Biomechanical features of a novel step-down-and-pivot task in football players with hip/groin pain

Abstract

Identifying biomechanical impairments in young, physically active populations with hip/groin pain is crucial for early intervention. This study characterized the biomechanical features of a novel task, the step-down-and-pivot, in competitive football players, comprising 36 individuals with hip/groin pain (28 ± 6 years) and 11 controls (26 ± 4 years). Experimental motion data and ground forces were input into biomechanical models to calculate joint angles and moments, then transformed into principal components and input into a feature selection pipeline. Ten main biomechanical features were identified for each limb, i.e. the pivot limb and the swing limb, that could discriminate between symptomatic and control groups with p < 0.05. In symptomatic individuals, the main features were as follows: pivot limb: smaller hip flexion and knee extension angles, delayed initiation of hip flexion and increased ankle dorsiflexion moment; swing limb: reduced hip flexion moment, increased hip internal rotation moment, delayed hip adduction and knee extension moments, and reduced ankle dorsiflexion angle. The largest group differences occurred during the transitions from step-down to pivot, and pivot to step-forward, supporting a potential role for multi-phase and/or multi-planar tasks when assessing biomechanical impairments due to hip/groin pain. Although biomechanical alterations in our symptomatic participants were small, they could be identified and characterized using feature selection.

Keywords: biomechanics; feature selection; femoroacetabular impingement; gait analysis; hip osteoarthritis; principal component analysis.

Figure 1.

The step-down-and-pivot task with left limb as the pivot limb and right limb as the swing limb. (A) The participant stands on the step, located 20 cm above the ground. (B) The participant steps down, with the pivot limb striking the force plate FP2 on the ground, and the swing limb remaining on FP1. (C) The participant initiates a 90° counter-clockwise pivot with the pivot limb, during which time the swing limb is lifted from FP1. (D) The swing limb swings around to strike the additional force plate on the ground FP3 before the participant steps forward by lifting the pivot limb from FP2.

Figure 2.

Pivot limb joint angles and moments for the step-down-and-pivot task for symptomatic and control groups. Ensemble waveforms were compared using t-tests with statistical parametric mapping. Grey shaded regions represent regions of statistical significance ($\alpha =0.003125$). Task phases: Step-down; Pivot; and Step-forward. Event labels have format [Limb][Event type][Sequence number]. Limb: P (pivot) or N (swing); Event type: FO (foot-off) or FS (foot-strike). E.g. PFO3 is a foot-off on the pivot limb which is the third event on the pivot limb.

Figure 3.

Swing limb joint angles and moments for the step-down-and-pivot task for symptomatic and control groups. Ensemble waveforms were compared using t-tests with statistical parametric mapping. Grey shaded regions represent regions of statistical significance ($\alpha =0.003125$). Task phases: Step-down; Pivot; and Step-forward. Event labels have format [Limb][Event type][Sequence number]. Limb: P (pivot) or N (swing); Event type: FO (foot-off) or FS (foot-strike). E.g. PFO3 is a foot-off on the pivot limb which is the third event on the pivot limb.

Figure 4.

High principal components (i.e. PC1−3) from the final set of main features on the pivot limb. Top row: waveforms of the pooled original data representing the upper (solid orange) and lower (dashed green) quartiles of the principal component scores for each respective feature. Arrows signify the tendency for symptomatic (red) or control (blue) groups to be associated with the upper or lower quartile, e.g. for θ_HIPFLEX PC1, the control group tended towards the upper quartile, while the symptomatic group tended towards the lower quartile. Bottom row: waveforms of the principal component coefficients (solid black) and variance in the original data explained by that principal component coefficient (dashed black). Task phases: Step-down; Pivot; and Step-forward. Event labels have format [Limb][Event type][Sequence number]. Limb: P (pivot) or N (swing); Event type: FO (foot-off) or FS (foot-strike). E.g. PFO3 is a foot-off on the pivot limb which is the third event on the pivot limb.

Figure 5.

High principal components (i.e. PC1−3) from the final set of main features on the swing limb. Top row: waveforms of the pooled original data representing the upper (solid orange) and lower (dashed green) quartiles of the principal component scores for each respective feature. Arrows signify the tendency for symptomatic (red) or control (blue) groups to be associated with the upper or lower quartile, e.g. for M_HIPROT PC1, the control group tended towards the lower quartile, while the symptomatic group tended towards the upper quartile. Bottom row: waveforms of the principal component coefficients (solid black) and variance in the original data explained by the respective principal component coefficient (dashed black). Task phases: Step-down; Pivot; and Step-forward. Event labels have format [Limb][Event type][Sequence number]. Limb: P (pivot) or N (swing); Event type: FO (foot-off) or FS (foot-strike). E.g. PFO3 is a foot-off on the pivot limb which is the third event on the pivot limb.