Shared and idiosyncratic neural processing of naturalistic communication in the default mode network

Abstract

Humans have a unique ability to share knowledge of the world with each other via communication. A finite set of spoken sounds and written symbols can be used in as diverse situations as telling a personal story, ordering coffee, or sharing scientific results. However, the remarkable flexibility in the form and content of communication is matched with inherent ambiguity in its interpretation. A listener might misunderstand a speaker, lack the appropriate context or background to process the information, or simply interpret the same information in a different way than is intended. This dissertation aims to elucidate neural processes underlying shared and idiosyncratic interpretation of naturalistic communication. We argue that shared understanding of information across forms of communication is reflected in the shared dynamics of neural responses in the default mode network, both among receivers of information and between senders and receivers. Using a novel method, intersubject representational similarity analysis, we first show that regions of the DMN are sensitive to individual differences in the interpretation of an abstract, ambiguous narrative: across individuals, similarity in narrative interpretation is correlated to similarity in neural activity (Chapter 2). At the same time though, we find that the DMN is remarkably invariant to the modality of the narrative. The same narrative presented in strikingly different forms elicits similar neural responses in the DMN, but only when interpretation of the narrative is shared across forms (Chapter 3). We then demonstrate that shared neural responses in DMN emerge from the interaction of brains in communication in a non-narrative context. We find that the neural activity of teachers and students learning abstract technical information are coupled during communication and that the magnitude of this coupling is related to learning outcomes (Chapter 4). We ultimately argue for a new view of the DMN as a system for integrating internal representations with incoming, external information, resulting in both shared and idiosyncratic neural representations over long timescales. This unique interplay between extrinsic and intrinsic forces enables us to both shape and be shaped by other minds and provides a mechanism for negotiating a shared neural code underlying communication (Chapter 5).