Abstract

In many discrimination tasks, performance becomes independent of contrast at very low contrast levels. In this study, we examine performance in a hyperacuity discrimination task (bisection) as a function of contrast level. The stimuli were three parallel vertical lines superimposed on a 34-cd/m2 uniform field. On any given trial, the center line was offset either to the right or to the left by a single fixed distance. The offset was chosen to yield a d' of~1.7 at the highest available contrast level. Observers rated, on a 6-pt. scale, their certainty that the line was offset to the left (a rating of 1) or to the right (a rating of 6). The lines were always brighter than the background and could appear at one of eight contrasts spanning the range from threshold to ~80 %. In any block of 100 trials, only a single contrast value was used. Each day, four contrast levels were run in counterbalanced order. The z transform of the area under the ROC curve was used as a measure of performance and plotted as a function of contrast level. The results provide some information about which spatial frequency components are important in determining performance in the bisection hyperacuity task.

© 1985 Optical Society of America