Visual3d

Barbell displacement-time data were exported to Visual 3D (V6.01.22, C-Motion, Rockville, USA), and barbell velocity was calculated using the finite difference method in Visual 3D. Displacement data were filtered using a fourth order, zero-lag, Butterworth low-pass filter with a cut-off frequency of 12 Hz. Data were visually inspected to assess the effect that different cut-off frequencies (6-20 Hz) had on vertical velocity and 12 Hz was selected because lower cut-off frequencies attenuated peak values. The start of the concentric phase of each repetition was determined as the first frame in which the marker displayed a positive vertical velocity following the eccentric phase (bar lowering), and the end of the concentric phase was identified as the first frame in which the marker displayed a negative vertical velocity after the end of the concentric lifting phase. Peak vertical velocity and mean vertical velocity were subsequently determined from the highest values in the concentric phase and by averaging data over the concentric phase, respectively.

Static, ROM and RVJ trials were reconstructed and labelled, and gaps filled in VICON Nexus (v2.10.2, VICON, Oxford, UK) with C3D files exported for processing in Visual3D (v6.0, c-motion, MD, USA). Functional joint centres (hip and knee) were calculated using the ROM trials in Visual3D (C-Motion, 2020). A fourth order zero-lag low-pass Butterworth filter with 15 Hz cut-off frequency was used to smooth marker trajectory data. GRF (Fx, Fy, Fz), segment angular kinematics (thorax and pelvis) and joint angular kinematics (ankle, knee and hip) and joint kinetics (moment and power for the ankle, knee and hip) for the three axes of rotation (X ¼ medial-lateral; Y ¼ anterior-posterior; Z ¼ longitudinal) were calculated. The thorax segment was defined as the axial rotation component of thorax motion relative to pelvis motion, otherwise known as 'X-Factor' in golf (Hume et al., 2005) . Two key events were determined; initial contact (INC), defined as vertical GRF >10 N, and initial impact (INI), defined as 100 ms after INC (Norcross et al., 2013) . Data was collected at these two time points for statistical analysis. Peak GRF metric for all three axes was exported for each trial. The data for the GRF, segment and joint angular kinematics and joint kinetics across the five trials were averaged for each session.

The COP trajectory was resolved into a virtual foot local coordinate system from contact with the force platform (Visual3D, C-Motion Inc., USA). Specifically, the forward progression COP was normalised (arbitrary unit, a.u) by the distance along the anteriorposterior (A-P) axis from the proximal end of the foot segment (ankle joint) to the 2 nd metatarsal head (distal joint centre) (O'Connell et al., 1998) . This meant that A-P COP range of motion was quantified on the order of -1 to 2, where a negative value indicates that COP is behind the ankle joint centre and a value > 1 reflects COP ahead of the metatarsals (O'Connell et al., 1998) . Similarly, the medio-lateral (M-L) COP was normalised (a.u) by its distance along the distal radius of the foot segment (1 st to 5 th metatarsal head) with respect to the longitudinal axis of the foot segment. An M-L COP equal to zero reflects a position located on the A-P axis, whereas a positive value indicates a laterally-directed trajectory (Visual3D, C-Motion Inc., USA). The data were expressed relative to subdivisions of stance phase, representing early (0-33%; COP33), mid-(34-66%; COP66) and late stance (67-100%; COP100) regions (Chang et al., 2008) .