Poster F30, Tuesday, March 28, 8:00 – 10:00 am, Pacific Concourse
Neural representations of person types overlap with Theory of Mind regions
Connor Lane1, Giulia V Elli1, Marina Bedny1; 1Johns Hopkins University
We routinely try to predict people’s behavior and mental states. When we do so, we consider the individual characteristics of the agent. I.e. how likely is this particular person to think and act in this way? We investigated the neural basis of this mental state reasoning using representational similarity analysis (RSA) . While undergoing fMRI, participants judged whether a person type (e.g. bride, veteran, toddler) was more likely, equally likely, or less likely than the average person to engage in a mental act (e.g. “nuns respect”, “fools realize”). Thirty-one person types occurred in all possible pairings with thirty attitude verbs. Participants also completed a Theory of Mind (ToM) localizer where they read stories that required reasoning about beliefs (Dodell-Feder et al., 2010, NeuroImage). A separate group of Amazon Mechanical Turk participants judged the semantic similarity of person types to each other in all possible pairings, generating a people similarity space. An fMRI similarity space was generated by (1) averaging fMRI activity during judgments for each person type across all possible verbs, and (2) computing correlations between activity patterns for all person-types. We used searchlight RSA to compare the behavioral and fMRI similarity spaces. We show that semantic similarity between person types predicts fMRI pattern similarity within the ToM network. These results suggest that brain regions underlying ToM reasoning are sensitive to distinctions between person types.
Topic Area: EMOTION & SOCIAL: Person perception