Share this article on:

Changes in Running Kinematics, Kinetics, and Spring-Mass Behavior over a 24-h Run


Medicine & Science in Sports & Exercise: May 2011 - Volume 43 - Issue 5 - pp 829-836
doi: 10.1249/MSS.0b013e3181fec518
Applied Sciences

Purpose: This study investigated the changes in running mechanics and spring-mass behavior over a 24-h treadmill run (24TR).

Methods: Kinematics, kinetics, and spring-mass characteristics of the running step were assessed in 10 experienced ultralong-distance runners before, every 2 h, and after a 24TR using an instrumented treadmill dynamometer. These measurements were performed at 10 km·h−1, and mechanical parameters were sampled at 1000 Hz for 10 consecutive steps. Contact and aerial times were determined from ground reaction force (GRF) signals and used to compute step frequency. Maximal GRF, loading rate, downward displacement of the center of mass, and leg length change during the support phase were determined and used to compute both vertical and leg stiffness.

Results: Subjects' running pattern and spring-mass behavior significantly changed over the 24TR with a 4.9% higher step frequency on average (because of a significantly 4.5% shorter contact time), a lower maximal GRF (by 4.4% on average), a 13.0% lower leg length change during contact, and an increase in both leg and vertical stiffness (+9.9% and +8.6% on average, respectively). Most of these changes were significant from the early phase of the 24TR (fourth to sixth hour of running) and could be speculated as contributing to an overall limitation of the potentially harmful consequences of such a long-duration run on subjects' musculoskeletal system.

Conclusions: During a 24TR, the changes in running mechanics and spring-mass behavior show a clear shift toward a higher oscillating frequency and stiffness, along with lower GRF and leg length change (hence a reduced overall eccentric load) during the support phase of running.

Université de Lyon, Saint Etienne, FRANCE; and Laboratoire de Physiologie de l'Exercice - LPE EA4338, Saint-Etienne, FRANCE

Address for correspondence: Jean-Benoit Morin, Ph.D., Médecine du Sport-Myologie, Laboratoire de Physiologie de l'Exercice (EA4338),CHU Bellevue,42055 Saint-Etienne cedex 2, France; E-mail:

Submitted for publication May 2010.

Accepted for publication September 2010.

During running, the mechanical behavior of the musculoskeletal structures of the legs is often described as that of a spring loaded by the runner's mass, constituting the "spring-mass model" (3,30). This model has been a widely used paradigm for describing and studying the mechanics and energetics of bouncing and running gaits (10,16,18,22,38). The model consists of a point mass supported by a single massless linear "leg spring," and the main mechanical parameter studied when using the spring-mass model is the stiffness of the leg spring, defined as the ratio of the maximal force in the spring to the maximum leg compression at the middle of the stance phase (16). Moreover, although it does not correspond to any physical spring, the vertical stiffness is used to describe the vertical motion of the center of mass (COM) during contact (16,30) and is defined as the ratio of the maximal force to the vertical displacement of the COM as it reaches its lowest point, that is, the middle of the stance phase.

Runners have been shown to consistently maintain their leg stiffness constant despite changes in running velocity (22,30) or gravity level (22). Even when running or forward hopping on surfaces with variable stiffness, subjects have been shown to adjust their leg stiffness to maintain the overall stiffness of the system (surface + leg) and their support mechanics close to constancy (18). These adjustments are almost instantaneous because they occur right from the first step on a changing surface (17). Further, this capacity to regulate the neuromuscular system to adjust leg stiffness to external changes in running conditions has been observed in other species of bipeds or quadrupeds, ranging from kangaroo rats to horses (15). This demonstrates the ability of running bodies to adjust the overall stiffness of their lower limbs toward constancy. To our knowledge, the only external change in the conditions of level running that has been shown to be systematically associated with a change in leg stiffness is when step frequency is changed (16,38) and more precisely when contact time is the variable responsible for this change in frequency or changed alone (38).

Beyond these mechanical external changes, one may wonder how internal changes in the physiological status of the neuromuscular system could affect leg stiffness. Because the spring-mass behavior of biped running locomotion is allowed by a complex neuromuscular system (14), the alterations induced by fatigue could be expected to modify its properties and to induce regulations, similar or not to those observed under external changes in running conditions. For instance, Kuitunen et al. (26) showed that joint stiffness (knee and ankle) was modified after exhausting stretch-shortening exercises consisting of performing rebound jumps until exhaustion. When broadening the focus of study to running and the overall lower limb stiffness, studies have been undertaken about sprint running (e.g., repeated 100-m runs (35)), but very little is known about how the mechanical behavior of the spring-mass system changes (if changing) in extreme fatigue conditions induced by ultraendurance runs.

When considering fatigue as an appropriate paradigm to study how changes in the physiological status of lower limbs neuromuscular system may affect running mechanics, the literature is limited to very simple mechanical parameters and most often the temporal characteristics of the running step: decrease in step frequency (1,4,6,13,20,24), increase in contact time (1), or no changes in these parameters (19,34). It was also reported an increase in COM downward displacement during contact (33) and an increase in ground reaction forces (GRF) and impact (12) or a decrease in impact with the constant peak vertical GRF (VGRF) (20). To our knowledge, only Dutto and Smith (13) and Slawinski et al. (40) studied changes in spring-mass behavior with fatigue but during rather short-lasting running exercises (∼57 and ∼6 min, respectively). Dutto and Smith (13) observed a decrease in both vertical and leg stiffness (mainly caused by a higher downward displacement of the COM during contact), whereas Slawinski et al. (40) observed no change in vertical and leg stiffness or any other mechanical parameter of the running step before and after exercise.

However, all these studies only considered exhausting but short-lasting runs, through typical "time-to-exhaustion" protocols in which subjects are required to run for as long as possible: typically 5-30 min (1,4,6,19,20,33,34) and more rarely 60 min (24) at a given velocity, set as a percent of that corresponding to their respective maximal aerobic capacity.

Although one cannot deny that these conditions allow to explore fatigue, they do not correspond to long or very long running efforts, such as the marathon or those longer than a marathon (ultramarathon distances). Very few studies focused on marathon and with little focus on running mechanics (21,27,39). In addition, distances longer than a marathon have not been investigated to our knowledge, although this type of extreme exercise is increasingly popular in Europe and in several other countries (United States, South Africa, Japan, Korea, etc.). This is rather intriguing because they may represent a good model as they induce a heavier and different load on the musculoskeletal and neuromuscular systems. In particular, it has been shown that maximal voluntary force is reduced to a great extent after a 24-h treadmill run (24TR) (29), which could potentially influence lower limb stiffness. Therefore, we thought of interest to study whether and how ultralong-distance running induced changes in running mechanics by the spring-mass model.

In a recent case study (32), we observed that after a 161-d, 8500-km run, an experienced ultraendurance runner ran with a significantly higher step frequency, reduced aerial time without change in contact time, and lower maximal vertical force and loading rate (LR) at impact. This has been interpreted as a smoother and safer running pattern used by this runner to limit the overall loading experienced at each step and/or the associated painful consequences after such an extreme running feat. We sought to determine whether similar modifications also occurred during an "acute" extreme running exercise of 24TR, which induced large impairments in the neuromuscular function (29).

In this study, the authors showed substantial impairments in the neuromuscular function and a strength loss (maximal voluntary contraction (MVC)) as high as ∼40% for knee extensors and ∼30% for plantarflexors in trained ultraendurance runners. Given that leg stiffness represents the overall behavior of the lower limb and integrates both the mechanical characteristics of its musculotendinous structures and the neuromuscular mechanisms of actions of the various muscle groups involved (14), the abovementioned alterations in the neuromuscular and force production functions may reasonably be expected to induce changes in the spring-mass behavior, either directly or indirectly (through runners modified running pattern). If so, running kinematics and kinetics may reflect and help describe accurately how these changes in spring-mass behavior and leg stiffness occur during the 24TR.

Our aim was therefore to investigate the changes in running kinematics, kinetics, and in spring-mass behavior of the lower limbs over a 24TR. These variables were measured and compared before and after exercise but also every 2 h to better describe how changes occurred within this 24-h period.

Given i) the results of previous studies, a higher step frequency after the Paris-Beijing run (32) and a marathon (27), an impairment of the neuromuscular function of force production of lower limb extensors during a 24TR (29), a tendency put forward after a marathon of "loss of tolerance to impact during the belt contact" (39), and a reduced VGRF after the Paris-Beijing run (32) and ii) the fact that step frequency, stiffness, and length change of the leg spring are inherently related in running (e.g., (30,38)), we hypothesized that the spring-mass system would be regulated toward a higher leg stiffness and oscillating frequency, along with reduced downward displacement of the COM, leg compression, and vertical forces during the support phase. Although it is almost exactly the opposite of what was observed during short-lasting exhausting runs, this would likely allow subjects to attenuate the overall load faced by their musculoskeletal system at each step in such extreme running conditions.

Back to Top | Article Outline


Back to Top | Article Outline

Subjects and experimental protocol.

Ten healthy men (mean ± SD: age = 40.4 ± 6.5 yr, body mass = 75.0 ± 7.9 kg, height = 176.9 ± 5.8 cm, body fat = 17.9% ± 4.9%, V˙O2max = 52.8 ± 5.7 mL·min−1·kg−1) participated in this experiment, which was a part of a larger project initially involving 14 subjects. However, only 12 were able to complete the 24TR, and among them, 2 were not able to complete all the mechanical measurements at the imposed velocity. Subjects were experienced ultraendurance runners (they reported more than 15 yr of training in running on average), and all of them had completed at least one 24-h or 100-km race. For full details on the physiological characteristics of these subjects, see Millet et al. (31). Written informed consent was obtained from the subjects, and the study was conducted according to the Declaration of Helsinki II, approved by the local ethics committee (CPP Sud-Est 1, France) and registered on (NCT 00428779).

The subjects reported to the laboratory approximately 3 wk before the 24TR for a first session, during which they underwent a complete medical examination and were fully informed about the experimental procedures. They also confirmed having run at least three times on a treadmill during the year preceding the protocol and were familiarized with the treadmill dynamometer used for mechanical measurements.

For the second session, the subjects reported to the laboratory to perform the 24TR. After appropriate welcoming and preparation, subjects underwent a muscle biopsy aiming at collecting a sample (∼120 mg) of their vastus lateralis muscle under local anesthesia using a percutaneous technique (for details, see Morin et al. (36)).

After a 2-h rest, the 24TR started (time of the day around 5:00 p.m.), and subjects ran on a calibrated level motorized treadmill (Gymrol S2500, HEF Tecmachine, Andrézieux-Bouthéon, France, or Proform 585 Perspective, Health & Fitness Inc., Logan, UT) at a freely chosen pace. Subjects were individually chaperoned by an experimenter who ensured the required pace setting, the appropriate cooling of subjects, and their food and drink intakes (meals with carbohydrates and energy bars and drinks, ad libitum), and wrote down subjects' exact time and running pace over the 24TR (for accurate performance analysis).

Just before, every 2 h, and at the end of the 24TR, subjects ran at 10 km·h−1 for 60 s on a treadmill dynamometer (HEF Tecmachine; for details, see Belli et al. (2)). Stride mechanics were studied from a 10-s sample of data recorded between the 50th and 60th second. As the present study was part of a larger experiment, subjects also underwent physiological and neuromechanical tests/samples (before, every 4 h, and after 24TR), which are fully described by Martin et al. (29) and Millet et al. (31) and which represented experimental "pit stops" of ∼25 min every 4 h. All mechanical parameters were averaged more than 10 consecutive steps, allowing us to calculate the coefficient of variation (CV in %) as the ratio of the SD to the mean of the parameter for the 10 consecutive steps analyzed. This coefficient was further used to test potential effects of the 24TR on the variability of the running step.

Back to Top | Article Outline

Running kinematics and kinetics.

Mechanical parameters were measured for each step (the period from the onset of one foot contact to the onset of the contralateral foot contact) using the treadmill dynamometer with anteroposterior and VGRF data sampled at 1000 Hz. Contact (tc) and aerial (ta) times were measured from VGRF(t) signals, expressed in seconds, and used to compute runners' step frequency f = 1/(tc + ta).

The intensity of impact at the onset of each contact phase was quantified through the impact LR (in body weight per second), which corresponded to the maximal value of the time derivative of vertical force signals within the first 50 ms of the contact phase (11), whereas Fmax was the intensity of the active force peak (expressed in body weight), that is, the maximal value of VGRF during contact.

Back to Top | Article Outline

Spring-mass characteristics.

A spring-mass model paradigm (for details, see Blickhan (3) and McMahon and Cheng (30)) was used to investigate the main mechanical parameters characterizing the lower limbs behavior during running. Two important assumptions of this model are that i) the leg length at the moment of ground contact is equal to the initial (standing) length and ii) the horizontal motion and displacement of the COM are equivalent before and after mid-stance. According to this model, the stiffness of the leg spring (kleg, in kN·m−1) was calculated from F(t) measurements as kleg = FmaxL, with ΔL as the maximum leg spring compression (m) calculated from values of initial leg length L (great trochanter to ground distance in a standing position), running velocity v (m·s−1), contact time tc (s), and vertical maximal downward displacement of the COM during contact Δz, as per (16,22,39)

with Δz being determined by double integration of the vertical acceleration of the COM over time, as proposed by Cavagna (7).

However, during level running, the point of force application is not a fixed point over a typical stance phase (28), and it has been shown that the distance of the point of force application translation (PFAT) should be taken into account when computing ΔL to increase the accuracy of the spring-mass model for describing the lower limbs mechanics during human running (5). Leg length variation during contact was therefore calculated on the basis of equation 1, incorporating the distance of PFAT into the "traditional" planar spring-mass model:

The distance of PFAT (d) was shown to be 0.157 ± 0.006 m in subjects running from 1.5 to 5.0 m·s−1 (28), and this value was estimated to be equal to approximately 18% of the mean leg length of these subjects (5). As we could not measure d in the present study, this parameter was assumed to be d = 0.18 L. Therefore, leg stiffness was calculated as kleg = FmaxL. Further, vertical stiffness was calculated as kvert = Fmaxz and expressed in kilonewtons per meter.

Back to Top | Article Outline

Data analysis and statistics.

Descriptive statistics are presented as mean ± SD. Normal distribution of the data was checked by the Shapiro-Wilk normality test, and the mechanical variables studied were compared before, every 2 h, and after 24TR using a one-way (time), repeated-measures ANOVA. When warranted, the Student-Newman-Keuls post hoc tests were performed to identify which conditions differed. The importance of the differences found between pre-24TR and post-24TR conditions was assessed through the effect size and Cohen's d coefficient (9). The interpretation of the effect size was as follows: according to Cohen (9), d < 0.2, small difference; 0.2 < d < 0.5, medium difference; and d > 0.8, large difference. The significant level was set at P < 0.05.

Back to Top | Article Outline


Back to Top | Article Outline

Running performance.

The average distance covered by the subjects was 153 ± 15 km (ranging from 130 to 173 km), for an average effective running time, that is, the actual time spent running or walking on the treadmill for 18 h 53 ± 36 min). The average speed was 40.4% ± 3.9% of subjects' velocity at V˙O2max when considering the effective running time. For detailed performance results of the entire group of subjects during this experiment, the reader can refer to Martin et al. (29) and Millet et al. (31).

Back to Top | Article Outline

Running kinematics and kinetics.

Step frequency increased significantly (P < 0.001) from the fourth hour of the 24TR until the end of the run. This was caused by a significant (P < 0.001) decrease in contact time, with aerial time remaining constant throughout the 24TR (Fig. 1).

Overall, Fmax decreased significantly (P < 0.05) with time, with high interindividual variations (mean and SD of individual percent changes: −4.42% ± 5.76%; Table 1), which could explain the significant ANOVA, yet without significant post hoc results for this variable. This interindividual variability was also found for the LR at impact, which only tended to increase from 55.8 ± 13.1 to 59.2 ± 11.0 body weight per second between pre-24TR and post-24TR (P = 0.062), with a mean ± SD increase for the group of 5.96% ± 16.1%.

Back to Top | Article Outline

Spring-mass parameters.

Both Δz and ΔL decreased significantly (P < 0.001) from the fourth hour of the 24TR until the end of the run (Fig. 2), and this decrease was associated with significantly (P < 0.05) higher vertical and leg stiffness (Fig. 3).

The effect sizes of the changes observed in all the mechanical parameters are described in Table 1, along with the magnitude of changes as a percent of the initial (PRE) values.

Concerning stride variability, none of the CV calculated here was significantly different after the 24TR. These CV ranged from 1.60% (tc post-24TR) to 13.0% (LR pre-24TR), all parameters considered.

Most of the changes observed in running mechanics are illustrated in Fig. 4 through a representative running step pre-24TR and post-24TR.

Back to Top | Article Outline


The main result of this study is that subjects modified their running pattern and spring-mass behavior toward a significantly higher leg stiffness (P < 0.05) during a 24TR. This change in kleg was mainly related to a significantly (P < 0.001; −13.0% on average, large effect size) lower ΔL (and Δz), despite a significantly lower (P < 0.05; −4.4% on average, medium effect size) Fmax during contact (Table 1). These changes were associated with a large and significant (P < 0.001) increase in the frequency of oscillation of the spring-mass system, the latter being directly induced by lower contact times (P < 0.001), because aerial times did not change over the 24TR (Fig. 1).

Because, to our knowledge, no study reports how spring-mass behavior of the lower limbs and running mechanical pattern change during such a long-duration run as the 24TR studied here, the interpretation of our data in light of existing studies is limited. That said, the modifications of the running pattern we observed are in line with the study of Kyröläinen et al. (27), who report a significant ∼4.2% increase in stride frequency after a marathon, although in their study frequency was the only kinematic parameter measured that significantly changed. The only other study that considered the changes in running mechanics (mainly contact and aerial times) after a marathon showed very high interindividual variations, without significant results for the entire group (39). When comparing the changes observed here with those of ultralong-distance running, only the case study recently published about the Paris-Beijing runner (31) is available to our knowledge and also reported a higher step frequency (although caused by a shorter ta, with no change in tc) and a lower Fmax yet in one single experienced ultraendurance runner.

Interestingly, very similar changes in running mechanics were observed after a biopsy of the vastus lateralis muscle in trained ultraendurance runners (36): subjects' step frequency was significantly higher after biopsy, whereas other factors were significantly lower: Fmax, LR, and Δz. As for the Paris-Beijing run, the higher step frequency was due to a shorter ta, with no change in tc, and these changes were interpreted as being associated with a tendency to limit of the overall loading underwent by the musculoskeletal system when the muscular function is impaired by extremely long exercise duration (Paris-Beijing) or a structural modification of the muscle tissue, which is potentially painful (biopsy). However, if one focuses on the cause of this increase in step frequency in the present study compared with the Paris-Beijing and the biopsy protocols, it was not due to a shorter ta but to a shorter tc, the decrease in ta over the 24TR being not significant. Possible explanations for this discrepancy in what caused the increased step frequency are that i) 24TR subjects may have faced intense pain and run with decreased tc (hence braking phase duration), whereas it was less or not the case in the two aforementioned protocols, and ii) they may not have run on the instrumented treadmill with the exact same pattern as on the 24TR one, their ta remaining unchanged during the measurements, although it was expected to decrease, and may have actually on the 24TR treadmill, that is, when subjects knew no measurement was performed (see the "sampling effect" described in the following paragraphs).

If one considers the aging process as a slow yet actual impairment of the muscular function, the results of Cavagna et al. (8) and Karamanidis and Arampatzis (25) also match most of those presented here: compared with their younger counterparts, old men run with a lower vertical displacement of the COM during contact, a higher step frequency (caused in both studies by a reduced aerial time with no change in contact time), and a reduced maximal VGRF (25). Karamanidis and Arampatzis (25) also interpreted these changes as a reaction to the reduced capacities of subject's muscle-tendon unit by increasing the running safety.

It appears that when the overall neuromuscular function is impaired, noticeably because of ultralong-duration running, the overall behavior of the lower limb considered as a spring-mass system shifts toward a higher stiffness and oscillating frequency, allowing for reduced vertical force during each support phase and lower vertical displacements of the COM. The latter reduced lowering of the body mass during contact is illustrated by the decrease in leg length change during contact (significant from the fourth hour; Fig. 1), which could logically be associated with a reduced lengthening of the main lower limb extensor muscle-tendon units (knee extensors and plantarflexors) during the eccentric (braking) phase of the contact. Further, these muscles have specifically been shown to face a substantial peripheral fatigue during the 24TR (29). It is also interesting to note that the leg length change during contact (ΔL) was reduced by ∼13% on average, for an ∼5% shorter contact time, which logically means that the rate of leg length change was reduced post-24TR. One may reasonably assume this to result in a lower overall lengthening velocity of the muscle-tendon units of lower limbs extensors during the first half of the contact phase. Further, it has been hypothesized (23) that an increase in kleg (mainly through a reduced ΔL) could be related to a decrease in oxygen consumption because according to these authors, ΔL may be a variable describing the deformation of the leg spring and the "collective effect of muscle force demands of the entire lower extremity." However, to our knowledge, no clear, consistent, and direct link between spring-mass model and the energetics of normal running (i.e., with no exaggerated running style) has been shown, and the experimental data of the 24TR do not show any relationship between changes in spring-mass characteristics and running energetics (data not presented). Lastly, the increase in LR at impact, although only tending to be significant (P = 0.062), could seem contradictory with the lower values of Fmax observed post-24TR. However, LR and Fmax are mechanical variables representing the initial impact peak of force (caused by the foot colliding with the ground) and the active one (at midstance), respectively. Therefore, we think that post-24TR, the subjects could no longer control the falling of their lower limbs onto the ground, which resulted in a heel impact shock (LR) that tended to increase (with a high intersubjects variability).

The changes in running mechanics observed in the present study are almost exactly opposite to what has been observed in other fatigue conditions during running, that is, in cases of much shorter (hence ran at higher velocities) running efforts. Indeed, in the majority of "time-to-exhaustion" protocols we are aware of, either running mechanics and spring-mass characteristics did not change before and after fatigue (19,34) or they changed toward a lower step frequency (1,4,6,13,20,24) and longer contact time (1), higher vertical downward displacement of the COM (33), and higher vertical forces during impact and support phases (12). Lastly, for the two studies measuring leg stiffness, this parameter was reported to decrease (running time of ∼57 min (13)) or remain unchanged (∼6 min 40).

Our main hypothesis to explain this manifest discrepancy between modifications reported after time-to-exhaustion runs and what we observed after a 24TR is that in the latter case, subjects may wish to preserve (at least partly) the safety of their musculoskeletal structures and avoid pain by adopting the running pattern observed, mainly through reduced Fmax, Δz, and ΔL. This may not be an issue of importance in short-lasting time-to-exhaustion runs. Indeed, although very intense, the duration of the time-to-exhaustion runs is ∼60 min at the most and ∼20 min on average and may not be long enough to induce the modifications observed during the 24TR. Further, the decrease in force production (plantarflexor MVC) after a 10-km run (19) was shown to be less than that recently reported in 24TR subjects (29): ∼19% versus 30.3% ± 12.5%. Therefore, during a 24TR (compared with exhausting yet shorter runs), there may be a cumulative effect of more important loss of force production in lower limbs muscles and a necessity for subjects to limit duration-induced traumas, which could help them to stay below a maximal tolerated activation level (29).

Although significant, these changes in running mechanics are of lower relative magnitude than those reported by Martin et al. (29) about the subjects' neuromuscular function, particularly maximal voluntary force production. One can therefore wonder whether the observed changes in running mechanics and spring-mass behavior are caused by the important losses in knee extensors and plantarflexor maximal force (and somewhat escaping subjects' voluntary control) or the consequence of an anticipated strategy (conscious or not) adopted by subjects to maintain their running speed and to avoid or limit the potential pain and the effects of impacts generated at each step (particularly during the braking phase of the step). Further, the subjects of the present study reported a rate of perceived exertion that was increasing right from the early phases of the 24TR (significantly higher perceived exertion from the second and fourth hours; see Fig. 3B in Martin et al. (29)). In the present study, some of the mechanical changes were also significant after only 4 h of running, especially Δz and ΔL (Figs. 1 and 2), which could illustrate a choice made by subjects rather than a change induced by an alteration of their force capacities. This is further supported by the total absence of correlation (data not presented) we observed in these subjects between the alterations of the neuromuscular function reported by Martin et al. (29), including the loss of force (MVC) in knee extensors and plantarflexor muscle groups and the changes in running mechanics.

Some methodological limitations should be addressed in this study. Between measurement "breaks," that is, during the almost entire 24TR, subjects ran on a treadmill that was not the same as the instrumented treadmill used for measurements. Although the difference in treadmills rigidity may not have influenced our results (running mechanics measured on the same treadmill throughout the 24TR), subjects were fully aware of the moment measurements were performed because they were specifically requested to move from one treadmill to the other for that purpose. Consequently, it is possible that they changed their running pattern for the only reason that they were aware it was measured and when. This "sampling effect" phenomenon has recently been shown (37) and might have affected the data of this study because we did not measure the actual running mechanics and spring-mass behavior of subjects over the 24TR but rather those they adopted when running on the instrumented treadmill once every 2 h and most of all when knowing measurements were performed. Once again, this may have characterized all the subjects, although we cannot rule out a potential attenuation of actual changes in running mechanics over the 24TR induced by the way we measured them.

Two hours before the beginning of the 24TR, the subjects underwent a unilateral (left leg) biopsy of their vastus lateralis muscle, which has been shown to affect their running mechanics (36). However, although significant, the changes in mechanical variables were much less important in the context of a before and after biopsy comparison than those we report after the 24TR. For instance, the biopsy did not affect contact time or leg stiffness (36), and the changes it induced on step frequency were rather small (∼1.4% on average, effect size of 0.29) compared with ∼4.9% (effect size of 1.05) induced by the 24TR. The vertical downward displacement of the COM during contact was ∼13.9% lower after the 24TR (effect size of 1.19) versus ∼3.4% (effect size of 0.26) after biopsy. Further, the biopsy did not affect the variability of the running pattern. Lastly, all the subjects of the present study ran the entire 24TR after having underwent a biopsy following the exact same procedure, and comparisons (t-tests) made between the left and the right legs (five steps for each in the 10 steps samples studied) for all the mechanical variables measured pre-24TR and post-24TR showed no significant difference, except for kleg pre-24TR (15.9 ± 3.8 kN·m−1 for the left leg vs 15.0 ± 3.4 kN·m−1 for the right leg). That said, we acknowledge that we cannot rule out crossed effects (biopsy × running time) potentially variable among subjects.

In such long-lasting efforts involving sleep deprivation and the potential effects of circadian rhythms, a control group could have helped us separate the specific effects of running fatigue from those of sleep deprivation. Although to our knowledge no study investigated the effects of circadian rhythm and sleep deprivation on running mechanics, the fact that Martin et al. (29) observed that knee extensors MVC did not vary significantly during the 24-h period in their control group makes us reasonably assume that 10 km·h−1 running mechanics was not affected by sleep deprivation.

The last limit of the present study is that we observed changes over a level 24-h run in laboratory conditions, that is, with no impact of climate change and night, uphill, and downhill running (and the associated eccentric load). Thus, further studies performed in field conditions would be helpful to confirm the present results and whether they apply to ultratrail performance. For instance, we could expect that studying typical ultratrail races such as the Ultra Trail du Mont Blanc (166 km, 9400 m of negative altitude change, performances ranging from ∼21 to 46 h) could allow us to verify the hypothesis according to which changes in running pattern and spring-mass behavior occur to attenuate painful support phases because i) the duration of such an event is longer than 24 h and ii) the eccentric load faced by the musculoskeletal system is obviously more important than during a level 24TR.

In conclusion, all the changes observed show that the impairment of the muscular function caused by the 24TR is associated with a modification of the spring-mass behavior that is regulated toward a higher stiffness and oscillating frequency and reduced VGRF and vertical downward displacement of the COM. It remains unclear whether this spectrum of changes characterizing an overall "safer" running technique is deliberately chosen by the subjects (by anticipation or adaptation) and/or imposed by the neuromuscular consequences of exercise duration. However, these changes all lead to the same fact: runners attenuate the potentially harmful eccentric phase and overall load faced by their lower limbs musculoskeletal system at each step; this event occurring roughly 200,000 times over a typical 24-h run.

Disclosure statement: No fund was received for this study from the National Institutes of Health, the Wellcome Trust, the Howard Hughes Medical Institute, or others.

Conflict of interest: The authors declare no conflict of interest.

The authors thank Damien Fournet for assistance in data collection and the subjects for their constant cooperation and good mood throughout the 24TR.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Back to Top | Article Outline


1. Avogadro P, Dolenec A, Belli A. Changes in mechanical work during severe exhausting running. Eur J Appl Physiol. 2003;90:165-70.
2. Belli A, Bui P, Berger A, Geyssant A, Lacour JR. A treadmill ergometer for three-dimensional ground reaction forces measurement during walking. J Biomech. 2001;34:105-12.
3. Blickhan R. The spring-mass model for running and hopping. J Biomech. 1989;22:1217-27.
4. Borrani F, Candau R, Perrey S, Millet GY, Millet GP, Rouillon JD. Does the mechanical work in running change during the V˙O2 slow component? Med Sci Sports Exerc. 2003;35(1):50-7.
5. Bullimore SR, Burn JF. Consequences of forward translation of the point of force application for the mechanics of running. J Theor Biol. 2006;238:211-9.
6. Candau R, Belli A, Millet GY, Georges D, Barbier D, Rouillon JD. Energy cost and running mechanics during a treadmill run to voluntary exhaustion in humans. Eur J Appl Physiol Occup Physiol. 1998;77:479-85.
7. Cavagna GA. Force platforms as ergometers. J Appl Physiol. 1975;39:174-9.
8. Cavagna GA, Legramandi MA, Peyré-Tartaruga LA. Old men running: mechanical work and elastic bounce. Proc Biol Sci. 2007;275:411-8.
9. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale (NJ): Lawrence Erlbaum Associates; 1988. p. 567.
10. Dalleau G, Belli A, Bourdin M, Lacour JR. The spring-mass model and the energy cost of treadmill running. Eur J Appl Physiol Occup Physiol. 1998;77:257-63.
11. De Wit B, De Clercq D, Aerts P. Biomechanical analysis of the stance phase during barefoot and shod running. J Biomech. 2000;33:269-78.
12. Derrick TR, Dereu D, McLean SP. Impacts and kinematic adjustments during an exhaustive run. Med Sci Sports Exerc. 2002;34(6):998-1002.
13. Dutto DJ, Smith GA. Changes in spring-mass characteristics during treadmill running to exhaustion. Med Sci Sports Exerc. 2002;34(8):1324-31.
14. Farley CT, Ferris DP. Biomechanics of walking and running: center of mass movements to muscle action. Exerc Sport Sci Rev. 1998;26:253-86.
15. Farley CT, Glasheen J, McMahon TA. Running springs: speed and animal size. J Exp Biol. 1993;185:71-86.
16. Farley CT, Gonzalez O. Leg stiffness and stride frequency in human running. J Biomech. 1996;29:181-6.
17. Ferris DP, Liang K, Farley CT. Runners adjust leg stiffness for their first step on a new running surface. J Biomech. 1999;32:787-94.
18. Ferris DP, Louie M, Farley CT. Running in the real world: adjusting leg stiffness for different surfaces. Proc Biol Sci. 1998;265:989-94.
19. Finni T, Kyrolainen H, Avela J, Komi PV. Maximal but not submaximal performance is reduced by constant-speed 10-km run. J Sports Med Phys Fitness. 2003;43:411-7.
20. Gerlach KE, White SC, Burton HW, Dorn JM, Leddy JJ, Horvath PJ. Kinetic changes with fatigue and relationship to injury in female runners. Med Sci Sports Exerc. 2005;37(4):657-63.
21. Hausswirth CA, Bigard X, Guezennec CY. Relationships between running mechanics and energy cost of running at the end of a triathlon and a marathon. Int J Sports Med. 1997;18:330-9.
22. He JP, Kram R, McMahon TA. Mechanics of running under simulated low gravity. J Appl Physiol. 1991;71:863-70.
23. Heise GD, Martin PE. "Leg spring" characteristics and the aerobic demand of running. Med Sci Sports Exerc. 1998;30(5):750-4.
24. Hunter I, Smith GA. Preferred and optimal stride frequency, stiffness and economy: changes with fatigue during a 1-h high intensity run. Eur J Appl Physiol. 2007;100:653-61.
25. Karamanidis K, Arampatzis A. Mechanical and morphological properties of different muscle-tendon units in the lower extremity and running mechanics: effect of aging and physical activity. J Exp Biol. 2005;208:3907-23.
26. Kuitunen S, Avela J, Kyrolainen H, Nicol C, Komi PV. Acute and prolonged reduction in joint stiffness in humans after exhausting stretch-shortening cycle exercise. Eur J Appl Physiol. 2002;88:107-16.
27. Kyrolainen H, Pullinen T, Candau R, Avela J, Huttunen P, Komi PV. Effects of marathon running on running economy and kinematics. Eur J Appl Physiol. 2000;82:297-304.
28. Lee CR, Farley CT. Determinants of the center of mass trajectory in human walking and running. J Exp Biol. 1998;201:2935-44.
29. Martin V, Kerhervé H, Messonnier L, et al. Central and peripheral contributions to neuromuscular fatigue induced by a 24-hour treadmill run. J Appl Physiol. 2010;108(5):1224-33.
30. McMahon TA, Cheng GC. The mechanics of running: how does stiffness couple with speed? J Biomech. 1990;23(1 suppl):65-78.
31. Millet GY, Banfi JC, Kerhervé H, et al. Physiological and biological factors associated with a 24 h treadmill ultra-marathon performance. Scand J Med Sci Sports. 2009 In press.
32. Millet GY, Morin JB, Degache F, et al. Running from Paris to Beijing: biomechanical and physiological consequences. Eur J Appl Physiol. 2009;107:731-8.
33. Mizrahi J, Verbitsky O, Isakov E. Fatigue-induced changes in decline running. Clin Biomech. 2001;16:207-12.
34. Morgan DW, Martin PE, Baldini FD, Krahenbuhl GS. Effects of a prolonged maximal run on running economy and running mechanics. Med Sci Sports Exerc. 1990;22(6):834-40.
35. Morin JB, Jeannin T, Chevallier B, Belli A. Spring-mass model characteristics during sprint running: correlation with performance and fatigue-induced changes. Int J Sports Med. 2006;27:158-65.
36. Morin JB, Samozino P, Feasson L, Geyssant A, Millet GY. Effects of muscular biopsy on the mechanics of running. Eur J Appl Physiol. 2009;105:185-90.
37. Morin JB, Samozino P, Peyrot N. Running pattern changes depending on the level of subjects' awareness of the measurements performed: a "sampling effect" in human locomotion experiments? Gait Posture. 2009;30:507-10.
38. Morin JB, Samozino P, Zameziati K, Belli A. Effects of altered stride frequency and contact time on leg-spring behavior in human running. J Biomech. 2007;40:3341-8.
39. Nicol C, Komi PV, Marconnet P. Effects of marathon fatigue on running kinematics and economy. Scand J Med Sci Sports. 1991;1:195-204.
40. Slawinski J, Heubert R, Quievre J, Billat V, Hannon C. Changes in spring-mass model parameters and energy cost during track running to exhaustion. J Strength Cond Res. 2008;22:930-6.


©2011The American College of Sports Medicine