Spatial contiguity determines overshadowing between global-shape representations and stimulus-response associations in human navigation

Recent failures to observe cue competition in different preparations, together with the long-standing debate about competition phenomena in the spatial learning literature, have casted doubts on their generality. A recent study (Herrera et al., in press) suggested that strong goal-landmark contiguity might be a critical factor for competition to occur in human spatial navigation. This study aimed to extend those findings and evaluate the effect of goal-landmark contiguity together with variations in the starting-point locations, and the length of training trials. Participants were requested to find an invisible goal (a Wi-Fi hotspot) in a virtual T-Maze. Each group was trained with a landmark placed at different distances from the goal (Proximal, Middle, Distal and Control - no landmark). During test, participants were instructed to look the Wi-Fi signal outside of the arena, to assess if the landmark impaired or not its global representation. In Experiment 1, participants were released from a fixed location over 15 training trials, in Experiment 2 participants were released from multiple locations, and in Experiment 3 participants were released from a fixed arm but received 3 trials only. Overshadowing was found under strong contiguity (Proximal Group), but only when participants were consistently released from the same location and received extended training (Experiment 1), and no other cue competition effects were found across experiments. These results add further support to the increasing literature that shows strong contiguity is a key determinant of overshadowing in human spatial navigation, and we characterise the conditions under which this happens.