ABSTRACT

Imagine that you are looking at two women. One has an oval face with striking green eyes framed by long blond hair. The other has a round face with piercing blue eyes framed by wavy red hair. Long before we reach the realms of social psychology, several potential problems present themselves to the visual system. Did those blue eyes go with the blond hair? Was that blond hair wavy? If one woman is Lynn and the other is Anne, which is which? Coherent perception of even a single object requires that the properties of that object be coordinated or bound together. As discussed elsewhere in this issue of Neuron, information about these properties appears to be distributed across many different brain areas. This separation of different types of information about a single object raises the possibility of a ''binding problem.'' This paper will review some of the psychophysical evidence indicating that this is a real problem that is faced and, under most circumstances, solved by the visual system. We will also discuss contrary evidence that suggests that the visual system has no such binding problem. Finally, we will provide a theoretical framework within which to understand these apparently contradictory data. (Other issues like texture grouping and contour completion might be considered to be examples of binding. In this paper, however, we are restricting ourselves to the binding of features to objects.)