Humans continuously modulate their control strategies during walking based on their ability to anticipate disturbances. However, how people adapt and use motor plans to create stable walking in unpredictable environments is not well understood. Our purpose was to investigate how people adapt motor plans when walking in a novel and unpredictable environment. We evaluated the whole-body center of mass (COM) trajectory of participants as they performed repetitions of a discrete goal-directed walking task during which a laterally-directed force field was applied to the COM. The force field was proportional in magnitude to forward walking velocity and randomly directed towards either the right or left each trial. We hypothesized that people would adapt a control strategy to reduce the COM lateral deviations created by the unpredictable force field. In support of our hypothesis, we found that with practice the magnitude of COM lateral deviation was reduced by 28% (left) and 44% (right). Participants adapted two distinct unilateral strategies that collectively created a bilateral resistance to the unpredictable force field. These strategies included an anticipatory postural adjustment to resist against forces applied to the left, and a more lateral first step to resist against forces applied to the right. In addition, during catch trials when the force field was unexpectedly removed, participants exhibited trajectories similar to baseline trials. These findings were consistent with an impedance control strategy that provides a robust resistance to unpredictable perturbations. However, we also found evidence that participants made predictive adaptations in response to their immediate experience that persisted for three trials. Due to the unpredictable nature of the force field, this predictive strategy would sometimes result in greater lateral deviations when the prediction was incorrect. The presence of these competing control strategies may have long term benefits by allowing the nervous system to identify the best overall control strategy to use in a novel environment.
