Abstract
Discrimination performance has become an important proxy for the analysis of visuospatial attention. In a typical paradigm, test stimuli such as characters or oriented Gabors are briefly presented at various locations in the visual field. One potential problem arising here is that these stimuli themselves constitute visual objects that may structure the visual field and thus affect what they are intended to measure, the spatial distribution of attention. We developed a novel full-field stimulus composed of orientation-filtered dynamic pink noise that allows to determine the spatio-temporal distribution of attention across the visual field, without the presence of object-like visual structures. As a remarkable property of this stimulus, we demonstrate that local discrimination performance is largely independent of visual eccentricity. This allows to directly compare attentional performance at foveal and peripheral locations. We used this stimulus to analyze the distribution of spatial attention before saccadic eye movements, and to study the effect of the presence or absence of a saccade target. Participants directed saccades according to a central cue either towards a target object or into an unstructured visual field, while simultaneously discriminating the tilt angle of an orientation filtered noise patch embedded in full screen unfiltered noise. The discrimination signal occurred briefly on a continuous range between fixation and saccade target before eye movement onset. Results show that saccades are preceded by shifts of attention even if they are directed into an unstructured visual field. The presaccadic attention shift, however, is much more focused when the saccade is directed to a target object, confirming the hypothesis that the presence of objects molds the distribution of visual attention. Results also demonstrate that the deployment of attention towards the saccade landing position is accompanied by a removal of processing resources from fixation prior to eye movement onset.
Meeting abstract presented at VSS 2018