McCollough effect

• Dr. Celeste McCollough Howard, University of Dayton Research Institute, Dayton, Ohio, USA (retired)

• Michael A. Webster, Department of Psychology, University of Nevada, Reno

Figure 1: An illustration of the basic McCollough Effect. After viewing red/vertical and green/horizontal stripes, white/vertical stripes appear slightly greenish and white/horizontal stripes appear slightly reddish.

The McCollough effect (ME) is a contingent color aftereffect which can be seen on ordinary ruled white paper. Someone doing word-processing of green text on a black background—as many of us did before full color came to computers—will afterwards see a sheet of ruled paper as slightly pinkish. The pink color disappears when the paper is turned sideways and reappears whenever the lines are horizontal. White chalk lines drawn horizontally on a blackboard will also appear pinkish—but only while the head is upright. The effect is therefore an illusion of color that is contingent on line orientation; it is an aftereffect of looking at colored lines with that same orientation; and its color is complementary to the color of those lines (Fig. 1).

An applet for experiencing the aftereffect can be found on the Project Lite website: McCollough effect Demonstration*. Previous reviews of the ME include Stromeyer (1978), Harris (1980), and Humphrey (1998).

Historical background

The first reported observations of a contingent color aftereffect were incidental to the use of prism spectacles by psychologists investigating how people respond to optical distortion of the light reaching the eyes. Prisms are easier to incorporate in eyewear than the devices used earlier to invert or reverse the visual image, yet the distortions they produce—lateral displacement of objects, curvature of lines parallel to the prism’s thicker base—raise the same questions about the change in perceptual experience (how the world looks) as the wearer wears the prisms for days, gradually regaining the ability to move normally in the visual environment. The first prism-wearers (Gibson, 1933, footnote 4; Kohler, 1951) were paying little attention to the colored fringes they saw along high-contrast vertical borders (like window edges), but they noticed a gradual fading of these colors. Then to their immense surprise, colors re-appeared along those same edges when the experiment was over and the prisms were removed! An orange-red “phantom fringe” blossomed at edges where the prism had produced bluish-violet fringes; where the prism had produced orange-red fringes, the phantom was bluish-violet. Over the next few days, these phantom fringes gradually faded and disappeared. With the discovery of cells in the cat’s visual cortex that had receptive fields selective for lines or edges oriented in particular directions (Hubel & Wiesel, 1962), it became possible to think that “edge-detector cells” selective for vertical lines might play a role in producing phantom fringes. Perhaps edge-detectors have some sort of sensitivity to wavelength. If indeed prism fringes change the chromatic sensitivity of edge-detectors selective for vertical borders, then an adaptation procedure directed toward edge-detectors selective for different orientations might result in different “phantom colors” on horizontal and vertical lines.

Figure 2: Examples of test patterns used to measure the McCollough effect.

Celeste McCollough showed orange/vertical and blue/horizontal gratings to her classes at Oberlin College, and students reported bluish and yellowish aftereffects. Her 1965 Science article encouraged application of an adaptation procedure to other combinations of features in vision and other modalities. The McCollough procedure—displaying two or more color-orientation pairings alternately for some period of time before testing for aftereffects—was varied creatively in later studies in order to advance understanding of the aftereffect’s underlying neural processes. The McCollough procedure initially employed patterns projected on screens; observers simply reported colors seen before and after several minutes of adaptation. Soon projection colorimeters were devised that enabled observers to adjust the stimulus color in a test pattern. Strength of the ME seen on a single-orientation test pattern (Fig. 2, left) could be measured by comparing pre- and post-adaptation settings of an adjacent matching field. Some devices enabled observers to cancel the color difference seen in a two-part test pattern (Fig. 2, right); the amount of color added to each part provided a measure of ME strength. By the mid-1980s, when workers at video display terminals began complaining that their children’s homework paper looked pink, they were simply told (Greenwald & Blake, 1985) that they were seeing the McCollough effect.

Significance

The ME is one of many visual aftereffects (Thompson & Burr, 2009). Color afterimages are a well-known example. Staring at a red square leads to an afterimage; viewed on a light background, it appears as a greenish square. Viewing lines physically tilted clockwise can make a subsequent vertical line appear to be tilted counterclockwise—the tilt aftereffect. In these examples, perceptual changes depend on simple features, color or orientation. The ME was the first clear demonstration of an aftereffect dependent on a combination of such simple features. Contingent aftereffects have subsequently played a significant role in understanding how the visual system represents information. Many combinations of color, form and motion have been tested (see Durgin, 1996), and more recent studies have applied contingent adaptation to study high level visual representations such as face perception.

Simple color afterimages primarily reflect response changes at very early stages of the visual system, including the initial photoreceptors. Yet cells tuned for orientation are not clearly evident until primary visual cortex. Therefore, a color aftereffect contingent on orientation has offered a way of studying how color and shape are represented at cortical levels. At the same time, the aftereffect exhibits properties pointing to relatively early stages in the visual pathway. Adapting to ME patterns with one eye does not alter the appearance of patterns seen with the other eye, even though binocular cells receiving signals from both eyes are a major feature of primary visual cortex. Finding the locus of the effect has drawn enduring interest.

Compared to simple color afterimages, the ME is unusually long lasting. For example, Jones and Holding (1975) found that 10 minutes of induction can lead to an effect that lasts 24 hours. The ME has consequently played a central role in studies ranging from sensory adaptation to learning, encouraging innovative ideas about perceptual plasticity. It has also raised the question whether aftereffects reflect general-purpose adjustments to calibrate visual coding or adjustments that solve specific problems, such as correcting for color fringes introduced by the eye’s optics.

Neural locus of the McCollough effect

Several lines of evidence suggest that the ME involves monocular pathways at an early stage of the visual cortex, probably V1.

• Opposite MEs can be produced on either side of a central fixation point when the observer looks at that point during adaptation and testing (Harris, 1969). Such retinotopic localization is consistent with response changes in mechanisms with the small receptive fields typical of early visual areas.
• Opposite MEs can be produced in the left and right eyes by presenting opposite color/orientation pairings to the two eyes. Columns of cells receiving dominant input from one eye (ocular-dominance columns) are a well-known feature of V1.
• The ME depends on the patterns’ retinal orientation and the chromaticities of the adapting colors, not on "real world coordinates" (Ellis, 1976) or apparent color (Thompson & Latchford, 1986; Webster, Day & Willenberg, 1988).
• Normal MEs lasting at least an hour can be acquired by persons with known impairments of recall and recognition due to temporal lobectomy or Alzheimer’s disease (Savoy & Gabrieli, 1991). Primary visual cortex is relatively intact in these persons.
• Brain-damaged individuals who have lost the ability to distinguish grating orientation may retain substantial color vision. Acquiring an ME that provides a color difference enables them to make this discrimination (Humphrey & Goodale, 1998).
• Unlike many pattern adaptation effects, the ME is relatively unaffected by the observer’s state of attention (Houck & Hoffman, 1986) or by changes in conscious awareness owing to binocular rivalry (White, Petry, Riggs & Miller, 1978).
• The ME can be generated by patterns that alternate at rates of up to 50 Hz, too rapidly for their color differences to be consciously perceived (Vul & MacLeod, 2006). These high flicker rates are consistent with the temporal properties of early cortical areas.

Although these results consistently point to early visual cortex as the site of ME response changes, they obviously do not preclude a role of other levels.

• An fMRI study of MEs to left and right oblique gratings showed some evidence of blood-oxygen level dependent (BOLD) activation in areas V1/V2 but more consistent activation in extrastriate areas (lingual and fusiform gyri) of observers viewing test patterns (Humphrey, James, Gati, Menon & Goodale, 1999).
• Higher level processes that determine the subjective organization of a figure can influence perception of an ME (Uhlarik, Pringle & Brigell, 1977; but see Broerse & Crassini, 1986).
• In a 6-step sequence of adaptation presentations, Vidyasagar (1976) displayed one color/orientation pairing to each eye alone and the opposite pairing to both eyes. The MEs reported with both eyes open had colors opposite those reported with either eye alone. An ME visible only with simultaneous use of the two eyes indicates neural change at a binocular site. This has been taken as evidence of activity beyond V1, possibly in V4 (Grossberg, Huang & Mingolla, 2002).

Interocular transfer (IOT) is found with many pattern aftereffects (tilt aftereffect, spatial frequency shift, motion aftereffect). It is not generally found with color aftereffects contingent on orientation, polarity, spatial frequency, movement, or direction of motion. The ME’s lack of IOT suggests that its underlying neural system draws on largely monocular pathways that may be distinct from systems underlying pattern aftereffects.

• Charles Stromeyer demonstrated interocular transfer of a pattern aftereffect without transfer of color. With both eyes open, his observers first developed an ME contingent on spatial frequency. They then covered the right eye and induced an apparent frequency shift (Blakemore & Sutton, 1969) with the left eye alone. With the left eye observers saw both frequency shift and color on a test pattern. With the right eye alone observers saw the frequency shift, but no color difference. "The site of the interocular transfer of the frequency shift may occur after the site of the McCollough effect" (Stromeyer 1972, p. 731).
• In a dichoptic variation of the McCollough procedure, one eye views the normal alternation of color/pattern combinations while the other eye views a uniformly illuminated area whose color alternates in tandem with the induction pattern. In one variation, this color-only eye always views the color simultaneously presented to the color/pattern eye; in another, it always views the opposite color. In each case, the ME seen with the color-only eye is complementary to the color seen with that eye during presentation of a particular orientation to the other eye (White et al., 1978), suggesting that it is the concurrent spatial orientation that undergoes interocular transfer.

However, anomalous variants of the ME have been reported that show apparent IOT of color information.

• In another dichoptic variation, neither eye receives both pattern and color. One eye gets an alternation of plain red and plain green fields, while the other eye gets an alternation of horizontal and vertical gratings (in black and white). Each eye is then tested separately. The color-only eye sees the expected ME, complementary to the color/orientation pairing during adaptation. The pattern-only eye sees an “anomalous” ME whose color is the same as the color paired with it during adaptation (MacKay & MacKay, 1973, 1975a). Both aftereffects are weaker than those obtained by the same observers with the standard McCollough procedure.
• Sheth and Shimojo (2008) describe an “anti-McCollough effect” induced by alternating adaptation between a colored grating and an achromatic grating at the same orientation. The aftereffect seen in an achromatic test is the same hue as the inducing color (thus a positive rather than negative effect) and shows complete IOT.

A recent study of single cells in cortical areas V1 and V2 of macaque found both monocular and binocular neurons sensitive to color. The authors suggest that color might be processed by two populations – “one that encodes binocular spatial detail at the expense of binocular chromatic detail and another that does the reverse” (Peirce, Solomon, Forte & Lennie, 2008, p. 1). If so, such populations might underlie the rare “anomalous” IOT of color information in the ME.

Accounting for the McCollough effect

What is the nature of the processes underlying the ME? Are they primarily sensory processes, or should we understand the ME as a form of associative learning? This question has arisen repeatedly because, compared to most visual aftereffects, the ME is remarkably long lasting, suggesting that it might involve distinct forms of visual plasticity. Many studies have been devoted to characterizing the growth and decline of the aftereffect.

Figure 3: a) An array of channels tuned to different combinations of color and orientation. Bars represent the preferred color and orientation of individual channels. Stimulation with a vertical or horizontal achromatic grating produces a graded distribution of responses which peaks in channels tuned to the grating orientation and the achromatic color. b) Adaptation to red-vertical and green-horizontal gratings reduces sensitivity in the channels stimulated by the adapting gratings (as shown by the decreased bar size). When the achromatic gratings are again shown after adaptation, these sensitivity changes skew the response distributions so that the peaks occur in channels tuned to colors that are complementary to the adapting colors.

• With adaptation times as long as 150 min, color matches show ME strength increasing with adaptation duration (Riggs, White, & Eimas 1974).
• MacKay and MacKay (1975b) found that ME strength decreases exponentially over time since induction. No decrease occurred during 8 hours of sleep in darkness or after wearing an eyepatch for as much as 25 hours (MacKay & MacKay, 1977).
• Decay of the ME begins when light is admitted to the eye. It follows the same time course in an eye covered with translucent paper as in the uncovered eye (MacKay, 1978).
• In the first efforts to track ME acquisition, color matches that required 30-40 sec were interpolated within a series of 10- or 20-sec adaptation trials. The acquisition function over 20 trials showed a sharp initial rise followed by a more gradual increase, ending at a level dependent on total induction duration, not on number of trials (Skowbo & Rich, 1982; Skowbo & White, 1983).
• Data obtained in a recent study (Vul, Krizay & MacLeod, 2008) suggest that the ME develops at two distinct timescales, a fast one that rises and falls with a time constant of about 30 sec and a slow one that shows no signs of decay within a 5- to 8-minute experiment.

Lack of decay in the absence of stimulation, together with long persistence, inspired accounts of the ME in terms of classical conditioning (e.g., Siegel & Allan, 1992). However, most investigators who view the ME as involving some kind of learning have remained skeptical about regarding it as a conditioned response.

• Conditioning models face difficulty accounting for the fact that MEs cannot be generated for any arbitrary association between color and pattern, such as oppositely oriented curves or angles, or colors paired with images of different faces (Yamashita, Hardy, De Valois, & Webster, 2005).
• Results obtained with complex adapting patterns turn out to be contingent on simple spatial features like local orientation and size (Broerse & O’Shea, 1995; McCollough, 2000).

Alternatively, the ME can be viewed from the perspective of sensory adaptation, a process whereby sensory mechanisms change their responses as a result of exposure to stimuli. Many visual aftereffects are explained by a common framework in which perceptual biases result from selective response changes within a set of multiple channels encoding the stimulus (Braddick, Campbell, & Atkinson, 1978). For example, the tilt aftereffect has been modeled by assuming that perceived orientation corresponds to the modal response in a set of neural channels responsive to stimulus orientation. By reducing sensitivity in channels tuned to the adapting pattern’s orientation, adaptation biases the response distribution away from that orientation, generating an opposite aftereffect. Similarly, color afterimages result from biasing the responses within wavelength-selective channels (the photoreceptors). Under this account, contingent aftereffects arise from biases in channels selective for both orientation and color (Fig. 3).

Multiple channel models frequently assume that each channel adjusts independently, altering the size of its response but not its tuning. In order to account for contingent aftereffects like the ME, Horace Barlow (1990) developed a model based instead on interactions between channels. In this model, mutual inhibition increases between two channels whenever their responses occur at the same time. The inhibition acts to decorrelate the responses, removing redundancies in the neural representation. Barlow’s model predicts contingent color aftereffects without assuming mechanisms intrinsically selective for both color and orientation (“double-duty detectors”). Instead, this joint selectivity is an emergent property of the connections formed through adaptation. Because it is based on synaptic plasticity, this model may overcome a problem that conventional multiple channel models face in explaining the long duration of the ME; sensitivity changes within channels typically appear more transient.

Although its underlying mechanism remains unknown, the ME is widely seen as an example of perceptual plasticity in which the distinction between sensory adaptation and learning is blurred. Whatever its nature, one recurrent proposal regards the mechanism as functioning to make relatively permanent compensation for the color/orientation correlations due to astigmatism and chromatic aberration (Held, 1978; MacKay, 2003). From this viewpoint, the ME’s long persistence simply reflects the lack of color-orientation contingencies in the environment. The aftereffect colors become noticeable only when induced by displays (like monochrome green CRTs or McCollough adaptation patterns) for which the real world gives few opportunities to de-adapt.

The ME and spatial coding

In the mid-1960s, when the ME was first reported, new models of spatial pattern-processing were beginning to emerge. Studies with sinusoidal gratings suggested that analysis of a spatial image proceeds through many different channels, each tuned not only to a particular orientation but also to a particular “size” or spatial frequency (Braddick et al., 1978). By showing that these spatial channels can also be selective for color, the ME extended the concept of channel tuning to combinations of color and pattern.

• With sinusoidal grating patterns, MEs can be made contingent not only on different spatial frequencies but also on different phase relations between harmonics of the same frequency (Stromeyer, Lange, & Ganz, 1973).
• An ME produced with upright and oblique checkerboard patterns is contingent on orientation of the Fourier fundamental, which differs by 45° from the edges’ orientation (May, Agamy, & Matteson, 1978).
• Stromeyer’s 1978 review documents additional color aftereffects dependent on size, movement, and direction of edge contrast (polarity) as well as form aftereffects—tilt, spatial frequency, movement—dependent on color.

These results have generalized the ME beyond orientation; the name now subsumes a whole class of aftereffects characterized by color changes contingent on the spatial or temporal structure of the stimulus. Other procedures have since verified neural responses that are jointly tuned for color and form.

• Adaptation and masking studies have shown spatial interactions selective for color (Bradley, Switkes, & De Valois, 1988; Switkes, Bradley, & De Valois, 1988).
• fMRI adaptation techniques have shown changes in BOLD responses in primary visual cortex selective for color and orientation (Engel, 2005).
• Single unit studies have revealed “double-opponent” cells in V1 that respond to specific combinations of pattern and color (Johnson, Hawken, & Shapley, 2008).
  

  

  Figure 4: Targets formed by a unique conjunction of center and surround orientations are difficult to detect based on the spatial differences (pre-adapt), but after ME adaptation stand out based on the color differences (post-adapt).

In addition, the ME implied that color may play a role in form perception, even though visual acuity is much poorer for color-varying patterns than for patterns that vary in luminance. Beginning in the 1980s, studies examining the visual capacities supported by pure-color (equiluminant) patterns have generally found that information about orientation, size, and position can all be effectively carried by color (Mullen & Kingdom, 1991; De Valois, 2003). Although these studies indicate a close coupling between color and form for some visual tasks, recall that interocular transfer dissociates color aftereffects from spatial pattern aftereffects. Moreover, color and form seem to be processed independently in other tasks.

• In visual search a red bar pops out from green bars regardless of orientation, but a unique conjunction of color and orientation is difficult to detect among other color-orientation combinations (Treisman & Gelade, 1980).
• First-grade children unable to discriminate left from right obliques in a delayed response task can perform the task correctly after acquiring an ME to these patterns (Over, Blackwell, & Axton 1984).
• Acquiring MEs to left and right obliques also helps normal adults on search tasks where the target is "left oblique in center" and distractors are other combinations such as "right oblique in center" or "left oblique in surround" (Fig. 4; Humphrey, Gurnsey, & Fekete, 1991). The ME emerges pre-attentively and transforms this conjunction search (oblique and location) into a feature search (color).

The ME and color coding

The conventional ME is contingent on luminance contrast in a spatial pattern.

• Stronger MEs are obtained with high luminance adaptation patterns and low luminance test patterns (White, 1976).
• Stronger ME colors are obtained with high contrast in both adaptation and test patterns (Stromeyer, 1978).
• Color aftereffects are not observed in black-and-white (or even gray-and-white) test gratings when the adapting gratings are equiluminant (e.g., red/gray and green/gray stripes varying in color but matched for luminance; Stromeyer & Dawson, 1978), nor are phantom fringes seen on edges defined only by color (Kohler, 1951).

From these results it appeared that the ME might be restricted to patterns with luminance contrast. However, subsequent studies (Flanagan, Cavanagh, & Favreau, 1990; Webster & Malkoc, 2000) showed that adapting patterns varying only in color can produce ME-like aftereffects, as long as the equiluminant test pattern contains appropriate color variation. It appears that demonstration of an ME requires similar color/luminance relationships within the adapting and test patterns.

Figure 5: A physiologically inspired color space in which stimuli are represented by an achromatic axis (bright-dark) and two chromatic axes that correspond to variations in the L and M cones or variations in the S cones at constant luminance. The plane on the right shows stimuli defined by the achromatic and L vs. M chromatic axis. Gratings around the circumference show examples of adapting patterns used in ME experiments, while arrows show how the color and luminance variations in these patterns are “oriented” in color space.

At the level of the retina and LGN, color is organized along three cardinal axes: 1) a luminance axis; 2) a roughly reddish to cyan chromatic axis defined by differences in the signals from the L and M cones; and 3) a purple to yellow-green chromatic axis defined by differences between S-cone and L+M-cone signals (Krauskopf, Williams, & Heeley, 1982). Adapting to any axis within this 3D space (Fig. 5) biases color appearance away from the adapting direction and toward the orthogonal direction (Webster & Mollon, 1994) – effectively inducing a “tilt aftereffect” in the perceived directions of color space. However, no hue changes are predicted when the adapting and test patterns lie along orthogonal directions, just as horizontal adapting lines cannot tilt the appearance of vertical test lines. Thus an ME is not predicted to occur when the adapting and test stimuli fall on different cardinal axes (e.g., pure chromatic adapting and pure luminance test), or more generally between any pair of orthogonal directions in the color sphere. Conversely, by this account aftereffects should occur whenever the adapting and test patterns lie along nonorthogonal color-luminance directions.

From this perspective the ME is contingent not only on orientation and color but also on color-luminance relationships. That is, the observer of Figure 1 is adapting not just to a red and vertical grating but to one that is bright red and vertical. Moreover, when the vertical adaptation grating is bright-red/dark-green, the bright bars of a medium-contrast achromatic test grating appear greenish and the dark bars appear reddish (Webster & Malkoc, 2000). This color-luminance contingency suggests a cortical color space paved by multiple channels analogous to the multiple channels for orientation (Fig. 3), with each channel tuned to a different color-luminance direction. Such tuning need not be inherent; as in the Barlow model, channels could become tuned to any direction through adaptation (Atick, Li, & Redlich, 1993; Zaidi & Shapiro, 1993). This interpretation departs from the classical opponent-color theory’s assumption of independent representations of color and luminance, but it is consistent with the chromatic selectivities of cells actually observed in primary visual cortex (Gegenfurtner & Kiper, 2003; Lennie & Movshon, 2005). ME studies have also shown other properties unexpected under opponent-color theory.

• Webster and Malkoc (2000) observed MEs generated with high spatial frequency adapting patterns that can be resolved only by L and M cones—but with color differences detectable only by S cones. Color and pattern information in the ME can thus be carried by different cardinal chromatic axes.
• The ME is stronger for red and green than for the blue and yellow adapting patterns originally used by McCollough (1965). Investigators who varied the chromaticity of adapting patterns typically found a preponderance of pinkish and greenish aftereffects (Stromeyer, 1969).

When color space is defined in terms of relative signals on the geniculate cardinal axes, variations along these axes do not appear pure red/green or blue/yellow, despite the perceptual salience of the opponent primaries. Because the neural basis for blue/yellow opponent colors is not known, the weaker blue/yellow ME is intriguing.

Edge-color and spread-color

Besides their dependence on contour direction, phantom fringes and MEs have two distinctive features in common: long persistence and little or no interocular transfer (IOT). However, phantom fringes and MEs do not look alike, and while both are contingent on orientation, only phantom fringes are manifestly contingent on edge polarity. Anton Hajos (1969), who took up the investigation of phantom fringes in Kohler’s Innsbrück laboratory and continued it after Kohler’s death in 1984, believed that MEs and phantom fringes are fundamentally different.

• Acquisition of phantom fringe aftereffects requires more prolonged adaptation, on the order of days rather than minutes or hours.
• Adaptation to vertical gratings that have a small yellow edge at the left and a small blue edge at the right of each black bar will produce an aftereffect similar to phantom fringe colors within 30 minutes (Held, 1978; Pieper, 1995) (Fig. 6).
• Color fringes have been generated by simulations of chromatic dispersion in which the red, green, and blue components are horizontally displaced from their usual positions (Koh, Lennie, & Williams 1990).

Figure 6: Examples of ME patterns in which the gratings vary in color only along the edges. Gratings at the corners have the same color (red, green, blue or yellow) on each edge. The middle gratings have opposite colors on the left and right side of each bar; red-green on the top, and blue-yellow (characteristic of prism fringes) on the bottom.

Wolfgang Pieper attributed phantom fringes (and the aftereffects he observed with yellow and blue edges) to neural elements with receptive fields for edges with vertical orientation; elements responsible for the ME, he said, have receptive fields for oriented stripes. However, when fine colored lines of the same hue are applied at both sides of the black bars, the aftereffect looks like an ME—the test pattern’s white stripes appear fully tinted with the complementary color, and the tint becomes more saturated as the white stripes are made narrower (Broerse, Vladusich, & O’Shea, 1999). ME saturation has long been known to increase with increasing black:white ratio in the test pattern (Stromeyer 1978), so there is little doubt that this aftereffect is an ME. These same-color fringe data suggest that the conventional ME—illusory color apparently filling the white stripe—is actually spread-color initiated from the edges, where the adapting pattern has left its aftereffect. MEs may consist of two spatial components: “the induction of illusory colour at the local luminance discontinuities in the test grating [and] the spreading of these induced colours away from the edges into uncontoured regions of the surface.” (Broerse et al., 1999, p. 1316). This spreading or “filling-in” effect is well-known as the watercolor illusion. Note that filling-in would not be expected to occur with adapting patterns that simulate prismatic dispersion, because the aftereffects at either side of the white area are not the same color.

Filling-in plays a significant role in the most detailed model yet proposed for the ME (Grossberg et al., 2002). This model includes a Boundary Contour System for detecting and making binocular adjustments to spatial borders defined by luminance contrast. A Feature Contour System then uses a filling-in process to spread color (or texture) within these boundaries. Such a “filling-in” model may also account for the appearance of ME color in certain cases where “top-down” influence on the ME has been claimed, such as the evidence of MEs on subjective transparent structures and on occluded, perceptually-continued edges (Watanabe, Zimmerman, & Cavanagh, 1992; Watanabe, 1995).

Future directions

More than four decades of intensive study have shed light on many features of the ME, yet the underlying processes are still poorly understood. Basic questions remain about the nature and function of these processes and how they relate to other visual phenomena.

• Do the many different color aftereffects contingent on form arise from the same process? Or does the ME involve multiple mechanisms, such as both adaptation to edges and adaptation to patterns; or direct changes in color responses and indirect changes through color spreading?
• Why are “color aftereffects contingent on form” so much more salient than “form aftereffects contingent on color”? Is this difference due to asymmetries in the visual tuning for color (broader) and form (narrower)? Or does a distinct process (like color spreading) account for the salience of color aftereffects?
• Why is the ME stronger for red-green than for blue-yellow adaptation patterns? Perhaps red-green opponency is more prominent than blue-yellow opponency in the mechanisms contributing to the ME. Or perhaps the red-green ME is stronger because observers are already strongly adapted to the blue-yellow bias characterizing most natural color environments (Webster & Mollon, 1997).
• Are contingent aftereffects really a special form of adaptation (driven by the correlations between channels), or are do they reflect an inherent multidimensional structure of visual mechanisms?
• Is the long persistence of the ME a special feature of the visual system’s response, or does it reflect a feature of the visual world—a world that normally does not covary in color and orientation?
• How is the ME related to other recent demonstrations of very long-term adaptation effects on color appearance, which can also be induced by adapting to edges (Belmore & Shevell, 2008; Neitz, Carroll, Yamauchi, Neitz, & Williams, 2002)?
• How is the ME manifested in the natural viewing of scenes that typically contain spatial structure at many orientations and contrast along many color directions? Can MEs occur in complex yet natural images that simultaneously expose the observer to different color biases at different orientations?

References

Atick, J.J., Li, Z. & Redlich, A.N. (1993). What does post-adaptation color appearance reveal about cortical color representation? Neural Computation, 33, 123-129.

Barlow, H. (1990). A theory about the functional role and synaptic mechanism of visual after-effects. In C. Blakemore (Ed.), Vision: Coding and efficiency, (pp. 363-375), Cambridge: Cambridge University Press.

Belmore, S. C., & Shevell, S. K. (2008). Very-long-term chromatic adaptation: test of gain theory and a new method. Vis Neurosci, 25, 411-414.

Blakemore, C. & Sutton, P. (1969). Size adaptation: A new after effect. Science, 166, 245-247.

Braddick, O., Campbell, F. W., & Atkinson, J. (1978). Channels in vision: Basic aspects. In R. Held, H.W. Leibowitz & H.-L. Teuber (Eds.), Handbook of Sensory Physiology (Vol. VIII, pp. 3-38). Berlin: Springer-Verlag. . Bradley, A., Switkes, E., & De Valois, K.K. (1988). Orientation and spatial frequency selectivity of adaptation to color and luminance gratings. Vision Research, 28, 841-856.

Broerse, J. & Crassini, B. (1986). Making ambiguous displays unambiguous: The influence of real colors and colored aftereffects on perceptual alternation. Perception & Psychophysics, 39, 105-116.

Broerse, J. & O'Shea, R.P. (1995). Local and global factors in spatially-contingent coloured aftereffects. Vision Research, 35, 207-226.

Broerse, J., Vladusich, T. & O'Shea, R.P. (1999). Colour at edges and colour spreading in McCollough effects. Vision Research, 39, 1305-1320.

De Valois, K.K. (2003). The role of color in spatial vision. In L.M. Chalupa & J.S. Werner (Eds.), The Visual Neurosciences (Vol 2, pp. 924-935), Cambridge: MIT Press.

Durgin. F.H. (1996). Visual aftereffect of texture density contingent on color of frame. Perception and Psychophysics, 58, 207-223.

Ellis, S.R. (1976). Orientation constancy of the McCollough effect. Perception and Psychophysics, 19, 183-192.

Engel, S.A. (2005). Adaptation of oriented and unoriented color-selective neurons in human visual areas. Neuron, 45, 613-623.

Flanagan, P., Cavanagh, P. & Favreau, O. (1990). Independent orientation-selective mechanisms for the cardinal directions of colour space. Vision Research, 30, 769-778.

Gegenfurtner, K.R. & Kiper, D.C. (2003). Color vision. Annual Review of Neuroscience, 26, 181-206.

Gibson, J.J. (1933). Adaptation, aftereffect and contrast in the perception of curved lines. Journal of Experimental Psychology, 16, 1-31.

Greenwald, M.J. & Blake, R. (1985). Prolonged complementary chromatopsia in users of video display terminals. American Journal of Ophthalmology, 99, 735-736.

Grossberg, S., Hwang, S., & Mingolla, E. (2002). Thalamocortical dynamics of the McCollough effect: boundary-surface alignment through perceptual learning. Vision Research, 42, 1259-1286.

Hajos, A. (1969). Verlauf formspezifischer Farbadaptationen im visuellen System des Menschen. Psychologische Beiträge, XI(1), 95-113.

Harris, C.S. (1969). Retinal localization of orientation-specific color aftereffects. Journal of the Optical Society of America, 59, 504.

Harris, C.S. (1980). Insight or out of sight? Two examples of perceptual plasticity in the human adult. In C.S. Harris (Ed.), Visual Coding and Adaptability, (pp. 69-94), Hillsdale, NJ: Erlbaum.

Held, R. (1980). The rediscovery of adaptability in the visual system: Effects of intrinsic and extrinsic chromatic dispersion. In C.S. Harris (Ed.), Visual Coding and Adaptability, (pp. 69-94), Hillsdale, NJ: Erlbaum.

Houck, M.R. & Hoffman, J.E. (1986). Conjunction of color and form without attention: Evidence from an orientation-contingent color aftereffect. Journal of Experimental Psychology, 12, 186-199.

Hubel, D.H. & Wiesel, T.N. (1962). Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. Journal of Physiology (London), 160, 106-154.

Humphrey, G.K. (1998). The McCollough effect: Misperception and reality. In V. Walsh & J. Kulikowski (Eds.), Perceptual Constancy: Why Things Look as They Do, (pp. 31-68), Cambridge: Cambridge University Press.

Humphrey, G.K. & Goodale, M.A. (1998). Probing unconscious visual processing with the McCollough effect. Consciousness & Cognition, 7, 494-519.

Humphrey, G.K., Gurnsey, R. & Fekete, E. (1991). Rapid discrimination of McCollough effects. Perception, 20, 467-482.

Humphrey, G.K., James, T.W., Gati, J.S., Menon, R.S. & Goodale, M.A. (1999). Perception of the McCollough effect correlates with activity in extrastriate cortex: A functional magnetic resonance imaging study. Psychological Science, 10, 444-448.

Johnson, E.M., Hawken, N.J. & Shapley, R. (2008). The orientation selectivity of color-responsive neurons in macaque V1. Journal of Neuroscience, 28, 8096-8106.

Jones, P. D., & Holding, D. H. (1975). Extremely long-term persistence of the McCollough effect. Journal of Experimental Psychology: Human Perception & Performance, 1, 323–327.

Koh, K., Lennie, P. & Williams, D.R. (1990). Mechanisms of adaptation to chromatic fringes. Investigative Ophthalmology and Visual Science, 31, 326.

Kohler, I. (1951). Über Aufbau und Wandlungen der Wahrnehmungswelt. Österr. Akad. der Wiss. Sitzungsber., Phil-Hist Klasse, 227, 5-118.

Krauskopf, J., Williams, D.R. & Heeley, D.W. (1982). Cardinal directions of color space. Vision Research, 22, 1123-1131.

Lennie, P. & Movshon, J.A. (2005). Coding of color and form in the geniculostriate visual pathway. Journal of the Optical Society of America A, 22, 2013-2033.

MacKay, David J.C. (2003). Deconvolution in humans. Information Theory, Inference, and Learning Algorithms, Cambridge: Cambridge University Press, pp. 552-554.

MacKay, D.M. & MacKay, V. (1973). Orientation-sensitive after-effects of dichoptically presented colour and form. Nature, 242, 477-479.

MacKay, D.M. & MacKay, V. (1975a). Dichoptic induction of McCollough-type effects. Quarterly Journal of Experimental Psychology, 27, 225-233.

MacKay, D.M. & MacKay, V. (1975b). What causes decay of pattern-contingent chromatic aftereffects? Vision Research, 15, 462-464.

MacKay, D.M. & MacKay, V. (1977). Retention of the McCollough effect in darkness: storage or enhanced read-out? Vision Research, 17, 313-315.

MacKay, V. (1978). Associative responses to colour and pattern in the human visual system. Doctoral dissertation, Univ. of Keele.

May, J.G., Agamy, G. & Matteson, H.H. (1978). The range of spatial frequency contingent color aftereffects. Vision Research, 18, 917-921.

McCollough, C. (1965). Color adaptation of edge detectors in the human visual system. Science, 149, 1115-1116.

McCollough, C. (2000). Do McCollough effects provide evidence for global pattern processing? Perception & Psychophysics, 62, 350-362.

Mullen, K.T. & Kingdom, F.A.A. (1991). Color contrast in form perception. In P. Gouras (Ed.), Vision and Visual Dysfunction 6: The Perception of Colour (pp. 198-217). London: MacMillan.

Neitz, J., Carroll, J., Yamauchi, Y., Neitz, M., & Williams, D. R. (2002). Color perception is mediated by a plastic neural mechanism that is adjustable in adults. Neuron, 35, 783-792.

Over, R., Blackwell, A. & Axton, C. (1984). Confusable gratings selectively elicit orientation-contingent colour aftereffects. Australian Journal of Psychology, 36, 337-342.

Peirce, J.W., Solomon, S.G., Forte, J.D. & Lennie, P. (2008). Cortical representation of color is binocular. Journal of Vision, 8, 6: 1-10.

Pieper, W. (1995). Color adaptation of edge-detectors is different from color adaptation of stripe-detectors. Perception, 24 (supplement), 63.

Riggs, L.A., White, K.D. & Eimas, P.E. (1974). Establishment and decay of orientation contingent aftereffects of color. Perception & Psychophysics, 16, 535-542.

Savoy, R.L. & Gabrieli, J.D.E. (1991). Normal McCollough effect in Alzheimer's disease and global amnesia. Perception & Psychophysics, 49, 448-455.

Siegel, S. & Allan, L.G. (1992). Pairings in learning and perception: Pavlovian conditioning and contingent aftereffects. In D. Medin (Ed.), The Psychology of Learning and Motivation (vol. 28, pp. 127-160).

Sheth, B. & Shimojo, S. (2008). Adapting to an aftereffect. Journal of Vision, 29, 1-10.

Skowbo, D. & Rich, J. (1982). Practice does not facilitate acquisition of McCollough effects: Evidence against a learning model. Perception & Psychophysics, 32, 551-554.

Skowbo, D. & White, K. (1983). McCollough effect acquisition depends on duration of exposure to inducing stimuli, not number of stimulus presentations. Perception & Psychophysics, 34, 549-551.

Stromeyer, C.F. (1969). Further studies of the McCollough effect. Perception and Psychophysics, 6, 105-110.

Stromeyer, C.F. (1972). Edge-contingent color after effects: Spatial frequency specificity. Vision Research, 12, 717-733.

Stromeyer, C.F. (1978). Form-color aftereffects in human vision. In R. Held, H.W. Leibowitz & H.-L. Teuber (Eds.), Handbook of Sensory Physiology (Vol. VIII, pp. 98-142). Berlin: Springer-Verlag.

Stromeyer, C.F. & Dawson, B.M. (1978). Form-color aftereffects: Selectivity for local luminance contrast. Perception, 7, 407-415.

Stromeyer, C.F., Lange, A.F & Ganz, L. (1973). Spatial frequency phase effects in human vision. Vision Research, 13, 2345-2360.

Switkes, E., Bradley, A., & De Valois, K. K. (1988). Contrast dependence and mechanisms of masking interactions among chromatic and luminance gratings. Journal of the Optical Society of America A, 5, 1149-1162.

Thompson, P. & Latchford, S. (1986). Colour-contingent aftereffects are really wavelength-contingent. Nature, 320, 525-526.

Thompson, P. & Burr, D. (2009). Visual aftereffects. Current Biology, 19, R11-R14.

Treisman, A., & Gelade, G., 1980. A feature integration theory of attention. Cognitive Psychology, 12, 97-136.

Uhlarik, J., Pringle, R. & Brigell, M. (1977). Color aftereffects contingent on perceptual organization. Perception & Psychophysics 22, 506-510.

Vidyasagar, T.R. (1976). Orientation specific color adaptation at a binocular site. Nature, 261, 39-40.

Vul, E., Krizay, E. & MacLeod, D.I.A. (2008). The McCollough effect reflects permanent and transient adaptation in early visual cortex. Journal of Vision, 8, 4: 1-12.

Vul, E. & MacLeod, D.I.A. (2006). Contingent aftereffects distinguish conscious and preconscious color processing. Nature Neuroscience, 9, 873-874.

Watanabe, T., Zimmerman, G. & Cavanagh, P. (1992). Orientation-contingent color aftereffects mediated by subjective transparent structures. Perception & Psychophysics, 52, 161-166.

Watanabe, T. (1995). Orientation and color processing for partially occluded objects. Vision Research, 35, 647-655.

Webster, M.A. & Malkoc, G. (2000). Color-luminance relationships and the McCollough effect. Perception & Psychophysics, 62, 659-672.

Webster, M.A. & Mollon, J.D. (1994). The influence of contrast adaptation on color appearance. Vision Research, 34, 1993-2020.

Webster, M. A., & Mollon, J. D. (1997). Adaptation and the color statistics of natural images. Vision Res, 37, 3283-3298.

Webster, W.R., Day, R.H. & Willenberg, K. (1988). Orientation-contingent color aftereffects are determined by real color, not induced color. Perception and Psychophysics, 44, 43-49.

White, K.D. (1976). Luminance as a parameter in the establishment and testing of the McCollough effect. Vision Research, 16, 297-302.

White, K.D., Petry, H.M., Riggs, L.A. & Miller, J. (1978). Binocular interactions during establishment of McCollough effects. Vision Research, 18, 1201-1215.

Yamashita, J.A., Hardy, J.L., De Valois, K.K., and Webster, M.A. (2005). Stimulus selectivity of figural aftereffects for faces. Journal of Experimental Psychology: Human Perception and Performance, 31, 420-437.

Zaidi, Q. & Shapiro, A.G. (1993). Adaptive orthogonalization of opponent color signals. Biological Cybernetics, 69, 415-428.

• McCollough effect applet courtesy of Kenneth Brecher, Science and Mathematics Education Center, Boston University, “Project Lite: Light inquiry through experiments,” bu.lite.edu.

Sponsored by: Robert P. O'Shea, Southern Cross University, Australia

Sponsored by: Eugene M. Izhikevich, Editor-in-Chief of Scholarpedia, the peer-reviewed open-access encyclopedia

Sponsored by: Naotsugu Tsuchiya, California Institute of Technology, Pasadena, CA, USA

Reviewed by: Dr. Frank H. Durgin, Psychology, Swarthmore College

Reviewed by: Anonymous

Accepted on: 2009-12-09 22:26:31 GMT