Brain Mechanisms of Vision

Author(s): David H. Hubel and Torsten N. Wiesel

Source: Scientific American , Vol. 241, No. 3 (September 1979), pp. 150-163
Published by: Scientific American, a division of Nature America, Inc.
Stable URL: https://www.jstor.org/stable/10.2307/24965293

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Scientific American, a division of Nature America, Inc. is collaborating with JSTOR to digitize,
preserve and extend access to Scientific American

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
 Brain Mechanisms of Vision
A functional architecture that may underlie processing of sensory
information in the cortex is revealed by studies of the activity and
the organization in space of neurons in the primary visual cortex

by David H. Hubel and Torsten N. Wiesel

feet. (In a 1963 article in Scientific Amer­

V
iewed as a kind of invention by of the location (at the surface of the
evolution, the cerebral cortex ican one of us gave the area as 20 square brain) of the electrical brain waves pro­
must be one of the great success feet and was quickly corrected by a neu­ duced when an animal was stimulated
stories in the history of living things. In roanatomist friend in Toronto, who said by touching the body, sounding clicks
vertebrates lower than mammals the ce­ he thought it was 1.5 square feet-"at or tones in the ear or flashing light in
rebral cortex is minuscule, if it can be least that is what Canadians have.") The the eye. Similarly, motor areas could
said to exist at all. Suddenly impressive folding is presumably mainly the result be mapped by stimulating the cortex
in the lowest mammals, it begins to of such an unlikely structure's having to electrically and noting what part of the
dominate the brain in carnivores, and it be packed into a box the size of the skull. animal's body moved.
increases explosively in primates; in A casual glance at cortical tissue un­
man it almost completely envelops the
rest of the brain, tending to obscure the
der a microscope shows vast numbers of
neurons: about 105 (100,000) for each T soon
his systematic mapping of the cortex
led to a fundamental realiza­
other parts. The degree to which an an­ square millimeter of surface, suggesting tion: most of the sensory and motor
imal depends on an organ is an index that the cortex as a whole has some 10lD areas contained systematic two-dimen­
of the organ's importance that is even ( 10 billion) neurons. The cell bodies are sional maps of the world they represent­
more convincing than size, and depen­ arranged in half a dozen layers that are ed. Destroying a particular small region
dence on the cortex has increased rapid­ alternately cell-sparse and cell-rich. In of cortex could lead to paralysis of one
ly as mammals have evolved. A mouse contrast to these marked changes in cell arm; a similar lesion in another small
without a cortex appears fairly nor­ density in successive layers at different region led to numbness of one hand or
mal, at least to casual inspection; a man depths in the cortex there is marked uni­ of the upper lip, or blindness in one
without a cortex is almost a vegetable, formity from place to place in the plane small part of the visual world; if elec­
speechless, sightless, senseless. of any given layer and in any direction trodes were placed on an animal's cor­
Understanding of this large and indis­ within that plane. The cortex is morpho­ tex, touching one limb produced a cor­
pensable organ is still woefully defi­ logically rather uniform in two of its respondingly localized series of electric
cient. This is partly because it is very dimensions. potentials. Clearly the body was system­
complex, not only structurally but also One of the first great insights about atically mapped onto the somatic senso­
in its functions, and partly because neu­ cortical organization came late in the ry and motor areas; the visual world
robiologists' intuitions about the func­ 19th century, when it was gradually re­ was mapped onto the primary visual
tions have so often been wrong. The out­ alized that this rather uniform plate of cortex, an area on the occipital lobe that
look is changing, however, as techniques tissue' is subdivided into a number of in man and in the macaque monkey (the
improve and as investigators learn how different regions that have very differ­ animal in which our investigations have
to deal with the huge numbers of intri­ ent functions. The evidence came from mainly been cond ucted) covers about 15
cately connected neurons that are the clinical, physiological and anatomical square centimeters.
basic elements of the cortex, with the sources. It was noted that a brain injury, In the primary visual cortex the map
impulses they carry and with the syn­ depending on its location, could cause is uncomplicated by breaks and discon­
apses that connect them. In this article paralysis or blindness or numbness or tinuities except for the remarkable split
we hope to sketch the present state of speech loss; the blindness could be total of the visual world down the exact mid­
knowledge of one subdivision of the or limited to half or less of the visual dle, with the left half projected to the
cortex: the primary visual cortex (also world, and the numbness could involve right cerebral cortex and the right half
known as the striate cortex or area 17), one limb or a few fingers. The consisten­ projected to the left cortex. The map of
the most elementary of the cortical re­ cy of the relation between a given defect the body is more complicated and is still
gions concerned with vision. That will and the location of the lesion gradually perhaps not completely understood. It is
necessarily lead us into the related sub­ led to a charting of the most obvious nonetheless systematic, and it is similar­
ject of visual perception, since the work­ of these specialized regions, the visual, ly crossed, with the right side of the
ings of an organ cannot easily be sepa­ auditory, somatic sensory (body sensa­ body projecting to the left hemisphere
rated from its biological purpose. tion), speech and motor regions. and the left side projecting to the right
In many cases a close look with a mi­ hemisphere. (It is worth remarking that

T he cerebral cortex, a highly folded croscope at cortex stained for cell bod­ no one has the remotest idea why there
plate of neural tissue about two mil­ ies showed that in spite of the relative should be this amazing tendency for ner­
limeters thick, is an outermost crust uniformity there were structural varia­ vous-system pathways to cross.)
wrapped over the top of, and to some tions, particularly in the layering pat­ An important feature of cortical maps
extent tucked under, the cerebral hemi­ tern, that correlated well with the clin­ is their distortion. The scale of the maps
spheres. In man its total area, if it were ically defined subdivisions. Additional varies as it does in a Mercator projec­
spread out, would be about 1.5 square confirmation came from observations tion, the rule for the cortex. being that

150

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
OCULAR-DOMINANCE COLUMNS, one of the two major sys­ ence. The autoradiograph was made by injecting a radioactively la­
tems that characterize the functional architecture of the primary vi­ beled amino acid into the right eye of an anesthetized animal. The
sual cortex, are revealed as periodic hright patches in this dark-field amino acid was taken up by cell bodies in the retina and transport­
autoradiograph of a section of macaque monkey cortex. The columns ed via the lateral geniculate nucleus, a way station in the brain, to
(actuaUy curving slabs of cortex, seen here in cross section in a brain cells in the cortex. A brain slice was coated with a photographic emul­
slice cut perpendicularly to the surface) are regions in which all neu­ sion, which was exposed for several months and then developed. Ex­
rons respond more actively to the right eye than to the left one; dark posed silver grains overlying the regions of radioactivity form the
regions separating the bright patches are columns of left-eye prefer- light-scattering patches that represent ocular-dominance columns.

DOMINANCE PATTERN is seen face on in an axonal-transport the label therefore accumulates. This section was cut in a plane tan­
autoradiograph of a brain section parallel, rather than perpendicu­ gential to the dome-shaped surface of the cortex and just below lay­
lar, to the surface of the primary visual cortex. As can be seen in the er IV, which therefore appears as a ring of roughly parallel bright
autoradiograph at the top of the page, the label is brightest in one bands. These are the radioactively labeled ocular-dominance regions,
layer of the folded cortex, layer IV. This is tbe level at which the ax­ which are now seen from above instead of edge on. The actual width
ons bringing visual information to the cortex terminate and where of the ocular-dominance regions is typically about .4 millimeter.

151

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
PRIMARY VISUAL CORTEX, also known as the striate cortex or the two occipital lobes. It also curves around the medial surface be­
area 17, is a region of the cerebral cortex: a layered plate of neurons tween the two cerebral hemispheres. It continues in a complex fold
that envelops the primate brain. In the macaque brain, seen here underneath the convex outer surface, as is shown in a parasagittal
from the side (left) and from above and behind (right), the primary section (see top illustration on opposite page) that was cut along the
visual cortex (colored areas) occupies most of the exposed surface of colored line and is viewed in the direction indicated by the arrows.

the regions of highest discrimination or have been revising their estimates up­ The implications of this are far-reach­
delicacy of function occupy relatively ward. The important basic notion is that ing. Whatever any given region of the
more cortical area. For the body sur­ information on any given modality such cortex does, it does locally. At stages
face, a millimeter of surface on the fin­ as sight or sound is transmitted first to a where there is any kind of detailed, sys­
gers, the lips or the tongue projects to primary cortical area and from there, tematic topographical mapping the anal­
more cortex than a millimeter of trunk, either directly or via the thalamus, to ysis must be piecemeal. For example,
buttocks or back; in vision the central successions of higher areas. A modern in the somatic sensory cortex the mes­
part of the retina has a representation guess as to the number of cortical areas sages concerning one finger can be com­
some 35 times more detailed than the might be between 50 and 100. bined and compared with an input from
far peripheral part. elsewhere on that same finger or with
Important as the advances in mapping
cortical projections were, they tended T he second major insight into cortical input from a neighboring finger, but
organization came from the work of they can hardly be combined with the
for some time to divert thought from the the anatomist Santiago Ramon y Cajal - influence from the trunk or from a foot.
real problem of just how the brain ana­ and his pupil Rafael Lorente de No. n\.e same applies to the visual world.
lyzes information. It was as though the This was the realization that the op­ Given the detailed order of the input to
representation could be an end in itself erations the cortex performs on the' in­ the primary visual cortex, there is no
instead of serving a more subtle pur­ formation it receives are local. What likelihood that the region will do any­
pose-as though what the cortex did was that means can best be understood by thing to correlate information coming
to cater to some little green man who sat considering the wiring diagram that in from both far above and far below
inside the head and surveyed images emerged from the Golgi method used by the horizon, or from both the left and
playing across the cortex. In the course Cajal and Lorente de No. In essence the the right part of the visual scene. It fol­
of this article we shall show that, for wiring is simple. Sets of fibers bring in­ lows that this cannot by any stretch of
vision at least, the world is represented formation to the cortex; by the time sev­ the imagination be the place where ac­
in a far more distorted way; any little eral synapses have been traversed the tual perception is enshrined. Whatever
green man trying to glean information influence of the input has spread verti­ these cortical areas are doing, it must be
from the cortical projection would be cally to all cell layers; finally several some kind of local analysis of the senso­
puzzled indeed. other sets of fibers carry modified mes­ ry world. One can only assume that as
The first major insight into cortical sages out of the area. The detailed con­ the information on vision or touch or
organization was nonetheless the recog­ nections between inputs and outputs dif­ sound is relayed from one cortical area
nition of this subdivision into areas hav­ fer from one area to the next, but with­ to the next the map becomes progres­
ing widely different functions, with a in a given area they seem to be rather sively more blurred and the information
tendency to ordered mapping. Just how stereotyped. What is common to all re­ carried more abstract.
many such areas there are has been a gions is the local nature of the wiring. Even though the Golgi-method stud­
subject of wide speculation. Anatomists' The information carried into the cortex ies of the early 1900's made it clear that
estimates have on the whole been rather by a single fiber can in principle make the cortex must'perform local analyses,
high-up to several hundred areas, de­ itself felt through the entire thickness in it was half a century before physiolo­
pending on the individual worker's sen­ about three or four synapses, whereas gists had the least inkling of just what
sitivity to fine differences in microscopic the lateral spread, produced by branch­ the analysis was in any area of the cor­
patterns and sometimes also on his abili­ ing trees of axons and dendrites, is lim­ tex. The first understanding came in the
ty to fool himself. Physiologists began ited for all practical purposes to a few primary visual area, which is now the
with lower estimates, but lately, with millimeters, a small proportion of the best-understood of any cortical region
more powerful mapping methods, they vast extent of the cortex. and is still the only one where the analy-

152

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
sis and consequent transformations of
information are known in any detail. Af­
ter describing the main transformations
that take place in the primary visual cor­
tex we shall go on to show how increas­
ing understanding of these cortical func­
tions has revealed an entire world of ar­
chitectural order that is otherwise inac­
cessible to observation.

e can best begin by tracing the vi­

W sual path in a primate from the
retina to the cortex. The output from
each eye is conveyed to the brain by
about a million nerve fibers bundled
together in the optic nerve. These fibers
are the axons of the ganglion cells of the
retina. The messages from the light-sen­
sitive elements, the rods and cones, have
already traversed from two to four syn­
apses and have involved four other
types of retinal cells before they arrive
at the ganglion cells, and a certain
amount of sophisticated analysis of the
information has already taken place.
A large fraction of the optic-nerve fi­
bers pass uninterrupted to two nests of
cells deep in the brain called the lateral
geniculate nuclei, where they make syn­
apses. The lateral geniculate cells in turn
send their axons directly to the primary
visual cortex. From there, after sever­
al synapses, the messages are sent to a
number of further destinations: neigh­
boring cortical areas and also several
targets deep in the brain. One contingent
even projects back to the lateral genicu­
late bodies; the function of this feedback
path is not known. The main point for
the moment is that the primary visual
cortex is in no sense the end of the visual
path. It is just one stage, probably an SECTION OF VISUAL CORTEX along the colored line in the illustration on the opposite
early one in terms of the degree of ab­ page was stained by the Nissl method, which makes cell bodies but not fibers visible. The visual
straction of the information it handles. cortex is seen to be a continuous layered sheet of neurons about two millimeters thick. The
As a result of the partial crossing of black rectangle outlines a section like the one that is furtber enlarged in the illustration below.
the optic nerves in the optic chiasm, the
geniculate and the cortex on the left side
are connected to the two left half retinas
and are therefore concerned with the
right half of the visual scene, and the
II
converse is the case for the right genicu­
late and the right cortex. Each genicu­
late and each cortex receives input from III

both eyes, and each is concerned with

the opposite half of the visual world. IVa
To examine the workings of this visu­
al pathway our strategy since the late
1950's has been (in .principle) simple. IVb
Beginning, say, with the fibers of the
optic nerve, we record with microelec­
trodes from a single nerve fiber and try
IVe
to find out how we can most effectively
influence the firing by stimulating the
retina with light. For this one can use V
patterns of light of every conceivable
size, shape and color, bright on a dark
VI
background or the reverse, and station­
ary or moving. It may take a long time,
but sooner or later we satisfy ourselves
that we have found the best stimulus for CROSS SECTION OF PRIMARY VISUAL CORTEX in the macaque, stained here by the
the cell being tested, in this case a gangli­ Nissl method and enlarged about 3S diameters, shows the layered structure and gives the con.
on cell of the retina. (Sometimes we are ventional designations of the six layers (left). The white gaps are sectioned blood vessels.

153

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
wrong!) We note the results and then go symmetry they respond well to such a the ones most immediately connected to
on to another fiber. After studying a few line stimulus whatever its orientation. the input.
hundred cells we may find that new To sum up, the retinal ganglion cells and Cells outside layer IV all respond best
types become rare. Satisfied that we the cells of the lateral geniculate-the to specifically oriented line segments. A
know roughly how the neurons at this cells supplying the input to the visual typical cell responds only when light
stage work, we proceed to the next stage cortex-are cells with concentric, cen­ falls in a particular part of the visual
(in this case the geniculate) and repeat ter-surround receptive fields. They are world, but illuminating that area dif­
the process. Comparison of the two sets primarily concerned not with assessing fusely has little effect or none, and small
of results can tell us something about levels of illumination but rather with spots of light are not much better. The
what the geniculate does. We then go on making a comparison between the light best response is obtained when a line
to the next stage, the primary cortex, level in one small area of the visual that has just the right tilt is flashed in the
and repeat the procedure. scene and the average illumination of region or, in some cells, is swept across
Working in this way, one finds that the immediate surround. the region. The most effective orien­
both a retinal ganglion cell and a genic­ tation varies from cell to cell and is usu­
ulate cell respond best to a roughly cir­
cular spot of light of a particular size in T hetionsfirstaccomplished
of the two major transforma­
by the visual
ally defined sharply enough so that a
change of 10 or 20 degrees clockwise or
a particular part of the visual field. The cortex is the rearrangement of incoming counterclockwise reduces the response
size is critical because each cell's recep­ information so that most of its cells re­ markedly or abolishes it. (It is hard to
tive field (the patch of retinal receptor spond not to spots of light but to specifi­ convey the precision of this discrimina­
cells supplying the cell) is divided, with cally oriented line segments. There is a tion. If 10 to 20 degrees sounds like a
an excitatory center and an inhibitory wide variety of cell types in the cortex, wide range, one should remember that
surround (an "on center" cell) or exactly some simpler and some more complex the angle between 12 o'clock and one
the reverse configuration (an "off cen­ in their response properties, and one o'clock is 30 degrees.) A line at 90 de­
ter" cell). This is the center-surround soon gains an impression of a kind of grees to the best orientation almost nev­
configuration first described by Stephen hierarchy, with simpler cells feeding er evokes any response.
W. Kuffier at the J ohns Hopkins Univer­ more complex ones. In the monkey Depending on the particular cell, the
;ity School of Medicine in 1953. A spot there is first of all a large group of cells stimulus may be a bright line on a dark
exactly filling the center of an on-center that behave (as far as is known) just background or the reverse, or it may
cell is therefore a more effective stimu­ like geniculate cells: they have circularly be a boundary between light and dark
lus than a larger spot that invades the symmetrical fields. These cells are all in regions. If it is a line, the thickness is
inhibitory area, or than diffuse light. A the lower part of one layer, called layer likely to be important; increasing it be­
line stimulus (a bar of light) is effective IV, which is precisely the layer that re­ yond some optimal width reduces the
if it covers a large part of the center ceives the lion's share of the geniculate response, just as increasing the diameter
region and only a small part of the sur­ input. It makes sense that these least of a spot does in the case of ganglion and
round. Because these cells have circular sophisticated cortical cells should be geniculate cells. Indeed, for a particu­
lar part of the visual field the geniculate
receptive-field centers and the optimal
cortical line widths are comparable.

N ty vary in their complexity.

eurons with orientation specifici­
The
simplest, which we call "simple" cells,
behave as though they received their
input directly from several cells with
center-surround, circularly symmetri­
cal fields-the type of cells found in lay­
er IV. The response properties of these
simple cells, which respond to an opti­
mally oriented line in a narrowly defined
location, can most easily be accounted
for by requiring that the centers of the
incoming center-surround fields all be
excitatory or all be inhibitory, and that
they lie along a straight line. At pres­
ent we have no direct evidence for this
scheme, but it is attractive because of
its simplicity and because certain kinds
of indirect evidence support it. Accord­
ing to the work of Jennifer S. Lund of the
University of Washington School of
Medicine, who in the past few years has
done more than anyone else to advance
RIGHT EYE the Golgi-stain anatomy of this corti­
cal area, the cells in layer IV project to
the layers just above, which is roughly
where the simple cells are found.
The second major group of orienta­
VISUAL PATHWAY is traced schematically in the human brain, seen here from below. The
tion-specific neurons are the far more
output from the retina is conveyed, by ganglion-cell axons bundled in the optic nerves, to the
numerous "complex" cells. They come
lateral geniculate nuclei; about half of the axons cross over to the opposite side of the brain, in a number of subcategories, but their
so that a representation of each half of the visual scene is projected on the geniculate of the main feature is that they are less par­
opposite hemisphere. Neurons in the geniculates send their axons to the primary visual cortex. ticular about the exact position of a line.

154

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
RIGHT EYE LEFT EYE RIGHT EYE

LA TERAL GENICULATE NUCLEUS of a normal monkey (left) any radius (black iiI/e) receive signals from the same part of the visu­
is a layered structure in which cells in layers 1,4 and 6 (numbered from al scene. The layered nature of the input is demonstrated in the two
bottom to top) receive their input from the eye on the opposite side geniculates of an animal that had vision in the left eye only (two micro­
and those in layers 2, 3 and 5 receive information from the eye on graphs at right): in each geniculate cells in the three layers with input
the same side. The maps are in register, so that the neurons along from right eye have atrophied. Geniculates are enlarged 10 diameters.

Complex cells behave as though they ysis of visual forms. It is worth asking tivated will be those whose optimal ori­
received their input from a number of which cells at this early stage would be entation happens to coincide with the
simple cells, all with the same receptive­ expected to be turned on by some very prevailing direction of the border. And
field orientation but differing slightly simple visual form, say a dark blob on among these the simple cells will be
in the exact location of their fields. a light background. Any cell whose re­ much more exacting than the complex
This scheme readily explains the strong ceptive field is entirely inside or outside ones, responding optimally only when
steady firing evoked in a complex cell as the boundaries of such an image will the border falls along a line separating
a line is kept in the optimal orientation be completely unaffected by the figure's an excitatory and an inhibitory region. It
and is swept across the receptive field. presence because cortical cells effective­ is important to realize that this part of
With the line optimally oriented many ly ignore diffuse changes in the illumina­ the cortex is operating only locally, on
cells prefer one direction of movement tion of their entire receptive fields. bits of the form; how the entire form is
to the opposite direction. Several possi­ The only cells to be affected will be analyzed or handled by the brain-how
ble circuits have been proposed to ex­ those whose field is cut by the borders. this information is worked on and syn­
plain this behavior, but the exact mecha­ For the circularly symmetrical cells the thesized at later stages, if indeed it is-is
nism is still not known. ones most strongly influenced will be still not known.
Although there is no direct evidence those whose center is grazed by a bound­
that orientation-sensitive cells have any­
thing to do with visual perception, it is
ary (because for them the excitatory and
inhibitory subdivisions are most un­ T he second major function of the
monkey visual cortex is to combine
certainly tempting to think they repre­ equally illuminated). For the orienta­ the inputs from the two eyes. In the lat­
sent some early stage in the brain's anal- tion-specific cells the only ones to be ac- eral geniculate nuclei a neuron may re-

a b

/
I
\
"

"
,
I

RECEPTIVE FIELDS of various cells in the visual pathway are creasingly complex response properties. The cortical cells that receive
compared. Retinal ganglion cells and neurons in the lateral genicu­ signals directly from the geniculate have circularly symmetrical fields.
late nucleus have circular fields with either an excitatory center and Cortical cells farther along the pathway, however, respond only to a
an inhibito ry surround (a) or the opposite arrangement. A spot of light line stimulus in a particular orientation. A "simple" cell (b) responds
falling on the center stimulates a response from such a cell; so does a to such a line stimulus only in a particular part of its field. A "com­
bar of light falling on the field in any orientation, provided it falls on plex" cell (c) responds to a precisely oriented line regardless of where
the center. In the v isual cortex there is a hierarchy of neurons with in- it is in its field and also to one moving in a particular direction(arrow).

155

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
spond to stimulation of the left eye or of If one maps the right-eye and left-eye should be present in two duplicate cop­
the right one, but no cell responds to receptive fields of a binocular cell (by ies. It is perhaps even more surpris­
stimulation of both eyes. This may seem stimulating first through one eye and ing that all of this can be observed in a
surprising, since each geniculate re­ then through the other) and compares newborn animal. The wiring is mostly
ceives inputs from both eyes, but the the two fields, the fields turn out to have innate, and it presumably is geneti­
fact is that the geniculates are construct­ identical positions, levels of complex­ cally determined. (In one particular re­
ed in a way that keeps inputs from the ity, orientation and directional prefer­ spect, however, some maturation of bin­
two eyes segregated. Each geniculate ence; everything one learns about the ocular wiring does take place mostly
body is divided into six layers, three left­ cell by stimulating one eye is confirmed after birth.)
eye layers interdigitated with three through the other eye. There is only one
right-eye ones. The opposite-side half of exception: if first one eye and then the e now turn to a consideration of
the visual world is mapped onto each other are tested with identical stimuli, W the way these cells are grouped in
layer (with the six maps in precise regis­ the two responses are usually not quan­ the cortex. Are cells with similar char­
ter, so that in a radial pathway travers­ titatively identical; in many cases one acteristics-complexity, receptive-field
ing the six layers the receptive fields of eye is dominant, consistently producing position, orientation and ocular domi­
all the cells encountered have virtual­ a higher frequency of firing than the nance-grouped together or scattered at
ly identical positions in the visual field). other eye. random? From the description so far it
Since any one layer has input from only From cell to cell all degrees of ocu­ will be obvious that cells of like com­
one eye, the individual cells of that layer lar dominance can be found, from com­ plexity tend to be grouped in layers, with
must be monocular. plete monopoly by one eye through the circularly symmetrical cells low in
Even in the visual cortex the neurons equality to exclusive control by the oth­ layer IV, the simple cells just above
to which the geniculate cells project di­ er eye. In the monkey the cells with a them and the complex cells in layers II,
rectly, the circularly symmetrical cells marked eye preference are somewhat III, V and VI. Complex cells can be fur­
in layer IV, are all (as far as we can tell) commoner than the cells in which the ther subcategorized, and the ones found
strictly monocular; so are all the simple two eyes make about equal contribu­ in each layer are in a number of ways
cells. Only at the level of the complex tions. Apparently a binocular cell in the very different.
cells do the paths from the two eyes primary visual cortex has connections These differences from layer to layer
converge, and even there the blending to the two eyes that are qualitatively vir­ take on added interest in view of the
of information is incomplete and takes tually identical, but the density of the important discovery, confirmed by sev­
a special form. About half of the com­ two sets of connections is not necessar­ eral physiologists and anatomists during
plex cells are monocular, in the sense ily the same. the past few decades, that fibers project­
that any one cell can be activated only It is remarkable enough that the elab­ ing from particular layers of the cortex
by stimulating one eye. The rest of the orate sets of wiring that produce speci­ have particular destinations. For exam­
cells can be influenced independently by ficity of orientation and of direction of ple, in the visual cortex the deepest lay­
both eyes. movement and other special properties er, layer VI, projects mainly (perhaps
only) back to the lateral geniculate
body; layer V projects to the superior
colliculus, a visual station in the mid­
brain; layers II and III send their pro­
2 jections to other parts of the cortex.
3
This relation between layer and projec­
tion site probably deserves to be ranked
4
as a third major insight into cortical or­
5 ganization.
The next stimulus variable to be con­
6
sidered is the position of the receptive
7 field in the visual field. In describing the
lateral geniculate nucleus we pointed
8 out that in each layer the opposite-half
visual field forms an ordered topograph­
9
ical map. In the projection from lateral
geniculate to primary visual cortex this
order is preserved, prod ucing a cortical
map of the visual field. Given this or­
dered map it is no surprise that neigh­
boring cells in this part of the cortex
always have receptive fields that are
4
close together; usually, in fact, they
overlap. If one plunges a microelectrode
into the cortex at a right angle to the
surface and records from cell after cell
(as many as 100 or 200 of them) in suc­
cessively deeper layers, again the recep­
tive fields mostly overlap, with each new
field heaped on all the others. The ex­
tent of the entire pile of fields is usually
several times the size of any one typi­
POSITIONS OF RECEPTIVE FIELDS (nllmbered from 1 10 9) of cortical neurons mapped
cal field.
by an electrode penetrating at roughly a right angle to the surface are essentially the same (iefl),
although the fields are different sizes and there is some scatter. In an oblique penetration(righl) There is some variation in the size of
from two to four cells were recorded, at .I-millimeter intervals, at each of four sites (nllmbered these receptive fields. Some of the varia­
from 1 (04) one millimeter apart. Each group includes various sizes aud some scatter, but now tion is tied to the layering: the largest
there is also a systematic drift: fields of each successive group of cells are somewhat displaced. fields in any penetration tend to be in

156

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
 150
120
SURFACE OF CORTEX
-
--
90 �
2120
Cii
w
w
60 : \
, �
1918
a:
CJ
w
30 �
1 ,-'1
���; 15167 � -30 ,
e- O ,
z "
� •

14 � -60 �. z
•

� -90 LAYER IV
5
-120 o

-150
-180� .5 o
__ � ____

1 1.5
-L__��__��__�
2 2.5
,

TRACK DISTANCE (MILLlME;TERS)

ORIENTATION PREFERENCES of 23 neurons encountered as a witb tbe observation tbat a microelectrode penetrating tbe cortex per­
microelectrode penetrated tbe cortex obliquely are cbarted (left); tbe pendicularly encounters only cells tbat prefer tbe same orientation
most effective tilt of tbe stimulus cbanged steadily in a counterclock­ (apart from tbe circularly symmetrical cells in layer IV, wbicb bave
wise direction. Tbe results of a similar experiment are plotted (center); no preferred orientation), suggested tbat tbe cortex is subdivided into
in tbis case, bowever, tbere were several reversals in direction of ro­ rougbly parallel slabs of tissue, witb eacb slab, called an orientation
tation. Tbe results of a large number of sucb experiments, togetber column, containing neurons witb like orientation specificity (right).

ORIENTATION COLUMNS are visualized as anatomical struc­ jection tbe animal was stimulated witb a pattern of vertical stripes, so
tures in a deoxyglucose autoradiograpb made by tbe autbors and tbat cells responding to vertical lines were most active and became
Micbael P. Stryker. Radioactively labeled deoxyglucose was injected most radioactive. In tbis section perpendicular to surface active-cell
into a monkey; it was taken up primarily by active neurons, and an regions are narrow bands about .5 millimeter apart. Layer IV (witb
early metabolite accumulated in tbe cells. Immediately after tbe in- no orientation preference) is, as expected, uniformly radioactive.

ORIENTATION PATTERN, seen face on, is unexpectedly com­ sent continuously labeled layer IV. In tbe otber layers tbe orientation
plex. Tbis deoxyglucose autoradiograpb is of a section tangential to regions are intricately curved bands, sometbing like tbe walls of a
tbe somewbat curved layers of tbe cortex. Tbe darker regions repre- maze seen from above, but distance from one band to next is uniform.

157

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
LEFT EYE RIGHT EYE where a block of cortex about one or
two millimeters in size is what is needed
to take care of a region of the visual
world equivalent to the size of an aggre­
gate field.
These observations suggest the way
the visual cortex solves a basic problem:
how to analyze the visual scene in detail
in the central part and much more
crudely in the periphery. In the retina,
which has the same problem, for obvi­
ous optical reasons the number of milli­
meters corresponding to a degree of vi­
HORIZONTAL HORIZONTAL sual field is constant. The retina handles
the central areas in great detail by hav­
VERTICAL VERTICAL ing huge numbers of ganglion cells, each
subserving a tiny area of central visual
BINOCULAR CELL in tbe cortex can be influenced independently by botb eyes or more
field; the layer of ganglion cells in the
strongly by botb eyes togetber. Here tbe left-eye and rigbt-eye fields are mapped for a com­
plex cell wbose receptive field is in tbe upper left quadrant of tbe visual field. (Tbe lines repre­
central part of the retina is thick, where­
sent tbe borizontal and vertical meridians of tbe field, intersecting at tbe point of fixation.) Tbe as in the outlying parts of the retina it is
two receptive fields are identical, but tbe amount of response may differ depending on wbetber very thin. The cortex, in contrast, seems
tbe left eye or tbe rigbt eye is stimulated. Preference for one eye is called ocular dominance. to want to be uniform in thickness ev­
erywhere. Here there are none of the
optical constraints imposed on the reti­
layers III, V and VI. The most important a consistent drift in field position, its di­ na, and so area is simply allotted in
variation, however, is linked to eccen­ rection dictated by the topographical amounts corresponding to the problem
tricity, or the distance of a cell's recep­ map of the visual fields. And an interest­ at hand.
tive field from the center of gaze. The ing regularity is revealed: it turns out The machinery in any square millime­
size of the fields and the extent of the that moving the electrode about one or ter of cortex is presumably about the
associated scatter in the part of the cor­ two millimeters always produces a dis­ same as in any other. A few thousand
tex that maps the center of gaze are tiny placement in visual field that is roughly geniculate fibers enter such a region, the
compared to the size and amount of enough to take one into an entirely new cortex does its thing and perhaps 50,000
scatter in the part that maps the far pe­ region. The movement in the visual fibers leave-whether a small part of the
riphery. We call the pile of superim­ field, in short, is about the same as the visual world is represented in great de­
posed fields that are mapped in a pene­ size of the aggregate receptive field. For tail or a larger part in correspondingly
tration beginning at any point on the the primary visual cortex this holds less detail. The uniformity of the cortex
cortex the "aggregate field" of that wherever the recording is made. At the is suggested, as we indicated at the out­
point. The size of the aggregate field is center of gaze the fields and their associ­ set, by the appearance of stained sec­
obviously a function of eccentricity. ated scatter are tiny, but so is the dis­ tions. It is compellingly confirmed when
placement corresponding to a one-milli­ we examine the architecture further,
fthe electrode penetrates in an meter movement along the cortex. With looking specifically at orientation and at
I oblique direction, almost parallel to increasing eccentricity (farther out in ocular dominance.
the surface, the scatter in field position the visual field) both the field and scat­
from cell to cell is again evident, but ter and the displacement become larger, or orientation we inquire about
now there is superimposed on the scatter in parallel fashion. It seems that every- F groupings of cells just as we did with
field position, looking first at two cells
sitting side by side. Two such cells al­
most invariably have the same optimal
3' 1 SURFACE OF CORTEX 2
stimulus orientation. If the electrode is

� L, R,
inserted in a direction perpendicular to
the surface,' all the cells along the path
f'.. of penetration have identical or almost
� identical orientations (except for the

�
R,
cells deep in layer IV, which have no
optimal orientation at all). In two per­
L,

� I'-....
R,
R, pendicular penetrations a millimeter or
so apart, however, the two orientations
observed are usually different. The cor­
LAYER IV L � L R
tex must therefore be subdivided by

�'"
some kind of vertical partitioning into
L, R,
regions of constant receptive-field ori­
entation. When we came on this sys­
I'-..... tem almost 20 years ago, it intrigued us
\ , II � because it fitted so well with the hier­
archical schemes we had proposed to
GROUPING OF CELLS according to ocular dominance was revealed by pbysiological explain how complex cells are supplied
studies. In one typical vertical penetration of tbe cortex (1) a microelectrode encounters only
by inputs from simple cells: the circuit
cells tbat respond preferentially to tbe left eye (Lr) and, in layer IV, cells tbat respond only to tbe
diagrams involve connections between
left eye (L); in anotber vertical penetration (2) tbe cells all bave rigbt-eye dominance (R /) or, in
layer IV, are driven exclusively by tbe rigbt eye (R). In an oblique penetration (3) tbere is a
cells whose fields cover the same part
regular alternation of dominance by one eye or tbe otber eye. Repeated penetrations suggest of the visual world and that respond to
tbat tbe cortex is subdivided into regions witb a cross-sectional widtb of about .4 millimeter and the same line orientation. It seemed em­
witb walls perpendicular to tbe cortical surface and layers: tbe ocular-dominance columns. inently reasonable that strongly inter-

158
© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
connected cells should be grouped to­ at the National Institute of Mental active the neuron, the greater the up­
gether. Health and described elsewhere in this take. The compound begins to be me­
If the cortex is diced up into small issue [see "The Chemistry of the Brain," tabolized, but for reasons best known
regions of constant receptive-field orien­ by Leslie L. Iversen, page 134]. The to biochemists the sequence stops with
tation, can one say anything more about method capitalizes on the fact that brain a metabolite that cannot cross the cell
the three-dimensional shape of the re­ cells depend mainly on glucose as a wall and therefore accumulates within
gions than that their walls are perpen­ source of metabolic energy and that the the cell.
dicular to the surface? Are neighboring closely similar compound 2-deoxyglu­ The Sokoloff procedure is to inject an
regions related in any systematic way or cose can to some extent masquerade as animal with deoxyglucose that has been
are regions subserving all the possible glucose. If deoxyglucose is injected into labeled with the radioactive isotope car­
orientations scattered over the cortex at an animal, it is taken up actively by neu­ bon 14, stimulate the animal in a way
random? We began to study these ques­ rons as though it were glucose; the more calculated to activate certain neurons
tions simply by penetrating the cortex
obliquely or parallel to the surface.
When we first did this experiment in
about 196 1, the result was so surpris­
ing that we could hardly believe it. In­
stead of a random assortment of succes­
sive orientations there was an amazing
orderliness. Each time the electrode
moved forward as little as 25 or 50 mi­
crometers (thousandths of a millimeter)
the optimal orientation changed by a
small step, about 10 degrees on the aver­
age; the steps continued in the same di­
rection, clockwise or counterclockwise,
through a total angle of anywhere from
90 to 270 degrees. Occasionally such a
sequence would reverse direction sud­
denly, from a clockwise progression to
a counterclockwise one or vice versa.
These reversals were unpredictable,
usually coming after steady progres­
ANATOMICAL CONFIRMATION of ocular-dominance columns came from various stain­
sions of from 90 to 270 degrees.
ing methods and from axonal-transport autoradiographs such as those shown in color on page
Since making this first observation we
151. This composite autoradiograph visualizing the pattern over an area some 10 millimeters
have seen similar order in almost every wide was made by cutting out and pasting together the regions representing layer IV in a num­
monkey. Either there is a steady pro­ ber of parallel sections: the one in bottom illustration on page 151 and others at different depths.
gression in orientation or, less frequent­
ly, there are stretches in which orien­
tation stays constant. The successive
changes in orientation are small enough
so that it is hard to be sure that the re­
gions of constant orientation are finite in
size; it could be that the optimal orienta­
tion changes in some sense continuously
as the electrode moves along the cortex.
We became increasingly interested in
the three-dimensional shape of these
regional subdivisions. From considera­
tions of geometry alone the existence of
small or zero changes in every direction
during a horizontal or tangential pene­
tration points to parallel slabs of tissue
containing cells with like orientation
specificity, with each slab perpendicular
to the surface. The slabs would not nec­
essarily be planar, like slices of bread;
seen from above they might well have
the form of swirls, which could easiiy
explain the reversals in the direction
of orientation changes. Recording large
numbers of cells in several parallel elec­
trode penetrations seemed to confirm
this prediction, but it was hard to exam­
ine more than a tiny region of brain with
the microelectrode.

ortunately an ideal anatomical meth­

F od was invented at just the right time RECONSTRUCTION of the ocular-dominance pattern over the entire exposed part of the
for us. This was the 2-deoxyglucose right primary visual cortex was made by the authors and Simon LeVay from a series of sec­
technique for assessing brain activity, tions stained by a reduced-silver method he developed. The left-hand margin is at the medial
devised by Louis Sokoloff and his group edge of occipital lobe, where cortex folds downward; pattern is enlarged about six diameters.

159

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
and then immediately examine the brain part of layer IV was uniformly radioac­ countered preferred the right eye, then
for radioactivity, which reveals active tive, as was expected from the fact that so did all the cells, right down to the
areas where cells will have taken up the cells in the layer have circularly bottom of layer VI; if the first cell pre­
more deoxyglucose than those in quies­ symmetrical receptive fields and show ferred the left eye, so did all the rest.
cent areas. The usual way of examining no orientation selectivity. Any penetration favored one eye or the
the brain for this purpose is to cut very Sections cut parallel to the surface other with equal probability. (Since the
thin slices of it (as one would for mi­ showed an unexpectedly complex set of cells of layer IV are monocular, there it
croscopic examination) and press them periodically spaced bands, often swirl­ was a matter not of eye preference but
against a photographic plate sensitive ing, frequently branching and rejoining, of eye monopoly.) If the penetration was
to the radioactive particles. When the only here and there forming regular par­ oblique or horizontal, there was an al­
film is developed, any areas that were allel slabs. What was particularly strik­ ternation of left and right preferences,
in contact with radioactive material are ing was the uniformity of the distance with a rather abrupt switchover about
seen as dark masses of developed silver from one band to the next over the en­ every half millimeter. The cortex thus
grains. Together with Michael P. Stry­ tire cortex. This fitted perfectly with the proved to be diced up into a second set
ker we adapted the Sokoloff method to idea of a uniform cortex. Moreover, the of regions separated by vertical walls
our problem, injecting an anesthetized distance between stripes fitted well with that extend through the full cortical
animal with deoxyglucose and then the idea that the cortical machinery thickness. The ocular-dominance sys­
moving a pattern of black and white ver­ must repeat itself at least every millime­ tem was apparently quite independent
tical stripes back and forth 1.5 meters in ter. If the distance were, for example, 10 of the orientation system, because in
front of the animal for 45 minutes. We millimeters from vertical through 180 oblique or tangential penetrations the
then cut the brain into slices, either per­ degrees and back to vertical, sizable two sequences had no apparent relation
pendicular to the surface of the cortex parts of the visual field would lack cells to each other.
or parallel to it. sensitive to any given orientation, mak­ The basis of these ocular-dominance
The autoradiographs quickly con­ ing for a sketchy and extremely bizarre columns, as they have come to be called,
firmed the physiological results. Sec­ representation of the visual scene. seems to be quite simple. The terminals
tions cut perpendicular to the surface of geniculate fibers, some subserving
showed narrow bands of radioactivity
about every 570 micrometers (roughly T he final variable whose associated
architecture needs to be considered
the left eye and others the right, group
themselves as they enter the cortex so
half a millimeter), extending through is eye preference. In microelectrode that in layer IV there is no mixing. This
the full thickness of the cortex. Evident­ studies neighboring cells proved almost produces left-eye and right-eye patches
ly these were the regions containing cells invariably to prefer the same eye. If in at roughly half-millimeter intervals. A
responsive to vertical lines. The deep vertical penetrations the first cell we en- neuron above or below layer IV receives
connections from that layer from up to
about a millimeter away in every direc­
tion. Probably the strongest connections
are from the region of layer IV closest
to the neuron, so that it is presumably

II
dominated by whichever eye feeds that

I' I1 I1 11 I 1 I 1 1
L region.
Again we were most curious to learn

111.' 1 '
what these left-eye and right-eye regions
I

111111 1 I 1 1 11 1, '1 /1 1 / 1. /1
might look like in three dimensions; any
L I.. of several geometries could lead to the
ill

i' ll 'I 1 I . / 1 .//

cross-sectional appearance the physi­
R ology had suggested. The answer first

1I I"I1 I1I
II11

1 I 1 1I1I 11
came from studies with the silver-degen­
.eration method for mapping connec­
L I tions, devised by Walle J. H. Nauta of

II ! I IIII!I I 'I 1 ,I 1 I 1 . 1 1
the Massachusetts Institute of Technol­
ogy. Since then we have found three oth­

!I
er independent anatomical methods for
I' I, '
1 demonstrating these columns.
A particularly effective method (be­
cause it enables one to observe in a sin­
,1 I I '
1'1 I'
,II
II gle animal the arrangement of columns
I I· I
over the entire primary visual cortex) is
1· 1· 1' 1, based on the phenomenon of axonal
;!;I transport. The proced ure is to inject a
1,1·
;tV
radioactively labeled amino acid into an
area of nervous tissue. A cell body takes
up the amino acid, presumably incorpo­
rates it into a protein and then transports
it along the axon to its terminals. When
we injected the material into one eye
of a monkey, the retinal ganglion cells
took it up and transported it along their
axons, the optic-nerve fibers. We could
BLOCK OF CORTEX about a millimeter square and two millimeters deep (light color) can be
considered an elementary unit of tbe primary visual cortex. It contains one set of orientation
then examine the destinations of these
slabs subserving all orientations and one set of ocular-dominance slabs subserving botb eyes.
fibers in the lateral geniculate nuclei by
Tbe pattern is reiterated tbrougbout tbe primary visual area. Tbe placing of tbe boundaries coating tissue slices with a silver emul­
(at tbe rigbt or tbe left eye, at a vertical, borizontal or oblique orientation) is arbitrary; rep­ sion and developing the emulsion; the
resentation of tbe slabs as flat planes intersecting at rigbt angles is an oversimplification. radioactive label showed up clearly in

160

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
the three complementary layers of the VERTICAL

\
geniculate on each side.
This method does not ordinarily trace
a path from one axon terminal across a
synapse to the next neuron and its termi­ R L R L
L R L R
L
nals, however, and we wanted to follow
the path all the way to the cortex. In
1 97 1 Bernice Grafstein of the Cornell
University Medical College discovered
that after a large enough injection in the
eye of a mouse some of the radioactive
material escaped from the optic-nerve
terminals and was taken up by the cells
in the geniculate and transported along
their axons to the cortex. We had the
thought that a similarly large injection
in a monkey, combined with autoradi­
ography, might demonstrate the genicu­
late terminals from one eye in layer IV
of the visual cortex.

O urnegative
first attempt yielded dismayingly
results, with only faint
hints of a few silver grains visible in lay­
er IV . It was only after several weeks HORIZONTAL
that we realized that by resorting to
HYPOTHETICAL PATTERN OF CORTICAL ACTIVITY that might result from stimula­
dark-field microscopy we could take ad­ tion of the left eye with a single short horizontal line, placed in the upper left quadrant of the
vantage of the light-scattering proper­ visual field, is shown by the colored patches on a diagram of an area of the right cortex, seen
ties of silver grains and so increase the face on. The area receiving input from the object in the visual field is indicated by the broken
sensitivity of the method. We borrowed black line. If ocular-dominance and orientation columns are arrayed as shown, activated cells
a dark-field condenser, and when we will be those that respond optimally to approximately horizontal stimuli from the left eye.
looked at our first slide under the mi­
croscope, there shining in all their glory
were the periodic patches of label in lay­ entire primary visual cortex, once more deep. To know the organization of this
er IV [see top illustration on page 151]. emphasizing the uniformity of the cor­ chunk of tissue is to know the organiza­
The next step was to try to see the tex. Again the widths fit perfectly with tion for all of area 17; the whole must be
pattern face on by sectioning the cortex the idea that all of the apparatus needed mainly an iterated version of this ele­
parallel to its surface. The monkey cor­ to look after an area the size of an aggre­ mentary unit. Of course the elementary
tex is dome-shaped, arid so a section par­ gate field must be contained within any unit should not be thought of as a dis­
allel to the surface and tangent to layer square millimeter of cortex. The two crete, separable block. Whether the set
IV shows that layer as a circle or an techniques, deoxyglucose labeling and of orientation slabs begins with a slab
oval, while a section below layer IV amino acid transport, have the great ad­ representing a vertical orientation, an
shows it as a ring. By assembling a series vantage of being mutually compatible, oblique one or a horizontal one is com­
of such ovals and rings from a set of so that we have been able to apply both pletely arbitrary; so too is whether an
sections one can reconstruct the pattern together, one to mark orientation lines ocular-dominance sequence begins with
over a wide expanse of cortex. and the other to see the ocular-domi­ a left-plus-right pair of dominance slabs
From the reconstructions it was im­ nance columns. The number of brains or a right-plus-left pair. The same thing
mediately obvious that the main overall examined so far is too small to justify is true for a unit crystal of sodium chlo­
pattern is one of parallel stripes repre­ any final conclusions, but the two sys­ ride or for any complex repetitive pat­
senting terminals belonging to the in­ tems appear to be quite independent, tern such as is found in wallpaper.
jected eye, separated by gaps represent­ neither parallel nor at right angles but What, then, does the visual scene real­
ing the other eye. The striping pattern is intersecting at random. ly look like as it is projected onto the
not regular like wallpaper. (We remind The function served by ocular-domi­ visual cortex? Suppose an animal fixes
ourselves occasionally that this is, after nance columns is still a mystery. We its gaze on some point and the only ob­
all, biology!) Here and there a stripe know there are neurons with all grades ject in the visual field is a straight line
representing one eye branches into two of eye preference throughout the entire above and a bit to the left of the point
stripes, or else it ends blindly at a point binocular part of the visual fields, and it where the gaze is riveted. If each active
where a stripe from the other eye may be that a regular, patterned system cell were to light up, and if one could
branches. The irregularities are com­ of converging inputs guarantees that the stand above the cortex and look down at
monest near the center of gaze and distribution will be uniform, with nei­ it, what would the pattern be? To make
along the line that maps the horizon. ther eye favored by accident in any one the problem more interesting, suppose
The stripes always seem to be perpen­ place. Why there should be all these the pattern is seen by one eye only. In
dicular to the border between the pri­ grades of eye preference everywhere is view of the architecture just described
mary visual cortex and its neighbor, itself not clear, but our guess is that it the pattern turns out to be not a line but
area 18, and here the regularity is great­ has something to do with stereoscopic merely a set of regularly spaced patches
est. Such general rules seem to apply to depth perception. [see illustration above]. The reasoning can
all macaque brains, although the details be checked directly by exposing a mon­

Given
of the pattern vary from one individual what has been learned about key with one eye closed to a set of verti­
to the next and even from one hemi­ the primary visual cortex, it is clear cal stripes and making a deoxyglucose
sphere to the other in the same monkey. that one can consider an elementary autoradiograph. The resulting pattern
The width of a set of two stripes is piece of cortex to be a block about a should not be a great surprise: it is a set
constant, about . 8 millimeter, over the millimeter square and two millimeters of regularly spaced patches, which sim-

16 1

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
ply represents the intersection of the two ed out above, the basic topography is a For most of these physiologically de­
sets of column systems. Imagine the sur­ map of the opposite half of the body, fined systems (except the visual ones)
prise and bewilderment of a little green but superimposed on that there is a two­ there are so far no anatomical corre­
man looking at such a vcrs ion of the fold system of subd ivisions, with some lates. On the other hand, in the past few
outside world! areas where neurons respond to the years several anatomists, notably Ed­
Why evolution has gone to the trouble movement of the joints or pressure on ward G. J ones of the Washington Uni­
of designing such an elaborate architec­ the skin and other areas where they re ­ versity School of Medicine and Nauta
ture is a que stion that continues to fasci­ spond to touch or the bend ing of hairs. and Patricia Goldman at M.I.T., have
nate us. Perhaps the most plausible no­ As in the case of the visual columns, a shown that connections from one region
tion is that the column systems are a complete set here (one area for each of the cortex to another (for example
solution to the problem of portray ing kind of neuron) occupies a distance of from the somatic sensory area on one
more than two dimensions on a two-di­ about a millimeter. These subdivisions side to the corresponding area on the
mensional surface. The cortex is dealing are analogous to ocular-dominance col­ other side) te r minate in patches that
with at least four sets of values: two for umns in that they are determined in the have a regular periodicity of about a
the x and y position variables in the visu­ first instance by inputs to the cortex millimeter. Here the columns are evi­
al field, one for orientation and one for (from either the left or the right eye and dent morphologically, but one has no
the different degrees of cye preference. from either deep receptors or receptors idea of the physiological interpretation.
The two surface coordinates are used up in the upper skin layers) rather than by It is clear, however, that fine periodic
in designating field position; the other connections within the cortex, such as subdivisions are a very general feature
two variables are accommodated by those that determine orientation selec­ of the cerebral cortex. Indeed, Mount­
dicing up the cortex with subd ivisions so tivity and the associated system of ori­ castle's original observation of that fea­
fine that one can run through a complete entation regions. ture may be said to supply a fourth pro­
set of orientations or eye preferences The columnar subdivisions associated found insight into cortical organization.
and meanwhile have a shift in visual­ with the visual and somatic sensory sys­
field position that is small with respect tems are the best-understood ones, but t would surely be wrong to assume
to the resolution in that part of the visu­ there are indications of similar vertical I that this account of the visual cortex
al world. subd ivisions in some other areas: several in any way exhausts the subject. Color,
The strategy of subdivid ing the cortex higher visual areas, sensory parietal re­ movement and stereoscopic depth are
with small vertical partitions is certainly gions recently studied by M ountcastle probably all dealt with in the cortex, but
not limited to the primary visual area. and the auditory region, where Thomas to what extent or how is still not clear.
Such subd ivisions were first seen in the J. Imig, H. O. Adrian and John F. There are indications from work we and
somatic sensory area by Vernon B. Brugge of the University of Wisconsin others have done on depth and from
M ountcastle of the Johns Hopkins Uni­ Medical School and their colleagues work on color by Semir Zeki of Uni­
versity School of Medicine about 10 have found subdivisions in which the versity College London that higher cor­
years before our work in the visual area. two ears seem alternately to add their tical visual areas to which the primary
In the somatic sensory area, as we point- information or to compete. area projects directly or indirectly may
be specialized to handle these variables,
but we are a long way from knowing
what the handling involves.
What happens beyond the primary vi­
sual area, and how is the information on
orientation exploited at later stages? Is
one to imagine ultimately find ing a cell
that responds specifically to some very
particular item? (Usually one's grand­
mother is selected as the particular item,
for reasons that escape us.) Our answer
is that we doubt there is such a cell, but
we have no good alternative to offer. To
speculate broadly on how the brain may
work is fortunately not the only course
open to investigators. To explore the
brain is more fun and seems to be more
profitable.
There was a time, not so long ago,
when one looked at the millions of neu­
rons in the various layers of the cortex
and wondered if anyone would ever
have any idea of their function. Did they
all work in parallel, like the cells of the
liver or the kidney, achieving their ob­
jectives by pure bulk, or were they each
doing something special? For the visu­
al cortex the answer seems now to be
known in broad outline: Particular stim­
uli turn neurons on or off; groups of neu­
rons do indeed perform particular trans­
formations. It seems reasonable to think
ACTUAL PATTERN of cortical activity was elicited by exposing only the left eye to a set of
that if the secrets of a few regions such
vertical stripes. The deoxyglucose autoradiograph is of a tangential section in the outer layers
as this one can be unlocked, other re­
of the cortex. The pattern of regularly spaced dark patches of radioactivity represents inter­ gions will also in time give up their se­
section of ocular-dominance and orientation systems. Magnification is about eight diameters. crets.

162

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms
 This S u m mer
We To u red T h e G a l axies W ith
1 00 , 0 0 0 P eo p l e . S o o n We're G o i n g To
J u p ite r's M o o n s , S aturn , T h e M i l k y Way . . .
I f you've been i m p ressed a l l your - osci l l at i o n s of the u n i verse - M a rs than the expert a stro n o m e rs of 50
life with the stars that p u n c h you - red variables - photo g ra p h i c as- years ago ever dreamed a b o u t .
i n the eye o n a wa rm s u m mer's n i g ht , tro n o m y - Voyager m i s s i o n s - ri n g s B ec o m e a s u bs c ri b e r to ASTR O N ­
you s h o u l d be read i n g ASTR O N O M Y around the p l a n ets - c l i mate c h a n g e s O M Y and join us in our travels
m a g a z i n e . I f yo u ' ve thought yo u c o u l d - s t a r swarms - b l a c k holes - a n d t h ro u g h the unfo l d i n g u n i ve r s e . I n
never a p p re c i ate the " m u s i c of t h e much more. its pages yo u ' l l find new worlds of
s p heres" you o u g ht to t r y ASTR O N ­ T h e r e a re p o p u l a r d e p a rtments wonders p i l i n g i n o n one a n other as
O M Y m a g a z i n e . I f t h e wonder a n d such a s E q u i pment Atl a s , w h i c h your knowledge and a p p reciation
excitement o f a b i l l i o n s p e a rs of d e l v e s i nto a stro n o m i c a l "tools o f of the heavens g rows g reater each
l i g ht s e e m s j u st beyond your u n d er­ the t r a d e " for a mateurs ; G a zers month .
sta n d i n g , you s h o u l d be s u b s c r i b i n g G a zette , a s h o rt c o u rse i n obser­
S pecial O ffer To S c ie ntific
to ASTR O N O M Y . v i n g e a c h m o nth ; S k y A l m a n a c , y o u r
American Readers
M o nth after month, with observa­ g u i d e t o the s k y ; Star O o m e , t o h e l p
If you send c h e c k o r money o rd e r
tory-q u a l ity color photographs, lavi s h you e x p l o re the s k i e s i n deta i l ; Fo­
w i t h y o u r s u bs c r i pti o n , w e ' l l a d d
p a i n t i n g s a n d crystal c l e a r writi n g , ru m , co ntrovers i a l and i ntri g u i n g ;
o n e b o n u s i s s u e to y o u r s u bs c r i p­
ASTR O N O M Y makes t h e adventure A stro-N ews f o r l ate-bre a k i n g item s ; tion - o u r " t h a n k Y o u " for h e l p i n g
of s p a c e as exciting as today's news . A stro-ma rt for free c l a s sified ads ; r e d u c e o u r pa perwo rk.
What was o n c e d ry-a s-dust physics a n d a reg u l a r book review g u i d e for
a n d mathematics beco m e trans­ further rea d i n g . ------ F R E E ------
formed i nto e a s i l y understood ex­ A st ro n o m y has come of a g e . After you s e n d y o u r s u bscri pti on
planations of the wonders a bove us. We've l e a rned m o re about the vast payment, w e ' l l send you F R E E
Today 1 00 , 000 readers have s p a c e s a r o u n d us in t h e l a s t 30 t h i s stu n n i n g f u l l c o l o r A N TA R E S
entered the u n i verse of ASTR O N ­ years t h a n i n all previ o u s hi story. poster. I t ' s a g i a n t 2 2 x34" repro­
O M Y , t h e c o u ntry's l a rgest a stron­ Young people today know more a­ d u ct i o n of J o h n C l a rk ' s v i s i o n of
omy m a g a z i n e . Y o u ' l l find a bsorbi n g bout g a l a x i e s , c l usters , q u a s a r s , d e e p-s p a c e e x p l o rati o n . A $ 3 . 9 5
a rti c l e s a bout g a l a xi e s - r e d dwarfs p l u s a rs a n d mysteri o u s b l a c k h o l e s va lue, it's yo urs free with payment!

1 Year $ 1 3 . 44
2 Years $ 2 2 .44

T H E WOR LD'S MOST

BEAUTI F U L ASTRO N O M Y
MAGAZI N E

P O B o x 92788, D E PT . SAJ4E
M I LW A U K E E , W I 53202

© 1979 SCIENTIFIC AMERICAN, INC

This content downloaded from

128.61.154.29 on Mon, 06 Jan 2020 20:23:07 UTC
All use subject to https://about.jstor.org/terms