On the filtering of intersegmental loads during running

W Brent Edwards,Karen L Troy,Timothy R Derrick

doi:10.1016/j.gaitpost.2011.06.006

W Brent Edwards, Karen L Troy + Show 1 more

Open Access

https://doi.org/10.1016/j.gaitpost.2011.06.006

Copy DOI

Abstract

When performing inverse dynamics analysis, smoothing kinematic and force platform data at different cutoff frequencies creates an “impact” like artifact that is visible in the joint moments during impulsive activity. Here we illustrate a processing technique in which inverse dynamics analysis is performed on the raw kinematic and force platform data and the joint reaction forces and moments are subsequently smoothed based on the frequency content of the distal reaction force. The effectiveness of this technique is illustrated on forward dynamics simulation data with known intersegmental loads. We then apply the technique to an experimental data set of 10 subjects running at three prescribed speeds. We show that performing inverse dynamics on the raw data and subsequently smoothing the intersegmental loads results in minimal attenuation of the joint reaction force and avoids impact artifacts in the joint moments. Artifacts that occur using a traditional filtering technique are systematic, become more pronounced with speed, and are most noticeable at the hip joint.

Highlights

Intersegmental forces and moments can provide insight into the internal loading of joints and coordination of movements during impulsive activity such as landing and running
In the first, ground reaction forces were smoothed at 100 Hz and kinematics were smoothed at a cutoff frequency corresponding to the mean 98th percentile frequency of all kinematic signals; inverse dynamics was performed
Cutoff percentiles were chosen from optimization, and corresponded to the cutoff frequencies that minimized the average root mean squared error (RMSE) for each joint between inverse dynamics results and known intersegmental loads

Summary

Introduction

Intersegmental forces and moments can provide insight into the internal loading of joints and coordination of movements during impulsive activity such as landing and running. Joint moments free from artifacts can be resolved if kinematic and force platform data are smoothed at a relatively low, identical cutoff frequency (e.g., 20 Hz) [5] This approach is not ideal when reaction forces are of interest because the impact portion of the reaction force, which has major frequencies between 10 and 30 Hz during running [8], is attenuated. The purpose of this study is to illustrate a filtering technique in which inverse dynamics is performed on the raw kinematic and force platform data, and the resulting moments and reaction forces are subsequently smoothed based on the frequency content of the reaction force (similar to White and Podraza [7]) This technique is first validated on forward dynamics simulation data with known intersegmental loads [6]. Joint moments and reaction forces are calculated for a group of ten subjects running at three selected speeds

Objectives

Methods

Results