Feature binding is the process whereby different features such as shape, colour, size, orientation, location, etc. are linked together to form a coherent representation of the object. It is a ubiquitous physiological sequence and an essential phase in information processing, for it provides the basis of mental representations, which in turn, are requisite for all cognitive functions.
It is important to realize though, that binding is not an isolated process. There are myriad stimuli impinging on our senses at all times, vying to gain entry into our consciousness. Further, not only does sensory input emanate from a complex, dynamic environment, but it also enters a neural system that is already activated by previous inputs and is oriented towards future goals. Which aspects of the momentary sensory input are selected for further processing depends as much on the state of the system as it does on the sensory input itself. Indeed, some fundamental questions one may ask about binding are whether, why, and how, some features are selected for binding at the cost of others.
The bottom-up view of information processing is that the input received by the brain is processed in a largely automatic way to the higher centers in the brain. The physiological basis of binding is postulated to be either conjunctively coding neurons, or synchrony among participating neural networks to encode features and out of phase neural activity to encode separate objects. But, mere perceptual integration of features, whether by synchrony or by specialized neurons, does not even begin to capture the implication that binding results in coherent objects, fundamental for further information processing. An object is not only a bundle of features. At the very least, the features need to be integrated so that the object can be distinguished from other objects. This implies selection and manipulation of the basic information supplied by separate features. The top-down view of information processing contends that binding is more influenced by the reentrant processes (the downward and lateral feedback to the lower areas, emanating from the higher centers of the brain). Reentrant processes not only help to confirm what is correct but also resolve competition. These top-down processes are linked to attention and higher cognitive functions help select relevant input.
We aim to debate what happens to the irrelevant information in the process of binding. Are irrelevant features simply lost from the system over time, or are they deliberately deleted? Is there any inhibitory process involved in binding? What is the empirical evidence for such a process at the behavioral level? Is such a process active and resource-demanding or relatively passive and automatic? What do neuropsychological studies show? What are the physiological underpinnings of such a process? How is it incorporated in computational models to increase our understanding of the binding process? The idea is to bring together diverse views on ‘Inhibition in Feature Binding’ with the ultimate aim of better understanding the process of binding and invoking informed and insightful future research.