Not What, But Where, Is Your “Self”?

Psychiatrists are familiar with patients who are convinced that they are surrounded by impostors who secretly replaced a spouse, a child, or someone else close to them. In a different but related disorder, a patient misidentified herself in the mirror, believing the image to be that of another person with whom she could chat. This woman was not psychotic, delusional, nor mentally unbalanced. She simply could not be persuaded, by any means, that the face she saw in the mirror was her own.* In this syndrome and others like it, people seem to suffer a partial loss of self or, at the very least, a strange transformation in their relationship to themselves. The patient continues to exist, to think, and to relate to others, but something that we call the “self” has begun to slip away.¹

To understand and help such patients, we must ask: What is the self? Is it real in the sense that a rock or a tree (or a brain) is real? That has been the subject of philosophical, religious, and scientific debates for centuries, even millennia. The 18th-century German philosopher Emmanuel Kant was a firm believer in the existence of the self; indeed, he placed a primordial and unified ego at the center of his philosophical beliefs.On the other hand, the giant of 19th-century American psychology and philosophy of mind, William James, argued in The Principles of Psychology that the “Ego” was nothing but “a ‘cheap and nasty’ edition of the soul.”² James did concede that something could be appropriately referred to as the self, but he denied that any single entity at the core of self could be called an “I” or “ego.” He held, instead, that passing states of consciousness constitute the sole reality of mental life and that our experience of ourselves as unified beings arises because we experience successive mental states, uniquely our own, in our stream of consciousness. In 1882, James wrote that “Metaphysics or theology may prove the Soul to exist; but for psychology the hypothesis of such a substantial principle of unity is superfluous.”³

The controversy over the existence of the self has persisted to our day. The contemporary philosopher Daniel Dennett, in his book Consciousness Explained, offers this summary:

Since the dawn of modern science in the seventeenth century, there has been nearly unanimous agreement that the self, whatever it is, would be invisible under a microscope, and invisible to introspection, too. For some, this has suggested that the self was a nonphysical soul, a ghost in the machine. For others, it has suggested that the self was nothing at all, a figment of metaphysically fevered imaginations. And for still others, it has suggested only that a self was in one way or another a sort of abstraction, something whose existence was not in the slightest impugned by its invisibility. After all, one might say, a center of gravity is just as invisible—and just as real…The question of whether there really are selves can be made to look ridiculously easy to answer, in either direction…⁴

Why should the simple fact of our existence as selves be so controversial? As pointed out by the French philosopher René Descartes, more than 300 years ago, and reiterated by Dennett, one reason may be that, unlike a brain or a body, the self cannot be visualized. The subject that represents the “I” in the sentences “I see,” “I hear,” “I experience,” “I exist” cannot be weighed, measured, or photographed with an electron microscope. Indeed, it can be directly experienced only by its possessor. This is the problem of the subjectivity of the self.

If we could pinpoint the brain structures critical to the self, we would go far toward proving its existence. Two different but complementary approaches supply tantalizing leads into this question:

• Some patients, like the woman mentioned at the outset of this article, experience dramatic transformations in their personal relatedness to themselves, other persons, and the world. Especially provocative are the clinical characteristics of several fascinating disorders that affect the self. Brain damage or so-called neurologic lesions discovered in these patients provide a window into the self and an opportunity to examine its physical structure.
• Modern techniques of neuroimaging, such as functional magnetic resonance imaging (fMRI), can be used to study the self. Some current experiments are using neuroimaging to explore the neurobiological basis of disorders that affect the self. In particular, it appears that the right (non-dominant) hemisphere plays a critical role in the creation and maintenance of the self. There is reassuring agreement between the neuroimaging findings and observations in patients. The areas of the brain that light up (appear to become active) in the scanner when subjects are somehow involved with the self seem to be the same areas that are damaged in patients that have a deficit of self.

The value of defining the brain locus of the self goes far beyond satisfying age-old theological and philosophical questions. Many disorders of the self, including some discussed here, are the result of demonstrable brain damage, easily visualized on scans of the brain. Other pathologic conditions, however, dramatically affect the self but are poorly understood in terms of brain dysfunction. For example, the neurobiological bases for schizophrenia, childhood autism, and Asperger’s syndrome—all conditions entailing prominent derangement of the self—remain baffling. Better understanding of the neurobiology of the self offers our best chance to achieve insight into these and other conditions in which the self is disturbed.

The Stranger in the Mirror

In 1923, French psychiatrist Joseph Capgras, along with one of his interns, J. Reboul-Lachaux, reported the case of Mme. M, a 53-year-old woman who complained that imposters had replaced her husband, children, and even herself. She explained that her husband was murdered and that at least 80 imposters, disguised to look like him, had appeared. She asserted: “I can assure you that the imposter husband that they are trying to insinuate is my own husband, who has not existed for 10 years, is not the person who is keeping me here.” She also claimed that there were more than 2,000 doubles of her daughter and that policemen, a concierge, doctors, nurses, and neighbors were impersonators as well. Eventually, she even claimed there were doubles of the doubles.

Capgras and Reboul-Lachaux called her disorder the “L’Illusion Des Sosies.” They took the name “Sosies” from the Greek myth in which Zeus, who wishes to seduce the mortal and married Alcmena, physically transforms himself to take on the appearance of her husband, Amphitryon. He also arranges for Mercury to impersonate Sosia, Alcmena’s servant. The word “sosie” took on the meaning of a “a double” in French prior to Capgras’ paper. Today, the disorder is known as Capgras syndrome.

In contrast to Mme. M, most people with Capgras syndrome misidentify only a small number of people, usually relatives or people significant in the patient’s life, or often one particular person. Married patients who develop Capgras syndrome frequently claim that an imposter has replaced the wife or husband, and, if the patient is debilitated, a primary caretaker such a daughter or son may be misidentified. Some patients develop paranoid attitudes toward the “imposter,” but there are instances when patients do not seem particularly upset about the substitution, and some patients actually seem happy about it. In 1931, researchers described a woman with Capgras syndrome who complained about her poorly endowed and sexually inadequate lover. She was happy to report that she had discovered he possessed a double who was rich, virile, handsome, and aristocratic and who competed with the original lover for her attention.

In a variant of the Capgras syndrome, the patient does not misidentify other people, only himself or herself as seen in a mirror. Susan was a woman in her sixties who developed Capgras syndrome for her mirror image. She became hearing impaired in childhood and learned to communicate by both sign language and lip reading. Her son, with whom she lived, became concerned when he observed Susan using sign language in front of the mirror in her bedroom. She told him that she was communicating with the “other” Susan. This other Susan was identical to her in most respects. They were the same age, came from similar backgrounds, and bore a striking physical resemblance. The other Susan even had a son identical in appearance to the patient’s own. However, when her actual son stood behind her and was reflected in the mirror, or when other people did so, she did not misidentify anyone but herself. This was a highly selective self-misidentification, making it clear that Susan does not suffer from a different syndrome known as prosopagnosia. The person with prosopagnosia has a generalized disturbance in face recognition and cannot identify anyone by looking at their face. Prosopagnosia is a specific visual identification disorder, in contrast to Capgras syndrome, which entails a denial or disavowal of identity.

In this narrative by Susan, she describes the other Susan in the mirror:

Well, she’s all right…sounds funny for me talking from one Susan to another, because you know she was a new person to me, and I’m surprised. She was all right, but she’s very nervous, she likes to do her own ways…she never knew that she couldn’t hear so good, and she’s not a very good lip reader. I had to do mostly in sign language for her, to make her understand… she copies every word I say like this, like this motion…she doesn’t even know the sign language very well, and I was confused a little bit, you know, because I wanted her. I thought she knew the sign language very well, so I won’t have to repeat it twice, but then I found out that she’s not that bright. I hate to say that…I don’t want to brag, but she’s a nice person.¹

“Wake Up, Little Susie”

Asomatognosia means literally “lack of recognition of the body.” In the most common form of this disorder, patients who had a recent stroke in the right hemisphere experience profound paralysis and sensory loss of the left arm but deny that the paralyzed left arm belongs to them. Some patients disown the left leg, as well, and some patients deny ownership of a head or a breast. Some asomatognosic beliefs can be explained by an inattention to the left side (called “hemispatial neglect”) or generalized confusion as a result of the diffuse effects of the stroke. But even if the error is repeatedly pointed out to them, some patients with asomatognosia cannot be convinced, and they continue adamantly to deny that the left side belongs to them. Their beliefs are more delusions than simple errors. Indeed, a patient with severe asomatognosia can trace the connection of his paralyzed arm to his own left shoulder with his good right hand, yet still maintain that it is not his own arm.

In some cases, asomatognosia is not apparent until one specifically asks the patient to identify his arm, but in others it is obvious that a patient has the condition. For example, patients have been observed attempting to throw the left arm out of the bed or they may complain to the hospital staff that someone is lying next to them in the bed. One patient we observed complained that her husband’s arm repeatedly fell on her while she was sleeping, so she threw it in the garbage pail.

When shown their left arm, some patients simply plead ignorance and say, simply, “I don’t know what it is” or “How should I know? You’re the doctor!” But one of the most interesting aspects of asomatognosia is the colorful misidentifications and personifications these patients may make of their left arm. Women have a tendency to claim the arm belongs to a man, often the patient’s husband. Men appear more likely to claim the misidentified arm belongs to a woman, most often claiming that the arm belongs to a daughter or mother-in-law.

She later went on to tell me why she had developed this idea about her left arm being little Susie. “It’s a coping mechanism. It’s like laughter is the best medicine. If you can’t laugh what have you got? I thought I could bring her back with some loving kindness. So I sang it, wake up little Susie, which is one of my favorite songs from the Everly Brothers.”¹

The Phantom Child Delusion

The phantom child delusion is a disorder of the self in which self-referential autobiographical material is misattributed to a fictitious child. In a 1956 article, “Delusions about Children Following Brain Injury” neurologist and psychoanalyst Edwin Weinstein, M.D., described patients with brain injuries who expressed the delusion that they were the parent of a fictitious child. One patient was a soldier with a head injury and retrograde amnesia whose wife left him after his injury. Although childless, he claimed he had two children, a boy six and a girl five. He said the girl’s name was “Bobby” which was his estranged wife’s name, but he could not recall the name of the boy.

Weinstein and collaborators observed a unique feature of the delusion: These patients often attributed to the phantom child the same illness or disability that they themselves had. For example, a woman with a pituitary tumor and blindness claimed she had a child who was “sick and blind,” and a 21-year-old soldier with a head injury and bilateral leg weakness claimed he had a three-year-old “daughter” who had paralysis of both legs from polio. In some patients the phantom child embodied personal concerns other than illness. For instance, a woman who felt abused by the nursing staff claimed she had a “baby” that the nurses had “harmed and even killed.”

Over the years, we encountered several patients with this fascinating form of confabulation. One 63-year-old man had a ruptured anterior cerebral artery aneurysm and bilateral damage to his frontal lobes. Neuropsychological testing showed impairments in attention, memory, and executive functioning (the ability to plan ahead, regulate his actions, and reason conceptually). Despite this, he denied any cognitive impairment and, when asked what he was doing at the rehabilitation facility, claimed he was “a guest” with the “optimists club” to “help out.” This man was the biological father of three children, but, in addition, he falsely claimed the existence of a child that he was planning to adopt. Although he denied his own impairments, he asserted that the adopted child “has problems” and complained about the way the doctors were treating “the child”:

I don’t think a kid who is 6 or 7 years old is capable of giving you the right answer…I know this kid has been in the hospital off and on for a couple of years, and they kind of rate them as far as progress goes or things like that…I guess they must rate them when they don’t hear the things they want to hear…like the kid is not accomplishing anything, which I think is very unfair to basically analyze a kid that way.

Another patient, a 65-year-old woman, had surgery for a brain aneurysm that resulted in damage to her frontal lobes on both sides. As a consequence, she suffered from significant memory impairments. She initially denied being ill in any way but later admitted she had an aneurysm; she still denied any neurologic impairments. She was divorced and had no children. When asked why she was in the hospital, she claimed that she was visiting a “niece” whom she described as a “child” and “a little girl” who had an aneurysm. She also insisted that she had an aunt and six cousins, all with “aneurysms on top of their heads.”

Alien Hand Syndrome: “Strangelove Effect” 

Another condition that involves a malfunction of the sense of self is what is known as alien hand syndrome. Here a patient has a hand that performs actions—such as strangling the patient’s own throat—that are beyond the patient’s voluntary control. The celebrated German neurologist Kurt Goldstein, M.D., provided the first description of the condition. In 1908, he described a 57-year-old woman who claimed that her left hand acted as if it had a will of its own. The patient, not knowing what caused this unusual symptom, concluded that her hand must be possessed.

On one occasion the hand grabbed her own neck and tried to throttle her, and could only be pulled off by force. Similarly, it tore off the bed covers against the patient’s will…She soon is complaining about her hand; that it is a law unto itself, an organ without will [willenloses Werkzeug]; when once it has got hold of something, it refuses to let go: “I myself can do nothing with it; if I’m having a drink and it gets hold of the glass, it won’t let go and spills [the drink] out. Then I hit it and say: ‘Behave yourself, hand’ [literally, mein Händchen]” (Smiling,) “I suppose there must be an evil spirit in it.”⁵

In a case we examined, one patient’s left alien hand answered the phone and refused to surrender the receiver to the right hand, dumped a glass of water into a bowl of cereal, tried to steal the remote control from the right hand, grabbed at the patient’s genitals, and tried to strangle the patient while he was asleep. In this and similar cases, the hand performs actions that are seemingly purposeful and voluntary, often aggressive, but the patient claims that the movements are unwanted, involuntary, and not consistent with their conscious intentions.

Most patients with this disorder display what we call “intermanual conflict,” in which the hands perform opposite or opposing actions. For example, the normal hand opens a drawer or closet and the alien hand closes it. Neurosurgeon Joseph Bogen, M.D., reported that one of his patients was seen buttoning up his shirt with one hand while the other hand was “coming along right behind it undoing the buttons.” This patient, like many others with the condition, attempted to restrain the alien hand when it got out of control. This is sometimes called the “Dr. Strangelove effect,” after the character in the movie by the same name, portrayed by Peter Sellers, who felt compelled to make a Nazi salute while his normal hand attempted to restrain the wayward limb.

Although there are various forms of alien hand syndrome, the most clear-cut examples are caused by a pure disconnection of the right and left hemispheres of the brain. For instance, the condition is known to occur in patients who had their corpus callosum—the main white matter pathway connecting the two hemispheres—surgically cut for the control of epileptic seizures. It is striking that in these cases, the alien hand is typically, if not always, on the left side.

Toward a Neuroanatomy of the Self

The patients we described all display a significant disturbance of the self. Whether it involves a misidentification of the reflected image of the self, the ownership of the left arm, the autobiographical history of the self (as in the phantom child syndrome), or the actions of an alien limb, some aspect of the patients’ selves before they developed the condition has gone awry.

What can these disorders teach us about the neurobiology of the self? When we investigated the cause of Susan’s mirror misidentification, neurologic examination revealed evidence of an abnormality of her right hemisphere, and a computed axial tomography (CAT) scan of her brain also demonstrated abnormality within the right hemisphere, a finding typical of patients with Capgras syndrome. In 1998, Karen Spangenberg, Ph.D., and her colleagues reported a similar case of a woman who misidentified herself in the mirror and claimed that there was “a little girl” there. A CAT scan of this patient’s brain also showed more significant pathology of the right hemisphere as compared with the left.⁶

The neuroanatomical underpinnings of asomatognosia are clear: Most patients have damage to the right hemisphere. Indeed, few if any well-documented cases of asomatognosia occur after left-hemisphere strokes, whereas the literature is replete with descriptions of the condition as a result of right-hemisphere pathology. The actual number of reported cases in which the anatomy of the phantom child syndrome is known is small, but all the patients identified have right-hemisphere as well as left-hemisphere lesions. Whether the syndrome can be caused by unilateral lesions of the right or left hemisphere is unknown; the answer must await the examination of additional cases.

We, however, recently reviewed 29 instances of patients with consistent and refractory delusional misidentification of people, places, the left arm after stroke, or the belief in phantom children (some patients had more than one type of misidentification). Of the 29 instances of these conditions, 28 had sustained damage to the right frontal lobe, whereas no other region of the brain within either hemisphere was consistently damaged. This finding further supports the hypothesis that the frontal regions of the right hemisphere are essential in the creation of personal relatedness and the self.⁷

Finally, when the alien hand syndrome is caused by the anatomical separation of the right and left hemispheres, it is always the left hand that is alien to the patient and perceived as disconnected from the patient’s sense of self. This suggests that the input from the right hemisphere to the left hemisphere is essential to the person’s sense of his or her existence as a unified person.

So we have evidence that, in the four clinical disorders of the self we have described (Capgras syndrome, asomatognosia, phantom child syndrome, and the alien hand syndrome), there is either a disruption of right-hemisphere functioning or a disconnection of the right hemisphere from the rest of the brain.

From the Hospital to the Laboratory: Imaging the Self

Although much can be learned about the self and brain by examining afflicted patients, the use of neuroimaging promises to increase significantly our understanding of the self. Only recently have neuroscientists acquired the tools to investigate questions as difficult as the self/brain relationship. It is worth asking why we might want to use techniques such as fMRI instead of relying exclusively on patient information.

The answer is that when investigating brain/behavior relationships, there are some drawbacks to studies involving patients. First, it is difficult to get together enough patients who have these rather unusual disorders, so most patient studies involve few participants. In the case of Capgras for the mirror image, there are fewer than five cases documented in the literature. Therefore, it can be helpful to check our results by using scanning. Second, when it comes to patients, little opportunity exists to replicate results. It is unacceptable to cause brain damage or trauma in healthy subjects just to test our favorite theories or hypotheses; thus, we are at the mercy of chance to provide cases. Finally, we usually lack basic information about most of our patients. It would be beneficial, for example, to have a record of how patients with asomatognosia reacted to their arms before the brain trauma.

Neuroimaging allows us to overcome many of these disadvantages. For example, we can test a large number of subjects under well-controlled circumstances. This is particularly useful when looking at the self when the variables are difficult to measure. By using neuroimaging, we can gain baseline data and manipulate our stimuli and variables cautiously.

The Self in the Scanner

In studies of the self, the face has special significance.⁸ Humans are one of the few species that recognize their own image. Chimpanzees and orangutans also have self-face recognition, but, as far as we know, all other animals, including dogs, cats, birds, and elephants, react to their own face as if it were another animal. Then, too, we humans tend to love to see our own faces, which provide a powerful stimulus in studying the self. For example, we present subjects with a “morph” of their own face, combining half of the self-face with half of the face of a famous person. No matter how many times they see this morph, many research participants still insist that it is their own face.

The brain appears to have a right-hemisphere bias for the self-face, which is processed in a manner different from other faces. One hint of this comes from studies of Capgras syndrome. In all of the patients with this disorder who have focal brain damage, the damage is primarily to the right hemisphere. On the basis of only the handful of patients studied with this disorder, however, can we conclude the right hemisphere is dominant for the self-face? Thanks to neuroimaging, the answer appears to be yes. In the first attempt to use neuro-imaging to locate the self-face in the brain, we used fMRI to scan subjects while they looked at their own faces. We found that the self-face did in fact activate (light up) regions of the right hemisphere (see illustration that follows). To make sure that the right hemisphere’s reaction was indeed to the self-face, we also compared reactions to the self-face and to famous faces, as well as to both familiar and novel faces. Regardless of the comparison, the self-face almost always engaged the right hemisphere. These early results were confirmed by other researchers, including Steven Platek, Ph.D., at Drexel University, who found that the right prefrontal region was particularly important for self-face recognition.⁹

We also used the 50/50 face morphs to help us confirm the regions of the brain that are involved when we see our own face. One study used the Wada test, a procedure in which half of the brain is anesthetized at a time—in effect, disabling or putting to sleep one hemisphere and then the other for about five minutes each. With the subjects under this anesthesia, we presented self-face morphs. What we found confirmed the right-hemisphere hypothesis. As you can see in the illustration on the next page, knocking out the left hemisphere with anesthesia did not disrupt self-face recognition; with only the right hemisphere still awake, subjects saw the morph as their own face. But when the right hemisphere was put to sleep, subjects no longer saw their own face in the morph. This reaction told us that the right hemisphere was indeed important for self-face recognition.

We turned to yet another neuroimaging technique to ascertain where in the brain the self-face might be recognized. Transcranial magnetic stimulation (TMS) is similar to Wada in enabling researchers to disrupt defined areas of the brain for brief periods. The big advantage to TMS, however, is that the regions that get disrupted are extremely small—just one square centimeter. We found that delivering TMS to the right frontal region (the same area identified by Platek) disrupted self-face recognition. We also found evidence that other self-related phenomena might be mediated by the right hemisphere. For example, hearing your own voice or thinking about yourself while you are in a fMRI scanner clearly activates regions of the right hemisphere. These studies confirm what researchers have observed in patients: The right hemisphere is essential for maintaining our selves.

The Brain, the Right Hemisphere, and the Self

Intriguing questions remain: What special functions of the right hemisphere lead to its dominance for self and self-related functions? And how or why did this lateralization of functions evolve in us and the few primates that share our recognition of self?

Some prescient comments of Sigmund Freud can serve as a useful starting point. Writing in 1930, Freud said:

Pathology has made us acquainted with a great number of states in which the boundary lines between the ego and the external world become uncertain or in which they are actually drawn incorrectly. There are cases in which parts of a person’s own body, even portions of his mental life—his perceptions, thoughts and feelings—appear alien to him and as not belonging to his ego; there are other cases in which he ascribes to the external world things that clearly originate in his own ego and that ought to be acknowledged by it. Thus even the feeling of our own ego is subject to disturbances and the boundaries of the ego are not constant.¹⁰

Under the conditions of brain dysfunction that we described, we see both types of disorder described by Freud. Without the mediation of the right frontal lobe, we observed patients who display an inappropriate alienation from items of personal relevance, such as a mirror reflection or a left arm. Conversely, some patients, like those who develop the phantom child syndrome, may project personal items and feelings onto fictitious others in the environment.

What missing functions of the right hemisphere could cause this to occur? One possibility is that right frontal damage results in a disturbance of ego boundaries and ego functions because this region of the brain is essential in establishing the normal relationship between the self and the world. Thus, in these conditions, there is a specific two-way disturbance between the self and the environment with regard to personal relatedness, which could lead to disorders of both under-and over-relatedness to the environment.

Why should the right hemisphere be specialized to do certain tasks while the left hemisphere is specialized to do other tasks? In humans, the left hemisphere is dominant for language functions. Could it be that, as language evolved in the left hemisphere, a concurrent sense of self was evolving in the right hemisphere, and that the two functions interacted to form what now seems like a seamless and integrated sense of self? As the evidence mounts that at least some apes have a sense of self and a similarly lateralized brain, it seems possible that our evolved brain became lateralized by our sense of self as much as from our language abilities. This would make for more efficient operation: As the left hemisphere is thinking verbally, the right hemisphere is thinking in terms of the self.

The muted, nonverbal, right hemisphere is only now beginning to reveal its secrets and provide scientists with new insights into such baffling disorders as Capgras syndrome and asomatognosia. Knowing who we are may be reserved for a select few species with highly specialized brains. Working in tandem with the speaking left hemisphere, our selves and our right hemisphere might be indicating that the question “Where am I?” may be as relevant as the question “Who am I?”

For a thorough review of face processing in the brain, see “About Faces,” elsewhere in this issue.

Some portions of this essay were adapted by Feinberg from Altered Egos: How the Brain Creates the Self and the forthcoming Lost Self: Pathologies of the Brain and Identity.