Neural encoding of perceived patch value during competitive and hazardous virtual foraging

Brian Silston,Xin Sui,Toby Wise,Song Qi,Peter Dayan,Dean Mobbs

doi:10.1038/s41467-021-25816-9

Abstract

Natural observations suggest that in safe environments, organisms avoid competition to maximize gain, while in hazardous environments the most effective survival strategy is to congregate with competition to reduce the likelihood of predatory attack. We probed the extent to which survival decisions in humans follow these patterns, and examined the factors that determined individual-level decision-making. In a virtual foraging task containing changing levels of competition in safe and hazardous patches with virtual predators, we demonstrate that human participants inversely select competition avoidant and risk diluting strategies depending on perceived patch value (PPV), a computation dependent on reward, threat, and competition. We formulate a mathematically grounded quantification of PPV in social foraging environments and show using multivariate fMRI analyses that PPV is encoded by mid-cingulate cortex (MCC) and ventromedial prefrontal cortices (vMPFC), regions that integrate action and value signals. Together, these results suggest humans utilize and integrate multidimensional information to adaptively select patches highest in PPV, and that MCC and vMPFC play a role in adapting to both competitive and predatory threats in a virtual foraging setting.

Highlights

Natural observations suggest that in safe environments, organisms avoid competition to maximize gain, while in hazardous environments the most effective survival strategy is to congregate with competition to reduce the likelihood of predatory attack
Anterior to the mid-cingulate cortex (MCC) is the dorsal cingulate cortex, which has been linked to foraging decisions and the difficulty of such decisions, while the vmPFC has been linked to representations of choice value during foraging tasks[14,16], which is in line with its role in monitoring action value, exploration and economic decisions[17,18]
Our results demonstrate that humans adaptively select social environments based on their perceived patch value, using a competition-avoidant strategy when threat is absent and switching to a risk-diluting strategy in the presence of threat

Summary

Introduction

Natural observations suggest that in safe environments, organisms avoid competition to maximize gain, while in hazardous environments the most effective survival strategy is to congregate with competition to reduce the likelihood of predatory attack. In a virtual foraging task containing changing levels of competition in safe and hazardous patches with virtual predators, we demonstrate that human participants inversely select competition avoidant and risk diluting strategies depending on perceived patch value (PPV), a computation dependent on reward, threat, and competition. Perceived patch value (“perceived patch value”, or “PPV”), is represented as the overall potential benefit, whether in the safe domain via selection of less competition dense patches or risk dilution in the threat domain via occupation of more competition dense patches It is distinct from the observed social density of a patch, as social density can result in greater or lesser value depending on the danger posed by predators. Patch switch costs were zero, allowing us to investigate pure contextual changes in the perceived patch value of the decision

Methods

Results

Conclusion