During the preamputation period, every time the motor cortex sent a movement command to the arm, the sensory cortex in the parietal lobe would receive negative feedback from the muscles, skin, joints, and eyes. The entire feedback loop had gone dead. Now, it is well established that experience modifies the brain by strengthening or weakening the synapses that link neurons together. This modification process is known as learning. When patterns are constantly reinforced—when the brain sees that event B invariably follows event A, for instance—the synapses between the neurons that represent A and the neurons that represent B are strengthened. On the other hand, if A and B stop having any apparent relationship to each other, the neurons that represent A and B will shut down their mutual connections to reflect this new reality.
So here we have a situation where the motor cortex was continually sending out movement commands to the arm, which the parietal lobe continually saw as having absolutely zero muscular or sensory effect. The synapses that used to support the strong correlations between motor commands and the sensory feedback they should generate were shown to be liars. Every new, impotent motor signal reinforced this trend, so the synapses grew weaker and weaker and eventually became moribund. In other words, the paralysis was learned by the brain, stamped into the circuitry where the patient’s body image was constructed. Later, when the arm was amputated, the learned paralysis got carried over into the phantom so the phantom felt paralyzed.
How could one test such an outlandish theory? I hit on the idea of constructing a mirror box (Figure 1.4). I placed an upright mirror in the center of a cardboard box whose top and front had been removed. If you stood in front of the box, held your hands on either side of the mirror and looked down at them from an angle, you would see the reflection of one hand precisely superimposed on the felt location of your other hand. In other words, you would get the vivid but false impression that you were looking at both of your hands; in fact, you would only be looking at one actual hand and one reflection of a hand.
If you have two normal, intact hands, it can be entertaining to play around with this illusion in the mirror box. For example, you can move your hands synchronously and symmetrically for a few moments—pretending to conduct an orchestra works well—and then suddenly move them in different ways. Even though you know it’s an illusion, a jolt of mild surprise invariably shoots through your mind when you do this. The surprise comes from the sudden mismatch between two streams of feedback: The skin-and-muscle feedback you get from the hand behind the mirror says one thing, but the visual feedback you get from the reflected hand—which your parietal lobe had become convinced is the hidden hand itself—reports some other movement.
FIGURE 1.4 The mirror arrangement for animating the phantom limb. The patient “puts” his paralyzed and painful phantom left arm behind the mirror and his intact right hand in front of the mirror. If he then views the mirror reflection of the right hand by looking into the right side of the mirror, he gets the illusion that the phantom has been resurrected. Moving the real hand causes the phantom to appear to move, and it then feels like it is moving—sometimes for the first time in years. In many patients this exercise relieves the phantom cramp and associated pain. In clinical trials, mirror visual feedback has also been shown to be more effective than conventional treatments for chronic regional pain syndrome and paralysis resulting from stroke.
Now let’s look at what this mirror-box setup does for a person with a paralyzed phantom limb. The first patient we tried this on, Jimmie, had an intact right arm, phantom left arm. His phantom jutted like a mannequin’s resin-cast forearm out of his stump. Far worse, it was also subject to painful cramping that his doctors could do nothing about. I showed him the mirror box and explained to him this might seem like a slightly off-the-wall thing we were about to try, with no guarantee that it would have any effect, but he was cheerfully willing to give it a try. He held out his paralyzed phantom on the left side of the mirror, looked into the right side of the box and carefully positioned his right hand so that its image was congruent with (superimposed on) the felt position of the phantom. This immediately gave him the startling visual impression that the phantom had been resurrected. I then asked him to perform mirror-symmetric movements of both arms and hands while he continued looking into the mirror. He cried out, “It’s like it’s plugged back in!” Now he not only had a vivid impression that the phantom was obeying his commands, but to his amazement, it began to relieve his painful phantom spasms for the first time in years. It was as though the mirror visual feedback (MVF) had allowed his brain to “unlearn” the learned paralysis.
Even more remarkably, when one of our patients, Ron, took the mirror box home and played around with it for three weeks in his spare time, his phantom limb vanished completely, along with the pain. All of us were shocked. A simple mirror box had exorcised a phantom. How? No one has proven the mechanism yet, but here is how I suspect it works. When faced with such a welter of conflicting sensory inputs—no joint or muscle feedback, impotent copies of motor-command signals, and now discrepant visual feedback thrown in via the mirror box—the brain just gives up and says, in effect, “To hell with it; there is no arm.” The brain resorts to denial. I often tell my medical colleagues that this is the first case in the history of medicine of a successful amputation of a phantom limb. When I first observed this disappearance of the phantom using MVF, I myself didn’t quite believe it. The notion that you could amputate a phantom with a mirror seemed outlandish, but it has now been replicated by other groups of researchers, especially Herta Flor, a neuroscientist at the University of Heidelberg. The reduction of phantom pain has also been confirmed by Jack Tsao’s group at the at the Walter Reed Army Medical Center in Maryland. They conducted a placebo-controlled clinical study on 24 patients (including 16 placebo controls). The phantom pain vanished after just three weeks in the 8 patients using the mirror, whereas none of the control patients (who used Plexiglas and visual imagery instead of mirrors) showed any improvement. Moreover, when the control patients were switched over to the mirror, they showed the same substantial pain reduction as the original experimental group.
More important, MVF is now being used for accelerating recovery from paralysis following stroke. My postdoctoral colleague Eric Altschuler and I first reported this in The Lancet in 1998, but our sample size was small—just 9 patients. A German group led by Christian Dohle has recently tried the technique on 50 stroke patients in a triple-blind controlled study, and shown that a majority of them regained both sensory and motor functions. Given that one in six people will suffer from a stroke, this is an important discovery.
More clinical applications for MVF continue to emerge. One pertains to a curious pain disorder with an equally curious name—complex regional pain syndrome–Type II (CRPS-II)—which is simply a verbal smoke screen for “Sounds awful! I have no idea what it is.” Whatever you call it, this affliction is actually quite common: It manifests in about 10 percent of stroke victims. The better-known variant of the disorder occurs after a minor injury such as an ordinarily innocuous hairline fracture in one of the metacarpals (hand bones). There is initially pain, of course, as one would expect to accompany a broken hand. Ordinarily the pain gradually goes away as the bone heals. But in an unfortunate subset of patients this doesn’t happen. They end up with chronic, excruciating pain that is unrelenting and persists indefinitely long after the original wound has healed. There is no cure—or at least, that’s what I had been taught in medical school.