1 INSERM U371, 18 avenue du Doyen Lépine, 69675 Bron Cédex,
² Ingénierie de la Vision, Site GIAT-industries, 3 rue Javelin-Pagnon, BP
505, F42007 Saint-Etienne Cédex 01,
³ Hôpital de Bellevue, Service d’ophtalmologie, Boulevard Pasteur, F 42055
Saint Etienne Cédex 02,
France.

Purpose.

To analyze the influence of color contrast on visual search behavior with visual field loss.

Method.

Subjects: 12 observers (mean age = 75 ± 6.8 years; mean far visual acuity = 0.7 logMAR) with exudative age-related macular degeneration, without cataract, and 2 control groups: 9 young (mean age = 32 ± 8.5 years) and 16 old observers (mean age = 66 ± 8.1 years) with normal vision (0 and 0.05 mean logMAR, respectively). A display of 2 or 8 disks, each measuring 2.7 deg diameter and located on a virtual circle of 12 deg diameter, was presented on each trial. On 50% of the trials, one disk (the target) was in a different color than the other(s) (the distractors). Four color axes (achromatic, protan, deutan or tritan confusion axes) and three levels of contrast (0.25, 0.50 and 0.75 of the maximum available on each axis) were tested. For each trial, the color difference was chosen randomly from one axis and at one contrast. Reaction time was measured for the detection of the presence of a target.

Results.

An ANOVA showed an effect of chromatic axis (reaction times were shorter on the achromatic axis than along the others) and an effect of contrast (performance was poorest for the lowest level of contrast). The most surprising result was that reaction times decrease with an increase in number of disks. Though not significant, a similar tendency was noted throughout the data in the low vision group. As the 8-disks stimuli formed a circle, either proximity or configuration effects might explain this unexpected result. The proximity effect would predict that reaction times would increase with the separation of the 2 disks. This hypothesis does account for the performance for the observers with central visual loss but not for the 2 normal groups.

Conclusion.

These results show the importance of central vision in the visual search behavior of a pattern even when the pattern can be visualized in peripheral vision.