A Dynamic Field Theory of Visual Recognition in Infant Looking Tasks

Infant looking tasks are fruitful sources of data on infant cognition. Researchers using these tasks measure differential looking times as infants look at multiple stimuli. These looks are used to infer stimulus recognition and visual processing of familiar and novel stimuli. We propose here that looking behavior in two of these tasks, infant categorization and visual paired comparison, arise from a common visual recognition mechanism. We propose a dynamic field theory (DFT) of visual recognition in infant looking tasks and describe simulations using a dynamic neural network that capture looking behavior associated with central phenomena in the infant categorization and infant recognition memory literatures. Critically, these phenomena are supported by a common recognition mechanism.