Author(s): REHORST, S., MAI, C., SIKELER, L., BRAUN, L., ANEDDA, B., VON DIECKEN, C., WILLWACHER, S., Institution: OFFENBURG UNIVERSITY OF APPLIED SCIENCES, Country: GERMANY, Abstract-ID: 997

INTRODUCTION:
During maximum-speed sprinting, athletes can temporarily maintain velocity despite accumulating fatigue and reduced muscle activation [1]. Fatigue may induce compensatory coordination changes [2,3], yet most research describes kinematic changes during speed loss rather than when speed is held constant. Therefore, this study examined whether fatigue-related kinematic adaptations during constant-speed sprinting follow specific temporal patterns and where within the stride cycle they occur.
METHODS:
Nineteen competitive sprinters performed two constant-speed treadmill runs at maximum speed until exhaustion. Markerless motion capture (200 Hz) recorded whole-body kinematics. Sagittal hip, knee, and ankle angles were extracted and strides time-normalized. Fatigue was operationalized as relative progression throughout the trial, with time-normalized stride angle curves sequentially aggregated to 10 equally spaced mean waveforms spanning 0–100% of trial and represented using cubic B-splines. Functional principal component analysis identified dominant variations in spline coefficients. PC scores were analyzed across bins using linear mixed-effects models testing linear (assuming continuous change), quadratic (assuming curvilinear change), and hinge (assuming piecewise linear change with estimated breakpoint) trends, with athlete and trial as random intercepts. Reconstructed waveforms quantified magnitude and timing of change. Variance components and ICC assessed between-athlete variability.
RESULTS:
The first three PCs captured most joint-angle variability (ankle 77.8%, knee 81.0%, hip 87.6%) and were analyzed further. Mixed-model comparisons showed predominant hinge-type temporal trends across joints with mean breakpoint at 34% ± 20% (mean ± SD) of trial. Only the 2nd PC at the knee favored a significant quadratic fit. Changes in reconstructed waveforms were concentrated between 38-78% (ankle; mean: +4.0 deg plantarflexion), 26-76% (hip; mean: +3.0 deg extension), and 60-88% (knee; mean: +2.0 deg flexion) of the stride cycle with peak values of +8.8 deg ankle plantarflexion, +6.0 deg hip extension, and +3.9 deg knee flexion at 57%, 55%, and 76% of stride. Between-participant differences accounted for most of the variability in PC scores (ICC 0.73-0.96).
CONCLUSION:
Kinematics changed in hinge-type temporal patterns with breakpoints around the first 3rd of trial, indicating linear fatigue-related shifts after a short accommodation phase. Fatigue-related adaptations were mainly attributable to propulsion and leg repositioning rather than to pre and early contact commonly associated with peak vertical force generation. These findings provide new insight into how sprint kinematics reorganize under accumulating fatigue when running speed is maintained.

References
[1] Ross A, Leveritt M, & Riek S. Sports Medicine 2001; 31(6):409-425.
[2] Hicks DS, McMillan S, Schollhorn W, van den Tillaar R. Sports Medicine 2026.
[3] Trezise J, Bartlett R, Bussey M. Int J Sports Sci Coach 2011; 6(3):357-363.