|
 Visual perception is an interactive
process involving prediction, identification, classification, and
decision or reaction. My
interest is in understanding how the visual cortex takes local
information about discontinuities (e.g., edges) and forms a global
percept. Specifically, how does
the cortex integrate distributed signals in order to identify correlations
and patterns among those discontinuities? The problem is compounded by the fact
that patterns and objects are typically embedded within an environment
that also contains structure and correlation. We use knowledge from natural scene
statistics, psychophysical data, and computer vision algorithms to
formulate hypotheses, and we employ rigorous statistical analyses on
simultaneous recordings from multiple neurons distributed across the
cortical network. Our results provide
us with clues on how the brain is able to segment and identify objects,
as well as reveal properties of the underlying cortical mechanisms.
|
