| Sumario: | Peak tibial acceleration (PTA) is a parameter that is often studied in runners. PTA may be associated with tibial stress fractures and is used to measure shock attenuation. In the field setting PTA is measured with sensors that have a lower sampling frequency than sensors used in the lab. A sensor with a lower sampling frequency has the capability for longer measurements due to longer battery life and lower data usage. However, this sacrifices signal detail and thus affects PTA measurements. PURPOSE: Show the effects of a reduced accelerometer sampling frequency on PTA during running by differentiating between a systematic and a random error. METHODS: 13 subjects (7 Female, 6 Male; Mean±1SD: age 22.5y ± 4.0y, height 1.70m ± 0.09m, mass 61.3kg ± 10.3kg) ran for 20 minutes at a self-selected pace on a treadmill. A high resolution accelerometer (1200Hz) was placed on the distal anteromedial tibia. Data were reduced using the decimation function, which applies a low pass filter prior to down sampling. The norm of the acceleration was calculated and PTA was detected at all sampling frequencies for the first 5 minutes of the measurements. Next, the systematic error at reduced sampling frequencies was calculated by subtracting the mean PTA of the reduced sampling frequency by the mean PTA of the original sampling frequency. The random error is calculated as the standard deviation of the differences in PTA at the reduced and original sampling frequency and were corrected for the systematic error to center the random error around zero. RESULTS: Lower sampling frequencies resulted in lower PTA estimates. Error metrics increased with greater data reduction (Table 1). CONCLUSION: Down to 80Hz, the mean random error (5.0 m/s²) is within the standard deviation of the detected peaks (5.1 m/s²). This means that sampling frequencies ≥80Hz may be sufficient to capture changes in PTA within a longer run, though the absolute PTA will be underestimated at 80Hz (systematic error of -13.2m/s²).
|
|---|