Author(s): KLEIN, J., OEPPERT, T.1,2, SCHAPSCHRÖER, M.1, ALT, T.3, WILLWACHER, S.4, ABEL, T.1, Institution: GERMAN SPORT UNIVERSITY COLOGNE, Country: GERMANY, Abstract-ID: 2311

INTRODUCTION:
World-class sprinters generate horizontal power outputs of up to 36 W/kg during acceleration and vertical ground reaction forces of > 4.5x body mass to reach speeds of more than 12 m/s. Sprint performance is therefore tightly governed by lower-limb neuromuscular capacities. Despite this mechanistic link, empirical studies frequently report inconsistent associations between isolated strength and power tests and sprint performance [1,2]. This suggests that group-level interpretation of single test metrics may obscure meaningful individual neuromuscular characteristics, while inter-individual variability may reflect distinct neuromuscular performance profiles (NPP) within elite sprinters. Accordingly, this study aimed to explore NPP to improve test interpretation and identify profile-specific performance indicators.
METHODS:
36 male (100m PR: 10.06 –10.78 s) and 28 female (100m PR: 11.40–12.03 s) elite sprinters completed two trials of countermovement jump (CMJ), drop jump (DJ) and run-specific hip- and ankle isometric push tests [3] on a dual force plate system. A matrix of biomechanical features (64 subjects x 24 features) was constructed, with metrics averaged across trials and normalized to body mass. Unsupervised machine learning was applied to identify NPP, with cluster quality evaluated across combinations of dimensionality reduction and clustering algorithms. Between-cluster differences were quantified using one-way ANOVA with significance defined as P < 0.05 and effect-size calculated as eta_p^2.
RESULTS:
Combining principal component analysis (PCA) with K-means clustering resulted in the highest cluster quality (Silhouette = 0.45, Davies–Bouldin = 0.82, Calinski–Harabasz = 75.87). PCA revealed five factors with an Eigenvalue > 1, explaining 86.96% of variance in the dataset. Three distinct NPP were identified: (1) high force / high restitution profile, characterized by rapid force production and high stiffness; (2) low velocity / high impulse profile, relying on prolonged, concentric-dominant force application; and (3) a low stiffness / low power profile, with low rate of force development. Clusters differed significantly in CMJ jump height, DJ reactive strength index and ankle and hip-push-test relative peak vertical force (p<0.001; eta_p^2= [0.36, 0.54]).
CONCLUSION:
Even within a homogenous cohort of elite sprinters, three distinct NPP were evident, highlighting meaningful inter-cluster differences in physical capacities. Acknowledging these distinct NPP may help explaining previously reported inconsistent associations between isolated metrics and sprint performance, while enabling researcher and coaches to identify cluster-specific performance patterns and establish NPP-specific benchmarks to guide training and rehabilitation.
REFERENCES:
1) Healy et al. (2019) J Strength Cond Res
2) Loturco et al. (2019) J Strength Cond Res
3) Ryan et al. (2025) Phys Ther Sport