Project 3. Project Summary. This Project 3 derives from Project 3 in our current Conte Center. It has the only new PI in our Conte Center, new faculty at Caltech Dean Mobbs (moving from Columbia University in summer, 2016), together with senior Co-PI Colin Camerer. Its overarching goal is to understand how social inference and context guide social decisions under threat, a counterpart to Project 2, which investigates how these processes instead guide prosocial altruistic decisions. There are four Aims, which trace this theme from the buffering safety value of other people, through investigations of social versus nonsocial threat, to investigations of how we behave when other people block our goals (closely related to the RDoC construct of “frustrative non-reward”). The last Aim, as in Projects 1 and 2, investigates individual differences. As with the other Projects, there are close links to Aims investigated with different approaches. For instance, Aim 1 under Project 4 will use single-unit recordings from the amygdala to investigate neuronal responses to specific types of threat, from basic to social. These links are also reflected in the personnel of Project 3, which includes PIs Adolphs and Hutcherson (from Project 2), O'Doherty (Project 1), and Rutishauser (Project 4, for the single-unit recording link), as well as several shared post-docs and students. The Aims leverage the expertise of PI Mobbs in constructing more ecologically valid protocols that can be used in the fMRI environment, have quantifiable parameters, but capture ecological contexts related to social threat in the real world. It will test 60 healthy participants recruited through Cores 2 and 3 in six fMRI experiments. Aim 1 investigates how the competence and reputation of other people can serve as a buffer against physical threats (electric shock). Aim 2 investigates dynamic decision-making in a virtual foraging task, to understand how people's choices trade reward probabilies against social or nonsocial threats. Aim 3 investigates frustrative nonreward in a social setting: what responses are elicited when other people block our goals-- and how this may vary depending on whether or not the other person has an excuse or is blocking us intentionally. Taken together, Projects 1, 2 and 3 build from a very computational framework with very quantitative tasks, to a more standard social neuroscience framework, to a framework maximizing ecological validity. They all involve fMRI studies in healthy people, and they all have links to Aims under Projects 4 and/or 5 that extend the investigation to single-unit studies or lesion studies.