Properties of the recombination of one-dimensional motion signals into a pattern motion signal

We have examined the human ability to determine the direction of movement of a variety of plaid patterns. The plaids were composed of two orthogonal sine-wave gratings. When the plaid components are of unequal spatial frequency or sometimes of unequal contrast, observers judge the direction of movement incorrectly. In terms of the two-stage model of Adelson and Movshon (1982), these errors may result from either a misjudgment in the perceived speeds of each of the components or a failure in the combination of one-dimensional-component movements into a coherent direction of motion of the two-dimensional plaid pattern, or both. A comparison of the perceived direction of motion of plaids with the relative perceived-speeds of the plaid component gratings suggests that both failures occur, but in different circumstances The relative perceived speed of the plaid components was measured with a spatial and a temporal forced-choice technique, the former leading to larger differences. Our results support the notion that the visual system decomposes a moving plaid into oriented components and subsequently recombines the component motions.     

Perceptual, oculomotor, and neural responses to moving color plaids

Moving plaids constructed from two achromatic gratings of identical luminance contrast are known to yield a percept of coherent pattern motion, as are plaids constructed from two identical chromatic (e.g. isoluminant red/green) gratings. To examine the interactive influences of chromatic and luminance contrast on the integration of visual motion signals, we constructed plaids with gratings that possessed both forms of contrast. We used plaids of two basic types, which differed with respect to the phase relationship between chromatic and luminance modulations (after Kooi et al, 1992 Perception 21 583-598). One plaid type ('symmetric') was made from component gratings that had identical chromatic/luminance phase relationships (e.g. both components were red-bright/green-dark modulation). The second plaid type ('asymmetric') was made from components that had complimentary phase relationships (i.e. one red-bright/green-dark grating and one green-bright/red-dark grating). Human subjects reported that the motion of symmetric plaids was perceptually coherent, while the components of asymmetric plaids failed to cohere. We also recorded eye movements elicited by both types of plaids to determine if they are similarly affected by these image cues for motion coherence. Results demonstrate that, under many conditions, eye movements elicited by perceptually coherent vs noncoherent plaids are distinguishable from one another. To reveal the neural bases of these perceptual and oculomotor phenomena, we also recorded the responses of neurons in the middle temporal visual area (area MT) of macaque visual cortex. Here we found that individual neurons exhibited differential tuning to symmetric vs asymmetric plaids. These neurophysiological results demonstrate that the neural mechanism for motion coherence is sensitive to the phase relationship between chromatic and luminance contrast, a finding which has implications for interactions between 'color' and 'motion' processing streams in the primate visual system.     

Coherence determines speed discrimination

The visual system must determine which elements in a scene to regard as parts of a single object and which to regard as different objects. We can create stimuli that are ambiguous, ie consistent with more than one interpretation, and ask in what situations the stimulus elements are interpreted as part of a single object and when they are interpreted as multiple objects. The ambiguous stimuli in this study were moving plaid patterns--the sum of two drifting gratings with different orientations. Observers may see a rigid coherent plaid object moving in one direction, or may see two gratings moving in different directions sliding over one another. When the gratings have similar contrasts they appear to cohere and only the plaid speed is perceptually available; when the gratings have different contrasts they appear to slide and only the speeds of the gratings are perceived. Coherence thus determines what speed information is passed to higher stages of motion processing. A two-stage model of plaid motion perception is presented which agrees with the model proposed by Adelson and Movshon and extends it, detailing the relationship between coherence and speed discrimination.     

'Squaring' is better at predicting plaid motion than the vector average or intersection of constraints

How do humans combine the velocity information from two moving gratings (plaids) to detect pattern motion direction?-We are still unable to answer this question. The 'intersection of constraints' rule (IOC-Adelson and Movshon, 1982 Nature 300 523-525), and the 'vector average' rule (VA-Wilson et al, 1992 Visual Neuroscience 9 79-97) have both been supported by results in the plaid literature, but could these results be predicted by a 'squaring' nonlinearity that now forms part of several influential spatiotemporal energy models (Wilson et al 1992, loco cit.; Lu and Sperling, 1995 Vision Research 35 2697-2722; Simoncelli and Heeger, 1998 Vision Research 38 743-761)? Spatiotemporal energy in these models predicts directions other than those predicted by standard spatiotemporal energy models and may underlie the results that support the combination rules. The two combination rules and predictions from 'squaring' were tested under identical conditions. In the first three experiments a plaid was randomly presented in one of 45 different orientations, and observers were asked to remember the direction. The stimulus was then replaced by an oriented line indicating the direction predicted by one of the hypotheses. The observer was unaware which hypothesis had generated the line and was asked to make a same/different judgment. Results showed that the 'squaring' hypothesis was better at predicting perceived direction than either the IOC or VA.     

Features derived from first-order motion mechanisms predict anomalies in motion perception

Current dominant hypotheses of how humans detect the movement of patterns assume that the pattern is divided into one-dimensional sinusoidally varying luminance patterns, referred to as gratings (first-order components). The speed of these gratings is independently encoded from predominantly spatial and temporal frequency information, and their direction is encoded from orientation information. This paper addresses the problem of how the individually encoded grating information is combined to give perceived pattern direction, given that real moving objects are generally made up of more than one component. More specifically, further evidence is presented for a combination based on the use of a feature derived from first-order components--'first-order feature hypothesis'. This hypothesis essentially implements a constraint on pattern direction called the intersection of constraints (IOC) proposed by Adelson and Movshon [1982, Nature 300 523-525]. A simulation of the model is used to make three new predictions about a perceived motion reversal reported by Derrington et al (1992, Vision Research 32 699-707); these predictions are tested and found to be consistent with the first-order feature hypothesis.     

Higher-order factors influencing the perception of sliding and coherence of a plaid.

The effect of several new stimulus parameters on the perception of a moving plaid pattern (the sum of two sine-wave gratings) were tested. It was found that: (i) the degree of perceived sliding is strongly influenced by the aperture configuration through which the plaid is viewed; (ii) the chromaticity of the sinusoidal components affects coherence in that more sliding is observed when the plaid components differ in hue, and there is less sliding when they are of the same hue; (iii) equiluminant plaids made of components equal in color almost never show any sliding; and (iv) sliding increases with viewing time. The coherence-sliding percept must therefore be influenced by color, by global interactions, and by adaptation or learning effects, thus suggesting a higher-level influence. These results are most easily modelled by separating the decision to carry out recombination from the process of recombination.     

Coherent plaids are preattentively more than the sum of their parts

We investigated whether plaids activate preattentive mechanisms that are distinct from those activated by their component gratings. Observers searched for a target plaid, the sum of two perpendicular components in a circular window (radius = 0.65°). The target was present on half the trials. On all trials, half of the distractors had the same frequency and orientation as one component of the plaid, and the rest were the same as the other component. The target and the distractors were arrayed evenly on a circle (radius = 2.36°) around fixation. Target and distractor contrasts were randomly perturbed up to ±30%. The following results held for each of the 6 participants tested. (1) When F1 = 2 c/deg and F2 = 5.25 c/deg, response times (RTs) increased significantly when set size (number of distractors plus target, if present) was increased from four to eight. (2) When the spatial frequencies of both plaid components were the same (i.e., both 2 c/deg or both 5.25 c/deg), RTs increased very slightly, if at all, when set size was increased from four to eight. These results suggest the existence of a preattentive, plaid-sensitive mechanism with band-limited input that does not respond to individual grating components.     

Psychophysical evidence for an extrastriate contribution to a pattern-selective motion aftereffect

A stationary vertical test grating appears to drift to the left after adaptation to an inducing grating drifting to the right, this being known as the motion aftereffect (MAE). Pattern-specific motion aftereffects (PSMAEs) induced by superimposed pairs of gratings in which the component gratings drift up and down but the observer sees a single coherent plaid drifting to the right have been investigated. Two experiments are reported in which it is demonstrated that the PSMAE is tuned more to the motion of the pattern than to the orientation and direction of motion of the component gratings. However, when subjects adapt to the component gratings in alternation, aftereffect magnitude is dependent upon the individual grating orientations and motion directions. These results can be interpreted in terms of extrastriate contributions to the PSMAE, possibly arising from the middle temporal area, where some cells, unlike those in striate cortex (V1), are tuned to pattern motion rather than to component motion.     

Two-dimensional tilt illusions induced by orthogonal plaid patterns: effects of plaid motion, orientation, spatial separation, and spatial frequency

Tilt illusions occur when a drifting vertical test grating is surrounded by a drifting plaid pattern composed of orthogonal moving gratings. The angular function of this illusion was measured as the plaid orientation (and therefore its drift direction) varied over a 180 degrees range. This was done when the test and inducing stimuli abutted and had the same spatial frequency, and when the test and inducing stimuli either differed in frequency by an octave, or were spatially separated by a 2 deg blank annulus, or both differed in frequency and were also separated by the annulus (experiments 1-4). The obtained angular function was virtually identical to that obtained previously with the rod and frame effect and other cases involving orthogonal inducing components, with evidence for illusions induced both by real-line components and by virtual axes of symmetry. Although the magnitude of the illusion was very similar in all four experiments, there was evidence to suggest that largest real-line effects occurred in the abutting same-frequency condition, with a pattern of results similar to that obtained previously with the simple one-dimensional tilt illusion. On the other hand, virtual-axis effects were more prominent with gaps between test and inducing stimuli. A fifth, repeated-measures, experiment confirmed this pattern of results. It is suggested that this pattern-induced tilt effect reflects both striate and extrastriate mechanisms and that the apparent influence of spatially distal virtual axes of symmetry upon perceived orientation implies the existence of AND-gate mechanisms, or conjunction detectors, in the orientation domain.     

New motion illusion caused by pictorial motion lines

Motion lines (MLs) are a pictorial technique used to represent object movement in a still picture. This study explored how MLs contribute to motion perception. In Experiment 1, we reported the creation of a motion illusion caused by MLs: random displacements of objects with MLs on each frame were perceived as unidirectional global motion along the pictorial motion direction implied by MLs. In Experiment 2, we showed that the illusory global motion in the peripheral visual field captured the perceived motion direction of random displacement of objects without MLs in the central visual field, and confirmed that the results in Experiment 1 did not stem simply from response bias, but resulted from perceptual processing. In Experiment 3, we showed that the spatial arrangement of orientation information rather than ML length is important for the illusory global motion. Our results indicate that the ML effect is based on perceptual processing rather than response bias, and that comparison of neighboring orientation components may underlie the determination of pictorial motion direction with MLs.     

Tracking without perceiving: a dissociation between eye movements and motion perception

Can people react to objects in their visual field that they do not consciously perceive? We investigated how visual perception and motor action respond to moving objects whose visibility is reduced, and we found a dissociation between motion processing for perception and for action. We compared motion perception and eye movements evoked by two orthogonally drifting gratings, each presented separately to a different eye. The strength of each monocular grating was manipulated by inducing adaptation to one grating prior to the presentation of both gratings. Reflexive eye movements tracked the vector average of both gratings (pattern motion) even though perceptual responses followed one motion direction exclusively (component motion). Observers almost never perceived pattern motion. This dissociation implies the existence of visual-motion signals that guide eye movements in the absence of a corresponding conscious percept.     

Plaid motion rivalry: correlates with binocular rivalry and positive mood state

Recently Hupé and Rubin (2003, Vision Research 43 531- 548) re-introduced the plaid as a form of perceptual rivalry by using two sets of drifting gratings behind a circular aperture to produce quasi-regular perceptual alternations between a coherent moving plaid of diamond-shaped intersections and the two sets of component 'sliding' gratings. We call this phenomenon plaid motion rivalry (PMR), and have compared its temporal dynamics with those of binocular rivalry in a sample of subjects covering a wide range of perceptual alternation rates. In support of the proposal that all rivalries may be mediated by a common switching mechanism, we found a high correlation between alternation rates induced by PMR and binocular rivalry. In keeping with a link discovered between the phase of rivalry and mood, we also found a link between PMR and an individual's mood state that is consistent with suggestions that each opposing phase of rivalry is associated with one or the other hemisphere, with the 'diamonds' phase of PMR linked with the 'positive' left hemisphere.     

Visual perception of motion in depth: Application of a vector model to three-dot motion patterns

The aim of the present study was to identify spatial properties of three-dot motion patterns yielding perceived motion in depth. A proposed vector model analyzed each pattern in terms of common and relative motion components of the moving parts. The dots moved in straight paths in a frontoparallel plane. The Ss reported verbally what they perceived. The common motion did not affect the kénd of perceived event (translation or rotation in depth). Relative motions toward or away from a common point, i.e. concurrent motions, yielded perceived translatory motion in depth. Parallel relative motions toward or away from a common line generally yielded perceived rotation in depth. Complex motion patterns, consisting of concurrent and parallel relative motion components combined, evoked simultaneously perceived translation and rotation in depth under certain phase conditions of the components. Some limitations of the model were discussed and suggestions made to widen its generality.     

Coherence and transparency of moving plaids composed of Fourier and non-Fourier gratings.

We examined the perceptual coherence of two-component moving plaids. The gratings that constituted the plaids were either standard Fourier gratings (F), in which luminance was determined by a drifting sinusoid, or non-Fourier gratings (NF), in which the contrast of a random background was modulated by a drifting sinusoid. These NF gratings are examples of stimuli that generate a compelling percept of motion, even though they fail to elicit a motion signal from motion analyzers based on standard cross-correlation (Chubb & Sperling, 1988). Naive observers viewed three types of stimuli consisting of superpositions of these two components: (1) two standard drifting gratings (F/F), (2) two non-Fourier drifting gratings (NF/NF), and (3) one standard and one non-Fourier drifting grating (F/NF). As expected, the F/F stimulus yielded a compelling percept of coherent motion. The dominant percept of all the observers for the NF/NF stimulus was one of coherent motion, provided that both gratings were visible and of approximately equal contrast. None of the observers reported a dominant percept of coherent motion for the F/NF condition, over a wide range of contrasts for the two grating components and across two varieties of NF gratings. In view of the results of Albright (1992) and Albright and Chaudhuri (1989), that show that single cells in macaque V1 and MT respond to both F and NF motion, one cannot interpret our findings as evidence that F and NF motion are processed independently. Alternative, "higher level" interpretations based on the intrinsically ambiguous nature of the stimuli and physical laws governing the appearance of transparent objects are discussed.     

Coherence and transparency of moving plaids composed of Fourier and non-Fourier gratings

We examined the perceptual coherence of two-component moving plaids. The gratings that constituted the plaids were either standard Fourier gratings (F), in which luminance was determined by a drifting sinusoid, or non-Fourier gratings (NF), in which the contrast of a random background was modulated by a drifting sinusoid. These NF gratings are examples of stimuli that generate a compelling percept of motion, even though they fail to elicit a motion signal from motion analyzers based on standard cross-correlation (Chubb & Sperling, 1988). Naive observers viewed three types of stimuli consisting of superpositions of these two components: (1) two standard drifting gratings (F/F), (2) two non-Fourier drifting gratings (NF/NF), and (3) one standard and one non-Fourier drifting grating (F/NF). As expected, the F/F stimulus yielded a compelling percept of coherent motion. The dominant percept of all the observers for the NF/NF stimulus was one of coherent motion, provided that both gratings were visible and of approximately equal contrast. None of the observers reported a dominant percept of coherent motion for the F/NF condition, over a wide range of contrasts for the two grating components and across two varieties of NF gratings. In view of the results of Albright (1992) and Albright and Chaudhuri (1989), that show that single cells in macaque V1 and MT respond to both F and NF motion, one cannot interpret our findings as evidence that F and NF motion are processed independently. Alternative, “higher level” interpretations based on the intrinsically ambiguous nature of the stimuli and physical laws governing the appearance of transparent objects are discussed.     

A spatial illusion from motion rivalry

A new dynamic visual illusion is reported: contrast reversal of a horizontal and vertical plaid pattern (produced by adding two orthogonal sinusoidal gratings) causes the pattern to appear as an array of lustrous diamonds, cut by sharp lines into a diagonal lattice structure. On the basis of computer simulations it is suggested that the illusion results from rivalrous interaction of motion detectors tuned to opposing directions of motion.     

Local and global mechanisms of one- and two-dimensional orientation illusions

One-dimensional (1-D) orientation illusions induced on a test grating by a tilted and-surrounding 1-D inducing grating have a well-known angular function that exhibits both repulsion and attraction effects. Two-dimensional (2-D) orientation illusions are those induced on a test grating by 2-D image modulation, such as a pair of superimposed inducing gratings at different orientations, usually orthogonal (a plaid). Given the known angular functions induced by the plaid component gratings, two hypotheses were developed that predicted different plaid-induced illusion functions. Hypothesis 1 states that the 1-D component-induced effects simply add linearly; Hypothesis 2 states that there is an additional mechanism that responds to the virtual axes of mirror symmetry of the plaid and adds to the effect. The data of two experiments were consistent with the predictions from the second hypothesis but not the first. Possible neural substrates of mechanisms that extract axes of symmetry are discussed; it is suggested that such global symmetry axes may underlie the perceived orientation of complex shapes.     

Motion over the retina and the motion aftereffect.

The motion aftereffect (MAE) was measured with retinally moving vertical gratings positioned above and below (flanking) a retinally stationary central grating (experiments 1 and 2). Motion over the retina was produced by leftward motion of the flanking gratings relative to the stationary eyes, and by rightward eye or head movements tracking the moving (but retinally stationary) central grating relative to the stationary (but retinally moving) surround gratings. In experiment 1 the motion occurred within a fixed boundary on the screen, and oppositely directed MAEs were produced in the central and flanking gratings with static fixation; but with eye or head tracking MAEs were reported only in the central grating. In experiment 2 motion over the retina was equated for the static and tracking conditions by moving blocks of grating without any dynamic occlusion and disclosure at the boundaries. Both conditions yielded equivalent leftward MAEs of the central grating in the same direction as the prior flanking motion, ie an MAE was consistently produced in the region that had remained retinally stationary. No MAE was recorded in the flanking gratings, even though they moved over the retina during adaptation. When just two gratings were presented, MAEs were produced in both, but in opposite directions (experiments 3 and 4). It is concluded that the MAE is a consequence of adapting signals for the relative motion between elements of a display.     

The sensing of retinal motion

The apparent relative motion of physically stationary objects that frequently occurs as the head is moved in a frontoparallel plane is almost always in the direction expected from the projection into the distal world of the relative motion of the images on the eye. It is hypothesized that this is the result of the perceptual underestimation of the depth between the objects. If a perceptual overestimation of the depth were produced, it was predicted that the apparent relative motion would be in a direction opposite to that expected from the projection of the retinal motions. This prediction was tested using the binocular disparity cue to produce perceptual overestimation of the slant (depth) of a luminous line. In this case, perceived slant was the indicator of perceived depth, and perceived rotation concomitant with the motion of the head was the indicator of perceived relative motion. The results support the prediction and also provide some support for a theoretically derived equation specifying the relation between these two perceptual variables.     

Local and global mechanisms of one- and two-dimensional orientation illusions.

One-dimensional (1-D) orientation illusions induced on a test grating by a tilted and surrounding 1-D inducing grating have a well-known angular function that exhibits both repulsion and attraction effects. Two-dimensional (2-D) orientation illusions are those induced on a test grating by 2-D image modulation, such as a pair of superimposed inducing gratings at different orientations, usually orthogonal (a plaid). Given the known angular functions induced by the plaid component gratings, two hypotheses were developed that predicted different plaid-induced illusion functions. Hypothesis 1 states that the 1-D component-induced effects simply add linearly; Hypothesis 2 states that there is an additional mechanism that responds to the virtual axes of mirror symmetry of the plaid and adds to the effect. The data of two experiments were consistent with the predictions from the second hypothesis but not the first. Possible neural substrates of mechanisms that extract axes of symmetry are discussed; it is suggested that such global symmetry axes may underlie the perceived orientation of complex shapes.