Flow Tree: A dynamic model for navigation paths and strategies
Keywords: dynamics, navigation, behavior, trees, mazes, cognitive representation, neuroscience
TL;DR: The Flow Tree is a novel graph-based way to capture the temporal and spatial dynamics of human navigation through mazes.
Abstract: Navigation is a dynamic process that involves learning how to represent the environment, along with positions in and trajectories through it. Spatial navigation skills vary significantly among individual humans. But what exactly differentiates a good navigator from a bad one, or an easy-to-navigate path from a hard one, is not well understood. Several studies have analysed exploration and navigation behaviour using static quantitative measures, like counts of positions visited or distance travelled. These static measures, however, are inherently limited in their ability to describe dynamic behaviors, providing a coarse quantification of the navigation process. To fill this gap, we introduce the \emph{Flow Tree}, a novel data structure, which quantifies the dynamics of a group of trajectories through time. This is a discrete adaptation of the Reeb graph, a mathematical structure from topology, computed from multiple trajectories (from different people or the same person over time). Each divergence in trajectory is captured as a node, encoding the variability of the collection of trajectories. A Flow Tree encodes how difficult it will be to navigate a certain path for a group of humans. We apply the Flow Tree to a behavioural dataset of 100 humans exploring and then navigating a small, closed-form maze in virtual reality. In this paper we (1) describe what a Flow Tree is and how to calculate it, (2) show that Flow Trees can be used to predict path difficulty more effectively than static metrics, and (3) demonstrate that a trajectory through the Flow Tree is predictive of that individual's success. We (4) introduce a hypothesis testing framework over Flow Trees to quantitatively differentiate between the strategies of the best navigators from those of worst. Thus, we show that Flow Trees are a powerful tool to analyse dynamic trajectory data.\footnote{The code will be made publicly available at [anon-github-link].}
Primary Area: applications to neuroscience & cognitive science
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 9575
Loading