High contextual interference in perturbation-based balance training leads to persistent and generalizable stability gains of compensatory limb movements.

Reactive responses to balance perturbations have been shown to be improved by training. This investigation aimed to compare the effects of block and random training perturbation schedules on stability gains of compensatory arm and leg movements in response to unpredictable large-magnitude balance perturbations. Perturbations were produced by means of sudden displacements of the support base, associating mode (rotation, translation, combined), direction, and velocity of platform motion. Healthy young participants were assigned to one of three groups: random, block, and control. For the random group, perturbation sequence was unpredictable. For the block group, each balance perturbation was repeated over blocks of four trials. Controls were tested only, serving as reference of first trial responses in the post-test. Evaluation was made through a scale rating stability of compensatory arm and leg movements (CALM). We probed immediate and persistence gains (1-week retention), in addition to generalizability to perturbations of higher velocity and to dual-tasking (mental subtraction). In the post-test both the block and random groups achieved higher leg and global scores in comparison with controls in the most challenging perturbations. In retention and transfer tests, results for the global score indicated higher values for the random than for the block and control groups. These results support the conclusion that high but not low contextual interference in perturbation-based balance training leads to enduring and generalizable increased stability gains of compensatory limb movements in response to unpredictable balance perturbations.