It strikes me that it's less specialized to processing sensory input from light and more specialized to constructing a model of where things can be expected to

It strikes me that it's less specialized to processing sensory input from light and more specialized to constructing a model of where things can be expected to be in relation to where the body is. So it's like a proprioceptory map. It's interesting because I can look around my bathroom in the light, turn the light off, then reach for and grasp EG a toothbrush from memory on the first try without feeling around. It's like the images are a map of where things are and that map doesn't require the sensory images once it's constructed, nor does it require visual stimulation to construct.