Influence of Threat and Anxiety on Cognitive Map Learning

Adaptive behavior often requires using structured internal representations about the external world, such as bringing to mind relevant side-streets when encountering a detour on your way home from work. These internal representations are collectively referred to as cognitive maps. Given their fundamental importance to representing spatial, temporal, and conceptual relationships, surprisingly little is known about factors that modulate map learning. Here, we tested how cognitive map learning is influenced by threat and anxiety. We developed a novel drag-and-drop network leaning task involving images of either threat or neutral valence. Our aim was to uncover how subject-level trait anxiety and image valence impacted three key metrics of map learning: number of attempts until correct placement during the drag-and-drop map learning task; detection of ‘invalid’ network transitions during a post-learning random walk sequence; and accuracy during a post-learning odd-one-out image community classification task. We found that there was no difference in whether threat or neutral images were learned first in the drag-and-drop task, and no significant difference between anxiety levels or image valence for odd-one-out classification accuracy. However, participants with high trait anxiety were better at detecting invalid network transitions for both image valences, and detection of invalid transition was overall worse when they involved threat images. These novel findings indicate that emotionally salient locations can impact the manner in which people build mental representations of their environment, and that high trait anxiety promotes cognitive map learning.