Now, how to explain the bizarre tendency to attribute touch sensations arising from the face to the phantom hand? The orphaned brain map continues to represent the missing arm and hand in absentia, but it is not receiving any actual touch inputs. It is listening to a dead channel, so to speak, and is hungry for sensory signals. There are two possible explanations for what happens next. The first is that the sensory input flowing from the facial skin to the face map in the brain begins to actively invade the vacated territory corresponding the missing hand. The nerve fibers from the facial skin that normally project to the face cortex sprout thousands of neural tendrils that creep over into the arm map and establish strong, new synapses. As a result of this cross-wiring, touch signals applied to the face not only activate the face map, as they normally do, but also activate the hand map in the cortex, which shouts “hand!” to higher brain areas. The net result is that the patient feels that his phantom hand is being touched every time his face is touched.
A second possibility is that even prior to amputation, the sensory input from the face not only gets sent to the face area but partially encroaches into the hand region, almost as if they are reserve troops ready to be called into action. But these abnormal connections are ordinarily silent; perhaps they are continuously inhibited or damped down by the normal baseline activity from the hand itself. Amputation would then unmask these ordinarily silent synapses so that touching the face activates cells in the hand area of the brain. That in turn causes the patient to experience the sensations as arising from the missing hand.
Independent of which of these two theories—sprouting or unmasking—is correct, there is an important take-home message. Generations of medical students were told that the brain’s trillions of neural connections are laid down in the fetus and during early infancy and that adult brains lose their ability to form new connections. This lack of plasticity—this lack of ability to be reshaped or molded—was often used as an excuse to tell patients why they could expect to recover very little function after a stroke or traumatic brain injury. Our observations flatly contradicted this dogma by showing, for the first time, that even the basic sensory maps in the adult human brain can change over distances of several centimeters. We were then able to use brain-imaging techniques to show directly that our theory was correct: Victor’s brain maps had indeed changed as predicted (Figure 1.3).
FIGURE 1.3 A MEG (magnetoencephalograph) map of the body surface in a right-arm amputee. Hatched area, hand; black areas, face; white areas, upper arm. Notice that the region corresponding to the right hand (hatched area) is missing from the left hemisphere, but this region gets activated by touching the face or upper arm.
Soon after we published, evidence confirming and extending these findings started to come in from many groups. Two Italian researchers, Giovanni Berlucchi and Salvatore Aglioti, found that after amputation of a finger there was a “map” of a single finger draped neatly across the face as expected. In another patient the trigeminal nerve (the sensory nerve supplying the face) was severed and soon a map of the face appeared on the palm: the exact converse of what we had seen. Finally, after amputation of the foot of another patient, sensations from the penis were felt in the phantom foot. (Indeed, the patient claimed that his orgasm spread into his foot and was therefore “much bigger than it used to be.”) This occurs because of another of these odd discontinuities in the brain’s map of the body: The map of the genitals is right next to the map of the foot.
MY SECOND EXPERIMENT on phantom limbs was even simpler. In a nutshell, I created a simple setup using ordinary mirrors to mobilize paralyzed phantom limbs and reduce phantom pain. To understand how this works, I first need to explain why some patients are able to “move” their phantoms but others are not.
Many patients with phantoms have a vivid sense of being able to move their missing limbs. They say things like “It’s waving goodbye” or “It’s reaching out to answer the phone.” Of course, they know perfectly well that their hands aren’t really doing these things—they aren’t delusional, just armless—but subjectively they have a realistic sensation that they are moving the phantom. Where do these feelings come from?
I conjectured that they were coming from the motor command centers in the front of the brain. You might recall from the Introduction how the cerebellum fine-tunes our actions through a servo-loop process. What I didn’t mention is that the parietal lobes also participate in this servo-loop process through essentially the same mechanism. Again briefly: Motor output signals to the muscles are (in effect) CC’ed to the parietal lobes, where they are compared to sensory feedback signals from the muscles, skin, joints, and eyes. If the parietal lobes detect any mismatches between the intended movements and the hand’s actual movements, they make corrective adjustments to the next round of motor signals. You use this servo-guided system all the time. This is what allows you, for instance, to maneuver a heavy juice pitcher into a vacant spot on the breakfast table without spilling or knocking over the surrounding tableware. Now imagine what happens if the arm is amputated. The motor command centers in the front of the brain don’t “know” the arm is gone—they are on autopilot—so they continue to send motor command signals to the missing arm. By the same token, they continue to CC these signals to the parietal lobes. These signals flow into the orphaned, input-hungry hand region of your body-image center in the parietal lobe. These CC’ed signals from motor commands are misinterpreted by the brain as actual movements of the phantom.
Now you may wonder why, if this is true, you don’t experience the same sort of vivid phantom movement when you imagine moving your hand while deliberately holding it still. Here is the explanation I proposed several years ago, which has been since confirmed by brain-imaging studies. When your arm is intact, the sensory feedback from the skin, muscles, and joint sensors in your arm, as well as the visual feedback from your eyes, are all testifying in unison that your arm is not in fact moving. Even though your motor cortex is sending “move” signals to your parietal lobe, the countervailing testimony of the sensory feedback acts as a powerful veto. As a result, you don’t experience the imagined movement as though it were real. If the arm is gone, however, your muscles, skin, joints, and eyes cannot provide this potent reality check. Without the feedback veto, the strongest signal entering your parietal lobe is the motor command to the hand. As a result, you experience actual movement sensations.
Moving phantom limbs is bizarre enough, but it gets even stranger. Many patients with phantom limbs report the exact opposite: Their phantoms are paralyzed. “It’s frozen, Doctor.” “It’s in a block of cement.” For some of these patients the phantom is twisted into an awkward, extremely painful position. “If only I could move it,” a patient once told me, “it might help alleviate the pain.”
When I first saw this, I was baffled. It made no sense. They had lost their limbs, but the sensory-motor connections in their brains were presumably the same as they had been before their amputations. Puzzled, I started examining some of these patients’ charts and quickly found the clue I was looking for. Prior to amputation, many of these patients had had real paralysis of their arm caused by a peripheral nerve injury: the nerve that used to innervate the arm had been ripped out of the spinal cord, like a phone cord being yanked out of its wall jack, by some violent accident. So the arm had lain intact but paralyzed for many months prior to amputation. I started to wonder if perhaps this period of real paralysis could lead to a state of learned paralysis, which I conjectured could come about in the following way.