Attraction Effect in Risky Choice Can Be Explained by Subjective Distance Between Choice Alternatives

Peter N C Mohr,Hauke R Heekeren,Jörg Rieskamp

doi:10.1038/s41598-017-06968-5

Abstract

Individuals make decisions under risk throughout daily life. Standard models of economic decision making typically assume that people evaluate choice options independently. There is, however, substantial evidence showing that this independence assumption is frequently violated in decision making without risk. The present study extends these findings to the domain of decision making under risk. To explain the independence violations, we adapted a sequential sampling model, namely Multialternative Decision Field Theory (MDFT), to decision making under risk and showed how this model can account for the observed preference shifts. MDFT not only better predicts choices compared with the standard Expected Utility Theory, but it also explains individual differences in the size of the observed context effect. Evidence in favor of the chosen option, as predicted by MDFT, was positively correlated with brain activity in the medial orbitofrontal cortex (mOFC) and negatively correlated with brain activity in the anterior insula (aINS). From a neuroscience perspective, the results of the present study show that specific brain regions, such as the mOFC and aINS, not only code the value or risk of a single choice option but also code the evidence in favor of the best option compared with other available choice options.

Highlights

Recently have researchers started to explicitly model the data of context effects using quantitative models, such as MDFT15. We go beyond these past approaches by (a) providing further evidence that decisions under risk can be influenced by context effects, such as the AE, (b) examining whether these effects can be described by evidence accumulation models like Multialternative Decision Field Theory (MDFT), and (c) determining how the neurobiological underpinnings of the AE are related to the decision process assumed by MDFT
To examine whether MDFT described the data well, we rigorously tested it against Expected Utility (EU) theory and against a baseline chance model
The results showed that MDFT outperformed the other two models on an aggregated level

Summary

Results

We further observed significant negative correlations between the choice probability and brain activity in the bilateral anterior insula (aINS), dorsomedial prefrontal cortex (DMPFC), thalamus, dorsolateral prefrontal cortex (DLPFC), and parietal cortex (see Fig. 3B) These regions have been identified in a meta analysis to code for risk in decision making[17]. Subjects with lower overweighting of the distance in the dominance direction showed a decreased correlation between brain activity and predicted choice probability of the chosen option

Discussion

Methods