Abstract
Crowding refers to the inability to identify an object when it is presented in clutter. Nandy and Tjan (2009, SfN) proposed that crowding is due to erroneously estimated image statistics acquired during saccadic eye movements, under the yet-to-be extinguished spotlight of attention that initiated the saccade. Changes in the patterns of saccades might therefore result in corresponding changes in the size and shape of crowding zones. An eccentric retinal location can become the preferred retinal locus (PRL) for landing saccades and for fixations as a result of central scotoma. Nandy and Tjan (2010, VSS) predicted that the crowding zone at such a PRL would become smaller and rounder than a normal crowding zone at the same eccentricity, which is elongated along the radial direction relative to the fovea. We tested this prediction using a gaze-contingent display to simulate a 10-degree central scotoma. Normally-sighted subjects performed demanding visual-search and object-tracking tasks for up to 15 hours (spread over weeks) without any instructions on how to use their peripheral vision. Three of four subjects developed a single PRL. For these three subjects, the landing site of the first saccade in the object-tracking task was close to the PRL and away from the fovea, suggesting the saccades had been re-referenced to the PRL. At the end of the experiment, subjects' crowding zones were measured with letter stimuli at their PRLs. We also measured crowding zones of four control subjects who did not participated in the experiment. Consistent with the crowding theory of Nandy and Tjan (2009, 2010), the size of the crowding zone at the PRL was significantly smaller (55%) and its aspect ratio indistinguishable from being round (1.0). These findings support the idea that a change in saccadic eye movements can bring about functional changes in peripheral vision, particularly at the PRL.
National Institutes of Health Grants R01-EY017707, R01-EY016093.