Visual Processing In The Primate Brain

National Eye Institute

Investigators

Linked publications & trials

Abstract

A complex visual scene requires us to inspect different parts in succession in order comprehend the entire scene. This inspection is performed by shifting visual attention to one region of the visual field and then to another until the areas of possible interest have been examined. Accompanying the shifts of attention to different regions of the visual field are rapid or saccadic eye movements that move the eyes quickly from one part of the visual scene to another and bring that part of the scene onto the fovea. The brain mechanisms that generate these saccadic eye movements have been studied in a model of the human visual and oculomotor system, the old world monkey, in this laboratory for over twenty years. The circuit in the brain that controls these movements is therefore known in considerable detail from the visual input into the brain from the retina of the eye through an extensive pathway within the brain and eventually to the muscles that move the eye. How the eye gets to the new target is known at least in outline. In contrast, what brain mechanisms underlie the shifts of attention to one part of the field or the other remain unknown. One hypothesis is that the same mechanisms that underlie the saccade generation underlie the shifts of attention to one part of the visual field; a shift of attention occurs as an integral part of the mechanisms that will lead to the saccade. We tested this hypothesis by determining whether increasing activity in the saccade pathway, without actually producing a saccade would produce an effect similar to that seen with a shift of attention. Using a task in which the monkey has great difficulty seeing a change in the visual scene unless its attention is drawn to the region of the field by a visual cue, we demonstrated that with this shift of attention the monkey could see the change but without the cue it rarely could. We then substituted a brief stimulation in the saccade pathway for the visual cue by stimulating the superior colliculus, a midbrain saccade related structure, to see if we could direct attention to particular regions of the visual field. We presented multiple fields of moving random dots radially positioned around a central fixation point. After a random delay while the monkey fixated, the direction of dot motion in just one of the fields changed on about half the trials. The monkey was rewarded for making a saccade to the target if it changed, and for maintaining fixation otherwise. During some trials, there was a 150ms blank at the time of the possible change during which only the fixation point remained, thus masking the change. We previously learned that a visual cue facilitates detection of changes, even in blanked "change-blindness" trials, presumably due to the directed allocation of attention. By overlapping fields of dots and movement fields in the SC, we now tested if sub-threshold stimulation of the SC facilitates detection of changes at a particular location. We found that the monkey could more easily detect changes in the areas of the visual field corresponding to the activation sites in the SC. We observed an increase in correct saccades to changes in both blanked and non-blanked trials - the proportion of incorrect saccades did not increase. This suggests that the monkey was not just making more saccades to the stimulated site, but was better able to detect changes there. This indicates that activating the SC facilitates detection of visual changes in particular locations, much like directing attention with a visual cue. The experiments lend strong support to the hypothesis that the neuronal basis of shifts of attention to different regions of visual space is closely related to the generation of an eventual eye movement to that same region of space, and opens the possibility of understanding the basic mechanisms in the brain that underlie visual spatial attention.