The Mind and the Brain (19 page)

But something else was lurking in the back of Taub’s mind. While doing the usual literature search for previous experiments in preparation for his own, Taub happened on a 1909 book by one H. Munk called
Ueber die Functionen von Him und Rückenmark
(On the functions of the brain and spinal cord). In it, the German scientist recounted how he, too, had performed unilateral deafferentation experiments on monkeys. Yet his results differed dramatically from Sherrington’s of fourteen years earlier. Munk claimed that it was possible to induce a hungry monkey to lift food to its mouth with the unfeeling arm under two conditions: if the intact arm were
restrained, and if the initial halting attempts by the deafferented arm were immediately rewarded. Until Taub stumbled on them, Munk’s observations had been essentially ignored and, in effect, lost to science.

In 1957 Taub and two colleagues at the Brooklyn hospital began a series of experiments designed to test Sherrington’s theory that sensation is necessary for intentional, goal-directed movement. They confirmed that monkeys whose arm had been deafferented failed to use it if given any say in the matter. But Taub suspected that the deafferented animals retained what he called “a latent capacity for purposive movement.” The experiments he designed were meant to “force the expression” of that capacity; in other words, the monkeys would be induced to use their “useless” limb. What was required, Taub suspected, were three things: motivating the monkeys, keeping the motor tasks simple, and repeating the motor trials.

Taub provided the motivation in the very first set of experiments, conducted two weeks after deafferentation surgery. The monkey sat, immobilized, in a “restraining chair.” He heard a tone. If he flexed his deafferented right arm so that it broke a light beam five inches above a waist-high board within 3.5 seconds of hearing the tone, nothing else happened. But if he did not move his senseless arm, usually by flexing the elbow and shoulder, to interrupt the beam of light, he experienced an intense electric shock that lasted up to 3.5 seconds. This type of behavioral conditioning is called
avoidance conditioning
; it uses what B.F. Skinner (the psychologist who first systematically described it) termed “primary negative reinforcers,” such as loud noise or electric shocks, to teach an organism new patterns of behavioral response. As Skinner put it, “A negative reinforcer strengthens any behavior that reduces or terminates it.” Conditioning of this sort was widely used in animal (and even human) research in the 1960s and 1970s, the period of Skinner’s greatest influence.

In the training paradigm Taub used, the conditioned stimulus
was the buzzer. Because it was paired with a primary negative reinforcer (the shock), it became, in Skinnerian lingo, “a conditioned negative reinforcer.” The conditioned response or “operant behavior” to be performed to avoid the shock was flexing the deafferented arm. Each training session lasted twenty trials. The entire training series, Taub said, “was to be of long duration, if necessary.” In fact, it would typically require more than nine weeks of testing, five days a week, spent conditioning and conditioning the animal through electric shocks. Within a few weeks Taub had collected results that threatened a key finding of the Sherringtonian canon. “What do you know: the monkeys with the deafferented arms could learn new conditioned responses,” he recalls more than thirty years later. To avoid the shock, the monkeys could move their deafferented arms.

Taub had another way to motivate the monkeys. In the second set of experiments, he and colleagues put six monkeys in straitjackets to restrain their good arm. This made the monkeys very upset—they either struggled obsessively to wriggle out of the constraint or refused to move at all. Nevertheless, five of the six animals got the point: they figured out that if they wanted to reach the food placed outside their cage, they would have to extend the deafferented arm. Two of the six also used their bad arm to support themselves, to move around “in the limited confines of their cages,” as Taub put it, and even to feed themselves a peanut. “If you use simple restraint of the unaffected limb, within a couple of hours the monkeys will begin using the affected limb,” Taub says, looking back on his results. “The use of that limb is clumsy, but it is permanent. This demonstrated unequivocally that the Sherringtonian reflexology position was incorrect.” Clearly, volitional movement did not require sensory feedback.

“I couldn’t imagine the incredible response when we began publishing in the early 1960s,” Taub recalls. His discovery that sensory feedback is not necessary for movement contradicted the position of a professor of his, one of the Columbia psychology department’s
leading researchers. Taub was about to experience the first, but far from the last, consequence of bucking the conventional scientific wisdom. When it came time for him to defend his proposal for his doctoral thesis, the professor whose views his work undercut showed up at an event usually attended only by the student’s thesis committee. “He was very angry,” Taub recalls. After debating Sherringtonian reflexology with Taub, the professor stormed out. Soon after, Taub learned that he had failed a course (taught by the same professor) required for his Ph.D. because he had not taken the final. He had expected to receive an incomplete, which could usually be converted into a grade as soon as a student took the exam. Instead, the professor told Taub he had failed—because of his “insolence.”

In 1962 Taub transferred to New York University. He continued his deafferentation work, under an NIH grant to the senior scientist in the group, A. J. Berman. Within a few years they found that monkeys could flex the fingers of a deafferented arm if properly motivated. To provide the motivation, Taub strapped a monkey into a restraining chair and taped a fluid-filled plastic cylinder into its bad hand. If the monkey squeezed the cylinder within a given amount of time, nothing happened to him; if he failed to do so, he received an electric shock. Just as the monkeys conditioned to break a light beam after hearing a tone did, these monkeys, too, learned (eventually) to grasp the container in order to avoid an electric shock, Taub reported in 1966.

By 1968 he had concluded that a monkey will not use a deafferented arm if it can get along reasonably well with the other three limbs, especially since trying to use an arm without feeling can lead to uncoordinated movement, falling, and dropping food. The monkey succumbs to what Taub dubbed
learned nonuse
. But conditioning (getting an electric shock if it doesn’t use the bad arm) and restraint (leaving the monkey no choice if it wants to eat or walk but to use the bad arm) force it to use the arm or to be subjected to electric shock or go hungry. The motivation to use the limb is increased, so the monkey uses it. “With specific training,” Taub
concluded, “deafferented monkeys can learn to perform almost any sequence of movements of which a normal monkey is capable, except the most precise.”

Something else was happening. Animals with two deafferented arms behaved quite differently from those deafferented of a single arm: the bilaterally deafferented monkeys were able, soon after surgery, to use both arms to grasp, walk, and climb, Taub and Berman reported in 1968. This counterintuitive result—the more crippling the surgery, the better the monkey did—was, Taub decided, “one of the central enigmas” in the field. Somehow, a smaller lesion (single limb deafferentation) was producing worse crippling than a lesion twice the size. When a single limb is deafferented, the monkeys don’t use the arm; when both limbs are deafferented, the monkeys exhibit almost normal movement. This, Taub decided, was “a paradoxical inversion of results.”

Twenty years later, Taub still professes shock at the reaction his results elicited.

In 1970 Taub finished his Ph.D. thesis in experimental psychology, “Prism Adaptation and Intermanual Transfer: An Application of a Learning Theory of Compensation for Sensory Rearrangement.” (The thesis described how monkeys whose view of the world was skewed as a result of looking through a prism could still direct a deafferented arm to a target, even though the prism made objects appear displaced.) In the fall of 1969 he had been offered a research job at the Institute for Behavioral Research in Silver Spring. There Taub decided to continue his deafferentation experiments, focusing on which movements would or would not be impaired after sensory deprivation, and under what conditions.

At IBR, Taub carried out numerous variations on the deafferentation experiments that would change neuroscience. From the first, he observed that immediately after surgery a deafferented monkey gets by with only its good arm. That made sense, for recovering function in a deafferented limb requires time. Unless the experimenter intervenes to induce use, “the monkeys never learn that, several months after the operation, the limb has become potentially useful,” Taub explained. Putting the good arm of unilaterally deafferented monkeys in a straitjacket for nine full weeks after the surgery, he found, overcame this learned nonuse. The monkeys continue to use the deafferented limb after being freed from the restraint. All the animal needed was some motivation to prevent him from coddling his deafferented arm. Depriving mon
keys of the use of their one good arm, or subjecting them to electric shocks until they learned to use the bad arm, did the trick. “Everything we did,” Taub said more than twenty years later, “was to answer an important question.” Eventually, monkeys could use their bad arm to climb to the top of an eight-foot bank of cages, as well as climb laterally and pick up raisins.

One problem was starting to crop up, however. “Deafferented monkeys have a tendency to sustain severe damage to their affected extremities, frequently as the result of self-mutilation,” Taub reported in 1977. “This tendency toward injury, self-inflicted or otherwise, constitutes one of the major difficulties in carrying out deafferentation experiments with monkeys.” “Difficulties,” he would find out, was putting it mildly.

So far, all of Taub’s experiments had been carried out with adolescent monkeys, who of course had been using their limbs for years before Taub severed their sensory nerves. That raised a key question: was somatic sensation early in life necessary for developing normal coordination? How early could an animal lose sensation to a limb and still learn, or relearn, to use it? Were certain movements hard-wired into the brain, or did they require the sensory feedback that Sherrington posited? If the latter, how much sensory feedback was necessary—days, or weeks, of movement in utero? To answer this, in the early 1970s Taub began to deafferent monkeys on the day of their birth. Amazingly, their ability to walk, climb, and reach was as good at three months as that of monkeys that had not been operated on.

The next logical step was to deafferent fetal monkeys. In 1975, Taub and colleagues at IBR performed the delicate surgery outside the womb, with the fetuses in a warm saline solution bath, and then returned the tiny things to the uterus. Unfortunately, the experiment had a high mortality rate: in one batch, six out of eleven fetuses died. Not even the survivors fared well. All were quadri-paretic; the best-off could stand for a few moments, but do little else. Their impairment worsened; one died; and the rest were “sac
rificed” at five to twelve months of age. But the reason for their immobility, Taub learned when he autopsied the animals, was that the surgery had made the fetuses’ vertebrae flare in such a way as to damage the spinal cord. After using a different surgical procedure to deafferent fetal monkeys, Taub found that (in the two monkeys that survived) the deafferented arm was impaired but not useless: the monkeys supported themselves, walked, and reached with it. Volitional movement, Taub concluded, was not dependent on sensory feedback; rather, it was preloaded into an animal’s brain like Windows XP on a laptop.

In the late 1970s, Taub began an experiment designed to test his hypothesis of learned nonuse directly: he restrained the deafferented arm of his monkeys for three months after surgery. This way, a monkey would never learn that the limb was useless during this period; it would simply assume that the restraint was holding it hostage. Taub also restrained the animal’s good arm, so the monkey would not become adept at living one-handed, carrying out its daily activities with a single forelimb. So for three straight months, monkeys were strapped into straitjackets, with their arms crossed over their chest, pinned to the side of their body, or tied behind the back. The limb position was changed every other day.

Almost as soon as the restraints came off, the monkeys managed to use the deafferented limb.

Taub had demonstrated the phenomenon, and the power, of learned nonuse. In the immediate aftermath of an injury, the animal learns to avoid using the affected extremity, because doing so gives the monkey only “negative feedback” when it attempts to walk, climb, or grasp with the arm. That is, those attempted movements are clumsy or ineffectual. At the same time, the monkey learns compensating moves that are, in stark contrast, successful and rewarding: it works around the injury, much as stroke patients with a useless arm learn to do. The combination of negative feedback if he tries to use the affected extremity and reward if he devel
ops compensatory moves suppresses use of the affected extremity. Although the condition is normally permanent, Taub had seen hints in the Silver Spring monkeys that learned nonuse of the affected limb could be reversed by restraining the intact limb so the monkey was compelled to use the deafferented arm. “The simplest method for evoking purposive use of a single deafferented limb is prolonged impersonal restraint of the intact limb,” Taub concluded. Failure to use a deafferented limb reflected learned helplessness, not a motor incapacity. It was 1980.