Greenlee M W, Thomas J P
Neurologische Universitätsklinik, Abteilung für Neurophysiologie, Universität Freiburg, Germany.
J Opt Soc Am A. 1993 Mar;10(3):395-404. doi: 10.1364/josaa.10.000395.
Discrimination thresholds for spatial frequency and contrast were measured as a function of (1) the spatial-frequency bandwidth of the stimuli, (2) the reference contrast of the stimuli, and (3) whether the observer knew on which dimension, spatial frequency or contrast, the stimuli would differ. Relative bandwidth was manipulated by a one-dimensional Gaussian window and varied from 0.25 to 2.0 octaves. The results of two different tasks were compared. In a single-judgment task the observer knew on which dimension the stimuli would differ, and discrimination thresholds were measured in a standard two-interval forced-choice staircase procedure. In a dual-judgment task the dimension on which the stimuli differed varied from trial to trial. In this task the observer made two judgments on each trial in a modified two-interval forced-choice procedure: (1) on which dimension the stimuli differed and (2) in which interval the stimulus with the higher spatial frequency or contrast appeared. Thresholds were measured by five interleaved staircases for five reference contrast levels that varied from 0.02 to 0.32. Thresholds for both spatial frequency and contrast increased with increasing stimulus bandwidth and decreased with increasing stimulus contrast. Thresholds were, on average, 1.8 times higher in the dual-judgment task than in the single-judgment task, independent of the stimulus contrast and the bandwidth. A stimulus-uncertainty model is described that accounts for the difference between thresholds in the single- and the dual-judgment tasks.
