The Disordered Mind: What Unusual Brains Tell Us About Ourselves

by Eric R. Kandel

Wernicke demonstrated not only that comprehension and expression are processed separately but that they are connected to each other by a pathway known as the arcuate fasciculus.

Wernicke predicted that someday, someone would find a disorder of language that involves simply a disconnect between the two areas. This proved to be the case: people with damage to the arcuate pathway connecting the two areas can understand language and express language, but the two functions operate independently. This is a bit like a presidential press conference: information comes in, information goes out, but there is no logical connection between them.

The search for localization of function in the brain was enhanced dramatically in the 1930s and ’40s by Canada’s renowned neurosurgeon Wilder Penfield, who operated on people suffering from epilepsy caused by scar tissue that had formed in the brain after a head injury.

Cajal determined that each neuron has the same four principal anatomical components (fig. 1.3): the cell body, the dendrites, the axon, and the presynaptic terminals, which end in what are now known as synapses.

Eventually, Cajal united these four principles in a theory now called the Neuron Doctrine (fig. 1.4). The first principle is that each neuron is a discrete element that serves as the fundamental building block and signaling unit of the brain. The second is that neurons interact with one another only at the synapses. In this way, neurons form the intricate networks, or neural circuits, that enable them to communicate information from one cell to another. The third principle is that neurons form connections only with particular target neurons at particular sites. This connection specificity accounts for the astonishingly precise circuitry that underlies the complex tasks of perception, action, and thought. The fourth principle, which derives from the first three, is that information flows in one direction only—from the dendrites to the cell body to the axon, then along the axon to the synapse. We now call this flow of information in the brain the principle of dynamic polarization.

Studies of twins show that autism also has a powerful genetic component: when one identical twin has autism, the other identical twin has a 90 percent chance of developing the disorder. A different sibling in that same family, including a fraternal twin, is considerably less likely to develop autism, while an individual in the general population has only a scant chance of developing the disorder (table 1).

When both genes and environment are involved, it is usually easier to find candidate genes first, by carrying out large-scale studies to determine which genes correlate with depression and which correlate with mania, and then try to sort out the environmental contribution.

Structural imaging looks at the anatomy of the brain. Computed tomography (CT) combines a series of X-ray images taken from different angles into a cross-sectional picture.

one step further, introducing the dimension of time. Functional imaging enables scientists to observe activity in the brain of a person who is carrying out a cognitive task, such as looking at a work of art, hearing, thinking, or remembering. Functional magnetic resonance imaging (fMRI) works by detecting changes in the concentration of oxygen in red blood cells. When an area of the brain becomes more active, it consumes more oxygen; to meet the demand for more oxygen, blood flow to the area increases. Thus, scientists can use fMRI to create maps showing which parts of the brain are active during a variety of mental tasks.

Functional imaging evolved from studies pioneered by Seymour Kety and his colleagues, who in 1945 developed the first effective way to measure blood flow in the living brain. In a series of classic studies, they measured blood flow in the brains of people who were awake and people who were asleep, thereby establishing the basis for subsequent studies using functional imaging.

An animal model of a disorder can be engineered in two ways. One, as we have seen, is by identifying the genes in an animal that are equivalent to the human genes thought to contribute to a disorder, altering those animal genes, and then observing the effects on the animal. The second is by inserting a human gene into an animal’s genome to see whether it produces the same effects in the animal as in people.

Neurological disorders tend to produce unusual behavior, or fragmentation of behavior into component parts, such as unusual movements of a person’s head or arms, or loss of motor control. By contrast, the major psychiatric disorders are often characterized by exaggerations of everyday behavior.

A second apparent difference is in how readily we can see actual physical damage to the brain. Damage resulting from neurological disorders, as we have learned, is often clearly visible at autopsy or in structural imaging. Damage resulting from psychiatric disorders is often less obvious, but as imaging techniques improve in resolution, we are beginning to detect changes resulting from these disorders.

The third apparent difference is location. Because of neurology’s traditional emphasis on anatomy, we know a great deal more about the neural circuitry of neurological disorders than of psychiatric disorders.

are anatomical findings regarding the timing of brain growth and development in autistic children. Before the age of two, the circumference of an autistic child’s head is often larger than that of a typically developing child. In addition, some regions of an autistic child’s brain may develop prematurely during the first years of life, particularly the frontal lobe, which is involved in attention and in decision making, and the amygdala, which is involved in emotions.

Autism was recognized as a separate disorder in the early 1940s by two scientists who had no contact with each other: Leo Kanner, working in the United States, and Hans Asperger, working in Austria.

Based on his analysis of Donald and the ten other children, Kanner presented a vivid picture of the three important features of classic autism in childhood: (1) profound aloneness, a strong preference for being by oneself; (2) a desire for things to be the same, not to change; and (3) islets of creative ability.

Matthew State, now at the University of California, San Francisco, has found that having an extra copy of one segment of chromosome 7 puts people at much greater risk of developing a disorder on the autism spectrum. When that same brain region is lost, however, the result is Williams syndrome.15 Williams syndrome is virtually the reverse of autism. Children with this genetic disorder are extremely social (fig. 2.9). They have a strong, almost irrepressible desire to speak and communicate. They are very friendly and trusting, even of strangers. Moreover, whereas some children with autism have strong drawing skills, children with Williams syndrome tend to be musical. In fact, children with Williams syndrome have difficulty constructing visuospatial relationships, which may account for their inability to draw well.

Four nearly simultaneous studies by scientists at Yale, the University of Washington, the Broad Institute at the Massachusetts Institute of Technology, and Cold Spring Harbor Laboratory have found that de novo mutations markedly increase the risk of autism.17

Applying those criteria to mental illnesses, Kraepelin distinguished two major groups of psychotic disorders: disorders of thought and disorders of mood. He called the disorders of thought dementia praecox—the dementia of young people—because they start earlier in life than other dementias, such as Alzheimer’s, and he called the disorders of mood manic-depressive illness because they manifest themselves as either depressed or elevated feeling states. We now refer to dementia praecox as schizophrenia, and we refer to manic-depressive illness as bipolar disorder. We refer to depressed states alone, with no manic component, as major depression, or unipolar depression. The majority of people with depressive disorders are unipolar.

Although twice as many women suffer from depression as men, and women attempt suicide three times more often than men, men are three or four times as likely to actually kill themselves. The reason is that men tend to choose more aggressive methods—guns, jumping off bridges, throwing themselves under a subway train—and such methods are more likely to be fatal.

In 1928 Mary Bernheim, a graduate student in the Department of Biochemistry at the University of Cambridge in England, discovered monoamine oxidase (MAO), an enzyme that breaks down a class of neurotransmitters known as monoamines.6 (Neurotransmitters, as we have seen, are chemical messengers that neurons release into the synapses to communicate with other neurons.) Her discovery led to the introduction of a drug called iproniazid, which was used to treat people with tuberculosis. In 1951 doctors and nurses working on the tuberculosis ward of Sea View Hospital on Staten Island, New York, noticed that their patients taking iproniazid seemed less lethargic and much happier than those who were not taking the drug. Subsequent clinical trials revealed that iproniazid had antidepressant properties. Shortly thereafter, imipramine, a drug developed initially to treat people with schizophrenia, was also found to relieve symptoms of depression, by blocking the reuptake of monoamines into nerve terminals. Reuptake is a process that recycles the neurotransmitters and stops the signaling.

Researchers discovered that monoamine oxidase breaks down and removes from the synapses two neurotransmitters: noradrenaline and serotonin. Without enough of these neurotransmitters, people experience symptoms of depression. The scientists reasoned that inhibiting the action of the enzyme that removes the monoaminergic transmitter from the synapse leaves more noradrenaline and serotonin in the synapses, thereby relieving the symptoms of depression. Thus, the idea of monoamine oxidase inhibitors as a treatment for depression was born.

In time, however, scientists realized that treating depression is more than a simple matter of flooding the synapses with serotonin. To begin with, boosting serotonin didn’t help all patients get better.

Ketamine works differently from traditional antidepressants. To begin with, it targets glutamate, not serotonin. To understand why this is important, we must first know that neurotransmitters fall into two categories: mediating and modulatory. Mediating neurotransmitters are released by a neuron at the synapse and act directly on the target cell, either exciting the target cell or inhibiting it. Glutamate is the most common excitatory transmitter, and GABA (gamma aminobutyric acid) is the most common inhibitory transmitter. Modulatory neurotransmitters, on the other hand, fine-tune the action of excitatory and inhibitory neurotransmitters. Dopamine and serotonin are modulatory neurotransmitters. Because ketamine acts on the excitatory neurotransmitter glutamate, which directly affects the target cell, the drug reduces depression more quickly than drugs that act on the modulatory transmitter serotonin. In addition, ketamine prevents the transmission of glutamate from one neuron to the next by blocking a particular glutamate receptor on the target cell. Since a receptor blocked by ketamine can’t bind glutamate, the neurotransmitter can’t affect the target cell. The demonstration of the antidepressant effect of ketamine profoundly changed the way we think about depression.

Electroconvulsive (shock) therapy gained a bad reputation during the 1940s and ’50s because patients were given high doses of electricity without any anesthesia, resulting in pain, fractured bones, and other serious side effects. Today, electroconvulsive therapy is painless. It is administered after the patient has been given general anesthesia and a muscle relaxant, it uses small electric currents to induce a brief seizure, and it is often very effective. Many patients have six to twelve sessions over a period of several weeks. Scientists are still not clear about how it works, but it is thought to relieve depression by producing changes in the chemistry of the brain. Unfortunately, the effects of electroconvulsive therapy usually do not last very long.

Once the first manic episode is initiated—usually at the age of seventeen or eighteen—the brain is changed in ways we do not yet understand, such that even minor events can trigger a later manic episode. After the third or fourth manic episode, a trigger may not be required. As a person with bipolar disorder grows older, the disease advances and the intervals between episodes may become shorter, particularly if he or she discontinues treatment. Bipolar disorder affects about 1 percent of Americans, or more than 3 million people. While depression affects more women than men, bipolar disorder affects men and women equally. The disorder takes several forms, but the most common forms are known as bipolar I and bipolar II. People with bipolar I disorder have manic episodes, and sometimes cross over into psychosis with symptoms such as delusions and hallucinations, whereas people with bipolar II disorder have less severe, hypomanic episodes. Some people experience symptoms of both mania and depression at the same time, a condition known as a mixed state.

Unlike other medications used to treat psychiatric illnesses, lithium is a salt; consequently, it does not bind to a receptor on the surface of a neuron. Rather, it is actively transported into the neuron through sodium ion channels in the cell membrane that open in response to an external stimulus (see chapter 1). When a sodium ion channel opens, both sodium and lithium move into the cell. The sodium is subsequently pumped out, but the lithium remains inside. There, lithium may stabilize mood swings by affecting the action of neurotransmitters, either directly or through interaction with a second messenger system. As we have seen, neurotransmitters bind to receptors on the cell membrane. This activates second messenger systems, which transmit signals from the receptors to molecules inside the neuron. Lithium may blunt the activation of second messenger systems, thus reducing signal transmission. Lithium may also damp down a neuron’s responsiveness to neurotransmitters inside the cell. This could explain why lithium works so effectively in bipolar disorder: it may decrease a neuron’s sensitivity to both external and internal stimuli. In addition, lithium affects the modulatory neurotransmitters serotonin and dopamine as well as the mediating neurotransmitter GABA. Thus, its efficacy may be attributable to its wide-ranging neurobiological effects rather than to a single mechanism.

Positive symptoms reflect disordered volition and thinking. Disordered thought detaches a person from reality, leading to altered perceptions and behavior, such as hallucinations and delusions. These psychotic symptoms can be terrifying, not just for people who experience them but also for people who witness them.

Hallucinations, the most common positive symptom, can be visual or auditory.

The negative symptoms of schizophrenia—social withdrawal and lack of motivation—are typically present before the positive symptoms, but they are generally overlooked until a person experiences a psychotic episode.

Brain scans of untreated people with schizophrenia reveal, over time, a subtle, but perceptible, loss of gray matter, which contains the cell body and dendrites of neurons in the cerebral cortex. This loss of gray matter, which contributes to the cognitive symptoms of schizophrenia, is thought to result from excessive pruning of dendrites during development, which leads to loss of synaptic connections among neurons, as we shall see later in this chapter.

It’s ten o’clock on a Friday night. I am sitting with my two classmates in the Yale Law School Library. They aren’t too happy about being here; it’s the weekend, after all—there are plenty of other fun things they could be doing. But I am determined that we hold our small-group meeting. We have a memo assignment; we have to do it, have to finish it, have to produce it, have to … Wait a minute. No, wait. “Memos are visitations,” I announce. “They make certain points. The point is on your head. Have you ever killed anyone?”

Schizophrenia is not a rare disorder. It affects about 1 percent of people worldwide and roughly 3 million people in the United States.

Paul Charpentier, a French chemist working for the pharmaceutical firm Rhône-Poulenc, had begun work on an antihistamine that he hoped would be effective against allergies but without producing the numerous side effects of existing antihistamines. The drug he developed in 1950 was called Thorazine (its generic name is chlorpromazine). As Thorazine went into clinical trials, everyone was amazed at its effect: it made people calmer, much more relaxed.

The first clue to how typical antipsychotics work came from analysis of their neurological side effects. Since these drugs produce the same effects on movement as Parkinson’s disease, which is caused by a deficiency in the modulatory neurotransmitter dopamine, scientists reasoned that the drugs might act by reducing dopamine in the brain. They also reasoned, by extension, that schizophrenia might result in part from excessive action of dopamine. In other words, reducing dopamine in the brain might account for both the drugs’ therapeutic effects and their adverse side effects.

When neurons release dopamine into a synapse, the dopamine ordinarily binds to receptors on target neurons. If those receptors are blocked by antipsychotics, the action of dopamine is attenuated. As it turns out, many typical antipsychotics act by blocking dopamine receptors. This finding bolstered the idea that either excessive dopamine production or an excessive number of dopamine receptors is an important factor in causing schizophrenia.

Most dopamine-producing neurons are located in two clusters in the midbrain: the ventral tegmental area and the substantia nigra. The axons that extend outward from these two clusters of neurons form the neural circuits known as the dopaminergic pathways. Two of these dopaminergic pathways—the mesolimbic pathway and the nigrostriatal pathway—are the neural pathways primarily affected in schizophrenia and are therefore the most important ones to examine in looking for treatments (fig. 4.3).

The mesolimbic pathway extends from the ventral tegmental area to parts of the prefrontal cortex, hippocampus, amygdala, and nucleus accumbens. These regions are important for thought, memory, emotion, and behavior—the mental functions that are adversely affected by schizophrenia.

Working memory develops from childhood through the late teens, getting progressively better over time. At age seven, children who will be diagnosed with schizophrenia ten or fifteen years later have normal working memory. But by age thirteen, their working memory has fallen well below where it should be at that stage of development. A key component of working memory is the pyramidal neurons of the prefrontal cortex, so called because the cell body of these neurons is shaped roughly like a triangle. In every other respect these cells are like other neurons, both structurally and functionally.

As we have seen, neurons send information outward along the axon, which forms synaptic connections with a target cell’s dendrites. Most of a pyramidal neuron’s synapses are located on small protrusions from the dendrites called dendritic spines. The number of dendritic spines on a neuron is a rough measure of the amount and richness of the information it receives. Dendritic spines begin to form on pyramidal neurons during the third trimester of pregnancy. From then through the first few years of life, the number of dendritic spines, and the number of synapses on them, expands rapidly. In fact, a three-year-old’s brain contains twice as many synapses as an adult’s brain. Beginning at about puberty, synaptic pruning removes the dendritic spines that the brain isn’t using, including spines that aren’t actually helping working memory. Synaptic pruning becomes particularly active during adolescence and early adulthood.

In schizophrenia, synaptic pruning appears to go haywire during adolescence, snipping off far too many dendritic spines (fig. 4.4). Consequently, the pyramidal neurons are left with too few synaptic connections in the prefrontal cortex to form the robust neural circuits we need for an adequate working memory and other complex cognitive functions. This excessive-pruning hypothesis for schizophrenia, first proposed by Irwin Feinberg, now at the University of California, Davis,2 has been documented by David Lewis and Jill Glausier at the University of Pittsburgh.3 A similar defect is thought to affect pyramidal neurons located in the hippocampus of people with schizophrenia, which would adversely affect memory.

If you had an identical twin with schizophrenia, you would have about a 50-50 chance of developing the disease, regardless of whether the two of you were raised together or apart. That risk of developing schizophrenia is much higher than the 1 in 100 risk for the general population. The twin data tell us two things: first, schizophrenia has a strong genetic component, regardless of environment; and second, those genes can’t be acting alone, because the risk isn’t 100 percent.

About 30 percent of adults with the syndrome are diagnosed with psychiatric disorders, including bipolar disorder and anxiety disorders. But schizophrenia is by far the most prevalent disorder. In fact, the risk of schizophrenia in a person with 22q11 deletion syndrome is twenty to twenty-five times greater than the risk of schizophrenia in the general population.

The researchers identified two genes that are disrupted by the translocation: DISC1 (disruption in schizophrenia 1) and DISC2 (disruption in schizophrenia 2). Although this particular translocation has been found in only one family, that family’s unusually high incidence of psychiatric disorders suggests that these two genes, and other genes close to where the chromosomes broke, may be responsible for psychotic symptoms in schizophrenia and in mood disorders.

Recently, Steven McCarroll, Beth Stevens, Aswin Sekar, and their colleagues at Harvard Medical School provided further evidence in support of this idea. They also described how and why pruning may go wrong, and they have identified the gene responsible.7 The researchers focused on a particular region of the human genome, a locus called the major histocompatibility complex (MHC). This complex of genes on chromosome 6 encodes proteins that are essential for recognizing foreign molecules, a critical step in the body’s immune response. The MHC locus, which had been strongly associated with schizophrenia in previous genetic studies, contains a gene called C4. The activity of the C4 gene—that is, its level of expression—varies significantly among individuals. The researchers wanted to find out how variations in the C4 gene are related to its level of expression and whether its level of expression is related to schizophrenia.

Earlier studies had found that proteins produced by genes in the MHC locus play a role in immunity and are involved in synaptic pruning during normal development. This raised a critical question: What exactly is the role of the protein product produced by the C4-A gene? To answer the question, scientists bred mice without the gene. They observed less-than-normal synaptic pruning in these mice, indicating that the protein’s role is to promote pruning and suggesting that too much of the protein leads to excessive pruning. In studies of these mice McCarroll, Stevens, Sekar, and their colleagues also found that during normal development the C4-A protein “tags” the synapses to be pruned. The more active the C4 gene is, the more synapses are deleted. Together, these studies suggest that overexpression of the C4-A variant leads to excessive synaptic pruning.

Imaging has also confirmed that psychotherapy is a biological treatment—that it physically changes the brain, as drugs do. Imaging has even predicted, in some cases of depression, which patients are best treated with drugs, with psychotherapy, or with both.

For a long time Milner thought that H.M.’s memory deficit applied to all areas of knowledge. Then she made a remarkable discovery. She asked H.M. to trace the outline of a star while looking at his hand, his pencil, and the paper in a mirror. Everyone who tries this tracing task makes errors on the first day, drawing outside the line of the star and having to adjust back in, but people with normal memory improve to almost perfect performance by the third day. If H.M.’s memory loss applied to all areas of knowledge, he should show no such improvement. Yet after three days, and despite having no memory of practicing the task or of having seen Milner before, H.M. had learned this motor task as well as anyone else (fig. 5.4).

Another person who challenged the idea that a complex neural circuit is required for learning was the Canadian psychologist Donald Hebb. Hebb suggested that associative learning could be produced by the simple interaction of two neurons: if neuron A repeatedly stimulates neuron B to fire an action potential—the electrical impulse that travels down the axon to the synapse—a change will take place in one or both of those cells. That change strengthens the synaptic connection between the two neurons. The strengthened connection creates and stores, for a short time, a memory of the interaction.3 Two researchers working at the University of Göteborg in Sweden, Holger Wigström and Bengt Gustafsson, later provided the first evidence suggesting that Hebb’s mechanism might be at work in the formation of explicit memory in the hippocampus.