Vision for Action

Vision is perhaps our richest sense. With it we perceive a world of objects defined by attributes that include texture, color and motion. The task of the visual brain is to interpret the world by processing light that enters the eyes and stimulates an array of rod and cone photoreceptors in the retina. How do we perceive this distributed pattern of information as distinct objects that compose a natural scene?  Segregation must occur in order to parse the world into distinct objects in the presence of multiple ones. However, an integrative process also operates on the multiple signals arising from the retina that correspond to a given object so that we can perceive and act upon that object as a whole.



Eye movements are actions that are closely coupled to visual perception, and depend upon the integrative and segregative processes of the visual system. We move the eyes to inspect detailed object features with the fovea, and to follow objects when they move.  Furthermore, eye movements often precede other types of movements such as reaching and locomotion, and therefore information obtained from them can be used by these other systems.  Since the eyes usually follow one object, a multiple-object visual scene must be first segregated into individual ones, following which a decision must be made as to whether to look at one of those objects or not. Finally, the motion information of that object must be integrated so that a single eye velocity signal can be generated from pursuit.  

The decision to initiate a movement to an object in a natural scene can be "bottom-up", i.e., driven by heightened salience imbued by low-level visual properties such as brightness, size or motion. Alternatively, the relevance of an object based on prior experience could contribute to the decision process in a "top-down" fashion. Imagine a situation where one is confronted with two objects, a tree swaying in the wind, and a stationary lion staring directly at them. Here, a conflict arises between the reflex to respond to the tree made salient by motion cues, and prior knowledge about the danger of the lion. In this situation, top-down processing should interact with simple sensory-motor processing and hopefully inhibit a reflexive action to the tree.

Our studies concentrate on the neuronal transformation of signals that arise from stimulation of multiple motion detectors to eventually result in a single eye movement command, and cognitive processes that intervene, specifically we wish to understand: 1) The role of frontal cortex in decision making for movement initiation and inhibition, 2) Integration of motion signals for smooth pursuit of natural objects .

| Home | Vision for Action | Movement Control | Timing | Integration & Segregation |

| People | Life at SKERI | Related Links | Publications | Contact | SK Home |