Predicting Decision in Human-Agent Negotiation using functional MRI

Abstract

The importance of human-agent negotiation, and the role of emotion in such negotiations, have been emphasized in human-agent interaction research. Thus far, studies have focused on behavioral effects, rather than examining the neural underpinnings of different behaviors shown in human-agent interactions. Here, we used a multi-round negotiation platform, instead of the more common single-shot negotiation, and were able to find distinct brain patterns in emotion-related regions of the brain during different types of offers. Using multi-voxel pattern analysis to analyze brain imaging data acquired during functional MRI scanning, we show that it is possible to predict whether the negotiator concedes, does not change, or asks for more during the negotiation. Most importantly, we demonstrate that left dorsal anterior insula, which is known to be an emotion-related brain region, shows a different pattern of activity for each of the three offer types.


Back to Table of Contents