Abstract
Immediately prior to an eye movement, spatial information about potentially relevant visual objects is updated to compensate for the retinal displacement caused by the saccade. Although much is known about the processes involved in such updating, it remains unclear whether predictive remapping preserves an object’s visual features across saccades. To investigate whether featural information of a probe object is updated during predictive remapping, we used a visual crowding paradigm. Observers executed a saccade to identify a probe letter presented at a known location in the periphery and at various intervals prior to the saccade. Flankers presented in the opposite visual field to the probe, but at screen positions that flanked the predicted post-saccadic location of the probe (the probe’s "remapped location") reduced observers’ ability to report the identity of the probe. This decrease was greater when probes and flankers shared elementary visual features compared with probes and flankers that were featurally distinct. Furthermore, probe identification was poorer when flankers appeared within the critical distance (half the eccentricity) of the probe’s remapped location, than when flankers appeared beyond this distance. Finally, the pre-saccadic time course of this "remapped crowding" effect closely matched the neurophysiological time course of activity in neurons that predictively remap spatial information, previously reported in single-cell studies. Our findings reveal a form of non-retinotopic crowding, in which visual features from different visual fields are integrated due to predictive remapping. Remapped crowding is consistent with the notion that predictive remapping not only updates location information, but also preserves an object’s featural information across saccades.
Meeting abstract presented at VSS 2012