developed children were used as reference. Moments and powers were normalized to the body mass. Results: At the baseline, kinematic data of six right gait cycles showed a reduction of hip extension and knee recurvatum, in late stance and at the ground contact an increase of ankle plantar flexion. Kinetic data of three right gait cycles showed the absence of moments and powers. Spatio-temporal parameters showed a reduction of walking speed (0.28 Æ 0.052 m/s). At the follow up, 15 days after, we observed an increase in hip extension, a reduction of knee recurvatum and ankle plantar flexion, and finally appeared moments and powers on all lower limb joints. Walking speed increased to 0.67 Æ 0.15 m/s. Discussion: The use of the described dynamic component improved gait pattern during a training of 15 days. This evidence is compatible with the hypothesis that gait pattern is constrained by the interaction between the external force field and the available resources of the organism. This interaction allowed the synergy assembling as a cooperativity of roles specific for the function [2]. Although the ribbon spring acted only on knee joint, the dynamic effects are evident on the whole lover limb pattern, involving kinematic and kinetic changes both on hip, knee and ankle joints; that is, it modified the internal structure of the motor synergy. References [1] Turvey MT. Hum Mov Sci 2007;26:657–97. [2] Turvey MT Fonseca S. New York: Springer; 2008. doi: 10.1016/j.gaitpost.2008.10.046 Arm and leg swing during overground and treadmill walking I. Carpinella 1, *, P. Mazzoleni 1 , P. Crenna 2 , M. Rabuffetti 1 , M. Ferrarin 1 1 Biomedical Technology Department, Fond. Don Carlo Gnocchi Onlus, Milan, Italy 2 Institute of Human Physiology I, Faculty of Medicine, University of Milan, Milan, Italy Introduction: Treadmill walking is frequently used in locomotion research to simulate overground ambulation. How- ever literature has shown significant differences between the two walking conditions in spatio-temporal-parameters and lower limb kinematics [1], kinetics and EMG pattern [2]. Treadmill has been also widely used to study arm swing movement and upper/lower limb coordination during gait. Results have shown that different arm swing patterns, strictly related to the walking speed, are adopted. In particular, for velocities lower than 0.8–1.0 m/s, rhythmic oscillations of the arms have a frequency of two cycle per stride while, above this range, arm swing frequency equals the leg frequency [3]. Moreover, increasing velocity generates a progressive decre- ment of the phase shift between arm and ipsilateral leg movement, toward a perfect counterphase coupling. Goal of the present work was to compare arm and leg swing movement during overground and treadmill walking at different velocities, in a group of healthy subjects. Moreover, we aimed to investigate the possible role of the inertial properties of upper limbs in modulating arm swing behaviour. Methods: Ten healthy subjects were tested by means of optoelectronic kinematic analysis during both overground (OW) and treadmill (TW) walking at three velocity levels (slow, natural and fast). Treadmill velocities were set according to the mean velocities maintained by the subjects during overground trials. Spatio-temporal gait parameters, arm and thigh ROM in the sagittal plane and upper/lower limb phase shift were computed and analyzed. Results: For each velocity level, cadence and swing phase were significantly higher in TW with respect to OW, while stride length was lower, despite the same velocity. Arm and thigh ROM were not significantly different between OW and TW. On the contrary, phase shift between arm and ipsilateral leg swing movements was consistently lower in TW with respect to OW. Ten of the twenty arms analyzed showed, in OW at low speed, a frequency ratio of arm to thigh swing equal to 2:1. Only three of these arms maintained this pattern also in TW, while the other seven arms adopted a coordination pattern based upon 1:1 frequency ratio. Abstracts of the SIAMOC 2008 Congress / Gait & Posture 29S (2009) e1–e31 e29