Investigating the influence of environmental predictability on route planning using a novel foraging paradigm

Abstract

During self-guided behaviors animals rapidly identify constraints of the problems they face and adaptively employ appropriate strategies. In the case of foraging, animals must balance sensory-guided exploration of an environment with memory-guided exploitation of known resource locations. Here we show that animals adaptively shift cognitive resources between sensory and memory systems during foraging to optimize route planning under uncertainty. We demonstrate this using a new, laboratory-based discovery method to define the strategies used to solve a difficult route optimization scenario, the probabilistic “traveling salesman” problem. Using this system, we precisely manipulated the strength of prior information as well as the complexity of the problem. We find that rats are capable of efficiently solving this route-planning problem, even under conditions with unreliable prior information and a large space of possible solutions. Through analysis of animals’ trajectories, we show that they shift the balance between exploiting known locations and searching for new locations of food based upon the predictability of food locations. When compared to a Bayesian search, we found that animal performance is consistent with an approach that adaptively allocates cognitive resources between sensory processing and memory, enhancing sensory acuity and reducing memory load under conditions in which prior information in unreliable. We also find that the complexity of this route planning problem can be finely titrated through manipulating the number of food locations, with greater numbers of locations leading to wider angles of approaches to pellets and increased head movements while foraging. Additionally, the influence of environmental predictability on hippocampal local field potentials is examined. We show that animals trained on unpredictable pellet distributions increase their theta power over the course of training, while animals trained on predictable pellet distributions exhibit location-specific decreases in theta power when they reach the first few pellets in an acquisition sequence. Our findings establish new approaches to understand neural substrates of natural behavior as well as the rational development of biologically inspired approaches for complex real-world optimization.