Sensorimotor coordination in polymodal sensory integration
Poster

André Cyr
Cognitive Informatics, UQÀM

Pierre Poirier
Philosophy, UQÀM

     Abstract ID Number: 72
     Full text: Not available
     Last modified: March 14, 2003

Abstract
Sensorimotor coordination is an old concept in neuroscience, where it is well understood (at least in its basic form). But it is a relatively new idea in the emerging discipline, or theoretical outlook, that is embodied in cognitive science (Pfeiffer, 2002). Another major theme currently investigated by scientists is the polymodal sensory integration mechanism (Stein and Meredith, 1993). Although it has been less investigated for various reasons, the search for that which “merges the senses” is gaining in popularity, mainly because of its purported link to one of this century’s main quests, the mystery of consciousness (Crick and Koch, 2003). We believe that polymodal integration should be studied in the context of sensorimotor coordination (or more generally, embodied cognitive science). In order to sustain this claim, we propose a model where an embodied neural network (i.e., a neural network embedded in a robot) receives continuous and real-time data from its visual (light sensors) and somesthetic (pressure) sensory modalities, and where the output, or the effective part, is a motor driving a wheel. We show that sensorimotor coordination (i.e., motor outputs that structure subsequent sensory inputs) constrains (hence facilitates) polymodal integration.