The structure of spatial knowledge: Do humans learn the geometry, topology, or stable properties of the environment?

Ericson, Jonathan D.

Full Metadata

Overview

Year:: 2014
Contributor:: Ericson, Jonathan D (creator); Warren, William (Director); Domini, Fulvio (Reader); Burwell, Rebecca (Reader); Brown University. Cognitive and Linguistic Sciences: Cognitive Sciences (sponsor)
Genre:: theses
Subject:: Spatial Knowledge; Cognitive Maps; Spatial Cognition; Spatial Memory; Graphs; Metric; Euclidean Geometry; Noneuclidean Geometry; Cognitive maps (Psychology); Topology; Graphic methods; Neighborhoods; Geometry, Non-Euclidean
Extent:: 11, 174 p.
DOI: https://doi.org/10.7301/Z0WD3XXP

Files

Description

Abstract:: This dissertation critically examined the structure of human spatial knowledge by testing three general hypotheses: the Euclidean, Topological, and Stability Hypotheses. The present experiments selectively destabilized three geometric properties of the environment during learning, specifically Euclidean metric structure, neighborhoods, or the place graph. In the learning phase, separate groups of participants walked in one of four virtual hedge mazes: (1) The Control Maze preserved all three properties. (2) Elastic Maze I preserved the place graph while destabilizing Euclidean and neighborhood structure, by stretching paths during learning so that a target could occupy two different metric locations in different neighborhoods. This maze provided metric (via path integration) but not visual information about neighborhood boundary relations. (3) Elastic Maze II added visual information for neighborhood boundaries, by having stretched paths visibly intersect paths they crossed during learning. (4) The Swap Maze preserved neighborhoods while destabilizing Euclidean and graph structure by swapping pairs of targets within a neighborhood during learning. In the test phase, spatial learning was probed by asking participants to perform one of three navigational tasks: (a) metric shortcut task, designed to assess Euclidean spatial knowledge, (b) neighborhood shortcut task, to assess knowledge of neighborhoods, (c) route task, to assess graph knowledge. This design aimed to test several specific predictions: (H1) if spatial knowledge is primarily Euclidean, then performance on all tasks should deteriorate when metric structure is destabilized during learning; (H2) if spatial knowledge is primarily topological, performance should reflect the place graph when metric and neighborhood structure are destabilized, and (H3) if navigators acquire whatever geometric properties remain stable during learning, performance should reflect the stable structure in each maze. The results are inconsistent with the Euclidean and Stability hypotheses, but support the Topological hypothesis. Metric performance is highly unreliable and neighborhood knowledge can suffer even when neighborhoods are stable. In contrast, the place graph is successfully learned in all environments, with over 96% accuracy. Spatial knowledge appears to be primarily topological, and is consistent with a labeled graph that incorporates approximate local distance and angle information.
Notes:: Thesis (Ph.D. -- Brown University (2014)

Content

Access Conditions

Rights: In Copyright
Restrictions on Use: Collection is open for research.

Citation

Ericson, Jonathan D., "The structure of spatial knowledge: Do humans learn the geometry, topology, or stable properties of the environment?" (2014). Cognitive Sciences Theses and Dissertations. Brown Digital Repository. Brown University Library. https://doi.org/10.7301/Z0WD3XXP

Relations

Collection:

Cognitive Sciences Theses and Dissertations

Theses and Dissertations for the Cognitive Sciences department.
...