Mechanisms Underlying Economic Revealed Preference Of Multi-Component Bundles In Primate Orbitofrontal Cortex

Abstract

Classic neuro-economic studies suggest that people and animals alike assign values to individual attributes arriving at a total summed valuation and therefore overall appraisal of the offer (Kahnt et al., 2011; Lak et al., 2014). However, further work is needed to determine whether such individual attribute appraisals can predict choice between multi-component bundles. In this paper we show the importance and applications of economic theory to decision neurophysiology concerning multi-component attribute options or bundles. We applied random utility modelling (McFadden, 1973) to choices amongst 10 different two-component bundles in non-human primates (Macacca mulatta). We found that behavioural random utility (RUM) complies well with revealed preference theory and recapitulates the fundamental properties of empirically obtained indifference curves (IC). Bundles positioned in the same IC display equal choice preference and similar utility whereas bundles positioned across different ICs display different choice preferences. RUM also captures the characteristic shape of IC on each bundle revealing the synergetic interactions between components. We extended RUM to neurophysiological data and focused our investigation to orbitofrontal cortex (OFC), area 13. Neuronal RUM complied with behavioural RUM at single neuron and population level (N: 54). Neural coding of utility was present at target onset, choice, and reward. However, a distinct and separate group of neurons (N: 26) showed partial compliance with utility in either the chosen or the unchosen bundle options and correlated with both utility and probability of choice. We show that relative chosen-utility coding constitutes a separate type of computation and suggest a role for this type of neuron in value-to-choice processing in OFC.