Abstract
The stimulus-driven contribution to the allocation of attention is bottom-up attention. Although bottom-up selection is typically quick and potent, it remains unclear how bottom-up attention will be distributed when multiple salient stimuli are presented simultaneously and whether this distribution depends on awareness. Here we performed psychophysical and functional magnetic resonance imaging (fMRI) experiments to examine these issues. In our experiments, each texture stimulus had a regular Manhattan grid of 13 × 37 low luminance bars, presented in the lower visual field on a dark screen. All bars were identically oriented except for the foreground region of 2 × 2 bars with another orientation. There were four different foregrounds with 0°, 25°, 50°, and 90° orientation contrasts between the foreground bars and the background bars. In each texture stimulus, a pair of foregrounds was centered in the lower left and lower right quadrants at 5.83° eccentricity, one of which was 90° foreground. Thus, there were three possible texture stimuli: 90° + 0°, 90° + 25°, and 90° + 50° in our experiments. Low- and high-contrast masks, which had the same grid as the texture stimuli, rendered the whole stimulus visible (Experiment 1) and invisible (Experiment 2, confirmed by a forced-choice test) to subjects, respectively. In the psychophysical experiment, the Posner cueing paradigm was adopted to measure the spatial cueing effect of 90° foreground on an orientation discrimination task and thus the other foreground (i.e., 0°, 25°, and 50°) served as the distractor. We found that, when orientation contrast of the distractor increased, both the cueing effect and the fMRI BOLD signal in areas V1–V4 of 90° foreground fell off monotonically in Experiment 1 (the “Gradient” mechanism), but was constant in Experiment 2 (the “Winner-take-all” mechanism). These findings indicate an awareness-dependent distribution of visual bottom-up attention in human visual areas.