Effects of Offset-Normalizing Techniques on Variability in Motion Analysis Data

in Journal of Applied Biomechanics
View More View Less
  • 1 Michigan State University
Restricted access

Purchase article

USD  $24.95

Student 1 year online subscription

USD  $88.00

1 year online subscription

USD  $118.00

Student 2 year online subscription

USD  $168.00

2 year online subscription

USD  $224.00

This study assessed the effect of offset normalizations on variability in kinematic data. The tarsal angles for 12 elderly horses, with mild lameness of the tarsal joint, were measured at the trot pre and post 2 weeks administration of a dietary supplement intended to promote joint health (Corta-Flx, Nature's Own, Aiken, SC). For five strides, pre- and postsupplement, the tarsal angles measured on the flexor side (full exten. = 180°) were smoothed, normalized to 101 data points, and averaged. Four offset normalizations were applied: minus standing tarsal angle (Tarsal); minus impact angle (Impact); minus mean angle (Average); multiplicative scatter correction (MSC). For 11 angle variables across the stride there were no significant differences pre- and postsupplement, p > 0.05. Normalization had no effect on the timing of variables or magnitude of angles, but generally the variability in the angles was reduced. Least to greatest reduction occurred with the Tarsal, Impact, Average, then MSC normalizations. The Average and MSC techniques resulted in two and three variables, respectively, becoming significantly different. These differences were small, emphasizing that significant findings should be interpreted for meaningfulness. Normalizations based on the data gave the largest reductions in variability, but these may introduce biases into the data. Thus, normalization with respect to measurements external to data capture may be preferable, but their theoretical and statistical relationship to the kinematic variables should be confirmed. MSC altered the shape of the kinematic trace, which may be misleading. Offset normalizations should be used with care, but they can reduce variability in kinematic data to increase statistical power in biomechanical studies.

The authors are with the Mary Anne McPhail Equine Performance Center, Large Animal Clinical Sciences, College of Veterinary Medicine, Michigan State University, East Lansing, MI 48824-1314.

All Time Past Year Past 30 Days
Abstract Views 169 137 76
Full Text Views 2 1 0
PDF Downloads 4 3 0