A model of V1 visual contrast processing utilizing long-range connections and recurrent interactions

We propose a functional architecture of recurrent long-range processing in primary visual cortex. Simulations of noisy artificial stimuli and natural stimuli show that the circuitry increases responses of cells coherent with a global structure while suppressing conflicting responses. Quantitatively, we demonstrate that orientation significance is increased at object borders while remaining unchanged in homogeneous regions. We suggest that the proposed circuit serves as a low-level neural implementation of the Gestalt principle of good continuation, thus building a preliminary stage for perceptual grouping and figure-ground segmentation