Muscle–tendon interaction and EMG profiles of world class endurance runners during hopping
- First Online:
- Cite this article as:
- Sano, K., Ishikawa, M., Nobue, A. et al. Eur J Appl Physiol (2013) 113: 1395. doi:10.1007/s00421-012-2559-6
- 1.8k Views
The present study examined the muscle–tendon interaction of ten international level Kenyan runners. Ultrasonography and kinematics were applied together with EMG recordings of lower limb muscles during repetitive hopping performed at maximal level. The ten Kenyans had longer gastro Achilles tendon at rest (p < 0.01) as compared with ten control subjects matched in height. Conversely, the stretching and shortening amplitudes of the tendinous tissues of the medial gastrocnemius (MG) muscle were significantly smaller in the Kenyans than in controls during the contact phase of hopping. This applied also to the fascicle length changes, which were smaller and more homogeneous among Kenyans. These limited musculo-tendinous changes resulted in higher maximal hopping height and in larger power despite their reduced body weight. The associated finding of a greater shortening to stretching ratio of the MG tendinous tissues during contact could imply that the Kenyan MG muscle–tendon unit is optimized to favor efficient storage and recoil of elastic energy, while operating at optimal muscle fascicle working range (plateau region).
KeywordsStretch-shortening cycleElastic energyRunning economyUltrasoundStiffness
The recent years have demonstrated the great success of the African runners in general and that of the Kenyan runners in particular. The success of the Kenyan runners has been so dramatic that scientists and practitioners alike wondered whether this success is not only as a result of superior training system and athletic selection, but also due to the genetic make-up specifically providing the runners with exceptional preconditions for high level performance. Contrasting the expectations, the reliable studies conducted on Kenyan boys and Kenyan elite runners have not shown that they would be markedly different in many basic physiological parameters from their e.g., European counterparts (Larsen 2003). Nevertheless, they seem to have high maximal aerobic capacity associated with exceptionally good running economy (Saltin et al. 1995a, b). This finding of high running economy could not be explained by any of the histochemical and/or biochemical parameters measured from the muscle biopsy samples. Consequently, it was a clear conclusion of these reports of Saltin et al. (1995b) that one of the possible reasons for both high mechanical efficiency and improved running economy could be the special biomechanical make up of these runners. In particular, their thin lower leg thickness was found as a more crucial factor than their slender body shape for running economy. The question can be asked whether the Kenyan runners possess also a particular type of triceps surae muscle–tendon complex, which could favor efficient storage and use of elastic energy in Achilles tendon during natural forms of locomotion.
To examine this hypothesis, basic information is first needed regarding the structural and functional specifics of the lower leg among Kenyan elite runners. As suggested by Saltin et al. (1995b), the biomechanical characteristics may play a decisive role in our efforts to seek reasons for the superior “biomechanics” of the Kenyan athletes, especially middle and long distance runners, including the marathon ones. However, before this reasoning can reach the level of mechanistic factors, the first step and effort should be made to test the performance and the muscle–tendon interaction of the Kenyan runners in hopping activity, which loads the Achilles tendon individually and heavily (for references see Komi and Ishikawa 2007). This task is simple to perform, but require also well-coordinated activation of both agonist and antagonist muscles. In addition the hopping task should be less sensitive than running to the possible influence of the thinner lower leg of Kenyans. Further requirements should be the easy use of ultrasonography for undisturbed recording of muscle–tendon interaction. We chose maximal repetitive hopping for this purpose. During the functional phase of hopping (from pre-contact to push-off), the muscle actions are expected to follow the basic concept of the stretch-shortening cycle (SSC), in which the muscles are optimally preactivated, actively stretched during the braking phase and subsequently shortened during the push-off. Hopping as performed by humans is also characterized as a locomotor activity, in which activation of the muscles is variable and under the influence of both central and reflex control. For example, during normal hopping the preactivated lower limb extensor muscles are very active during the braking phase. The final push-off phase can then be variably activated depending on the task, intensity of effort and the muscles in question (for reviews, see Komi and Nicol 2011; Ishikawa and Komi 2008).
The present study therefore focused on the muscle–tendon interaction of Kenyan runners belonging to the elite category as measured by their success in international achievement as medal winners or running records. It was hypothesized that the Kenyan runners would present a specific pattern of SSC function, which could explain partly their performance excellence. To help testing this hypothesis, the repetitive maximal hopping tasks were monitored with simultaneous recordings of ultrasonography, kinematics, and with EMG recordings of the relevant lower limb muscles.
Kenyan (n = 10)
Control (n = 10)
21 ± 4
25 ± 4
1.75 ± 0.06
1.74 ± 0.05
Body mass (kg)
57.9 ± 5.1
71.3 ± 5.0*
18.9 ± 1.5
23.4 ± 0.9**
Shank length (m)
0.40 ± 0.02
0.43 ± 0.02
Gastro Achilles tendon length (mm)
264.2 ± 24.5
196.6 ± 12.8**
MG MTU (mm)
437.3 ± 27.4
468.0 ± 18.7
MG fascicle length (mm)
54.2 ± 4.0
56.8 ± 9.4
Pennation angle (°)
20.0 ± 2.1
21.8 ± 0.9*
MG Achilles tendon structures (mm)
395.6 ± 24.9
419.9 ± 11.0*
Contact time (s)
0.187 ± 0.029
0.215 ± 0.015*
Stretching time (s)
0.078 ± 0.020
0.101 ± 0.011**
Shortening time (s)
0.109 ± 0.021
0.114 ± 0.006
Flight time (s)
0.451 ± 0.044
0.352 ± 0.047**
Rebound height (m)
0.251 ± 0.048
0.154 ± 0.040**
Vertical stiffness (kN m−1)
21.7 ± 5.2
22.2 ± 3.0
Jumping power (W)
After several minutes of habituation to repetitive bilateral hopping exercise, the subjects took a standing position for measurements of the following parameters: shank length, medial gastrocnemius (MG) fascicle length and its pennation angle as well as the gastro Achilles (GA) tendon length from its insertion point on the calcaneus to the junction with Achilles tendon, medial and lateral gastrocnemii muscles (gastro Achilles tendon junction). The MG fascicle lengths and their pennation angles as well as GA tendon length in the upright position were measured by ultrasonography (linear array probe with scanning frequency of 13 MHz, Hitachi-Aloka Inc., Japan). The GA tendon length was tape measured along the line from the calcaneus insertion point to the gastro-Achilles tendon junction. After these anthropometric measurements and hopping practices, the subjects were asked to hop with gradually increasing height from low to maximum so that the final recording included at least five stable maximal hops. Subjects were instructed to hop with maximal effort, active plantar flexion and limited knee flexion throughout the hop test. They were also instructed to look straight ahead.
Kinematics and ultrasonography
All hops were recorded with a high-speed video camera at 240 fps (HDR-CX550V, SONY, Japan) from the left side perpendicular to the line of motion. Reflective markers were placed on trochanter major, the approximate center of rotation of the knee, lateral malleolus, calcaneus and fifth metatarsal head. These points were then digitized automatically and filtered with Butterworth fourth-order filter (cut-off frequency 10 Hz) using Motus software (Peak Performance Inc, USA) to calculate knee and ankle joint angles.
RNCC is the averaged normalized correlation coefficient for images during the contact phase. T is the total images during the contact phase of hopping. The i and j are X–Y coordinates of image. The calculated size (i, j) of the image area is a resolution of 424 × 326 pixels. The Ft(i,j) and St(i,j) are image brightness for the first and second steps, respectively. The correlation coefficient for images during the contact phase was on an average 0.90 ± 0.04.
Electromyographic (EMG) recordings
EMG activity was recorded from tibialis anterior (TA), soleus (SOL) and MG muscles of the left leg using bipolar surface active electrodes (electrode material: Ag/AgCl, electrode shape: parallel-bar electrodes, size 2 mm width × 9 mm length, inter-electrode distance: 10 mm) (NM-512G, Nihon Koden, Japan) with a multi-telemeter AD converter system (WEB-5000, NIHON KODEN, Japan) (sampling frequency 1 kHz, input impedance >10 MΩ, common mode rejection ratio >80 dB, time constant 0.03 s). The electrode placement followed the SENIAM guidelines (Hermens et al. 2000) as accurately as possible. The MG electrodes were placed slightly lateral to the muscle midbelly to accommodate the ultrasound probe. Before electrode placement, the skin was cleaned with alcohol. The EMG signals were band-pass filtered (20–450 Hz).
To synchronize the above kinematics, ultrasound and EMG data, all recordings were triggered by a foot switch sensor (SEN-08713, FlexiForce, USA) positioned under the ball part of the left foot.
During hopping, the stretching and shortening phases were determined based on the peak LMTU stretch during the contact phase. The pre-activation (PRE) phase was defined as the 100 ms period preceding ground contact (Komi et al. 1987). For each phase of hopping, the rectified time course EMG signals were integrated and then averaged individually (aEMG). In addition, the first 25 ms period of the stretching phase (early 25 ms) was considered to be an additional preprogrammed phase that could not be influenced by any afferent signal initiated at ground impact (Avela et al. 1996; McDonagh and Duncan 2002). Amplitudes of MG LFa, LTS and LMTU changes during stretching and shortening phases were calculated from the time of initial foot contact to the peak LMTU stretch and from this to toe-off, respectively.
Values are presented as means and standard deviations. Statistical analysis was performed using the paired t test for comparing Kenyan and control groups, unless stated otherwise. An one-way repeated-measures ANOVA, with a post hoc Fisher’s least significant difference multiple comparison, was used to compare the different phases of hopping for each group. If the normality test failed for the whole data, the Wilcoxon’s signed-rank test and Spearman’s rank correlation coefficient were used. The ANOVA for repeated measurements on two factors and post hoc Bonferroni’s were used to test the main effects of length (MG LMTULTS and LFa) and groups as well as interactions on different parameters. Relationships between variables were investigated using Pearson’s product–moment correlation coefficient. The probability level for statistical analysis was set at p < 0.05.
The anthropometric data of the two groups were similar in body height, shank length, MG LMTU and LFa (Table 1). The Kenyans had longer GA tendon length (p < 0.01) at rest, but lower MG pennation angle. The body mass index of the Kenyans was also clearly lower than that of the controls (p < 0.01). In maximal hopping, the contact time was shorter (p < 0.05) and the flight time longer (p < 0.01) for the Kenyans (Table 1). Consequently, their maximal rebound height (hmax) was 73.7 ± 59.2 % higher (p < 0.01), which then translated into a 25.7 ± 31.4 % greater maximal jumping power (Pmax) (p < 0.05) when compared with the controls. None of these two functional parameters was significantly correlated to the GA tendon length at rest within each subject groups: Kenyans (hmax: r = −0.23, p = 0.52; Pmax: r = 0.01, p = 0.97, N = 10) and controls (hmax: r = −0.34, p = 0.34; Pmax: r = −0.20, p = 0.58, N = 10), respectively. When the subjects were combined together, the r values remained low, but reached the statistical significance for both parameters (hmax: r = 0.58, p = 0.01; Pmax: r = 0.45, p = 0.04, N = 20).
Originally this study was designed as an attempt to examine if the elite Kenyan middle and long distance runners possess “such biomechanical qualities” that would help identifying reasons why they are consistently demonstrating performance of excellence in top level international competitions including the World Championship and Olympic Games. Saltin et al. (1995b) referred specifically to the possibility that long-term high-intensity training might have resulted in favorable adaptation in the Achilles tendon to store and utilize elastic energy. In the course of the present hopping testing of the Kenyan athletes, it became soon evident that the true answer to the challenging question of performance excellence would not come easily and not evidently from the first study of this kind. However, it was rewarding to observe that the available research tools (ultrasound unit and EMG recording system together with high-speed cinematography) were sufficient to allow the primary hypothesis, the “muscle–tendon interaction” as expressed above, to remain tenable.
From the structural points of view, it was shown that the Kenyan endurance runners had very long GA tendons (expressed in both absolute and relative values) as compared to the young healthy controls. The structural muscle–tendon characteristics of the entire population correlated weakly, however, with the maximal hopping performance. This observation should not be such a surprise as the Achilles tendon moment arm (Scholz et al. 2008) and the tendon–fascicle interaction are expected to be more relevant during hopping than the isolated length of Achilles tendon.
In maximal hopping, ultrasound analysis showed for both groups a typical SSC type of MG fascicle–tendon behavior, with Kenyans having the smaller MG fascicle length changes (Figs. 1, 2, 3) and the higher shortening/stretching ratio of MG LTS (Fig. 4). In normal forms of ground locomotion, the reported range of MG LTS stretch varies from 5.8 to 8.3 % of their total length (Lichtwark and Wilson 2005, 2006; Ishikawa et al. 2007). The current values of 5.0 versus 8.1 % of MG LTS stretch amplitudes for the Kenyans as compared to the controls are in line with these estimates. Interestingly, however, the Kenyans presented an overall smaller range of MG LTS and LFA changes during the ground contact (Figs. 2, 3). Consequently, one could have expected that they would not be so superior over the controls in hopping height and power. The greater shortening to stretching ratio for the MG LTS in Kenyans (Fig. 4) implies an efficient use of tendinous elasticity, especially for a better hysteresis. This notion challenges the often proposed assumption that longer stretch should automatically lead to greater storage and recoil of elastic energy, but is in good agreement with most of the animal studies in hopping (for a review see Alexander 2003). It must be noted at this specific point that the tendons are not operating in isolation but instead in a continuous control by the activation of the muscle and under the influence of external loading, such as impact in the present study. The present protocol was not ideal in this regard as we measured hopping in the maximal condition only. However, some remarks can be made regarding the fascicle–tendon interaction among the Kenyan runners of the present study.
Based on the reports of the EMG profiles during the SSC performance of the present kind (Komi and Nicol 2011), the triceps surae muscle should be very active during the braking (stretching phase) and probably less or more variably active in the subsequent concentric (push off) phase. This was also the case in the present study, and particularly for the Kenyans (Fig. 6), whose shortening to stretching SOL aEMG ratio was about 40 % smaller than in the controls. This was coupled with the strong MG EMG activity of the braking (eccentric phase), as shown by the ratio MG EMG stretching/preactivation, which was significantly greater among the Kenyan group of subjects as compared to the controls (Fig. 6). Associated with a higher rebound performance, this activation pattern gives additional support to the hypothesis of a high Kenyan efficacy in the use of the elastic energy recoil from the tendinous structures. The present results cannot, however, be applied directly to the maximal drop jump, which was used by McBride et al. (2008) and resulted in their conclusion that increased levels and distribution of EMG profiles of the agonist muscles for the preactivation and eccentric phases are very relevant for enhancing the hopping performance. It would be attractive to interpret that these proposed mechanisms could imply that the shifts in the activation during the preactivation—eccentric phases could result in favorable conditions to make the Achilles tendon stiffer than what could be expected from the long GA tendon of the Kenyans. Again, the present findings can give only partial support for this proposal and it needs to be looked not only from the EMG profiles, but especially simultaneously with the length changes in the fascicles and tendon.
It has been reported that the working range of sarcomeres during running and jumping exercises is near the ascending limb of the plateau region (Fukunaga et al. 2002; Ishikawa et al. 2007). With increasing intensity (e.g., running speed), however, the working length range of the MG fascicle can shift from the plateau region to the ascending limb (Ishikawa and Komi 2007). In the present study, the LFa during preactivation phase was on the average shorter in Kenyans than in controls (Fig. 5). The subsequent stretching phase was characterized for the Kenyan group by a significant increase in MG activity associated with smaller muscle and tendon length variations (Figs. 2, 3). The MG LFa working range during the stretching phase of hopping in the Kenyans remained also closer to the resting length measured at rest as compared to the controls (Fig. 2). These observations suggest that the working length of the MG fascicles could be closer to the plateau region of force–length relation in the Kenyans than in the controls and thus, may contribute to increase the force output of cross-bridge formation effectively. In addition, the smaller fascicle length changes and preactivation of MG muscles but shorter stretching phase time indicate that the Kenyans are stiffer than the controls. These certifications will be needed in future studies.
The above discussion could reflect possible mechanistic events and their reasons for excellent performance among the Kenyan subjects of the study. Some words of caution should, however, be exercised before drawing the final conclusions. In this regard certain methodological issues should be addressed. Relevant at this point is selection and use of the CTRL group in the present study. One may objectively state that the comparison of the Kenyan top class athletes with active Caucasian controls only is not meaningful to answer the question whether or not the neuro-mechanical behavior of the Kenyans may contribute to their general SSC efficacy. However, the mechanical properties of the Achilles tendon of Caucasian long distance runners are not different from those of moderate active subjects (Arampatzis et al. 2007). Therefore, the comparison of the present study can be considered as valid. Nevertheless, the body mass of Kenyans was significantly lower than controls. Although their hopping power was significantly higher, their lower body mass may still have contributed to the observed smaller contact times and muscle–tendon behavior during contact phase. These limitations should be considered in the future studies.
In summary, the study showed that Kenyan runners possessed a very long GA tendon as compared to the young controls. This long tendon length at rest was only slightly related to hopping performance in the whole group comparison, and not at all in the single groups. Suggesting a high efficacy in the use of the recoil of elastic energy from the MG tendinous structures, the Kenyans reached higher hopping performance despite an overall smaller range of fascicle and tendinous length changes during the contact phase. This phenomenon was accompanied by shorter stretching and total contact times as well as higher stretching to preactivation EMG ratio. It was also observed that in the push-off phase, Kenyans showed greater tendon recoil ratio (greater MG LTS shortening to stretching ratio) as compared to controls. This specific muscle–tendon characteristic in the Kenyans is likely to have contributed to the effective SSC performance and it may imply that there can be a specific fascicle–tendon behavior to utilize elastic energy effectively.
This work was supported by MEXT/JSPS KAKENHI Grant Number 23700756 and 23500729.