BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI
The Perception of Motion in Chromatic Stimuli
Simon J. Cropper
University of Melbourne, Australia
Sophie M. Wuerger
University of Liverpool, United Kingdom
The issue of whether there is a motion mechanism sensitive to
purely chromatic stimuli has been pertinent for the past 30 or
moreyears. The aim ofthis review is toexamine whysuch differ-
ent conclusions have been drawn in the literature and to reach
some reconciliation. The review critically examines the behav-
ioral evidence and concludes that there is a purely chromatic
motion mechanism but that it is limited to the fovea. Examina-
tion of motion performance for chromatic and luminance stim-
uli provides convincing evidence that there are at least two dif-
ferent mechanisms for the two kinds of stimuli. The authors
lar disadvantage when the integration of multiple local motion
signals is required. Finally, the authors present a descriptive
model that may go some way toward explaining the reasons for
the differences in collected data outlined in this article.
The fundamental structure of many computational
theories of vision, particularly those developed and
wealth of anatomical and physiological evidence
(reviewed in Zeki, 1993), has strongly encouraged us to
think of the visual system as processing different aspects
of the image in largely parallel and separate streams.
parallel processing is the claim that the color and the
motion within a natural image are processed independ-
ently and that there is no motion mechanism that
is to assess whether published experimental work sup-
ports this hypothesis. We ask two basic questions:
1. Can a chromatic signal be used for motion discrimina-
a low-level chromatic motion mechanism? Are there in-
stances in which luminance and color information
always processed by separate pathways? Or does color
act like a low-contrast luminance stimuli?
Definitions and Scope
To put these questions in the context of a motion-
processing hierarchy, we define “low level” as the direct
sensitivity to the spatiotemporal orientation of the input
and “high level” as an extrapolation of the temporal ori-
entation from successive spatial analysis; this definition
amounts to also meaning anything that cannot be
defined as “low level.” We will review both kinds of
motion perception, although the question posed above
refers directly to the low-level analysis. Furthermore, we
ulus (e.g., first order, second order, etc.) strictly to describe
the stimulus structure, independent of the mechanism
of stimulus detection or discrimination (Cavanagh &
Mather, 1989; Julesz, 1971).
We focus strongly on the behavioral evidence in this
review as this forms the vast majority of the published
tion posed. The Speculations section outlines a descrip-
tive model to explain the data and is preceded by a brief
outline of the most relevant neurophysiological data in
to thank sincerelyourcolleagueswho respondedto ourgenerale-mail
request for references and hope that we have cited you appropriately.
Randolph Blake, Bradley Wolfgang, Caroline Andrews, Michael
Johnston, LeanneKennedy, SueHarding, and Alexa Ruppertsbergfor
comments on the article and for general assistance and encourage-
ment. The work was supported by a Wellcome Trust Biomedical Re-
search Collaboration grant to the authors and by Australian Research
Council QE2 Fellowship and large grant awarded to S.J.C. and
Wellcome Trust Grant 5013 2919 to S.M.W.
Behavioral and Cognitive Neuroscience Reviews
Volume 4 Number 3, September 2005 192-217
© 2005 Sage Publications
to several other recent pieces of work that review the
neurophysiological literature well (Cavanagh & Anstis,
1991; Derrington, 2000; Dobkins & Albright, 1993a,
1993b; Dobkins & Albright, 2004; Hawken &
The description of a stimulus as chromatic, as
opposed to luminance, means that the critical proper-
ties of that stimulus allowing any system to detect its
motion are coded only by a change in the stimulus color
and not by any change in its luminance; this is termed
equiluminant. Thus, only a mechanism sensitive to color
change, in any way, can detect that motion. A stimulus is
defined in this sense on three levels. Primarily, the pho-
tometric properties of the stimulus are described in
terms of the stimulus as it is displayed on the stimulus
generation equipment, most commonly a cathode ray
tube. Thus, photometric equiluminance means that
there is no luminance change when measured by a pho-
tometer equipped with a photometric head: a filter
approximating the human visual systems sensitivity pro-
file as a function of wavelength.1Once defined as such,
the chromatic properties of the stimulus must be con-
trolled at the level of the projection of the stimulus on
the retina. The ocular properties of the eye can cause a
deviation from photometric equiluminance through
chromatic aberration under many conditions, and this
deviation must somehow be accounted for. This is usu-
ally achieved through both physical and behavioral
means. Physically, controls such as keeping the spatial
frequency (the rate of change over space) below 1 cycle
per degree of visual angle and the temporal frequency
(rate of change over time) correspondingly low to mod-
erate minimizes the chromatic aberration. Behaviorally,
minimal flicker, minimally distinct border, and mini-
mum motion are all commonly used measures to nor-
malize the stimulus to the observer’s equiluminant
point, which varies between individuals as does almost
every other aspect of the visual system. Finally, there is
the neural representation of the stimulus that has been
transduced attheretina andbecome apatternof neural
firing within the system. At this point, the definition of
equiluminance really falls down as the stimulus is repre-
sented, however, the system is able to do it, and the defi-
nition becomes a property of the result. It has also long
been argued that because individual neurons have
our overarching question as follows: Are the critical
properties of the internal signal mediating the motion
by a static purely chromatic modulation, and is that
signal independent of any luminance-coded motion
signal? In other words, is the internal signal functionally
CAN WE SEE CHROMATIC STIMULI MOVE?
Motion Discrimination Data
Although the answer to this question is empirically
relatively simple, data showing our basic ability to dis-
criminate motion in purely chromatic stimuli is
able, so is its direction of motion (Levinson & Sekuler,
1975; Watson, Thompson, Murphy, & Nachmias, 1980),
nance grating of moderate to low spatial frequency
remains fairlyconstantwithincreasing stimuluscontrast
(Boulton, 1987; Boulton & Hess, 1990; Cropper &
Derrington, 1994; Johnston & Wright, 1985). A purely
chromatic grating, however, requires a significantly
higher drift rate to discriminate its direction of motion
at a low contrast (0.5 log units above detection thresh-
old) compared to a luminance grating.Itis not until the
chromatic grating is increased in contrast to about 10
times detection threshold that the motion threshold,
measured in terms of the minimum drift rate, becomes
equivalent for chromatic and luminance stimuli
(Cropper & Derrington, 1994).
Alternatively, if one expresses performance as the
ratio between the contrast required to detect the stimu-
lus and the contrast required to discriminate its motion,
be much higher for a chromatic stimulus (Cavanagh &
Anstis, 1991; Derrington & Henning, 1993; Kooi & De
Valois, 1992; Lindsey & Teller, 1990; Metha & Mullen,
1998; Metha, Vingrys, & Badcock, 1994; Mullen &
Boulton, 1992a, 1992b; Palmer, Mobley, & Teller, 1993).
When this measure is parametrically examined, the
shape of contours for detection and for discrimination,
plotted in color space, imply that for red-green chro-
matic stimuli, two separate mechanisms may operate at
of a stimulus and another tuned to discriminate its
motion (Stromeyer, Kronauer, Ryu, Chaparro, & Eskew,
1995). This is revealed because as the temporal drift-
ination increases concomitantly (Stromeyer et al., 1995),
hue mechanism and the luminance mechanism remain
approximately constant with temporal frequency
This result suggests the existence of at least two alterna-
tively constructed mechanisms, at least in terms of their
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 193
ratio data are very hard to interpret clearly (also see the
Speculations section). This can be a problem when it is
so often referred to as a critical data set supporting the
absence of a chromatic motion mechanism.
The significant differences in the detection:discrimi-
measured under roughly comparable conditions
(although see the Reconciling the Noise and Grating
variation in the data between studies is significant: the
highest estimation of the detection:discrimination ratio
being around 8.5 (log = 0.93; Yoshizawa, Mullen, &
Baker, 2000) the lowest being close to 1 (log = 0;
Gegenfurtner & Hawken, 1995).
Slight differences between methodologies have been
cited as an explanation for the discrepant data
(Derrington & Henning, 1993; Lindsey & Teller, 1990;
Metha et al., 1994; Metha & Mullen, 1998). Derrington
and Henning (1993) suggested that the disparity
is due to the retinal eccentricity; if the stimulus were
located foveally, the detection:discrimination ratio
would be about unity, whereas if the stimulus were pre-
sented parafoveally, then the ratio would rise steeply. As
will become clear later in this section, a careful meta-
analysis of the stimulus size and retinal location reveals
much about the differences in data.
Summary and Critique of Basic Direction-
Even if one takes into account the differences out-
lined above,todrawconclusions abouttheproperties of
underlying motion mechanisms from thesedataisprob-
is usually contrast dependent. Luminance and chro-
matic stimuli are defined by two directions in a three-
dimensional cone space, and depending on what
contrast measure one adopts, one might reach different
conclusions as to how effective both kinds of stimuli are
in a given task (Switkes & Crognale, 1999). Second, if
one compares the ratio between detection and motion
discrimination thresholds for luminance and equi-
luminant red-green stimuli to gauge how effective chro-
the implicit assumption is thatthe mechanisms involved
in each task have the same relationship to one another
194 BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
Cropper (1992) SJC
Cropper (1992) CL
Cropper (1992) Luminance
Palmer et al (1993) LM
Palmer et al (1993) JS
Cavanagh & Anstis (1991) RG 0.5cpd
Cavanagh & Anstis (1991)
Derrington & Henning (1993)
Mullen & Boulton (1992)
Stromeyer et al (1995) CFS
Stromeyer et al (1995) RTE
Gegenfurtner & Hawken (1995) CT
Gegenfurtner & Hawken (1995) KG
Gegenfurtner & Hawken (1995) CT 2Hz
Gegenfurtner & Hawken (1995) KG 2Hz
Metha & Mullen (1998) ABM
Metha & Mullen (1998) KTM
Metha & Mullen (1998) ABM 2Hz
Metha & Mullen (1998) KTM 2Hz
Lindsey & Teller (1990) CA 3.75Hz
Lindsey & Teller (1990) DL 3.75Hz
Yoshizawa et al (2000) TY foveal
Yoshizawa et al (2000) TY perifoveal
Yoshizawa et al (2000) RPP foveal
Yoshizawa et al (2000) RPP perifoveal
NOTE: For comparison, one data set for luminance-defined stimuli is included (closed circles). The ratio between the contrast requiredto detect
the stimulus and the contrast required to discriminate its direction of motion is plotted against the duration of the stimulus for several different
studies. Straight lines connect the data points from a single study and observer. Different symbols indicate different studies as listed in the key.
Detection:Discrimination Threshold Ratios for L-M Chromatic Stimuli as a Function of the Duration of the Stimulus.
for a given stimulus subtype (color or luminance). This
raises the possibility that the detection:discrimination
contrast ratio may be different not because of different
motion mechanisms for luminance and color but
because of entirely different types of detection mecha-
nisms for luminance and chromatic stimuli. In other
words, it is the numerator in the ratio calculation that is
critical in determining the differences between chro-
matic and luminance stimuli rather than the denomina-
tor. This is expanded in the Speculations section.
In summary, it is clear thatwe can discriminate stimu-
lus motion when that stimulus is coded by color alone,
but this does not answer the critical question posed in
the introduction: Do we have a low-level motion mecha-
nism independently sensitive to chromatic change? To
address this question, the stimuli and tasks have been
manipulated such that particular hypothetical charac-
teristics of low-level motion detection may be isolated.
WHAT IS THE NATURE OF THE CHROMATIC
Adaptation to Motion: A Specific
The definition of stimulus properties on the basis of
long-range and short-range motion endured as useful
empirical categories for a significant period as a distinc-
tion between low-level and high-level processes (Anstis,
1970; Braddick, 1974; Pantle & Picciano, 1976), and the
classical phenomenon of a motion aftereffect was con-
sidered to be a defining property of the short-range
motion mechanism. Three coincident studies inde-
pendently showed that chromatic stimuli, patterns that
previously had been assumed to be processed by long-
range motion mechanisms, did induce and null a
motion aftereffect (Cavanagh & Favreau, 1985;
Derrington & Badcock, 1985; Mullen & Baker, 1985).
The conclusion of all three articles was that chromatic
stimuli may be detected by a low-level motion mecha-
nism and that the chromatic motion output was at some
subsequent stage combined with a luminance motion
signal; an observation consistent with the earlier study
on perceived speed by one of the authors (Cavanagh,
Tyler, & Favreau, 1984). The existence of a chromatic
Derrington, 1992; Webster, Day, & Cassell, 1992) and
extended such that not only does there seem to be a
strong contrast-dependent cross-adaptation between
color and luminance stimuli (Cropper & Derrington,
across spatial orientations when a luminance adapting
and discrimination of chromatic and luminance stimuli
showed thatadaptation to a drifting luminance modula-
tion increased the contrast required selectively to
discriminate motion of a chromatic stimulus (Willis &
Anderson, 1998). Although these data may be taken to
imply a common mechanism for chromatic and lumi-
nancemotion discrimination upto2cpd and4Hz,or at
tasks being affected differently from spatial (detection)
issue in the following section.
Masking of Motion
Although motion aftereffect studies are somewhat
limited in their scope, the related paradigm of simulta-
neous masking has proven to be more flexible and the
majority of the recent work on color and motion has
used some kind of masking paradigm.
Noise Masks: A Generic Masking Approach
One of the most commonly used forms of mask is
noise. The theory behind this is that the system has to
deal with noise all the time, both externally and inter-
nally generated, and that adding a measurable amount
of external noise, some of the characteristics of the
both qualitatively (Breitmeyer,1984), and quantitatively
A significant body of work examined the effect of
noise on the perceived motion of a chromatic Gabor
Yoshizawa, Mullen, & Baker, 2003). The examination of
chromatic stimuli built on earlier work examining the
density on luminance micropattern stimuli (Boulton &
Baker, 1993a, 1993b), which resulted in the proposal
that two motion mechanisms were responsible for the
measured performance. These correspond to motion
order (envelope) spatial modulation, respectively2
(Cavanagh & Mather, 1989; Julesz, 1971). The first-
order-mediated properties of a chromatically defined
micropattern were, however, masked by superimposed
luminance noise, leaving the second-order percept of
envelope motion relatively unaffected. Furthermore,
there was no effect of the noise on the detection of a
chromatic carrier (Baker et al., 1998; Yoshizawa et al.,
2000), consistent with previous work (Mullen, Cropper,
& Losada, 1997; Sankeralli & Mullen, 1997; Switkes,
Bradley, & De Valois, 1988; Willis & Anderson, 1998),
suggesting that the luminance noise was selectively dis-
rupting the percept of low-level chromatic motion. Fur-
ther work solidified this conclusion and, overall, the
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 195
series of studies combining luminance noise with chro-
matic test stimuli (Baker et al., 1998; Mullen et al., 2003;
Yoshizawa et al., 2000, 2003) concluded that an inter-
nallygenerated luminance-like signalmediated anylow-
level properties of the motion of chromatic stimuli and
that independent chromatic motion detection was
mediated only by a high-level mechanism.
Grating Masks: A Specific Masking Approach
An alternative approach to noise masking is to use a
tial and temporal structure, such as a grating. There are
two published interpretations of the effect of adding a
theoretical framework, the static grating mask will have
an effect only on a mechanism sensitive to the displace-
ment of the features (in this case, the shape of the struc-
tured contrast envelope) in the stimulus. Thus, immu-
nity to the mask (termed a pedestal in these experiments
and henceforth when this framework is assumed) indi-
cates a mechanism insensitive to precise stimulus struc-
ture and, by assumption, sensitive to the motion energy
in the composite stimulus (Lu, Lesmes, & Sperling,
1999; Lu & Sperling, 1995, 2001; Zaidi & DeBonet,
2000). In the alternative framework, the combination of
displaced test and static mask (the term thatwill be used
when discussing work adopting this theoretical frame-
work) is treated as a compound stimulus by the system if
common mechanisms are used at some stage in the pro-
cessing of the two stimulus components (Cropper,
2005b; Cropper & Derrington, 1996; Stromeyer,
Chaparro, & Kronauer, 1996; Zemany, Stromeyer,
Chaparro, & Kronauer, 1998). Empirical data exist to
oversimplifications of the story told by the data.
The pedestal experiments rely on the stimulus struc-
ture itself to define the underlying detection mecha-
nisms. Lu and colleagues (1999) combined a similarly
modulated pedestal andtest,thatis,both defined either
an effect only on the perceived motion of the chromatic
stimulus, leaving the luminance stimulus untouched.
Zaidi and DeBonet (2000) adopted the same pedestal
framework and found data partially in agreement with
Lu and colleagues (1999) yet also found independence
of chromatic motion when in the presence of a lumi-
nance pedestal and chromatic independence of a chro-
matic pedestal under some conditions. In a third study
further examining the pedestal approach in a more
parametric vein, Zemany and colleagues (1998) found
motion was affected by the pedestal, questioning the
narrow interpretation originally offered.
The alternative framework for the grating masking
paradigm takes a more traditional view and considers
the composite stimulus as a test and mask. The theory
does not rely on the stimulus structure itself to discrimi-
nate between the formats of the mechanism underlying
the motion performance but states that if the motion
mechanism is sensitive to both test and mask compo-
nents of the stimulus, then an interaction will be seen in
ilarly defined mask and test (luminance:luminance or
color:color) where common mechanisms in the system
will be sensitive to both test and mask. Although this
approach initially tests only for independence of the
underlying chromatic and luminance sensitive mecha-
nisms, an underlying low-level motion mechanism for a
of the stimulus such that a feature-sensitive mechanism
would not be able to signal the stimulus unambiguously
because it is simply not present for long enough
(Cavanagh, 1992; Cropper & Derrington, 1994, 1996).
When a short-duration (17 ms) chromatic grating is
displaced by 0.25 cycles in the presence of a static lumi-
nance mask, the motion is clearly discriminable, leaving
performance unaffected by the mask (Cropper &
by the strong reversal in perceived motion seen when
either color gratings or luminance gratings are com-
mask provides a critical intermediate condition, bridg-
ing the gap between the noise mask condition and the
grating mask condition. The result indicated independ-
ence of a chromatic motion response from a jittered
luminance mask to be selective to low stimulus contrasts
(Stromeyer et al., 1996).
Reconciling the Noise and Grating Masking Data
The reasons for differences in the data cited above
become clear only in the context of a full parametric
examination of the effects (Cropper, 2005a, 2005b).
One critical disparity between studies is the size and the
placement of the stimuli and another is the mean lumi-
nance of the display. These differences are summarized
in Figure 2. If the stimulus is 4 degrees or less and cen-
trally placed and the mean luminance is greater than 30
perception of chromatic motion; as the stimulus size is
increased or the pattern is placed outside the fovea,
interactions between color and luminance signals
become more pronounced (Cropper, 2005a). This
result can be explained by the suggestion that an inde-
pendent chromatic motion mechanism is present only
in the central portion of the visual field and can there-
fore be isolated only with foveal stimuli. Furthermore,
the stimulus must be presented at a clearly photopic
196BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
mean luminance to avoid the intrusion of any
We suggest that the effect of stimulus size and place-
ment is a by-product of the very specific wiring demands
of such a mechanism. Whatwe mean by this is thatthere
must be a spatially and temporally discrete input to any
motion detector. As one moves away from the fovea and
receptive field sizes increase, the spatial and temporal
acuity of the subsystems decreases and color and lumi-
nance signals become mixed. Although we obviously
have the ability to later make these mixed signals dis-
crete (Lennie & D’Zmura, 1988), it is probable that this
tion, the behavioral symptom of which is the masking of
color motion by temporal luminance noise. This reduc-
for luminance stimuli (Metha et al., 1994; Mullen, 1985,
1987, 1991), and it makes ecological sense to maintain
the acuity of luminance sensitive mechanisms as far as
mechanisms. Therefore, if there are any low-level chro-
matic motion detectors, they are more likely to exist in
the fovea than in any other part of the visual field.
From an empirical perspective, the use of a dynamic
noisemaskhasamore general effectthanperhapsorigi-
nally considered because any temporal judgment seems
to be impaired by the presence of a dynamic luminance
mask (Cropper, 2005a). This observation, in conjunc-
tion with the size dependency, can explain the discrep-
ancy in data both from adaptation studies (Willis &
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI197
Figure 2: Depiction of the Variation in Stimulus Size, Dimension, and Mean Luminance Across the Studies Cited in the Nature of the Chromatic
Motion Mechanism Section.
NOTE: Of the three figures expressing stimulus dimension, the top left shows stimulus dimensions where the respective authors concluded there
anism, and top right shows conditions where no conclusionwas reached. The bottom left graph plots mean display luminance against the study in
question and indicates whether the conclusion was for (4) or against (6) a chromatic motion mechanism.
Anderson, 1998) and the masking studies cited above. A
similar effect of stimulus size and mean luminance is
seen when both test and mask are gratings (Cropper,
between the grating mask and grating pedestal data is a
of the empirical evidence is in favor of the test:mask
interpretation of the stimulus combination, and there is
little evidence that pedestal immunity is a general result
for stimuli, the motion of which one would strongly
assume is being detected by an energy-sensitive mecha-
nism (Cropper, 2005b; Zaidi & DeBonet, 2000; Zemany
et al., 1998).
Reverse Masking: Color Masking Luminance
An alternative approach to examining the contribu-
tion of color to motion is to examine the effect the chro-
a luminance-based motion task. Morgan and colleagues
(Morgan & Cleary, 1992; Morgan & Ingle, 1994) used a
random-dot kinematogram to examine the chromatic
and luminance input to apparent motion perception
ulus. The effect of reversal of the chromatic polarity
frequency–dependent effect such that at low spatial fre-
quencies, albeit in a spatially broadband stimulus, the
chromatic properties of the stimulus dominate the per-
ceived direction of motion (Morgan & Cleary, 1992;
given the role of the random-dot kinematogram in the
classical short-range/long-range distinction (Braddick,
1974). A similar result was found with a grating stimulus
whose chromatic contrast reversed in synchrony with its
displacement (Dobkins & Albright, 1993a). This gener-
both signed and unsigned chromatic borders were
important when there was no consistently moving lumi-
nance cue in the stimulus. At larger displacements, the
signed border dominated over the unsigned; at smaller
displacements (<51.4 degrees of spatial phase angle),
the unsigned border dominated unless luminance con-
trast was added to the stimulus when the luminance-
signed border consistently governed perceived direc-
tion.Thegreaterdisplacement required toproduce this
effect is in line with the spatial frequency (and displace-
ment size) data shown with a random-check kinemato-
gram (Morgan & Cleary, 1992; Morgan & Ingle, 1994).
A particular form of masking that generally falls out-
side the commonly defined and accepted breadth of
masking studies is known as motion nulling. Primarily,
the nulling paradigm has been an effective way of assess-
ing the subjective equiluminant nature of a chromatic
stimulus. However, in addition to operating as a calibra-
Cavanagh, 1983; Cavanagh, Macleod, & Anstis, 1987)
lus structure varies between studies, but the basic idea
combined at some point in the system, then the relative
strengths of the two opposing inputs may be said to be
equal when there is no net percept of motion. If one of
the components is then changed in some characteristic,
such as the axis of color space along which it is modu-
said to represent equal contributions of the two compo-
nents to the percept of drift (Anstis & Cavanagh, 1983;
Cavanagh & Anstis, 1991; Cavanagh et al., 1987).
Cavanagh and Anstis (1991) used a motion-nulling
tribution to the percept of motion. Their stimuli were
the sum of two oppositely moving gratings, one modu-
lated in luminance (fixed at 10% contrast), the other
tribution of color to the percept of motion in the com-
pound stimulus, the luminance contrast in the com-
tively null the motion. The equivalent luminance con-
trast of the variable (test) color/luminance grating was
both nonzero and constant across varying levels of
added luminance contrast, as long as they were below
levels that would swamp the independent chromatic sig-
nal. A potential neurophysiological correlate of this
result (Thiele, Dobkins, & Albright, 1999), however,
reveals different data and will be further detailed in the
Neural Correlate section. Nonetheless, from a behav-
ioral perspective, in this comprehensive set of experi-
ments on both normal and color-deficient subjects,
Cavanagh and Anstis (1991) showed that the contribu-
percept could not be explained by any luminance arti-
fact and appeared to be a genuine chromatic input to
the motion mechanism underlying performance in the
Chichilnisky, Heeger, and Wandell (1993) used a
motion-nulling paradigm similar to that of Cavanagh et
al. (1987) in that each of the components was drifting
ral frequency (2 Hz). The aim of their investigation was
motion promoted from a variety of sources was indeed
198 BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
an issue of luminance signals contributing only to the
a product of the difference between transformation of
thethree cone signalsfor thetaskofchromatic discrimi-
nation compared to the task of motion discrimination
within the same primary (cone) color space
concluded that motion discrimination was not confined
tostimulusmodulated alongtheluminancedirection of
color space, thus rejecting motion monochromacy and
therefore functional segregation on a color/motion
Aslightlymodified version ofthemotion-nulling par-
adigm was adopted by Webster and Mollon (1997).
of the stimulus components was changed. The construc-
than each component carrying a distinct motion signal,
neither component alone elicited a directional
response. The stimulus was constructed such that there
was an imbalance only in the temporal modulation,
resulting in a net drift, if the two components contained
a common signal that combined to produce motion.
The stimulus composition places the phases of the two
components, each a counterphasing grating, in spatial
and temporal quadrature (Cavanagh et al., 1987; Lu
et al., 1999; Webster & Mollon, 1997). The results sug-
gest that signals are combined within the three cardinal
combining luminance and chromatic stimuli whatever
diate chromatic axes, there was a contrast-dependent
effect such that two colors could elicit a motion percept
if each of their cardinal components were appropriately
at a given relative contrast (equated in threshold multi-
ples) led to the conclusion that color and luminance
motion signals were qualitatively different from each
other (Webster & Mollon, 1997).
Reconciling the Motion-Nulling Data
issue of the assumptions underlying what initially
appears to be only a small difference in paradigm. The
major difference between the studies is highlighted by
the fact that in one (Cavanagh & Anstis, 1991), each
component of the compound stimulus elicits a percept
of motion in isolation; however, in the second study
(Webster & Mollon, 1997), neither component alone is
sufficient to give a directional response. This raises the
different processes. The stimulus in which each compo-
nent elicits a percept of motion alone (Cavanagh &
Anstis, 1991) is likely to be examining the interaction of
two motion signals, each of which may be considered to
be the output of a motion-detection mechanism. In the
cited case, one “mechanism” is sensitive to luminance
modulation; the other is sensitive to purely chromatic
modulation. The second paradigm (Webster & Mollon,
1997) examines the input to a single mechanism
because the two components need to be combined to
create a motion percept. Neither component alone will
give a directional response. In this second case, the two
filters providing an input to any motion mechanism
must either be sensitive to both chromatic and lumi-
or must be mismatched in their sensitivity (see also
Derrington & Badcock, 1985) to elicit a percept of
Combining Input Attributes to Attain
A body of work that occupies an unclear position
within that discussed so far is that dealing with the per-
ception of motion in a stimulus that historically would
have been considered to be long range (Agonie &
Gorea, 1993; Gorea, Lorenceau, Bagot, & Papathomas,
1992; Gorea & Papathomas, 1989, 1991, 1993; Gorea,
Papathomas, & Kovacs, 1993; Papathomas & Gorea,
1991). However, as pointed out by the authors of these
into a higher level of motion; examination of the work
across the individual articles reveals a story consistent
with several of the studies cited thus far.
The stimuli employed were a series of bars varying in
color, luminance, or orientation (or some combination
of these three) from their background. The displace-
ment of the bars between frames of the display is half of
a display that is drift balanced as a whole (Chubb &
Sperling, 1988) because there is no consistent motion
signalifallthebarswere coded byexactlythesamecom-
bination of attributes. Such a stimulus allows the exami-
sen coding properties of the stimulus, in conjunction
with the pattern of displacement of the stimulus
sequence as a whole (Gorea & Papathomas, 1989).
Gorea andhiscolleagues (Agonie&Gorea, 1993;Gorea
et al., 1992; Gorea et al., 1993; Gorea & Papathomas,
1989, 1991, 1993; Papathomas & Gorea, 1991) found
that motion was perceived both within and between
attributes (see also Cavanagh, Arguin, & von Grünau,
1989) but that the temporal parameters of a spatially
consistent stimulus controlled the relative weight of the
elicit a motion percept in an otherwise drift-balanced
display. Increasing the velocity of the drift makes the
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 199
motion percept more specific to a given attribute, but at
the higher rates of drift (7.5 degrees/s), both color and
luminance contribute to the motion detected, although
sizes showing a fall off for chromatic correspondence–
mediated motion perception. At smaller displacements
& Albright, 1993b), luminance and chromatic-coded
the form of a contrast-polarity reversal) does not exces-
sively disrupt the chromatic correspondence, and the
lent stimuli, possibly indicating the role of a higher level
mechanism in the task.
The velocity and displacement dependence of
motion based on this chromatic correspondence agrees
well with studies addressing these issues with stimuli
designed to tap into low-level mechanisms (Dobkins &
Albright, 1993b; Gegenfurtner & Hawken, 1995;
Hawken, Gegenfurtner, & Tang, 1994). The analogy
drawn with texture segregation and the ability to inte-
grate elements across different coding attributes seem
also to imply a higher level process more generally con-
sidered tobe mediating motion detection inthiskind of
stimuli. The stimuli used by Gorea and colleagues elicit
behavior associated with both sources of motion input,
tent approach to the problem from both a theoretical
and a structural stimulus-based point of view.
Finally, an alternative method of examining the
nature of the underlying mechanisms involved in
motion perception is to use an oscillating stimulus and
sure a just-detectable oscillation (Krauskopf & Li, 1999;
Seiffert & Cavanagh, 1998, 1999).
First, two studies used oscillating stimuli to examine
whether the mechanism underlying motion perception
in different types of stimuli were sensitive to velocity (or
temporal frequency) or to the displacement of the stim-
a stimulus can be independently manipulated by chang-
ing the oscillation frequency of a stimulus of given spa-
tial period. Changing the frequency of oscillation (in
hertz) for a given spatial frequency of, in this case, a
radial grating will increase the (peak) velocity of the
stimulus for a given constant displacement (Figure 1 in
of using such stimuli as opposed to unidirectional stim-
location (for a given section of the stimulus) over a
reasonable range of stimulus parameters.
The conclusions drawn from this first pair of studies
fellrelatively well intostepwithotherwork.Itwasshown
ency on position change, whereas for luminance grat-
ings, a dependency on velocity was shown. However,
once the chromatic contrast rose above approximately
40 to 80 times the detection threshold, then the motion
mechanism dealing with high-contrast chromatic stim-
uli showed velocity dependence. The change from one
characteristic to the other was also relatively even and
gradual, indicating the operation of both types of mech-
the oscillation was added to a base drift speed and the
observers had to perform the same task. The chromatic
data were also shown to be very similar to texture-coded
motion at low contrast in agreement with previous work
(Cropper, 1994; Cropper & Derrington, 1994).
Adopting a slightly different approach, an oscillating
stimulus was used to assess whether chromatic motion
mechanisms used absolute retinal motion as a primary
input, as we assume is the case for basic luminance-
coded motion, or whether they primarily used relative
position sense (Krauskopf & Li, 1999). The stimuli were
4-degree Gabor patches (giving an envelope frequency
of approximately 0.25 c/degree), with a carrier of 1 c/
degree, the position of which was oscillated sinusoidally
at 1 Hz. In one condition, the Gabor carrier was oscil-
lated rigidly (retinal motion); in the second, only the
lower half of a horizontally bisected carrier was moved
(relative motion). The oscillation amplitude depend-
ence of the ability of observers to discriminate the
motion was measured as a function of carrier (peak)
contrast. When the task was to detect absolute retinal
motion and the stimulus was defined by a chromatic
modulation only, performance was contrast dependent.
Performance for luminance-defined absolute motion
was, on the other hand, contrast independent. The
detection of relative motion was shown to be contrast
dependent for both kinds of stimuli. Thus,the principal
result was the difference in behavior for chromatic and
luminance carriers, as has been a common thread
throughout the work.
Overall Summary of the Motion
The majority of the studies on motion discrimination
suggest that there is genuine chromatic input to motion
processing and that the luminance and the chromatic
input are processed by separate mechanisms. The rela-
tive weight of the chromatic mechanism (in relation to
the luminance mechanism) is different for detection
200 BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
and motion discrimination tasks and depends on the
temporal and spatial stimulus parameters and to some
extent on the specific task (e.g., relative motion vs. abso-
lute motion judgment). Whether chromatic motion is
computed by extracting the spatiotemporal energy or
sial issue, but we conclude that there are conditions
under which a low-level chromatic motion mechanism
operates. We also suggest that this low-level input to
chromatic motion probably has two components, a
purely chromatic component and a mixed luminance
and color component. We will return to this in the
Neural Correlate section and the Speculations section.
THE ROLE OF COLOR IN THE
PERCEPTION OF SPEED
The estimation of the velocity (i.e., direction and
speed) of a particular attribute or feature in an image is
vital for many visual tasks, such as extracting three-
dimensional structure from motion, signaling depth
ofmoving chromatic stimuli wasone ofthefirstobserva-
tions that led to the hypothesis that chromatic motion
processing is qualitatively different from luminance-
defined motion. Numerous studies have corroborated
this basic finding that the perceived velocity of chro-
for luminance-defined motion. Although perceived
speed can be thought of as being mediated by a higher
level process, and at some stage the interdependence of
contrast and velocity in relation to the mechanism tun-
ply considered as the accurate assessment of the ampli-
tude of a motion signal (Johnston, McOwan, & Benton,
All of the studies on perceived speed that we have
reviewed reported that the perceived speed for chro-
ceived for luminance-defined stimuli (Boulton &
Mullen, 1990; Burr, Fiorentini, & Morrone, 1998;
Cavanagh & Anstis, 1991; Cavanagh et al., 1984;
Cavanagh & Favreau, 1985; Cavanagh, Saida, & Rivest,
1995; Dougherty, Press, & Wandell, 1999; Derrington,
2000; Farell, 1999; Gegenfurtner & Hawken, 1996;
Hawken et al., 1994; Henning & Derrington, 1994; Kooi
Moreland, 1980; Mullen & Boulton, 1992a, 1992b;
Troscianko & Fahle, 1988).
InFigures 3and4,wereplot datafromseveral studies
comparing the perceived speed of luminance gratings
with the perceived speed of chromatic gratings. The
lus (LUM; Figure 2), and the comparison was always a
chromatic grating, with (LUM+RG) or without (RG)
luminance contrast. The data plotted in Figure 4 show
independent of the spatial frequency of the gratings.
The relative speed of the chromatic grating varies
between 0.1 (indicating a large reduction in perceived
speed) and 1.0 (indicating no reduction in perceived
speed). The contrast was either matched in terms of
detection thresholds (Farell, 1999; Henning &
Derrington, 1994) or the chromatic contrast was con-
ing varied (Boulton & Mullen, 1990; Cavanagh et al.,
(1999), using a rating procedure, reported similar
results: A high-contrast equiluminant red-green grating
an 8% luminance grating moving at 4 degrees/s. Lu
et al. found conditions of “motion standstill,” and
together with their other findings (see the Grating
chromatic input to perceived speed estimation but that
chromatic motion is mediated by a feature-based system
rather than a low-level motion mechanism: They claim
that if the feature-based motion system is silenced, chro-
matic motion cannot be perceived. Several other
researchers have reported phenomena similar to
motion standstill (Cavanagh et al., 1984). Mullen and
Boulton (1992a) observed that equiluminant stimuli
appear to be displaced rather than moving smoothly.
ers’ ability to discriminate sampled motion depends on
the presence of temporal replicas induced by the sam-
pling regime (Cropper & Badcock, 1994). However, an
important caveat to the observation that moving chro-
matic stimuli appear static is that the measure is always
made relative to a much stronger luminance-based
motion percept, either by adjusting the stimulus from
luminance to color coding or by providing a luminance
The presence or absence of an artifact, or the use of
the task (Stromeyer, Chaparro, Tolias, & Kronauer,
1994), is harder to establish for perceived speed mea-
sures thanit is for direction measures. Ittherefore tends
to be more critical to examine the methods of removing
any (detectable) luminance artifact due to optical fac-
tors and the setting of subjective equiluminance with
perceived speed and speed discrimination studies. The
presence of luminance signals in the nominally chro-
matic stimulus must always be considered, and to this
end, most studies in thisreview have used low spatial fre-
quencies (<2 cpd) to minimize luminance artifacts due
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI201
gratings are less affected by chromatic aberration, trans-
verse chromatic aberration induces still a significant
cies as low as 1 cpd (Cavanagh & Anstis, 1991; Faubert,
Bilodeau, & Simonet, 2000). Based on their calculation
that only gratings below 0.5 cpd are free of significant
transverse chromatic aberration. When scaled to detec-
tively high stimulus contrasts because the disparity in
detection threshold is so high between luminance and
chromatic stimuli at low spatial and temporal frequen-
cies (Chaparro, Stromeyer, Huang, Kronauer, & Eskew,
1993; Cropper & Derrington, 1994; Switkes & Crognale,
In addition to using low spatial frequencies, all the
studies on perceived speed have determined the
equiluminant point for each individual observer using
heterochromatic flicker photometry (Wyszecki & Stiles,
imum perceived speed by measuring the percept as a
independent variable. This minimum is usually found
within 10% of what is expected from heterochromatic
For luminance-defined gratings, the perceived speed
depends on the contrast of the moving stimuli and not
only on the physical speed. This misperception of speed
is evident across a wide range of contrasts (up to 50%),
although the exact conditions of slowed-down motion
for luminance-defined gratings are still not entirely
known (Stone & Thompson, 1992; Thompson, 1982). It
hasbeen suggested thatthereduction inspeed forchro-
matic gratings is due to the low effective contrast of the
chromatic gratings (Troscianko & Fahle, 1988) and that
there are no qualitative differences in speed processing
for chromatic and luminance-defined stimuli.
It is clear from Figure 4 that the contrast gain for per-
ceived speed is different for chromatic and luminance
gratings. For low speeds (<1.2 degree/s; see Figure 4a),
the slope relating perceived speed to contrast is much
higher for gratings that contain luminance and chro-
matic information when compared to the slope of pure
luminance gratings.Forhighspeeds (4.8-10degrees/s),
the contrast gains are almost identical (Figure 4c). For
the intermediate speeds (1.6 degrees/s), the slope for
chromatic gratings is intermediate (slope = 2.12; Figure
4b). All these matches were performed for low spatial
frequencies (0.8-1.05 cpd). These data are consistent
with more recent investigations of the gain functions at
low and high speeds. For chromatic stimuli, the contrast
lar for both kinds of gratings (Burr et al., 1998;
Gegenfurtner & Hawken, 1996; Hawken et al., 1994;
Krauskopf & Li, 1999).
Speed matches for stimuli containing both lumi-
nance and chromatic contrast reveal a similar pattern
expressed as multiples of detection threshold, the con-
trast gain for stimuli of both color and luminance con-
trast is similar to the slope for pure luminance modula-
tions.Stimuli were gratingsof1cpdandmoved atadrift
rate of 1.6 Hz. For these spatial and temporal parame-
to luminance variations, and detection ratios are on the
order of 10:1. Because the authors do not provide the
cone-contrast needed for detection, we can only assert
that the slope relating perceived speed to luminance
202 BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
Relative Perceived Speed of Chromatic Grating
Slope = 0.04 (0.05)
LUM/LUM+RG, Cavanagh et al. (1984)
LUM/LUM+RG, Mullen & Boulton (1992a)
LUM/RG, Henning & Derrington (1994)
LUM/LUM+RG, Farell (1999)
LUM/RG, Farell (1999)
LUM/LUM+RG, Mullen & Boulton (1992b)
Spatial Frequency (cpd)
Figure 3: Perceived Speed of the Chromatic Grating as a Function of
Spatial Frequency of Both Test and Comparison Stimuli.
NOTE: The linear regression slope indicates the effect of spatial fre-
quency on the perceived speed. The key indicates the standard/test
compared to a mixed luminance/chromatic contrast test. Different
symbols indicate different studies as outlined fully in the text.
contrast is shallower for the luminance-defined gratings
than for the gratings of both luminance and color con-
trast. Again, this result is consistent with the findings
shown in Figure 4.
From a more general perspective, when a chromatic
grating is added to a luminance grating, the compound
grating moves slower than the luminance grating in iso-
lation (Boulton & Mullen, 1990; Cavanagh & Anstis,
1991; Cavanagh et al., 1984; Cavanagh et al., 1995;
Cavanagh & Favreau, 1985; Mullen & Boulton, 1992a,
1992b). When the result is placed in the context of the
potential for separate mechanisms extracting direction
and speed, as well as the potential segregation of color
and motion, the most parsimonious conclusion is that
color does indeed have a significant contribution to
motion processing. Furthermore, this manipulation of
adding color contrast to a constant drifting lumi-
nance stimulus (Cavanagh et al., 1984) is the closest
the potential luminance input to speed perception in a
nominally chromatic stimulus. It is not dependent on
accepting the calibration measures of equiluminance,
and it is very difficult to argue for no chromatic input to
motion perception alongside this result.
One suggestion to explain the slowed chromatic
motion is that it is caused by a miscalibration in the
decoding of velocity (Cavanagh & Anstis, 1991). The
response of a directionally selective neuron does not
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI203
0.000.04 0.080.120.160.20 0.24
0.000.040.08 0.120.16 0.200.24
0.00 0.040.08 0.120.160.20 0.24
LUM slope = 1.0 (0.3)
LUM+RG slope = 3.0 (0.3)
(a) LOW SPEEDS (0.25- 1.2 deg/sec)
LUM/LUM, Cavanagh et al. (1984)
LUM/LUM+RG, Cavanagh et al. (1984)
LUM/LUM, Burr et al. (1998)
LUM/RG, Burr et al. (1998)
Relative Perceived Speed
LUM+RG slope = 2.1 (0.3)
LUM/LUM+RG, Mullen & Boulton (1992)
LUM/LUM+RG, Boulton & Mullen (1990)
(b) MEDIUM SPEEDS (1.6 deg/sec)
LUM = 0.6 (0.5)
LUM+RG = 1.1 (0.5)
LUM/LUM, Cavanagh et al. (1984)
LUM/LUM+RG, Cavanagh et al. (1984)
LUM/LUM, Burr et al. (1998)
LUM/RG, Burr et al. (1998)
(c) HIGH SPEEDS (4.8 - 10 deg/sec)
Relative Perceived Speed
Figure 4: RelativePerceivedSpeedofLuminanceGratingsandEquiluminantRed-GreenGratingsasa FunctionofContrastforLow(a), Medium
(b), and High Speeds (c).
NOTE: Linear regression lines are the slopes of perceived speed as a function of contrast for luminance tests (solid lines) and chromatic tests
ambiguously signal speed but increases both with stimu-
lus velocity and with contrast. To extract these con-
founded variables, the visual system needs another
mechanism to estimate contrast independently. This
independent of stimulus contrast by implementing a
ratio calculation (Thompson, 1982). One possible sub-
strate for the contrast detector are non–directionally
tuned subunits, which then take the place of the low-
temporal frequency tuned units in a ratio model of
velocity normalization (Thompson, 1982). If the
nondirectional units are more sensitive to contrast than
the directional units are, then the speed is underesti-
idea is corroborated by the relationship between detec-
tion and discrimination thresholds for chromatic stim-
uli: For low temporal and spatial frequencies,
nondirectional units are relatively more sensitive to
chromatic than to luminance modulation. A specific
implementation of the ratio model (Metha & Mullen,
1997) assumed thatperceived speed is computed by tak-
ing the ratio between two temporal filters (one station-
ary, one tuned to higher temporal frequencies) and has
been used successfully to predict the contrast
dependence of perceived speed in chromatic and
luminance gratings across a range of speeds and
An alternative explanation for the slowing of chro-
matic motion is that proposed by Derrington and
Badcock (1985), who argued that chromatic motion is
seen via a common pathway shared with luminance
information and that the slowed response is due to
reversed motion responses at and near equiluminance.
This comes about in their scheme because of mis-
constituting the input to a Reichardt-type detector, that
is, +L-M paired with –L+M. Thus, a given motion detec-
tor will signal correct motion for a luminance edge but
may give a reversed signal for an equiluminant edge.
They argue this mismatch is caused through random
pairing of chromatic receptive fields with only the lumi-
nance signature matched. If the overall output of the
ensemble of mixed-input detectors provided some mea-
sure of the stimulus speed, then it is possible that the
the overall output and therefore the perceived speed.
However,theirmodel does notexplain thereduced per-
ceived speed for chromatic motion stimuli processed by
a truly chromatic pathway unless this forms a subset of
these inputs such that chromatically matched inputs are
somehow used independently of the luminance
response. Because the population response may be a
critical factor within the sensory cortical coding
(Georgopoulos, Schwartz, & Kettner, 1986; Wray &
Edelman, 1996; Zohary, 1992) and color and luminance
signals must be disassociated at some point (Lennie &
D’Zmura, 1988), this is still a viable proposal.
Perceived speed and its dependence on color and
ies have compared speed discrimination performance
for luminance and chromatic stimuli (Cropper, 1994;
Wuerger & Morgan, 1997). The critical difference
between the measurement of a perceived speed and
absolute estimate of signal magnitude needs to be
extracted from the stimulus; the latter task simply
requires an estimate of the greater or lesser signal. It is
therefore the case that we cannot simply assume perfor-
mance to be the same for the two tasks in the same way
that we cannot assume direction and speed discrimina-
tion to yield the same results. Speed discrimination
thresholds were assessed with equiluminant chromatic
stimuli and luminance-defined stimuli for spatial fre-
to 4 degrees/s (Cropper, 1994; Wuerger & Morgan,
In Figure 5, Weber fractions are replotted as a func-
tion of the average LM cone contrast (Cropper, 1994;
Wuerger & Morgan, 1997): Luminance (filled symbols)
and equiluminant red-green stimuli (open symbols)
were matched approximately as multiples of detection
threshold. When plotted as a function of LM cone con-
trast, it is evident that speed discrimination for
equiluminant stimuli is contrast dependent and
improves up to a cone contrast of 2% for 1-cpd gratings.
For the 2-cpd gratings, Weber fractions for speed dis-
crimination decrease up to a cone contrast of 10%. On
the other hand, for luminance-defined stimuli, speed
discrimination performance does not exhibit this con-
trast dependence, and the minimum Weber fraction is
reached at detection threshold. It is particularly perti-
nent for the argument contained in this article that
stimuli in the task of speed discrimination are removed
when the contrast scaling is changed from multiples of
detection threshold to LM cone modulation, a point
made clear in Figure 5.
THE INTEGRATION OF CHROMATIC MOTION
SIGNALS: PLAIDS, RANDOM-DOT
KINEMATOGRAMS, GLOBAL MOTION, AND
STRUCTURE FROM MOTION
The integration of motion signals is an essential pro-
cess in the hierarchy of motion detection, and it is this
stage that will combine the output of both low-level and
204BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
high-level motion detection mechanisms to form some
criminating between low-level and high-level motion
However, if one assumes that the principal luminance
input is from low-level motion mechanisms, then com-
parison between luminance and chromatic perfor-
mance in tasks of motion integration can be very
Arguably the simplest two-dimensional pattern is a
plaid, created by adding two gratings of different spatial
orientation. An important early example of work using
plaid patternsbrings outinconsistencies inthehypothe-
although predating any explicit suggestion of such a
modular process. The observations of Wallach (1935)
and the dependence of that percept on the pattern
motion. Wallach observed that when two red/green
square wave gratings were crossed at different orienta-
tions, a composite pattern was created that had alternat-
ing red/green stripes and, at the intersections, gray
checks. When the pattern was static, or appeared to
move rigidly as a two-dimensional structure, the gray
checks persisted in their color, consistent with the com-
posite perceived structure. However, when this pattern
motion broke down, two transparently moving gratings
were perceived to slide over one another, and “nothing
sistent with the complimentary colors at the intersec-
tions disappeared as soon as those intersections became
inconsistent with the overall perceived structure of the
moving pattern (Wuerger, Shapley, & Rubin, 1996, pp.
1362-1363). His observations demonstrated that color,
not computed in isolation; when several interpretations
of the form and color exist and several directions of
motion are possible solutions, the human observer
perceives only certain combinations of form, color, and
Krauskopf and Farell (1990) examined whether the
motion of two gratings modulated along different cardi-
nal axes would be perceived as a coherent plaid pattern.
Ifthe gratings are perpendicular to each other in spatial
orientation, then different-colored gratings do not
cohere, whereas similar-colored gratings do cohere
(Albright, 1991; Krauskopf & Farell, 1990) when modu-
lation is along the cardinal axes. Greater coherence was
seen if the components were modulated between cardi-
nal axes. With these interaxis grating components, the
cardinal projection of the combined pattern is itself a
ported by a fair amount of subsequent work (Dobkins,
Valois, 1992; Kooi, De Valois, Switkes, & Grosof, 1992;
Krauskopf, Wu, & Farell, 1996).
This initial result was then expanded to examine its
generality and whether the relationship of the two com-
This latter piece of work used the perceived direction of
the pattern as a measure of the interaction rather than
the perceived coherence and used both Type I and Type
II plaids. It was found that if the area of interaction
spatial angle between the two, then the plaid elicited a
greater pattern motion percept (taken to be indicative
of coherence between the components) when those
nents was modulated along the luminance axis (see also
Webster & Mollon, 1997). Both Type I and Type II chro-
matic plaids elicited perceived directions equivalent to
their luminance counterparts (Ferrera & Wilson, 1990,
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 205
LUM/LUM, Cropper (1994), 1 cpd
RG/RG, Cropper (1994), 1 cpd
LUM/LUM, Wuerger & Morgan (1997), 2 cpd
RG/RG, Wuerger & Morgan (1997), 2 cpd
LM Cone Contrast
Figure 5: Speed Discrimination for Luminance and Equiluminant L-M
Stimuli as a Function of LM Cone Contrast.
NOTE: The Weber fraction measured in a velocity discrimination task
stimulus. All stimuli are at detection threshold at the lowest plotted
arate the first-order and second-order components in
their percept and response depending on which aspect
of the stimulus they were instructed to attend. This
clearly relates well to the effect of color on attentionally
modulated motion perception (Cavanagh, 1992;
Culham & Cavanagh, 1994).
a Type II plaid, then the perceived direction of that pat-
tern changes with exposure duration in a similar way to
Banton, 1993) given the limitations imposed on short
presentation duration and chromatic stimuli (Cropper,
Badcock, & Hayes, 1994; Cropper & Derrington, 1994).
When the contrast threshold for the detection of direc-
tion of motion is measured in a plaid, then the perfor-
mance can be predicted accurately from that of the
components alone when the plaid is made from lumi-
nance or from fast-moving (>4 Hz) chromatic compo-
nents (Gegenfurtner, 1998). This is an extension of the
detection:discrimination threshold paradigm (discrimi-
nation of component:discrimination of pattern motion),
and the data show again that slow-moving (<4 Hz)
equiluminant patterns appear to be different and possi-
bly do not fit into the apparent two-stage structure of
plaid motion discrimination. Combined with the cross-
axis data (Cropper et al., 1996) this implies that the two-
when the components are chromatic, possibly because
rather than some fundamental structural difference.
This conclusion is supported by the similarities between
chromatic and luminance motion when assessed by a
secondary measure such as an oculomotor response
evidence against fundamentally different processes for
chromatic and luminance motion.
ties (Braddick, 1980; Ramachandran & Gregory, 1978),
lated components in a composite pattern (Newsome &
Pare, 1988). The observer’s task is to indicate a single
direction alone. Thus, the system must integrate local-
ized motion vectors to give a single dominant direction.
The independent variable is usually the number of local
elements moving in the same direction in any pair of
frames, that is, the motion coherence. This has become
known as “global motion,” and the dependence of the
global percept and motion integration on chromatic
dot kinematograms (Britten, 1999; Croner & Albright,
1997; Edwards & Badcock, 1996; Moller & Hurlbert,
1997a, 1997b; Ramachandran, 1987; Ramachandran &
Gregory, 1978; Ruppertsberg, Wuerger, & Bertamini,
2003; Snowden & Edmunds, 1999). The studies were
nance and the chromatic motion signals and the role of
chromatic information in aiding the segmentation of
Several studies found that when motion coherence is
correlated with the color of the dot, the coherence
threshold for motion detection is reduced.
Edwards and Badcock (1996) employed the same
sparse random-dot kinematograms to investigate the
interactions between the chromatic and luminance sig-
able was, as in all of the above studies, the percentage of
coherently moving dots that are required to reliably
judge the direction of motion. Among other findings,
they report that adding a chromatic motion signal to
vation. This finding provides another piece of evidence
against the hypothesis that chromatic motion acts like
low-contrast luminance motion. If the model arrange-
actually also providing the basic single motion-vector
input to the global-motion sensitive mechanisms, the
chromatic mismatch of receptive field input could
explain the negative effect of chromatic motion vectors
on luminance-coded global-motion extraction.
Chromatic signals as a direct input for global motion
perception havebeen studied onlyrecently (Bilodeau &
Faubert, 1999; Ruppertsberg et al., 2003). Bilodeau and
Faubert (1999)showed thatwhenstimuli are equated in
an L and M cone contrast, the relative coherence
required to identify the direction of motion was on aver-
age twice as high for the equiluminant stimuli. In some
cases, the observer could not correctly identify the
motion even for a coherence level of 100%.
matic input to global motion for different color direc-
tions in the equiluminant color plane and showed that
no S cone input for the set of spatial parameters used in
their experiment. Together, both studies suggest that
when the integration of local motion signals is required,
which is consistent with the conclusions drawn regard-
ing the extraction of structure from motion.
Detailed studies measuring theability todiscriminate
direction in a dense random-dot kinematogram
(Barbur,2004; Barbur,Harlow,& Fahle, 2001) provide a
link between studies using the more traditional stimuli
206BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
work using Gabor micropatterns (Baker et al., 1998;
Yoshizawa et al., 2000) and the work using sparse-dot
was configured such that luminance-defined and chro-
matically defined motion could be independently sig-
naled by elements of the display. Configuring the stimu-
lus as incoherent luminance noise and coherent
chromatic motion showed that movement defined by
ments defining the motion were moving consistently, as
would a bar or an edge. When the elements were inco-
herent or distributed across the display, there was little
apparent size dependency of the elements coding the
plaid data and the suggestion that color information is
critical to the structural qualities of the image.
Structure From Motion and Motion Parallax
A task in which local motion speed signals need to be
integrated to extract the three-dimensional properties
(SFM). Differential velocities in a moving stimulus give
rise to the kinetic depth effect (Wallach & O’Connell,
1953). These differential retinal velocities can be gener-
ated either by a head movement and a stationary stimu-
lus (motion parallax) or by a moving stimulus and a
stationary observer (SFM).
The three studies (Cavanagh et al., 1995; Li & King-
dom, 1998; Wuerger & Landy, 1993) that investigated
the chromatic input to depth from motion found a sig-
nificant color contribution. In the structure-from-
motion task, purely chromatic stimuli always yielded
above-chance performance, and performance
dom, 1998; Wuerger & Landy, 1993). The stimulus in
one of the SFM tasks (Wuerger & Landy, 1993) was a 6 ?
6–degree field of small blurred blobs with a cutoff fre-
quency at2cpd,moving ataspeed rangingfrom 0.6to4
degrees/s with an average speed of about 1.6 degrees/s.
The contrast dependence in the structure-from-motion
task differs for chromatic (open diamonds in Figure 6a)
and achromatic stimuli (solid diamonds). In Figure 6,
the relative percentage correct in another structure-
indicated by open circles. These stimuli contained lumi-
nance and chromatic contrast, and the performance is
contrast dependence for this stimulus is very similar to
the contrast dependence for the purely chromatic
stimulus (Wuerger & Landy, 1993).
A similar pattern is found in the motion parallax task
(Cavanagh et al., 1995). Stimuli were vertical sine wave
gratings, and the task of the observers was to report the
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI207
0.00 0.040.08 0.120.16 0.20
0.00 0.040.08 0.120.16 0.20
LUM, Cavanagh et al (1995)
LUM+RG, Cavanagh et al (1995) 1.3
Relative Perceived Depth
LUM, Wuerger & Landy (1993)
RG, Wuerger & Landy (1993)
LUM+RG, Li & Kingdom (1998)
Performance in SFM Task
LM Cone Contrast
NOTE: (a) Relative performance in a structure-from-motion task as a
function of the LM cone contrast. (b) Relative perceived depths from
parallax motion for luminance and red-green stimuli as a function of
perceived depth. Figure 6b shows the replotted data
from Cavanagh et al. (1995): At high luminance con-
depth = 1 cm). Cavanagh et al. also assessed perceived
relating perceived velocity to luminance contrast is
steeper for luminance-defined targets than for targets
ceived depth and perceived velocity are highly corre-
lated,suggesting thatsimilar mechanisms underlie both
tasks or at least thatthe primary input is an indication of
the speed or temporal drift frequency.
At low-luminance contrast, adding a chromatic grat-
ing to a luminance grating increases perceived depth in
the motion parallax task. This is illustrated in Figure 6b:
chromatic grating to a luminance grating has the oppo-
site effect; that is, the perceived depth is decreased. We
that the contrast gain is much higher (by a factor of 2.5)
for luminance stimuli when compared with stimuli that
contain chromatic or both luminance and chromatic
information is inconsistent with the hypothesis that
chromatic stimuli act like low-contrast luminance
There is clearly a strong link between velocity dis-
crimination and the extraction of an apparent three-
dimensional structure purely from motion information.
Thestructure-from-motion taskrequires theobserverto
discriminate small velocity differences to identify the
correct three-dimensional shape.
In one of the SFM tasks (Wuerger & Landy, 1993) the
speeds vary from 0.6 to 4 degrees/s, with an average
cpd. The velocity discrimination task (Wuerger & Mor-
degrees/s) and similar spatial frequency (2 cpd grat-
ings). Hence, we would expect to find similar contrast
dependence in the velocity discrimination and the SFM
task. Comparing Figure 5 with Figure 6a reveals that the
contrast dependence for chromatic and luminance-
defined stimuli is different between the two tasks.Speed
discrimination performance for luminance-defined
stimuli asymptotes at 2% LM cone contrast, whereas a
ination improves for LM cone contrasts up to 0.1 and
similarly, SFM performance improves at least up to 12%
LM cone contrast (which is the largest contrast avail-
luminance-defined stimuli in the two tasks suggests that
the integration of local speed signals is different for
chromatic and luminance stimuli. The major difference
between the two tasks is that the SFM task requires the
rapid integration of local speed signals across space,
whereas in the speed discrimination task, only a single
speed signal needs to be processed. The luminance
mechanism seems to be integrating more effectively
across space than the chromatic red-green mechanism
does: At 12% cone contrast, speed discrimination per-
formance is very similar for luminance and chromatic
stimuli, but the SFM performance for luminance-
defined stimuli is greater by a factor or 4 compared to
the chromatic SFM performance. Comparison of the
ilar point: The luminance versus chromatic contrast
dependence in these two tasks is reversed, suggesting
that a different mechanism is involved in the structure-
from-motion task. The steep slope observed for the
REACTION TIMES AND RESPONSE LATENCIES
FOR MOTION IDENTIFICATION
Reaction Times and Latencies
Several studies have investigated reaction times for
luminance-defined and equiluminant red-green stimuli
(Burr & Corsale, 2001; Burr et al., 1998; Troscianko &
for stimuli moving at about 1 degree/s, the reaction
times required to identify the direction of motion of the
stimuli increase significantly when no luminance con-
trast is present. They also measured reaction times for
vernier displacement and for the onset of stationary tar-
gets. In both cases, the effect of equiluminance on reac-
and they concluded that the absence of luminance con-
trast affects mainly the temporal and not the spatial pro-
cesses. This hypothesis is in accordance with their find-
ings that adding temporal jitter to luminance-defined
gratings yields reaction times identical to the ones for
equiluminant stimuli. However, it has also been shown
that the reduction in perceived speed seen in
equiluminant red-green gratings is not predicted by a
reduction in the perceived temporal and spatial
frequency of the gratings (Henning & Derrington,
Burr et al. (Burr & Corsale, 2001; Burr et al., 1998)
measured reaction times for motion onset in
equiluminant stimuli and found three major character-
istics. First, the reaction times for chromatic and lumi-
nance stimuli differ only for very slow speeds and con-
verge rapidly for speeds larger than1 degree/s. Second,
208BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
for slow speeds (<1 degree/s), the contrast dependence
of the reaction time for chromatic stimuli is much stron-
ger than for luminance stimuli. Finally, reaction times
for chromatic and for luminance stimuli are predicted
by their perceived rather than their physical speed.
These data are consistent with the idea that reaction
times depend on integrating information over time and
that the temporal filters have longer time constants for
chromatic than for luminance stimuli (Troscianko &
Fahle, 1988). The different contrast dependency for
luminance and chromatic stimuli for reaction times is
entirely predicted by the contrast dependence of per-
ger latency for involuntary eye movements to chromatic
stimuli (Guo & Benson, 1999).
The degree to which vision is segregated
preattentionally is a controversial issue. However, it is
apparent that there are some nominal attentional fac-
tors that may influence the perception of motion
(Britten, 1999; Cavanagh, 1991, 1992; Thiele, Rezec, &
Dobkins, 2002), including that of chromatic stimuli.
There is a clear role of attention within the process of
motion detection in that there has to be a link between
the detection of retinal motion and the tracking of a
moving signal of interest. The fundamental difference
of interest because its retinal motion is detected and rel-
evant; once that feature is being tracked, it is the only
aspect of the image that is not moving (given the errors
inherent in the tracking process). It makes sense thatan
attention-modulated motion mechanism bridges this
Cavanagh (1992) concluded that the attentive track-
ing of motion signals cannot be due to the sole input of
low-level motion mechanisms because chromatic grat-
ings that are powerful stimuli for tracking do not neces-
sarily have a significant input for the computation of
motion perception. Furthermore, for the attentive
motion-tracking task, the chromatic input is stronger
the relative efficacy of color and luminance signals for
attentive tracking depends on the motion task. In a
global motion task that assessed motion capture, the
color and luminance contrast necessary for attentive
tracking was identical to the contrast that was required
for motion capture (Culham & Cavanagh, 1994).
a drifting chromatic grating (see the Motion Nulling
Section) in the presence and absence of attention
revealed surprisingly little effect of attention on the
strength of the percept of purely chromatic motion
added in to the chromatic component of the stimulus,
however, a strong effect of withdrawing attention was
noted. Although this result implies that chromatic
motion must, in part, be mediated by low-level motion
mechanisms largely unaffected by moderate attentional
load, the authors interpret their data in the context of a
residual chromatic signal in the M retino-cortical path-
way so the mechanism was not considered to be purely
chromatic in its nature (Dobkins & Albright, 1993b,
1994, 2004; Thiele et al., 1999). This, then, begs the
broader question of what the neural basis is of our
chromatic motion perception.
THE NEURAL CORRELATE?
As with many apparently simple aspects of vision, the
precise neural mechanism of chromatic motion percep-
tion remains unknown. This, however, is not surprising
given that the exact mechanism of luminance-coded
motion is not known either. Even the neural correlate
(i.e.,theneural response correlated toabehavioral abil-
ity but not necessarily the causal mechanism) of the per-
cept of chromatic motion is relatively obscure, although
a few studies have gone some way toward its definition.
a few studies measuring the output of single neurons to
chromatic moving stimuli (Barberini, Cohen, Wandell,
& Newsome, 2005; Dobkins & Albright, 1994; Dobkins
et al., 1998; Gegenfurtner et al., 1994; Gegenfurtner,
Kiper, & Levitt, 1997; Hawken & Gegenfurtner, 1999;
Horwitz & Albright, 2005; Saito, Tanaka, Isono, Yasuda,
& Mikami, 1989; Seidemann, Poirson, Wandell, &
Newsome, 1999; Thiele et al., 1999; Thiele, Dobkins, &
Albright, 2001). What data do exist are somewhat equiv-
ocal as to whether there are any neurons that are suffi-
ciently directionally and chromatically selective to
explain our ability to discriminate chromatic motion.
Furthermore, the result seems to depend on the precise
stimulus employed, even more so than the
psychophysics, and this compounds the obvious prob-
lems inherent in crossing species and methodology.
Largely, the problem has been the presence of qualita-
tive support for the sensitivity of extrastriate neurons to
the motion of chromatic stimuli, mainly in MT, but the
lack of quantitative data to support the mediation of the
far from a problem isolated to the study of chromatic
Similarly, there have been few neural imaging studies
addressing the same question (Claeys, Lindsey, De
Schutter, & Orban, 2003; Wandell et al., 1999). Having
said this, there is some work that has specifically com-
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 209
significant light on the nature of the mechanism of
purely chromatic motion perception. It is these studies
on which we will focus below.
The signed/unsigned border stimuli developed by
effect of combining luminance and chromatic modula-
ulus ambiguous (see the Nature of the Chromatic
Motion Mechanism section). As such, they provided an
response arisinginmagnocellular neuronsofthelateral
were also used as the stimulus for individually recorded
and showed some strong agreement with the psycho-
physics. It was found that the output of MT neurons
could qualitatively explain the behavioral observations
previously made by showing a directionally selective
response to both signed and unsigned border displace-
ment and that which direction was signaled depended
on the luminance contrasts of the combined stimulus in
a way similar to that measured behaviorally (Dobkins &
Albright, 1993b). This result showed that although the
luminance-defined border was clearly dominant in the
directional signal of an MT neuron, the same neuron
could potentially signal the direction of motion of an
unsignedchromaticborder intheabsence ofdetectable
luminance variation, when displacements were small.
Furthermore, some neurons showed response proper-
ties consistent with a signed-border discrimination at a
displacement at which the unsigned border elicited an
ambiguous signal, that is, a 90-degree phase shift. Thus,
be the likely location of the neural substrate for chro-
further studies comparing psychophysical and
neurophysiological data in the rhesus and macaque
monkey (Thiele et al., 1999, 2001). The principal result
supported the earlier work by showing that MT neurons
sufficient luminance modulation, but the studies also
made the important point that quantitative agreement
between behavioral and neurophysiological data was
dependent on both the task and the species involved, as
one may intuitively expect. It was also shown that the
the output from a directionally selective neuron and by
the psychophysical response of the monkeys, decreased
with increasing luminance contrast in the mixed color-
luminance grating. This result is consistent with the pre-
diction generated by a residual chromatic response in
the largely luminance-sensitive M-pathway (Cavanagh &
Anstis, 1991; Dobkins, 2005; Dobkins & Albright, 2004;
Thiele et al., 1999) but still inconsistent with other
behavioral measures of equivalent luminance contrast
and purely chromatic motion detection (Cavanagh &
Anstis, 1991; Cavanagh et al., 1998; Cropper, 2005a,
2005b; Cropper & Derrington, 1996). One of these
results (Cropper & Derrington, 1996) showed proper-
ties of both a mixed color-luminance signal and a purely
chromatic signal, through the residual phase depend-
ence of the luminance mask. This result supports the
existence of both types of low-level input to the motion
percept. However, it was suggested that this chromatic
motion signal measured in MT was due to a mixed chro-
to cortex and so although low level in its nature, not
purely chromatic in its properties (Dobkins, 2005;
Dobkins & Albright, 1993b, 1994, 2004; Dobkins et al.,
1998; Thiele et al., 1999, 2001; Thiele et al., 2002).
A particularly useful trio of studies was performed
specifically to link the data from the three complemen-
tary methodologies of psychophysics, neurophysiology,
and imaging (Dougherty et al., 1999; Seidemann et al.,
1999; Wandell et al., 1999). Performance in a speed dis-
crimination task was compared for an S-cone isolating
stimulus and a luminance (L+M) modulating stimulus,
and remarkable congruence between each data set was
seen across the studies. This was the first time that the
responses of single neurons in macaque MT had been
shown to be sufficient to underlie both the behavioral
response and the blood-oxygenation-level-dependent
ically or luminance modulated. Indeed, it is worth not-
ing that these three studies constitute the few examples
of work comparing data across methodologies for the
same stimulus set for any specific task, and as such, they
work by the same extended group quantitatively ana-
lyzed the S and L/M cone inputs to MT neurons, which
perception, at least in these S-cone-modulating stimuli,
connection from the interlaminar layers of the LGN
(Sincich,Park,Wohlgemuth, &Horton,2004)or viathe
more established pathway through V1. This latter route
may be generated differently for color and luminance
signals because there appear to be few directionally
Albright, 2005) compared to luminance modulation.
This conclusion, however, is based on a particular form
of measurement and analysis (reverse correlation) that
does appear to generate some odd results for chromatic
210 BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
stimuli, which are significantly at odds with the psycho-
physical data (Cottaris & De Valois, 1998; Horwitz &
As a final link back to the psychophysical evidence
that has formed the majority the published work that we
have reviewed, the more traditional examination of
visual deficits has shown that central achromatopsics,
which show severely disrupted cortical color processing,
havecolor motionperception (measuredagaininterms
of equivalent luminance contrast) for high-contrast
stimuli comparable to color-normal subjects and are far
stronger than congenitally L-M deficient subjects
(Cavanagh et al., 1998). This result implies a role for a
purely chromatic motion signal that remains separate
from a purely chromatic signal because the achroma-
topsics do not have a usable chromatic signal and also
over and above the influence of a mixed chromatic and
luminance signal into MT (Cavanagh et al., 1998;
Dobkins & Albright, 1993b, 1994; Thiele et al., 1999,
What limited links exist between the psychophysics
and the neurophysiology suggest that we are able to dis-
criminate the motion of purely chromatic stimuli
through the use of simple, low-level mechanisms sensi-
are not all of one common structure and comprise both
mixed chromatic and luminance-sensitive receptive
shall expand on this in the final section, but first we will
make some interim conclusions.
Our analysis leads us to believe that much of the dis-
agreement between workers in the area of chromatic
motion may well be due to relatively subtle structural
stimulus differences andtheconsequent disparity in the
data sets, rather than fundamental differences in the
underlying neural mechanisms. We reach the following
processing from low-level motion detection to
attentionally modulated motion isolation.
2. Receptive fields for low-level chromatic motion mecha-
nisms are likely to be confined to the fovea and remain
independent of luminance signals when stimuli are
3. There are different motion mechanisms for chromatic
and luminance stimuli; these differences have been re-
vealed in numerous studies by comparing the contrast
dependence for both kinds of stimuli. We think that
these differences become obvious under conditions in
tive or understimulated, rather than nonexistent. It is
also important to note that in many cases, there are
probably both mixed and purely chromatic low-level
signals contributing to the percept of motion.
4. The chromatic input to motion processing depends on
a variety of spatial and temporal parameters as does
luminance-coded motion. Chromatic motion is more
easily and independently detected in foveally located
stimuli of lower spatial frequencies and for low tempo-
ral frequencies. We suggest that one of the commonly
cited pieces of evidence against a chromatic motion
mechanism, the higher contrast ratio for detection
compared to direction discrimination, is a product of
fundamentally different structures involved in the de-
tection of chromatic stimuli rather than the
discrimination of their motion.
5. Comparing performance for single motion signals with
the performance in a global motion tasks suggests that
the luminance mechanism is more effective in rapidly
matic mechanism is.
Finally, we wish to suggest a hypothetical underlying
motion in chromatic and luminance stimuli. We dis-
cussed the evidence for and against a mechanism that is
sensitive to the spatiotemporal orientation of a purely
chromatic modulation or its motion energy (Adelson &
nism was the discrepancy between the contrast required
to discriminate its direction of motion in a chromatic
stimulus. We think that this result is a product of the
underlying structure dealing with the detection of the
stimulus rather than the structure of the mechanisms
discriminating the direction of motion and wish to sug-
gest the following hypothetical structural arrangement
that explains this and many other aspects of the data.
The proposed structure relies on the suggestion that
are mediated by a mechanism of fundamentally differ-
ent properties than a luminance detection mechanism.
The data reviewed above have shown that there is some
evidence foraseparate chromatic detection mechanism
anism (Gegenfurtner & Hawken, 1995; Metha et al.,
1994; Stromeyer et al., 1995; Willis & Anderson, 1998).
underlying detection ofachromatic stimulusisverysen-
sitive when expressed interms ofacone modulation, far
more sensitive than a luminance detection mechanism
at low spatial and temporal stimulus frequencies
(Chaparro et al., 1993; Cropper & Derrington, 1994;
Eskew, Stromeyer, & Kronauer, 1994; Gegenfurtner &
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 211
Hawken, 1995; Metha et al., 1994; Switkes & Crognale,
1999). Furthermore, it has been argued that this detec-
tion mechanism may provide a contrast-normalization
stage if a ratio model of velocity coding were imple-
tion in the velocity of a chromatic stimulus (Cavanagh
et al., 1984; Metha & Mullen, 1997; Thompson, 1982).
of detection:discrimination ratios as the principal data
leveled against the existence of a chromatic motion
Furthermore, there is now growing evidence that the
system may rely on the product of a population of neu-
rons acting together rather than any particular property
(Georgopoulos et al., 1986; Pouget et al., 2000; Wray &
Edelman, 1996; Zohary, 1992). If the nominal detection
mechanisms for chromatic stimuli were actually the
at the level of V1, then several anomalies regarding the
detection and discrimination of luminance and chro-
matic stimuli may be dealt with. Not least of these is the
observation that no single neuron has been found that
in terms of the contrast sensitivity or in terms of the car-
characteristics (Cropper, 1992; Cropper et al., 1996;
& Farell, 1990; Krauskopf, Williams, & Heeley, 1982;
Krauskopf, Williams, Madler, & Brown, 1986) and
neurophysiological properties (Chan, De Valois, &
1978; De Monasterio & Gouras, 1975; De Monasterio,
Gouras, & Tolhurst, 1975; De Valois, Abramov, & Jacobs,
1966; De Valois, Smith, Kitai, & Karoly, 1958;
Derrington, Krauskopf, & Lennie, 1984; De Valois,
Cottaris, Mahon, Elfar, & Wilson, 2000; Gegenfurtner et
al., 1994; Lennie, Krauskopf, & Sclar, 1990; Lennie,
Lankheet, & Krauskopf, 1994; Solomon, Peirce, Dhruv,
& Lennie, 2004; Solomon, Peirce, & Lennie, 2004;
Tootell, Silverman, Hamilton, De Valois, & Switkes,
1988) of which have been extensively studied.
The selective implementation of such a mechanism
for the detection of a chromatic modulation is likely if
one considers the demands that must be placed on a
luminance-coding mechanism by the evolution and
development of a visual system. There is little doubt that
our ability to detect and discriminate luminance modu-
lation is spatially and temporally discrete and relatively
finely tuned in both domains (Campbell, Nachmias, &
Jukes, 1970; Campbell & Robson, 1964, 1966; Watson,
Barlow, & Robson, 1983; Watson & Robson, 1981). It is
also the case that these simple behavioral abilities can
potentially be mediated within the bounds of the prop-
erties ofsingle neurons actingessentially independently
(Blakemore & Campbell, 1969; Derrington & Lennie,
1984; De Valois, Morgan, & Snodderly, 1974; Enroth-
Cugell & Robson, 1966; Harwerth, Boltz, & Smith, 1980;
Robson, 1980). This allows a more simple approach to
based performance is mediated through the properties
of single neurons, albeit arranged in functional groups.
discriminate motion and form effectively and discretely.
This fine spatiotemporal tuning, however, is not a prop-
erty of a population of neurons in which the output is
arrangement very much like the chromatic sensitivity
profile ofV1 simple cells (Lennie etal.,1990)and poten-
spatiotemporal chromatic contrast-sensitivity function
(Mullen, 1985). If such a mechanism were selectively
implemented for chromatic signals, or those carrying
some chromatic response, this might have the effect of
decreasing thedetection threshold for achromatic grat-
maticallytunedneuron (Derrington etal.,1984;Lennie
et al., 1990; Lennie et al., 1994) and altering the appar-
ent behavioral tuning function to one predicted only
(Krauskopf et al., 1982).
The evidence supporting this tentative proposal lies
not only in the lack of a specific neural substrate to
explain behavior but also in several aspects of the anat-
omy and physiology of the retino-cortical stream. First,
there is a specific pattern of expansion of receptive field
numbers per given spatial location such that in V1, the
foveal region is overrepresented in the retino-cortical
map (Azzopardi & Cowey, 1993). The critical aspect of
processing streams expand differentially; the P stream
principally expanding between retina and LGN, the M
stream expanding between LGN and cortex (Azzopardi
& Cowey, 1996; Azzopardi, Jones, & Cowey, 1999). This
opens up the possibility for different mechanisms of sig-
nalcoding to operate veryearly on color andluminance
gestion of different mechanisms for detection of these
Second, there must be some role for the extensive
that this may have a role in our proposed structure. If
these cortico-geniculate efferent connections exert
some control over the firing pattern in the LGN such
that individual outputs remained unchanged in fre-
quency and amplitude of response, but the outputs
212 BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
they may significantly contribute to the overall cortical
response. This suggestion is partially consistent with
recent data that have shown that cooling of the cortex
does have some effect on the output of geniculate neu-
rons and that this effect may be selective for particular
aspects of the receptive field organization
(Przybyszewski, Foote, & Pollen, 2000). If part of this
cortico-geniculate feedback modified the outputs of
parvocellular neurons, which given the anatomical
expansion are more likely to be firing coherently than
the magnocellular units (Azzopardi & Cowey, 1996;
ity of the cortex to use that input signal for the purposes
of chromatic detection, particularly given the peculiar
nature of the threshold response.
Finally, at the level of V1, there is a significant prob-
lem reconciling what we know about the neurophysio-
logical properties of the constituent units and what we
know about the functional and behavioral properties of
the system as a whole. This is exemplified within studies
of color vision by the apparently robust result of three
independent cardinal mechanisms specifically for the
detection of chromatic and luminance stimuli (Krauskopf
et al., 1982), which, despite a great deal of data on the
issue (for a critical analysis, see Krauskopf, 1997, 1999),
the physiological substrate for this most basic of tasks
remains unresolved. If a population-based coding
scheme were adopted to explain the properties of chro-
matic detection, problems such as the absence of any
would also predict that the detection threshold for a
chromatic stimulus would be far lower than the motion
discrimination threshold (if it were to involve low-level
apopulation-code model because ofitsnecessarydefini-
tion in space and time. The only motion signal available
at the chromatic detection threshold would be severely
smeared inspaceandtime,resulting inapoorly defined
tinuous. Furthermore, a property consistent with this
proposed structure is that the contrast at which a chro-
low-level motion mechanism is the threshold thatwould
be predicted from a single neuron. It is also consistent
with the result that the postadaptation contrast thresh-
the same axis of color space is approximately equivalent
to that estimated from the response of a single neuron,
as if the adaptation procedure did not affect individual
outputs but affected only the population response as a
whole (Cropper & Derrington, 1994, 1996; Derrington
1982; Krauskopf et al., 1986; Lennie et al., 1990).
1. This process is actually most accurately achieved when the spec-
tral output from the individual monitor phosphors are known and the
corrected human sensitivity profiles are used (Macleod & Boynton,
1979; Wyszecki & Stiles, 1982).
statistical structure of the modulation of the stimulus in question and
make no assumptions about the underlying detection mechanisms.
fortheperceptionof motion.Journal oftheOptical Society ofAmerica
A, 2, 284-299.
Adelson, E. H., & Movshon, J. A. (1982). Phenomenal coherence of
moving visual patterns. Nature, 300, 523-525.
green drifting stimuli: Dependency on luminance-colour interac-
tions and on the psychophysical task. Journal of the Optical Society of
America A, 10, 1341-1352.
Albright, T. D. (1991). Color and the integration of motion signals.
Trends in Neurosciences, 14, 266-269.
Anstis, S. M. (1970). Phi movement as a subtraction process. Vision
Research, 10, 1411-1430.
Anstis, S. M., & Cavanagh, P. (1983). A minimum motion technique
for judging equiluminance. In J. D. Mollon & L. T. Sharpe (Eds.),
Colour vision: Physiology and psychophysics (pp. 154-166). London:
Azzopardi, P., & Cowey, A. (1993). Preferential representation of the
fovea in the primary visual cortex. Nature, 361, 719-721.
Azzopardi, P., & Cowey, A. (1996). Models of ganglion call topogra-
phy in the retina of macaque monkeys and their application to
sensory cortical scaling. Neuroscience, 73, 617-635.
Azzopardi, P., Jones, K. E., & Cowey, A. (1999). Uneven mapping of
magnocellular and parvocellular projections from the lateral
geniculate nucleus to the striate cortex in the macaque monkey.
Vision Research, 39, 2179-2190.
Baker, C. L., Jr., Boulton, J. C., & Mullen, K. T. (1998). A nonlinear
chromatic motion mechanism. Vision Research, 38, 291-302.
Barberini, C. L., Cohen, M. R., Wandell, B. A., & Newsome, W. T.
(2005). Cone signal interactions in direction-selective neurons in
Barbur, J. L. (2004). “Double-blindsight” revealed though processing
of color and luminance contrast defined motion signals. Progress
in Brain Research, 144, 243-259.
defined motion perception in human vision. Investigative Ophthal-
mology and Visual Science, 42(4 ARVO abstracts).
Bilodeau, L., & Faubert, J. (1999). Global motion cues and the chro-
Blakemore, C., & Campbell, F. W. (1969). On the existence of
neurones in the human visual system selectively sensitive to the
orientation and size of retinal images. Journal of Physiology (Lon-
don), 203, 237-260.
Boulton, J. C. (1987). Two mechanisms for the detection of slow
motion. Journal of the Optical Society of America A, 4, 1634-1642.
Boulton, J. C., & Baker, C. L. (1993a). Different parameters control
motion perception above and below a critical density. Vision
Research, 33, 1803-1811.
Boulton, J. C., & Baker, C. L., Jr. (1993b). Dependence on stimulus
onset asynchrony in apparent motion: Evidence for two mecha-
nisms. Vision Research, 33, 2013-2019.
Boulton, J. C., & Hess, R. F. (1990). Luminance contrast and motion
detection. Vision Research, 30, 175-179.
Boulton, J. C., & Mullen, K. T. (1990). A case in which colour vision is
no longer motion blind. Clinical Vision Science, 5, 175-184.
Braddick, O. (1974). A short-range process in apparent motion.
Vision Research, 14, 519-527.
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 213
motion. Philosophical Transactions of the Royal Society London B, 290,
Breitmeyer, B. (1984). Visual masking: An integrative process. Oxford,
UK: Clarendon Press.
Britten, K. (1999). Motion perception: How are moving images seg-
mented? Current Biology, 9, 728-730.
Burr, D. C., & Corsale, B. (2001). Dependency of reaction times to
motion onset on luminance and chromatic contrast. Vision
Research, 41, 1039-1048.
Burr, D. C., Fiorentini, A., & Morrone, C. (1998). Reaction time to
motion onset of luminance and chromatic gratings is determined
by perceived speed. Vision Research, 38, 3681-3690.
Campbell, F. W., Nachmias, J., & Jukes, J. (1970). Spatial-frequency
ica, 60, 555-559.
sis to the modulation response of the eye [Abstract]. Journal of the
Optical Society of America, 54, 581.
sis to the visibility of gratings. Journal of Physiology (London), 197,
Cavanagh, P. (1991). No slowing for active motion perception at
equiluminance. Investigative Ophthalmology and Visual Science, 32,
Cavanagh, P. (1992). Attention-based motion perception. Science,
in normal and color-deficient observers. Vision Research, 31, 2109-
Cavanagh, P., Arguin, M., & von Grünau, M. (1989). Interattribute
apparent motion. Vision Research, 29, 1197-1204.
Cavanagh, P., & Favreau, O. E. (1985). Colour and luminance share a
common motion pathway. Vision Research, 25, 1592-1601.
Cavanagh, P., Henaff, M.-A., Michel, F., Landis, T., Troscianko, T., &
Intrilligator, J. (1998). Complete sparing of high-contrast color
roscience, 1, 242-247.
Cavanagh, P., MacLeod, D. I. A., & Anstis, S. M. (1987).
Equiluminance: Spatial and temporal factors and the contribu-
Cavanagh, P., & Mather, G. (1989). Motion: The long and short of it.
Spatial Vision, 4, 103-129.
Cavanagh, P., Saida, S., & Rivest, J. (1995). The contribution of color
Cavanagh, P., Tyler, C. W., & Favreau, O. E. (1984). Perceived velocity
of moving chromatic gratings. Journal of the Optical Society of Amer-
ica A, 1, 893-899.
Chan, H., De Valois, K. K., & Switkes, E. (1986). Binocular spatial
interactions of equiluminant colour and luminance patterns.
Investigative Ophthalmology & Visual Science, 27, 75.
Chaparro, A., Stromeyer, C. F., III, Huang, E. P., Kronauer, R. E., &
Eskew, R. T., Jr. (1993). Colour is what the eye sees best. Nature,
Chichilnisky, E.-J., Heeger, D., & Wandell, B. A. (1993). Functional
segregation of color and motion perception examined in motion
nulling. Vision Research, 33, 2113-2125.
Chubb, C., & Sperling, G. (1988). Drift-balanced random stimuli: A
of the Optical Society of America A, 5, 1986-2006.
Claeys, K. G., Lindsey, D. T., De Schutter, E., & Orban, G. T. (2003). A
higher order motion region in human inferior parietal lobule:
Evidence from fMRI. Neuron, 40, 631-642.
Cottaris, N., & De Valois, R. L. (1998). Temporal dynamics of chro-
matic tuning in macaque primary visual cortex. Nature, 395, 896-
Croner, L. J., & Albright, T. D. (1997). Image segmentation enhances
discrimination of motion in visual noise. Vision Research, 37, 1415-
detectors? Unpublished doctoral dissertation, University of New-
castle Upon Tyne, UK.
Cropper, S. J. (1994). Velocity discrimination in chromatic gratings
and beats. Vision Research, 34, 41-48.
Cropper, S. J. (2005a). The detection of motion in chromatic stimuli:
First-order and second-order spatial structure. Vision Research, 45,
Cropper, S. J. (2005b). The detection of motion in chromatic stimuli: Pedes-
tals and masks. Manuscript submitted for publication.
Cropper, S. J., & Badcock, D. R. (1994). Discriminating smooth from
sampled motion: Chromatic and luminance stimuli. Journal of the
Optical Society of America A, 11, 515-530.
Cropper, S. J., Badcock, D. R., & Hayes, A. (1994). On the role of
second-order signals in the perceived direction of motion of type
ii plaid patterns. Vision Research, 34, 2609-2612.
Cropper, S. J., & Derrington, A. M. (1992). Motion after effects in
chromatic stimuli are dependent on stimulus contrast. Perception,
Cropper, S. J., & Derrington, A. M. (1994). Motion of chromatic stim-
uli: First-order or second-order? Vision Research, 34, 49-58.
Cropper, S. J., & Derrington, A. M. (1996). Rapid colour-specific
detection of motion in human vision. Nature, 379, 72-74.
Cropper, S. J., Mullen, K. T., & Badcock, D. R. (1996). Motion coher-
ence across cardinal axes. Vision Research, 36, 2475-2488.
Culham, J. C., & Cavanagh, P. (1994). Motion capture of luminance
stimuli by equiluminous color gratings and by attentive tracking.
Vision Research, 34, 2701-2706.
De Monasterio, F. M. (1978). Center and surround mechanisms of
opponent-color x and y ganglion cells of retina of macaques. Jour-
nal of Neurophysiology, 41, 1418-1434.
De Monasterio, F. M., & Gouras, P. (1975). Functional properties of
ganglion cells of the rhesus monkey retina. Journal of Physiology,
De Monasterio, F. M., Gouras, P., & Tolhurst, D. J. (1975). Trichro-
matic colour opponency in ganglion cells of the rhesus monkey
retina. Journal of Physiology, 251, 197-216.
Derrington, A. M. (2000). Can colour contribute to motion? Current
Biology, 10, 268-270.
Derrington, A. M., &Badcock, D. R. (1985). Thelow level motion sys-
temhas both chromaticandluminanceinputs. VisionResearch, 25,
Derrington, A. M., & Henning, G. B. (1993). Detecting and discrimi-
nating the direction of motion of luminance and colour gratings.
Vision Research, 33, 799-811.
Derrington, A. M., Krauskopf, J., & Lennie, P. (1984). Chromatic
mechanisms in lateral geniculate nucleus of macaque. Journal of
Physiology, 357, 241-265.
sensitivities of neurones in lateralgeniculatenucleus of macaque.
Journal of Physiology (London), 357, 219-240.
De Valois, R. L., Abramov, I., & Jacobs, G. H. (1966). Analysis of
response patterns of LGN cells. Journal of the Optical Society of Amer-
ica A, 56, 966-977.
De Valois, R. L., Cottaris, N. P., Mahon, L. E., Elfar, S. D., & Wilson, J.
A. (2000). Spatial and temporal receptive fields of geniculate and
cortical cells and directional selectivity. Vision Research, 40, 3685-
De Valois, R. L., Morgan, H. C., & Snodderly, D. M. (1974).
Psychophysical studies of monkey vision—iii. Spatial luminance
contrast sensitivity tests of macaque and human observers. Vision
Research, 14, 75-82.
De Valois, R. L., Smith, C. J., Kitai, S. T., & Karoly, A. J. (1958).
Responses of single cells in different layers of the primate lateral
Dobkins, K. R. (2005). Enhanced red/green color input to motion
processing in infancy: Evidence for increasing dissociation of
color and motion information during development. In
Y. Munakata & M. Johnson (Eds.), Attention and performance xxi
(pp. 401-423). Oxford, UK: Oxford University Press.
214BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
Dobkins, K. R., & Albright, T. (1993a). Color, luminance, and the
detection of visual motion. Current Directions in Psychological Sci-
ence, 2, 189-193.
Dobkins, K. R., & Albright, T. (1993b). What happens if it changes
chromatic input to motion detectors. Vision Research, 33, 1019-
Dobkins, K. R., & Albright, T. D. (1994). What happens if it changes
color when it moves? The nature of chromatic input to macaque
area MT. Journal of Neuroscience, 14, 4854-4870.
Dobkins, K. R., &Albright, T. D. (2004). Mergingprocessingstreams:
Color cues for motion detection and interpretation. In
L. Chalupa & J. Werner (Eds.), The visual neurosciences (pp. 1217-
1228). Cambridge, MA: MIT press.
Dobkins, K. R., Stoner, G. S., & Albright, T. D. (1998). Perceptual,
oculomotor, and neural responses to moving color plaids. Percep-
tion, 27, 681-709.
Dougherty, R. F., Press, W. A., & Wandell, B. (1999). Perceived speed
of colored stimuli. Neuron, 24, 893-899.
Edwards, M., & Badcock, D. R. (1996). Global motion perception:
Interaction of chromatic and luminance signals. Vision Research,
Enroth-Cugell, C., & Robson, J. G. (1966). The contrast sensitivity of
Eskew, R. T., Stromeyer, C. F., & Kronauer, R. E. (1994). Temporal
Farell, B. (1999). Color and luminance in the perception of 1- and 2-
dimensional motion. Vision Research, 39, 2633-2647.
Faubert, J., Bilodeau, L., & Simonet, P. (2000). Transverse chromatic
aberration and colour-defined motion. Ophthalmic and Physiologi-
cal Optics, 20, 274-280.
Ferrera, V. P., & Wilson, H. R. (1990). Perceived direction of moving
two-dimensional patterns. Vision Research, 30, 273-287.
Ferrera, V. P., & Wilson, H. R. (1991). Perceived speed of moving two-
dimensional patterns. Vision Research, 31, 877-893.
Freedland, R. L., & Banton, T. (1993). Type ii near-isoluminant
Ophthalmology & Visual Science, 34, 1618.
Gegenfurtner, K. R. (1998). Thresholds for the identification of the
direction of motion of plaid patterns defined by luminance or
chromatic contrast. Vision Research, 38, 881-888.
Gegenfurtner, K. R., & Hawken, M. J. (1995). Temporal and chro-
matic properties of motion mechanisms. Vision Research, 35, 1547-
Gegenfurtner, K. R., & Hawken, M. J. (1996). Perceived velocity of
luminance, chromatic and non-Fourier stimuli: Influence of con-
trast and temporal frequency. Vision Research, 36, 1281-1290.
Gegenfurtner, K. R., Kiper, D. C., Beusmans, J. M. H., Carandini, M.,
Zaidi, Q., & Movshon, J. A. (1994). Chromatic properties of
neurones in MT. Visual Neuroscience, 11, 455-466.
Gegenfurtner, K. R., Kiper, D. C., & Levitt, J. B. (1997). Functional
properties of neurons in macaque area v3. Journal of
Neurophysiology, 77, 1906-1923.
Georgopoulos, A. P., Schwartz, S. B., & Kettner, R. E. (1986).
Neuronal population coding of movement direction. Science, 233,
Gorea, A., Lorenceau, J., Bagot, J.-D., & Papathomas, T. V. (1992).
Sensitivity to colour- and to orientation-carried motion respec-
tively improves and deteriorates under equilumiant background
conditions. Spatial Vision, 6, 285-302.
Gorea, A., & Papathomas, T. V. (1989). Motion processing by chro-
ica A, 6, 590-602.
Gorea, A., & Papthomas, T. V. (1991). Texture segregation by chro-
matic and achromatic visual pathways: An analogy with motion
processing. Journal of the Optical Society of America A, 8, 386-393.
ized concept in texture segregation illustrated with stimuli
defined by color, luminance and orientation. Journal of the Optical
Society of America A, 10, 1450-1462.
Gorea, A., Papathomas, T. V., & Kovacs, I. (1993). Motion perception
with spatiotemporally matched chromatic and achromatic infor-
mation reveals a “slow” and a “fast” motion system. Vision Research,
influences the suppressed ocular following response. Neuroreport,
Harwerth, R. S., Boltz, R. L., & Smith, E. L., III. (1980).
Psychophysical evidence for sustained and transient channels in
the monkey visual system. Vision Research, 20, 15-22.
Hawken, M. J., & Gegenfurtner, K. R. (1999). Interactions between
color and motion in the primate visual system. In K. R.
Gegenfurtner & L. T. Sharpe (Eds.), Colour vision: From genes to per-
ception (pp. 283-299). Cambridge, UK: Cambridge University
Hawken, M. J., Gegenfurtner, K. R., & Tang, C. (1994). Contrast
dependence of colour and luminance motion mechanisms in
human vision. Nature, 367, 268-270.
and temporal frequency comparisons with luminance and colour
gratings. Vision Research, 34, 2093-2101.
Horwitz, G. D., & Albright, T. D. (2005). Paucity of linear motion
detectors in macaque v1. Journal of Vision, 5, 525-533.
Johnston, A., McOwan, P. W., & Benton, C. P. (1999). Robust velocity
computation from a biologically motivated model of motion per-
ception. Proceedings of the Royal Society of London B, 266, 509-518.
Johnston, A., McOwan, P. W., & Buxton, H. (1992). A computational
model for the analysis of some first-order and second-order
motion patterns by simple and complex cells. Proceedings of the
Royal Society of London B, 250, 297-306.
Johnston, A., & Wright, M. J. (1985). Lower thresholds of motion for
gratings as a function of eccentricity and contrast. Vision Research,
Julesz, B. (1971). Foundations of cyclopean perception. Chicago: Univer-
sity of Chicago Press.
Kooi, F. L., & De Valois, K. K. (1992). The role of color in the motion
system. Vision Research, 32, 657-668.
Kooi, F. L., De Valois, K. K., Switkes, E., & Grosof, D. H. (1992).
Higher order factors influencing the perception of sliding and
coherence of a plaid. Perception, 21, 583-598.
Krauskopf, J. (1997). The paucity of evidence for cardinal mecha-
nisms. In C. Dickinson, I. Murray, & D. Carden (Eds.), John Dal-
ton’s colour vision legacy (pp. 431-440). London: Taylor & Francis
Krauskopf, J. (1999). Higher order color mechanisms. In G. K. R. &
L. T. Sharpe (Eds.), Color vision, from genes to perception (pp. 149-
164). Cambridge, UK: Cambridge University Press.
Krauskopf, J., & Farell, B. (1990). Influence of colour on the percep-
tion of coherent motion. Nature, 348, 328-331.
Krauskopf, J., & Li, X. (1999). Effect of contrast on detection of
motion of chromatic and luminance targets: Retina-relative and
object-relative movement. Vision Research, 39, 3346-3350.
Krauskopf, J., Williams, D. R., & Heeley, D. W. (1982). Cardinal direc-
tions of color space. Vision Research, 22, 1123-1131.
Krauskopf, J., Williams, D. R., Madler, M. B., & Brown, A. M. (1986).
Higher order color mechanisms. Vision Research, 26, 23-32.
Krauskopf, J., Wu, H. J., & Farell, B. (1996). Coherence, cardinal
directions and higher-order mechanisms. Vision Research, 36,
tion, 27, 889-935.
Lennie, P., & D’Zmura, M. (1988). Mechanisms of color vision. CRC
Critical Reviews in Neurobiology, 3, 333-400.
Lennie, P., Krauskopf, J., & Sclar, G. (1990). Chromatic mechanisms
in striate cortex of macaque. Journal of Neuroscience, 10, 649-669.
selective habituation in monkey striate cortex. Investigative Oph-
thalmology and Visual Science, 35, 1885.
Levinson, E., & Sekuler, R. (1975). The independence of channels in
human vision selectivefor direction of movement. Journal of Physi-
ology (London), 250, 347-366.
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 215
luminance facilitate the detection of structure-from-motion in
noise? Perception, 27, 769-784.
Lindsey, D. T., & Teller, D. Y. (1990). Motion at isoluminance: Dis-
crimination/detection ratios for moving isoluminant gratings.
Vision Research, 30, 1751-1761.
Lu, Z.-L., Lesmes, L. A., & Sperling, G. (1999). The mechanism of
isoluminant chromatic motion. Proceedings of the National Academy
of Sciences of the U S A, 96, 8289-8294.
Lu, Z.-L., & Sperling, G. (1995). The functional architecture of
human visual motion perception. Vision Research, 35, 2697-2722.
Lu, Z.-L., & Sperling, G. (2001). Three systems theory of human
visualmotion perception: Review and update. Journal ofthe Optical
Society of America A, 18, 2331-2370.
Macleod, D. I. A., & Boynton, R. M. (1979). Chromaticity diagram
showing cone excitation by stimuli of equal luminance. Journal of
the Optical Society of America, 69, 1183-1186.
Marr, D. (1982). Vision: A computational investigationinto the humanrep-
resentation and processing of visual information. San Francisco: W. H.
Freeman and Company.
Metha, A. B., & Mullen, K. T. (1997). Red-green and achromatic tem-
poral filters: A ratio model predicts contrast-dependent speed
perception. Journal of the Optical Society of America A, 14, 984-997.
Metha, A. B., & Mullen, K. T. (1998). Failure of direction discrimina-
tion at detection threshold for both fast and slow chromatic
motion. Journal of the Optical Society of America A, 15, 2945-2950.
Metha, A. B., Vingrys, A. J., & Badcock, D. R. (1994). Detection and
discrimination of moving stimuli: The effects of color, luminance
and eccentricity. Journal of the Optical Society of America A, 11, 1697-
Moller, P., & Hurlbert, A. (1997a). Interactions between colour and
motion in image segmentation. Current Biology, 7, 105-111.
Moller, P., & Hurlbert, A. (1997b). Motion edges and regions guide
don B, 264, 1571-1577.
Moreland, J. D. (1980). Spectral sensitivity measured by motion pho-
tometry. In J. G. Verrist (Ed.), Color deficiencies (Vol. 5, pp. 299-
305). Bristol, UK: Hilger.
Morgan, M. J., & Cleary, R. (1992). Effects of colour substitutions
upon motion detection in spatially random patterns. Vision
Research, 32, 815-821.
if luminance but not color contrast is reversed during displace-
detection. Vision Research, 34, 2527-2535.
red/green and blue/yellow chromatic gratings. Journal of Physiol-
ogy (London), 359, 381-400.
Mullen, K. T. (1987). Spatial influences on colour opponent contri-
butions to pattern detection. Vision Research, 27, 829-840.
Mullen, K. T. (1991). Colour vision as apost-receptoral specialization
of the central visual field. Vision Research, 31, 119-130.
Mullen, K. T., & Baker, C. L. (1985). A motion aftereffect from an
isoluminant stimulus. Vision Research, 25, 685-688.
Mullen, K. T., & Boulton, J. C. (1992a). Absence of smooth motion
perception in color vision. Vision Research, 32, 483-488.
Mullen, K. T., & Boulton, J. C. (1992b). Interactions between colour
and luminance contrast in the perception of motion. Ophthalmic
and Physiological Optics, 12, 201-205.
subthreshold summation between red-green and luminance
mechanisms over a wide range of spatio-temporal conditions.
Vision Research, 37, 1157-1167.
Mullen, K. T., Yoshizawa, T., & Baker, C. L., Jr. (2003). Luminance
based mechanisms mediate the motion perception of
“isoluminant” chromatic gratings: The role of “temporal chro-
matic aberration.” Vision Research, 43, 1235-1247.
Newsome, W. T., & Pare, E. B. (1988). A selective impairment of
motion perception following lesions of the middle temporal
visual area (MT). Journal of Neuroscience, 8, 2201-2211.
Palmer, J., Mobley, L. A., & Teller, D. Y. (1993). Motion at
isoluminance: Discrimination/detection ratios and the summa-
tion of luminance and chromatic signals. Journal of the Optical Soci-
ety of America A, 10, 1353-1362.
Pantle, A. J., & Picciano, L. (1976). A multistable movement display:
Evidence for two separate systems in human vision. Science, 193,
Papathomas, T. V., & Gorea, A. (1991). Two carriers for motion per-
ception: Color and luminance. Vision Research, 31, 1883-1891.
Pelli, D. G. (1981). Effects of visual noise. Cambridge, UK: University of
Pouget, A., Dayan, P., & Zemel, R. (2000). Information processing
with population codes. Nature Reviews Neuroscience, 1, 125-132.
Przybyszewski, A. W., Foote, W., & Pollen, D. A. (2000). Striate cortex
increases contrast gain of macaque LGN neurons. Visual Neurosci-
ence, 17, 485-494.
Ramachandran, V. S. (1987). Interaction between color and motion
in human vision. Nature, 328, 645-647.
Ramachandran, V. S., & Gregory, R. L. (1978). Does colour provide
an input to human motion perception? Nature, 275, 55-56.
vision. In C. S. Harris (Ed.), Visual coding and adaptability (pp. 177-
214). Hillsdale, NJ: Lawrence Erlbaum Associates.
Ruppertsberg, A., Wuerger, S. M., & Bertamini, M. (2003). The chro-
matic selectivity of global motion perception. Visual Neuroscience,
Saito, H., Tanaka, K., Isono, H., Yasuda, M., & Mikami, A. (1989).
Directionally selective response of cells in the middle temporal
area (MT) of the macaque monkey to the movement of
equiluminous stimuli. Experimental Brain Research, 75, 1-14.
tion mechanisms revealed by noise masking in three-dimensional
cone contrast space. Journal of the Optical Society of America A, 14,
Schiller, P. H., & Colby, C. L. (1983). The responses of single cells in
the lateral geniculate nucleus of the rhesus monkey to color and
luminance contrast. Vision Research, 23, 1631-1641.
Seidemann, E., Poirson, A. B., Wandell, B. A., & Newsome, W. T.
(1999). Color signals in area MT of the macaque monkey. Neuron,
Seiffert, A. E., & Cavanagh, P. (1998). Position displacement, not
velocity, is the cue to motion detection of second-order stimuli.
Vision Research, 38, 3569-3582.
Seiffert, A. E., & Cavanagh, P. (1999). Position based motion percep-
tion for color and texture stimuli: Effects of contrast and speed.
Vision Research, 39, 4172-4185.
Sincich, L. C., Park, K. F., Wohlgemuth, M. J., & Horton, J. C. (2004).
ence, 7, 1123-1128.
Snowden, R. J., & Edmunds, R. (1999). Colour and polarity contribu-
Solomon, S. G., Peirce, J. W., Dhruv, N. T., & Lennie, P. (2004). Pro-
found contrast adaptation early in the visual pathway. Neuron, 42,
Solomon, S. G., Peirce, J. W., & Lennie, P. (2004). The impact of sup-
pressive surrounds on chromatic properties of cortical neurons.
Journal of Neuroscience, 24, 148-160.
Stiles, W. S. (1939). The directional sensitivity of the retina and the
spectral sensitivities of the rods and cones. Proceedings of the Royal
Society (London), B, 127, 64-105.
Stiles, W. S. (1959). Color vision: The approach through increment-
the U S A, 45, 100-114.
Stiles, W. S. (1978). Mechanisms of colour vision. London: Academic
trast dependent. Vision Research, 32, 1535-1549.
Stromeyer, C. F., Chaparro, A., & Kronauer, R. E. (1996). The colour
and motion of moving patterns are processed independently.
Investigative Ophthalmology and Visual Science, 37(Suppl.), s916.
Stromeyer, C. F., III., Chaparro, A., Tolias, A. S., & Kronauer, R. E.
(1994). Coloured fields produce large l vs. M phase shifts in lumi-
nance motion mechanisms. Investigative Ophthalmology and Visual
Science, 35, 1812.
216BEHAVIORAL AND COGNITIVE NEUROSCIENCE REVIEWS
Stromeyer, C. F., III, Kronauer, R. E., Ryu, A., Chaparro, A., & Eskew, Download full-text
R. T., Jr. (1995). Contributions of human long-wave and middle-
Switkes, E., Bradley, A., & De Valois, K. K. (1988). Contrast depend-
ence and mechanisms of masking interactions among chromatic
and luminancegratings. Journal of the Optical Society of AmericaA, 5,
Switkes, E., & Crognale, M. A. (1999). Comparison of color and lumi-
Thiele, A., Dobkins, K. R., & Albright, T. D. (1999). The contribution
of color to motion processing in macaque middle temporal area.
Journal of Neuroscience, 19, 6571-6587.
Thiele, A., Dobkins, K. R., & Albright, T. D. (2001). Neural correlates
of chromatic motion perception. Neuron, 32, 351-358.
Thiele, A., Rezec, A., & Dobkins, K. R. (2002a). Chromatic input to
motion processing in the absence of attention. Vision Research, 42,
Thompson, P. (1982). Perceived rate of movement depends on con-
trast. Vision Research, 22, 377-380.
Tootell, R. B. H., Silverman, M. S., Hamilton, S. L., De Valois, R. L., &
Switkes, E. (1988). Functionalanatomyof macaquestriatecortex.
Iii. Color. Journal of Neuroscience, 8, 1569-1593.
Troscianko, T., & Fahle, M. (1988). Why do isoluminant stimuli
Wallach, H. (1935).Uber
bewegungsrichtung. Psychologische Forschung, 20, 325-380.
Wallach, H., & O’Connell, D. (1953). The kinetic depth effect. Jour-
nal of Experimental Psychology, 45, 205-217.
Wandell, B. A., Poirson, A. B., Newsome, W. T., Baseler, H. A.,
Boynton, G. A., Huk, A., et al. (1999). Color signals in human
motion-selective cortex. Neuron, 24, 901-909.
Watson, A. B., Barlow, H. B., & Robson, J. G. (1983). What does the
eye see best? Nature, 302, 419-422.
Watson, A. B., & Robson, J. G. (1981). Discrimination at threshold:
Watson, A. B., Thompson, P. G., Murphy, B. J., & Nachmias, J. (1980).
Summation and discrimination of gratings moving in opposite
directions. Vision Research, 20, 341-347.
effects: Evidence for luminance- and color-movement pathways.
Vision Research, 32, 2187-2189.
Webster, M. A., & Mollon, J. D. (1997). Motion minima for different
directions in color space. Vision Research, 37, 1479-1498.
nisms underlie the detection and discrimination of moving chro-
matic targets. Proceedings of the Royal Society of London B, 265, 2435-
Wray, J., & Edelman, G. M. (1996). A model of color vision based on
cortical reentry. Cerebral Cortex, 6, 701-716.
Wuerger, S. M., & Landy, M. S. (1993). Role of chromatic and lumi-
nance contrast in inferring structure from motion. Journal of the
Optical Society of America A, 10, 1363-1372.
Wuerger, S. M., & Morgan, A. H. (1997). Speed discrimination for lumi-
nance and colour stimuli as a function of contrast. Paper presented at
the European Conference on Visual Perception, Finland.
Wuerger, S. M., Shapley, R., & Rubin, N. (1996). “On the Visually Per-
tion, 25, 1317-1367.
Wyszecki, G., & Stiles, W. S. (1982). Color science: Concepts and methods,
quantitative data and formulae. New York: John Wiley & Sons.
Yoshizawa, T., Mullen, K. T., & Baker, C. L., Jr. (2000). Absence of a
chromatic linear motion mechanism in human vision. Vision
Research, 40, 1993-2010.
Yoshizawa, T., Mullen, K. T., & Baker, C. L., Jr. (2003). Failure of
signed chromatic apparent motion with luminance masking.
Vision Research, 43, 751-759.
Zaidi, Q., & DeBonet, J. S. (2000). Motion energy versus position
tracking: Spatial, temporal and chromatic parameters. Vision
Research, 40, 3613-3635.
Zemany, L., Stromeyer, C. F. S., III, Chaparro, A., & Kronauer, R. E.
rons tune to more than one dimension. Biological Cybernetics, 66,
Cropper, Wuerger / PERCEPTION OF MOTION IN CHROMATIC STIMULI 217