Seeing the Big Picture
1999; Cell Press; Volume: 24; Issue: 4 Linguagem: Inglês
10.1016/s0896-6273(00)81026-0
ISSN1097-4199
AutoresLisa J. Croner, Thomas D. Albright,
Tópico(s)Neurobiology and Insect Physiology Research
ResumoThe effortlessness of visual perception belies the complexity of the task. For example, our perceptions of objects depend upon the integration of many different sources of visual information, or visual cues, such as color, motion, texture, and brightness (throughout this review, the term "cue" refers to a light parameter, such as color, texture, etc., that can vary to produce visible contrast; the term "feature" refers to a spatially localized instance of such contrast, possibly in a particular form—e.g., a dot, edge, or corner). Integration of these cues pervades all levels of visual processing in the primate brain, from early detection and representation of image features to object recognition and the transformation of sensory information into motor output. Befitting the central role the integration of visual cues plays, its study has a long and rich history—riddled with controversy. Recently, however, the development of experimental approaches that allow for parallel acquisition of perceptual and neurophysiological data and the emergence of technology for controlling complex visual stimuli have begun to reveal a coherent picture of both the functional significance of cue integration and the underlying neuronal mechanisms. The relationship between the cues of color and motion illustrates the topic and, owing to the availability of new data, serves to introduce a family of concepts, methods, and conclusions. The predominant view of visual cue integration has roots in two fundamental concepts of nineteenth century neuroscience: (1) elemental qualities of sensation and (2) localization of function within the brain. The former concept, known as "elementism," holds that perception results from an association of sensations and is reducible to a set of physically independent sensory "elements," such as color, brightness, and motion (10Boring E.G Sensation and Perception in the History of Experimental Psychology. D. Appleton-Century Company, New York1942Google Scholar). Localization of function, which is the more familiar of the two nineteenth century concepts, embraces the now extensively documented fact that specific functional processes can be localized to specific brain regions. In the twilight of the twentieth century, these two concepts have been explicitly joined in the idea that there are specific neuronal representations for the elemental qualities of vision—i.e., some cortical areas code for color processing, others for motion processing, etc. As we shall see, several lines of anatomical and physiological data appear to support this modular view, which leads to an important question: how are categories of image information that are thought to be processed independently linked together to yield a unified perceptual experience? Our goal in this review is to consider a number of observations that address this question, in the light of a theoretical framework for visual cue integration. In doing so, we lean heavily upon research from our laboratory, where several studies have been directed toward understanding the relationships between the neuronal representations of color and motion in the primate visual system. We can begin to consider how cue integration occurs by first noting that the simple modular processing scheme outlined above is flawed, inasmuch as it considers—as did the proponents of elementism—that information about different cues is combined linearly, with no interactions between cues. As has been documented repeatedly by psychologists of the Gestalt tradition (e.g.29Koffka K Principles of Gestalt Psychology. Harcourt, Brace, and Company, New York1935Google Scholar), the association of a set of sensory cues frequently leads to perceptual consequences that are not reducible to the individual effects of each cue. On the contrary, the perceptual interpretation of a cue is often swayed by the context of other cues with which it appears. Constraints on the nature of the interactions among visual cues arise when one considers how the visual system functions as a whole—as evinced by the rules of retinal image formation, the structure of the visual environment, and the behavioral goals of the observer. For visual motion, these constraints lead to the following operational principles for cue interactions, which suggest, in turn, specific testable hypotheses. An observer's window on the world is the pattern of light cast upon the retinal surface, i.e., the "retinal image." Motion of the observer, or motion of objects in the observer's environment, leads to changes in the retinal image, or retinal image motion. But retinal image motion is not sensed directly. Detection of such motion is secondary to the detection of some type of spatial contrast in the pattern of retinal illumination, such as that caused by changes in the intensity, pattern, or spectral composition of light. To perform optimally in a variable environment, a motion detector should generalize across spatial contrast cues, thus encoding the motion of an image feature regardless of the cue that enables it to be seen. For exampleFigure 1A illustrates that a thrown ball has the same motion regardless of its surface coloration or patterning; the ideal motion-sensing system would thus give the same signal about motion regardless of the cues that distinguish the ball from the background. We perceive the motions of objects, not the motions of retinal image features. It is only through retinal image features, however, that we can infer object motion, and that inference is indirect. Specifically, images of our world are formed on the retinal surface by optical projection, which by dimensional reduction precludes a unique relationship between any single retinal image feature and the object that gave rise to it. To construct a neuronal representation of object motion, the visual system must overcome this ambiguity, which generally can be done by evaluating contextual cues—i.e., local patterns of brightness, color, texture, etc.—in which each moving retinal image feature appears. For example, to evaluate the motion of the trotting horse shown in Figure 1B, a motion-sensing system must use visual context to determine that the regions segregated by foliage are parts of the same moving animal. We distinguish and recognize objects, and select them as targets for action, on the basis of their visual uniqueness, which can often be found solely in the conjunction of image cues, such as color and motion. Neuronal representations of cues that arise from a common object should be bound together, thereby forming both a representation of the whole object and a neuronal signal to guide actions directed at a target. For example, if we wish to select among the horses shown in Figure 1C, neuronal signals about colors, textures, shapes, and motions must be combined to distinguish the individual animals. Because they rely upon visual cues in different ways, these three principles are likely manifested as different kinds of interactions between the neuronal representations of distinct cue types. We will address these proposed interactions further below. First, we will briefly describe the anatomical and physiological evidence for modular processing. Considerable anatomical and physiological evidence supports the claim that different image cues—particularly color and motion—are processed independently in the primate visual system. These data and their implications have been thoroughly reviewed elsewhere (e.g.38Livingstone M.S Hubel D.H Segregation of form, color, movement, and depth anatomy, physiology, and perception.Science. 1988; 240: 740-749Crossref PubMed Scopus (2277) Google Scholar, 57Schiller P.H Logothetis N.K The color-opponent and broad-band channels of the primate visual system.Trends Neurosci. 1990; 13: 392-398Abstract Full Text PDF PubMed Scopus (240) Google Scholar, 42Merigan W.H Maunsell J.H How parallel are the primate visual pathways?.Annu. Rev. Neurosci. 1993; 16: 369-402Crossref PubMed Scopus (1224) Google Scholar, 62Stoner G.R Albright T.D Image segmentation cues in motion processing implications for modularity in vision.J. Cogn. Neurosci. 1993; 5: 129-149Crossref PubMed Scopus (78) Google Scholar, 20Dobkins K.R Albright T.D The influence of chromatic information on visual motion processing in the primate visual system.in: Watanabe T High-Level Motion Processing—Computational, Neurobiological, and Psychophysical Perspectives. MIT Press, Cambridge, MA1998Google Scholar); here, we summarize a few key points and introduce some relevant reservations. Parallel processing for color and motion is thought to originate in the retina. In primates, two morphologically distinct classes of retinal ganglion cells, termed parasol and midget ganglion cells, project exclusively to two different sets of layers in the lateral geniculate nucleus (LGN) of the thalamus, known as the magnocellular (M) and parvocellular (P) layers. Cells of the M and P layers project, in turn, to two distinct sublayers of the primary visual cortex (Figure 2A). Physiological studies at each processing stage have shown that, beginning in the retina, neurons in the M and P pathways possess characteristic responses to stimulus features: M pathway neurons exhibit properties indicative of a role in motion processing; whereas P pathway neurons, by contrast, exhibit properties that suggest contributions to both color and form processing. A second avenue of research has focused on the connections and functional contributions of cortical visual areas. Anatomical studies have revealed two principal processing streams ascending from primary visual cortex (V1): a "dorsal stream" that extends from V1 into the parietal lobe, and a "ventral stream" that terminates in the temporal lobe (Figure 2B). One view, promoted by 70Ungerleider L.G Mishkin M Two cortical visual systems.in: Ingle D.J Goodale M.S Mansfield R.J.W Analysis of Visual Behavior. MIT Press, Cambridge, MA1982Google Scholar and expanded upon by 44Milner A.D Goodale M.A The Visual Brain in Action. Oxford University Press, New York1995Google Scholar and others, holds that the dorsal and ventral streams serve visual–spatial ("where") and object recognition ("what") functions, respectively. Thus, two separate lines of research each led to proposed functional dichotomies—the first present at relatively early stages of visual processing and consisting of segregated color/form versus motion pathways, the second present at later (cortical) processing stages and manifested as separate representations of the "what" and "where" properties of objects. It was tempting to conclude that these two dichotomies were related; not surprisingly, a proposal of this sort by 37Livingstone M.S Hubel D.H Psychophysical evidence for separate channels for the perception of form, color, movement, and depth.J. Neurosci. 1987; 7: 3416-3468PubMed Google Scholar attracted considerable attention. The proposal was based, in part, upon two types of findings: (1) evidence for anatomical continuity between components of early- (M versus P) and late-stage (dorsal versus ventral) dichotomies, and (2) physiological and neuropsychological evidence suggesting that the dorsal and ventral cortical streams are specialized—mirroring the early M and P pathways—for motion and color processing, respectively. Not only does this view offer an attractive conjunction of elementism and modularity, but it provides a seamless extension of M and P channels through the highest levels of visual cortex. This proposal has stimulated a plethora of studies attempting to uncover perceptual consequences of segregated color and motion processing. Despite such efforts, however, there is a striking lack of consensus in the field. We argue that this lack of consensus has arisen, at least in part, because the proposal is based upon faulty principles of visual processing, as described above. The constraints identified above, as well as the recent findings summarized below, suggest that the truth is more complex. Most studies of the interactions between visual cues have focused on the contributions of nonmotion cues (e.g., color) to motion processing. Hence, the discussion here will concentrate on neuronal populations that are specialized for motion processing and the routes by which nonmotion information may be accessible to those neurons. In primates, motion direction is first signaled by neurons in V1. It is, however, at a subsequent cortical stage, known as the middle temporal visual area (area MT), that motion processing comes into its own: ∼90% of MT neurons exhibit pronounced selectivity for the direction and speed of a stimulus moved through the receptive field (75Zeki S.M Functional organization of a visual area in the posterior bank of the superior temporal sulcus of the rhesus monkey.J. Physiol. 1974; 236: 549-573PubMed Google Scholar, 71Van Essen D.C Maunsell J.H.R Bixby J.L The middle temporal visual area in the macaque myeloarchitecture, connections, functional properties and topographic connections.J. Comp. Neurol. 1981; 199: 293-326Crossref PubMed Scopus (475) Google Scholar, 41Maunsell J.H Van Essen D.C Functional properties of neurons in middle temporal visual area of the macaque monkey. I. Selectivity for stimulus direction, speed, and orientation.J. Neurophysiol. 1983; 49: 1127-1147PubMed Google Scholar, 1Albright T.D Direction and orientation selectivity of neurons in visual area MT of the macaque.J. Neurophysiol. 1984; 52: 1106-1130PubMed Google Scholar, 6Albright T.D Desimone R Gross C.G Columnar organization of directionally selective cells in visual area MT of the macaque.J. Neurophysiol. 1984; 51: 16-31PubMed Google Scholar). The responses of a typical MT neuron to moving stimuli are shown in Figure 3. Motion up and to the right evoked the largest responses from this neuron and would be referred to as the neuron's "preferred" direction; motion in other directions either evoked smaller responses or suppressed the neuron's spontaneous (unstimulated) activity. Different MT neurons have different preferred directions, and the range of preferred directions spans 360° without bias. As diagramed in Figure 2B, MT receives direct cortical input from areas V1, V2, and V3 and also has connections with areas thought to lie at the same level of the visual hierarchy, such as area V4. MT neurons integrate information over a larger region of visual space than do their cortical afferents; on average, the receptive field diameter of an MT neuron is equal to its distance from the center of gaze (5Albright T.D Desimone R Local precision of visuotopic organization in the middle temporal area (MT) of the macaque.Exp. Brain Res. 1987; 65: 582-592Crossref PubMed Scopus (204) Google Scholar). While the layout of receptive fields across MT preserves the neighbor relations of visual space, a much larger portion of MT is devoted to processing information arising from the center, rather than the periphery, of the visual field (23Gattass R Gross C.G Visual topography of striate projection zone (MT) in posterior superior temporal sulcus of the macaque.J. Neurophysiol. 1981; 46: 621-638PubMed Google Scholar, 71Van Essen D.C Maunsell J.H.R Bixby J.L The middle temporal visual area in the macaque myeloarchitecture, connections, functional properties and topographic connections.J. Comp. Neurol. 1981; 199: 293-326Crossref PubMed Scopus (475) Google Scholar, 5Albright T.D Desimone R Local precision of visuotopic organization in the middle temporal area (MT) of the macaque.Exp. Brain Res. 1987; 65: 582-592Crossref PubMed Scopus (204) Google Scholar). MT neurons project to higher cortical areas that appear to encode more complex forms of motion. A variety of findings have supported a role for MT in the analysis and perception of motion. As described above, the vast majority of MT neurons respond selectively to the direction and speed of a moving stimulus. In addition, MT is organized such that motion direction is represented in a columnar fashion, with all directions represented for each location in visual space (6Albright T.D Desimone R Gross C.G Columnar organization of directionally selective cells in visual area MT of the macaque.J. Neurophysiol. 1984; 51: 16-31PubMed Google Scholar, 5Albright T.D Desimone R Local precision of visuotopic organization in the middle temporal area (MT) of the macaque.Exp. Brain Res. 1987; 65: 582-592Crossref PubMed Scopus (204) Google Scholar). The importance of MT to motion perception is further supported by the fact that its damage in either human or animal observers severely compromises motion perception and may even lead to motion blindness (76Zihl J Von Cramon D Mai N Selective disturbance of movement vision after bilateral brain damage.Brain. 1983; 106: 313-340Crossref PubMed Scopus (634) Google Scholar, 47Newsome W.T Paré E.B A selective impairment of motion perception following lesions of the middle temporal visual area (MT).J. Neurosci. 1988; 8: 2201-2211PubMed Google Scholar). Finally, electrical microstimulation alters perceived motion in a predictable manner (54Salzman C.D Britten K.H Newsome W.T Cortical microstimulation influences perceptual judgments of motion direction.Nature. 1990; 346: 174-177Crossref PubMed Scopus (582) Google Scholar, 55Salzman C.D Murasugi C.M Britten K.H Newsome W.T Microstimulation in visual area MT effects on direction discrimination performance.J. Neurosci. 1992; 12: 2331-2355PubMed Google Scholar). The full range of MT response characteristics, anatomical organization, and suggested contributions to perception are reviewed elsewhere (4Albright T.D Cortical processing of visual motion.Rev. Oculomot. Res. 1993; 5: 177-201PubMed Google Scholar). MT's dominant role in motion processing would seem to place it solidly in the so-called M pathway through visual cortex. As such, MT is an ideal cortical location for investigating the contributions of nonmotion cues to motion processing in terms of the three principles outlined above. In the remainder of this review, we will address how each of these principles is manifested in the response properties of MT neurons. Most retinal image cues, such as luminance (a measure related to brightness), chrominance (related to hue), and texture, arise from spatial variations in the light cast upon the retinae. Motion, of course, takes place over time as well as space and is thus a secondary cue that depends upon some primary form of spatial definition. Under most circumstances, that spatial definition is the result of multiple coincident cues—the light reflected from a journal, for example, may be brighter, redder, and less textured than that reflected from the desk on which it lies—but the primate visual system can detect spatial contrast arising from any one of these cues alone. For them to be broadly utilitarian, the responses of neurons sensitive to motion must likewise be independent of the particular spatial cues present in the image. The sensitivity of the primate visual system to the motion of patterns defined by luminance contrast with the background is well established, both at perceptual and neuronal levels (reviewed by 4Albright T.D Cortical processing of visual motion.Rev. Oculomot. Res. 1993; 5: 177-201PubMed Google Scholar, 20Dobkins K.R Albright T.D The influence of chromatic information on visual motion processing in the primate visual system.in: Watanabe T High-Level Motion Processing—Computational, Neurobiological, and Psychophysical Perspectives. MIT Press, Cambridge, MA1998Google Scholar). By contrast, much has been made of the fact that moving stimuli defined solely by chromatic contrast (such as a pattern of red and green stripes of the same luminance) sometimes appear to move more slowly than their luminance-defined counterparts (e.g.14Cavanagh P Tyler C.W Favareau O.E Perceived velocity of moving chromatic gratings.J. Opt. Soc. Am. (A). 1984; 1: 893-899Crossref PubMed Scopus (313) Google Scholar, 37Livingstone M.S Hubel D.H Psychophysical evidence for separate channels for the perception of form, color, movement, and depth.J. Neurosci. 1987; 7: 3416-3468PubMed Google Scholar). This result has been hailed as evidence that chromatic contrast does not serve as a reliable spatial cue for motion processing, thus indicating a lack of interaction between the neuronal representations of color and motion. The direction of motion of chrominance-defined stimuli is generally perceived veridically, however, particularly if the stimuli are designed to elicit sufficiently contrasting responses from different cone photoreceptor populations (18Dobkins K.R Albright T.D Color, luminance, and the detection of visual motion.Curr. Dir. Psychol. Sci. 1993; 2: 189-193Crossref Scopus (9) Google Scholar, 50Palmer J Mobley L.A Teller D.Y Motion at isoluminance discrimination/detection ratios and the summation of luminance and chromatic signals.J. Opt. Soc. Am. (A). 1993; 10: 1353-1362Crossref PubMed Scopus (39) Google Scholar, 26Hawken M.J Gegenfurtner K.R Tang C Contrast dependence of colour and luminance motion mechanisms in human vision.Nature. 1994; 367: 268-270Crossref PubMed Scopus (121) Google Scholar, 67Stromeyer III, C.F Kronauer R.E Ryu A Chaparro A Eskew Jr., R.T Contributions of human long-wave and middle-wave cones to motion detection.J. Physiol. (Lond.). 1995; 485: 221-243Google Scholar). Neurophysiological studies of the contribution of chromatic cues to motion processing have focused largely on responses in area MT. In seeming support of MT's proposed position in the "color blind" M pathway, early studies found that MT neurons have little selectivity for either the chrominance or form of moving stimuli (75Zeki S.M Functional organization of a visual area in the posterior bank of the superior temporal sulcus of the rhesus monkey.J. Physiol. 1974; 236: 549-573PubMed Google Scholar, 8Baker J.F Petersen S.E Newsome W.T Allman J.M Visual response properties of neurons in four extrastriate visual areas of the owl monkey (Aotus trivirgatus) a quantitative comparison of medial, dorsomedial, dorsolateral and middle temporal areas.J. Neurophysiol. 1981; 45: 387-405Google Scholar, 41Maunsell J.H Van Essen D.C Functional properties of neurons in middle temporal visual area of the macaque monkey. I. Selectivity for stimulus direction, speed, and orientation.J. Neurophysiol. 1983; 49: 1127-1147PubMed Google Scholar, 1Albright T.D Direction and orientation selectivity of neurons in visual area MT of the macaque.J. Neurophysiol. 1984; 52: 1106-1130PubMed Google Scholar). However, more recent studies have shown that some MT neurons continue to signal the direction of moving patterns defined only by chromatic contrast (53Saito H Tanaka K Isono H Yasuda M Mikami A Directionally selective response of cells in the middle temporal area (MT) of the macaque monkey to the movement of equiluminous opponent color stimuli.Exp. Brain Res. 1989; 75: 1-14Crossref PubMed Scopus (132) Google Scholar, 24Gegenfurtner K.R Kiper D.C Beusmans J.M Carandini M Zaidi Q Movshon J.A Chromatic properties of neurons in macaque MT.Vis. Neurosci. 1994; 11: 455-466Crossref PubMed Scopus (127) Google Scholar). Thus, MT apparently receives chromatic information that, according to the segregated pathways hypothesis, should be available only to the P stream. To determine whether this chromatic information reaches MT via the P pathway18Dobkins K.R Albright T.D Color, luminance, and the detection of visual motion.Curr. Dir. Psychol. Sci. 1993; 2: 189-193Crossref Scopus (9) Google Scholar, 19Dobkins K.R Albright T.D What happens if it changes color when it moves? The nature of chromatic input to macaque visual area MT.J. Neurosci. 1994; 14: 4854-4870PubMed Google Scholar studied whether motion is processed in such a way as to strongly preserve the chromatic identity of moving features. Specifically, if the motion of a border defined by chromatic contrast is visible only when the colors on either side of the border stay the same, this would support P pathway input to MT. However, if a chromatic border is seen to move even when the colors change, then M pathway input suffices (see below). To study this, Dobkins and Albright used visual stimuli known as chromatic "sinusoidal gratings" (Figure 4) that could be configured to convey two opposing motion signals simultaneously—one direction visible if motion correspondence (the pairing of two spatially and temporally disparate stimuli to form a motion cue) preserves chromatic identity, and the opposite direction visible if motion correspondence ignores chromatic identity (Figure 4C). The perceptual reports of human observers revealed that the motion system generally disregards chromatic identity; i.e., it utilizes image features defined by chromatic contrast, independent of the specific colors involved. When individual MT neurons in monkeys were presented with the stimuli diagramed in Figure 4, the pattern of responses mirrored those of human observers (19Dobkins K.R Albright T.D What happens if it changes color when it moves? The nature of chromatic input to macaque visual area MT.J. Neurosci. 1994; 14: 4854-4870PubMed Google Scholar). Dobkins and Albright argued that these results could be accounted for by signals traveling to MT via the M layers of the LGN. Specifically, the observed lack of sensitivity to chromatic identity is readily explained by the fact that many M cells encode the presence of chromatic contrast but not its sign (56Schiller P.H Colby C.L The responses of single cells in the lateral geniculate nucleus of the rhesus monkey to color and luminance contrast.Vision Res. 1983; 23: 1631-1641Crossref PubMed Scopus (128) Google Scholar, 32Lee B.B Martin P.R Valberg A The physiological basis of heterochromatic flicker photometry demonstrated in the ganglion cells of the macaque retina.J. Physiol. 1988; 404: 323-347PubMed Google Scholar, 33Lee B.B Martin P.R Valberg A Sensitivity of macaque retinal ganglion cells to chromatic and luminance flicker.J. Physiol. 1989; 414 (a): 223-243PubMed Google Scholar, 34Lee B.B Martin P.R Valberg A Amplitude and phase of response of macaque retinal ganglion cells to flickering stimuli.J. Physiol. 1989; 414 (b): 245-263PubMed Google Scholar, 35Lee B.B Martin P.R Valberg A Nonlinear summation of M- and L-cone inputs to phasic retinal ganglion cells of the macaque.J. Neurosci. 1989; 9 (c): 1433-1442PubMed Google Scholar, 39Logothetis N.K Schiller P.H Charles E.R Hurlbert A.C Perceptual deficits and the activity of the color-opponent and broad-band pathways at isoluminance.Science. 1990; 247: 214-217Crossref PubMed Scopus (122) Google Scholar). Thus, signals traveling through the M layers of the LGN and extending to area MT provide a potential route by which chromatic information can influence detection of motion. While these findings confirm a contribution of chromatic cues to motion processing and suggest underlying mechanisms, recent debate has centered on the strength of sensitivity to chromatic contrast. To address this issue68Thiele A Dobkins K.R Albright T.D The contribution of color to motion processing in macaque area MT.J. Neurosci. 1999; in pressGoogle Scholar adopted an established paradigm (13Cavanagh P Anstis S The contribution of color to motion in normal and color-deficient observers.Vision Res. 1991; 31: 2109-2148Crossref PubMed Scopus (211) Google Scholar) that enables one to calibrate the sensitivity of motion detectors to chrominance-defined stimuli, relative to sensitivity to a luminance-defined standard. The stimulus used consisted of two superimposed sinusoidal gratings moving in opposite directions (Figure 5). One grating was defined solely by luminance contrast (achromatic), while the other had both chromatic contrast (alternating stripes were red and green) and varying amounts of luminance contrast (red stripes could be brighter than green or vice versa). In general, the neuronal responses elicited by the achromatic and chromatic gratings moving simultaneously in opposite directions were dominated by the more salient of the two gratings. For example, if the stimulus consisted of a salient achromatic grating moving in a neuron's preferred direction and a weak chromatic grating (with low luminance contrast) moving in the nonpreferred direction, then the neuron's response was relatively large. If, on the other hand, the preferred-direction achromatic grating was less salient than the nonpreferred-direction chromatic grating (with high luminance contrast), then the response was relatively small. By systematically varying the luminance contrast of the chromatic grating, it was possible to determine the contrast value for which responses to the two oppositely moving components were equivalent. When chromatic and achromatic inputs were thus balanced, the difference in their luminance contrasts provided a precise measure of strength of sensitivity to the chromatically defined stimulus, in units of luminance contrast. The measure obtained, termed "equivalent luminance contrast," revealed that chromatic contrast has a powerful influence over the responses of MT neurons when luminance contrast in the moving stimulus is low (<5%–8% contrast). Increasing the luminance contrast of the moving stimulus, however, subordinated the effects of the chromatic component. Thus, the chromatic properties of an object convey important information about motion direction to MT neurons only when insufficient luminance contrast exists; otherwise, motion of luminance-defined components drives the neuronal responses. In further support of the proposal of 19Dobkins K.R Albright T.D What happens if it changes color when it moves? The nature of chromatic input to macaque visual area MT.J. Neurosci. 1994; 14: 4854-4870PubMed Google Scholar summarized above, these results can be accounted for by simple pooling of the responses of LGN M c
Referência(s)