Colored aftereffects contingent on patterns generated by Lie transformation groups

The McCollough effect and facial emotion discrimination in patients with schizophrenia and their unaffected relatives

Abnormalities in visual processing have been found consistently in schizophrenia patients, including deficits in early visual processing, perceptual organization, and facial emotion recognition. There is however no consensus as to whether these abnormalities represent heritable illness traits and what their contribution is to psychopathology. Fifty patients with schizophrenia, 61 of their first-degree healthy relatives, and 50 psychiatrically healthy volunteers were tested with regard to facial affect (FA) discrimination and susceptibility to develop the color-contingent illusion [the McCollough Effect (ME)]. Both patients and relatives demonstrated significantly lower accuracy in FA discrimination compared with controls. There was also a significant effect of familiality: Participants from the same families had more similar accuracy scores than those who belonged to different families. Experiments with the ME showed that schizophrenia patients required longer time to develop the illusion than relatives and controls, which indicated poor visual adaptation in schizophrenia. Relatives were marginally slower than controls. There was no significant association between the measures of FA discrimination accuracy and ME in any of the participant groups. Facial emotion discrimination was associated with the degree of interpersonal problems, as measured by the Schizotypal Personality Questionnaire in relatives and healthy volunteers, whereas the ME was associated with the perceptual-cognitive symptoms of schizotypy and positive symptoms of schizophrenia. Our results support the heritability of FA discrimination deficits as a trait and indicate visual adaptation abnormalities in schizophrenia, which are symptom related.

Color constancy and the functional significance of McCollough effects

A central problem in visual perception concerns how humans perceive stable and uniform object colors despite variable lighting conditions (i.e. color constancy). One solution is to 'discount' variations in lighting across object surfaces by encoding color contrasts, and utilize this information to 'fill in' properties of the entire object surface. Implicit in this solution is the caveat that the color contrasts defining object boundaries must be distinguished from the spurious color fringes that occur naturally along luminance-defined edges in the retinal image (i.e. optical chromatic aberration). In the present paper, we propose that the neural machinery underlying color constancy is complemented by an 'error-correction' procedure which compensates for chromatic aberration, and suggest that error-correction may be linked functionally to the experimentally induced illusory colored aftereffects known as McCollough effects (MEs). To test these proposals, we develop a neural network model which incorporates many of the receptive-field (RF) profiles of neurons in primate color vision. The model is composed of two parallel processing streams which encode complementary sets of stimulus features: one stream encodes color contrasts to facilitate filling-in and color constancy; the other stream selectively encodes (spurious) color fringes at luminance boundaries, and learns to inhibit the filling-in of these colors within the first stream. Computer simulations of the model illustrate how complementary color-spatial interactions between error-correction and filling-in operations (a) facilitate color constancy, (b) reveal functional links between color constancy and the ME, and (c) reconcile previously reported anomalies in the local (edge) and global (spreading) properties of the ME. We discuss the broader implications of these findings by considering the complementary functional roles performed by RFs mediating color-spatial interactions in the primate visual system.

Do McCollough effects provide evidence for global pattern processing?

Contingent color aftereffects (CAEs, or McCollough effects) were induced using two pairs of orthogonally related patterns (horizontal/vertical and concentric/radial) to determine whether the CAEs of the four patterns are independent. Tests using composite test patterns (like those employed by Emerson, Humphrey, & Dodwell, 1985) suggested independent aftereffects. However, tests using unitary patterns indicated additive or competing effects of the four patterns in regions where line orientations were similar, and tests isolating such regions showed clear interactions between the pattern aftereffects. The results fail to support the claim that global (rather than local) features of the patterns control these CAEs.

The Effects of Adapting to Complex Motions: Position Invariance and Tuning to Spiral Motions

Recent neurophysiological evidence (e.g., Graziano, Andersen, & Snowden, 1994) suggested that some cells in the medial superior temporal area (MST) of the Old World monkey are sensitive to complex motions such as those brought about by a surface moving in depth or rotating. Two important findings were that these cells show position invariance (i.e., their preferred stimulus does not change across the receptive field), and that some cells were selective for “spiralling” stimuli rather than pure rotations or pure expansion/contractions. This paper attempts to provide evidence for similar processes in the human visual system by employing the technique of selective adaptation. We have simulated surfaces undergoing a motion in depth (div) or a rotation (curl), but have removed any cues that are not related to global motion. After adapting to a large pattern undergoing, say, an expansion, an aftereffect that contained an element of contraction could be elicited by placing small test patterns anywhere in the adapted area. This suggests that the global structure of the motion field must have been encoded as well as the local motion. Likewise thresholds for detecting motions similar to the adapting motion were elevated across the adapted area, while thresholds for other motions were not. Hence the effects of adaptation are both selective and show a degree of position invariance. Adaptation to pure div or pure curl stimuli was compared with adaptation to spiralling stimuli. Threshold elevation was always selective for the adapting motion and the shape and broadness of tuning did not vary. In simulations we could not reproduce our results using a model that had only div and curl detectors, but we could reproduce them if we allowed for detectors tuned for a broad range of spiral pitches. Our results suggest that humans encode the complex motion of surfaces by detectors tuned to many different types of motion and that the detectors are invariant across space in their properties.

Visual aftereffect of texture density contigent on color of frame

An aftereffect of perceived texture density contingent on the color of a surrounding region is reported. In a series of experiments, participants were adapted, with fixation, to stimuli in which the relative density of two achromatic texture regions was perfectly correlated with the color presented in a surrounding region. Following adaptation, the perceived relative density of the two regions was contingent on the color of the surrounding region or of the texture elements themselves. For example, if high density on the left was correlated with a blue surround during adaptation (and high density on the right with a yellow surround), then in order for the left and right textures to appear equal in the assessment phase, denser texture was required on the left in the presence of a blue surround (and denser texture on the right in the context of a yellow surround). Contingent aftereffects were found (1) with black-and-white scatter-dot textures, (2) with luminance-balanced textures, and (3) when the texture elements, rather than the surrounds, were colored during assessment. Effect size was decreased when the elements themselves were colored, but also when spatial subportions of the surround were used for the presentation of color. The effect may be mediated by retinal color spreading (Pöppel, 1986) and appears consistent with a local associative account of contingent aftereffects, such as Barlow's (1990) model of modifiable inhibition.

Constraints on perceptual learning: objects and dimensions

The article addresses two questions about perceptual learning: What are the circumstances which produce learning? What is the content of learning? For each question, a critical principle is suggested: (1) Objects are constrained to behave in certain ways. If a violation is detected, an internal malfunction is assumed and subsequently corrected. (2) Learning involves mappings between entire perceptual dimensions rather than associations between individual stimuli. The principles are applied to two phenomena: the classic adaptation to prism distorted vision and the more recent, but equally elusive, McCollough effect. The view suggests a new interpretation of the McCollough effect and accounts for findings difficult to account for in other interpretations including which stimuli can successfully lead to contingent after-effects, the outcome of correlation manipulations, and why the effect exists at all. In addition, the phenomenon is linked to prism adaptation, usually regarded as a distinct type of plasticity. In general, the view advanced is that the two principles help distinguish perceptual learning from other types of learning processes.

Local and global factors in spatially-contingent coloured aftereffects

Dodwell and O'Shea's [(1987) Vision Research, 27, 569-580] conclusions that contingent coloured aftereffects (CAEs) depend on gobal pattern organization were investigated in four experiments. In Expt 1, we replicated findings that CAEs can be induced with complex patterns (concentric circles; radial spokes) under conditions of systematic eye movements. Contrary to Dodwell and O'Shea's argument that eye movements should uniformly cancel local orientation-colour contingencies, leaving only global effects, we reduced CAE magnitude by halving the diameter of the test stimuli. This suggests that cancellation did not occur uniformly over whole patterns, and that CAEs observed on these patterns are the residuals of uncancelled local orientation-colour contingencies. In Expt 2 we used central-fixation induction procedures to demonstrate that it is possible to induce CAEs with randomly-organized and locally-orthogonal orientation components. These findings are inconsistent with Dodwell and O'Shea's failure to observe CAEs under these conditions, and with their conclusion that global organization is necessary for CAE induction. However, CAEs induced with randomly-organized components were significantly weaker than those induced with globally-organized components. We examined the contribution of global organization in two additional experiments. In Expt 3 we induced CAEs with randomly-organized components under conditions in which the need for central fixation was removed, and found that CAE strength was directly related to the organization as well as the density of local-orientation components. In Expt 4, we found that the global organization of local-orientation components enhanced CAE strength only in regions away from the edges of these components: pattern organization did not affect the strength of CAEs at edges. We interpret these findings as evidence that CAEs may involve separate edge- and spread-colour components, and conclude that such components may account for observations previously attributed to global pattern geometry.

Selectivity for polar, hyperbolic, and Cartesian gratings in macaque visual cortex

The neural basis of pattern recognition is a central problem in visual neuroscience. Responses of single cells were recorded in area V4 of macaque monkey to three classes of periodic stimuli that are based on spatial derivative operators: polar (concentric and radial), hyperbolic, and conventional sinusoidal (Cartesian) gratings. Of 118 cells tested, 16 percent responded significantly more to polar or hyperbolic (non-Cartesian) gratings than to Cartesian gratings and only 8 percent showed a significant preference for Cartesian gratings. Among cells selective for non-Cartesian gratings, those that preferred concentric gratings were most common. Cells selective for non-Cartesian gratings may constitute an important intermediate stage in pattern recognition and the representation of surface shape.

Spatial-frequency-contingent color aftereffects: adaptation with two-dimensional stimulus patterns

The spatial-frequency theory of vision has been supported by adaptation studies using checkerboards in which contingent color aftereffects (CAEs) were produced at fundamental frequencies oriented at 45 degrees to the edges. A replication of this study failed to produce CAEs at the orientation of either the edges or the fundamentals. Using a computer-generated display, no CAEs were produced by adaptation of a square or an oblique checkerboard. But when one type of checkerboard (4 cpd) was adapted alone, CAEs were produced on the adapted checkerboard and on sine-wave gratings aligned with the fundamental and third harmonics of the checkerboard spectrum. Adaptation of a coarser checkerboard (0.80 cpd) produced CAEs aligned with both the edges and the harmonic frequencies. With checkerboards of both frequencies, CAEs were also found on the other type of checkerboard that had not been adapted. This observation raises problems for any edge-detector theory of vision, because there was no adaptation to edges. It was concluded that spatial-frequency mechanisms are operating at both low- and high-spatial frequencies and that an edge mechanism is operative at lower frequencies. The implications of these results are assessed for other theories of spatial vision.

Rapid discrimination of McCollough effects

The McCollough effect is a colour aftereffect that is contingent on pattern orientation. Three experiments were conducted to establish whether such aftereffect colours could serve as a basis for discrimination in several rapid discrimination tasks. In the first experiment it was investigated whether aftereffect colours could act like a simple 'feature' in a visual search task involving a difficult orientation discrimination. Without McCollough adaptation, the time taken to detect a 'target' among 'distractors' increased substantially as the number of distractors increased. With adaptation, detection time was essentially independent of the number of distractors, indicating that the nature of the task changed from a difficult orientation discrimination to a simple discrimination based on differences in aftereffect colours. The second and third experiments employed a difficult four-alternative forced-choice procedure in which subjects were required to discriminate a monochromatic patch of square-wave grating oriented at 45 degrees from three others oriented at 135 degrees (and vice versa). The gratings were presented very briefly (67-333 ms) followed by a 500 ms mask. Subjects performed the task with and without McCollough adaptation. Performance was strikingly better after adaptation: colour aftereffects could be used to make the discrimination even at exposure durations as short as 67 ms. The third experiment demonstrated that this enhanced performance was indeed due to perceived colour differences (rather than a possible contrast difference). The results of the three experiments are discussed in relation to proposals about the locus of the McCollough effect.

Color aftereffect contingent on text

During adaptation, two different letter strings (each five or six letters) were presented to subjects alternately, one in green and the other in magenta. The extent to which these letter strings subsequently elicited a color aftereffect was assessed. In different experiments, the chromatic letter strings consisted of words and nonwords. The results indicated that letter strings that form English words can contingently elicit a color aftereffect. This was the case even when the words were anagrams. There was no evidence that nonword letter strings could contingently elicit such an aftereffect, even when the nonwords conformed to English orthography. The results are relevant to understanding other contingent color aftereffects (McCollough effects), illusory color noted by computer operators who work at monochrome (green or amber) displays, and the processing of text.

Display organization and the detection of horizontal line segments

Observers searched for a horizontal line segment through displays containing varying numbers of elements differing from the target and from each other in terms of orientation. These elements were always positioned on imaginary concentric circles centered in the middle of the display. They were allocated to these positions either randomly or in such a way that their orientation was equal to that of the tangent to the circle at that position. The search for the target line appeared to proceed spatially in parallel with the latter class of displays, and serially with the former. These findings are explained and discussed within the context of the attentive-preattentive dichotomy that characterizes spatial vision.

Pattern-contingent color aftereffects on noninduced patterns

In a series of experiments, we found that in addition to expected reports of color aftereffects on patterns viewed during induction, reliable and predictable reports of color were given by subjects to patterns they did not view during induction. These reports to noninduced patterns were generally to patterns that were orthogonal to the patterns seen during induction. Induction with, for example, a red vertical grating led to appropriate aftereffects (i.e., green) on that vertical pattern and to the complementary aftereffect (i.e., pink) on a horizontal grating. We suggest that such color aftereffects on noninduced patterns are based on a shift in the activity of orientation coding mechanisms as a result of viewing the inducing patterns. We further propose that the results are consistent with the Lie transformation group theory of neuropsychology and that they add to a growing body of research demonstrating the applicability of this theory to the understanding of pattern-contingent color aftereffects.

Overprediction and blocking in the McCollough aftereffect

The Rescorla-Wagner theory (Rescorla & Wagner, 1972) of associative learning offers specific predictions about the associative strength of CS-US pairs when two or more CSs are conditioned to the same US separately and subsequently paired in a compound with the same US. The magnitude of orientation-contingent color aftereffects (AEs) was used as an index of associative strength in this study. The results of experiments using an "overprediction" (Rescorla, 1970) and a "blocking" (Kamin, 1969) paradigm conformed to the predictions of the Rescorla-Wagner theory. In Experiment 1, AEs were established simultaneously for horizontal-vertical and diagonal patterns. When observers subsequently viewed compound induction patterns, AE magnitude was found to be significantly decreased, relative to a condition in which observers did not view such an induction pattern. In Experiment 2, AE magnitude for a given test pattern following inspection of compound induction stimuli was significantly reduced by inspection of the other component prior to viewing the compound induction stimuli. The applicability of associative learning and feature-adaptation models of the McCollough effect is discussed.

Global factors generate the McCollough effect

It has recently been demonstrated that certain globally orthogonal pattern pairs derived from Hoffman's Lie Transformation Group Model of Neuropsychology (LTG/NP) are sufficient to generate McCollough effects (pattern-contingent colour aftereffects: PCCAES). We now provide evidence that the global factors postulated to generate these particular PCCAEs are also necessary. In Experiment 1 observers made controlled eye movements over the induction patterns and subsequently showed PCCAEs which were as strong as those obtained under central fixation. The induction process therefore did not depend on selectively adapting local oriented and colour-specific edge detectors. The global structure of the inducing patterns appears to be a necessary component of the generation of such aftereffects. In Experiment 2, a number of pattern pairs having various degrees of global organization were used as induction patterns, but in every case there was an equal degree of local orthogonality. Local orthogonality was shown not to be sufficient to induce PCCAEs, but global orthogonality was. These results lend strong support to the notion that McCollough-type PCCAEs are generated in a structure of the visual system somewhat above the level of strictly local operations, whether defined as edge and bar detectors or, equivalently, as units selectively tuned to different spatial frequencies and orientations. The global nature of the effects, and the particular pattern structures involved, lead us to suggest that Hoffman's LTG/NP provides a sound basis for explaining our findings.