Why triathletes need to worry about their cycling cadence
One of the most crucial and difficult moments in a triathlon competition is the transition between cycling and running and various triathlon ‘experts’ have offered their opinions on how this transition should proceed. For example, Brick suggested that the cycling stage of the competition should conclude with a low-resistance, fast-cadence spin, while the consequent run portion should begin with long, slow strides. Along similar lines, Niles argued that fast cycling cadences towards the end of the cycling stage tend to reduce force production by the muscles and thus overall oxygen consumption, leaving the athlete in a less-stressed physical state at the beginning of the run stage. By contrast, veteran triathlon coach Joe Friel called for high-resistance (i.e. using high gears) and low cadence frequencies combined with stretching out the legs on the bike during the final moments of the cycling stage. Friel indicated that the calves and hamstrings could be stretched properly if triathletes stood up on their bikes and made strenuous efforts to flex their hips while riding; this routine, he claimed, would simulate the muscle actions involved in running, thus preparing the neuromuscular system for the subsequent running phase of competition. Note, though, that these apparently reasonable hypotheses are not based on scientific evidence. Unfortunately, scientific research on the effects of various types of cycling on subsequent running is somewhat impoverished, and the studies that do exist are somewhat contradictory. Key findings from these studies are as follows:
- Running efficiency tends to be reduced after cycling;
- Running form is significantly altered after cycling, with an increase in hip flexion and a decrease in stride length.
One important investigation showed that cycling before running elicits specific adaptations during a very intense run, including an increase in stride frequency and a decrease in stride length. Somewhat paradoxically, efficiency during this maximal run was greater than when running very fast without prior cycling.
Non sports-related research in the exciting field of neuroscience has shown that when people perform a rhythmic activity for an extended period of time, they tend to continue this movement pattern involuntarily, even after the activity is officially ‘over’. For example, when a suspended human leg is artificially stimulated to produce a rhythmic stride pattern, the leg tends to continue to move at the original frequency for many ‘cycles’, even after the stimulation is removed. This finding could be extremely useful for triathletes: if a triathlete could establish a pattern of neuromuscular control late in the cycling stage which would lead to an optimal gait during the running stage, performance might improve significantly. Conversely, if a triathlete established neuromuscular regulation during cycling which led to sub-optimal mechanics, stride frequency and/or stride length during the subsequent run, performance would be hindered.
Clearly, though, triathletes must strive for balance in this situation and there is no point in finding a cycling mode which optimises running to the detriment of cycling performance.
In an effort to find the right cycling strategy for the late portion of the cycling stage, researchers from the University of Colorado recently studied 13 male athletes from the university triathlon team. The average age of the triathletes was 25, and each had engaged in triathlon-specific training for a minimum of two years. The researchers hypothesised that cycling cadence during a triathlon would influence subsequent running speed by means of a direct effect on running-stride frequency. Specifically, they anticipated that a fast cycling cadence would induce a high stride frequency during consequent running, with a positive effect on overall running speed. Conversely, they believed, a slow cycling cadence would lead to a slow stride rate, reducing running speed. Each study participant completed three sessions on consecutive weeks at an indoor 200m running track, each session consisting of a cycling workout immediately followed by a run. The athletes rode on their own bikes, using clipless pedals, mounted on a stationary ‘windtrainer’. For the first session (the ‘control’ condition), the triathletes completed a 30-minute bout of cycling, followed by an intense 3,200m run; they were instructed to perform as they would in a racing situation, choosing their own cycling cadence. After the 30-minute bout, the athletes quickly dismounted, removed their cycling gear and put on their running shoes, in accordance with their normal transition routines, before setting off for the run. The two remaining sessions consisted of a ‘fast condition’ (FC) and a slow one (SC). For the former, the triathletes completed the 30-minute bike phase of the workout using a cadence 20% faster than the control condition, and for the latter they went for a cadence 20% slower. Since the control situation involved a cycling cadence of 90 revolutions per minute that meant that FC used 109 revs per minute, while SC backed down to 71 revs. The order of fast and slow conditions was randomised, and cycling cadences were ‘enforced’ by metronomes. Each athlete matched his FC and SC heart rate to that of the control condition by adjusting bike resistance. Both conditions were followed by the same all-out 3,200m run performed in the control condition.
Same heart rate, better performance
Since a 200m track was used for all three sessions, each 3,200m max effort involved running 16 laps. As it turned out, using the fast, low-resistance cadence during cycling enhanced running speed on eight of the 16 laps by comparison with SC. Overall, the triathletes ran 4% faster in FC than in the control condition and 7% faster than in SC: running velocity during the 3,200m test was 4.45m per second for FC (6:02 per mile), 4.19 m/sec for control (6:24 per mile) and 4.08m/sec for SC (6:34 per mile). Interestingly enough, the triathletes managed their improved performances during the fast condition without any corresponding rise in heart rate, which averaged about 164 beats per minute during all conditions! The participants ran faster after the fast-cadence cycling primarily by increasing stride frequency, which averaged 1.52 strides per second during FC, 1.42 during control and just 1.35 strides following the slow-cadence, high-resistance cycling. Stride frequency in each condition closely paralleled cycling cadence during the preceding bout of cycling, being 5% greater during FC than control and 10% greater than during SC. Amazingly enough, stride length did not differ between the three conditions, which meant that the faster running pace achieved during the fast condition was accomplished solely by manipulation of stride frequency. Following the fast-cadence cycling, the athletes, in effect, needed less time to generate the muscular force required for a typical stride length; a more abbreviated stance phase of the gait cycle accounted for the increase in stride rate. In fact, during the initial running lap of FC, the stance time was 17% shorter than during SC and 11% shorter than during control, while during the second lap stance time was 12% shorter during FC than during SC. Stance time, measured at eight points during the 3,200m run, was quicker during FC than SC on five occasions. Meanwhile, swing time – 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 – was briefer during FC than SC on only two occasions. Biomechanically, there were no differences between FC, SC and control at any point for any leg-joint angles during cycling or at foot strike or toe-off during running. In addition, the maximal flexion of each leg joint did not differ during the mid-stance and mid-swing phases of the gait cycle between the three conditions. What, then, was the mechanism which allowed FC to produce higher stride rates and faster running speeds? The researchers suggested that a nervous system ‘rhythm generator’ might have been responsible for this effect, and indeed there is evidence in the scientific literature that rhythmic signals generated by visual or proprioceptive frequencies can be sent via special sensory nerves to the central nervous system and can thus ‘entrain’ a very fast stride pattern. This rapid stride pattern is created and ‘enforced’ by motor-control signals emanating from the central nervous system with the same frequency as the incoming sensory messages. The rhythm generator hypothesis is certainly a reasonable one. In the Colorado study, the athletes cycled at a fast, medium or slow cadence for 30 minutes without stopping; during the subsequent runs, the fast cycling cadence always produced quick strides, the medium cycling cadence intermediate strides and the slow cadence lethargic ones, suggesting that the athletes’ nervous systems had ‘learned’ rates of motor control during the bouts of cycling which they put to use during their runs.
To summarise, running stride frequency is based on neural firing rates, and neural firing rates are dependent on ‘prior task patterns’. In this study, the prior task patterns were the cadences established during the cycling bouts, but in other situations the task patterns may simply be the ones used repeatedly in training. For example, endurance runners cannot build their training around a foundation of prolonged running at slow-to-moderate paces then expect to be fast during competitions; if their task patterns possess a slow bias, their neural firing rates will be slower during major races than those of athletes who have repeatedly used explosive task patterns during training. So what’s the bottom line for triathletes? The exact time needed to establish an optimal task pattern during cycling for the running phase of competition is not known (remember that the Colorado researchers used 30 minutes); but it is clear, based on this new evidence, that high cycling cadences of perhaps 100-110 revs per minute, carried out for a significant amount of time, may well lead to the best-possible performances during the ensuing run phase of the triathlon.
The truly amazing feature of this study is that the athletes were able to cover 3,200m about a minute faster after a fast cycling cadence than after a slow one, with no apparent corresponding increase in physiological effort (as indicated by heart rate). Certainly, there is a need for more research in this area. For, although the differences in run quality were strikingly large in this study (from 6:02 to 6:34 per mile), the advantages of fast-cadence cycling might be less marked if the ensuing run was as long as 42.2 kilometres – as in an ‘Ironman’ competition – rather than the much shorter distance used in the Colorado research.