Overcoming the limitations of motion sensor models by considering dendritic computations

The estimation of motion is a fundamental process for any sighted animal. Computational models for motion sensors have a long and successful history but they still suffer from fundamental shortcomings, as they disagree with physiological evidence and each model is dedicated to a specific type of motion, which is controversial from a biological standpoint. In this work we propose a new approach for modeling motion sensors that considers dendritic computations, a key aspect for predicting single-neuron responses that had previously been absent from motion models. We show how, by taking into account the dynamic and input-dependent nature of dendritic nonlinearities, our motion sensor model is able to overcome the fundamental limitations of standard approaches.