Agreement of Wearable Inertial Sensors with Optoelectronic Motion Capture for Felge-to-Support Temporal Segmentation

Background: Wearable inertial measurement units (IMUs) provide a portable alternative to optoelectronic motion capture, but their performance in complex gymnastics movements requires task-specific evaluation. This study evaluated agreement between IMU- and Vicon-derived temporal phase durations during the felge-to-support on parallel bars. Methods: Four elite male gymnasts performed 10 executions recorded with an 11-camera Vicon system (100 Hz) and three Vicon Blue Trident IMUs (225 Hz). IMU signals were resampled to 100 Hz. Five phases were identified independently in each system from hip flexion–extension events and expressed as percentages of the movement cycle. Agreement was assessed using repeated-measures Bland–Altman analysis, with ICC(A,1), MAE, and RMSE as complementary metrics; cumulative-event analysis examined internal phase boundaries. Results: No phase-by-side comparison showed evidence of systematic mean bias, although limits of agreement varied across phases and limbs. The narrowest limits occurred for right Phase 3 (−2.256 to 2.277%). ICC(A,1) ranged from 0.432 to 0.822, MAE from 0.899% to 3.315%, and RMSE from 1.079% to 3.890%. Cumulative-event agreement was also event- and side-dependent. Conclusions: IMU–Vicon agreement for percentage-based temporal segmentation was phase- and side-dependent. IMUs may support portable monitoring of normalized phase timing under the evaluated conditions, but should not be considered universally interchangeable with Vicon.