Validity Of Peak Tibial Acceleration Using Wearable Accelerometers During Running

Abstract

With wearable technology becoming more popular, more companies are creating sensors to assess biomechanical parameters including peak tibial acceleration (PTA) during running. However, validity of data from wearable technology relative to “gold-standard” research-grade instruments is highly important. PURPOSE: Assess the difference in PTA obtained from wearable inertial measurement units (IMU) and a research-grade accelerometer at different running speeds. METHODS: Six participants completed 1-2min treadmill running bouts at 3.0 m/s and 4.0 m/s while wearing standardized footwear (1080, New Balance). A research-grade tri-axial accelerometer (ACC; 1200Hz, model 356A26, PCB Piezotronics) and a 9-axis IMU (1000Hz, Blue Trident, IMeasureU) were secured to the distal tibia to capture PTA during running. The testing at both speeds was completed with the ACC below and above the IMU to account for possible position effects. Data were collected for the final 15 seconds of each running bout and the average of both positions for 10-15 peaks of axial tibial acceleration were used for analyses. Paired t-tests and Cohen’s d effect sizes were calculated to compare instrument PTA means at different speeds. RESULTS: At 3.0 m/s, mean PTA was not different between ACC (5.2±1.9 G) and IMU (5.9±2.2 G; p = 0.33; d = 0.34). At 4.0 m/s, mean PTA was not different between ACC (8.5±3.7 G) and IMU (9.8±4.2 G; p = 0.33; d = 0.33). Figure 1 demonstrates the individual variability in the difference in PTA obtained from the ACC and IMU at different speeds (A) and different positions (B). CONCLUSION: Despite the statistically similar PTA means obtained from both instruments, the difference in PTA between ACC and IMU appears to be highly variable among individuals. This variability in PTA may be due to differences between devices, device placement, attachment method, or individual running style. This highlights the difficulty in controlling sources of variability during testing.