Mapping the Self: A Network Approach for Understanding Psychological and Neural Representations of Self-Concept Structure

How people self-reflect and maintain a coherent sense of self is an important question that spans from early philosophy to modern psychology and neuroscience. Research on the self-concept has not yet developed and tested a formal model of how beliefs about dependency relations amongst traits may influence self-concept coherence. We first develop a network-based approach, which suggests that people's beliefs about trait relationships contribute to how the self-concept is structured (Study 1). This model describes how people maintain positivity and coherence in self-evaluations, and how trait interrelations relate to activation in brain regions involved in self-referential processing and concept representation (Study 2 and Study 3). Results reveal that a network-based property theorized to be important for coherence (i.e., outdegree centrality) is associated with more favorable and consistent self-evaluations and decreased ventral medial prefrontal cortex (vmPFC) activation. Further, participants higher in self-esteem and lower in depressive symptoms differentiate between higher and lower centrality positive traits more in self-evaluations, reflecting associations between mental health and how people process perceived trait dependencies during self-reflection. Together, our model and findings join individual differences, brain activation, and behavior to present a computational theory of how beliefs about trait relationships contribute to a coherent, interconnected self-concept.