Coordination of the Ankle Joint Complex During Walking

U. Rattanaprasert1, R. Smith1, & N. O’Dwyer2

1. School of Exercise and Sport Science, University of Sydney, Sydney, Australia
2. School of Physiotherapy, University of Sydney, Sydney, Australia


A simple model of the joint between the leg and the rearfoot is that of two hinges each with one degree of freedom. The superior hinge at the talo-crural joint caters for most of the sagittal plane motion due to its medio-lateral orientation. Motion in the transverse and coronal planes is largely taken up by the subtalar joint which has an inclination of 42° to the transverse plane and deviates 23° to the medial side. It could be expected from the orientation of this joint that motion in the coronal plane (inversion/eversion) would be tightly coupled to rearfoot motion in the transverse plane (adduction/abduction). A simple, non-viscous mechanical hinge joint would produce an in-phase, fixed gain relationship between inversion/eversion and adduction/abduction. This relation could be adequately quantified by Pearson’s product moment correlation (rP). In contrast, muscular control of the joint, which introduces visco-elastic elements and time-delayed feedback loops and control lines, could be characterised by more complex phase and gain relations among the frequency components of the signals. This would require a dynamic analysis to quantify the relation between the angles. In the present study, both Pearson’s correlation and linear systems analysis were employed.

Five trials of one walking stride (stance and swing phases) were videographed (at 30 Hz) with three markers each placed on the leg and rearfoot of 43 normal adults. Using inversion/eversion as the input and adduction/abduction as the output, the distribution of the variance of the angle time series according to their frequency spectra was computed. The output was separated into a coherent component which is linearly correlated with the input and an incoherent component not correlated with the input. The ratio of the coherent component to the total output is measured by the coherence square function. This quantifies the proportion of the total variance accounted for by the linear dynamic relation between the angles and is analogous to the correlation squared (rP2).

The mean rP2 was 0.23, whereas the mean overall coherence was 0.78, thus accounting for at least three times the variance. This result suggests that while there may be a degree of simple mechanical coupling between the angles, muscular control takes the major role.