A Dynamic Bayesian Actor Model explains Endpoint Variability in Homing Tasks

Abstract

Goal-directed navigation requires integrating information from a variety of internal and external spatial cues, representing them internally, planning, and executing motor actions sequentially. However, a comprehensive computational account of how these processes interact in an ambiguous, uncertain, and noisy environment giving rise to biases and variability observed in navigation behavior is currently unavailable. In this paper, we introduce an optimal control under uncertainty model, which provides a computational-level explanation of how landmarks and path integration interact and are combined to reduce variability in navigation. We apply our model to trajectory and end-point data from three previously published studies that employ a variant of the triangle completion task with landmarks. Contrary to observer models, which attribute human endpoint variability to perceptual cue combination processes only, this dynamic Bayesian actor model provides a unifying account of a wide range of phenomena found in this task by considering variability in perception, action, and internal representations jointly. Taken together, these findings have wide-ranging implications for the analysis and interpretation of human navigation behavior, including resolution of seemingly contradictory results on cue integration in navigation.

Spatial uncertainty and environmental geometry in navigation

“…1c). Indeed, recent work using a more formal model of the planning process 20 has been able to account for apparent violations of the Bayes-optimal combination of landmark and self motion cues 27,28 and reconcile them with earlier results suggesting optimal cue combination 26 . These results lend further support to our core suggestion: that Bayesian principles play a fundamental role in navigation.…”

“…Conversely, models that formalized navigation in a Bayesian inference framework have not incorporated an image-computable ideal observer. Thus, these models did not study how environmental geometry affects spatial uncertainty and, in turn, homing behavior or grid cell responses 16,20,21,25,[51][52][53][54][55][56] .…”

“…Here we show instead that all these data can be accounted for by a single, unifying principle: that allocentric location is fundamentally uncertain, and this uncertainty thus pervades both behavioral and neural responses. While the importance of representing uncertainty about spatial location is well accepted in engineering solutions to navigation 18,19 (even mobile phone navigation apps show a simple form of it by the shaded blue disc around the currently estimated location), its role in biological navigation has only rarely been considered [20][21][22][23][24][25] . In particular, the fundamental effects of environmental geometry on spatial uncertainty have not been studied.…”

“…Formalizing the effects of environmental geometry on navigation is challenging because it can affect visual input, a main source of information for navigation, in complex ways that may not be well captured by simple, discrete point-like landmarks as used in some virtual-reality experiments [26][27][28] and assumed to provide direct distance and bearing information in related modeling studies 20,26,29 . Thus, we constructed the Location posterior when all inputs are available, P t |v0:t , w0:t , u θ 0:t , u 0:t .…”

Finding your Way Out: Planning Strategies in Human Maze-Solving Behavior

“…Therefore, for analyzing decision-making and its underlying planning mechanisms, mazes have been of particular popularity in studying humans, animals and machines in various fields including Cognitive Science [1][2][3], Neuroscience [4,5], and Robotics [6,7]. Real world naturalistic navigation tasks usually require the integration of internal and external cues, the execution of motor actions, and internal planning [8]. However, one major advantage of mazes as experimental environments is that they can be clearly defined and generated in terms of their topology [9].…”