An arduous and strategic part of a triathlon is the transition from cycling to running. Authors of lay publications have suggested various cycle-to-run methods to optimize this transition. For example, Brick (2) recommended concluding the cycling stage with a low-resistance, fast cadence spin and beginning the run with long, slow strides. Similarly, Niles (15) speculated that fast cycling cadence reduces force production and oxygen consumption during cycling. In contrast, Friel (5) advocated high-resistance, low cadence frequencies in conjunction with stretching during the final moments of the cycling bout. Further, he promoted stretching the calves and hamstrings by standing on the pedals and flexing the hips in an effort to simulate the muscle actions involved in running (6). However, these ideas remain controversial and are not based on scientific evidence.
Some research studies have investigated the effects of cycling on subsequent running, but they have also been contradictory. Hausswirth (10) found that running efficiency decreases following a session of cycling. In addition, Hausswirth et al. (11) showed that after cycling running stride length decreased and hip flexion increased. In a previous study, we compared running kinematics and performance during a maximal run after a high-intensity cycling bout versus after a high-intensity running bout (8). Cycling before running elicited unique transitory adaptations during the maximal run—stride frequency increased, stride length decreased, and, in contrast to Hausswirth (11), efficiency increased as indicated by heart rate—yet none of these studies directly investigated how cycling cadence affects running speed and running biomechanics.
More extensive past research has focused on cycling and running as individual disciplines. For instance, Gaesser and Brooks (7) and Cavanagh and Williams (4) explained that altering cycling cadence or running stride frequency affects economy and hence performance. Specifically, Cavanagh and Williams (4) noted that deviations from preferred stride frequency would increase energy cost in most individuals. But the manipulation of running stride frequency could, in some individuals, improve economy and theoretically enhance race performance (1,14,17). Thus, previous studies have documented the energetics and biomechanics of preferred cycling and running kinematics, but little is known about how the manipulation of kinematics affects performance.
It is plausible that cycling cadence could influence subsequent running stride frequency and, therefore, speed. Classic studies in neuroscience have demonstrated that persons performing a rhythmic activity for an extended period of time will involuntary continue this movement pattern (3). This phenomenon is called perseveration. For example, Gurfinkel, et al. (9) showed that when a suspended human leg is stimulated to produce a rhythmic stride pattern, the leg would continue to move at the prescribed frequency for numerous cycles, even after stimulation ceased. In the context of multi-sport events such as triathlons, it is possible that perseveration would cause individuals to unintentionally begin the running bout with a stride frequency similar to the cadence of the previous cycling bout.
The rationale for the present study was that cycling cadence might influence subsequent running speed via changes in stride frequency. We anticipated that, compared with the preferred cadence, a fast cycling cadence would increase stride frequency and subsequent running speed. In contrast, a slow cycling cadence would decrease stride frequency, thereby decreasing running speed.
Thirteen male athletes of the University of Colorado triathlon team volunteered (average age = 24.78 ± 1.20 yr; average mass = 72.69 ± 1.42 kg; average height = 1.80 ± 0.02 m). Each participant had at least 2 yr of triathlon-specific training with varying levels of racing experience. All athletes gave written informed consent that followed the guidelines of the University of Colorado Human Research Committee.
Upon waking on the morning of a testing session, each participant determined his resting heart rate, monitoring for excessive training or illness (13). If resting heart rate was more than 5 beats per min (bpm) faster than usual, the participant postponed the session for 48 h. Of the 39 completed trials, this occurred on only one occasion.
Each participant completed three sessions on consecutive weeks at an indoor 200-m running track facility. Each session consisted of a cycling bout immediately followed by a running bout. The participants rode their own racing bikes with clipless pedals, and mounted on a stationary windtrainer (Performance Corporation, Chapel Hill, NC). The participants performed a self-directed 20-min warm-up consisting of cycling, running, and stretching. During the first session this routine was noted and duplicated before the second and third sessions.
For the first session (used as the control condition [CC]) each participant completed a 30-min cycling bout immediately followed by a 3200-m running bout. We asked the participants to perform at an intensity that simulated that of a race by selecting an appropriate resistance on the racing bike. Heart rate (HR) was monitored during each session using an electronic heart rate device with a chest electrode (Polar Accurex II, Polar Electronics, Woodbury, NY) and was recorded every 2 min during the control condition so the participant could monitor and maintain the same intensity during the second and third sessions.
The cadence of the cycling bout during the control condition was freely chosen. However, the participants were asked to cycle at a frequency that they would use during a race. At the conclusion of the cycling bout, the participants quickly dismounted their bikes, removed their cycling gear, and put on running shoes. They were instructed to perform their accustomed routine of transitioning from cycling to running during a triathlon. The time for each individual to complete this routine was recorded.
During both the cycling and running bouts, a 120-Hz video camera recorded the locations of reflective markers that were placed on key anatomical landmarks on the left side of each participant: head of the humerus, iliac crest, lateral condyle of the femur, lateral malleolus of the fibula, calcaneus, and the fifth metatarsal. The camera was mounted on a tripod three meters away from the first inside lane of the indoor track. After each session, the videotape recording of the cycling bout and the beginning of each lap of the running bout was analyzed utilizing a video digitization and kinematics analysis system (Peak Motus, Peak Performance Corporation, Englewood, CO).
During the fast condition (FC) and slow condition (SC), each participant completed a 30-min high-intensity cycling bout at a cadence 20% faster or 20% slower than the control condition. The order of the FC and SC was randomized. Cycling cadence was enforced using a metronome set to the particular cadence for that trial (Wittner Quartz Metronome, Korea). Each participant also matched his heart rate intensity of the control cycling bout (average heart rate of 13 participants in FC = 165 ± 3, CC = 164 ± 4, SC = 162 ± 2) by modifying bike resistance to alter the intensity of their cycling. The cycling bout was immediately followed by a 3200-m run at the same heart rate as during the control condition run.
Key kinematic variables were calculated for the running trials as follows. Stride frequency (SF) was defined as the period of time measured in seconds from the initial contact of a given foot to the next initial contact of the same foot. Stride length (SL) was defined as the distance traveled in the direction of progression from initial contact of a foot to the next contact of the same foot measured in meters. Stance time was defined as the portion of the running stride during which one foot is in contact with the ground, specifically from foot strike to toe off, measured in milliseconds (ms). Swing time was defined as the portion of the running stride during which one foot is no longer in contact with the ground, specifically from toe off to foot strike, measured in milliseconds (ms). Two-dimensional joint angles in the sagittal plane of the hip, knee, and ankle were measured at each 45-degree crank-angle segment during the cycling bout and at foot strike and toe off during the running bout. In addition, the maximal flexion of each joint was recorded during the stance phase and the swing phase.
All data are reported as mean values ± standard error. Cadence and heart rate during the cycling bout and the defined dependent measures and total time to complete 3200 m during the running bout were analyzed by repeated measures ANOVA. The design was a 3 × 8 factorial with the condition (fast, control, slow) and lap number (1,2,3,4,6,8,12,16) as factors. Post hoc analysis of significant means and appropriate interactions were performed using a Newman-Kuels test for simple effects. The level of statistical significance was set at P < 0.05.
Our most important finding was that faster cadence cycling substantially increased the subsequent average running speed of the 3200-m race effort (Figure 1). During the fast condition, participants ran 4% faster (P < 0.05) than the control condition and 7% faster (P < 0.01) than the slow condition (Table 1). This remarkable increase in running speed after cycling with a fast cadence occurred with heart rates equivalent to the values of the control condition during both the cycling and running bouts. Furthermore, the speed increases were consistent throughout the 3200-m race effort for the fast condition.
The participants ran faster primarily by increasing stride frequency (Figure 2). The mean stride frequency of the fast condition was 5% greater (P < 0.05) than the control condition and 10% greater (P < 0.05) than the slow condition. Immediately after each cycling bout, the participants ran with a stride frequency similar to the previous cycling cadence (Table 1). During the fast condition, stride frequency was elevated above the control and slow conditions throughout the 3200-m race effort.
In contrast, the mean stride length did not differ between conditions by more than 2% (Figure 3). In all three conditions the participants initially ran with short strides, but by the conclusion of the second lap, stride length had returned to normal and stabilized.
Stance times were significantly shorter for the fast condition during the first three laps of the 3200-m race effort (Figure 4). During the initial lap of the fast condition, the participants ran with a stance time 17% shorter than the slow condition (P < 0.01) and 11% shorter than the control condition (P < 0.05). During the second lap, the fast condition stance time was 12% shorter than the slow condition (P < 0.05). Throughout the run, stance time during the fast condition progressively increased whereas it remained fairly constant during the control and slow conditions. Swing times did not differ between conditions by more than 4% (Figure 5). However, in all conditions, swing time steadily increased throughout the run.
There were no significant differences between conditions at any time for any of the leg joint angles during cycling or at foot strike or toe off during running. Also, the maximal flexion of each joint did not differ during the mid-stance and mid-swing phase between conditions.
After cycling with a fast cadence, participants ran the 3200-m race simulation nearly a minute faster than after cycling with a slow cadence. This increased speed was accomplished with similar physiological effort, increased SF, unchanged SL, and similar leg angular displacements. Thus, our null hypothesis that cycling cadence would have no effect on running speed or stride frequency was rejected.
These differences in running speed did not occur as a result of varying joint angles during the cycling bout. Past research has shown that a cyclist can alter which muscles are the agonists, antagonists, or synergists by changing the amount of ankle flexion at the bottom of the crank revolution (13,16). But in the present study, the hip, knee, and ankle joint angles were not significantly different at any crank angle during the cycling conditions. Rather, it appears that cycling cadence was the major determinant influencing subsequent running performance and kinematics.
Perseveration is a likely mechanism responsible for the increased running stride frequency of the fast condition. Hudson determined that cycling and running depend on different neural firing rates owing to the specific cyclic frequencies of each movement (12). It is possible that the neural firing rate after each cycling condition biased the firing rate used subsequently for running. For example, the high frequency firing rates during the fast cadence cycling bout appear to have translated into an increased stride frequency during the running bout. In short, the coordinated neural control during cycling strongly influenced the generation of the running frequency.
A central nervous system rhythm generator may help to explain how cycling cadence influenced subsequent running stride frequency. Rhythmic signals generated by visual, auditory, or proprioceptive frequencies are sent to a neural rhythm generator and can entrain the stride pattern (16,18). The generator is a locomotor control in which movement patterns emerge from the tempo of the neural system and from the rhythmic movements of the musculoskeletal system (19). In our study, the participants cycled at a specific cadence for 30 min, and it seems that the neural rhythm generator produced a program that influenced subsequent running kinematics. In summary, running stride frequency based on neural firing rates is dependent on prior task patterns (18).
In conclusion, we found that cadence during a cycling bout immediately before a running bout influenced running performance and stride kinematics. Maintaining an unusually high cadence while cycling resulted in substantially faster running speed. The neurological bases for this phenomenon are likely a combination of perseveration and central pattern generator effects. On a practical note, triathletes may benefit from adopting an increased cycling cadence before running.
Address correspondence to: Jinger S. Gottschall, Department of Kinesiology and Applied Physiology, University of Colorado, Campus Box 354, Boulder, CO 80309; E-mail: [email protected]
1. Anderson, T. Biomechanics and running economy. Sports Med. 22: 76–89, 1996.
2. Brick, M. Short-course training for an athlete with a cycling background. Inside Triathlon 11: 76, 1996.
3. Brugger, P., and M. Gardener. Perseveration
in healthy subjects: an impressive classroom demonstration for educational purposes. Percept. Mot. Skills. 78: 777–778, 1994.
4. Cavanagh, P. R., and K. R. Williams. The effect of stride length and oxygen uptake during distance running. Med. Sci. Sports Exerc. 14: 30–35, 1982.
5. Friel, J. Duathlon training and racing. Inside Triathlon. 12: 21–23, 1998.
6. Friel, J. Power and the master athlete. Inside Triathlon. 13: 18–19, 1999.
7. Gaesser, G. A., and G. A. Brooks. Muscular efficiency during steady-rate exercise: effects of speed and work rate. J. Appl. Physiol. 32: 1132–1139, 1975.
8. Gottschall, J. S., and B. M. Palmer. Effects of cycling on running stride length and stride frequency. J. Sport Cond. Res. 14: 97–101, 2000.
9. Gurfinkel, V. S., Y. S. Levik, O. V. Kazennikov, and V. A. Selionov. Locomotor-like movements evoked by leg muscle vibration in humans. Eur. J. Neurosci. 10: 1608–1612, 1998.
10. Hausswirth, C., A. X. Bigard, M. Berthelot, M. Thomaidis, and C. Y. Guezennec. Variability in energy cost of running at the end of a triathlon and a marathon. Int. J. Sports Med. 17: 572–579, 1996.
11. Hausswirth, C., A. X. Bigard, and C. Y. Guezennec. Relationships between running mechanics and energy cost of running at the end of a triathlon and a marathon. Int. J. Sports Med. 18: 330–339, 1997.
12. Hudson, A. J. Perseveration
. Brain. 91: 571–582, 1968.
13. Krieder, P. B., D. E. Cundiff, J. B. Hammett, C. W. Corter, and K. W. Williams. Effects of cycling on running performance in triathletes. Ann. Sports Med. 3: 220–225, 1988.
14. Kyrolainen, H., P. V. Komi, and A. Belli. Mechanical efficiency of athletes during running. Scand. J. Med. Sports Sci. 5: 200–208, 1995.
15. Niles, R. Bike-to-run crossover. Triathlete Magazine,
18:38–42, February 1998.
16. Nilsson, J., and A. Thorstensson. Adaptability in frequency and amplitude of leg movements during human locomotion at different speeds. Acta. Phsyiol. Scand. 129: 107–114, 1987.
17. Shields, S. L. The effect of varying lengths of stride on performance during submaximal treadmill stress testing. J. Sports Med. Phys. Fit. 22: 66–72, 1982.
18. Taga, G. A model of the neuro-musculo-skeletal system for human locomotion. Biol. Cybern. 73: 113–121, 1995.
19. Taga, G., Y. Yamaguchi, and H. Shimizu. Self-organized control of bipedal locomotion by neural oscillators in unpredictable environment. Biol. Cybern. 65: 147–159, 1993.