.61 , 4.17 ms21 28.42%). The decreased CT (254 6 11.2 milliseconds) at 7% slope with 4.17 ms21 contributed to the increased SF (3.95 6 0.18 Hz). Conversely, during level walking at 3.61 ms21 the greater CT (302 6 15.4 milliseconds) contributed to the decreased SF (3.32 6 0.18 seconds). Similar to SL, CT gradually decreased with increasing treadmill slopes: a decrease of 3.46% (p = n.s.) between level and 2% slope and 27.81% (p , 0.01) between level and 7% slope. The SL and SF increased with the increase in walking speeds at all treadmill slopes, and the CT decreased with regard to all speed and slope conditions (Table 1).

DISCUSSION
The study aimed to investigate the characteristics of kinematic variables, SL, SF, and CT, during 3 RW speeds and 3 different slopes (0, 2, and 7%). This study was conducted as step kinematic characteristics of RW have previously been studied at slower speeds than speeds elicited during competitions. The second important reason for this study was the acknowledgment that slopes play a crucial role in the RW training methodology. We found that SL increased gradually as the speed increased, and it decreased as the slopes increased. The SF increased linearly as walking speeds and slopes increased. Accordingly, the CT also decreased linearly with regard to the increase of speed and slope. As was expected, SL was longer during level walking than on slope walking at each walking speed. At increased incline, to maintain a constant speed, more force should be produced, which would provide a longer SL. However, increases of the SL are hindered by the elevation of the treadmill belt, in which the swing leg is forced to contact the belt in a shorter time as a result of the elevation. To maintain constant walking speed, the athlete should therefore
VOLUME 0 | NUMBER 0 | MONTH 2012 |

The SF increased signicantly with changing treadmill slopes (r = 0.50, 0.1 to 0.83, p , 0.0001) and velocities (r = 0.56, 0.02 to 0.86, p , 0.0001). As shown in Table 1, a 3.59% (p = n.s.) difference between level and 2% slope was found, which increased to 8.39% (p , 0.01) between level and 7%. Similarly to the SL, SF increased (Table 1) linearly from 4.56% (3.61 , 3.89 ms21) to 9.11% (3.61 , 4.17 ms21) with the increase in walking velocity. With relation to the slopes, compared with level, SF increased approximately 3.5% at each speed at 2% slope, whereas it increased more (;8.5%) at 7% slope in all the speeds studied.
Contact Time

The CT decreased signicantly with the increased treadmill speeds (r = 0.57, 0.860.01, p , 0.0001) and slopes (r = 0.50, 0.83 to 0.1, p , 0.0001). The decrease in CT was

Uphill Race Walking on Treadmill


increase SF and decrease CT, which would presumably also produce a shorter SL. This was supported within this study, SF increased gradually as slopes increased at each walking speed. A similar adaptation was revealed in CT. Namely, the CT became shorter when walking on slopes indicating that a constant speed can be maintained by decreasing the CT, which causes greater SF when athletes walk on slopes. Interestingly, the changes in the 3 variables occurred at the same slope progression when the treadmill belt speed was increased independently of the different kinds of slope chosen. For example, the SL increased uniformly by 0.03 m as the slope progressed from zero to 2 and 7%. Also, the SF increased by 1518 steps per second and CT decreased by 1214 milliseconds uniformly. The present results support previous observations in level RW. Hoga et al. reported that walking speed was signicantly linked to SF; however, SF did not affect speed during a 20-km race in which the RW speed ranged between 3.50 and 4.25 ms21 (11). In their study, SL was approximately 1.031.13 m for those who walked at a lower velocity and 1.25 m for the fastest athletes. In this study, we determined the SL for the same subject at 3 different speeds falling approximately within the same range of those in the experiment by Hoga et al. (11). We found a similar increase in SL, that is, 0.09 m increased from 3.61 to 4.17 ms21. Cairns et al. support the current results in which they showed an increase in both SL and SF when the selected RW speed increased from 2.89 to 3.61 ms21 (4). A 0.83-ms21 increase in velocity resulted in an 11.3% increase in SL, 13.2% increase in SF, and a 23.5% decrease in CT. Within this study, the increase in speed from the slowest to the fastest speed was 0.56 ms21 in which the SL was 5.5% longer at 4.17 ms21 than during the slowest speed. The SF was greater by 10.9% at 4.17 ms21 than at 3.61 ms21: double the increase in SL. The CT decreased 10.8% between the slowest and fastest walking speed conditions. According to Cairns et al. (4), these results may indicate that the percentage difference in the selected kinematic variables is lower in highest speeds. Furthermore, these results may indicate that SF does inuence the increase of walking speed in faster walking races more than the SL. This is supported by previous research ndings, which suggested that increases in speed are reliant upon increases in SF, once an optimal SL is achieved (4). As shown in Table 1, it appears that SL decreased and SF increased gradually with the increase of treadmill slopes. Finally, from the results of this study, it can be inferred that the biomechanical patterns used by athletes to adapt to an increased treadmill speed include increased SL, conversely during increased treadmill incline athletes SL proportionally decrease. In fact, to keep the speed constant at each slope, the SL decrease is compensated through an increase in SF. Furthermore, the reduction of the SL during slope conditions is presumably also because of a reduction of the swing phase. If we also consider that, on the slope and at increased velocities, a reduction of CT corresponds to an increase of the propulsive phase (10), it is clear that the complete step cycle is reduced. Therefore, to increase the speed on the slopes, a greater emphasis should be placed on increasing the SF, rather than the SL. This study shows that adaptations in basic biomechanical parameters may occur more at faster speeds than at slow speeds. To maintain a steady speed on different slopes, it is necessary to alter a combination of SL, SF, and CT. During uphill treadmill incline, race walkers tend to increase speed via increased SF and decreased CT so as to compensate for the reduced SL. Considering that training on slopes is used by athletes to increase performance, the current data suggest that using slopes .7% (where the greatest changes were found, compared with level) could considerably alter the neuromechanical behavior of athletes. Therefore, the changes elicited, within all studied biomechanical parameters, at 7% incline could negatively inuence RW techniques. Consequently, this could increase energetic cost and weigh heavily on the SL/SF ratio on race performance. Conversely, this study found less evident changes at 2% incline at faster speeds; therefore, these conditions may positively inuence RW performance without negatively altering the RW technique.

PRACTICAL APPLICATIONS
In RW competition, the average number of steps is reportedly 27,786 during 10-km races (3) and 52,661 during 50-km races (19) (estimated in relation to their race performances at 4.36 and 3.89 ms21 and reported SFs). Small decreases in SL could produce greater loss of distance over the course of a race, which would adversely affect race performance. Therefore, it is necessary for coaches to know the SL decrease incurred at different slopes, and at a constant speed, to estimate an athletes efciency level. For example, by knowing the speed of the treadmill (the best performance RW race speed) during level and the number of steps per minute, SL can be estimated (an approximation) with equation 1.

Step Length m Speed at level m s1 360 min=SF steps=second: 1


Furthermore, coaches can use changes in slope and SF as a coaching method. For example, via knowledge of speed and SL during level and on a slope, coaches can calculate the optimal SF through the equation for optimal SF equation 2 (21). The athlete can therefore be coached and trained to perform a precise number of steps per minute to produce the desired SL (21). This would be benecial to an athlete during race situations in which athletes could use their knowledge of optimal SF to maintain speed during long distance events.

Optimal step frequency Speed in slope m s1 = SL at level m360 min:2 ACKNOWLEDGMENTS


The authors are grateful to the Gruppo Sportivo Fiamme Gialle in particular to Coach Patrizio Percesepe, for supporting

Journal of Strength and Conditioning Research

the

TM

Journal of Strength and Conditioning Research


this research. J. Padulo and G. Annino have contributed equally to this article.

the

TM

| www.nsca-jscr.org

12. Hoga, K, Ae, M, Enomoto, Y, Yokozawa, T, and Fujii, N. Joint torque and mechanical energy ow in the support legs of skilled race walkers. Sports Biomech 5: 167182, 2006. 13. La Torre, A. Lallenamento di Ivano Brugnetti per la 20 km di Atene. Edizione Fidal: Atletica Studi 2010. 14. Margaria, R, Cerreteri, P, Aghemo, P, and Sassi, G. Energy cost of running. J Appl Physiol 18: 367370, 1963. 15. Matsas, A, Taylor, N, and McBurney, H. Knee joint kinematics from familiarized treadmill walking can be generalized to overground walking in young unimpaired subjects. Gait Posture 11: 4653, 2000. 16. Minetti, AE, Ardigo, LP, and Saibene, F. Mechanical determinants of the minimum energy cost of gradient running in humans. J Exp Biol 195: 211225, 1994. 17. Minetti, AE, Moia, C, Roi, GS, Susta, D, and Ferretti, G. Energy cost of walking and running at extreme uphill and downhill slopes. J Appl Physiol 93: 10391046, 2002. 18. Murray, MP, Guten, GN, Mollinger, LA, and Gardner, GM. Kinematic and electromyographic patterns of Olympic race walkers. Am J Sports Med 11: 6874, 1983. 19. Nizhegorodov, D. Word record race walkers 50 km. Available at: http://www.iaaf.org/statistics/records/inout=o/discType=6/disc= 50KR/detail.html. Accessed September 13, 2011. 20. Owings, TM and Grabiner, MD. Measuring step kinematic variability on an instrumented treadmill: How many steps are enough? J Biomech 36: 12151218, 2003. 21. Padulo, J, Annino, G, Migliaccio, GM, DOttavio, S, and Tihanyi, J. Kinematics of running at different slopes and speeds. J Strength Cond Res. In Press. Doi: 10.1519/JSC.0b013e318231aafa. 22. Vernillo, G, Agnello, L, Drake, A, Padulo, J, Piacentini, MF, and Torre, AL. An observational study on the perceptive and physiological variables during a 10,000-m race walking competition. J Strength Cond Res. In Press. Doi: 10.1519/JSC.0b013e318242a33c.

REFERENCES
1. Belli, A, Rey, S, Bonnefoy, R, and Lacour, JR. A Simple device for kinematic measurements of human movement. Ergonomics 35: 177186, 1992. 2. Bobbert, AC. Energy expenditure in level and grade walking. J Appl Physiol 15: 10151021, 1960. 3. Burayev, V. Word record race walkers 10 km. Available at: http:// www.iaaf.org/statistics/records/inout=o/discType=5/disc=10KW/ detail.html. Accessed September 13, 2011. 4. Cairns, MA, Burdett, RG, Pisciotta, JC, and Simon, SR. A biomechanical analysis of racewalking gait. Med Sci Sports Exerc 18: 446453, 1986. 5. Cavagna, GA and Franzetti, P. Mechanics of competition walking. J Physiol 315: 243251, 1981. 6. Cavagna, GA, Thys, H, and Zamboni, A. The sources of external work in level walking and running. J Physiol 262: 639657, 1976. 7. Cavanagh, PR, Pollock, ML, and Landa, J. A biomechanical comparison of elite and good distance runners. Ann N Y Acad Sci 301: 328345, 1977. 8. Cavanagh, PR and Williams, KR. The effect of stride length variation on oxygen uptake during distance running. Med Sci Sports Exerc 14: 3035, 1982. 9. Dill, DB. Oxygen used in horizontal in grade walking and running in the treadmill. J Appl Physiol 20: 1922, 1965. 10. Gottschall, JS and Kram, R. Energy cost and muscular activity required for leg swing during walking. J Appl Physiol 99: 2330, 2005. 11. Hoga, K, Ae, M, Enomoto, Y, and Fujii, N. Mechanical energy ow in the recovery leg of elite race walkers. Sports Biomech 2: 113, 2003.

AU7

VOLUME 0 | NUMBER 0 | MONTH 2012 |

You might also like