This pilot study aims to demonstrate that social cognition and underlying social and non-social cognitive processes can be measured with fMRI during real social (verbal) interactions. Social cognition designates the cognitive processes that support social interactions. These processes are highly affected in people with schizophrenia and are the best predictors of these patients’ functional outcome. Despite the impact of social cognition deficits, we however know little about the neural mechanisms that lead from difficulties in understanding others to difficulties in interacting with others as previous social cognition studies have exclusively relied on non-interactive tasks. Though measuring social cognition during interactive tasks is a great challenge, a SSHRC-funded research project led by Amelie Achim (Université Laval, psychiatrie et neurosciences) and Marion Fossard (Université de Neuchâtel, Département de logopédie) has recently demonstrated that social cognition can be measured behaviourally during real, controlled social interactions using a referential communication paradigm (see project’s methods) and a fine analysis of verbal/referential productions. This project also showed that schizophrenia patients have difficulties adapting their verbal/referential productions to their interlocutor’s likely knowledge (Achim et al., 2012). Our new collaboration with Pascale Tremblay (Université Laval, Rehabilitation), a specialist of verbal communication and cognition who has a unique expertise in assessing speech production using fMRI, is central to bringing this method to fMRI to understand the nature of the neurocognitive mechanisms that are at play during real social interactions and that impact effective communication. As part of this pilot project, ten participants with no psychiatric illness and four patients with schizophrenia will be scanned while performing an adapted version of our referential communication paradigm. This interactive task requires the participants to present a series of characters to an interlocutor (who will be in the control room), so that the interlocutor can identify the character from a larger set of images. All the characters will be known to the participant but will be selected so that only half on them can safely be assumed to be known to the interlocutor, a woman in her twenties. As observed in our previous behavioural study, we expect that our participants will adapt their referential expressions to the state of knowledge of their interlocutor (i.e. use more names for likely known character and more description for likely unknown characters), but that this pattern will not be as pronounced in schizophrenia patients. A sparse sampling fMRI protocol, which involves periods of silence (MRI gradients turned off) interleaved with periods of data acquisition, will be used to mitigate motion concerns, and to record verbal responses using an MRI compatible microphone. This project relies on a very innovative and interactive protocol whose design results from the combined expertise of three young researchers with complementary background in social cognition (AA), in linguistics (MF) and in speech and hearing research (PT). Though there is a recognized need to study social cognition with interactive paradigms to truly understand how these processes influence everyday interactions, few such tasks exist and those that do typically allow for a very limited set of behaviour (ex: press one of three buttons), limiting the ecological validity that is required to understand the impact of social cognition processes on real social interactions. Once our paradigm is validated through the current RBIQ pilot project, we are confident that this will lay the foundation for successful CIHR or NARSAD applications for funding to understand the neural bases of social deficits in schizophrenia, and eventually lead to new collaborations to study these deficits in other clinical populations with recognized social cognition and social interaction deficits.