Eye, Brain, and Vision

this page, allowed us to see the difference at a glance. Clearly something had
gone wrong, with a vengeance.
We soon repeated the experiment in more kittens and in baby monkeys. In
kittens, a larger series soon showed that if an eye is closed at birth, on the
average only 15 percent of cells prefer the eye that was closed, instead of about
50 percent. The same results were found in monkeys (see the bottom histo-
gram). Of the few cells that did respond through the eye that had been closed,
many seemed abnormal; they fired sluggishly, fatigued easily, and lacked the
normal precise orientation tuning.
A result like this raises many questions. Where in the visual path was the
abnormality? In the eye? The cortex? Could the cat see with the eye that had
been closed, despite the cortical abnormality? Was it light or form deprivation
that produced the abnormality? Was the age at which we closed the eye impor-
tant? Was the abnormality a result of disuse or of something else? To get
answers to such questions took a long time, but we can state the results in a
few words.
The obvious way to determine the site of the abnormality was to record at
lower levels, starting, say, in the eye or the geniculate. The results were une-
quivocal: both the eye and the geniculate had plenty of cells whose responses
were virtually normal. Cells in the geniculate layers that received input from
the eye that had been closed had the usual center-surround receptive fields;
they responded well to a small spot and poorly to diffuse light. The only hint
of abnormality was a slight sluggishness in the responses of these cells, com-
pared with the responses of cells in the layers fed by the normal eye.
Given this relative normality, we were amazed when we first saw the Nissl-
stained lateral geniculate under the microscope. It was so abnormal that a
microscope was hardly needed. The cat's geniculate has a somewhat more
simple organization than the monkey's; it consists mainly of two large-cell
layers, which are on top rather than on the bottom, as in the monkey. The
upper layer receives input from the contralateral eye, the lower from the ipsi-
lateral. Beneath these layers is a rather poorly defined small-cell layer with
several subdivisions, which I will ignore here. On each side, the large-cell
layer receiving input from the closed eye was pale and clearly thinner than its
companion, which looked robust and perfectly normal. The cells in the abnor-
mal layers were not only pale but were shriveled to about two-thirds their
normal cross-sectional area. This result for a right-eye closure is shown in the photo-
graphs on the next page. Similar results were found in the macaque monkey
k for a right-eye closure, as shown in the photograph below. We thus faced
a paradox that took us a few years to explain: a lateral geniculate whose cells
seemed relatively normal physiologically but were manifestly pathological his-
tologically. Our original question was in any case answered, since cortical
cells, although virtually unresponsive to the closed eye, were evidently receiv-
ing a substantial and, on the face of it, practically normal geniculate input. This
seemed to exonerate both the eye and the geniculate as primary sites of the
damage and placed the main abnormality in the cortex. When we looked at the
cortex histologically, we saw absolutely nothing to suggest any abnormality.
As we will see, the cortex does show anatomical defects, but they do not show
up with these staining methods.
We next asked what it was about the eye closures that produced the abnor-
mality. Closing the eye reduces the light reaching the retina by a factor of
about ten to fifty; of course, it also prevents any images from reaching the
retina. Could it be simply the reduction in light that was causing the trouble?
To help decide, we inserted in one eye of a newborn kitten an opalescent
contact lens made of a plastic with the consistency of a ping pong ball. In some
animals we instead surgically sewed over one eye a thin, translucent, opales-
cent membrane, in effect, an extra eyelid called the nictitating membrane that
cats possess and we don't. The plastic or the membrane reduced the light by
about one-half but prevented the formation of any focused images. The results
were the same: an abnormal cortical physiology; an abnormal geniculate his-
tology. Evidently it was the form deprivation rather than light deprivation
that was doing the damage.
In a few kittens we tested vision before recording by putting an opaque
black contact lens over the eye that had not been closed and then observing
how the animal made out. The animals were clearly blind in the eye that had
been deprived: they would walk confidently over to the edge of a low table, go
past the edge, and fall to a mattress placed on the floor. On the floor they
would walk into table legs. These are things no normal, self-respecting cat
ever does. Similar tests with the eye that had not been closed showed that
vision was entirely normal.
Next we did a protracted study in both cats and monkeys to learn whether
the age at which the closures were done and the duration of the closures were
important. It soon became clear that age of onset of deprivation was critical.
An adult cat deprived of vision in one eye for over a year developed no blind-
ness in that eye, no loss of responses in the cortex, and no geniculate pathol-
ogy. (The first cat we deprived, the mother of our first litter of kittens, was an