This study incorporated a dual-task paradigm, in which participants were asked to perform basic locomotion tasks with one of three interfaces while remembering a sequence of either spatial or verbal items. Interfaces varied in similarity to natural body movements. Stopping performance was compromised when concurrently remembering a spatial, but not verbal, sequence. Also users exhibited lower performance on spatial memory tasks while using more unnatural locomotion interfaces. These results confirm that semi-natural locomotion interfaces require spatial working memory resources and thus locomotion interfaces compete with ongoing spatial tasks, as opposed to those requiring verbal resources or general attention resources.