Accuracy of the Formation of Spatial Representations of Dynamic Scenes in Working Memory

Abstract

<p>The cognitive representation of the environment is formed using cognitive systems that process data on spatial representations of two types: egocentric, encoding the position of environmental objects relative to the observer, and allocentric, encoding the position of objects relative to each other, regardless of the position of the observer. Data on spatial representations were studied mainly in problems of memorization and reconstruction of static scenes. However, the task of processing information about dynamic scenes in everyday life has a higher ecological validity. We used HMD virtual reality technologies to study the accuracy of the formation of egocentric and allocentric spatial representations of static and dynamic scenes in working memory. The subjects were presented 8 three-dimensional virtual scenes of 4 objects each for 10 seconds in static and dynamic conditions for memorization and reconstruction. Identification accuracy (number of correctly reconstructed objects) and localization accuracy (accuracy of spatial scene reconstruction) were assessed. Localization accuracy was assessed in topological units, corresponding to the accuracy of the representation of the general configuration of objects in the scene (global topological information), and in metric units, corresponding to the accuracy of the representation of the spatial coordinates of each object (local metric information). The results showed that object identification accuracy was similar in static and dynamic conditions; the processes of encoding metric local information during the formation of both types of representations of dynamic scenes worsen compared to static ones; the accuracy of encoding topological global information remains stable compared to the static condition. We can conclude that the visual and spatial systems operate independently as part of a general cognitive system that processes data on spatial representations in time-limited working memory, as well as the redistribution of its resource in dynamic condition for supporting topological data of the holistic configuration of moving objects more, than metric data. The results highlight the importance of topological spatial characteristics of spatial representations for processes of early spatial perception, decision making, and action in the environment.</p>