Functional characterization of visual inhibitory mechanisms in motion perception

Motion perception in its early stages has traditionally been conceived as a process in which different sensors —orientation, spatial and temporal frequency tuned— operate in parallel and independently over the incoming external information. The key notion of independent and parallel processing developed by the sensors was challenged in 1987 when Derrington & Henning reported systematic errors in motion direction discrimination when human observers were presented with short duration compound stimuli. These compound stimuli were comprised of coarse and fine-scale components that excited both low and high spatial frequency tuned motion sensors at the same time. The fact that a joint activation of differently tuned sensors resulted in motion discrimination errors could have never been predicted under an independent and parallel processing scheme. Indeed, in 2007, Serrano-Pedraza, Goddard & Derrington proposed a computational model in which the information coming from low and high spatial frequency tuned sensors underwent inhibitory interactions in later stages of motion processing. The model was successful in replicating the main newly discovered results, thus unseating the prevailing notion of independent and parallel processing. Numerous efforts have been made trying to shed light on the nature of the interaction phenomenon. This way, several parameters (i.e. stimulus size, spatial frequency, contrast, etc.) have been seen to have some modulatory effect on it. In this sense, the present thesis aims at obtaining a deeper functional characterization of some of the aspects exerting such modulatory effect. Specifically, this thesis exhaustively explores the effect that the relative spatial frequencies in a compound stimulus (composed of two 2D Gabor functions drifting at the same time with the same speed) and their contrast have on the interaction under study; additionally, the effect of drifting speed or temporal frequency is exhaustively explored too..