BackgroundWhen moving in public space, individuals are challenged with having to master multiple cognitive and motor demands, either simultaneously or in short succession. Empirical evidence suggests that cognitive-motor multi-tasking during walking may impact one or both, cognitive and motor performance. These performance changes may result from unintentional task-interference effects, but also from strategic behavioral changes to cope with the multiple task demands. Strategic changes in human walking behavior have been uncovered in experimental scenarios, in which individuals avoid colliding with other individuals or objects in the environment. However, whether collision avoidance behavior is sensitive to cognitive-motor multi-task demands has remained underexplored, yet. Thus, with this study, we aimed at systematically studying cognitive-motor multi-task effects on collision avoidance during human locomotion. MethodsTen healthy participants (23.9 ± 4.3 years, 4 female) were walking at their preferred speed from a predefined start to end position under four experimental conditions: walking only (BL), walking while having to avoid a collision with another person (IO), writing a text message on a mobile phone while walking (cognitive-motor dual-task, DT), and writing while walking with collision avoidance demand (multi-task, MT). Parameters quantifying locomotor as well as collision avoidance behavior (path length, walking speed, minimum distance, path and speed adjustment) were assessed using optical motion tracking. In addition, performance in the writing task (errors, writing speed) was examined. ResultsParticipants' locomotor behavior was significantly affected by experimental conditions, with additive effects of multi-task demands on both path length (BL = DT < IO < MT) and walking speed (BL > IO > DT > MT). Further, participants showed an increased error rate and writing speed in the writing task when walking as compared to when standing still, independent of the presence of an interferer. Importantly, collision avoidance behavior was selectively influenced by cognitive-motor multi-task demands, with an increased minimum distance to the other person in the MT-condition, but no differences in path or speed adjustment. DiscussionOur results suggest significant multi-tasking effects of writing a message on the mobile phone while walking on both locomotor behavior and writing task performance. Collision avoidance behavior seems to be selectively affected by multi-task demands, reflected in an increased minimum passing distance, without overt changes in path or speed adjustments. This may be indicative for a strategic change in collision avoidance behavior towards a more cautious strategy to account for altered attention allocation and less visual feedback when writing while walking.