Accelerometric analysis of gait in young and elderly subjects: Frequency analysis

Research question: Are there significant differences in gait acceleration signal spectrum between young and elderly subjects? And could these improve the understanding of motor control? Introduction: From signals theory the knowledge of spectrum features is the starting point for the signal processing. In particular, know features of gait acceleration signal spectrum of young and elderly subjects could improve the understanding of the multiple factors that contribute to achieve the specific motor task. Studies about the range of gait acceleration frequency are presented in literature; these start from: kinematic data [1], acquired with stereo photogrammetry, and from ground reaction force data [2], acquired with force platform. The first method is affected by the well-known drifts problem; instead the second measures the acceleration of the center of pressure (COP) and not those of the center of mass that is a follower of the COP. The study aims to assess the features of gait acceleration signal spectrum, directly obtained from inertial measurement unit (IMU), in young and elderly subjects. Materials and methods: The study was conducted on 10 healthy young people and 10 elderly subjects. The participants performed an instrumented over ground gait task wearing an IMU located on the trunk at the height of fifth lumbar vertebra. The acceleration signals in antero-posterior, medio-lateral and vertical directions were acquired with a sample frequency of 128 Hz. The Fast Fourier Transform (FFT) of the signals was used to determine: the frequency corresponding to the maximum of the signal spectrum (fMAX) and the frequency corresponding to the median value of the signal spectrum (fMEDIAN). The potential spectrum density (psd) was obtained with the matlab built-in function pwelch. The signal power was calculated integrating the psd. The percentage power at 32, 16, 10, 8 Hz was calculated. Results: For the young population fMAX varies from 1.9 to 3 Hz and fMEDIAN from 28.9 to 32.4 Hz. Instead for the elderly population fMAX is between 1.0 and 2.6 Hz and fMEDIAN from 26.8 to 34.3 Hz. The power analysis showed for young subjects that the signal power is between 99.9% at 32 Hz and 72.9% at 8 Hz. For the elderly population, instead, it varies form 99.2% at 32 Hz to 44.9% al 8 Hz. Discussion: On the medio-lateral axis the frequency range of the elderly is lower than young subjects; this could be explained by the fact that medio-lateral frequencies are influenced by stride time. Instead, for antero-posterior and vertical directions frequency range of the elderly are shifted to the higher frequencies than those of young; this could be due to the muscularity stiffness present in elderly subjects. The percentage power, in all the directions, is lower in elderly than in young subjects; this could be due to the no smooth body control during gait that introduces components at the high frequencies.