On the visual system's architecture underlying binocular rivalry and motion perception

Abstract

Our everyday visual perception is supported by a complicated set of interactions between different brain areas. These areas often have a specific function. A lot of communication takes place between, and also within, these areas. The total set of interactions between and within the different brain areas is the "architecture of the visual system". Although a lot is known about the visual system's architecture from histological research, there is still much unknown about the functional architecture (about how the interactions actually work). In the first part of the thesis, I have investigated functional architecture underlying human perception of motion, and in particular the speed of motion. In the second part, I have researched the human perception in circumstances in which a large conflict exists between the inputs in the two eyes (a circumstance in which binocular rivalry results). Motion consists of two components: a direction and a speed. It is generally accepted that the direction-component is processed by a single system (that is, a collection of interconnected brain areas). However, for the speed of motion no such general agreement exists. It has often been proposed that speed-processing is supported by two (or more) systems: one for low speeds and one for high speeds. In this thesis I have proposed that the data know up to now also permit a simpler explanation, namely that a single system processes the entire range of visible speeds. I have tested this prediction with computational models and experiments. The results point out that a single speed-sensitive system indeed may explain the data in the literature and the newly-obtained data of this thesis. Binocular rivalry results when our two eyes receive conflicting images. In stead seeing an average of the two images, we see a continuing alternation of the images. The images fight, as is were, for dominance, hence the name binocular rivalry. It has been well-studied how spatial aspects of stimulation influence the rivalry behaviour. However, temporal aspects are not so well-studied. In this thesis I started the investigation on the influence of temporal aspects of stimulation on binocular rivalry. Remarkably, binocular rivalry is quite insensitive to a temporary blanking of the stimulation. The two images need not even be presented simultaneously to obtain rivalry, a long as they are repeatedly shown within 350 milliseconds. This temporal limit to rivalry is dependent on form information and independent of motion information, and is even found when the two images are shown to a single eye. The experiments add to the discussion on how rivalry takes place; one theory says that rivalry takes place between the neurons specific to information from different eyes, while another theory says that the rivalry takes place between neurons coding for different patterns. The data in this thesis are mainly support for the second theory. The data provide new insights into the architecture of the visual system. Firstly, the thesis shows that a single system suffices for human speed perception, opposing theories that claim that at least two systems are necessary. Secondly, the thesis strongly suggests that binocular rivalry is dependent on pattern-based competition and not eye-based competition, and therefore probably takes place at a level within the visual system where eye-specific information is lost.