Crowding in central and eccentric vision: the effects of contour interaction and attention

To quantify the crowding effect with eccentric stimuli and to determine the relative contribution of neural interaction and attention to crowding in central and eccentric vision.

Monocular visual acuity was measured with computer-generated tumbling E and Landolt C targets presented centrally or at 2 degrees in the right visual field. Crowding distractors were designed to cause increasing contour interaction or increasing need for focused attention. A four-alternative forced-choice method of constant stimuli was used. In experiment 2 the distance between the target and the distractors was varied.

Data are presented in terms of normalized visual acuity. Crowding in central vision was minimal, mainly caused by contour interaction, and did not occur with distractors more than four stroke widths distant. Crowding in eccentric vision was far greater in magnitude and extent (occurring for distractors as far as 16 stroke widths distant) and was caused by contour interaction and attentional factors.

The results indicate that eccentric vision differs quantitatively and qualitatively from central vision. The extent of contour interaction effects are consistent with the proposed size of cortical processing zones. The results are discussed with reference to current theories of preattentive and attentive tasks and with reference to subjects with low vision due to central scotoma who use an eccentric retinal locus for fixation.