Item

Abstract

Computational modeling of visual information process- ing can lead to important new insights about the function of visual cortex. Here we asked whether we can build a proof-of-concept model that implicitly learns known cor- tical organization principles. We chose cortical magnifi- cation, which refers to the fact that more cortical tissue is dedicated to the processing of the foveal as compared to peripheral visual field. We built a brain-optimized convo- lutional neural network model trained to predict brain ac- tivity across twelve retinotopic regions as measured with functional MRI. We treated cortical magnification as a free parameter, using multivariate Gaussian distributions act- ing on the network’s feature representations. Our results demonstrate that cortical magnification can, indeed, be learned implicitly, demonstrating the general feasibility of our computational modeling approach.