Abstract
Recent work suggests that, for some tasks, selection among multiple narrow-band SF channels is determined ‘bottom-up’ by stimulus size (e.g. Majaj et al., 2002, V. Res., 42, 1165–1184). In contrast, work on spatial scale processing suggests that ‘top-down’ factors can determine the selection of spatial scale (e.g. Schyns et al., 2002, Psych. Sci., 13, 402–409). We have been exploring whether such effects result from top-down attentional modulation of early SF processing. When an observer was cued top-down to detect a sinusoidal grating presented at one SF their detection of an unexpected SF was impaired compared with when the same SF was expected. Further, the degree of impairment scaled with the difference between the expected and presented SF (Sowden et al., 2001, Perception, 30, 91a). Here we further explored tuning of ‘expectancy’ effects. On the first and last sessions, using ZEST, we measured 5 observers’ ‘baseline’ contrast sensitivity to sinusoidal gratings presented, in single frequency blocks, at one of 18 different SF's spanning an 8-octave range: 2 ‘primary’ SFs (0.5 & 8 cpd) and 16 ‘test’ SFs (0.25, 0.5, 1 or 2 octaves above or below the cued primary). During 8 intervening ‘cueing’ sessions, each trial, a top-down cue signalled the observer to expect one of the two primary SFs, with a different cue for each. However, on 25% of trials one of the test SFs was presented. Compared to baseline, observers' detection of the test SFs was relatively impaired. The degree of impairment increased as a function of difference in SF from the cued primary. The pattern of this tuning resembled the SF tuning seen in early visual analysis. Thus, our findings indicate that when explicit cues are used endogenous factors can induce attentional modulation of narrow-band SF channels in early vision. Further they suggest that the task dependent selection of spatial scale could similarly involve attention to diagnostic SF channels.