Development of a universal measure of quadrupedal forelimb-hindlimb coordination using digital motion capture and computerised analysis

Abstract

Background: Clinical spinal cord injury in domestic dogs provides a model population in which to test the efficacy of putative therapeutic interventions for human spinal cord injury. To achieve this potential a robust method of functional analysis is required so that statistical comparison of numerical data derived from treated and control animals can be achieved.

Results: In this study we describe the use of digital motion capture equipment combined with mathematical analysis to derive a simple quantitative parameter - 'the mean diagonal coupling interval' - to describe coordination between forelimb and hindlimb movement. In normal dogs this parameter is independent of size, conformation, speed of walking or gait pattern. We show here that mean diagonal coupling interval is highly sensitive to alterations in forelimb-hindlimb coordination in dogs that have suffered spinal cord injury, and can be accurately quantified, but is unaffected by orthopaedic perturbations of gait.

Conclusion: Mean diagonal coupling interval is an easily derived, highly robust measurement that provides an ideal method to compare the functional effect of therapeutic interventions after spinal cord injury in quadrupeds.

Figure 1

QTM plot illustrating sine wave-like movement of fore and hind paws in the x-plane during treadmill walking in a normal dog. SCD indicates the step cycle duration, measured between consecutive treadmill contact by the left forelimb paw; HCI indicates the left homolateral coupling interval, measured between consecutive contact between the left hindlimb and left forelimb paws; DCI indicates diagonal coupling interval, measured between treadmill contact made by the left hindlimb paw and the right forelimb paw.

Figure 2

Histograms illustrating the variability in normalised (a) left homolateral and (b) diagonal (left hind-right fore) coupling intervals between normal individuals. One-way ANOVA confirms a significant difference amongst the individuals [for lateral coupling: F(10,681) = 267.7; p < 0.0001; for diagonal coupling: F(10,684) = 198.9; p < 0.0001). Tukey's post hoc tests confirmed significant difference in this parameter between many specific pairs of dogs.

Figure 3

QTM plot to illustrate the calculation of the mean diagonal coupling interval. T1 illustrates the interval between left hindlimb and right forelimb paw touching the treadmill at the beginning of the recorded sequence and Tn illustrates the interval between the same limb pair at the end of this (short) sequence. a: in a normal dog there is ~0 time difference between T1 and Tn. b: in a SCI dog there is a large change between T1 and Tn – note that the hindlimb paw movement is considerably delayed. The lines between forelimb and hindlimb plots indicate sequential diagonal pairs and highlight the change in temporal sequence during the recording period, which in this instance is sufficient to alter the sequence of hind and fore paw contact within the period of this recording. The mean diagonal coupling interval is derived by dividing the total difference between T1 and Tn in a sequence by the number of steps (see Methods).

Figure 4

a: Box and whisker plot illustrating normalised diagonal coupling intervals in normal, lame and SCI dogs. Kruskal-Wallis test revealed a significant difference (p < 0.0001) amongst the groups; post hoc testing confirmed a significant difference between SCI dogs and both other groups (Dunn's multiple comparison tests, p < 0.001) but no difference between lame and normal dogs. b: Scatterplot of the mean diagonal coupling intervals in normal dogs with and without abdominal band support, lame and SCI dogs. Kruskal-Wallis test revealed a significant difference (p < 0.0001) amongst the groups; there were significant differences between SCI dogs and all other groups (Dunn's multiple comparison tests: p < 0.001 SCI versus normal and dogs with support; p < 0.01 SCI dogs versus lame dogs) but no differences between lame dogs and normal dogs, or between normal dogs with and without band support (Wilcoxon signed ranks test, p = 0.813).

Figure 5

Representative plots of mean diagonal coupling interval in (a) a normal dog and (b) a SCI dog – note differences in y-axis scale. In the normal dog, although there is some variability in both step cycle duration (SCD) and diagonal coupling (LHRF) there is minimal change in this interval between step 1 and step 41. In contrast, in the SCI dog the interval is greatly prolonged, reaching ~800 ms during the same period of 40 step cycles.

Figure 6

'Screenshot' image obtained from Qualisys software incorporating a representation of a dog walking on the treadmill and the orientation of the x-, y- and z- planes in relation to the direction of forward motion of the dog.