Amplitude spectra of line-motion stimuli

Does Allocation of Attention Influence Relative Velocity and Strength of Illusory Line Motion?

In illusory line motion, presentation of a cue is followed by presentation of a nearby stationary line, and the line is perceived to “unfold,” “expand,” or “extend” away from the cue. Effects of the allocation of attention regarding where the cue or the line would be presented were measured in three experiments, and ratings of relative velocity and relative strength of illusory motion were collected. Findings included (a) relative velocity and relative strength decreased with increases in SOA from 50 to 450 ms, (b) relative velocity and relative strength were not influenced by whether illusory motion moved from one end of the line to the other or from both ends toward the middle of the line, (c) increased uncertainty regarding where the line would appear did not influence relative velocity or relative strength, and (d) valid pre-cues regarding the location of a cue resulted in faster relative velocity than did invalid pre-cues, but pre-cue validity did not influence relative strength. Implications of these findings for the relationship of such illusory motion and attention (e.g., divided attention, shifts in attended location) are considered.

Flash-induced forward and reverse illusory line motion in offset bars

Illusory line motion (ILM) refers to perception of motion in a bar that onsets or offsets all at once. When the bar onsets or offsets between two boxes after one of the boxes flashes, the bar appears to shoot out of the flashed box (_flashILM). If the bar offsets during the flash, it appears to contract into the flashed box (reverse ILM; rILM). Onset bars do not show rILM. Moreover, rILM and _flashILM are not correlated, indicating they are two different illusions. To date, rILM has only been studied using a 50-ms flash where the bar offsets 16.7 ms after flash onset. It is not clear if rILM is due to the 16.7-ms flash-bar-removal stimulus onset asynchrony (SOA) or due to the flash offsetting after the bar. The current studies explore these parameters to better understand the conditions that lead to rILM. The results suggest that _flashILM is sensitive to the temporal interval between flash onset and bar offset, while rILM appears to arise when the flash offsets after the bar has been removed regardless of the temporal interval between flash onset and bar removal. These results are consistent with _flashILM reflecting visual exogenous attention while rILM may reflect the low-level spreading of subthreshold activation radiating from the flashed box. The findings are incorporated into the recent work that suggests that the literature concerning ILM is possibly conflating a number of different illusions of line motion, including polarized gamma motion (PGM), transformational apparent motion (TAM), and exogenous attention induced motion (_flashILM).

A few observations on metacontrast stimuli

Metacontrast is a form of visual masking in which the target and mask are non-overlapping. In metacontrast, the masking effect is typically largest when the mask is presented some time after the target. This is known as Type-B masking. The present report examines to what extent Type-B metacontrast masking can be explained based on the stimuli involved. The assumption is made that the visibility of the target is, at least in part, determined by the correlation between the amplitude spectrum of the target-and-mask combination and that of the target alone. It is found that the correlation is higher when the stimuli are presented at the same time relative to when they are presented at different times. This relationship follows from the stimuli alone. Thus, one would expect the masking to be weakest when the two stimuli are simultaneous. Type-B correlation functions, in which the largest reductions occur only when the mask is presented after the target, can be obtained by further assuming a temporal integration window with a rapid onset and a shallow decline. In agreement with psychophysical masking studies; the analyses yield functions that are most similar to Type-B masking for moderate mask intensities and become less Type-B like at higher mask intensities. The effects of dark adaptation and spatial separation of target and mask are also modeled.

Subjective criteria and illusions in visual testing: some methodological limitations

It is argued that illusions cannot generally be investigated with criterion-independent methods. This limits the value of the data obtained from them. This is particularly important when the results are compared between groups of subjects, for example, between dyslexic readers and controls, since it is possible that the differences between the groups reflect differences with regard to criteria rather than real perceptual differences.

Remapping of the line motion illusion across eye movements

Although motion processing in the brain has been classically studied in terms of retinotopically defined receptive fields, recent evidence suggests that motion perception can occur in a spatiotopic reference frame. We investigated the underlying mechanisms of spatiotopic motion perception by examining the role of saccade metrics as well as the capacity of trans-saccadic motion. To this end, we used the line motion illusion (LMI), in which a straight line briefly shown after a high contrast stimulus (inducer) is perceived as expanding away from the inducer position. This illusion provides an interesting test of spatiotopic motion because the neural correlates of this phenomenon have been found early in the visual cortex and the effect does not require focused attention. We measured the strength of LMI both with stable fixation and when participants were asked to perform a 10° saccade during the blank ISI between the inducer and the line. A strong motion illusion was found across saccades in spatiotopic coordinates. When the inducer was presented near in time to the saccade cue, saccadic latencies were longer, saccade amplitudes were shorter, and the strength of reported LMI was consistently reduced. We also measured the capacity of the trans-saccadic LMI by varying the number of inducers. In contrast to a visual-spatial memory task, we found that the LMI was largely eliminated by saccades when two or more inducers were displayed. Together, these results suggest that motion perceived in non-retinotopic coordinates depends on an active, saccade-dependent remapping process with a strictly limited capacity.