Purchase this article with an account.
Alexander Muryy, Andrew Glennerster; Models of navigation and pointing in non-metric environments. Journal of Vision 2018;18(10):1040. doi: https://doi.org/10.1167/18.10.1040.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Different models of spatial representation have been suggested to explain the behaviour of people as they navigate and remember the location of objects: full metric reconstruction, topological representation and labelled graphs (Warren, 2016). Some tasks require a topological representation, while others require a metric representation and different representations might be used for different tasks. In our experiment, human participants (n=8) were asked to perform two spatial tasks in a virtual labyrinth: navigational and pointing. The navigational task was to collect 4 coloured targets in a specified order and then, from the last location, to point to all other targets, which were not currently visible. There were 5 repeats with the same target order ('learning' phase) and three repeats with a different target order ('test phase'). We tested participants both in physically-realisable configurations ('metric') and impossible, non-metric configurations with 'wormholes' that increased the length and number of turns between pairs of locations in the maze but did not alter the topological structure. The participants navigated efficiently even in the most difficult non-metric condition. We predicted participants' routes in the test phase using data on successful routes from the learning phase. For long paths in the wormhole conditions, this model predicted participants' routes significantly better than assuming they followed the shortest metric route or the shortest topological route. This is what one would expect if they learn the topological structure of the maze gradually. For the pointing results, we computed the most likely metric configuration of the targets that would be consistent with the participant's pointing directions. In the wormhole conditions (but not in the metric condition), this provides a better explanation of pointing responses than ground truth (using Akaike information criterion) and suggests ways in which participants might create a distorted metric representation of the maze containing wormholes.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only