Abstract
Recent data (e.g., Zhaoping 2008; Zhaoping and Zhe, 2015) support the hypothesis that the primary
visual cortex (V1) in primates creates a saliency map from visual input to guide attention
exogenously (Li, 1999, 2002). This map can be read out through monosynaptic projections from
V1 to the superior colliculus (SC), which can select the most salient location as the target of a gaze
shift. Across species, the SC, called optic tectum in lower vertebrate, receives retinal input.
Meanwhile, non-mammals lack neocortex, with only a small fraction of their forebrains devoted
to visual processing. I propose that evolution leads to a migration of the saliency map from the
tectum in lower vertebrates such as fish and birds to V1 in primates. Phylogenetically across
vertebrates, including fish, frogs, birds, mice, rats, cats, and primates, this migration is
accompanied by an expansion of the forebrain, a migration of visual analysis from the optic
tectum to the forebrain (in particular to V1), a smaller percentage of retinal projections to the
tectum, and a smaller percentage of retinal ganglion cells tuned to complex features such as motion
direction. Effects of lesioning V1 or the tectum are consistent with this hypothesis.
