Abstract

In most spatial hyperacuity tasks, a measure of relative position can be computed, when superior precision for just one spatial coordinate is available. We have designed a 2-D nonalignment task wherein subjects have to localize a dot relative to the virtual intersection of vertical and horizontal directions defined by two line segments. This task can be executed if and only if the subject extracts both spatial coordinates of the dot with hyperacuity precision. The lines were 15 min of arc long and separated from the dot by gaps of 8 min of arc. A 4AFC measurement paradigm (top-left; bottom-right, etc.) was employed. Short stimulus presentation time (200 ms) resulted in 2-D nonalignment thresholds exceeding 25 s of arc, clearly above the hyperacuity range. The thresholds were significantly improved for long stimulus duration (2000 ms), when a sequential processing strategy was presumably employed. The same results were obtained by employing a different, 2AFC (staircase) measurement paradigm. In a control experiment, Vernier acuity did not present significant improvement with presentation time. We conclude that spatial hyperacuity tasks have access to a limited processing capacity of the visual system. The results further emphasize the importance of the comparison process previously concluded from perturbation experiments.¹

© 1985 Optical Society of America