Spatial-frequency difference thresholds for central and peripheral viewing.

Abstract

We measured spatial-frequency difference thresholds with a two-alternative forced-choice technique. On each trial a given standard frequency was presented in one temporal interval and a higher or lower frequency in the other. The subject indicated which interval contained the higher frequency of the two presented in that trial. The variable frequencies were chosen using a decision rule combined with a double-random-alternating staircase that converged on stimulus values that produced 70.7% correct choices. The proportional difference in frequency for criterion responding was computed from the mean of all stimulus values presented in a block of trials. This fraction was plotted as a function of standard spatial frequency. The resulting function shows local maxima and minima that are greater than the standard errors of the means across blocks of trials at a given standard. Functions obtained at 10 deg off the fovea showed a greater range of values than those obtained with central viewing. The results are consistent with the notion that there is a limited number of spatial-frequency channels whose bandwidths are narrow relative to their separation in the spatial-frequency spectrum.