BrainMap: Jorie Koster-Hale, PhD. Inside the social brain: representations in the neural basis of theory of mind

Wednesday, December 9, 2015 - 12:00 to 13:00
149 13th Street (Building 149), main second floor seminar room (2204)

Jorie Koster-Hale, PhD; Postdoctoral Fellow, Harvard University

Talk title: Inside the social brain: representations in the neural basis of theory of mind

Abstract: Social life depends on understanding why people do the things they do and what they are likely to do next. A cornerstone of the human capacity for social cognition is the ability to reason about the unobservable, internal causal structure underlying these actions: the person’s intentions, beliefs, and goals. A remarkable body of evidence has demonstrated that social cognition reliably and selectively recruits a specific group of brain regions. Building on prior work, which has largely focused on where in the brain mental state reasoning occurs, the research here investigates how neural populations encode mental state concepts. I find that “social brain” regions contain neural signatures of abstract mental state reasoning, and provide an initial characterization of their representational content and computational format.

First, I argue that distinct features of mental states are represented in social brain regions: intention, knowledge source, and emotion. These codes are specific to regions implicated in social cognition, differ across regions within the social network, and predict behavioral responses at the level of single items. Second, I argue that these neural representations are encoded by a hierarchical, generative model. Specifically, I adapt a computational model developed for mid- and high-level vision to social cognition: hierarchical predictive coding (HPC).  HPC models posit that the brain generates continuous predictions of upcoming events and then adjusts these predictions by computing an error signal that tracks deviations between predicted and observed events. I show that the HPC model provides a unified framework for many seemingly heterogeneous findings in social neuroscience, and I provide novel evidence supporting key predictions of the HPC model in social reasoning.