Vision Research 49 (2009) 943–959
Contents lists available at ScienceDirect
Vision Research
journal homepage: www.elsevier.com/locate/visres
Retinal parallel pathways: Seeing with our inner fish
Christina Joselevitch, Maarten Kamermans *
Retinal Signal Processing, The Netherlands Institute for Neuroscience (KNAW), The Royal Netherlands Academy of Arts and Sciences, Meibergdreef 47,
1105 BA Amsterdam, The Netherlands
a r t i c l e
i n f o
Article history:
Received 3 October 2007
Received in revised form 21 June 2008
Keywords:
Retina
Vision
Visual pathways
Photoreceptors
Neurons
a b s t r a c t
The general organization of the vertebrate retina is highly conserved, in spite of structural variations that
occur in different animal classes. The retinas of cyprinid fish, for example, differ in many aspects from
those of primates. However, these differences are in the same order of magnitude as those found among
mammalian species. Therefore, it is important to consider whether these changes are minor variations on
the same theme or whether they lead to fundamentally different functions. In this light, we compare the
retinal organization of teleost fish and mammals as regards parallel processing and discuss their many
similarities.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
The first intracellular recordings of vertebrate retinal neurons
were performed in teleosts (Svaetichin, 1953, 1956). Paramount
discoveries about the organization of the visual system of vertebrates, such as the existence of multiple retinal channels, resulted
from such studies in fish and other cold-blooded animals. Since
then, many differences were described between fish and mammalian retinas. One can however ask how much these structural differences are functionally significant, as even within the
mammalian class there is a large variation as far as retinal structure and wiring are concerned. This review compares the information coding schemes and transmission pathways in the fish and
mammal and discusses that, despite species-specific architectural
adaptations, the function of various retinal circuits is in principle
very similar.
2. Why have multiple pathways?
Why does the retina use parallel streams to convey information
to higher areas? It seems intuitively simpler to have a one-to-one
connection from the photoreceptors to the brain and leave the processing of information to the latter, such as in the auditory system.
However tempting, this reasoning has caveats, as discussed briefly
below and in detail elsewhere (Barlow, 1981; Laughlin, 2001;
Sterling, 2004).
* Corresponding author. Fax: +31 20 566 6121.
E-mail address: [email protected] (M. Kamermans).
0042-6989/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.visres.2008.07.019
First, exclusive lines from the retina to the brain would imply a
very thick optic nerve, which would increase the size of the blind
spot. Second, it would also impair eye movements, which are crucial for retinal fixation and to avoid photoreceptor adaptation (Barlow, 1952; Martinez-Conde, Macknik, & Hubel, 2004). Third, the
total ganglion cell population and consequently the retinal energy
consumption would increase, since generating spikes in ganglion
cells has metabolic costs (Laughlin, 2001; Lennie, 2003).
2.1. The need for retinal convergence raises problems
In order to diminish the absolute cell number, the visual system
makes photoreceptor signals converge onto second-order neurons.
Convergence, in turn, has both advantages and disadvantages. It
can have negative effects on many visual aspects such as sensitivity and acuity, as well as spectral and temporal resolution, because
different visual functions have conflicting needs in terms of signal
transmission (Ashmore & Falk, 1980b; Falk, 1988; Sterling, 2004;
Warrant, 2004).
For example, motion detection needs fast transmission but does
not rely on spatial detail, whereas visual acuity has exactly the
opposite needs (Koch et al., 2006; Sterling, 2004). Vision at different light levels has also multiple requirements. The visual system
needs to perform well both at night time and day time, even
though photon levels differ enormously from one condition to
the other (Barlow, 1981; Sterling, 2004). This requires transmission
with high gain at scotopic levels and with low gain at photopic levels. If all photoreceptors would converge onto a single pathway,
these needs would not be entirely met and, as a result, visual perception would suffer.
944
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
2.2. Divergence offers solutions
A solution to this design dilemma implies the concomitant use
of a second strategy: divergence. By having photoreceptor signals
travel to the brain via multiple channels with different absolute
and spectral sensitivities, as well as spatial and temporal resolutions, the visual system makes sure that relevant information does
not get lost due to the convergence necessary to keep the retina
economically and functionally viable.
Another advantage of signal divergence is efficient coding. Because a spiking neuron has a limited bandwidth (i.e. it can only
transmit a finite amount of information per unit time), dividing
signals into different channels with slightly distinct properties
might increase both the rate of transmission and the total amount
of transmissible information (Barlow, 1981; Koch et al., 2006;
Laughlin, 2001). This would, for instance, favor contrast sensitivity
and increase the amount of discriminable gray levels (Barlow,
1981).
Finally, the existence of several retinal pathways might create a sort of ‘‘neural backup” and prevent large deficits when
one particular system is compromised (Heiligenberg, 1987).
One such example can be found in the autosomal recessive form
of congenital stationary night blindness. This disease results
from a defect in the glutamate receptor of ON bipolar cells,
which renders the whole ON pathway silent (Dryja et al.,
2005; Zeitz et al., 2005). One would expect affected individuals
to show major deficits not only at scotopic levels, when all roddriven signals are conveyed to the inner retina by an ON bipolar
cell (discussed in Sterling, 2004 and in the next sections), but
also at mesopic and photopic levels, since cone-driven ON bipolar cells also use the same receptor (Vardi et al., 2002). However
decreased, mesopic and photopic visual functions in these patients are consistent with the existence of alternative pathways
from both rods and cones to the inner retina (Dryja et al., 2005;
Zeitz et al., 2005).
Similarly, mice lacking the ON bipolar cell receptor perform as
well as wild-type animals in tests of visually-guided behavior
(Masu et al., 1995). Together, these results suggest that other retinal pathways compensate for the absence of ON bipolar cell activity. Parallel retinal pathways seem to be a need shared by all
vertebrate species, fish and mammals included.
3. Are fish and mammalian channels that different?
The overall architecture of the vertebrate retina is very similar
among species (Cajal, 1893), which reflects the fact that the tasks
performed by their visual systems are in many ways alike. There
are nonetheless differences between fish and mammals as regards
retinal structure. In the next sections, we will show that these anatomical variations are in fact quite comparable to the ones found
among mammalian species, leaving however the function of the
retinal subsystems involved more or less unchanged. This indicates
that, as far as the functional organization of the retina is concerned,
fish and primates are not so far apart.
But how many retinal channels are there? This is not a straightforward question, because it depends on the criteria used to analyze retinal organization. Although it is tempting to directly
relate visual percepts such as motion, form, texture, color and
brightness to the activity of individual neurons or pathways (Livingstone & Hubel, 1988), the diversity of retinal cell types indicates
that, at this level, more than one channel might be involved in each
of these sensations.
At the same time, at each retinal stratum and beyond neurons converge and diverge, making the adjective ‘‘parallel”
somewhat inappropriate (Merigan & Maunsell, 1993). Evidence
of this is the fact that the number of cell types changes with
retinal level. Although one might find about 10–12 bipolar cell
types in the primate (Fig. 1C, Boycott & Wässle, 1999; Wässle,
2004), at the ganglion cell level the number of channels is already 17–18 (Field & Chichilnisky, 2007; Kolb, Linberg, & Fisher,
1992).
The same mismatch between the number of cell types at each
retinal level apparently also holds for the fish retina, although
the morphology of fish retinal neurons has not been completely
elucidated yet. The zebrafish, for example, seems to have at least
17 distinct bipolar cell types (Connaughton, Graham, & Nelson,
2004), but only about 11 ganglion cell types have been identified
(Mangrum, Dowling, & Cohen, 2002) so far. In the closely related
goldfish, about 14–15 bipolar cell types (Fig. 1A, Sherry & Yazulla,
1993) and 15 ganglion cell subtypes were described (Hitchcock &
Easter, 1986).
Since the physiology of most of the mammalian retinal neurons—and also of their fish counterparts—is still largely unknown,
we will concentrate on those pathways whose properties are reasonably well understood: rod and cone, ON and OFF and broadband
and opponent channels.
4. Rods use both ON and OFF channels
Fish and mammals with duplex retinas have chosen apparently
different strategies to convey rod and cone signals to second-order
neurons. As discussed in the next paragraphs, however, these strategies are, functionally speaking, quite similar. The classical picture
is that rod signals flow from the outer to the inner retina via different structures in mammals and fish.
These pathways are summarized in Fig. 1A for the goldfish, in
Fig. 1B for the mouse and in Fig. 1C for the monkey. Mammals have
an exclusive rod-driven channel comprised by an ON bipolar cell,
shown in green (Boycott, Dowling, & Kolb, 1969; Cajal, 1893). Fish,
on the other hand, do not have a bipolar cell exclusively dedicated
to rods. Rather, these animals use bipolar cells of both ON (Fig. 1A,
yellow) and OFF types (Fig. 1A, brown) that receive mixed rod–
cone input to transmit information (Scholes, 1975; Stell, 1967).
When examined closer, however, these dissimilarities are not substantial: rod pathways in both animal classes are actually conveyed
to the inner retina by ON and OFF pathways with distinct gains, as
discussed below.
Anatomical connections between rods and OFF bipolar cells
were described in a number of mammals (mouse: Tsukamoto,
Morigiwa, Ueda, & Sterling, 2001; rat: Hack, Peichl, & Brandstatter,
1999; squirrel: Li & DeVries, 2007; West, 1978; cat: Fyk-Kolodziej,
Qin, & Pourcho, 2003; rabbit: Li, Keung, & Massey, 2004). These
mixed-input OFF bipolar cells are indicated for the mouse in
Fig. 1B (brown neurons). In addition, mixed-input seems to be
present in the ON pathway of the mammalian retina as well, since
an ON ‘‘cone” bipolar cell of the mouse was shown to contact rods
directly (Tsukamoto et al., 2007). This cell is represented in yellow
in Fig. 1B.
Some of the mammalian OFF bipolar cells that contact rods
are analogous to the primate DB2 OFF bipolar cell (Euler & Wässle, 1995; Fyk-Kolodziej et al., 2003), depicted in gray in Fig. 1C.
This suggests that mixed-input bipolar cells might also exist in
the primate retina. Although such rod–bipolar cell contacts have
not been described in primates, it is possible that the bipolar cell
connectivity in the primate retina—as well as in a number of
other mammalian species—is simply not completely solved yet
(i.e. see discussion in Protti, Flores-Herr, Li, Massey, & Wässle,
2005). Alternatively, there might be indeed no mixed-input bipolar cells in the primate retina. In this case, one has to realize that
as far as rod-driven pathways are concerned, the mouse, rat,
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
945
Fig. 1. Bipolar cells in the vertebrate retina and their connections to rods. Thin horizontal lines through the inner plexiform layer represent the border between the OFF
sublamina (upper half) and the ON sublamina (bottom half). (A) In the goldfish, rod signals are conveyed to the inner retina via multiple OFF (brown) and ON bipolar cell types
(yellow) that receive contributions from rods and cones in different proportions. Redrawn from Sherry and Yazulla (1993). (B) In the mouse retina, ON (yellow) and OFF
mixed-input bipolar cells (brown) were also found (ON: Tsukamoto et al., 2007; OFF: Mataruga, Kremmer, & Muller, 2007), in addition to the classical rod ON bipolar cell
(green). Adapted from Ghosh, Bujan, Haverkamp, Feigenspan, and Wässle (2004), Mataruga et al. (2007) and Tsukamoto et al. (2007). (C) In the monkey retina, so far only the
rod ON bipolar cell (green) was described. However, if the primate DB2 bipolar cell (gray) is indeed homologous to the rabbit CBa2 and cat cb1 bipolars (Euler & Wässle, 1995;
Fyk-Kolodziej et al., 2003), it is expected to contact rods as well as cones (Fyk-Kolodziej et al., 2003; Li et al., 2004). Redrawn from Mariani (1989) and Boycott and Wässle
(1999). OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer. (For interpretation of the references to color in this figure legend, the reader is referred
to the web version of this paper.)
squirrel, rabbit and cat might be as far from the primate as the
goldfish.
4.1. ON bipolar cells form the high sensitivity pathway
Scotopic vision in fish, similarly to what happens in mammals,
is also subserved by the ON pathway. Mixed-input ON bipolar cells
are, regardless of their name, predominantly driven by rods in the
dark-adapted retina (Joselevitch & Kamermans, 2007; Kaneko &
Tachibana, 1978; Saito, Kondo, & Toyoda, 1979). Fig. 2A compares
the sensitivity of 38 mixed-input ON bipolar cells (dark circles), 8
mixed-input OFF bipolar cells (open circles) and 18 cones (dark triangles) recorded from the goldfish retina. Mixed-input ON bipolar
cells are more sensitive than OFF cells and cones, suggesting that
ON cells receive a large rod contribution. Accordingly, the intensity–response relation of these neurons (Fig. 2B, dashed black line)
is as steep as those of cones (colored symbols and lines) but at the
same time far off the intensity range in which cones are active. This
indicates that these cells receive predominantly inputs from a single photoreceptor type and that this receptor is more sensitive
than cones (Joselevitch & Kamermans, 2007). Not only are fish
mixed-input ON bipolar cells rod-driven in the dark-adapted retina, but they also make use of the same signaling cascade as the
mammalian rod bipolar cell (Shiells, Falk, & Naghshineh, 1981;
Slaughter & Miller, 1981), namely via a group III metabotropic
glutamate receptor (Nakajima et al., 1993).
The cone input onto fish mixed-input ON bipolar cells, discussed in more detail in the next sections, could in principle
shunt rod-driven signals (Falk, 1988; Wong, Cohen, & Dowling,
2005), because it modulates a conductance that is open in darkness (Saito, Kondo, & Toyoda, 1981; Saito et al., 1979). This
could, in turn, render mixed-input ON bipolar cells less sensitive
than their mammalian counterparts. However, psychophysical
measurements of absolute thresholds in goldfish (Powers & Easter, 1978) yielded figures not far from those reported for humans
under similar experimental conditions (see discussion in Powers
& Easter, 1978).
It seems possible that a number of mechanisms such as the
high gain of the rod-ON bipolar cell synapse (Ashmore & Falk,
1980a; Falk, 1988), the large intrinsic photosensitivity of the
rods (Baylor, Lamb, & Yau, 1979) and even electrical coupling
within the photoreceptor layer (Ashmore & Falk, 1980b; Tessier-Lavigne & Attwell, 1988) might compensate for the deleterious effects of cone convergence and enable the visual system
to perform relatively well at scotopic levels. Functionally
speaking, therefore, fish mixed-input ON bipolar cells behave
much like mammalian rod-ON bipolar cells in the dark-adapted
retina.
946
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
4.2. OFF bipolar cells form the low sensitivity pathway
Rod inputs at the transition between scotopic and mesopic vision, on the other hand, are more efficiently conveyed by mixed-input OFF bipolar cells in fish. Fig. 2A and B shows that fish mixedinput OFF bipolar cells are less sensitive and have a broader dynamic range than mixed-input ON bipolar cells (Joselevitch &
Kamermans, 2007), partly due to their comparatively large cone inputs (Ishida, Stell, & Lightfoot, 1980; Joselevitch & Kamermans,
2007; Stell, 1976, 1978).
Fig. 2C illustrates that the mammalian rod-to-OFF bipolar cell
pathway, just like its fish equivalent, also has a lower sensitivity
than the main rod-to-ON bipolar cell channel (Soucy, Wang, Nirenberg, Nathans, & Meister, 1998; Volgyi, Deans, Paul, & Bloomfield,
2004). In connexin 36 knock-out mice, in which gap junctions between photoreceptors (discussed in the next section) and between
AII ACs and cone-driven ON BCs in the inner retina are abolished,
the additional block of ON bipolar cell responses with an agonist
for group III mGluRs isolates the pathway from rods to mixed-input OFF bipolar cells (Fig. 2C). OFF ganglion cell responses in this
condition (dark lines) are less sensitive than those in wild-type retinas (green lines). The lower sensitivity of this rod pathway is
probably due to the smaller gain of the rod–OFF bipolar cell synapse (Soucy et al., 1998) and to the relatively large proportion of
cone contacts that mammalian mixed-input OFF bipolar cells make
(Li et al., 2004; Tsukamoto et al., 2001).
Since mammalian rod-driven ON bipolar cells saturate at very
low light levels (Berntson, Smith, & Taylor, 2004; Field & Rieke,
2002) and completely stop responding to light as background levels increase (Dacheux & Raviola, 1986), the lower gain of this alternative rod-driven channel enables the visual system to keep
responsive in the scotopic-mesopic transition (Soucy et al., 1998;
Volgyi et al., 2004), much like what happens in the cyprinid fish
retina.
4.3. Rods ‘‘piggyback” on the cone system
Fig. 2. Sensitivity of rod-driven ON and OFF pathways. (A) The sensitivity of 18 cones
(triangles, mean = 1.42 log ± 0.69), 8 mixed-input OFF (open circles, mean =
0.51 log ± 0.56) and 38 mixed-input ON (closed circles, mean = 0.84 log ± 0.67)
bipolar cells of the goldfish for full-field stimulation at 550 nm, plotted in ascending
order for comparison. Dashed lines indicate the mean values for each category. Both ON
and OFF neurons receive substantial rod input and are therefore more sensitive than the
cones; ON cells are more sensitive than OFF cells. From Joselevitch and Kamermans
(2007). (B) Hill fits for the intensity–response relations of a mixed-input ON (dashed
black line) and OFF (continuous black line) bipolar cell of the goldfish to full-field
stimulation at 550 nm, superposed on cone responses for similar stimuli. The curve for
the OFF cell has a shallower slope (n = 0.41) and lower sensitivity (log K 1 = 0.27)
than the ON cell. From Joselevitch and Kamermans (2007). (C) Comparison of the
normalized intensity–response profiles of OFF ganglion cells in wild-type mice (green
curves) and in Cx36 KO animals after pharmacological block of ON bipolar cells with
50 lm L-AP-4 (black curves). Under these conditions, only two pathways are active: the
direct rod ? OFF bipolar cell ? OFF ganglion cell pathway and the cone ? OFF bipolar
cell ? OFF ganglion cell pathway (black curves), both with low sensitivity. Redrawn
from Volgyi et al. (2004). (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this paper.)
Finally, there is a third strategy common to fish and mammals
to transmit rod signals to the inner retina: the electrical coupling
between rods and cones (DeVries & Baylor, 1995; Schneeweis &
Schnapf, 1995). The presence of gap junctions between these two
types of photoreceptors has been documented in a number of fish
and mammalian species (guinea pig: Sjostrand, 1958; squirrel: Cohen, 1964; human: Cohen, 1965; monkey and rabbit: Dowling &
Boycott, 1966; Raviola & Gilula, 1973, 1975; carp: Witkovsky, Shakib, & Ripps, 1974; cat: Kolb, 1977).
Fig. 3A shows that electrical coupling between rods and cones
changes the spectral sensitivity of goldfish cone-driven monophasic horizontal cells (Wang & Mangel, 1996). During the day, when
rod–cone gap junctions are closed, the light responses of these cells
follow the L-cone absorption spectrum (red line). At night, when
these gap junctions are open, rod signals reach monophasic horizontal cells and their spectral sensitivity peak shifts towards the
rod absorption spectrum (green line). A similar effect happens in
the mammalian retina. Monkey M-cones (Fig. 3B, open triangles)
and L-cones (Fig. 3B, open circles), for instance, are more sensitive
to 500 nm (green) light than they actually should be (the expected
spectral sensitivity is indicated by the dotted lines) if there were no
interactions between photoreceptors (Schneeweis & Schnapf,
1999). This experiment clearly shows that cone light responses
are directly influenced by rods (dark circles) in the primate retina
as well.
This pathway allows rods to diverge to cone systems and
change the response properties of cones and cone-driven neurons
(Dacheux & Raviola, 1982; Nelson, 1977; Wang & Mangel, 1996),
with the interesting side-effect of improving signal-to-noise ratio
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
947
Nonetheless, as far as these three rod systems are concerned,
fish and mammalian retinas are not, therefore, universes apart.
These pathways enable the visual system to bridge seamlessly a
large range of intensities, by combining rod and cone inputs to different degrees at the bipolar cell level.
4.3.1. Can fish see single photons?
This brings us to the question: Is the rod system in fish as sensitive to light as the mammalian one? The answer is: it depends.
The absolute sensitivity of an animal is driven by its own evolutionary history and ecological needs. Even among mammals, the
perception of single-photon events has variable ecological importance. The cat, for instance, is a predominantly nocturnal animal
with a large number of rods per unit area (Steinberg, Reid, & Lacy,
1973) and depends on vision at low light levels. It is not surprising,
therefore, that responses to single light quanta were measured
from cat ganglion cells (Barlow, Levick, & Yoon, 1971; Mastronarde, 1983b). The ground squirrel, on the other hand, is a diurnal
species with a cone-dominated retina (Anderson & Fisher, 1976;
West & Dowling, 1975). Single-photon events mean probably
much less for its survival than for the cat.
Similarly, the habitat and visual needs of deep sea fish might require the perception of single-photon events (Warrant, 2000,
2004), while those of the cone-dominated goldfish might not. In
fact, ON bipolar cells of the dogfish, an elasmobranch with an allrod retina, respond reliably to light flashes delivering less than
0.05 Rh* (Ashmore & Falk, 1980b). The largest limit to single photon detection in this species is the spontaneous activity of rod
machinery itself (Ashmore & Falk, 1982), a constraint common to
all vertebrate species.
Ecological needs, therefore, determine ultimately the retinal
characteristics that bear consequences for visual sensitivity, such
as the size of the eyes, the range of pupil apertures, the density
of photoreceptors, the existence and degree of electrical coupling
between retinal neurons and the amount of convergence in each
species (Barlow, 1981; Walls, 1942), as well as the relative proportion of centers in the brain dedicated to processing visual information (Rapaport & Stone, 1984).
Fig. 3. Rod–cone coupling in the vertebrate retina. (A) In the goldfish, the spectral
sensitivity of cone-driven monophasic horizontal cells, which do not receive direct
rod input, matches that of rod-driven cells (green) during the night. This
demonstrates that there is rod–cone coupling in the fish retina, and that it is
under circadian control. From Wang and Mangel (1996). (B) In the monkey, the
relative spectral sensitivity of M- and L-cones (empty triangles and circles),
depicted here as the ratio of sensitivity to light stimulation at 500 and 660 nm,
deviates from the expected values (dashed lines): both cells are more sensitive to
500 nm than they should be, indicating rod–cone coupling. From Schneeweis and
Schnapf (1999). (For interpretation of the references to color in this figure legend,
the reader is referred to the web version of this paper.)
by averaging out uncorrelated activity (Ashmore & Falk, 1980b;
Lamb & Simon, 1976).
Rod–cone coupling seems to be regulated both by a circadian
clock and by the adaptive state of the retina in fish (Mangel, Baldridge, Weiler, & Dowling, 1994; Wang & Mangel, 1996), whereas
such regulation in the mammalian retina is still under debate, as
well as the exact intensity range in which rod–cone coupling
should be the most efficient (see discussion in Hornstein, Verweij,
Li, & Schnapf, 2005; Schneeweis & Schnapf, 1999; Smith, Freed, &
Sterling, 1986).
4.3.2. The fish depolarizing receptor: A special case?
Fig. 4A shows that depolarizing light responses in teleost
mixed-input ON bipolar cells have two conductance mechanisms
in the mesopic range (Saito et al., 1979, 1981). When hyperpolarized by current injection (bottom traces), these cells increase their
responses throughout the spectrum as compared to control (upper
traces), whereas only the responses to 675 nm (red) light reverse
polarity. This indicates that one conductance mechanism drives
them at short wavelengths and another one at long wavelengths.
Additional evidence for a rod- and a cone-driven depolarizing
receptor in mixed-input ON bipolar cells is found in Fig. 4B, in
which a group III antagonist (CPPG) is not able to completely block
depolarizing responses of mixed-input ON bipolar cells to mesopic
light stimulation (Wong et al., 2005).
This cone-driven depolarizing receptor has long been regarded
as a peculiarity of the fish retina, in spite of sporadic reports of
depolarizing responses in mammalian rod bipolar cells that could
not be abolished pharmacologically (Cohen & Miller, 1999) or biphasic glutamate-induced conductance changes in mammalian
ON cone bipolar cells (i.e. Fig. 2 in Huang et al., 2003) and rod bipolar cells (Wersinger et al., 2006).
The fish receptor is probably an excitatory amino acid transporter, or EAAT (Grant & Dowling, 1995; Grant & Dowling, 1996)
that modulates a conductance with a negative reversal potential
in response to a change in the cone glutamate release (Saito et al.,
1979, 1981). The major candidate for this receptor, EAAT5, is found
in putative bipolar cell dendrites in the goldfish cone synaptic
948
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
quasilinearly in these cells (Falk, 1988), which would, as a result,
extend their dynamic range. Indeed, the dynamic range of goldfish
mixed-input ON bipolar cells is slightly larger than that of a photoreceptor (Joselevitch & Kamermans, 2007) or of a mammalian rod
bipolar cell (Berntson & Taylor, 2000; Euler & Masland, 2000).
In the scotopic range, three types of interaction between EAAT5
and the rod-driven conductance are possible, but there is no quantitative data available to show which one is predominant. First,
EAAT5 could help generate the driving force for the rod-driven
conductance by hyperpolarizing mixed-input ON bipolar cells in
darkness (Joselevitch & Kamermans, 2005). Alternatively, it could
also shunt the rod-driven conductance (Falk, 1988; Wong et al.,
2005), which would lead to a rise in threshold and decrease of
rod-driven light responses. A third possibility is that, depending
on the geometry of the cells and of the precise localization of
EAAT5 ionic channels (i.e. at the dendritic tips contacting cones
in invaginating processes as opposed to along the dendrites), they
could electrically isolate the cone-driven dendrites in the darkadapted state from the rest of the cell by creating a large local leak,
rendering mixed-input ON bipolar cells purely rod-driven at scotopic levels (Joselevitch & Kamermans, 2005, 2007).
Whatever might be the case, the existence of two types of glutamate receptor in fish mixed-input ON bipolar cells enables these
cells to function also at photopic levels, in which mammalian rod
bipolar cells are thought to be irresponsive (Dacheux & Raviola,
1986; Shimbo, Toyoda, Kondo, & Kujiraoka, 2000; Wong et al.,
2005). Since the output of mixed-input ON bipolar cells is not entirely characterized (see for instance Marc & Liu, 2000 and Zimov &
Yazulla, 2008), it is still to be determined how their activity at different light levels influences ganglion cells and the visual behavior
of the whole animal. Furthermore, until the functional significance
of mammalian EAAT5 is fully understood, claims that the fish
depolarizing receptor is a special case with no mammalian counterpart should be regarded with caution.
5. Cone ON and OFF channels: Mirror images are not quite what
they seem
Fig. 4. Light responses of mixed-input ON bipolar cells in the teleost retina are
mediated by two receptors at mesopic levels. (A) Responses of a carp mixed-input
ON bipolar cell to light stimuli at different wavelengths in control (top), and during
the injection of hyperpolarizing current (bottom). Each spectral response was
differently affected by hyperpolarization, indicating that at least two mechanisms
are involved in their generation. From Saito et al. (1979). (B) Left: Light responses of
mixed-input ON bipolar cells of the giant danio cannot be completely abolished by
the group III antagonist CPPG, but are completely blocked with the addition of
TBOA, a glutamate transporter blocker, to the bath. Right: Summary results for 11
cells. From Wong and Dowling (2005). (C) Pre-embedding immuno EM for EAAT5 in
the goldfish outer plexiform layer. Invaginating structures with the appearance of
bipolar cell dendrites (arrows) are positively labeled. Lateral and central elements
of the triads are unlabeled. SR, synaptic ribbon. Scale bar = 0.25 lm. From
Kamermans, Joselevitch, and Klooster (2004).
complex (Fig. 4C, arrows). It has also been described in mammalian
rod and cone bipolar cells (Pow & Barnett, 2000; Wersinger et al.,
2006). A role for EAAT5 as a post-synaptic receptor in mammalian
bipolar cells cannot be fully excluded yet (Wersinger et al., 2006).
At least in the teleost retina, the need for such a receptor may
have arisen from the existence of rod and cone contributions to
ON bipolar cells, since it offers certain advantages to an ON cell
that receives mixed-input. In the mesopic range, EAAT5 could in
principle improve the synaptic transmission efficiency between
cones and mixed-input ON bipolar cells by counteracting the
shunting effect of tonically open rod-driven channels (Falk, 1988;
Saito et al., 1979). This way, rod and cone signals could add up
ON and OFF channels may have multiple roles in photopic vision. The functional significance of the ON–OFF division, however,
is not entirely clear. It was argued that dividing the visual signals in
channels signaling with opposite polarities would effectively double the bandwidth of the retinal output for each piece of an image
(Barlow, 1981). Such doubling could be beneficial for contrast perception, which relies on the somewhat sparse code of ganglion
cells (Barlow, 1981; Schiller, 1992).
Although tempting, this idea has requirements that do not always seem to hold. First, it needs symmetric and coextensive representations of the visual world in both channels (as discussed in
Vaney & Hughes, 1990). Second, it needs the channels to remain
parallel in the retina and converge only at a post-retinal site. Literature is however filled with examples of asymmetries and retinal
cross-talk between ON and OFF channels in both fish and mammals, as discussed below.
5.1. Asymmetries between retinal ON and OFF pathways are common
in fish and mammals
Asymmetries between ON and OFF cells are found at each retinal stage. One example is the stratification pattern of bipolar cells.
The number of cone-driven bipolar cell types that stratify in sublamina a of the inner plexiform layer differs from that stratifying
in sublamina b in a number of species. This is the case in teleosts
such as the rudd (Scholes, 1975), goldfish (Sherry & Yazulla,
1993, Fig. 1A) and zebrafish (Connaughton et al., 2004), and also
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
in mammals such as the rat (Euler & Wässle, 1995), cat (Famiglietti, 1981; Kolb, Nelson, & Mariani, 1981), monkey (Mariani,
1983; Wässle, 2004) and human (Kolb et al., 1992). If one assumes
that sublamina a contains exclusively ON cells and sublamina b contains only OFF cells, this would mean that not all cone-driven ON
bipolar cells have an OFF counterpart.
Asymmetries at the ganglion cell level are plenty. The dendritic
tree size and branching of ON ganglion cells are dissimilar to those
of OFF ganglion cells belonging to the same class (goldfish: Vallerga & Djamgoz, 1991; rat: Peichl, 1989; cat: Wässle, Peichl, & Boycott, 1981; rabbit: Tauchi, Morigiwa, & Fukuda, 1992; human:
Dacey & Petersen, 1992), suggesting that these neurons have also
different numbers of synaptic contacts per unit area and receptive
field sizes. Indeed, receptive field sizes of ON versus OFF ganglion
cells of the same sort are different in goldfish (Vallerga & Djamgoz,
1991) and monkey (Chichilnisky & Kalmar, 2002).
Further, ON and OFF ganglion cell pairs respond differently to
contrast in goldfish (Bilotta & Abramov, 1989), guinea pig (Zaghloul, Boahen, & Demb, 2003) and monkey (Chichilnisky & Kalmar,
2002). ON and OFF responses from goldfish optic fibers differ in
absolute (Russell & Wheeler, 1983) and spectral sensitivity
(Wheeler, 1979). A similar asymmetry regarding the spectral sensitivity of ON and OFF responses was also observed in recordings
of optic nerve fibers of the trout (Beaudet, Browman, & Hawryshyn,
1993) and monkey (de Monasterio, 1979a).
In the mouse, neither the absolute sensitivity nor the dynamic
range of ON and OFF a-ganglion cells mirror each other (Wu,
Gao, & Pang, 2004). This can be seen in Fig. 5, in which the dynamic
ranges of spiking activity (red dotted bars), excitatory (red bars)
and inhibitory (black bars) currents measured in different neurons
in the mouse dark-adapted retina are depicted (Wu et al., 2004).
ON a-ganglion cells are more sensitive than transient OFF a-cells.
In fact, OFF a-ganglion cells might actually comprise two physiological types (one transient and one sustained, Pang, Gao, & Wu,
2003), which transforms the ON/OFF dichotomy into a trichotomy.
Fig. 5. Dynamic ranges of excitatory (red lines) and inhibitory (black lines)
currents, as well as spiking activity (dashed lines) of different mouse retinal
neurons to stimulation at 500 nm. Note that rod-driven bipolar cells can be
classified in two different groups as regards inhibitory inputs (DBCR1 and DBCR2).
Further, there is considerable asymmetry between ON and OFF cells: cone-driven
OFF bipolar cells (HBCMC and HBCSC) and are less sensitive than cone-driven ON
bipolar cells (DBCC1 and DBCC2). While its is possible that not all bipolar cell types
were recorded in this study, at the ganglion cell level, an extra type of OFF aganglion cell is described (tOFFaGC) that has no ON counterpart. From Pang et al.
(2003) and Wu et al. (2004). (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this paper.)
949
The fact that these asymmetries are found in different animal
classes means that they might share underlying reasons. It is conceivable that the distinct response properties of ON and OFF ganglion cells reflect different information content (Arnett, 1978;
Meister, 1996). This would mean that the ultimate function of
ON and OFF pathways at photopic levels might not be bandwidth
doubling, but rather complementation. In this case, cross-talk between these channels could be beneficial for visual processing, as
discussed below, and one would expect such cross-talk to happen
from fish to primates.
5.2. Cross-talk between ON and OFF channels improves signal-to-noise
ratio
In the late 60s, Henk Spekreijse observed that the responses of
ganglion cells in the goldfish retina are rectified versions of the input, that is, they distort the stimulus waveform in a non-linear
manner (Spekreijse, 1969). Rectification was also observed in the
responses of carp amacrine cells (Toyoda, 1974; Toyoda, Hashimoto, & Otsu, 1973), cat ganglion cells (Enroth-Cugell & Robson, 1966;
Hochstein & Shapley, 1976) and rabbit ganglion cells (see Fig. 10b
from van Wyk, Taylor, & Vaney, 2006). It was shown to happen at a
site posterior to the summation of photoreceptor inputs (Spekreijse & Van den Berg, 1971), which is approximately linear and
probably happens at the bipolar cell level (Hochstein & Shapley,
1976).
Later studies of goldfish (Levine & Shefner, 1977) and rabbit
ganglion cells (Roska, Molnar, & Werblin, 2006) confirmed that this
rectification occurs indeed in the transmission between bipolar
and ganglion cells: even though the membrane potential of bipolar
cells can be modulated in both directions, signal transmission
across this synapse is most efficient for stimuli that depolarize
these neurons (Levine & Shefner, 1977; Roska et al., 2006; but
see Demb, Zaghloul, Haarsma, & Sterling, 2001). The rectifying
mechanism might be the spike generator of amacrine and/or ganglion cells, because impulse rates cannot go negative (Hochstein &
Shapley, 1976).
Two important consequences derive from this rectification:
first, the bipolar cell surrounds produced in the outer retina are reduced in the bipolar-to-ganglion cell transmission (Roska et al.,
2006; but see McMahon, Packer, & Dacey, 2004). Second, the retinal output is distorted and no longer represents accurately the input. The convergence of ON and OFF channels at the ganglion cell
level would counteract both these effects to a certain extent by
shaping ganglion cell center responses and actively creating their
surrounds. Indeed, evidence in favor of retinal cross-talk between
ON and OFF channels was first found in studies of goldfish ganglion
cells (Levine & Shefner, 1977; Shefner & Levine, 1979), being later
described in ganglion cells of the cat (Mastronarde, 1983a), guinea
pig (Zaghloul et al., 2003), mouse (Renteria et al., 2006) and rabbit
(Roska & Werblin, 2001; Roska et al., 2006), as well as in amacrine
neurons of carp (Kujiraoka, Saito, & Toyoda, 1986; Kujiraoka, Saito,
& Toyoda, 1988) and rabbit (Hsueh, Molnar, & Werblin, 2007).
This cross-talk can happen in various ways, some of which are
illustrated in Fig. 6. First, it can take place via active sign-inverting
synapses from amacrine cells (Fig. 6A and B), as suggested by studies in cat (Mastronarde, 1983a; McGuire, Stevens, & Sterling, 1986;
Wässle, Schafer-Trenkler, & Voigt, 1986) and rabbit (Massey, Redburn, & Crawford, 1983) and demonstrated by recordings of amacrine–ganglion cell pairs in the carp (Toyoda, Shimbo, Kondo, &
Kujiraoka, 1992). Second, cross-talk can also take place via direct
excitatory synapses of ON and OFF bipolar cells onto bistratified
ganglion cells (Fig. 6C), as in the case of carp ON–OFF neurons
(Famiglietti, Kaneko, & Tachibana, 1977) and rabbit ON–OFF DS
cells (Fried, Munch, & Werblin, 2002, 2005; Roska et al., 2006). A
third possibility is the direct convergence of ON and OFF bipolar
950
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
Fig. 6. Some possible modes of cross-talk between ON (yellow) and OFF (brown) channels in the inner plexiform layer. GC, ganglion cell; AC, amacrine cell; BC, bipolar cell.
Excitatory synapses are symbolized by the green arrows, inhibitory synapses by the red arrows. (A) Indirect convergence via amacrine cells. Circuits involved in the
generation of center responses of an ON ganglion cell. This neuron receives direct excitatory input from an ON bipolar cell and inhibitory input from an amacrine cell which, in
turn, is driven by OFF bipolar cells. Adapted from Levine and Shefner (1977). (B) Indirect convergence via amacrine cells. Circuits involved in the generation of center
responses of an OFF ganglion cell. This neuron receives direct excitatory input from an OFF bipolar cell and inhibitory input from an amacrine cell which, in turn, is driven by
ON bipolar cells. Adapted from Levine and Shefner (1977) and Wässle et al. (1986). (C) Direct convergence onto a ganglion cell. Circuits involved in the generation of center
responses from an ON–OFF ganglion cell. This neuron stratifies in both sublamina of the inner plexiform layer, where it receives direct excitatory inputs from ON and OFF
bipolar cells. Adapted from Calkins (2001). (D) Direct convergence onto an ON–OFF amacrine cell. This neuron stratifies in both sublamina of the inner plexiform layer, where
it receives direct excitatory inputs from ON and OFF bipolar cells. Adapted from Toyoda et al. (1973) and Toyoda (1974). (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this paper.)
cells onto ON–OFF amacrine cells (Fig. 6D, Toyoda, 1974; Toyoda et al.,
1973), whose output is not well characterized.
In all of these cases, the result might be an improvement in
signal-to-noise ratio. Schellart and Spekreijse (1973) noted that
post-synaptic sum or subtraction of inputs at the ganglion cell
level can decrease the noise present in the inputs: sum of inputs
would eliminate uncorrelated noise, whereas subtraction would
eliminate correlated noise. However, both sum and subtraction
of inputs could actually improve signal-to-noise ratio, depending
on whether the noise is positively or negatively correlated (for a
detailed description of the types of correlated activity at the ganglion cell level and their origins, see Arnett, 1978 and Meister,
1996) and whether the inputs themselves are modulated in
the same direction or in opposite directions. For instance, the
noise in ON and OFF bipolar cells with partially overlapping
receptive fields is decrementally correlated, since these cells
share at least part of their inputs—this noise would be selectively suppressed if the outputs of these neurons would combine
through direct convergence onto a ganglion cell via excitatory
synapses.
Indirect (via amacrine cells) cross-talk, on the other hand,
sharpens response timing (Demb et al., 2001; McGuire et al.,
1986; Wässle et al., 1986) and reinforces center–surround interactions (Levine & Shefner, 1977; McGuire et al., 1986; Roska et al.,
2006), which also increase signal-to-noise ratio (Meister, 1996;
Roska et al., 2006), help decorrelate photoreceptor signals (Wachtler, Doi, Lee, & Sejnowski, 2007), prevent ganglion cell saturation
(Barlow & Levick, 1976) and decrease spatial blur by constraining
the spread of excitation across the ganglion cell dendritic field
(Roska et al., 2006).
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
The combined result of direct and indirect cross-talk between
ON and OFF pathways, which might be the reason for creating
ON and OFF channels at the bipolar cell level in the first place, is
the improvement of contrast perception, much like in the ‘‘bandwidth doubling” hypothesis (but with a twist). Such a fundamental
need justifies the existence of the ON/OFF division in all vertebrate
retinas and explains why goldfish, guinea pig, rabbit, cat and monkey ganglion cells behave similarly in many ways.
5.3. Overlap between ON and OFF channels might improve visual
acuity
Visual acuity can be defined and measured in a number of ways,
suggesting that different physiological mechanisms might underlie
what we generically call the spatial resolution of the visual system.
In primates, it is believed that midget ganglion cells underlie high
acuity vision, due to their small receptive fields (Wässle & Boycott,
1991; for an alternative view, see Crook, Lange-Malecki, Lee, & Valberg, 1988). Midget ganglion cells are also called P cells and correspond roughly to the b-ganglion cells of the cat (Leventhal,
Rodieck, & Dreher, 1981), the goldfish 2.2 and 2.4 ganglion cells
(Hitchcock & Easter, 1986; Mangrum et al., 2002) and zebrafish
type III and IV neurons (Mangrum et al., 2002).
However, the physiology of the midget/P/b systems and alike
cannot account for all aspects of behaviorally measured visual acu-
951
ity (for examples, see Lennie, Pokorny, & Smith, 1993). This raises
the possibility that spatial resolution might result from the interplay of multiple processing strategies: additionally to the use of
cells with small receptive fields in order to scan the images with
a high number of sampling points, the visual system might also
interpolate larger retinal receptive fields in order to retrieve spatial
information (Eurich & Schwegler, 1997; Heiligenberg, 1987; Meister, 1996). The relative contribution of each of these strategies
to visual acuity may vary with retinal eccentricity and species,
since photoreceptor size and spacing, as well as the degree of retinal convergence, depend on these variables.
Within a single retinal channel, the receptive fields of ganglion
cells already overlap to a certain extent (Cleland, Levick, & Wässle,
1975; Wässle et al., 1981). The overlap is however increased by the
superposition of multiple parallel channels which tile the same retinal area and create intersecting representations of the visual
world with different spatial filtering properties (Eurich & Schwegler, 1997; Vaney & Hughes, 1990). It seems therefore plausible
that visual acuity may profit from the existence of ON and OFF
pathways. In addition to the improved contrast perception yielded
by active retinal cross-talk, as explained in the previous section,
post-retinal comparison of ON and OFF receptive fields might be
a potentially useful source of spatial information.
Hughes (1981), for instance, calculated that the retinal image of
the cat could suffer aliasing if ON and OFF b-ganglion cells would
Fig. 7. Visual acuity might profit from receptive field overlap. (A) The electrophysiologically measured receptive field centers of goldfish ganglion cells in young (bottom) and
adult animals (top) are extremely large. The smallest fields correspond to about 5 deg of visual angle in both cases. From Macy and Easter (1981). (B) Behavioral
measurements of goldfish contrast sensitivity to stationary gratings (left) and gratings drifting at 1 Hz (right). The contrast sensitivity threshold at high luminance can be
regarded as the resolving power of the goldfish visual system. It lies at around 3 cycles/deg, which would correspond to an acuity of about 9 arcmin. From Bilotta and Powers
(1991). (C) Two independent measures of bluegill sunfish vision. The closed symbols show the intercone spacing, which decreases as animals grow. The open symbols are
behaviorally measured reaction times of fish of four sizes to four sizes of prey. As animals grow older and larger, their resolving power seems to increase, which is not what
one would expect based on the receptive field size of ganglion cells, shown in (A). From Hairston et al. (1982).
952
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
sample independently, that is, if the information from ON and OFF
channels would not be combined to extract spatial information.
When treated as a single heterogeneous population instead of
two independent ones, on the other hand, the b-ganglion cell array
yielded cut-off frequencies up to 9 cycles/deg in the area centralis,
which corresponds to the highest acuity values measured psychophysically (8–9 cycles/deg: Jacobson, Franklin, & McDonald, 1976)
and electrophysiologically (8.5 cycles/deg: Harris, 1978).
There is also evidence of such interpolation in fish. The visual
acuity of some species, as measured by different groups, can be
reasonably well explained based on photoreceptor spacing (Bilotta
& Powers, 1991; Hairston, Li, & Easter, 1982; Neumeyer, 2003;
Northmore & Dvorak, 1979). However, since most teleost fish do
not have a fovea with one-to-one connections between photoreceptors and ganglion cells, photoreceptors converge onto ganglion
cells (Kock, 1982; O’Connell, 1963; Stell & Kock, 1984). This convergence could limit visual acuity considerably.
This should not be much of a problem if the ganglion cells involved in visual acuity had small receptive fields, as pointed by
Neumeyer (2003). However, when one examines what happens
with the fish eye in the course of its lifetime, a large paradox
arises: convergence from photoreceptors to ganglion cells either
increases (Kock & Reuter, 1978; Kock & Stell, 1985; Macy & Easter, 1981; Stell & Kock, 1984) or stays the same (Johns & Easter,
1977; Kock, 1982), but acuity actually gets better (Hairston
et al., 1982).
This is illustrated in Fig. 7. Fig. 7A shows that receptive field
sizes (in degrees of visual angle) of goldfish ganglion cells scale
as the animals grow, such that the distribution does not vary considerably with age (Macy & Easter, 1981). In both young and adult
animals, the smallest receptive fields measured electrophysiologically cover about 5 deg of visual angle. This is clearly a much higher
value than the psychophysically determined visual acuity for this
species (Fig. 7B, Bilotta & Powers, 1991). If one calculates spatial
resolution from the contrast sensitivity measurements for stationary and drifting gratings at high luminance, the resulting visual
acuity is about 3.2 cycles/deg or 9 arcmin.
Further evidence that the size of ganglion cell receptive fields
might not be solely responsible for determining visual acuity is
found in Fig. 7C. In the perciform bluegill sunfish, visual acuity
(as measured by the visual angle subtended by prey of different
sizes) improves as the animals grow (Hairston et al., 1982). This
is consistent with acuity being better matched by the spacing in
the photoreceptor array (filled symbols in Fig. 7C), which decreases as fish grow due to the addition of new photoreceptor
cells and the increase in the lens magnification factor (Hairston
et al., 1982; Johns & Easter, 1977; Macy & Easter, 1981; Stell
& Kock, 1984). Taken together, these results point to the possibility that the partly coextensive receptive fields of adjacent cells
be interpolated somewhere downstream in the visual system.
6. Broadband and opponent channels: Strategies for optimal
coding
So far we have examined the retinal processing of photoreceptor signals without taking the spectral dimension into account.
Cone photoreceptors come however in different flavors as far as
their spectral sensitivity is concerned. It is the differential processing of their signals that ultimately leads to what we call color vision. It is therefore instructive to study how the retina copes
with colored stimuli, and what kind of information it can extract
from them.
When one observes the behavior of retinal neurons regarding
wavelength information, two large cell classes emerge: broadband
neurons encompass both ON and OFF types and respond with the
same polarity to stimulation throughout the visible spectrum,
while opponent neurons do not: the response polarity of these cells
depends on the spectral characteristics of the stimulus.
The ubiquity of spectrally opponent neurons in the most diverse
animal classes indicates that this coding strategy has a long evolutionary history. However, the literature regarding the retinal
broadband and opponent channels is somewhat confusing, because
it is permeated by a number of implicit assumptions that may or
may not hold. This makes interclass comparisons difficult, espe-
Fig. 8. Color vision and spectral opponency. (A) Theoretic chromatic and achromatic response functions of the CIE standard observer. The chromatic functions were calculated
based on hue cancellation experiments, in which a color sensation (for instance, blue) elicited by a given wavelength is cancelled by the superposition of a certain amount of
its complementary color (in this case, yellow). The relative energy of the yellow stimulus needed to cancel the blue sensation at each wavelength is then taken as a measure of
the strength of blue perception throughout the spectrum. The achromatic function is simply the foveal achromatic brightness threshold (or luminosity function) of the human
eye. From Hurvich and Jameson (1957). (B) Spectral coding of carp cone-driven horizontal cells. The action spectrum of monophasic horizontal cells (black line) resembles the
human achromatic functions, whereas the responses of biphasic (green/red line) and triphasic cells (blue/yellow line) resemble the two human chromatic channels. From
Norton et al. (1968). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
cially because the underlying assumptions of many studies vary
according to the animal class studied. In the next sections, we will
evaluate a number of these assumptions and examine some possible functions for these visual channels in both fish and mammalian
visual systems.
6.1. Are these channels really related to color and brightness?
The idea of opponent or antagonistic processes as the basic
physiological substrate of our ability to perceive colors was originally developed by Hering (1874). Color vision is defined as an
opponent process because certain color percepts such as red and
green, blue and yellow or black and white cannot take place concomitantly (Hurvich & Jameson, 1957).
Fig. 8A illustrates this idea. The graph plots the average human
achromatic (black line) and chromatic (colored lines) functions,
measured with two different experimental paradigms. The achromatic curve is simply the human photopic luminance function
(also called spectral sensitivity) measured by a threshold technique
(Hurvich & Jameson, 1955; Lennie et al., 1993). The chromatic
functions were measured through cancellation experiments, which
take advantage of the fact that opponent colors cannot be experienced at the same time (Hurvich & Jameson, 1957). For each test
wavelength that elicits a certain hue sensation (for instance,
green), a second test light of its opposite color (in this case, red)
is presented concomitantly and the energy of this second light is
varied until the sensation of ‘‘greenness” is cancelled. The amount
of red light needed to cancel the green sensation is then plotted as
the intensity of the green process for that particular wavelength.
By repeating this procedure throughout the whole spectrum, the
green/red and blue/yellow opponent curves are obtained.
The initial finding of opponent neurons in the teleost retina
(Svaetichin, 1956; Svaetichin & MacNichol, 1958; Wagner, MacNichol, & Wolbarsht, 1960) and the relative similarity between
the responses of these neurons with the human color opponent
channels (Fig. 8B, redrawn from Norton, Spekreijse, Wolbarsht,
& Wagner, 1968) led to the assumption that opponent neurons
are the physiological correlate of Hering’s theory and therefore
directly underlie color perception (Hurvich & Jameson, 1960;
Naka & Rushton, 1966a; Orlov & Maximova, 1965; Svaetichin
& MacNichol, 1958; Wagner et al., 1960). This idea was
strengthened by the subsequent finding of opponent cells in
the primate LGN (De Valois, 1960; De Valois, Abramov, & Jacobs, 1966; Wiesel & Hubel, 1966), although primate horizontal
cells were later shown not to be spectrally opponent, as discussed in the following paragraphs (Dacey, Lee, Stafford, Pokorny, & Smith, 1996).
It is however not entirely clear whether retinal broadband and
opponent channels directly underlie the perception of brightness
and color. There is plenty of evidence that they may not. Whereas
opponent neurons do carry some information about the wavelength composition of the stimulus, the behavior of such cells is
not consistent with that of a color detector. Examples of mismatches between the physiology of retinal neurons and their putative perceptual correlates can be found in both fish and
mammalian retinas.
For instance, the response polarity of fish retinal opponent neurons depends not only on wavelength, but also on the intensity of
the stimulus. When one increases the intensity of the light stimulation, red–green opponency is lost both at the horizontal (Gottesman & Burkhardt, 1987; Joselevitch, 1999; Naka & Rushton,
1966b), bipolar (Joselevitch & Kamermans, 2007; Kaneko & Tachibana, 1981; Kaneko & Tachibana, 1983) and ganglion cell levels
(Van Dijk & Spekreijse, 1983; Wagner et al., 1960), but color vision
is still present in the same animals (Neumeyer, 1986; Neumeyer &
Arnold, 1987). Primate horizontal cells are not spectrally opponent
953
(Dacey et al., 1996), and yet color vision in primates is well
developed.
There is further evidence in the primate visual system against
the idea that the different types of spectrally opponent neurons
correspond to the psychophysical opponent channels described
by Hering (1874). Ganglion cells with red/green opponency, for instance, receive also inputs from S-cones (de Monasterio, 1979b),
which should not participate in the red–green perceptual axis
(Hurvich & Jameson, 1957). These opponent ganglion cells, furthermore, behave as broadband detectors at high temporal frequencies
(Gouras & Zrenner, 1979; Ingling & Martinez-Uriegas, 1983), suggesting they might also inform the visual system about brightness
(for more on this matter, see Stockman & Plummer, 2005; Stockman, Plummer, & Montag, 2005). This also holds for primate opponent LGN units, which respond to both chromatic and achromatic
stimuli (De Valois, 1972; De Valois & Pease, 1971).
On top of this, primate opponent ganglion cells that correspond
to the P/midget type discussed in the previous sections have small
receptive field centers and therefore seem to contribute to visual
acuity as well, which is essentially achromatic (Ingling & Martinez-Uriegas, 1983; Lennie et al., 1993). The perceptual color vision
system, on the other hand, has a much lower spatial resolution
(Mullen, 1985). Lastly, no spectrally opponent neuron described
so far can explain the invariability of yellow, the perception of unique blue (Abramov & Gordon, 2005), unique green (Kaplan, Lee, &
Shapley, 1990) or the sensation of redness at the short end of the
visible spectrum (Abramov & Gordon, 2005; Stockman & Plummer,
2005).
It is nonetheless indisputable that the creation of broadband
and opponent channels forms a fundamental step in the coding
process that leads to visual perception in all its richness. However,
these early retinal pathways may not be directly related to visual
percepts on a one-to-one basis, since they are most likely involved
in transmitting information about multiple perceptual dimensions
concomitantly, as suggested by a number of studies (Calkins &
Sterling, 1999; De Valois, 1972; Eurich & Wilke, 2000; Ingling &
Martinez-Uriegas, 1983; Zhang & Sejnowski, 1999).
6.2. Was color vision really reinvented in evolution?
When comparing species, one can look at similarities or at differences. In this review, we have chosen the first option. Primate
literature, on the other hand, is filled with examples of the last
strategy, in which the evolutionary gap between cold-blooded vertebrates and primates (or even between dichromatic mammals and
primates) is emphasized (see for instance Kaplan et al., 1990; Shapley & Perry, 1986). Therefore, the idea that color vision has developed independently many times during evolution has been called
forth repeatedly (Bowmaker & Hunt, 2006; De Valois, 1960; Jacobs
& Rowe, 2004; Mollon, 1989; Walls, 1942).
During the course of evolution, most mammals became monoor dichromats. Some species re-evolved an L-cone opsin through
gene duplication shortly after the African and American continents
separated (Mollon, 1989; Nathans, 1999). An interesting question
is how mammals first adapted to dichromacy and then readapted
to the newly acquired photoreceptor type. Did they first lose the
post-synaptic machinery needed for trichromacy and had to create
new coding strategies when the new opsin appeared, or did they
keep the old trichromatic machinery and used it when the L-cone
opsin became available?
The physiological substrate for the ability to discriminate colors
is thought to lie not only in the existence of multiple types of
cones, but in cell-specific connectivity patterns between photoreceptors and second-order neurons. For instance, the existence of
bipolar cells that connect exclusively to S-cones or M-cones as opposed to collecting from both sorts of photoreceptors is well docu-
954
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
mented in cyprinid fish (Scholes, 1975; Stell, 1980), rodents (Haverkamp et al., 2005; Li & DeVries, 2006), lagomorphs (Liu & Chiao,
2007) and primates (Klug, Herr, Ngo, Sterling, & Schein, 2003;
Kouyama & Marshak, 1992; Mariani, 1984).
Even though the response properties of these neurons is not
known, one assumes that this photoreceptor-specific connectivity
reflects photoreceptor-specific responses (i.e. that there is no
cross-talk between channels in the outer or inner nuclear layers).
Given the apparent specificity of this ‘‘primordial” color vision system, one has long believed that the ‘‘late” trichromatic system of
primates would also depend on cone-specific connections in addition to the appearance of a third cone type (Lee, 1996; Shapley &
Perry, 1986). This idea would indeed require that the primate trichromatic system be reinvented from scratch and could explain,
for instance, why women carriers for anomalous trichromacy,
who have four types of cones, do not develop tetrachromacy (Jordan & Mollon, 1993; but see Wachtler et al., 2007 for an alternative
explanation).
The specific-connectivity hypothesis for this ‘‘late” color vision
system is however still a matter of debate. One intriguing result
is the emergence of trichromatic vision in knock-in mice expressing a third cone opsin, with no obvious retinal rearrangement (Jacobs, Williams, Cahill, & Nathans, 2007; Onishi et al., 2005;
Smallwood et al., 2003).
The results from these groups point to the possibility that,
although most mammals lack one or more opsin as compared to
fish and primates, the rest of the machinery and of the strategies
used for color processing was not lost through evolution. In fact,
there is evidence that color vision might have appeared in a common ancestor to teleosts and mammals, long before terrestrial vertebrates evolved from aquatic forms (Burkhardt, Gottesman,
Levine, & MacNichol, 1983). This would mean that color vision
was not invented many times from scratch, bringing the chimp a
little closer to the goldfish.
In view of these results, one should consider the possibility that
there may be more to color vision than hardwired selective photoreceptor connections and retinal spectrally opponent cells. Color
vision might still be possible without them, as long as the ratios
of cone inputs to different types of second- and third-order neurons differ, and that these ratios be retrieved and compared somewhere downstream.
6.2.1. Horizontal cells as a substrate for color constancy in fish and
primates
A long held discussion concerns the function of horizontal cells.
While it is generally accepted that they form the receptive field
surrounds of bipolar cells (Toyoda & Tonosaki, 1978a, 1978b) and
may contribute indirectly to the surrounds of some ganglion cells
(Mangel & Miller, 1987; McMahon et al., 2004), it has been argued
over and over that teleost and primate horizontal cells cannot possibly have the same functions, since the first can be spectrally
opponent (for review, see Kamermans & Spekreijse, 1995), while
the second does not have the same characteristics (see for instance
Dacey et al., 1996; Kaplan et al., 1990). One needs however to realize that this argument is embedded in two assumptions that we
have just falsified: first, that retinal spectrally opponent neurons
are directly related to color vision while broadband cells are not;
second, that primates are ‘‘different” and recreated trichromacy
from scratch.
Some studies have addressed this issue and demonstrated that
the differences between primate and teleost horizontal cells might
be more of a qualitative matter than a functional one. For instance,
Kraaij, Kamermans, and Spekreijse (1998) have shown that
although goldfish horizontal cells are spectrally opponent, their
output is not: the feedback signal they send to cones is broadband.
Subsequently, Kamermans, Kraaij, and Spekreijse (1998) demon-
strated that because this feedback signal corrects the output of
the photoreceptors to the spectral composition of the illuminant,
it is the ideal neuronal substrate for color constancy. Psychophysical experiments had already shown that the goldfish is a color
constant animal (Dorr & Neumeyer, 1996; Neumeyer, Dorr, & Fritsch, 1997) and performs much like humans in similar behavioral
tasks (Neumeyer, Dorr, Fritsch, & Kardelky, 2002).
Finally, VanLeeuwen, Joselevitch, Fahrenfort, and Kamermans
(2007) used primate cone and horizontal cell spectra and extended
these results from Kamermans et al. (1998) to the primate retina,
showing that essentially the same conclusions apply for fish and
mammals. So although primate horizontal cells are not spectrally
opponent, they correct the output of the photoreceptors much in
the same way as fish horizontal cells do.
But why are fish horizontal cells selectively connected to specific cone types, and why are some of them spectrally opponent?
There are multiple explanations. One of them is that the spectral
opponency of fish horizontal cells might result from the ecological needs of these animals. Not only is the spectral composition of
aquatic environments much more variable than that of terrestrial
ones (Lythgoe, 1979), but the visual system of these fish might require larger corrections in the photoreceptor output to remain
color constant, since they are more sensitive to the UV and red
parts of the spectrum than mammals and therefore more susceptible to spectral changes in the environment (VanLeeuwen et al.,
2007).
6.3. Types of opponency: The architecture of outer and inner retinal
channels
This brings us to the question: How is the division of the visual
system into opponent and broadband channels achieved? Spectral
opponency may have different origins and appear at many retinal
sites. For instance, it can result from direct excitatory inputs to the
receptive field center of a cell, as in the case of teleost opponent
bipolar neurons in the outer retina (Wong & Dowling, 2005), or it
can emerge from center–surround interactions, as in the case of
teleost horizontal cells (Kamermans & Spekreijse, 1995), primate
opponent bipolar (Dacey & Lee, 2001) and ganglion cells (for reviews, see Kaplan et al., 1990; Wässle, 2004).
This division of photoreceptor signals into broadband and opponent channels was however long thought to happen at different
retinal locations in fish and mammals. It was argued that since fish
species have a smaller brain when compared to mammals, they
might need to ‘‘compute more” in their retinas (see for instance
Bilotta & Abramov, 1989). This would be the reason for the emergence of opponent horizontal and bipolar cells already at the first
retinal synapse in these animals. Such an argument, however, appears more and more to be a fallacy. Although the spectral properties of mammalian bipolar cells are to a great extent unknown,
recent evidence indicates that also in the primate retina some
opponent channels are already formed at the photoreceptor–bipolar cell synapse (Dacey & Lee, 2001). It will not be surprising to see
more similarities appear with further study of mammalian bipolar
cell physiology.
6.3.1. Fish double-opponent neurons: Another special case?
Spectrally opponent ganglion cells are quite common in the retinas of fish (Daw, 1967, 1968; Wagner et al., 1960). In these neurons, opponent interactions can already take place within the
receptive field center. In primates, on the other hand, spectral
opponency at the bipolar and ganglion cell level was found to result mostly from spatial interactions between center and surround
processes with different spectral sensitivities (Calkins & Sterling,
1999; Wässle, 2004): opponent processes, therefore, do not seem
to be spatially coextensive.
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
There are, however, exceptions to this rule. One of them is the
bistratified ganglion cell of the primate, which receives S-cone input (Dacey & Lee, 1994). This cell has spectrally opponent light responses throughout its whole receptive field center and no clear
surround antagonism (Dacey, 2000; Dacey & Lee, 1994). Its opponency, therefore, is most likely a result of direct antagonistic inputs
at the synapse between cones and the bipolar cells that project to
this ganglion cell (Calkins, Tsukamoto, & Sterling, 1998). Other single-opponent units with no center–surround organization were recorded sporadically from the cat (Cleland & Levick, 1974; Ringo &
Wolbarsht, 1986) and primate retina (de Monasterio, 1978), suggesting that the dissimilarities found so far between fish and mammalian single-opponent ganglion cells might reflect more a
sampling bias than deep species-specific differences.
A substantially large part of fish ganglion cells have yet more
complex receptive fields. In these cells, both the center and the surround processes are spectrally opponent (Daw, 1967, 1972; Spekreijse, Wagner, & Wolbarsht, 1972). Such neurons have not been
observed in the primate retina (Daw, 1972, 1984), but only in higher visual areas (see for instance Hubel & Wiesel, 1968; Michael,
1978a, 1978b). This difference in spectral sensitivity was interpreted as evidence that fish and mammalian visual systems are
fundamentally different. The properties of primate double-opponent neurons were suggested to arise from the convergence of single-opponent cells (Daw, 1973, 1984; Michael, 1978a, 1978b),
whereas the double-opponency of fish ganglion cells was suggested to derive directly from bipolar cells, which can also be double-opponent in the fish retina (Daw, 1973; Kaneko & Tachibana,
1981, 1983; Shimbo et al., 2000).
However, the fact that double-opponent ganglion cells have
been sporadically recorded from in the cat (Ringo & Wolbarsht,
1986) makes it difficult to justify why only double-opponent cells
in the primate should be the result of higher processing and in
other mammals not. Rather, it suggests that the spectral properties
of the ganglion cells described so far depend on the luck of the
experimenter, the absorption spectra of the photoreceptors involved and their retinal distribution (Buchsbaum & Gottschalk,
1983; Paulus & Kroger-Paulus, 1983; Ruderman, Cronin, & Chiao,
1998), and does not really point to fundamental differences in
the organization of visual pathways in fish and mammals.
6.4. Broadband and opponent channels in evolution: Similar tasks lead
to similar solutions
Even though opponent and broadband channels might be created via different mechanisms at distinct retinal locations, the
underlying reason for this creation does not seem to differ: the
optimization of retinal bandwidth. Because the absorption spectra
of the different cone types overlap to a great extent, their activity is
in great part correlated (Buchsbaum & Gottschalk, 1983; Ruderman et al., 1998). This is true for all vertebrate species, although
the number of cone types and their spectral sensitivities differ. If
this correlation would not be at least partially removed from the
retinal code, ganglion cells would need to produce more costly
spikes to transmit visual information (Laughlin, 2001, 2003).
Part of this correlation and of the redundancy contained herein
is removed by having signals from distinct photoreceptor types
converge onto second- and third-order neurons with different
signs (Barlow, 1981; Buchsbaum & Gottschalk, 1983; Ruderman
et al., 1998; Schellart & Spekreijse, 1973). As a result, visual pathways are subdivided into broadband and opponent channels. The
specific types of spectral coding that result from these combinations depend on a number of factors, such as the spectra of the
photoreceptors involved and the organization of the cone mosaics
in different species and/or retinal areas (Buchsbaum & Gottschalk,
1983; Paulus & Kroger-Paulus, 1983; Ruderman et al., 1998).
955
This coding strategy enables the visual system not only to reduce redundancy in its neural code, but also to achieve great spectral resolution. Even though photoreceptors have broadly
overlapping spectral sensitivities, this post-receptoral processing
enables primates to discriminate wavelength differences of up to
2 nm (De Valois & Morgan, 1974; Hurvich & Jameson, 1957). In
the goldfish, wavelength discrimination can reach astounding
4 nm acuity in some parts of the spectrum (Neumeyer, 1986), even
though the photoreceptor spectra in this species do not overlap as
much as in primates (Marks, 1965; Mooij & Van den Berg, 1983;
Palacios, Varela, Srivastava, & Goldsmith, 1998; Stell & Harosi,
1976; Tsin, Liebman, Beatty, & Drzymala, 1981). One should note
here, however, that the retinal opponent channels might not support color perception alone; the combined activity of broadband
and opponent cells is needed for sending color information.
7. How parallel is parallel?
As observed by Merigan and Maunsell (1993), the adjective
‘‘parallel” is rather inadequate to describe visual pathways, since
their segregation is not complete. Although they were actually
talking about post-retinal processing, we hope it has by now become clear that the same reasoning applies to within the retina, given the extensive cross-talk between its many channels.
Not only rods converge onto cone systems and ON and OFF
pathways cross-talk, as discussed in the previous paragraphs, but
also broadband and opponent channels intermingle. After all,
opponency is the result of the convergence of broadband inputs
of opposing sign onto a neuron, whether in the center or surround
of its receptive field. Broadband channels can also ‘‘piggyback” on
an opponent system: the convergence of rod and cone inputs onto
cyprinid (Beauchamp & Daw, 1972; Raynauld, 1972), cat (Cleland &
Levick, 1974) and primate (Daw, 1972) ganglion cells, for example,
make these cells behave as broadband in the dark-adapted, and as
spectrally opponent units in the light-adapted state.
What all of these retinal channels have in common is that, in
addition to being present in all vertebrates, they also cross-talk
in a conserved manner. Retinal coding strategies, therefore, do
not differ fundamentally in the zebrafish or carp as compared to
the squirrel, cat or gorilla.
Acknowledgments
The authors thank Dr. Jan Klooster for providing the EM picture
in Fig. 4C. C.J. was financially supported by Conselho Nacional de
Pesquisa (CNPq), Brazil (Grant 200915/98-3 to C.J.). M.K. is supported by grants from ALW-NWO, ZonMW-NWO, EOARD and
AFOSR.
References
Abramov, I., & Gordon, J. (2005). Seeing unique hues. Journal of the Optical Society of
America A, Optics, Image Science, and Vision, 22, 2143–2153.
Anderson, D. H., & Fisher, S. K. (1976). The photoreceptors of diurnal squirrels:
Outer segment structure, disc shedding, and protein renewal. Journal of
Ultrastructure Research, 55, 119–141.
Arnett, D. W. (1978). Statistical dependence between neighboring retinal ganglion
cells in goldfish. Experimental Brain Research, 32, 49–53.
Ashmore, J. F., & Falk, G. (1980a). Response of rod bipolar cells in the dark-adapted
retina of the dogfish, scyliorhinus canicula. The Journal of Physiology, 300,
115–150.
Ashmore, J. F., & Falk, G. (1980b). The single-photon signal in rod bipolar cells of the
dogfish retina. The Journal of Physiology, 300, 151–166.
Ashmore, J. F., & Falk, G. (1982). An analysis of voltage noise in rod bipolar cells of
the dogfish retina. The Journal of Physiology, 332, 272–297.
Barlow, H. B. (1952). Eye movements during fixation. The Journal of Physiology, 116,
290–306.
Barlow, H. B. (1981). The Ferrier Lecture, 1980. Critical limiting factors in the design
of the eye and visual cortex. Proceedings of the Royal Society of London Series B,
Containing Papers of Biological Character, 212, 1–34.
956
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
Barlow, H. B., & Levick, W. R. (1976). Threshold setting by the surround of cat retinal
ganglion cells. The Journal of Physiology, 259, 737–757.
Barlow, H. B., Levick, W. R., & Yoon, M. (1971). Responses to single quanta of light in
retinal ganglion cells of the cat. Vision Research, Suppl. 3, 87–101.
Baylor, D. A., Lamb, T. D., & Yau, K. W. (1979). Responses of retinal rods to single
photons. The Journal of Physiology, 288, 613–634.
Beauchamp, R. D., & Daw, N. W. (1972). Rod and cone input to single goldfish optic
nerve fibers. Vision Research, 12, 1201–1212.
Beaudet, L., Browman, H. I., & Hawryshyn, C. W. (1993). Optic nerve response and
retinal structure in rainbow trout of different sizes. Vision Research, 33,
1739–1746.
Berntson, A., Smith, R. G., & Taylor, W. R. (2004). Transmission of single photon
signals through a binary synapse in the mammalian retina. Visual Neuroscience,
21, 693–702.
Berntson, A., & Taylor, W. R. (2000). Response characteristics and receptive field
widths of on-bipolar cells in the mouse retina. The Journal of Physiology, 524,
879–889.
Bilotta, J., & Abramov, I. (1989). Spatiospectral properties of goldfish retinal
ganglion cells. Journal of Neurophysiology, 62, 1140–1148.
Bilotta, J., & Powers, M. K. (1991). Spatial contrast sensitivity of goldfish: Mean
luminance, temporal frequency and a new psychophysical technique. Vision
Research, 31, 577–585.
Bowmaker, J. K., & Hunt, D. M. (2006). Evolution of vertebrate visual pigments.
Current Biology, 16, R484–R489.
Boycott, B. B., Dowling, J. E., & Kolb, H. (1969). Organization of the primate retina:
Light microscopy. Philosophical transactions of the Royal Society of London Series
B, Biological Sciences, 255, 109–184.
Boycott, B. B., & Wässle, H. (1999). Parallel processing in the mammalian retina: The
proctor lecture. Investigative Ophthalmology & Visual Science, 40, 1313–1327.
Buchsbaum, G., & Gottschalk, A. (1983). Trichromacy, opponent colors coding and
optimum colour information transmission in the retina. Proceedings of the Royal
Society of London Series B, Containing Papers of Biological Character, 220, 89–113.
Burkhardt, D. A., Gottesman, J., Levine, J. S., & MacNichol, E. F. (1983). Cellular
mechanisms for color-coding in holostean retinas and the evolution of color
vision. Vision Research, 23, 1031–1041.
Cajal, S. R. (1893). ‘La retine des vertebres’ translated by Maguire, D. and Rodieck, R.
W. In R. W. Rodieck (Ed.), The vertebrate retina (pp. 775–904). San Francisco:
Freeman.
Calkins, D. J. (2001). Seeing with S cones. Progress in Retinal and Eye Research, 20,
255–287.
Calkins, D. J., & Sterling, P. (1999). Evidence that circuits for spatial and color vision
segregate at the first retinal synapse. Neuron, 24, 313–321.
Calkins, D. J., Tsukamoto, Y., & Sterling, P. (1998). Microcircuitry and mosaic of a
blue–yellow ganglion cell in the primate retina. The Journal of Neuroscience, 18,
3373–3385.
Chichilnisky, E. J., & Kalmar, R. S. (2002). Functional asymmetries in ON and OFF
ganglion cells of primate retina. The Journal of Neuroscience, 22, 2737–2747.
Cleland, B. G., & Levick, W. R. (1974). Properties of rarely encountered types of
ganglion cells in the cat’s retina and an overall classification. The Journal of
Physiology, 240, 457–492.
Cleland, B. G., Levick, W. R., & Wässle, H. (1975). Physiological identification of a
morphological class of cat retinal ganglion cells. The Journal of Physiology, 248,
151–171.
Cohen, A. I. (1964). Some observations on the fine structure of the retinal receptors
of the American gray squirrel. Investigative Ophthalmology, 3, 198–216.
Cohen, A. I. (1965). Some electron microscopic observations on inter-receptor
contacts in the human and macaque retinae. Journal of Anatomy, 99, 595–610.
Cohen, E. D., & Miller, R. F. (1999). The network-selective actions of quinoxalines on
the neurocircuitry operations of the rabbit retina. Brain Research, 831, 206–228.
Connaughton, V. P., Graham, D., & Nelson, R. (2004). Identification and
morphological classification of horizontal, bipolar, and amacrine cells within
the zebrafish retina. The Journal of Comparative Neurology, 477, 371–385.
Crook, J. M., Lange-Malecki, B., Lee, B. B., & Valberg, A. (1988). Visual resolution of
macaque retinal ganglion cells. The Journal of Physiology, 396, 205–224.
Dacey, D. M. (2000). Parallel pathways for spectral coding in primate retina. Annual
Review of Neuroscience, 23, 743–775.
Dacey, D. M., & Lee, B. B. (1994). The ‘blue-on’ opponent pathway in primate retina
originates from a distinct bistratified ganglion cell type. Nature, 367, 731–735.
Dacey, D. M., & Lee, B. B. (2001). Functional architecture of cone signal pathways in
the primate retina. In K. Gegenfurtner & L. T. Sharpe (Eds.), Color vision. From
genes to perception (pp. 181–202). Cambridge: Cambridge University Press.
Dacey, D. M., Lee, B. B., Stafford, D. K., Pokorny, J., & Smith, V. C. (1996). Horizontal
cells of the primate retina: Cone specificity without spectral opponency. Science,
271, 656–659.
Dacey, D. M., & Petersen, M. R. (1992). Dendritic field size and morphology of
midget and parasol ganglion cells of the human retina. Proceedings of the
National Academy of Sciences of the United States of America, 89, 9666–9670.
Dacheux, R. F., & Raviola, E. (1982). Horizontal cells in the retina of the rabbit. The
Journal of Neuroscience, 2, 1486–1493.
Dacheux, R. F., & Raviola, E. (1986). The rod pathway in the rabbit retina: A
depolarizing bipolar and amacrine cell. The Journal of Neuroscience, 6, 331–345.
Daw, N. W. (1967). Goldfish retina: Organization for simultaneous color contrast.
Science, 158, 942–944.
Daw, N. W. (1968). Colour-coded ganglion cells in the goldfish retina: Extension of
their receptive fields by means of new stimuli. The Journal of Physiology, 197,
567–592.
Daw, N. W. (1972). Color-coded cells in goldfish, cat, and rhesus monkey.
Investigative Ophthalmology, 11, 411–417.
Daw, N. W. (1973). Neurophysiology of color vision. Physiological Reviews, 53,
571–611.
Daw, N. W. (1984). The psychology and physiology of colour vision. Trends in
Neurosciences, 7, 330–335.
de Monasterio, F. M. (1978). Properties of ganglion cells with atypical receptivefield organization in retina of macaques. Journal of Neurophysiology, 41,
1435–1449.
de Monasterio, F. M. (1979a). Asymmetry of on- and off-pathways of blue-sensitive
cones of the retina of macaques. Brain Research, 166, 39–48.
de Monasterio, F. M. (1979b). Signals from blue cones in ‘‘red–green” opponentcolour ganglion cells of the macaque retina. Vision Research, 19, 441–449.
De Valois, R. L. (1960). Color vision mechanisms in the monkey. The Journal of
General Physiology, 43(Suppl. 6), 115–128.
De Valois, R. L. (1972). Processing of intensity and wavelength information by the
visual system. Investigative Ophthalmology, 11, 417–427.
De Valois, R. L., Abramov, I., & Jacobs, G. H. (1966). Analysis of response patterns of
LGN cells. Journal of the Optical Society of America A, Optics, Image Science, and
Vision, 56, 966–977.
De Valois, R. L., & Morgan, H. C. (1974). Psychophysical studies of monkey vision—I.
Macaque luminosity and color vision tests. Vision Research, 14, 53–67.
De Valois, R. L., & Pease, P. L. (1971). Contours and contrast: Responses of monkey
lateral geniculate nucleus cells to luminance and color figures. Science, 171,
694–696.
Demb, J. B., Zaghloul, K., Haarsma, L., & Sterling, P. (2001). Bipolar cells contribute to
nonlinear spatial summation in the brisk-transient (Y) ganglion cell in
mammalian retina. The Journal of Neuroscience, 21, 7447–7454.
DeVries, S. H., & Baylor, D. A. (1995). An alternative pathway for signal flow from rod
photoreceptors to ganglion cells in mammalian retina. Proceedings of the
National Academy of Sciences of the United States of America, 92, 10658–10662.
Dorr, S., & Neumeyer, C. (1996). The goldfish—a colour—constant animal. Perception,
25, 243–250.
Dowling, J. E., & Boycott, B. B. (1966). Organization of the primate retina: Electron
microscopy. Proceedings of the Royal Society of London Series B, Containing Papers
of Biological Character, 166, 80–111.
Dryja, T. P., McGee, T. L., Berson, E. L., Fishman, G. A., Sandberg, M. A., Alexander, K.
R., et al. (2005). Night blindness and abnormal cone electroretinogram ON
responses in patients with mutations in the GRM6 gene encoding mGluR6.
Proceedings of the National Academy of Sciences of the United States of America,
102, 4884–4889.
Enroth-Cugell, C., & Robson, J. G. (1966). The contrast sensitivity of retinal ganglion
cells of the cat. The Journal of Physiology, 187, 517–552.
Euler, T., & Masland, R. H. (2000). Light-evoked responses of bipolar cells in a
mammalian retina. Journal of Neurophysiology, 83, 1817–1829.
Euler, T., & Wässle, H. (1995). Immunocytochemical identification of cone
bipolar cells in the rat retina. The Journal of Comparative Neurology, 361,
461–478.
Eurich, C. W., & Schwegler, H. (1997). Coarse coding: Calculation of the resolution
achieved by a population of large receptive field neurons. Biological Cybernetics,
76, 357–363.
Eurich, C. W., & Wilke, S. D. (2000). Multidimensional encoding strategy of spiking
neurons. Neural Computation, 12, 1519–1529.
Falk, G. (1988). Signal transmission from rods to bipolar and horizontal cells: A
synthesis. Progress in Retinal Research, 8, 255–279.
Famiglietti, E. V. (1981). Functional architecture of cone bipolar cells in mammalian
retina. Vision Research, 21, 1559–1563.
Famiglietti, E. V., Kaneko, A., & Tachibana, M. (1977). Neuronal Architecture of On
and Off pathways to ganglion cells in carp retina. Science, 198, 1267–1269.
Field, G. D., & Chichilnisky, E. J. (2007). Information processing in the primate retina:
Circuitry and coding. Annual Review of Neuroscience, 30, 1–30.
Field, G. D., & Rieke, F. (2002). Nonlinear signal transfer from mouse rods to bipolar
cells and implications for visual sensitivity. Neuron, 34, 773–785.
Fried, S. I., Munch, T. A., & Werblin, F. S. (2002). Mechanisms and circuitry
underlying directional selectivity in the retina. Nature, 420, 411–414.
Fried, S. I., Munch, T. A., & Werblin, F. S. (2005). Directional selectivity is formed at
multiple levels by laterally offset inhibition in the rabbit retina. Neuron, 46,
117–127.
Fyk-Kolodziej, B., Qin, P., & Pourcho, R. G. (2003). Identification of a cone bipolar cell
in cat retina which has input from both rod and cone photoreceptors. The
Journal of Comparative Neurology, 464, 104–113.
Ghosh, K. K., Bujan, S., Haverkamp, S., Feigenspan, A., & Wässle, H. (2004). Types of
bipolar cells in the mouse retina. The Journal of Comparative Neurology, 469,
70–82.
Gottesman, J., & Burkhardt, D. A. (1987). Response properties of C-type horizontal
cells in the retina of the bowfin. Vision Research, 27, 179–189.
Gouras, P., & Zrenner, E. (1979). Enhancement of luminance flicker by coloropponent mechanisms. Science, 205, 587–589.
Grant, G. B., & Dowling, J. E. (1995). A glutamate-activated chloride current in conedriven ON bipolar cells of the white perch retina. The Journal of Neuroscience, 15,
3852–3862.
Grant, G. B., & Dowling, J. E. (1996). On bipolar cell responses in the teleost retina
are generated by two distinct mechanisms. Journal of Neurophysiology, 76,
3842–3849.
Hack, I., Peichl, L., & Brandstatter, J. H. (1999). An alternative pathway for rod signals
in the rodent retina: Rod photoreceptors, cone bipolar cells, and the localization
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
of glutamate receptors. Proceedings of the National Academy of Sciences of the
United States of America, 96, 14130–14135.
Hairston, N. G., Jr., Li, K. T., & Easter, S. S. Jr., (1982). Fish vision and the detection of
planktonic prey. Science, 218, 1240–1242.
Harris, L. R. (1978). Contrast sensitivity and acuity of a conscious cat measured by
the occipital evoked potential. Vision Research, 18, 175–178.
Haverkamp, S., Wässle, H., Duebel, J., Kuner, T., Augustine, G. J., Feng, G., et al.
(2005). The primordial, blue-cone color system of the mouse retina. The Journal
of Neuroscience, 25, 5438–5445.
Heiligenberg, W. (1987). Central processing of sensory information in electric fish.
Journal of Comparative Physiology A, Sensory, Neural, and Behavioral Physiology,
161, 621–631.
Hering, E. (1874). Outlines of a theory of the light sense (Translated by Hurvich, L. M.
and Jameson, D. in 1964). Cambridge, Mass.: Harvard University Press.
Hitchcock, P. F., & Easter, S. S. (1986). Retinal ganglion cells in goldfish: A qualitative
classification into four morphological types, and a quantitative study of the
development of one of them. The Journal of Neuroscience, 6, 1037–1050.
Hochstein, S., & Shapley, R. M. (1976). Linear and nonlinear spatial subunits in Y cat
retinal ganglion cells. The Journal of Physiology, 262, 265–284.
Hornstein, E. P., Verweij, J., Li, P. H., & Schnapf, J. L. (2005). Gap-junctional coupling
and absolute sensitivity of photoreceptors in macaque retina. The Journal of
Neuroscience, 25, 11201–11209.
Hsueh, H.-A., Molnar, A. C., & Werblin, F. S. (2007). Mammalian amacrine cells
receive only four forms of inhibition. ARVO Abstracts, E-abstract 1940.
Huang, L., Max, M., Margolskee, R. F., Su, H., Masland, R. H., & Euler, T. (2003). G
protein subunit G gamma 13 is coexpressed with G alpha o, G beta 3, and G beta
4 in retinal ON bipolar cells. The Journal of Comparative Neurology, 455, 1–10.
Hubel, D. H., & Wiesel, T. N. (1968). Receptive fields and functional architecture of
monkey striate cortex. The Journal of Physiology, 195, 215–243.
Hughes, A. (1981). Cat retina and the sampling theorem; the relation of transient
and sustained brisk-unit cut-off frequency to alpha and beta-mode cell density.
Experimental Brain Research, 42, 196–202.
Hurvich, L. M., & Jameson, D. (1955). Some quantitative aspects of an opponentcolors theory. II Brightness, saturation and hue in normal and dichromatic
vision. Journal of the Optical Society of America A, Optics, Image Science, and Vision,
45, 602–616.
Hurvich, L. M., & Jameson, D. (1957). An opponent-process theory of color vision.
Psychological Review, 64(Part 1), 384–404.
Hurvich, L. M., & Jameson, D. (1960). Perceived color, induction effects, and
opponent-response mechanisms. The Journal of General Physiology, 43(Suppl. 6),
63–80.
Ingling, C. R., Jr., & Martinez-Uriegas, E. (1983). The relationship between spectral
sensitivity and spatial sensitivity for the primate r–g X-channel. Vision Research,
23, 1495–1500.
Ishida, A. T., Stell, W. K., & Lightfoot, D. O. (1980). Rod and cone inputs to bipolar
cells in goldfish retina. The Journal of Comparative Neurology, 191, 315–335.
Jacobs, G. H., & Rowe, M. P. (2004). Evolution of vertebrate colour vision. Clinical &
Experimental Optometry, 87, 206–216.
Jacobs, G. H., Williams, G. A., Cahill, H., & Nathans, J. (2007). Emergence of novel
color vision in mice engineered to express a human cone photopigment. Science,
315, 1723–1725.
Jacobson, S. G., Franklin, K. B., & McDonald, W. I. (1976). Visual acuity of the cat.
Vision Research, 16, 1141–1143.
Johns, P. A., & Easter, S. S. (1977). Growth of the adult goldfish eye. II Increase in
retinal cell number. The Journal of Comparative Neurology, 176, 331–342.
Jordan, G., & Mollon, J. D. (1993). A study of women heterozygous for colour
deficiencies. Vision Research, 33, 1495–1508.
Joselevitch, C. (1999). Visão no ultravioleta em Carassius auratus (Ostariophysi,
Cypriniformes, Cyprinidae): estudo eletrofisiológico do sistema cones-células
horizontais (pp. 1–94). M.Sc. Dissertation, Universidade de São Paulo.
Joselevitch, C., & Kamermans, M. (2005). Dendritic potassium channels in mixed–
input ON bipolar cells. ARVO Abstracts, 46, E-Abstract 1123.
Joselevitch, C., & Kamermans, M. (2007). Interaction between rod and cone inputs in
mixed-input bipolar cells in goldfish retina. Journal of Neuroscience Research, 85,
1579–1591.
Kamermans, M., Joselevitch, C., & Klooster, J. (2004). Rod-driven light-responses in
mixed-input bipolar cells. ARVO Abstracts, 45, E-Abstract 2198.
Kamermans, M., Kraaij, D. A., & Spekreijse, H. (1998). The cone/horizontal cell
network: A possible site for color constancy. Visual Neuroscience, 15, 787–797.
Kamermans, M., & Spekreijse, H. (1995). Spectral behavior of cone-driven horizontal
cells in teleost retina. Progress in Retinal and Eye Research, 14, 313–360.
Kaneko, A., & Tachibana, M. (1978). Convergence of rod and cone signals to single
bipolar cells in the carp retina. Sensory Processes, 2, 383–387.
Kaneko, A., & Tachibana, M. (1981). Retinal bipolar cells with double colouropponent receptive fields. Nature, 293, 220–222.
Kaneko, A., & Tachibana, M. (1983). Double color-opponent receptive fields of carp
bipolar cells. Vision Research, 23, 381–388.
Kaplan, E., Lee, B. B., & Shapley, R. M. (1990). New views of primate retinal function.
Progress in Retinal Research, 9, 273–336.
Klug, K., Herr, S., Ngo, I. T., Sterling, P., & Schein, S. (2003). Macaque retina
contains an S-cone OFF midget pathway. The Journal of Neuroscience, 23,
9881–9887.
Koch, K., McLean, J., Segev, R., Freed, M. A., Berry, M. J., II, Balasubramanian, V., et al.
(2006). How much the eye tells the brain. Current Biology, 16, 1428–1434.
Kock, J. H. (1982). Dendritic tree structure and dendritic hypertrophy during growth
of the crucian carp eye. The Journal of Comparative Neurology, 209, 275–286.
957
Kock, J. H., & Reuter, T. (1978). Retinal ganglion cells in the crucian carp: I. Size and
number of somata in eyes of different size. The Journal of Comparative Neurology,
179, 535–548.
Kock, J. H., & Stell, W. K. (1985). Formation of new rod photoreceptor synapses onto
differentiated bipolar cells in goldfish retina. The Anatomical Record, 211, 69–74.
Kolb, H. (1977). The organization of the outer plexiform layer in the retina of the
cat: Electron microscopic observations. Journal of Neurocytology, 6, 131–153.
Kolb, H., Linberg, K. A., & Fisher, S. K. (1992). Neurons of the human retina: A Golgi
study. The Journal of Comparative Neurology, 318, 147–187.
Kolb, H., Nelson, R., & Mariani, A. P. (1981). Amacrine cells, bipolar cells and
ganglion cells of the cat retina: A golgi study. Vision Research, 21, 1081–1114.
Kouyama, N., & Marshak, D. W. (1992). Bipolar cells specific for blue cones in the
macaque retina. The Journal of Neuroscience, 12, 1233–1252.
Kraaij, D. A., Kamermans, M., & Spekreijse, H. (1998). Spectral sensitivity of the
feedback signal from horizontal cells to cones in goldfish retina. Visual
Neuroscience, 15, 799–808.
Kujiraoka, T., Saito, T., & Toyoda, J. (1986). Bipolar–amacrine synaptic transmission:
Effect of polarization of bipolar cells on amacrine cells in the carp retina.
Neuroscience Research Supplement, 4, S111–S119.
Kujiraoka, T., Saito, T., & Toyoda, J. I. (1988). Analysis of synaptic inputs to on–off
amacrine cells of the carp retina. The Journal of General Physiology, 92,
475–487.
Lamb, T. D., & Simon, E. J. (1976). The relation between intercellular coupling and
electrical noise in turtle photoreceptors. The Journal of Physiology, 263, 257–286.
Laughlin, S. B. (2001). Energy as a constraint on the coding and processing of
sensory information. Current Opinion in Neurobiology, 11, 475–480.
Lee, B. B. (1996). Receptive field structure in the primate retina. Vision Research, 36,
631–644.
Lennie, P. (2003). The cost of cortical computation. Current Biology, 13, 493–497.
Lennie, P., Pokorny, J., & Smith, V. C. (1993). Luminance. Journal of the Optical Society
of America A, Optics, Image Science, and Vision, 10, 1283–1293.
Leventhal, A. G., Rodieck, R. W., & Dreher, B. (1981). Retinal ganglion cell classes in
the Old World monkey: Morphology and central projections. Science, 213,
1139–1142.
Levine, M. W., & Shefner, J. M. (1977). Variability in ganglion cell firing patterns:
Implications for separate ‘‘ON” and ‘‘OFF” processes. Vision Research, 17,
765–776.
Li, W., & DeVries, S. H. (2006). Bipolar cell pathways for color and luminance vision
in a dichromatic mammalian retina. Nature Neuroscience, 9, 669–675.
Li, W., & DeVries, S. H. (2007). Synapse between rods and off cone bipolar cells in a
mammalian retina. Investigative Ophthalmology & Visual Science, 48, E-Abstract
3229.
Li, W., Keung, J. W., & Massey, S. C. (2004). Direct synaptic connections between
rods and OFF cone bipolar cells in the rabbit retina. The Journal of Comparative
Neurology, 474, 1–12.
Liu, P. C., & Chiao, C. C. (2007). Morphologic identification of the OFF-type blue cone
bipolar cell in the rabbit retina. Investigative Ophthalmology & Visual Science, 48,
3388–3395.
Livingstone, M., & Hubel, D. (1988). Segregation of form, color, movement, and
depth: Anatomy, physiology, and perception. Science, 240, 740–749.
Lythgoe, J. N. (1979). The ecology of vision. Oxford: Clarendon Press.
Macy, A., & Easter, S. S. (1981). Growth-related changes in the size of receptive field
centers of retinal ganglion cells in goldfish. Vision Research, 21, 1497–1504.
Mangel, S. C., Baldridge, W. H., Weiler, R., & Dowling, J. E. (1994). Threshold and
chromatic sensitivity changes in fish cone horizontal cells following prolonged
darkness. Brain Research, 659, 55–61.
Mangel, S. C., & Miller, R. F. (1987). Horizontal cells contribute to the receptive field
surround of ganglion cells in the rabbit retina. Brain Research, 414, 182–186.
Mangrum, W. I., Dowling, J. E., & Cohen, E. D. (2002). A morphological classification
of ganglion cells in the zebrafish retina. Visual Neuroscience, 19, 767–779.
Marc, R. E., & Liu, W. (2000). Fundamental GABAergic amacrine cell circuitries in the
retina: Nested feedback, concatenated inhibition, and axosomatic synapses. The
Journal of Comparative Neurology, 425, 560–582.
Mariani, A. P. (1983). Giant bistratified bipolar cells in monkey retina. The
Anatomical Record, 206, 215–220.
Mariani, A. P. (1984). Bipolar cells in monkey retina selective for the cones likely to
be blue-sensitive. Nature, 308, 184–186.
Mariani, A. P. (1989). Synaptic organization of classical neurotransmitter
phenotypes in the primate retina. Neuroscience Research Supplement, 10,
S101–S116.
Marks, W. B. (1965). Visual pigments of single goldfish cones. The Journal of
Physiology, 178, 14–32.
Martinez-Conde, S., Macknik, S. L., & Hubel, D. H. (2004). The role of fixational eye
movements in visual perception. Nature Reviews in Neuroscience, 5, 229–240.
Massey, S. C., Redburn, D. A., & Crawford, M. L. (1983). The effects of 2-amino-4phosphonobutyric acid (APB) on the ERG and ganglion cell discharge of rabbit
retina. Vision Research, 23, 1607–1613.
Mastronarde, D. N. (1983a). Correlated firing of cat retinal ganglion cells. I.
Spontaneously active inputs to X- and Y-cells. Journal of Neurophysiology, 49,
303–324.
Mastronarde, D. N. (1983b). Correlated firing of cat retinal ganglion cells response of
X- and Y-cells. II. To single quantal events. Journal of Neurophysiology, 49,
325–349.
Masu, M., Iwakabe, H., Tagawa, Y., Miyoshi, T., Yamashita, M., Fukuda, Y., et al.
(1995). Specific deficit of the ON response in visual transmission by targeted
disruption of the mGluR6 gene. Cell, 80, 757–765.
958
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
Mataruga, A., Kremmer, E., & Muller, F. (2007). Type 3a and type 3b OFF cone bipolar
cells provide for the alternative rod pathway in the mouse retina. The Journal of
Comparative Neurology, 502, 1123–1137.
McGuire, B. A., Stevens, J. K., & Sterling, P. (1986). Microcircuitry of beta ganglion
cells in cat retina. The Journal of Neuroscience, 6, 907–918.
McMahon, M. J., Packer, O. S., & Dacey, D. M. (2004). The classical receptive field
surround of primate parasol ganglion cells is mediated primarily by a nonGABAergic pathway. The Journal of Neuroscience, 24, 3736–3745.
Meister, M. (1996). Multineuronal codes in retinal signaling. Proceedings of the
National Academy of Sciences of the United States of America, 93, 609–614.
Merigan, W. H., & Maunsell, J. H. (1993). How parallel are the primate visual
pathways? Annual Review of Neuroscience, 16, 369–402.
Michael, C. R. (1978a). Color vision mechanisms in monkey striate cortex: Dualopponent cells with concentric receptive fields. Journal of Neurophysiology, 41,
572–588.
Michael, C. R. (1978b). Color vision mechanisms in monkey striate cortex: Simple
cells with dual opponent-color receptive fields. Journal of Neurophysiology, 41,
1233–1249.
Mollon, J. D. (1989). ‘‘Tho’ she kneel’d in that place where they grew...” The uses and
origins of primate colour vision. The Journal of Experimental Biology, 146, 21–38.
Mooij, J. E. M., & Van den Berg, T. J. T. P. (1983). The spectral shape of A2 visual
pigments. Vision Research, 23, 701–705.
Mullen, K. T. (1985). The contrast sensitivity of human colour vision to red–green
and blue–yellow chromatic gratings. The Journal of Physiology, 359, 381–400.
Naka, K. I., & Rushton, W. A. H. (1966a). An attempt to analyse colour reception by
electrophysiology. The Journal of Physiology, 185, 556–586.
Naka, K. I., & Rushton, W. A. H. (1966b). S-potentials from colour units in the retina
of fish (cyprinidae). The Journal of Physiology, 185, 536–555.
Nakajima, Y., Iwakabe, H., Akazawa, C., Nawa, H., Shigemoto, R., Mizuno, N., et al.
(1993). Molecular characterization on of a novel retinal metabotropic glutamate
receptor mglur6 with a high agonist selectivity for l-2-amino-4phosphonobutyrate. The Journal of Biological Chemistry, 268, 11868–11873.
Nathans, J. (1999). The evolution and physiology of human color vision:
Insights from molecular genetic studies of visual pigments. Neuron, 24,
299–312.
Nelson, R. (1977). Cat cones have rod input: A comparison of the response
properties of cones and horizontal cell bodies in the retina of the cat. The Journal
of Comparative Neurology, 172, 109–135.
Neumeyer, C. (1986). Wavelength discrimination in the goldfish. Journal of
Comparative Physiology A, Sensory, Neural, and Behavioral Physiology, 158,
203–213.
Neumeyer, C. (2003). Wavelength dependence of visual acuity in goldfish. Journal of
Comparative Physiology A, Sensory, Neural, and Behavioral Physiology, 189,
811–821.
Neumeyer, C., & Arnold, K. (1987). Tetrachromatic colour vision in goldfish and
turtle. In J. J. Kulikowski, C. M. Dickinson, & I. J. Murray (Eds.), Seeing contour and
colour (pp. 617–631). Oxford: Pergamon Press.
Neumeyer, C., Dorr, S., & Fritsch, J. (1997). Color constancy in goldfish: Effects of
surround size and lightness. Investigative Ophthalmology & Visual Science
Supplementum, 38, S898.
Neumeyer, C., Dorr, S., Fritsch, J., & Kardelky, C. (2002). Colour constancy in
goldfish and man: Influence of surround size and lightness. Perception, 31,
171–187.
Northmore, D. P., & Dvorak, C. A. (1979). Contrast sensitivity and acuity of the
goldfish. Vision Research, 19, 255–261.
Norton, A. L., Spekreijse, H., Wolbarsht, M. L., & Wagner, H. G. (1968). Receptive field
organization of the S-potential. Science, 160, 1021–1022.
O’Connell, C. P. (1963). The structure of the eye of Sardinops caerulea, Engraulis
mordax, and four other pelagic marine teleosts. Journal of Morphology, 113,
287–329.
Onishi, A., Hasegawa, J., Imai, H., Chisaka, O., Ueda, Y., Honda, Y., et al. (2005).
Generation of knock-in mice carrying third cones with spectral sensitivity
different from S and L cones. Zoological Science, 22, 1145–1156.
Orlov, O. Y., & Maximova, E. M. (1965). S-potential sources as excitation pools. Vision
Research, 5, 573–582.
Palacios, A. G., Varela, F. J., Srivastava, R., & Goldsmith, T. H. (1998). Spectral
sensitivity of cones in the goldfish, Carassius auratus. Vision Research, 38,
2135–2146.
Pang, J. J., Gao, F., & Wu, S. M. (2003). Light-evoked excitatory and inhibitory
synaptic inputs to ON and OFF alpha ganglion cells in the mouse retina. The
Journal of Neuroscience, 23, 6063–6073.
Paulus, W., & Kroger-Paulus, A. (1983). A new concept of retinal colour coding.
Vision Research, 23, 529–540.
Peichl, L. (1989). Alpha and delta ganglion cells in the rat retina. The Journal of
Comparative Neurology, 286, 120–139.
Pow, D. V., & Barnett, N. L. (2000). Developmental expression of excitatory amino
acid transporter 5: A photoreceptor and bipolar cell glutamate transporter in rat
retina. Neuroscience Letters, 280, 21–24.
Powers, M. K., & Easter, S. S. (1978). Absolute visual sensitivity of the goldfish. Vision
Research, 18, 1137–1147.
Protti, D. A., Flores-Herr, N., Li, W., Massey, S. C., & Wässle, H. (2005). Light signaling
in scotopic conditions in the rabbit, mouse and rat retina: A physiological and
anatomical study. Journal of Neurophysiology, 93, 3479–3488.
Rapaport, D. H., & Stone, J. (1984). The area centralis of the retina in the cat and
other mammals: Focal point for function and development of the visual system.
Neuroscience, 11, 289–301.
Raviola, E., & Gilula, N. B. (1973). Gap junctions between photoreceptor cells in the
vertebrate retina. Proceedings of the National Academy of Sciences of the United
States of America, 70, 1677–1681.
Raviola, E., & Gilula, N. B. (1975). Intramembrane organization of specialized
contacts in the outer plexiform layer of the retina. Journal of Cell Biology, 65,
192–222.
Raynauld, J. P. (1972). Goldfish retina: Sign of the rod input in opponent color
ganglion cells. Science, 177, 84–85.
Renteria, R. C., Tian, N., Cang, J., Nakanishi, S., Stryker, M. P., & Copenhagen, D. R.
(2006). Intrinsic ON responses of the retinal OFF pathway are suppressed by the
ON pathway. The Journal of Neuroscience, 26, 11857–11869.
Ringo, J. L., & Wolbarsht, M. L. (1986). Spectral coding in cat retinal ganglion cell
receptive fields. Journal of Neurophysiology, 55, 320–330.
Roska, B., Molnar, A., & Werblin, F. S. (2006). Parallel processing in retinal ganglion
cells: How integration of space–time patterns of excitation and inhibition form
the spiking output. Journal of Neurophysiology, 95, 3810–3822.
Roska, B., & Werblin, F. S. (2001). Vertical interactions across ten parallel, stacked
representations in the mammalian retina. Nature, 410, 583–587.
Ruderman, D. L., Cronin, T. W., & Chiao, C. C. (1998). Statistics of cone responses to
natural images: Implications for visual coding. Journal of the Optical Society of
America A, Optics, Image Science, and Vision, 15, 2036–2045.
Russell, P. W., & Wheeler, T. G. (1983). Scotopic sensitivity to ON and OFF stimulus
transients. Vision Research, 23, 525–528.
Saito, T., Kondo, H., & Toyoda, J. I. (1979). Ionic mechanisms of two types of Oncenter bipolar cells in the carp retina. I. The responses to central illumination.
The Journal of General Physiology, 73, 73–90.
Saito, T., Kondo, H., & Toyoda, J. I. (1981). Ionic mechanisms of two types of Oncenter bipolar cells in the carp retina. II. The responses to annular illumination.
The Journal of General Physiology, 78, 569–589.
Schellart, N. A., & Spekreijse, H. (1973). Origin of the stochastic nature of ganglion
cell activity in isolated goldfish retina. Vision Research, 13, 337–345.
Schiller, P. H. (1992). The ON and OFF channels of the visual system. Trends in
Neurosciences, 15, 86–92.
Schneeweis, D. M., & Schnapf, J. L. (1995). Photovoltage of rods and cones in the
Macaque retina. Science, 268, 1053–1056.
Schneeweis, D. M., & Schnapf, J. L. (1999). The photovoltage of macaque cone
photoreceptors: Adaptation, noise, and kinetics. The Journal of Neuroscience, 19,
1203–1216.
Scholes, J. H. (1975). Colour receptors, and their synaptic connexions, in the retina
of a cyprinid fish. Philosophical transactions of the Royal Society of London Series B,
Biological Sciences, 270, 61–118.
Shapley, R. M., & Perry, V. H. (1986). Cat and monkey retinal ganglion cells and their
visual functional roles. Trends in Neurosciences, 9, 229–235.
Shefner, J. M., & Levine, M. W. (1979). An analysis of receptor inputs to and
spatial distribution of ganglion cell on and off processes. Vision Research, 19,
647–653.
Sherry, D. M., & Yazulla, S. (1993). Goldfish bipolar cells and axon terminal patterns:
A golgi study. The Journal of Comparative Neurology, 329, 188–200.
Shiells, R. A., Falk, G., & Naghshineh, S. (1981). Action of glutamate and aspartate
analogues on rod horizontal and bipolar cells. Nature, 294, 592–594.
Shimbo, K., Toyoda, J. I., Kondo, H., & Kujiraoka, T. (2000). Color-opponent responses
of small and giant bipolar cells in the carp retina. Visual Neuroscience, 17,
609–621.
Sjostrand, F. S. (1958). Ultrastructure of retinal rod synapses of the guinea pig eye as
revealed by three-dimensional reconstructions from serial synapses. Journal of
Ultrastructure Research, 2, 122–170.
Slaughter, M. M., & Miller, R. F. (1981). 2-Amino-4-phosphonobutyric acid: A new
pharmacological tool for retinal research. Science, 211, 182–185.
Smallwood, P. M., Olveczky, B. P., Williams, G. L., Jacobs, G. H., Reese, B. E., Meister,
M., et al. (2003). Genetically engineered mice with an additional class of cone
photoreceptors: Implications for the evolution of color vision. Proceedings of the
National Academy of Sciences of the United States of America, 100, 11706–11711.
Smith, R. G., Freed, M. A., & Sterling, P. (1986). Microcircuitry of the dark-adapted
cat retina: Functional architecture of the rod–cone network. The Journal of
Neuroscience, 6, 3505–3517.
Soucy, E., Wang, Y., Nirenberg, S., Nathans, J., & Meister, M. (1998). A novel signaling
pathway from rod photoreceptors to ganglion cells in mammalian retina.
Neuron, 21, 481–493.
Spekreijse, H. (1969). Rectification in the goldfish retina: Analysis by sinusoidal and
auxiliary stimulation. Vision Research, 9, 1461–1472.
Spekreijse, H., & Van den Berg, T. J. T. P. (1971). Interaction between colour and
spatial coded processes converging to retinal ganglion cells in goldfish. The
Journal of Physiology, 215, 679–692.
Spekreijse, H., Wagner, H. G., & Wolbarsht, M. L. (1972). Spectral and spatial coding
of ganglion cell responses in goldfish retina. Journal of Neurophysiology, 35,
73–86.
Steinberg, R. H., Reid, M., & Lacy, P. L. (1973). The distribution of rods and cones in
the retina of the cat (Felis domesticus). The Journal of Comparative Neurology,
148, 229–248.
Stell, W. K. (1967). The structure and relationships of horizontal cells and
photoreceptor–bipolar synaptic complexes in goldfish retina. American Journal
of Anatomy, 121, 401–424.
Stell, W. K. (1976). Functional polarization of horizontal cell dendrites in goldfish
retina. Investigative Ophthalmology, 15, 895–908.
Stell, W. K. (1978). Inputs to bipolar cell dendrites in goldfish retina. Sensory
Processes, 2, 339–349.
C. Joselevitch, M. Kamermans / Vision Research 49 (2009) 943–959
Stell, W. K. (1980). Photoreceptor-specific synaptic pathways in goldfish retina: A
world of colour, a wealth of connections. In G. Verriest (Ed.), Colour vision
deficiencies V (pp. 1–14). Bristol: Adam Hilger.
Stell, W. K., & Harosi, F. I. (1976). Cone structure and visual pigment content in the
retina of the goldfish. Vision Research, 16, 647–657.
Stell, W. K., & Kock, J. H. (1984). Structure, development and visual acuity in the
goldfish retina. In S. Robert Hilfer & J. B. Sheffield (Eds.), Molecular and cellular
basis of visual acuity (pp. 79–105). New York: Springer-Verlag New York Inc..
Sterling, P. (2004). How retinal circuits optimize the transfer of visual information.
In L. M. Chalupa & J. S. Werner (Eds.), The visual neurosciences (pp. 234–259).
Cambridge: MIT Press.
Stockman, A., & Plummer, D. J. (2005). Spectrally opponent inputs to the human
luminance pathway: Slow +L and M cone inputs revealed by low to moderate
long-wavelength adaptation. The Journal of Physiology, 566, 77–91.
Stockman, A., Plummer, D. J., & Montag, E. D. (2005). Spectrally opponent inputs to
the human luminance pathway: Slow +M and L cone inputs revealed by
intense long-wavelength adaptation. The Journal of Physiology, 566, 61–76.
Svaetichin, G. (1953). The cone action potential. Acta Physiologica Scandinavica, 29,
565–600.
Svaetichin, G. (1956). Spectral response curves from single cones. Acta Physiologica
Scandinavica Supplementum, 39, 17–46.
Svaetichin, G., & MacNichol, E. F. (1958). Retinal mechanisms for chromatic and
achromatic vision. Annals of the New York Academy of Sciences, 74, 385–404.
Tauchi, M., Morigiwa, K., & Fukuda, Y. (1992). Morphological comparisons between
outer and inner ramifying alpha cells of the albino rat retina. Experimental Brain
Research, 88, 67–77.
Tessier-Lavigne, M., & Attwell, D. (1988). The effect of photoreceptor coupling and
synapse nonlinearity on signal:noise ratio in early visual processing. Proceedings
of the Royal Society of London Series B, Containing Papers of Biological Character,
234, 171–197.
Toyoda, J., Hashimoto, H., & Otsu, K. (1973). Bipolar-amacrine transmission in the
carp retina. Vision Research, 13, 295–307.
Toyoda, J., Shimbo, K., Kondo, H., & Kujiraoka, T. (1992). Push–pull modulation of
ganglion cell responses of carp retina by amacrine cells. Neuroscience Letters,
142, 41–44.
Toyoda, J. I. (1974). Frequency characteristics of retinal neurons in the carp. The
Journal of General Physiology, 63, 214–234.
Toyoda, J. I., & Tonosaki, K. (1978a). Effect of polarisation of the horizontal cells on
the on-center bipolar cell of carp retina. Nature, 276, 399–400.
Toyoda, J. I., & Tonosaki, K. (1978b). Studies on the mechanisms underlying
horizontal-bipolar interaction in the carp retina. Sensory Processes, 2, 359–365.
Tsin, A. T. C., Liebman, P. A., Beatty, D. D., & Drzymala, R. (1981). Rod and cone visual
pigments in the goldfish. Vision Research, 21, 943–946.
Tsukamoto, Y., Morigiwa, K., Ishii, M., Takao, M., Iwatsuki, K., Nakanishi, S., et al.
(2007). A novel connection between rods and ON cone bipolar cells revealed by
ectopic metabotropic glutamate receptor 7 (mGluR7) in mGluR6-deficient
mouse retinas. The Journal of Neuroscience, 27, 6261–6267.
Tsukamoto, Y., Morigiwa, K., Ueda, M., & Sterling, P. (2001). Microcircuits for night
vision in mouse retina. The Journal of Neuroscience, 21, 8616–8623.
Vallerga, S., & Djamgoz, M. B. A. (1991). Ganglion cells in the goldfish retina:
Correlation of light-evoked response and morphology. Vision Research, 31,
487–497.
Van Dijk, B. W., & Spekreijse, H. (1983). Ethambutol changes the color coding of carp
retinal ganglion cells reversibly. Investigative Ophthalmology & Visual Science, 24,
128–133.
van Wyk, M., Taylor, W. R., & Vaney, D. I. (2006). Local edge detectors: A substrate
for fine spatial vision at low temporal frequencies in rabbit retina. The Journal of
Neuroscience, 26, 13250–13263.
Vaney, D. I., & Hughes, A. A. (1990). Is there more than meets the eye? In C.
Blakemore (Ed.), Vision: Coding and efficiency (pp. 74–83). Cambridge:
Cambridge University Press.
VanLeeuwen, M. T., Joselevitch, C., Fahrenfort, I., & Kamermans, M. (2007). The
contribution of the outer retina to color constancy: A general model for color
constancy synthesized from primate and fish data. Visual Neuroscience, 24,
277–290.
959
Vardi, N., Dhingra, A., Zhang, L., Lyubarsky, A., Wang, T. L., & Morigiwa, K. (2002).
Neurochemical organization of the first visual synapse. The Keio Journal of
Medicine, 51, 154–164.
Volgyi, B., Deans, M. R., Paul, D. L., & Bloomfield, S. A. (2004). Convergence and
segregation of the multiple rod pathways in mammalian retina. The Journal of
Neuroscience, 24, 11182–11192.
Wachtler, T., Doi, E., Lee, T., & Sejnowski, T. J. (2007). Cone selectivity derived from
the responses of the retinal cone mosaic to natural scenes. Journal of Vision, 7, 6.
Wagner, H. G., MacNichol, E. F., & Wolbarsht, M. L. (1960). The response properties
of single ganglion cells in the goldfish retina. The Journal of General Physiology,
43, 43–62.
Walls, G. L. (1942). The vertebrate eye and its adaptive radiation. Bloomfield Hills:
Cranbrook Institute of Science.
Wang, Y., & Mangel, S. C. (1996). A circadian clock regulates rod and cone input to
fish retinal cone horizontal cells. Proceedings of the National Academy of Sciences
of the United States of America, 93, 4655–4660.
Warrant, E. J. (2000). The eyes of deep-sea fishes and the changing nature of visual
scenes with depth. Philosophical transactions of the Royal Society of London Series
B, Biological Sciences, 355, 1155–1159.
Warrant, E. J. (2004). Vision in the dimmest habitats on earth. Journal of Comparative
Physiology A, Sensory, Neural, and Behavioral Physiology, 190, 765–789.
Wässle, H. (2004). Parallel processing in the mammalian retina. Nature Reviews in
Neuroscience, 5, 747–757.
Wässle, H., & Boycott, B. B. (1991). Functional architecture of the mammalian retina.
Physiological Reviews, 71, 447–480.
Wässle, H., Peichl, L., & Boycott, B. B. (1981). Morphology and topography of on- and
off-alpha cells in the cat retina. Proceedings of the Royal Society of London Series
B, Containing Papers of Biological Character, 212, 157–175.
Wässle, H., Schafer-Trenkler, I., & Voigt, T. (1986). Analysis of a glycinergic
inhibitory pathway in the cat retina. The Journal of Neuroscience, 6,
594–604.
Wersinger, E., Schwab, Y., Sahel, J. A., Rendon, A., Pow, D. V., Picaud, S., et al. (2006).
The glutamate transporter EAAT5 works as a presynaptic receptor in mouse rod
bipolar cells. The Journal of Physiology, 577, 221–234.
West, R. W. (1978). Bipolar and horizontal cells of the gray squirrel retina: Golgi
morphology and receptor connections. Vision Research, 18, 129–136.
West, R. W., & Dowling, J. E. (1975). Anatomical evidence for cone and rod-like
receptors in the gray squirrel, ground squirrel, and prairie dog retinas. The
Journal of Comparative Neurology, 159, 439–460.
Wheeler, T. G. (1979). Retinal ON and OFF responses convey different chromatic
information to the CNS. Brain Research, 160, 145–149.
Wiesel, T. N., & Hubel, D. H. (1966). Spatial and chromatic interactions in the lateral
geniculate body of the rhesus monkey. Journal of Neurophysiology, 29,
1115–1156.
Witkovsky, P., Shakib, M., & Ripps, H. (1974). Interreceptoral junctions in the teleost
retina. Investigative Ophthalmology, 13, 996–1009.
Wong, K. Y., Cohen, E. D., & Dowling, J. E. (2005). Retinal bipolar cell input
mechanisms in giant danio II. Patch–clamp analysis of on bipolar cells. Journal of
Neurophysiology, 93, 94–107.
Wong, K. Y., & Dowling, J. E. (2005). Retinal bipolar cell input mechanisms in giant
danio. III. ON–OFF bipolar cells and their color-opponent mechanisms. Journal of
Neurophysiology, 94, 265–272.
Wu, S. M., Gao, F., & Pang, J. J. (2004). Synaptic circuitry mediating light-evoked
signals in dark-adapted mouse retina. Vision Research, 44, 3277–3288.
Zaghloul, K. A., Boahen, K., & Demb, J. B. (2003). Different circuits for ON and OFF
retinal ganglion cells cause different contrast sensitivities. The Journal of
Neuroscience, 23, 2645–2654.
Zeitz, C., van Genderen, M., Neidhardt, J., Luhmann, U. F., Hoeben, F., Forster, U., et al.
(2005). Mutations in GRM6 cause autosomal recessive congenital stationary
night blindness with a distinctive scotopic 15-Hz flicker electroretinogram.
Investigative Ophthalmology & Visual Science, 46, 4328–4335.
Zhang, K., & Sejnowski, T. J. (1999). Neuronal tuning: To sharpen or broaden? Neural
Computation, 11, 75–84.
Zimov, S., & Yazulla, S. (2008). Novel processes invaginate the pre-synaptic terminal
of retinal bipolar cells. Cell and Tissue Research, 333, 1–16.