Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 105
The Distribution of Cones
in the Primate Retina
V. HUGH PERRY
Every student of vision is aware that the sampling of an image by the
photoreceptors imposes important limitations on the information that is
available at every subsequent stage of processing along the visual pathways.
Despite the importance of knowing the receptor distribution, the actual
data base for the primate retina has been remarkably small until the recent
publication of a number of papers that have sought to rectify this and that
have brought to light new and interesting points. This paper will bring
together some of the main points, but no attempt is made to give all the
quantitative values since these are readily available in the primary sources.
TISSUE PREPARATION
Assessing the distribution of photoreceptors across the retina would
appear to be a relatively trivial problem. However, it must be recognized
that the preparation of any tissue for microscopic examination usually
involves some sort of processing, and this processing commonly results in
shrinkage of the tissue. Unless the shrinkage is taken into account by
making appropriate measurements at the various stages of preparation, the
value of receptor counts from such material is limited. The fact that the
receptor density changes dramatically with eccentricity also requires that the
sampling eccentricity be exactly known. The use of whole-mounted retinas
greatly reduces many of the problems associated with using sectioned retinas
and allows accurate estimates of retinal shrinkage. The identification of
rods and cones in the primate retina poses few problems (Figure 1~.
The classic study of Osterberg (1935) described the distribution of
rods and cones over a large part of a single human retina. Polyak (1957)
described the density of cones in monkey (M. mulatta) and human retinas
105
OCR for page 106
106
A
Kit HUGH PERRY
FIGURE 1A Photomicrograph of foveal cones in a horizontal section through the inner
segments. The retina is from an M. fasciculans monkey. The field is a little more than 1
deg across (201 M/degree from Peny and Cowey, 1985~.
but largely restricted his observations to the fovea and perifovea. More
recent studies on a small number of retinas have also been restricted to the
fovea or central few degrees (Miller, 1979; de Monasterio et al., 1985~. The
foveal cone density has attracted much attention, in no small part because
of the interest of many visual scientists in measurements of threshold vision.
What was required was a study of cone density distribution in a sufficient
number of retinas to give an idea of the variability between individuals.
Rolls and Cowey (1970) examined a number of retinas from M. mulatto and
S. sciureus along the horizontal meridian, but they used vertical sections of
retina and were well aware of the problems in extracting accurate estimates
from such material.
My co-workers and I examined, in whole-mounted retinas, the distri-
bution of cones along the horizontal and vertical meridians in four retinas
from M. mulatta (Perry and Cowey, 1985~. In these preparations the shrink-
age was minimal. Estimates of the foveal cone density were made from
horizontal sections that passed through the inner segments. The shrinkage
of the sections was known. In one case the entire retina was sampled
except for the fovea itself. A number of important points came out of this
study. The cone density changes very rapidly over the central 1 deg of the
visual field (Figure 1A) and continues to decline rapidly over the central
OCR for page 107
THE DIST~B=ION OF CONES IN THE PLATE ~TINA 107
Bma ~ _
~~ i.
FIGURE 1B A field centered on 1 deg from the center of the fovea (M. fasciculans).
5 degrees (Figures 1B and 1C). The change in the size of the cones and
the increasing number of rods are easily seen in these photomicrographs.
There was a clear nasotemporal asymmetry in the cone distribution outside
the central 10 deg or so of the retina; the cone density declined more
slowly along the nasal horizontal meridian of the retina than along the
other axes. The asymmetry in cone density was somewhat smaller than
that seen for the ganglion cells (Perry et al., 1984~. This asymmetry is
consistent with psychophysical observations of acuity and sensitivity being
superior in the temporal field. In addition, we found differences in foveal
cone density between species (see also Rolls and Cowey, 1970; Borwein
et al., 1980~. While these data provide a useful basis for comparison with
physiological and psychophysical studies on monkeys, it remained to be
shown how the cone density varied in the primate for which we have so
much psychophysical data the human.
CONE DENSITY
Curcio et al. (1987b) have provided the first complete maps of the
retinal cone distribution in humans. The same laboratory has also shown
that the density of cones varies with age (Yuodelis and Hendrickson, 1986~.
Bearing in mind that the number of retinas examined is small, Yuodelis
OCR for page 108
~-
a: ~
~ HUGH PERRY
FIGURE 1C A field centered on a point 5 deg Tom the center of the fovea (M. mulatta).
The estimated linear shrinkage of A and B during processing was about 15 percent; no
correction has been applied to the scaling. Scale bar = 25 ,um.
and Hendrickson have shown that the foveal cone density in the newborn is
relatively low; density then increases for at least the next 4 or more years.
The density may fall again in aged subjects. It is believed that the increase
in cone density in the postnatal period is accomplished by the migration
of cones toward the center of the fovea (Hendrickson and Kupfer, 1976;
Yuodelis and Hendrickson, 1986~. The loss of cones in the aging eye may
be a result of cell death.
Not only is there a change in cone density with age, Curcio et al.
(1987b) have shown that the foveal density may vale by as much as a factor
of 3 in normal adult human eyes. A sample of four retinas from subjects
aged 27 to 44 years was studied. There was considerably less variation in
density outside the central 1 deg. The cone distribution in the human eye
has a similar but not as pronounced nasotemporal asymmetry as described
for the monkey with the higher densities in the nasal retina.
The large variation in foveal cone density cannot be readily attributed
to differential tissue shrinkage. The method of clearing the tissue and
viewing it under Normaski optics greatly reduced processing distortion
(Curcio et al., 1987a) and the authors estimate the shrinkage as varying
from 2 to 12 percent. It could be argued, however, that the large variation
in cone density was due to the fact that human eyes are not always obtained
OCR for page 109
THE DISTRIBUTION OF CONES IN THE PRIMATE RETINA
D
I.,
109
---~d~D ~
~~.~.
.~
~ Act_ - ~ ~ Y f A \~ ' ~ V - hi. ' · ~1~ ~
~ ~ t~ Ad' ~~-._~N
A. ~~
/~ . · ~ t! /~ - ~ . - ~ ~ Y Y ~ ~ ~ ~ ~ ~ v - v
He's'' ~ ~ .~ ~ ~? ~~
~ _~J ~ f · ~ MA ~ ~ ~ ~ I ~ ~ - (<
_ V1~ -7 /~ ~ ~ ~ ~ \. ~ _ ~) - #_ ~ ~ ~
4~-~
Y<\
2 it,/ ff i' · ~
FIGURE ID, E Analysis of cone mosaics to demonstrate the local triangular lattice at
the fovea and how it degenerates with eccentricity. See text for details.
OCR for page 110
110
~ mu_ _
~ HUGH PERRY
~~ r
Afar
FIGURE IF Analysis of cone mosaics to demonstrate the local triangular lattice at the
fovea and how it degenerates with eccentncity. See text for details.
under conditions for tissue histology, and it is generally agreed that the
fovea is particularly vulnerable to postmortem artefacts. We therefore
undertook a study of eyes from M. fascicular~s, where the eyes could all
be obtained immediately postmortem following fixation in situ by perfusion
(Hawker et al., 1988~. Each eye was subsequently processed in as near
identical fashion as possible following the method of Curcio et al. (1987a).
Counts of the cone density were made throughout the central degree and
at selected intervals out to the periphery. There would appear to be no a
priori reason why the size of the sampling area chosen for counting should
be of any particular dimensions. In Figure 1A it is clear that the density
is changing very rapidly across the central 1 deg. The sample area used
by Perry and Cowey (1985) of About 0.45 x 0.45 deg was almost certainly
slightly too large. Some of the variability among different studies can, in
part, be accounted for by differences in the sampling area. Pokorny (1968)
has examined how grating acuity changes with the size of the target viewed,
and his results suggest that an area about 0.125 x 0.125 deg would be
appropriate. This was the area we used.
A sample of our data is shown in Figure 2, where the cone density
along the temporal horizontal meridian in eight retinas is plotted. In six of
the retinas there is a rapid decline in cone density with eccentricity, as we
OCR for page 111
THE DISTRIBUTfON OF CONES IN THE PRIMATE RETINA
~ o 5 -
-
sin
_'
i 104
o
10
111
, , , , I, . . .
0 1 10 100
ECCENTRI C I TY ~ deg. arc ~
FIGURE 2 Density of cones at the fovea and along the temporal horizontal meridian in
retinas from M. fasciculans monkeys.
expected from previous studies. The variation is similar in these six retinas
across the eccentricities we have examined. In contrast, two of the retinas
(- and o) showed quite clear reductions in density within the central 0.5 deg.
In these animals we could observe no pathology or distortion of the tissue,
and indeed it was clear that the inner segments of the cones were themselves
larger than in the retinas with higher densities. If these low counts resulted
from distortion of the tissue, we might expect to see anomalies at further
eccentricities; this was not the case. Despite having low densities in the
foveal center, these two animals would still theoretically have a cone density
at O.S deg. sufficient to resolve about 30 cycles/degree. The highest and
lowest foveal counts give about a sixfold range in peak density. It would be
interesting to know how these anatomical results compared with behavioral
measures of acuity in M. fascicularis, but unfortunately no large sample is
available. However, behavioral measures of acuity in M. mulatta, which
OCR for page 112
112
~ HUGH PERRY
were refracted and optically corrected, showed considerable variation (N
Cowey, personal communication).
Thus, we are able to confirm the results of Curcio et al. namely, that
cone density varies substantially at the fovea between individual eyes. The
reason that some animals have a low density is unclear, but it may well be
that the late migration of cones into the foveal region (Hendrickson and
Kupfer, 1976; Yuodelis and Hendrickson, 1986) is particularly susceptible
to environmental, nutritional, and mechanical factors. How the variation
in the foveal cone density is related to the refractive state of the eye was,
unfortunately, not examined in this study but certainly merits further study.
The anatomical data suggest that we can give no single value as the
cone spacing at the human fovea. Ideally, we would like to look at the cone
separation in the living eye and compare the results obtained with those
found in whole-mount preparations. Using laser interferometric methods,
it is possible to do just this (Williams, 1985, 1988~. Using this technique
the data suggest that there is rather little variability in cone spacing. The
discrepancy between the anatomical and psychophysical results may be
more apparent than real at this time, since only small numbers of subjects
have been examined in both instances. Williams (1988) does report that
at least one subject found the moire patterns produced by the interference
fringes difficult to see, which may reflect some anomaly in this subject's
cone packing.
IMAGE SAMPLING
The way in which cones are packed at the fovea is also a matter of
some interest since positional disorder in the elements sampling an image
will degrade the high frequencies in an image (French et al., 1977; Hirsch
and Hylton, 1984~. Once again, for this analysis we chose only those
preparations where the distortion was minimized. Figure 1A shows that
the cones are regularly packed but gives the impression of short-range
order and little long-range order. This is simply demonstrated in Figure
ID. A spot was placed in the center of each cone shown in Figure 1, and
a straight line was drawn through as many spots as could be joined. If the
cone mosaic was perfectly regular, the lines would cross the entire field.
This is clearly not the case; the cone mosaic divides into "islands" in which
the cones are arranged in a triangular lattice with fractures between the
islands. The largest islands are about 10 x 20 cones. We adopted the same
procedure for cones centered on a region 1 and 5 deg from the center of
the fovea. Larger spots were used at 5 deg to allow for the fact that the
cones themselves were substantially larger that at the fovea. The regularity
of the mosaic clearly degenerates with eccentricity. The islands are less
obvious, and the number of cones contributing to each of them is greatly
OCR for page 113
THE DISTRIBUTION OF CONES IN THE PRIMATE RETINA
113
reduced. Williams (1988) has shown that in the human eye, again using
interference fringes, the orientations of the moire patterns observed are
of a form consistent with triangular packing of the cones with local but
not long-range order. The extent of the regular packing predicted by his
studies is in good agreement with the size of the largest islands and the
rapid degeneration of the regularity with distance from the foveal center.
If the cones are arranged in a crystalline lattice in the fovea, it would
be of interest to know whether the different types of cones long-, medium-,
and short-wavelength sensitive (here referred to as red, green, and blue
cones) are also distributed in regular patterns. The blue cones, which can
be identified by staining with Procion dyes, are organized In a regular lattice
(de Monasterio et al., 1985~. Unfortunately, there is no simple anatomical
method for discriminating between red and green cones in the primate.
The only study that successfully differentially labeled the red and green
cones in a primate was in the baboon retina (Marc and Sperling, 1977~.
Their results showed that the two types of cones were randomly arrayed.
It has been suggested that not only are the red and green cones
randomly arrayed but also that the cells of the inner retinal layers are
not able to distinguish between the two cone types during development
(Shapley and Perry, 1986~. How then are color-opponent ganglion cells
generated? We hypothesize that a midget bipolar cell contacting a single
cone and connected to a single one of the smallest P ganglion cells (so-
called midget ganglion cells; Polyak, 1941) would provide the central input
to a cell with color-opponent receptive field. The surround of this cell
might receive inputs from both red and green cones and, depending on
the relative proportions of them, would appear as a more or less color-
opponent cell. Across the whole population of color-opponent P ganglion
cells, the neutral point to these cells would vary, as is found to be the case
(Zrenner, 1983~. It will be of considerable interest to know just how and
in what proportions these two cone types are distributed in the retina and
whether this is consistent with this the "hit-or-miss" hypothesis.
At the fovea the cells of the inner nuclear and ganglion cell layers are
laterally displaced; in order that the cones might contact the cells of the
inner nuclear layer, there is a long process, known as a fiber of Henle,
that courses in the outer fiber layer and terminates in a pedicle to make
synaptic contact in the outer plexiform layer (Polyak, 1941~. The length
of these fibers and their arrangement around the fovea will not only be
important for the preservation of the sampled image but will also determine
the numerical relationship between the number of cones and ganglion cells
sampling a given point of visual space.
The fibers of Henle of foveal cones in the monkey retina are at least
400 Em in length (Boycott et al., 1987; Perry and Cowey, 1988~. Cones
immediately adjacent to the central foveal cones have slightly shorter fibers
OCR for page 114
114
~ HUGH PERRY
and so on, such that the length of the fibers decreases in a systematic fashion
from the fovea out to about 15 deg eccentricity. The fibers radiate out
from the center like the spokes from a wheel, and it is clear that adjacent
cones have very similar lengths (PerIy and Cowey, 1988), so the topography
of the cone mosaic at the level of the inner segments is preserved at the
pedicles. There is, however, an interesting problem for those cones in the
most central part of the fovea since the fibers of these cones project to all
four quadrants around the fovea. Initially the cells to which these fibers
are connected lay over the foveal center and migrated laterally during
development; thus, it would appear that the topography was preserved.
However, while this is in large part correct, there remains the problem that
some fibers pass to either side of the vertical meridian. Ganglion cells on
either side of the meridian send their axons to opposite sides of the brain.
This poses a problem as to how the image is then retrieved, particularly
with high spatial frequencies since these are poorly transmitted across the
corpus callosum (Bernard) et al., 1987~. Williams (1988) has found that in
some observers the cone spacing is larger in the horizontal direction than
in the vertical direction. It seems possible that this may be related to the
behavior of the fibers of Henle.
GANGLION/FOVEAL CELL RATIO
From our knowledge of the length of the fibers of Henle or by taking
into account the fact that cone pedicles are at least twice the diameter of
the cone inner segment (Boycott et al., 1987), it is possible to estimate the
numerical relationship of ganglion cells to cones for the foveal retina. The
present estimates show that the ratio of ganglion cells to cones is slightly
greater than 2:1 for the central 1 dog (Perry and Cowey, 1988; Schein,
1988~. It has been shown that each cone pedicle is contacted by at least
two bipolar cells, which themselves terminate at different levels in the inner
plexiform layer (Kolb et al., 1969~. From what is known of the organization
of other retinas, it seems likely that one of these bipolar cells contacts an
ON-center ganglion cell, the other an OFF-center cell. A 2:1 ratio may
be of little advantage with respect to the spatial sampling but may be an
advantage for producing a system with a large dynamic range.
The foveal cones have a fiber of Henle of the order of 400 Am long
and about 2 to 3 sum in diameter. The conduction of a potential along
such a process will result in both a decrement and temporal smearing of
the signal. This is not likely to be an advantage. Polyak (1941) pointed
out that Muller's fibers, the radial glia of the retina, appear to partially
wrap themselves around the fibers of Henle. We have recently shown that
adjacent Muller's fibers and fibers of Henle have very similar lengths at the
same retinal eccentricities. It would appear that the Muller's fibers may act
OCR for page 115
THE DISTRIBUTION OF CONES IN TlIE PRIMATE RETINA
I15
as ensheathing glia and thus help to prevent the decrement of the signal as
it passes from the cell body to the pedicle.
A description of the distribution of the cones in the monkey and human
eye is now available. This information should prove useful in showing the
extent to which cone density limits visual performance across the visual field.
The next important task is to discover from anatomical and physiological
experiments how the information from the cones is used to construct the
receptive field characteristics of the different types of ganglion cells.
REFERENCES
Bernadi, N., S. Bisti, and L. Maffei
1987 The transfer of visual information across the corpus callosum: spatial
characteristics. Joumal of Physiology (London) 384:619-632.
Borwein, B., D. Borwein, J. Medeiros, and J.W. McGowan
1980 The ultrastructure of monkey fovea photoreceptors, with special reference
to the structure, shape, size and spacing of foveal cones. American Journal
of Anatomy 159:125-146.
Boycott, B.B., J.M. Hopkins, and H.G. Sperling
1987 Cone connections of the horizontal cells of the rhesus monkey's retina.
Proceedings of the Royal Society B 229:345-379.
Curcio, C.A., O. Packer, and R.E. Kalina
1987a A whole mount method for sequential analysis of photoreceptor and ganglion
cell topography in a single retina. Vision Research 27:9-15.
Curcio, C.A., K.R. Sloan, Jr., O. Packer, A.E. Hendrickson, and R.E. Kalina
1987b Distribution of cones in human and monkey retina: individual variability
and radial asymmetry. Science (New York) 236:597-582.
de Monasterio, F.M., E.P. McCrane, J.K, Newlander, and J.S. Schein
1985 Density profile of blue sensitive cones along the horizontal meridian of
macaque retina. Investigative Ophthalmology and Visual Science 26:289-302.
French, AS., A.W. Snyder, and D.G. Stavenga
1977 Image degradation lay an irregular retinal mosaic. Biological Cybemetics
27:22~233.
Hawken, M.J., V.H. Perry, and A.J. parker
1988 Structural relationships of photoreceptors to VI receptive fields in primate.
Invest. Ophthalm. V's. Sci 29(Supp.~:297.
Hendrickson, A., and C. Kupfer
1976 The histogenesis of the fovea in the macaque monkey. Investigative Oph-
thalmolo~y 15:746-756.
Hirsch, J., and R. Hylton
1984 Quality of the primate receptor lattice and limits of spatial vision. Illusion
Research 24:347-356.
Kolb, H., B.B. Boycott, and J.E. Dowling
1%9 A second type of midget bipolar cell in the primate retina. Philosophical
Transactions of the Royal Society B 225:177-184.
Marc, R.E., and H.G. Sperling
1977 Chromatic organization of primate cones. Science 196:454-456.
OCR for page 116
116 I! HUGH PERRY
Miller, W.H.
1979 Ocular optica filtering. In Handbook of Sensory Physiology, vol. V11/6A.
Berlin: Springer.
Osterberg, G.
1935 Topography of the layer of rods and cones in the human retina. Acta
OpAthalmologica Supplement 6:1-103.
Perry, V.H., and A. Cowey
1985 The ganglion cell and cone distributions in the monkey's retina; implications
for central magnification factors. I^sion Research 25:1795-1810.
1988 The lengths of the fibres of Henle in the retina of macaque monkeys:
implications for vision. Neuroscience 25:225-236.
Perry, V.H., R. Oehler, and ~ Cowey
1984 Retinal ganglion cells that project to the dorsal lateral geniculate nucleus in
the macaque monkey. Neuroscience 12:1101-1023.
Pokorny, J.
1968 The effect of target area on grating acuity. Vision Research 8:54~554.
Polyak, S.
1941 The Retina. Chicago, Ill.: University of Chicago Press.
1957 The Herterbrate Usual System. Chicago, Ill.: University of Chicago Press.
Rolls, E.T., and A Cowey
1970 Topography of the retina and striate cortex and its relationship to visual
acuity in rhesus monkeys and squirrel monkeys. Experimental Brain Research
10 19~311.
Schein, S.J.
1988 Anatomy of the macaque fovea and spatial densities of neurons in foveal
representation. Journal of Comparative Neurology 269:479-505.
Shapley, R.M., and V.H. Perry
1986 Cat and monkey retinal ganglion cells and their visual functional roles.
Trends in Neuroscience 9:229-235.
Williams, D.R.
1985 Aliasing in human foveal vision. Vision Research 25:195-205
1988 Topography of the foveal cone mosaic in the living human eye. Vision
Research 28:433-454.
Yuodelis, C., and ~E. Hendrickson
1986 A qualatative and quantitative analysis of the human fovea during develop-
ment. Vision Research 26:847~55.
Zrenner, E.
1983 Neurophysiological aspects of colour vision in primates. In Studies of Brain
Function, vol. 9. Berlin: Springer-Verlag.
Representative terms from entire chapter:
ganglion cells
