Acceleration Coordination
- helenbayne
- Jun 4
- 7 min read
Over a decade at the University of Pretoria, I was fortunate to work with sprinters across the development pathway, from the TukSport High School academy program to senior international medallists and Olympians. I was equally fortunate to have access to research-grade technology and an environment that supported the integration of research and applied sport science. The final key ingredient for impact in science and practice was a group of willing and capable collaborators – coaches, athletes and scientists – that shared their knowledge of and enthusiasm for sprint mechanics (and patience with my constant questions and technological experimentation!). This post will describe the motivation, methods, and main findings of a recent paper from that group.
Initial acceleration in sprinting requires force application to move the centre of mass (CoM) rapidly forwards and gradually upwards with each step. Sprinters use a complex sequence of muscle actions to translate and rotate the segments of the body to position themselves for the effective application of these forces during ground contact and to reposition during flight in preparation for the next contact. Watching, assessing and refining the way that athletes move is at the crux of coaching sprint technique.
A kinogram-type approach is a core tool for coaches when visually reviewing an athlete’s technique, and one that I’ve used extensively in athletics and other sports. However, after assessing the position of an athlete at key events, the discussion tends to move on to how they are transitioning in and out of these positions. Until recently, the science behind sprint acceleration kinematics has also largely focused on describing individual joint angles at key points in time, ignoring both the periods in between these time points and the important dynamic relationships between segments that occur during these periods. So, to improve our measurement and understanding of sprinters’ movement patterns, I began to explore methods of coordination analysis. This led to the development of a PhD program, excellently conducted by Byron Donaldson and co-supervised by Neil Bezodis, involving a series of studies to (i) quantify coordination in highly trained to world class sprinters, (ii) categorise athletes based on multi-segment coordination patterns and explore the implications for acceleration performance, and (iii) investigate a range of strength capacities as individual constraints that influence coordination during initial sprint acceleration. This blog post will summarise the first of those three studies.
(If you want to read ahead, Byron’s full thesis is available to download at the University of Pretoria’s research repository)
Quantifying coordination
A common way to visualise intersegmental coordination is an angle-angle plot, (Figure 1, left) where each datapoint represents the orientation of the two segments of interest (distal segment on x-axis and proximal segment on y-axis, starting at the diamond symbol and reading anti-clockwise in the pictured example). Although this can be a useful visual tool, it doesn’t immediately provide objective data to quantify inter-segmental coordination.
The method we chose to address this is known as vector coding, which can ultimately be used to classify coordination across the movement cycle into one of eight colour coded “bins”. To get there, we start with the angle-angle plot, look at the vector formed between two consecutive datapoints and measure the angle of this vector relative to the right horizontal. This “coupling angle” can lie anywhere between 0° to 360° and four examples are shown in the figure below. When two segments are rotating in the same direction, we refer to the coordination pattern as “in-phase”. Conversely “anti-phase” coordination is when the two segments are rotating in opposite directions. If both segments are rotating in the positive direction, the vector will fall between 0° an 90° and if they’re moving in the negative direction the vector will be between 180° and 270°. The two remaining quadrants represent anti-phase coordination, with proximal segment positive rotation and distal negative (90° to 180°) and vice versa for coupling angles between 270° and 360°. The description can be a bit of a mouthful so, to simplify communication, we colour code the four quadrants! Finally, we add one more important level of detail, which is shown by the light and dark colour shading within each quadrant. This tells us which of the two segments is rotating more than the other and is valuable because at either edge of the quadrant one of the two segments is close to stationary (e.g., when the vector is 45° both segments are rotating through the same range, at 1° the distal segment dominates rotation while the proximal segment is almost stationary, and vice versa at 89°).
Figure 1. Left: example angle-angle plot with the distal segment on the y-axis and proximal segment on the x-axis; four quadrants of coupling angles (CA) shown: in-phase ++, anti-phase +-, in-phase --, anti-phase -+. Right: Coupling angle chart with eight colour coded bins.
Measuring world class sprinters
Now to the sprinting study, where we collected biomechanical data during a regular training session. In our applied research model, ongoing biomechanical support in the sprint programs enabled a rare opportunity for research in highly-trained to world class athletes, simultaneously providing coaches with objective data to guide and inform decisions, thereby supporting athlete development and performance.
Twenty-one sprinters, with 100 m PBs ranging from 9.89 s – 11.19 s in the males and 11.46 s – 12.14 s in the females, took part in the study. During a training session involving maximum acceleration efforts from a block start, athletes wore Noraxon inertial measurement units (IMUs) on the feet, shanks, thighs and trunk. We identified the best trial for each athlete (fastest time to 30 m), extracted the sagittal plane segment orientation data and ran vector coding procedures (as described above) for thigh-thigh, trunk-shank and shank-foot coordination for each of the first four steps. A step was defined from toe off to toe off (block clearance being the start of step 1) and trail and lead legs were defined based on their position at the start of the step (noting that the limbs switch during the step so that the leg denoted as the trail leg is actually “leading” at the end of the step). Positive rotation direction was defined as clockwise, if viewing the sprinter from the right hand side.
Figure 2. Visual summary of the research methods – study population, measurement tools, data processing and reporting. The vertical black line within each step shows the instant of ground contact, and the dashed vertical line shows when the thighs switched. BC: block clearance, TD: touchdown, TO: toe off.
The way they move
The results are summarised in Figure 3, where the coloured bars represent the group mean and the black bar charts below them show the standard deviation (representing variance between athletes) at each corresponding instant in time. Byron's PhD research drew on dynamical systems theory, where coordination patterns emerge from interacting task, individual and environmental constraints. With this framework in mind, periods of low between-athlete variance could be thought of as sections of the step with stronger task constraints, limiting the available movement solutions. Also, the unique coordination patterns occurring in step 1 compared to steps 2-4 show the effects of the distinct constraints of the block exit in sprinting.
Figure 3. Summary of the coordination results. Within each step, the vertical black line shows the instant of ground contact, and the dashed vertical line shows when the thighs switched.
Thigh-thigh coordination was predominantly anti-phase and trail leg dominant (dark red), but became lead leg dominant (light red) in the last ~20% of the step. In other words, the switching motion of the thighs is asymmetrical as the trail leg rotates more than the lead leg between any two timepoints for most of the step, reducing prior to the next toe off where the thigh of the leg that’s now in contact with the ground dominates the thigh-thigh coordination pattern. Step 1 differed from steps 2-4 just after toe off (block clearance in step 1), where there was a brief period of anti-phase coordination as thigh separation increased before they reverse direction. The standard deviation across the group was high at the very start and end of the step, demonstrating that athletes vary in the relative timing of reversing the direction of thigh motion around toe off (we’ll explore this in more detail in part 2 of this blog series!).
There were three key periods of shank-foot coordination: preparation for ground contact, ankle dorsiflexion in early stance, and plantarflexion in late stance. In steps 2-4, there is more in-phase clockwise (purple) rotation before contact compared to step 1, where the leg swings out more (in-phase anticlockwise, blue) with late if any change in direction of shank rotation (“shin block”). The foot is at a more vertical angle and the shin more horizontal at contact in step 1 compared to the later steps, and so dorsiflexion is foot dominant (dark green). Then, dorsiflexion occurs mostly through shank motion (dark purple) in steps 2-4, after a relatively more horizontal foot and vertical shin at contact. i.e. more “shin roll”. At the end of stance the shin maintains its relatively low angle and propulsion occurs through foot dominant plantarflexion (light red or light purple).
Trunk-shank coordination was mostly shank dominant, except during the mid to late stance phase in step one, where the trunk rotated relatively more than the shank. This reflects the unique task constraints of coming out of the block start position, where the greatest vertical change in CoM position has to occur in the first step.
Similar, but not all the same
This was part one of the investigation and described the general coordination patterns observed when averaged across the group of high level sprinters. The areas of individual variation sparked further interest and tied in with what we’d observed in the field. That is, different sprinters who are capable of achieving similar performance outcomes can actually adopt quite different movement patterns to one another. But, the question remained as to what types and how much variation should be allowed or encouraged without compromising performance. Coaches have also used categorisation as a way to tailor their methods to similar “types” of athletes based on musculoskeletal strengths and/or whole body movement strategies. We wanted to find out whether athletes could be categorised based on coordination patterns and whether different patterns were associated with differences in acceleration performance, or not. This led us to the second part of the research study, which will be summarised in the next post.
--
I’ve only captured the main coordination results here. For more in-depth analysis and additional data, including spatiotemporal parameters and continuous and discrete measures of individual segment kinematics, read the full (open access) paper:
Kommentare