YOU PROBABLY LOOK in a mirror every day without thinking about it. But mirrors can reveal a great deal about the brain, with implications for psychology, clinical neurology and even philosophy. They can help us explore the way the brain puts together information from different sensory channels such as vision and somatic sensations (touch, muscle and joint sense). In doing so, they can reveal a lot about our sense of self. Would a person who has never looked at his reflection—even in a pool—ever develop a sophisticated self-representation?

Using two bricks, or some duct tape, prop up an 18-inch-square mirror vertically on a table. Sit so that the edge faces you. Now put your left hand on the table at the left side of the mirror (either palm up or down) and match your right-hand position on the right side. If you now look into the right side of the mirror, you will see the right hand’s reflection optically superimposed in the same place where you feel your left hand to be. (You may need to adjust the position of the left hand to achieve this sensation.) It will now look like you are viewing your own left hand, but of course you are not. Now try the following experiments.

While continuing to look in the mirror on the right side and keeping your left hand perfectly still, move your right hand, wiggle its fingers or make a fist. The “left hand” in the mirror will appear to move in perfect synchrony with the right but, paradoxically, feel completely still. The conflict creates a slight jolt; it feels spooky, sometimes mildly uncomfortable. The brain abhors discrepancies.

Now do the opposite; keep the right hand still and move the left hand. The left hand appears still but feels like it is moving. You will feel the same kind of jarring sensation, but it will be less powerful than in the preceding case. The reason for the asymmetry is not clear.

Why the jolt? The answer resides in the right superior and inferior parietal lobules (located above your right ear), where signals from your various senses—visual, somatic—converge to create your internal sense of a body image. Stand up now and close your eyes. Either raise your arms or let them dangle by your side. Obviously you have a vivid sense of being “anchored” in your body except under special circumstances (such as ketamine anesthesia). Now open your eyes, and you have visual confirmation of what your other senses are telling you: you see your hand where you felt it to be. In short, your senses normally blend different sensory inputs to create a vivid dynamic image of your body moving in space and time.

The mirror experiment you did earlier disrupts this consistency of signals in the right superior parietal lobule. The discrepancy is picked up in part by the right insular cortex (buried in the temporal lobe), and that information is then relayed to the right frontal lobe, where it can be picked up through brain imaging (as shown by Richard Frackowiak, Ray Dolan and Chris Frith, all at University College London, and Peter Halligan of Cardiff University in Wales).

Is That My Hand?
You do not need fancy brain-imaging gizmos to try out some additional experiments that can give you insights into brain function.

Return your hands to either side of the mirror. Now have a friend touch, stroke, pinch, tap or rub your right hand while you look at its reflection. Obviously it will look like your own left hand is being stroked, pinched, tapped or rubbed. But because it is not actually being touched, you will experience one or all of the following (the response varies from person to person).

First, the hand may feel numb, anesthetized or asleep, and it will still feel as if it belongs to you. (Your brain is in effect saying, “I see my hand being rubbed but don’t feel it, so it must be asleep.”) This perception is unaffected by your higher-level intellectual knowledge of the optics of the situation. Your perceptual systems integrating vision and touch are on autopilot, as it were, applying their own statistical rules.

Second, you may see the hand as not belonging to you. Your brain is then ignoring its proprioceptive (muscle and joint feedback) congruence with the visual image of your hand. It is as if the brain is concluding that “because I see the touch but don’t feel it, that hand must be someone else’s.” Sometimes you will “see” the hand as a cadaver’s hand or a realistic plastic dummy. Interestingly, the brain does not settle on “halfway” ambiguities—at any given time you clearly experience one of the percepts.

Last, if you are lucky, you will actually feel some tingling touch sensations in the left hand—even though nothing is being done to it. This effect is a striking example of the brain “filling in” the missing information. Two sources of information (proprioception and vision) are internally consistent in telling you that it is your hand. But the third piece of information—that the hand looks like it is being stroked—is inconsistent with lack of touch sensations. So the brain “flags” the discrepancy as tingling—as if to say, “I’m feeling something odd.” Very infrequently, you may actually feel the touch—as though the brain fills in the blanks to create an internally consistent package to higher centers. We call this phenomenon intermanual touch referral.

Clues to Managing Pain
Try the following experiments. Before the stroking begins, look into the mirror and wiggle the fingers of your two hands in perfect synchrony. Nothing odd so far. Now have a friend deliver strokes, taps or pinches as before, but this time to the visible hand only. All of a sudden you start feeling intermanual referral (that is, feeling the actual touch in the hidden hand) much more vividly and less fleetingly than when your hands were stationary. Why?

In constructing a picture of the world, the brain assigns various weights to different sensory inputs based on a lifetime’s experience of their statistical reliability, as well as ongoing patterns of activation. In short, the brain does not average the signals—it looks for improbable internal consistencies.

When you start wiggling the fingers synchronously, the brain suddenly gets extra information that the hand is really yours. These data force your brain to accept the hand as your own, so you lean toward experiencing intermanual referral with or without tingling. The flood of proprioceptive signals coming in from the hidden hand vetoes any attempt by your brain to engage in disownership. So your brain adopts the next available strategy: accept the hand and feel intermanual referral.

The same effect occurs if you wiggle right- and left-hand fingers nonsynchronously. This time the tendency to think of the reflection as your own left hand is slightly mitigated by the incongruity between vision and proprioception. (The sight of wriggling is somewhat desynchronized from the felt position of the fingers.) Consequently, the intermanual referral is halfway between our previous two experiments.

One last experiment you—the reader—can do. Drop some itching powder on the (hidden) left hand so that it begins to itch. Next have the right hand vigorously stroked and scratched while wiggling both hands synchronously (that is, generate intermanual referral). Question: Does the illusory stroking and scratching felt in the left hand relieve the real itch? It worked better on one of us (Ramachandran) than the other (Rogers-Ramachandran), but you should try it on yourself. If it can be replicated on a large number of subjects, it would be the first example of a purely visual input (which creates an illusory touch) relieving a real itch in a normal hand. Write to us ( or

These effects are more than amusing curiosities; they may be clinically useful for treating pain and paralysis in existing limbs as well as phantom ones, as we discovered in the early 1990s.

Consider the curious but tragic pain disorder called complex regional pain syndrome (CRPS). If you suffer a fracture after your finger is jammed in a doorway, pain ensues. Chronic pain results in a reflex immobilization of the hand to prevent further injury and promote healing. In a few days or weeks the tissue swelling and inflammation subside, along with the pain. But in a small percentage of cases, the immobilization turns into permanent paralysis, and the hand becomes progressively more swollen, painful, inflamed and dysfunctional. The pain and paralysis spread upward to involve the entire arm. There is no known treatment.

In a lecture we gave in 1996 at the University of California, San Diego, Decade of the Brain Symposium, we referred to this phenomenon as learned pain. Every time the motor command centers sent a command to move the hand, excruciating pain accompanying the command blocked further movement. In a few unlucky individuals, an unconscious association—or memory link—is established between the initial command itself and pain, so the brain just gives up: learned pain. Speaking metaphorically, the hand becomes immobilized by fear; it is paralyzed. Admittedly, a hand-wave argument, but nonetheless it is about as compelling an example of mind-body interactions that you can find in all of clinical medicine.

More than 20 treatments, many of them involving drugs or surgery, have been tried for CRPS. What they all have in common is they do not work. (One technique, sympathetic ganglion block, works to some extent but involves an invasive procedure.)

Can the pain be “unlearned”? Prompted by our successful pain-relief treatment using mirrors for patients with phantom limbs, Candy McCabe, now at the University of the West of England, Bristol, and her colleagues tried mirror therapy. The patient looks at the reflection and moves both hands symmetrically so that it appears to the brain that the affected arm—the left, for example—is moving but not painful after all. Similarly, stroking or hitting the right hand creates the optical illusion that the dystrophic hand is being stroked and hit with impunity. Perhaps these two bits of evidence remove the “block” on the affected arm leading to a positive cycle of pain reduction, accompanied by a reduction of swelling and redness.

Taken collectively, these were the first demonstrations that “real” chronic pain can be reduced by visual input; indeed, even intense visual imagery may turn out to be partially effective, but this is hard to do. We first tried mirror therapy on patients with phantom pain from amputated limbs. Sometimes the missing hand feels “locked” in a painfully awkward cramp that can be excruciating, and the patient cannot volitionally move the phantom. When he looks at the reflection, a series of things may happen. First, he “sees” his phantom and recognizes that it is not being poked or held in a vice after all; there is no reason for it to be painful. Second, merely seeing the phantom may be beneficial because the brain can attribute the pain to the arm and, paradoxically, a pain whose source is known may be less troubling than “disembodied,” inexplicable pain (caused by discordant visual and proprioceptive signals). Third, seeing the cramped, paralyzed hand move seems to animate it in such a way as to relieve the cramp, an example of successful clinical application of visual capture. Repeated use may lead to an unlearning of learned paralysis. In placebo-controlled clinical trials on returning war veterans, mirror visualization feedback has since been found to be strikingly successful in some patients and moderately so in others. (Jack Tsao and his colleagues at Walter Reed Army Medical Center conducted the trials.)

Remarkably, in controlled clinical trials, we and others have found mirror therapy to relieve paralysis from cerebrovascular stroke. This relief may be partly because the paralysis could be learned and partly because many paralyzed limbs also have a form of CRPS associated with them. Both these effects contribute to the limb paralysis, which would explain the relief provided by the mirrors.

Litmus Test for Self-Awareness
Let us return to normal perception again and describe an observation we made in collaboration with Eric Altschuler of UMDNJ–New Jersey Medical School.

Have a friend sit behind an ordinary writing desk. In front of the desk, place a mirror so that it covers it completely and you can see only your friend’s torso behind the desk. Now stand at a distance of 20 feet from the desk, look at her and carefully align her torso with the reflection of your lower trunk and feet. Now walk toward the desk, and you will see your friend “walking toward you” with her feet moving in perfect synchrony with your own. If you are among the lucky 75 percent of subjects, you will have a spooky sensation of an out-of-body experience with “you” out there inhabiting your friend’s body, presumably because this is the only way your brain can interpret the perfect synchrony of her legs and yours. Try having her move her face a bit. Does that enhance or diminish the effect?

You may ask why this effect does not occur when you simply walk toward the mirror looking at your own reflection. The answer is twofold. First, can you really be sure it does not? When you shave or put on makeup do you not, at least to a limited extent, “project” yourself into the mirror? Perception is a multilayered phenomenon—hence, it is prone to endless paradoxes in contrived situations. Second, given your lifelong experience with mirrors, you have become habituated; just as horses are not normally scared of their own shadows. A feral child (or man) seeing himself in a mirror the very first time might indeed experience himself inhabiting the stranger in the mirror.

Finally, Gordon Gallup, Jr., now at the University at Albany, has suggested that mirrors can provide a litmus test for self-awareness, a topic that has been much discussed by philosophers for two millennia. When a chimp is asleep, dab a splotch of paint on its forehead. If you show it a mirror when it wakes up, it will spontaneously reach for its forehead to remove the splotch; it does not reach into the mirror. This response may or may not tell us that the chimp is self-conscious, but it does show that the chimp knows that it is looking at itself in the mirror and that it is looking at a reflection—a capacity that eludes monkeys. They fail the test.

We saw a 70-year-old neurological patient recently who, despite her progressive Alzheimer’s-type dementia, remained fairly intelligent and articulate. Her main presenting symptom, disturbing to her family members, was that she was terrified of seeing her own reflection: mirror phobia. She kept referring to it as a malevolent phantom twin who was following her. So all reflecting surfaces in her house had to be covered. Yet when we did the Gallup mirror splotch test on her, she passed, reflexively removing the splotch. This experience shows that merely passing the test does not indicate that you (whether you are a person or a chimp) are aware at a conscious level (“believe”) that what is in the mirror is really “you.”

Thus, mirrors have vast implications, whether for demonstrating the role of visual feedback in treating pain and paralysis or for the psychological and philosophical issues surrounding construction of body image and sense of self by your brain. There’s plenty to reflect on.