W. R. Kenan Jr. Professor of Psychology
Propagating covariation and segregating mechanism: Using modeling to understand brain evolution
Two observations stand out in mammalian forebrain evolution. First, the number of “cortical areas” making up the neocortex, typically topographic maps of sensory, motor, or computed representations, increases systematically with cortex size. Second, the volume of a distinct set of regions, including olfactory cortex, the hippocampus, and adjacent areas covary highly with each other but negatively with the neocortex. Historically, a species’ relative reliance on visual or olfaction has been the explanation offered for this variation, but is an increasingly unsatisfactory one. Modeling visual cortical area proliferation shows that nearest-neighbor covariation arising from the spontaneous activity of a fixed primary area is adequate for a series of visual maps to self-organize in adjacent unoccupied cortex, increasing in number with available area. The developmental segregation and negative correlation of the olfactory-hippocampus complex with neocortex volume may reflect a basic mechanistic challenge rather than the needs of sensory specialization alone. The efficiency of recovering temporal correlations from spatially distributed input, as is seen for odorants in olfaction or events in the hippocampus, versus the extraction of spatially-correlated information as in vision or somesthesis, is much improved if the two operations do not share the same substrate.