The Best Versus The Rest

The pursuit of optimal technique in sport is guided by biomechanical principles—laws of motion and anatomy that help us understand what efficient and effective movement looks like. But while these laws provide a framework for good technique, they don’t prescribe a single ideal model for everyone. The challenge for coaches is knowing which movement patterns are acceptable variations and which ones may compromise performance or increase injury risk. This article summarises a recent study that categorised athletes based on their coordination patterns during initial sprint acceleration and described the association between certain technical features and performance.

In sport, technique refers to the movement strategy an athlete uses to perform a task. Coordination is a significant component of technique, describing how related body segments move together to accomplish a specific task goal. Developing a particular coordination pattern involves managing the abundance of ways different body parts can move. According to the dynamical systems theory of motor control, coordination patterns emerge from the self-organisation of body segments within constraints set by an individual's characteristics, the environment, and the task.

While recognising that systems self-organise, this does not suggest all techniques are equally effective. Dynamical systems theory describes how coordination patterns emerge but does not evaluate the effectiveness of task execution, which is vital in sports contexts. Therefore, when evaluating the effectiveness an athlete’s technique, it is necessary to consider both biomechanical principles and performance outcomes.

Coaching sprinters at a high level involves understanding each athlete and applying an individualised approach. Nonetheless, categorising athletes by specific features can provide a practical method when coaching large groups, and ultimately serve as an initial step towards this individualisation.

In the first study of Byron Donaldson’s PhD we described the coordination patterns of sprinters, ranging from highly trained to world class competitors (part 1 of this blog series). For the next step, we focused on two key objectives: (i) categorising sprinters into groups based on similar coordination strategies and (ii) comparing performance outcomes between the groups to investigate whether certain coordination patterns may be preferable over others. The influence of individual strength-based constraints will be discussed in part three.

Technique clusters

The methods for quantifying inter- and intralimb coordination were described in part one of this blog series. Briefly, we captured segment kinematics of 21 athletes using IMUs during a training session where they did at least three maximum effort accelerations from a block start. We then applied a modified vector coding method to produce a coupling angle (that could range from 0 to 360) at 101 time points across each of the first four steps, and categorised the coupling angle into one of eight colour-coded bins. The bin designates the direction of rotation of the two segments, whether they are moving in the same or opposite direction to each other, and which one is rotating further between time points. For this categorisation study, we focused on the thigh-thigh and shank-foot coordination and ran separate analyses for step 1 and steps 2-4 combined.

In order to identify subgroups of athletes with similar technique, we calculated the similarity of coordination patterns using a coupling angle distance score (CAdist) for every paired combination of sprinters, where a score of 0 meant they were identical and 1 meant they were in completely opposite coordination patterns throughout the movement. We did this for the thigh-thigh and shank-foot segment couplings, separately for step 1 and steps 2-4, ultimately producing a matrix that showed how similar each sprinter was to every other sprinter. For a detailed explanation of the method, see this paper presented at ISBS 2022.

Cluster analysis was then performed on the CAdist scores to group sprinters who were most similar to each other, resulting in three clusters based on the step 1 coordination patterns (referred to as A, B and C), and two clusters for steps 2-4 (X and Y).

Figure 1. Dendograms and mean coupling angle profiles for step 1 (three clusters: A, B and C) and steps 2-4 (two clusters: X and Y). A step is defined from toe off (0%) to the next toe off (100%). Black vertical line indicates touchdown and dashed vertical line indicates thigh crossover. The main periods that differentiate between clusters are highlighted in the yellow boxes and the coloured arrows show the direction of segment motion with the larger of the two arrows indicating the dominant segment, where relevant.

For step 1, there were some interesting findings about coordination patterns and positions at key time points. The clusters effectively represented two different patterns of thigh interchange after block exit and balance of foot or shank contribution to dorsiflexion in early stance. Increasing thigh separation after block exit was either driven by lead limb flexion (A & B) or trail limb extension (C). This was followed by a brief “float” period where both limbs were flexing, before trail limb dominant anti-phase occurred until late stance for A&B, whereas in C, there was no float period and front limb retraction dominated the thigh interchange until just after ground contact. The shank-foot of the lead limb either continued to swing out until making contact with the ground after which ankle dorsiflexion occurred through “heel drop” during early stance (A), or there was a reversal of lower leg rotation -  “shin block” - before contact (B & C). Athletes in cluster A (yellow stick figure below) featured a more “tucked” lead limb at block exit and the shin in a more horizontal orientation at touchdown, which explains why the heel drop dominated over shin roll. Athletes in cluster C (blue stick figure) had a more vertical lead shin at block clearance and touchdown, with the lead limb appearing to “plant” on the ground further ahead of the centre of mass. They therefore relied more on shin roll to translate their centre of mass over the contact point, and required longer ground contact times (194 ± 25 ms, versus A: 150 ± 6 ms and B: 174 ± 22 ms). Athletes in cluster B shared thigh-thigh patterns with A and the shin block pattern before contact with C.

Figure 2. Typical body orientations at key events for clusters in step 1. Dashed lines indicate the limb that was not analysed. BC = block clearance; TD = Touchdown; TO = toe off.

In steps 2-4, clusters were separated by differences in the timing of thigh rotation reversal and the duration of different parts of the shank-foot rotation patterns. Athletes in cluster X displayed an anti-phase thigh-thigh pattern immediately after toe off and started lead leg retraction while still in contact with the ground at the end of the step. They had a shorter anti-clockwise (swinging out) of the shank-foot in the flight phase, and an earlier heel drop after contact which persisted for longer before the final shin roll started. Conversely, athletes in Y continued to extend (“push”) with the trail limb after toe off and the switching action of the thighs was therefore delayed. They spent longer in anti-clockwise rotation of the shank-foot in flight and relied more on shin roll than heel drop during stance.

Is there a “best” strategy?

The sets of clusters identified for step 1 and steps 2-4 allowed us to categorise sprinters into one of six possible combinations. The full breakdown of how sprinters were categorised is summarised in the image below, including male/female, 100m PB times and performance level classification (world class, elite, highly trained).

The B-X strategy was associated with higher performing athletes – including both world class and two out of the five elite sprinters. In other words, higher level sprinters with faster initial acceleration times exhibited strategies characterised by lead thigh dominant thigh separation after block exit and greater foot dominant coordination during early stance dorsiflexion in step 1 combined with early swing thigh retraction in later steps.

The remaining two elite athletes (one male, one female) and another highly trained male, whose acceleration performance was comparable with the elites, were categorised as A-Y. This strategy is characterised by lead thigh flexion dominant coordination in early flight and a tucked lower leg without a shin block before contact in step 1, combined with a delayed swing leg recovery in later steps. This less common strategy may therefore be an effective alternative for certain athletes, demonstrating the influence of individual constraints on emergent coordination patterns.

* A note for researchers considering similar types of research questions: The cluster analysis approach was a key element in allowing us to tease out these different technique strategies. The alternative approach, which has been used more often in biomechanics research, would be to define two groups upfront and compare coordination patterns between them (i.e. highly trained athletes vs elite and world class). This method would likely have washed out the subtle variations that we found, because highly trained athletes fell across all 6 categories, and the B-X and A-Y split would have been lost. Both approaches can be appropriate – but worth considering depending on how your specific research question is framed.

Figure 3. Matrix demonstrating step 1 and steps 2–4 cluster combinations. Gold, silver and bronze colours denote performance level classification according to criteria of McKay et al. (2022)

Taking it to the field

This series of coordination studies was designed to quantify the dynamic and interrelated motion of different segments during sprinting, to supplement the more typical assessment of positions at specific time points. The findings from this second study in the series of three give coaches a framework to assess both positions AND patterns, each of which can be analysed with high speed video.

Key things to look for:

• After block exit, there will typically be a brief period of increasing thigh separation. Notice whether this is dominated by pushing with the back leg or driving the front knee, and whether the “float” period occurs when the back leg starts to flex before the front leg reverses direction. The float followed by trail limb dominant switching was a feature of the better sprinters in our study.

• You will probably observe shin block pre-contact and shin roll after touchdown in step 1. This can be effective in combination with the thigh interchange described above. Check the position of the athlete at ground contact, looking to see that the shin is still angled forwards and that the contact point is directly below or slightly behind the centre of mass.

• If the lead leg is tucked at block exit (shin horizontal, foot vertical), shin roll might be absent because the athlete doesn’t not need to use much shin rotation before contact or during stance. This will look more like a piston action of the shin, with ankle dorsiflexion occurring mostly through motion of the foot. This is less common and may rely on specific musculoskeletal strength capacities.

• In steps 2-4, there will likely be an asymmetrical scissor action where one limb reverses direction before the other. So, there’s a brief period of in-phase rotation either just after toe off, where the trail leg continues extending, or just before the next toe off where the swing limb has reached maximum flexion and starts to retract. The latter was present in most of our elite and world class sprinters, but some did display the former with equal effectiveness.

This study allowed us to point to some features of sprint acceleration coordination that are preferable over others. However, there is no single movement pattern that works equally well for every athlete, because of differences in individual constraints. So, the next study in this series went on to investigate how strength interacts with coordination and acceleration performance.

--

References

Alt, T., Oeppert, T., Zedler, M., et al. (2022). A novel guideline for the analysis of linear acceleration mechanics – outlining a conceptual framework of ‘shin roll’ motion. Sports Biomechanics, 24:2, 215-232. DOI: 10.1080/14763141.2022.2094827

Donaldson, B., Bezodis, N., & Bayne, H. (2023). Characterising coordination strategies during initial acceleration in sprinters ranging from highly trained to world class. Journal of Sports Sciences, 41(19), 1768–1778. DOI: 10.1080/02640414.2023.2298100

Kimura, A., Yokozawa, T., Ozaki, H. (2021). Clarifying the biomechanical concept of coordination through comparison with coordination in motor control. Frontiers in Sports and Active Living. DOI: 10.1080/14763141.2022.2094827

McKay, A., Stellingwerf, T., Smith, E., et al. (2022). Defining training and performance caliber: A participant classification framework. International Journal of Sports Physiology and Performance. 17, 317-331. DOI: 10.1123/ijspp.2021-0451