Assessing Systematic Distortions in Visuospatial Mental Representations with Use of Non-Linear Dimensionality Reduction: An Explorative Study
The following research focuses on an explorative analysis of cognitive maps that are externalized via sketch maps. The information that is encoded in sketch maps possibly includes spatial information that is stored in memory. Participants were asked to walk through a fictional virtual environment after which they were required to draw a map of the route. To analyse these maps and possible distortions in spatial mental representations that are shared among participants, we explore a geometrical analysis of sketch maps. Assuming that, in pixel space, the sketch map data have intrinsic dimensionality, we will use machine learning algorithms to reduce the dimensionality non-linearly so that we are able to inspect and interpret the possible manifold underlying these sketch map data. By doing so we hope we will be able to retrieve more knowledge about the processing and storage of spatial knowledge and to possibly provide a tool for assessing in what way spatial information is stored in memory.
Lise Stork, "Assessing Systematic Distortions in Visuospatial Mental Representations with Use of Non-Linear Dimensionality Reduction: An Explorative Study", Master's Thesis for the Media Technology programme, Leiden University (The Netherlands), 2016