The Neural Organization of Social Knowledge

Abstract

Humans enjoy social lives of terrific complexity, but this sociality exacts high demands on the individual. The typical adult must interact with hundreds of others on an ongoing basis. Each such person in our life embodies a unique constellation of mental states and traits. How can we possibly navigate this social complexity? Two crucial tools for any navigator are their map and compass: knowledge of the terrain, and a set of cardinal directions they use to organize this knowledge. Across three studies, this dissertation investigates the social analogs of these navigational tools: knowledge of other people, and the psychological dimensions the brain relies on to organize this information. Using advanced neuroimaging techniques, these studies systematically map domains of social knowledge to reveal their organizing principles. Paper 1 investigates the neural representational of others’ mental states. Results suggest that three psychological dimensions – rationality, valence, and social impact – explain nearly half of the variance in how the brain makes sense of others’ thoughts and feelings. Paper 2 investigates the domain of personally familiar others – how people represent individuals, such as friends and family, with whom they interact on a regular basis. The data suggest these individuals are represented by both coarse and fine-grained patterns distributed throughout brain regions involved in social cognition. Paper 3 investigates the representation of famous others, with the aim of identifying dimensions which explain how the brain represents well-known individuals with whom one lacks direct contact. Results indicate that four established theories of person perception can each accurately predict patterns of brain activity associated with famous individuals. Moreover, a synthetic model combining these established theories achieved two-thirds of the accuracy of a hypothetical ideal theory. Combining data from Papers 1 and 3 yields another insight: encoding models trained to predict person-specific patterns of brain activity can also predict state-specific activity patterns. This suggests that the brain represents others’ mental states and traits in a partially-shared representational space. Together, these Papers advance our understanding of how the brain organizes social knowledge, and contribute to synthesizing more general and accurate theories of social cognition.