Agreement Level of Running Temporal Measurements, Kinetics, and Force-Time Curves Calculated from Inertial Measurement Units

Inertial measurement units (IMUs) and wearable sensors have enabled athlete monitoring and research to become more ecologically valid due to their small size and low cost. IMUs and accelerometers that are placed on the body close to the point of impact and that record at sufficiently high frequencies have demonstrated the highest validity when measuring temporal gait event moments such as ground contact time (GCT) and flight time (FT) as well as peak forces (PF) during upright running. While the use of IMUs has increased in the sport performance and athlete monitoring realm, the potential of the technology’s ability to estimate running force-time curves utilizing the two-mass model (TMM) remains unexplored. The purpose of this study was two-fold. First, was to determine the validity of measuring temporal gait events and peak forces utilizing a commercially available shank-mounted inertial measurement unit. Second, was to determine the validity of force-time curves generated from the TMM utilizing data from shank-mounted inertial measurement units. Ten subjects voluntarily completed submaximal treadmill tests equipped with a force plate while wearing shank-mounted IMUs on each leg. Using the raw data from the IMUs, GCT, FT, total step time (ST), PF, and two-mass model-based force-time (F-t) curves were generated for 25 steps at 8 different speeds. Paired sample T-tests were performed on the gait events and peak force between the IMU and treadmill with both individual step comparison and averages per each speed. 95% confidence intervals were calculated for each timepoint of the force time curves. No statistically significant differences (p \u3e 0.05) and nearly perfect relationships were observed for the step averages for each speed with FT, ST, and PF. Confidence intervals of the corrected mean difference suggest that F-t curves calculated from the TMM may not be valid when assessing the running population as a whole. When performing a sub-group analysis of skilled runners and recreational runners, F-t curves derived from shank-mounted IMUs may be more valid in skilled runners than recreational runners. In skilled runners, the 95% CI for the mean difference contained zero within the first 60% of the GCT duration, whereas the 95% CI recreational runners contained a zero-value in a smaller percentage of the GCT located only in the middle of the GCT at the curve peak height. The results of this study suggest that interchangeability between shank-mounted IMUs and force plates may be very limited when estimating temporal gait events and kinetics. While agreement was low between F-t curves after the peak in skilled runners, use of shank-mounted IMUs to estimate F-t curves may have several benefits still in skilled runners when assessing peak forces and force development from initial contact until peak force