The kinematic data were collected using the VICON system (Oxford Metrics Inc., Oxford, UK) with five CCD (charge-coupled device) cameras. The cameras were calibrated in a volume extending 2.0 m in the vertical (Z), 1.2 m in the anterior (Y) and 0.6 m in the lateral (X) direction. Camera system calibration and three-dimensional target reconstruction were done by AMASS (Adtech, Adelphi, MD, USA) software. Average residual errors did not exceed 3 mm for each camera. A camera non-linearity process was performed before the camera calibration and was applied to all kinematic data. The co-ordinate system of the laboratory had a positive sense in the directions: Z= up, Y= forward, X= right. Retroreflective spherical markers covered with 3 M high gain 7610 retroreflective tape of 25.4 mm diameter were affixed bilaterally to the fifth metatarsal head, lateral malleolus, lateral femoral epicondyle, greater trochanter, acromion, cheekbone in front of the tragus, forehead, elbow, and the styloid process of the radius. Root-mean-square (RMS) noise level of the system measured at the right ankle marker did not exceed 0.2 mm for X and Z and 0.3 mm for Y components.
Kinetic data were collected by a force platform AMTI type OR6-3 (Advanced Mechanical Technology, Inc., Newton, MA, USA), that measures the X, Y and Z forces (Fx, Fy and Fz) as well as the X, Y and Z moments (Mx, My and Mz). The centre of pressure (COP) (at the force transducers) was determined based on the cross product:
From this, the X (CPx) and Y (CPy) co-ordinates of the COP at the surface of the plate were determined:
where CPz is the distance from transducers to the surface of the plate.
Bipolar electrodes with preamplifiers (frequency band, 10-40 kHz; gain, × 300) were fixed to the right anterior tibial, lateral gastrocnemius, femoral quadriceps (vastus lateralis) and femoral biceps (lateral head) muscles. Two simultaneous recordings were made from each muscle with leading-off sites separated by about 3 cm along and 2 cm across the muscle fibres. The EMG signal was preamplified, then sequentially high-pass filtered, rectified and low-pass filtered with a 50 Hz Bessel filter and amplified. Video and force platform data were synchronously sampled (sampling rate, 50 Hz) with the EMG data (sampling rate, 200 Hz) by a PDP-11/73 computer (10-bit ADC) and transferred to a VAX 11/750 computer. The collected kinematic and kinetic data underwent a smoothing procedure with a second order Butterworth filter (6 Hz) with the data filtered in the forward and reverse direction to avoid phase lag (Winter, 1990).
The task of the subject was to step on the force platform after a command and to maintain balance holding a very light wooden bar (weight, 50 g; length, 0.9 m) in his hands with the arms freely hanging along the body and forearms in supination. This arm position was chosen to avoid both masking the markers and having the arms enter into the dynamics. Subjects were instructed to stand quietly.
Four experimental conditions were studied: (1) EO, standing with eyes open and natural support area (approximately 1-2 in between the heels and 10 in between the toes); (2) EC, standing with eyes closed and natural support area; (3) EOR, standing with eyes open and with feet close together (Romberg stance); and (4) ECR, standing with eyes closed and with feet close together. The duration of each whole trial was 50 s. The initial epoch of the first 13 s was discarded from the analysis. There were five trials for each condition; sufficient time for rest was allowed between the trials and between the conditions.