Proceedings of the 7th Annual Computational Neuroscience Meeting
July 26-30 - 2000 - Santa Barbara (California, USA).
 
Motion processing using one spike per neuron
 
William Paquier, Arnaud Delorme & Simon J. Thorpe
 

Centre de Recherche Cerveau et Cognition (UMR 5549) Faculté de Médecine de Rangueil, 133 route de Narbonne, 31062 Toulouse, France.

william@cerco.ups-tles.fr, arno@cerco.ups-tlse.fr, thorpe@cerco.ups-tlse.fr

 
    The rapidity of motion processing poses problems for classical rate coding models because individual neurons may only have enough time to generate one spike. An alternative coding scheme based on the order of firing of neurons has previously been used successfully to process static images. Here we show that the same form of rank order coding can be used to process dynamic sequences of natural images. Starting with transient and sustained ON- and OFF-center retinal ganglion cells, the model generates responses selective to complex motion patterns that include rotations, expansions, contractions and deformations similar to those seen in MST.