Spectral Analysis of Impact Accelerations Using Bone Versus Surface Mounted Accelerometers

Abstract

Listen

Translate article

The heel-strike transient, during running, results in an impact acceleration that gets propagated up the musculoskeletal system. Depending on its magnitude, rate, and frequency the acceleration may have positive or negative physiological effects. Traditionally, impact acceleration has been measured by skin-mounted accelerometers fixed to the distal anteriomedial tibia. Previous research using accelerometers attached to pins inserted in bone suggested that skin-mounted accelerometers may overestimate impact acceleration by 8 to 50%. PURPOSE: To determine if skin-mounted accelerometers result in attenuation or gain in impact acceleration when compared to bone-mounted accelerometers. METHODS: Two human cadaver lower extremities were mounted via an intra-medullary rod to a dynamic gait simulator. Ankle angle was controlled prior to impact by applying tension to the anterior tibialis muscle. Two uniaxial accelerometers were attached to the distal anteriomedial tibia, one directly to the bone and another to the skin. A Fast Fourier Transformation was used to determine the power spectra of the impact phase accelerations for 24 trials. The power spectral densities (PSD) of the bone-mounted accelerometer (BA) and skin-mounted accelerometer (SA) were used to calculate transfer functions for impact frequencies between 15 and 25 Hz. The overall transfer was calculated as the average transfer for the impact frequencies. Effect size was also calculated to determine the strength of difference between the average PSDs of the BA and SA for the impact frequencies. RESULTS: The dynamic gait simulator produced peak impact accelerations equivocal to those observed during normal walking/running ranging from 1.6 to 5.3 g. Peak impact acceleration for BA was 0.5 ± 0.4 g larger than SA and the transfer functions revealed that the SA impact attenuation was −1.9 dB. Effect size for average PSDs was −0.56 indicating a moderate difference between BA and SA. CONCLUSION: Although an overall impact attenuation was found in SA, the difference is small. If dynamic mechanical loading on bone is an underlying mechanism in bone adaptation, these preliminary results suggest that SAs may be a good predictor of skeletal impact acceleration. Future research will look to increase sample size as well as peak impact accelerations to see if this relationship can be applied to all subjects and impact accelerations typically seen at faster running velocities.