Abstract
We describe experimental and modeling work related to a recently discovered orientation afterimage (Vidyasagar, et al, 1999). In this afterimage, sequential viewing of two orthogonally related patterns produces an afterimage of the first pattern. In recent work (Francis & Rothmayer, in press), we reported experimental measurements of this afterimage and modeled the effects with Grossberg's BCS-FCS theory. We now report new experimental results that both validate the basic explanation provided by the BCS-FCS theory and suggest that new mechanisms are required for the theory to remain consistent with the data. In particular, the experimental results suggest that the filling-in process in the FCS cannot be isotropic diffusion. We show how an alternative neural mechanism can account for our new results and continue to act much like diffusive filling-in for other contexts.