Obstacle avoidance during human walking: learning rate and cross-modal transfer

1. The aim of this study was to investigate the significance of specific afferent information during motor learning. Blindfolded subjects stepped over an obstacle on a treadmill while different stimuli (acoustic (ACU), somatosensory (SOM) and light flash (LED)) signalled the approaching obstacle. The effect of the above stimuli was then evaluated and compared to full vision (VIS) locomotion. In the non-visual conditions feedback information about the performance was provided by an acoustic signal. 2. Using each of the different stimuli for information the level of subject performance was assessed by noting foot clearance and analysing both leg muscle electromyographic activity and movement trajectories during three successive runs. Each of these runs consisted of 100 steps over the obstacle. 3. The best performance at the onset of the first run was achieved during the VIS condition. When the VIS condition (run 1 + 2) was followed by ACU or SOM information or when the ACU condition (run 1 + 2) was followed by LED, little cross-modal transfer (CMT) occurred, i.e. adaptation in run 3 started again at a low level of performance. In contrast, if adaptation started with ACU stimuli followed by SOM stimuli, almost full CMT occurred. The absolute level of performance achieved after the second or third runs was similar in the VIS and non-VIS conditions. 4. In conclusion, the course of motor learning depends on specific afferent information, and feedforward control has a special influence on the performance only at the onset of the experiment but not on the rate of learning. The fact that little CMT occurs from visual to nonvisual stimuli and from ACU to LED suggests that visual afferent input is processed in a different way to non-visual stimuli.