The top-down nature of visual processing is also evident in Gestalt figures such as the reversing duck-rabbit (figure 4.1). If you are expecting to see a duck, then you will see the lines on the left as constituting a bill, whereas if you shift your focus and look for a rabbit, they will appear as ears. This kind of inference consists not of serial linguistic steps but rather of the parallel dynamic interaction of neurons that encode sensory information with neurons that encode expectations and knowledge of
what ducks and rabbits look like.