[Segmentation, grouping and accentuation during stimuli perception]

Abstract

The paper is concerned with grouping, segmentation and accentuation occurring in the processes of stimuli perception. An universal model of these events is based on vector coding in neuronal networks. Grouping is unification of objects or events into collections according to their similarity. Segmentation is separation of such groups up to small ensembles of units. In neuroscience grouping and segmentation are regarded as referred to neural mechanisms underlying perceptual and semantic processes resulting in a phenomenal attachment or separation. It is assumed that stimuli in neuronal nets are encoded by combinations of excitations of cardinal neurons constituting excitation vectors. Differences among stimuli are formed as absolute values of their excitation vector differences. The more different are stimuli the separate are their perceptual and semantic representations. The more similar are respective stimuli, the less is their separation. It suggests that stimuli having similar excitation vectors would be grouped together. On the contrary stimuli with opposed excitation vectors would be segmented and pushed to different ensembles. The vector encoding is expressed also for location in space. Thus spatial separation of objects is increasing with the increasing of their spatial excitation vector differences. The universal principle of vector encoding of differences can be illustrated by color contrast: differences of contrast colors rise with increase of their excitation vector differences. Objects having similar excitation vectors constitute a group accentuated due to summation of their excitation vectors. Groups of objects characterized by different excitation vectors are mutually accentuated by a contrast mechanism. A plastic accentuation depends on novelty of stimulation being habituated during repeated stimulus presentations.