LAUSR

The purpose of this study was to compare three different methods of data analysis: conventional, Statistical Non-Parametric Mapping (SnPM) and Functional Data Analysis (FDA) for a series of kinematic data acquired from soccer players from their dominant and non-dominant leg at the late swing phase while sprinting. Kinematic data was acquired via Vicon motion analysis system at 200 Hz. For conventional computation, local maximum, mean and minimum values were extracted for further analysis. A two-tailed dependent t-test was implemented for statistical analysis for the conventional method. A two-tailed SnPM and a one-tailed FDA permutation test was performed on the same data (α=0.05). All the trials where filtered using zero-lag 4th order Butterworth filter. The cutoff frequency (13 Hz) was calculated using the optimum cutoff frequency formula. Late-swing phase was defined as the moment where the knee started to extend from its flexed position in the 2nd half of swing phase till foot contact. The conventional results show there was a difference in the mean knee angle in the sagittal plane between the dominant (θo=73.0) and non-dominant leg (θo=74.7) (p=0.04). Both SnPM (% late swing phase found significant=28.9%–86.1%, t-value=2.52, p<0.001) and FDA (% late swing phase found significant=28.1%–87.7%, t-value=2.47–2.55, p<0.001) exhibited a significant difference between the knee angle of the dominant and non-dominant leg at the late swing phase. However, the SnPM demonstrated reliability in producing result while FDA presented a fluctuation, although minor, in the calculation of p value. SnPM would produce more reliable and comprehensive result in comparison to FDA and conventional methods. SnPM enables scientist to identify deviation in movement pattern in soccer players which is not limited to a single point in time and can identify deficiency in movement pattern which may hinder performance or lead to injury.