Distance measuring as a model of decision-making under risk

Summarise with VIRAT
Authors: Daniel Shani, M. Sable-Meyer, S. Veselic, E. Gutierrez, Kristopher Torp Jensen, Janahan Selvanayagam, Laurence Tudor Hunt, Peter Dayan, Steven Wayne Kennerley, Timothy E.J. Behrens
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Recent findings of grid coding of probability (P) and magnitude (M) in non-human primate (NHP) frontal cortex (Veselic et al. 2025, Bongioanni et al. 2021) are puzzling. Grid cells have been shown to be an optimal code for space (Dorrell et al. 2023), but why would they be used in risky decision making? Here, we hypothesise that NHPs are using spatial computations to solve these problems. That is, we propose that NHPs choose between options by comparing their distances from a suitable reference point in an internally generated metric space. First, we show that varying the parameters of the metric makes the distance measuring choice emulate both classical findings in human and animal literature regarding the shapes of the component functions of prospect theoretic choice. Then, we analyse the behaviour in two datasets of NHPs choosing between risky options. Having fit parameters of the metric, we show that the distance measuring choices match NHP behaviour more closely than various other common models of risky choice, including variants of Prospect Theory (PT) (Kahneman & Tversky 1979). We also use the distance measuring framework to generate new model-agnostic predictions, which we validate in the data. We provide a possible explanation for the phenomenon of abstract frontal grid cells, and provide novel predictions in terms of both behaviour and neuronal activity.

Fields of Study & Subjects

Decision-Making and Behavioral Economics Neural and Behavioral Psychology Studies Memory and Neural Mechanisms

Publication details

Journal / source
MPG.PuRe (Max Planck Society)
Publication date
1 March 2026
Language
EN