Comparisons of flashILM, transformational apparent motion, and polarized gamma motion indicate these are three independent and separable illusions

Abstract

Illusory line motion (ILM) refers to perceived motion in a bar when it is presented all at once. Explanations for ILM include lowlevel visual accounts, visual attention, and object tracking. These explanations tend to arise from studies using different protocols to induce ILM, based on the assumption that the same illusion is being generated. Using real motion in the same and in the opposite direction as the ILM quantifies the illusions from all protocols as the area between response curves for the left-and rightside inducers. This common measure enables testing of the assumption that two display configurations result in the same illusion. If there is a common underlying cause, an individual who shows a strong illusion in one situation should show a strong illusion in the other, but illusions that arise through different systems should not correlate. This approach has differentiated ILM induced by a flash (flash ILM) from ILM induced by matching the bar to an attribute of the inducing stimuli (transformational apparent motion, TAM). The former is thought to reflect attention, while the latter is thought to reflect object processing. Low-level visual explanations are often offered based on ILM that occurs when the bar is adjacent to only a single inducer (polarized gamma motion, PGM) rather than between two stimuli (flash ILM and TAM). The present study replicates the independence of flash ILM and TAM and shows that neither is related to PGM, suggesting that all three explanations for ILM are warranted and that the debates in the literature are conflating at least three different illusions.

Cancelling Flash Illusory Line Motion by Cancelling the Attentional Gradient and a Consideration of Consciousness

“…In addition, participants’ ILM area from PGM and flash ILM displays are not correlated with each other nor are ILM area measures between PGM and TAM. This pattern is consistent with the separate explanations for flash ILM, TAM, and PGM [12].…”

“…For example, TAM can arise when the boxes and bar match based upon colour or based upon size, and the ILM area measures for these two illusions are highly correlated [6] as both are thought to reflect the tracking of objects over time [3,4]. The ILM area from the TAM illusion, however, is not correlated with the ILM area from flash ILM [6,12], despite variations on the flash ILM paradigm also showing high test-retest reliability on this measure [6,10,11,12].…”

“…The illusory motion signal is thought to arise because exogenous attention is asymmetrically distributed to be focused upon the flashed box, and so is asymmetrically focused at one end of the bar upon its appearance. It is also possible to generate flash ILM in an existing bar by changing its colour upon the offset of the flash [7,12], which is also explained due to the asymmetric distribution of exogenous attention. When the colour change paradigm needs to be distinguished from the onset/offset bar paradigm we will refer to the former as paint ILM and the latter as flash ILM, but the former will be used only when it is necessary to highlight the paradigm difference, otherwise the term flash ILM will be used as paint ILM and flash ILM produce high test-retest reliability in the ILM area measure and neither produces any logical challenges to the common attention based explanation current in the literature.…”

“…If attention is redistributed to be less asymmetric following an instance of flash ILM [20] then a subsequent change should produce weaker flash ILM. Because flash ILM can be examined using the paint ILM paradigm [12,22], if an existing bar changes colour twice following a flash, flash ILM should be reduced for the second colour change as the first instance of flash ILM should have disrupted the asymmetry of the attentional gradient through redistribution. Hamm and Klein’s [20] results found no significant difference in the cuing effect at either end of the bar following flash ILM, suggesting that flash ILM may even drop below detectable levels.…”

“…While NHST p values tend to be described as either significant or non-significant, with some occasionally using phrases such as “marginally significant” (which really should be phrased “marginally non-significant” if it is to be used at all), Bayesian pH0|D values tend to be described using more graded language. We will employ the descriptions suggested by Raftery [25], with the addition of the equivocal description for the range (0.475–0.525) [8,12] if necessary. The Bayesian analysis may be viewed as guidance towards future investigations for hypothetically possible effects that may exist but are below detectable levels.…”