"Understanding brain function is one of the outstanding challenges of modern biology. Many studies focus on mammalian neocortex, a modular and versatile structure that operates equally well with different sensory inputs and for perception, planning as well as action. Neocortex, however, is remarkably complex. It contains many cell types, six layers, networks with local and long-range connections, and its study is technically challenging. We propose here to address central issues of cortical computation using a simpler experimental system. Neocortex evolved from a more primitive cortex, likely present in the ancestors of all amniotes. Extant reptiles are closest to this putative ancestor: their cortex contains only three layers, two of which are nearly exclusively neuropilar. Reptilian cortex is also closest to mammals’ old cortices (piriform and hippocampus). Like in mammals, reptilian cortex is modular. Its design, however, is considerably simpler and more ubiquitous than in mammals. Indeed, so far as we know, reptilian primary olfactory and visual cortices are very similar to one another. Finally, certain reptiles such as turtles have evolved biochemical and metabolic adaptations to resist long periods of anoxia. Thus, their brains can be studied ex vivo over long periods, giving experimenters access to the entire brain with an intact retina or nasal epithelium. We will use this system to study cortical computation, primarily in visual and olfactory areas. Using electrophysiological, imaging, molecular, behavioral and computational methods, we will discover the representational strategies of these two cortices in vivo, the functional architecture of their microcircuits and the computations that they carry out. This understanding of generic and ancient units of cortical computation will illuminate our studies of more complex and sophisticated cortical circuits."