Modeling the projection geometry of the retinotopic visual hierarchy

The original model of the geometry of human primary visual cortex (V1) was the complex log polar model (Daniel & Whitteridge, 1961; Schwartz, 1977, Duncan & Boynton, 1973).  However, this model predicts a foveal meridional anisotropy that is not observed in human fMRI data (Fig. 1). The same anisotropy predicted by the later Schwartz complex log dipole model (Schwartz, 1984) is corrected by the parametric shear model of human V1 by Schira, Wade & Tyler (2007), as seen in Fig. 2. The extension of these analyses to the higher retinotopic regions V2-3 using the wedge-dipole concept (Polimeni et al., 2006) reveals further distortions in the foveal mapping, which are corrected with a banded double sech model (Fig. 3), providing a much better match to human retinotopic mapping data (Tootell et al., 1998).