Abstract
Studies that have explored our sensitivity to natural image properties, in particular to the amplitude spectrum slope, and have demonstrated that we are more sensitive to a specific range of the spectrum (1–1.3), while insensitive to steep (>1.6) or shallow (<0.7) slopes. In a recent study (Johnson et al, submitted), where we assessed the perception of the amplitude spectrum, stimuli presented without a surround did not show increased sensitivity to values between 1 and 1.3. This lack of change in the thresholds could be due to the influence of the mask, therefore, we assessed the effects of masks on the amplitude spectrum discrimination tasks. The physical properties of masks, which are necessary to prevent direct comparison of stimuli, can have complex interactions with discrimination, changing individual sensitivity and improperly reflecting the visual systems' ability to discriminate changes. To address this issue, we introduced saccades - which also generate a suppression mechanism - as a potential replacement for physical masks when measuring sensitivity to image statistics. Retinal blurring, caused by saccade, can be used to remove stimuli from the retina. We measured discrimination for 5 amplitude spectrum values (0.4–1.6) with 5 different types of masks: no-mask, random noise, amplitude slope values of 1.0 and 2.0, and finally an adaptation of the trans-saccadic masking procedure (Irwin & Zelinsky, 1992). The results showed that thresholds are lowest when no mask was used (0.13–0.20), and when the stimulus and mask values where similar in amplitude, discrimination was facilitated and thresholds reduced (0.15, 0.11, 0.19). In addition, thresholds for the trans-saccadic mask were between the no-mask and physical masks values (0.08–0.16). This finding suggests that trans-saccadic masking is capable of effectively masking the stimuli, and may reflect a more ecologically valid sensitivity to discriminating changes in the amplitude spectrum slope.