Influence of bilateral GM activity on stability and swing foot position as a function of perturbation onset timing. Linear regression between (a) stride width, (b) margin of stability, (c) foot height and (d) inter-feet distance during mid swing and GM activity of the inward (blue line) and outward leg (green line). The intercept of each regression shows the effect of the perturbation on the outcome variable without an active muscle response (i.e. passive response simulation).The slope of the regression shows the potential of the reactive muscle activity (i.e. active response) to adjust the four outcome variables. The stride width and margin of stability was evaluated at first heelstrike of the outward leg for the perturbations with onset at 7.5% and 22.5% of the gait cycle, and at first heelstrike of the inward leg for the perturbations with onset at 37.5% and 52.5% (i.e. since the outward leg is already on the ground). The red areas indicate when there is a potential loss of stability: (b,f) negative margin of stability (c,g) ground-foot collision during midswing (d,h) inter-feet collision during mid swing. The regression was only computed when a significant GM response of the inward or outward leg was observed (Fig. 3).