Updating Local and Global Probability Events During Maze Navigation
MetadataShow full item record
Our mental models consist of relational knowledge. We apply this knowledge about whether something is near to or far from something else to solve tasks. As a specific exam- ple, when we navigate in our environment, we have global (far) location goals that we could navigate to using local (near) landmarks. The question for the present study is whether relational knowledge can be probabilistically and differently represented at global and local levels. To test this, we had participants navigate a maze in which the wall structure was hidden, but in which participants were given global and local cues. We manipulated the reliability of the global and local cues across experimental trials and experiments. Our results demonstrated separable effects for global and local cues. Participants made good estimates of global and local cues’ reliability, however, their estimates of global cue re- liability were less accurate than their estimates of local ones potentially due to inherent differences in how global and local information is represented. Their use of local cues roughly matched the ground truth local cues reliability whereas their use of global cues did not match the ground truth global cue reliability. In addition, participants relied on both local and global cues when they navigated in the mazes but with local cues dominant possibly because of their confidence in local cue reliability estimates, preference for cues associated with more immediate reward, and feedback proximity. Altogether, this study characterizes the mental representations of uncertain global and local cues and suggests that people negotiate between different probabilistic information when making decisions in maze navigation.
Cite this version of the work
Sixuan Chen (2022). Updating Local and Global Probability Events During Maze Navigation. UWSpace. http://hdl.handle.net/10012/18658