Factors influencing the quick onset of stepping following postural perturbation.

It has been shown that the stepping to recover balance following a forward fall occurs at a constant time (on average 293 ms) (Do et al. Journal of Biomechanics 15, 1982, 933-939). In this study, we tested the hypothesis according to which programming to make fast movement could trigger the movement earlier than when programming self-pace movement. The same experimental paradigm of forward fall was used (see Do et al., 1982) to induce stepping. Different extents of stepping were manipulated by instructions: Subjects were instructed to step to recover their balance naturally (control condition); to make shorter steps than in the control condition; longer steps; faster steps. Lastly, a fast step was also induced by the biomechanical constraint on the initial posture, i.e. by inclining the subject forward at his maximum capacity. Data were collected from 12 subjects. The variables analyzed were the onset latency of step execution and other classical parameters (time of heel-contact, duration of the swing phase, step length, center of mass progression velocity, and step velocity). The results showed that the onset of stepping was unchanged in the longer- and faster-step conditions, relative to the control condition (mean control value = 280 ms). In contrast, the onset of stepping was significantly earlier in the short-step condition, and when the initial inclination was greater (250 and 252 ms, respectively). The swing phase duration in these two conditions averaged 140 and 185 ms, was significantly shorter than in the other conditions, whereas step length was obviously expected to be shorter in the shorter-step condition and longer in the longer-step condition than in the other conditions. Step length was similar between the other conditions. We conclude that neither step length or step velocity programming could induce an earlier onset latency of stepping. Step programming in relation to these specific instructions seemed to concern the extent of step execution and not the time of triggering of the stepping. We suggest that the control of short swing phase duration resulted in an earlier onset latency of stepping to recover the balance. This control depends on the combination of biomechanical constraints and cognitive processes, including subject's interpretation of the instructions and evaluation of the risk of fall.