Embodied decision biases during walking

For many years, research of psychology has viewed cognition as a sequential and modular process, with action being only as the output of a higher-level cognitive process. This is a problem when studying decision-making in everyday behavior, where lower-level motor control (such as walking) often needs to occur at the same time as higher-level decision processes (such as avoiding an obstacle to the left or right). For these situations, research in multitasking, embodiment, and specifically embodied decision-making suggests that information-processing at the level of motor control can influence decision-making through crosstalk between the processes and parallel feedback of action costs. This challenges the traditional view of decision-making as a modular and sequential process. Thus, we set out to investigate the assumed influences of lower-level actions on higher-level decisions when action and decision-making must run concurrently. To do that, we first implemented a novel experimental paradigm for studying decision-making during the whole-body movement of walking. Participants were asked to walk toward an obstacle and to concurrently decide to turn toward a target on the left or right for reward. We manipulated the action of walking toward the obstacle by predetermining the swing leg before turning in front of the obstacle. Results revealed that the body dynamics of concurrent action influenced decision-making. More specifically, participants preferred turning toward the side of the swing leg, even at the expense of receiving less reward. After validating the experimental paradigm, we investigated the type of embodied decision bias present in the walking task. Thereby, we focused on the bias by action costs during action. If the decision process receives parallel feedback during movement, the cost dynamics during movement should influence decision-making. In four experiments, we manipulated the action costs of turning under various conditions.