Retina is made of brain tissue. (Or could we say brain is made of retinal tissue?)
The most primitive cephalochordates such as amphioxus have a single frontal eye, suggesting that our first chordate ancestors may also have been cyclopean. A cyclops' visual input is more absolutely inverted than binocular input, as shown in the schematic diagram above. (Though even in a binocular arrangement an absolute crossover is still required to correlate dorsal with ventral, with only a partial crossover to correlate left with right.)
In embryonic development, optic placodes form very early on, as the diencephalon forms. The greatest amount of brain development occurs after the eyes start forming.
All embryonic central nervous system anterior to the optic placodes is inverted while all CNS posterior to the optic placodes is not. The eyes seem to define a cutoff point between inverted CNS and the non-inverted CNS, suggesting they have some importance in the matter.
Chordate forebrains are inverted. Not only do the right and left hemispheres govern and sense opposite sides of the body, but the dorsal and ventral halves of the somatosensory and motor cortex also correspond to the lower and upper regions of the body respectively. This arrangement would seem to only complicate things and add extra distance to signal pathways, unless you assume it evolved by necessity to correlate some inverted sensory input with the rest of the body. Input from the eyes, and only the eyes, is inverted in exactly this way. (Yes, modern forebrains do a lot more than process vision, but they may at one time have had only this function.)
Mollusc eyes evolved in parallel to ours, so they work the same way but do not share the same heritage. Where does the signal cross over in octopi and squid? What about their embryonic development?
Bat optic nerve paths cross over more simply than ours. Is their visual layout more common in mammals, or is it just deevolution caused by dependency on sonar?
After writing this I realize that dorsal and ventral areas of human cortex correspond to posterior and anterior regions of the body, when the theory would predict that they correspond to ventral and dorsal regions. This isn't a show-stopper, but I'd feel a lot better if I could find a cortical map of a rat, for whom upper is dorsal.
November 30 2005