Leveraging human-computer interactions and social presence with physiological computing. (Améliorer les interactions homme-machine et la présence sociale avec l'informatique physiologique)

This thesis explores how physiological computing can contribute to human-computer interaction (HCI) and foster new communication channels among the general public. We investigated how physiological sensors, such as electroencephalography (EEG), could be employed to assess the mental state of the users and how they relate to other evaluation methods. We created the first brain-computer interface that could sense visual comfort during the viewing of stereoscopic images and shaped a framework that could help to assess the over all user experience by monitoring workload, attention and error recognition.To lower the barrier between end users and physiological sensors,we participated in the software integration of a low-cost and open hardware EEG device; used off-the shelf webcams to measure heart rate remotely, crafted we arables that can quickly equip users so that electrocardiography, electrodermal activity or EEG may be measured during public exhibitions. We envisioned new usages for our sensors, that would increase social presence. In a study about human-agent interaction, participants tended to prefer virtual avatars that were mirroring their own internal state. A follow-up study focused on interactions between users to describe how physiological monitoringcould alter our relationships. Advances in HCI enabled us to seam lesslyintegrate biofeedback to the physical world. We developped Teegi, apuppet that lets novices discover by themselves about their brain activity. Finally, with Tobe, a toolkit that encompasses more sensors and give more freedom about their visualizations, we explored how such proxy shifts our representations, about our selves as well as about the others.