This modeling study investigates whether an orderly convergence of neuronal selectivities from cortical areas V1 and V2 can produce curvature selective receptive fields found in area V4 recordings. A model of fast object recognition in the ventral visual pathway is made of spiking neurons based on recent findings about the convergent properties of V2 neurons on V1 afferents. A novel proposal is made for how V4 neurons may create selectivity for local curvature through the orderly convergence of different types of afferent inputs from V2. The model demonstrates a novel method for simulating spiking neurons using tensor programming and GPU hardware: Assuming that convergent functional micro-architectural patterns repeat in topographically organized visual space, the details of individual unit depolarization and spike time is modeled using convolution operations combined with a model of post-synaptic potentials. The few parameters in the model are set manually, and this is sufficient to qualitatively reproduce the V4 recordings. This demonstrates that convolution technologies can be used for realistic neuron simulations.
