Spanning set defines variability in locomotive patterns.

The purpose of the investigation was to use the spanning set methodology to quantify variability in locomotive patterns and to compare this method with traditional measures of variability. Subjects ran on a treadmill while sagittal plane kinematic data were collected with a high-speed (180 Hz) camera. Changes in variability were evaluated as the subject ran barefoot and in shoes. Mean ensemble curves for the knee angle during the stance period were created for each condition. From these curves, traditional measures of variability were calculated using the coefficients of variation (CVs), and the mean deviation (MD). Spanning set vectors were defined from the coefficients of polynomials that were fitted to the respective standard deviation curves. The magnitude of the spanning set was determined by calculating the norm of the difference between the two vectors. The normalised difference between the two conditions was 6.6%, 6.9% and 98%, for the MD, CV and spanning sets, respectively. The results indicated that the spanning set was capable of statistically (p < 0.05) determining differences in variability between the two conditions. CV and MD measures were unable to detect statistical differences (p > 0.05) between the two conditions. The spanning set provides an alternative, and sensitive measure for evaluating differences in variability from the mean ensemble curve.