Spatial Categorization and Computation - Empirical Evidence from Artificial Label

It is well known that spatial perception is a basic ability in our daily life, while we compute spatial relationship between two objects universally. This study examined how people perceive spatial categories using three tasks, learning task, producing task, and rating task. Three different kinds of spatial configurations were manipulated. 27 subjects were assigned randomly to each kind of spatial configuration. The results showed that response times (RTs) for spatial category judgment were, just as we have expected, slower near the boundary and faster near prototype, in learning task. RT pattern, in rating task, also suggested that people formed the expected configurations. People´s rating patterns were well captured the spatial configurations predefined. Performances in all three tasks suggested that people can form artificial spatial categories and make spatial computation with enough practices. Implication for Bayesian analysis and application were briefly discussed.