Abstract
Purpose. Human visual system is sensitive to both luminance (the first-order) modulations and contrast (the second-order) modulations in an image. A linear-nonlinear-linear (LNL) model is commonly used to explain the visual process of the 2nd-order stimuli. Here we used pattern masking paradigm to compare the 1st-order and the 2nd-order visual mechanisms and to characterize the nonlinear properties of their underlying mechanisms. Methods. The stimuli were a high frequency horizontal grating (8 cyc/d) either added to (1st-order stimuli) or multiplied with (2nd-order stimuli) a vertical low frequency (2 cyc/d) Gabor function. The discrimination threshold of the target was measured with pedestals whose spatial properties as that of the target except contrast (1st-order pedestal) or modulation depth (2nd-order pedestal) of either the low or the high frequency components. Results. The threshold function showed a typical dipper shape for both the 1st- and the 2nd-order stimuli: the threshold first decreased (facilitation) and then increased (suppression) with pedestal contrast or modulation depth. The results for the 1st-order stimuli were well explained by divisive inhibition model in which the facilitatory input was divided by the sum of broadband inhibitory inputs. The results for the 2nd-order stimuli were also well explained by a modified that operated on modulation depth rather than contrast in the input images. Conclusion. Our result suggests that divisive inhibition is required to explain visual discrimination in both the first- and the second-order patterns. However, the source and the nonlinearity of the divisive inhibition may be different for these two types of patterns.
Supported by NSC 96-2413-H-002-006-MY3 and NSC 097-2811-H-002-005- to CCC.