A Torque-Event Decomposition (TED) Framework for the Analysis of Pedaling Kinetics Using High-Resolution Signals

Laboratory and field systems can provide high-resolution torque and crank-angle signals, but analyses commonly reference fixed crank angles that may not coincide with mechanically meaningful events. This study introduced torque-event decomposition (TED), a framework that localizes signal-defined events in unilateral- or total-torque profiles and, when bilateral signals are available, quantifies their unilateral composition. TED was applied to 2393 cycles from 12 trained cyclists at ~91 rpm and ~221 W. Ten events were localized in every analyzed cycle, whereas two left-torque zero-crossing events could not be identified in two cycles because no corresponding crossings were detected within the predefined search intervals. Positive unilateral torque began ~15° before top dead center and ended ~20° after bottom dead center. Total-torque peaks occurred 6–10° after 90°/270°, whereas minima occurred 15–16° after 0°/180°. Total-torque peaks showed closer agreement with unilateral peaks (mean absolute difference: 4.1–4.5°; Lin’s concordance correlation coefficient: 0.718–0.813) and occurred when the extending lower limb produced near-maximal torque. Minima showed weaker agreement with unilateral positive-to-negative transitions (6.2–7.6°; 0.355–0.381), reflecting more complex inter-limb transitions. These findings support the use of signal-defined events to characterize kinetic timing beyond fixed crank positions and demonstrate the feasibility of the TED framework under the conditions studied, providing a basis for further validation across cycling conditions, populations and measurement systems.