adult by definition.) We concluded after many experiments that somewhere between birth and adulthood there must be a period of plasticity during which deprivation produces the cortical deficit. For the cat, this critical period turned out to be between the fourth week and the fourth month. Not surprisingly, closing an eye had little effect prior to the fourth week, because a cat uses its vision hardly at all during the first month of its life: the eyes open only around the tenth day, and for the next few weeks the kittens are behind the sofa with their mother. The susceptibility to deprivation comes on quickly and reaches a maximum in the first few weeks of the critical period. During that time, even a few days of closure result in a palpably distorted ocular-dominance histogram.
Over the ensuing four months or so, the duration of closure required to give obvious effects increases steadily; in other words, the susceptibility to deprivation tapers off.
The histograms summarizing some of the results in monkeys can be seen in the three graphs on the facing page. The left graph shows the severe effects of a six-week closure done at five days; almost no cells could be driven from the eye that was closed. A much briefer early closure (middle graph) also gave severe effects, but clearly not quite as severe as the longer closure. At four months the susceptibility begins to wane, so much so that even a closure of five years' duration, as shown in the righthand graph, although giving pronounced effects, is no match for the results of the earlier closure.
In these studies of the time course of the sensitive period, cats and monkeys gave very similar results. In the monkey, the sensitive period began earlier, at birth rather than at four weeks, and lasted longer, gradually tapering off over the first year instead of at around four months. It reached its peak in the first two weeks of life, during which even a few days of closure was enough to give a marked shift in ocular dominance. Closing the eye of an adult monkey produced no ill effects, regardless of the duration of closure. In one adult monkey, we closed an eye for four years, with no blindness, cortical deficit, or geniculate-cell shrinkage.
We next asked whether any recovery in physiology in a monkey could be obtained by opening the eye that had been closed. The answer was that after a week or more of eye closure, little or no physiological recovery ever occurred if the closed eye was simply opened and nothing else done. Even a few years later, the cortex was about as abnormal as it would have been at the time of reopening the eye, as shown in the left graph on the next page. If at the time of reopening, the other, originally open eye was closed, in a procedure called eye reversal, recovery did occur but only if the reversal was done when the monkey was still in the critical period, as shown in the middle and righthand graphs above for early and late eye reversal. After the critical period, even an eye reversal followed by several years with the second eye closed failed to bring about anything more than slight recovery in the anatomy or physiology.
The monkey's ability to see did not necessarily closely parallel the cortical physiology. Without reversal, the originally closed eye never recovered its sight. With reversal, sight did return and often approached normal levels, but this was so even in late reversals, in which the physiology remained very abnormal in the originally closed eye. We still do not understand this disparity between the lack of substantial physiological or anatomical recovery and what in some cases seemed to be considerable restoration of vision. Perhaps the two sets of tests are measuring different things. We tested the acuity of vision by measuring the ability to discriminate such things as the smallest detectable gap in a line or circle. But testing this type of acuity may yield an incomplete measure of visual function. It seems hard to believe that such florid physiological and anatomical deficits in function and structure would be reflected behaviorally by nothing more than the minor fall in acuity we measured.
THE NATURE OF THE DEFECT
The results I have been describing made it clear that a lack of images on the retina early in life led to profound long-lasting defects in cortical function. These results nevertheless left open two major questions concerning the nature of the underlying process. The first of these was a "nature versus nurture" question: Were we depriving our animals of an experience that they needed in order to build the right connections, or were we destroying or disrupting connections that were already present, prewired and functional at the time the animal was born? The dark-rearing experiments done in the decades prior to our work had practically all been interpreted in the context of learning—or failure to learn. The cerebral cortex, where most people thought (and still think) that memory and mental activity resides, was looked upon in roughly the same way as the 1-megabyte memory board for which we pay so much when we buy our computers: these contain many elements and connections, but no information, until we put it there. In short, people regarded the cortex as a tabuia rasa.
1: One eye was closed at birth for nine days in this monkey and then opened. The recordings were done four years later, during which time the animal had had much testing of its vision. Even that long a period with both eyes open produced little recovery in the physiology.
2: The right eye in this macaque monkey was closed at birth. At five and a half weeks the right eyelids were opened and the left closed. When at six months the recording was made from the right hemisphere, most of the cells strongly favored the right, originally closed eye.
3: In this macaque monkey the right eye was closed at seven days, for one year, at which time the right eye was opened and the left was closed. After another year, the left eye was opened, and both remained open. When finally the recording was made at three and a half years, most cells favored the eye that was originally open. Evidently one year is too late to do a reverse suture.