Abstract
Wearable technology can nowadays be used to improve para-swimming coaching; however, the extent to which individual anatomy affects features of swimming variability is unclear. Six paralympic swimmers were recruited, their upper-limb segment lengths were measured, and their absolute bilateral limb asymmetry indices (AbsLAIUL) were calculated. They were instrumented with a sacrum-worn inertial measurement unit and performed an in-water, fatiguing, freestyle aerobic test at incrementally faster paces. Stroke-to-stroke outcome and execution variability were calculated, respectively, using sample entropy (SampEn) and fractal dimension (FD) on forward and mediolateral linear acceleration signals. Significantly increased perceived exertion scores (F(4,28) = 154.1, p < 0.001) were observed. Execution and outcome variability increased in the forward (SampEn = F(4,25) = 11.86, p < 0.001; FD = F(4,24) = 6.17, p = 0.001) and mediolateral (SampEn = F(4,25) = 9.46, p < 0.001; FD = F(4,24) = 27.64, p < 0.001) directions. Modelling of FD (only) improved with AbsLAIUL as a covariate (forward = F(1,24) = 9.68, p = 0.005; mediolateral = F(1,24) = 8.57, p = 0.021), suggesting that AbsLAIUL affects only execution, but not outcome, variability. This information could help coaches determine which coordination indices should be personalized when monitoring variability during para-swimming training.