Failure of signed chromatic apparent motion with luminance masking

It has been suggested that there are two types of chromatic motion mechanisms: signed chromatic motion, in which correspondence across successive frames is based on chromatic content of image regions, and unsigned chromatic motion based on movement of chromatically-defined borders. We investigate whether signed and unsigned red-green chromatic motion are mediated by a genuinely chromatic mechanism. Direction discrimination of signed and unsigned red-green chromatic motion were measured in the presence of a dynamic luminance masking noise. Increasing the luminance noise contrast systematically impaired signed motion, regardless of contrast and speed. This result suggests that signed red-green chromatic motion is derived from a luminance-based signal, rather than a genuinely chromatic motion mechanism. In the case of unsigned chromatic motion, there is no effect of luminance masking noise, indicating there exists a genuine chromatic mechanism for second-order motion perception.     

Temporal asymmetry in motion masking: a shortening of the temporal impulse response function

Morgan and Chubb observed a striking temporal asymmetry in motion masking (Vis. Res. 39 (1999) 4217). Motion was produced with a two-frame sequence of gratings presented in spatial quadrature phase; the second grating (100 ms) was presented immediately after the first grating (100 ms), with no temporal overlap. The contrast threshold for detecting the direction of motion of the stimulus pair was facilitated when the first grating was of low-contrast and the second grating was of high-contrast, but strong masking occurred when the order was reversed, so the high-contrast grating came first. We replicated this result, but showed that the masking mostly disappeared when the two gratings temporally overlapped only slightly. The high sensitivity to the precise temporal pattern of the stimulus can be explained by a small temporal 'shortening' of the temporal impulse response function (IRF) as stimulus contrast is increased. The IRF is biphasic with a negative inhibitory lobe. When the first grating has high-contrast, its flash response (owing to the shortening of the IRF) may be in a fairly strong negative phase by the time that the positive response to the second, lower-contrast grating has reached appreciable strength--this reduces the magnitude of the motion signal generated by the two flashes and can account for the masking. A shortening of the IRF with increased contrast (a nonlinearity) is supported by psychophysical studies in humans and by recordings of magnocellular retinal ganglion cells in macaque, and the present results bolster this concept.     

Chick eyes compensate for chromatic simulations of hyperopic and myopic defocus: Evidence that the eye uses longitudinal chromatic aberration to guide eye-growth

Longitudinal chromatic aberration (LCA) causes short wavelengths to be focused in front of long wavelengths. This chromatic signal is evidently used to guide ocular accommodation. We asked whether chick eyes exposed to static gratings simulating the chromatic effects of myopic or hyperopic defocus would “compensate” for the simulated defocus.We alternately exposed one eye of each chick to a sine-wave grating (5 or 2 cycle/deg) simulating myopic defocus (“MY defocus”: image focused in front of retina; hence, red contrast higher than blue) and the other eye to a grating of the same spatial frequency simulating hyperopic defocus (“HY defocus”: blue contrast higher than red). The chicks were placed in a drum with one eye covered with one grating, and then switched to another drum with the other grating with the other eye covered. To minimize the effects of altered eye-growth on image contrast, we studied only the earliest responses: first, we measured changes in choroidal thickness 45 min to 1 h after one 15-min episode in the drum, then we measured glycosaminoglycans (GAG) synthesis in sclera and choroid (by the incorporation of labeled sulfate in tissue culture) after a day of four 30-min episodes in the drum.The eyes compensated in the appropriate directions: The choroids of the eyes exposed to the HY simulation showed significantly more thinning (less thickening) over the course of the experiment than the choroids of the eyes exposed to the MY simulation (all groups mean:−17 μm; 5 c/d groups: −24 μm; paired t-test (one-tailed): p = 0.0006). The rate of scleral GAG synthesis in the eye exposed to the HY simulation was significantly greater than in the eye exposed to the MY simulation (HY/MY ratio = 1.20; one sample t-test (one-tailed): p = 0.015). There was no significant interaction between the sign of the simulated defocus and either the spatial frequency or the presence of a +3 D lens used to compensate for the 33 cm distance of the drum.Although previous work has shown that chromatic cues to defocus are not essential for lens-compensation, in that chicks can compensate in monochromatic light, our evidence implies that the eye may be able to infer whether the eye is myopic or hyperopic from the different chromatic contrasts that result from different signs of defocus.     

The detection of motion in chromatic stimuli: first-order and second-order spatial structure

This study provides evidence for the existence of a low-level chromatic motion mechanism and further elucidates the conditions under which its operation becomes measurable in an experimental stimulus. Observers discriminated the direction of motion of amplitude modulated (AM) gratings that were defined by luminance or chromatic variation and masked with spatiotemporally broadband luminance or chromatic noise. The size and retinal location of the stimuli were varied and the effects of broadband noise and grating masks were both compared with the cohort of stimuli. Some significant disparities in the published literature were well explained by the results. In conclusion, evidence for a chromatically sensitive motion mechanism that evades the, detrimental effects of a luminance mask was found only at the fovea and only when the stimulus was small and centrally placed.     

Factors affecting footsteps: contrast can change the apparent speed, amplitude and direction of motion

Contrast can affect the apparent speed of a moving stimulus. Specifically, when a grey square drifts steadily across stationary black and white stripes, it appears to stop and start as its contrast changes--the so-called 'footsteps illusion'. We now show that what matters is the contrast of the leading and trailing edges, not of the lateral edges. The stripes act by altering the stimulus contrast, and are not merely stationary landmarks. Back and forth apparent motion appears smaller in amplitude at low contrasts, even on a spatially uniform (non-striped) surround, and this is a specific motion phenomenon, not a result of misjudging static position. Contrast also affects the perceived direction of a moving stimulus. A vertically jumping grey diamond on a surround of black and white quadrants appears to change its direction of movement depending on the relative contrast of its left-oblique versus right-oblique edges against the surround. Thus, the perceived direction, amplitude and speed of moving objects depend greatly on their luminance contrast against the surround. A model of motion coding is proposed to explain these results.     

Forward displacements of fading objects in motion: the role of transient signals in perceiving position

Visual motion causes mislocalisation phenomena in a variety of experimental paradigms. For many displays objects are perceived as displaced 'forward' in the direction of motion. However, in some cases involving the abrupt stopping or reversal of motion the forward displacements are not observed. We propose that the transient neural signals at the offset of a moving object play a crucial role in accurate localisation. In the present study, we eliminated the transient signals at motion offset by gradually reducing the luminance of the moving object. Our results show that the 'disappearance threshold' for a moving object is lower than the detection threshold for the same object without a motion history. In units of time this manipulation led to a forward displacement of the disappearance point by 175 ms. We propose an explanation of our results in terms of two processes: Forward displacements are caused by internal models predicting positions of moving objects. The usually observed correct localisation of stopping positions, however, is based on transient inputs that retroactively attenuate errors that internal models might otherwise cause. Both processes are geared to reducing localisation errors for moving objects.