Abstract
Background: We have recently shown that, contrary to the widespread assumption, vision is not uniform within the foveola, the tiny (~0.8 deg2 in area) high-acuity region of the fovea where rods are absent and cones most densely packed. Performance in fine spatial discrimination drops just a few arcminutes away from the center of the preferred retinal locus of fixation, an effect normally compensated by microsaccades, which precisely bring this locus on the object of interest. Humans can covertly shift attention away from the center of gaze toward objects in the visual periphery. However, it is unknown whether attention can also be selectively controlled within the foveola. Purpose: To test the hypothesis that, like eye movements, attention can also be selectively allocated at this microscopic scale, and facilitate target detection within the foveola. Methods: We manipulated endogenous (voluntary) attention. Observers were asked to detect the presence of targets, as fast and accurately as possible, at cued and uncued locations, with valid, neutral and invalid cues. Measuring attention shifts within the foveola is challenging because fixational eye movements continually displace the projection of the stimulus on the retina. We circumvented this problem by relying on gaze-contingent display control techniques to selectively stimulate a desired eccentricity within the foveola. We presented very small targets (5 arcminutes) at only 10 arcminutes from the center of gaze. Results: Reaction times were faster when targets were presented at cued than uncued locations. The attentional benefit (~25ms) was highly similar to that observed when stimuli were displayed outside the foveola, at 3o eccentricity. Accuracy did not differ significantly across attended and unattended locations for either eccentricity. Conclusion: These findings indicate that we are capable of selectively allocating attention to objects separated by only 20 arcminutes, revealing a remarkably high resolution of endogenous attention within the foveola.
Meeting abstract presented at VSS 2015