アイテム詳細

要旨

In contrast to spatial memory acquired in enclosed spaces (i.e., rooms), memory for navigable spaces (e.g., buildings) is less well understood. We had subjects learn eight target objects spread across a virtual environment, consisting of two apparent regions. Separation of regions was triggered by visual and semantic similarity (wall colour, object categories), walking distance and turning angle complexity (longest and most complex path at transition point), and spatio-temporal contingency (regions learned subsequently). A subsequent pointing task revealed increased pointing
latency with increasing corridor distance to the target, and a facilitative effect of being bodily aligned with the local corridor. Moreover, pointing to targets located within one’s current region was faster compared to pointing to targets in the other region, and alignment with region specific
orientations enhanced pointing latency. In sum, our results indicate the formation of local and regional memory units that manifest in the form of spatial reference frames. Our local and regional effects give way to the interpretation that, when confronted with a complex navigable space, multiple levels of spatial integration are stored, potentially in a hierarchical fashion. Importantly, our results cannot be explained by concepts of exclusive local corridor units or exclusive global, all-embracing memory units.