Guillaume Dumas – Operationalizing Social Neuroscience through Human-Human and Human-Machine Interactions

Speaker: Guillaume Dumas (Institut Pasteur & Hopital Robert-Debré, Paris)

Title: Operationalizing Social Neuroscience through Human-Human and Human-Machine Interactions

How are neural, behavioural and social scales coordinated in real time so as to make possible the emergence of social cognition? Answering this question requires to study the dynamics of coordination in real human interactions. However, even at the simplest dyadic scale, methodological and theoretical challenges remain. Several theories have been proposed to infer the link between neurobiology and social psychology, but the dynamical components of human interaction are still poorly explored because of the difficulty to record simultaneously the brain activity from several subjects. This is the goal of hyperscanning methodology. I will first present how the combination of situated social paradigms with hyperscanning allowed demonstrating that states of interactional synchrony at the behavioural level correlate with the emergence of inter-individual synchronization at the brain level (Dumas et al. PLoS ONE 2010). The inter-brain synchronization in different frequency bands appeared to reflect different aspects of social interaction, such as interactional synchrony, anticipation of other’s actions and co-regulation of turn-taking. Then, I will present how such phenomena can be simulated with biologically inspired numerical simulations (e.g. using direct measures of brain connectivity with DTI) and how the human connectome facilitates inter-individual synchronizations and thus may partly account for our propensity to generate dynamical couplings with others (Dumas et al. PLoS ONE 2012). Finally, I will present another tool called the Human Dynamic Clamp (HDC) (Dumas et al. PNAS 2014). This HDC integrates equations of human motion at the neurobehavioral level. A human and a “virtual partner” are then reciprocally coupled in real-time, which allow controlling the dynamical parameters of the interaction while maintaining the continuous flow of interaction. This technique scaled up to the level of human behaviour the idea of dynamic clamps used to study the dynamics of interactions between neurons. Combining human-human and human-machine interactions thus presents new approaches for investigating the neurobiological mechanisms of social interaction, and for testing theoretical/computational models concerning the dynamics at the neural, behavioural and social scales.