Several years ago, while I was on a sabbatical leave, a colleague stopped by my office and asked whether I would like to see an interesting patient. The patient, a 72-year-old man, had suffered a massive stroke in his right hemisphere that had paralyzed the left side of his body.

Mr. V. was seated in a wheelchair equipped with a large tray on which his right arm was resting; his left arm was immobilized in a sling, to keep it out of the way. He greeted us politely, almost formally, articulating his words carefully with a slight European accent.

Mr. V. seemed intelligent, and this impression was confirmed when we gave him some of the subtests of the Wechsler Adult Intelligence Test. His verbal intelligence appeared to be in the upper 5 percent of the population. The fact that English was not his native language made his performance even more remarkable.

The most interesting aspect of Mr. V.’s behavior after his stroke was his lack of reaction to his symptoms. After we had finished with the testing, we asked him to tell us a little about himself and his lifestyle. What, for example, was his favorite pastime?

“I like to walk,” he said. “I walk at least two hours each day around the city, but mostly I like to walk in the woods. I have maps of most of the national forests in the state on the walls of my study, and I mark all of the trails I’ve taken. I figure that in about six months I will have walked all of the trails that are short enough to do in a day.”

“You’re going to finish up those trails in the next six months?” asked Dr. W.

“Yes, and then I’ll start over again!” he replied.

“Mr. V., are you having any trouble?” asked Dr. W.

“Trouble? What do you mean?”

“I mean physical difficulty.”

“No.” Mr. V. gave him a slightly puzzled look.

“Well, what are you sitting in?”

Mr. V. gave him a look that indicated he thought that the question was rather stupid—or perhaps insulting. “A wheelchair, of course,” he answered.

“Why are you in a wheelchair?”

Now Mr. V. looked frankly exasperated; he obviously did not like to answer foolish questions. “Because my left leg is paralyzed!” he snapped.

Mr. V. clearly knew what his problem was, but he failed to understand its implications. He could verbally recognize his disability, but he was unable to grasp its significance. Thus, he blandly accepted the fact that he was confined to a wheelchair. The implications of his disability did not affect him emotionally or figure into his plans.

The word emotion refers to positive or negative reactions to particular situations. There is no such thing as a neutral emotion. For example, being treated unfairly makes us angry, seeing someone suffer makes us sad, and being close to a loved one makes us feel happy. Emotions consist of patterns of physiological changes and accompanying behaviors—or at least urges to perform these behaviors. These responses are accompanied by feelings. In fact, most of us use the word emotion to refer to the feelings, not to the behaviors. But it is behavior, not private experience, that has consequences for survival and reproduction. Therefore, the useful purposes served by emotional behaviors are what guided the evolution of our brain.

This chapter is divided into three major sections. The first considers the patterns of behavioral and physiological responses that constitute the negative emotions of fear and anger. It describes the nature of these response patterns, their neural and hormonal control, and the role of emotions in moral judgments and social behavior. The second section describes the communication of emotions—their expression and recognition. The third section examines the nature of the feelings that accompany emotions.

Emotions as Response Patterns

An emotional response consists of three types of components: behavioral, autonomic, and hormonal. The behavioral component consists of muscular movements that are appropriate to the situation that elicits them. For example, a dog defending its territory against an intruder first adopts an aggressive posture, growls, and shows its teeth. If the intruder does not leave, the defender runs toward it and attacks. Autonomic responses facilitate the behaviors and provide quick mobilization of energy for vigorous movement. In this example the activity of the sympathetic branch increases while that of the parasympathetic branch decreases. As a consequence the dog’s heart rate increases, and changes in the size of blood vessels shunt the circulation of blood away from the digestive organs toward the muscles. Hormonal responses reinforce the autonomic responses. The hormones secreted by the adrenal medulla—epinephrine and norepinephrine—further increase blood flow to the muscles and cause nutrients stored in the muscles to be converted into glucose. In addition, the adrenal cortex secretes steroid hormones, which also help to make glucose available to the muscles.

This section discusses research on the control of overt emotional behaviors and the autonomic and hormonal responses that accompany them. Special behaviors that serve to communicate emotional states to other animals, such as the threat gestures that precede an actual attack and the smiles and frowns used by humans, are discussed in the second section of the chapter. As you will see, negative emotions receive much more attention than positive ones do. Most of the research on the physiology of emotions has been confined to fear and anger—emotions associated with situations in which we must defend ourselves or our loved ones. The physiology of behaviors associated with positive emotions—such as those associated with lovemaking, caring for one’s offspring, or enjoying a good meal or a cool drink of water (or an alcoholic beverage)—is described in other chapters but not in the specific context of emotions. And Chapter 18 discusses the consequences of situations that evoke negative emotions: stress.

Fear

As we saw, emotional responses involve behavioral, autonomic, and hormonal components. These components are controlled by separate neural systems. The integration of the components of fear appears to be controlled by the amygdala.

RESEARCH WITH LABORATORY ANIMALS

The amygdala plays a special role in physiological and behavioral reactions to objects and situations that have biological significance, such as those that warn of pain or other unpleasant consequences or signify the presence of food, water, salt, potential mates or rivals, or infants in need of care. Researchers in several different laboratories have shown that single neurons in various nuclei of the amygdala become active when emotionally relevant stimuli are presented. For example, these neurons are excited by such stimuli as the sight of a device that has been used to squirt either a bad-tasting solution or a sweet solution into the animal’s mouth, the sound of another animal’s vocalization, the sound of the opening of the laboratory door, the smell of smoke, or the sight of another animal’s face (O’Keefe and Bouma, 1969 ; Jacobs and McGinty, 1972 ; Rolls, 1982 ; Leonard et al., 1985 ). And as we saw in Chapter 10 , the amygdala is involved in the effects of olfactory stimuli on reproductive physiology and behavior. This section describes research on the role of the amygdala in organizing emotional responses produced by aversive stimuli.

The amygdala (or, more precisely, the amygdaloid complex) is located within the temporal lobes. It consists of several groups of nuclei, each with different inputs and outputs—and with different functions (Amaral et al., 1992 ; Pitkänen, Savander, and LeDoux, 1997 ; Stefanacci and Amaral, 2000 ). The amygdala has been subdivided into approximately twelve regions, each containing several subregions. However, we need concern ourselves with just three major regions: the lateral nucleus, the basal nucleus, and the central nucleus.

The lateral nucleus (LA) receives information from all regions of the neocortex, including the ventromedial prefrontal cortex, the thalamus, and the hippocampal formation. The lateral nucleus sends information to the basal nucleus (B) and to other parts of the brain, including the ventral striatum (a brain region involved in the effects of reinforcing stimuli on learning) and the dorsomedial nucleus of the thalamus, whose projection region is the prefrontal cortex. The LA and B nuclei send information to the ventromedial prefrontal cortex and the central nucleus (CE) , which projects to regions of the hypothalamus, midbrain, pons, and medulla that are responsible for the expression of the various components of emotional responses. As we will see, activation of the central nucleus elicits a variety of emotional responses: behavioral, autonomic, and hormonal. (See Figure 11.1 . )

lateral nucleus (LA) A nucleus of the amygdala that receives sensory information from the neocortex, thalamus, and hippocampus and sends projections to the basal, accessory basal, and central nucleus of the amygdala.

basal nucleus (B) A nucleus of the amygdala that receives information from the lateral nucleus and sends projections to the ventromedial prefrontal cortex and the central nucleus.

central nucleus (CE) The region of the amygdala that receives information from the basal, lateral, and accessory basal nuclei and sends projections to a wide variety of regions in the brain; involved in emotional responses.

The central nucleus of the amygdala is the single most important part of the brain for the expression of emotional responses provoked by aversive stimuli. When threatening stimuli are perceived, neurons in the central nucleus become activated (Pascoe and Kapp, 1985 ; Campeau et al., 1991 ). Damage to the central nucleus (or to the nuclei that provide it with sensory information) reduces or abolishes a wide range of emotional behaviors and physiological responses. After the central nucleus has been destroyed, animals no longer show signs of fear when confronted with stimuli that have been paired with aversive events. They also act more tamely when handled by humans, their blood levels of stress hormones are lower, and they are less likely to develop ulcers or other forms of stress-induced illnesses (Coover, Murison, and Jellestad, 1992 ; Davis, 1992 ; LeDoux, 1992 ). Normal monkeys show signs of fear when they see a snake, but those with amygdala lesions do not (Amaral, 2003 ). In contrast, when the central amygdala is stimulated by means of electricity or by an injection of an excitatory amino acid, the animal shows physiological and behavioral signs of fear and agitation (Davis, 1992 ), and long-term stimulation of the central nucleus produces stress-induced illnesses such as gastric ulcers (Henke, 1982 ). These observations suggest that the autonomic and endocrine responses controlled by the central nucleus are among those responsible for the harmful effects of long-term stress, which are discussed in Chapter 18 . Rather than describing the regions to which the amygdala projects and the responses these regions control, I will refer you to Figure 11.2 , which summarizes this information. (See Figure 11.2 . )

FIGURE 11.1 The Amygdala

This figure shows a much-simplified diagram of the major divisions and connections of the amygdala that play a role in emotions.

A few stimuli automatically activate the central nucleus of the amygdala and produce fear reactions—for example, loud unexpected noises, the approach of large animals, heights, or (for some species) specific sounds or odors. Even more important, however, is the ability to learn that a particular stimulus or situation is dangerous or threatening. Once the learning has taken place, that stimulus or situation will evoke fear: heart rate and blood pressure will increase, the muscles will become more tense, the adrenal glands will secrete epinephrine, and the animal will proceed cautiously, alert and ready to respond.

FIGURE 11.2 Outputs of the Central Nucleus of the Amygdala

Shown here are some important brain regions that receive input from the central nucleus of the amygdala and the emotional responses controlled by these regions.

(Adapted from Davis, M., Trends in Pharmacological Sciences, 1992, 13, 35–41.)

The most basic form of emotional learning is a conditioned emotional response , which is produced by a neutral stimulus that has been paired with an emotion-producing stimulus. The word conditioned refers to the process of classical conditioning, which is described in more detail in Chapter 13 . Briefly, classical conditioning occurs when a neutral stimulus is regularly followed by a stimulus that automatically evokes a response. For example, if a dog regularly hears a bell ring just before it receives some food that makes it salivate, it will begin salivating as soon as it hears the sound of the bell. (You probably already know that this phenomenon was discovered by Ivan Pavlov.)

conditioned emotional response A classically conditioned response that occurs when a neutral stimulus is followed by an aversive stimulus; usually includes autonomic, behavioral, and endocrine components such as changes in heart rate, freezing, and secretion of stress-related hormones.

FIGURE 11.3 The Procedure Used to Produce Conditioned Emotional Responses

Several laboratories have investigated the role of the amygdala in the development of classically conditioned emotional responses. For example, these responses can be produced in rats by presenting an auditory stimulus such as a tone, followed by a brief electrical shock delivered to the feet through the floor on which the animals are standing. (See Figure 11.3 . ) By itself the shock produces an unconditionalemotional response: The animal jumps into the air, its heart rate and blood pressure increase, its breathing becomes more rapid, and its adrenal glands secrete catecholamines and steroid stress hormones. After several pairings of the tone and the shock, classical conditioning is normally established.

The next day, if the tone is presented alone—not followed by a shock—physiological monitoring will show the same physiological responses the animals produced when they were shocked during training. In addition, they will show behavioral arrest—a species-typical defensive response called freezing. In other words, the animals act as if they were expecting to receive a shock. The tone becomes a conditional stimulus (CS) that elicits freezing: a conditional response (CR).

Research indicates that the physical changes responsible for the classical conditioning of a conditioned emotional response take place in the lateral nucleus of the amygdala (Paré, Quirk, and LeDoux, 2004 ). Neurons in the lateral nucleus communicate with neurons in the central nucleus, which in turn communicate with regions in the hypothalamus, midbrain, pons, and medulla that are responsible for the behavioral, autonomic, and hormonal components of a conditioned emotional response. More recent studies indicate that the neural circuitry responsible for the process of classical conditioning is actually more complex than that (Ciocchi et al, 2010 ; Haubensak et al., 2010 ; Duvarci, Popa, and Paré, 2011 ), but the details of this process will be postponed until Chapter 13 , which discusses the physiology of learning and memory.

The neural mechanisms responsible for classically conditioned emotional responses evolved because they play a role in an animal’s survival but increase the likelihood that an animal can avoid dangerous situations, such as places where it encountered aversive events. In the laboratory, if the CS (tone) is presented repeatedly by itself, the previously established CR (emotional response) eventually disappears—it becomes extinguished. After all, the value of a conditioned emotional response is that it prepares an animal to confront (or, better yet, avoid) an aversive stimulus. If the CS occurs repeatedly but the aversive stimulus does not follow, then it is better for the emotional response—which itself is disruptive and unpleasant—to disappear. And that is exactly what happens.

Behavioral studies have shown that extinction is not the same as forgetting. Instead, the animal learns that the CS is no longer followed by an aversive stimulus, and, as a result of this learning, the expression of the CR is inhibited; the memory for the association between the CS and the aversive stimulus is not erased. This inhibition is supplied by the ventromedial prefrontal cortex (vmPFC) . Evidence for this conclusion comes from several studies (Amano, Unal, and Paré, 2010 ; Sotres-Bayon and Quirk, 2010 ). For example, lesions of the vmPFC impair extinction, stimulation of this region inhibits conditioned emotional responses, and extinction training activates neurons there. Besides playing an essential role in extinction of conditioned emotional responses, the vmPFC can modulate the expression of fear in different circumstances. Depending on the situation, one subregion of the prefrontal cortex can become active and suppress a conditioned fear response, and another subregion can become active and enhance the response.

ventromedial prefrontal cortex (vmPFC) The region of the prefrontal cortex at the base of the anterior frontal lobes, adjacent to the midline.

RESEARCH WITH HUMANS

We humans also acquire conditioned emotional responses. Let’s examine a specific (if somewhat contrived) example. Suppose you are helping a friend prepare a meal. You pick up an electric mixer to mix some batter for a cake. Before you can turn the mixer on, the device makes a sputtering noise and then gives you a painful electrical shock. Your first response would be a defensive reflex: You would let go of the mixer, which would end the shock. This response is specific; it is aimed at terminating the painful stimulus. In addition, the painful stimulus would elicit nonspecific responses controlled by your autonomic nervous system: Your eyes would dilate, your heart rate and blood pressure would increase, you would breathe faster, and so on. The painful stimulus would also trigger the secretion of some stress-related hormones, another nonspecific response.

Suppose that a while later you visit your friend again and once more agree to make a cake. Your friend tells you that the electric mixer is perfectly safe. It has been fixed. Just seeing the mixer and thinking of holding it again makes you a little nervous, but you accept your friend’s assurance and pick it up. Just then, it makes the same sputtering noise that it did when it shocked you. What would your response be? Almost certainly, you would drop the mixer again, even if it did not give you a shock. And your pupils would dilate, your heart rate and blood pressure would increase, and your endocrine glands would secrete some stress-related hormones. In other words, the sputtering sound would trigger a conditioned emotional response.

Evidence indicates that the amygdala is involved in emotional responses in humans. One of the earliest studies observed the reactions of people who were being evaluated for surgical removal of parts of the brain to treat severe seizure disorders. These studies found that stimulation of parts of the brain (for example, the hypothalamus) produced autonomic responses that are often associated with fear and anxiety but that only when the amygdala was stimulated did people also report that they actually feltafraid (White, 1940 ; Halgren et al., 1978 ; Gloor et al., 1982 ).

Many studies have shown that lesions of the amygdala decrease people’s emotional responses. For example, Bechara et al. ( 1995 ) and LaBar et al. ( 1995 ) found that people with lesions of the amygdala showed impaired acquisition of a conditioned emotional response, just as rats do.

Most human fears are probably acquired socially, not through firsthand experience with painful stimuli (Olsson, Nearing, and Phelps, 2007 ). For example, a child does not have to be attacked by a dog to develop a fear of dogs: He or she can develop this fear by watching another person being attacked or (more often) by seeing another person display signs of fear when encountering a dog. People can also acquire a conditioned fear response through instruction. For example, suppose that someone is told (and believes) that if a warning light in a control panel goes on, he or she should leave the room immediately because the light is connected to a sensor that detects toxic fumes. If the light does go on, the person will leave the room and is also likely to experience a fear response while doing so.

We saw that studies with laboratory animals indicate that the vmPFC plays a critical role in extinction of a conditioned emotional response. The same is true for humans. Phelps et al. ( 2004 ) directly established a conditioned emotional response in human subjects by pairing the appearance of a visual stimulus with electric shocks to the wrist and then extinguished the response by presenting the squares alone, without any shocks. As Figure 11.4 shows, increased activity of the medial prefrontal cortex correlated with extinction of the conditioned response. (See Figure 11.4 . )

FIGURE 11.4 Control of Extinction

These graphs show the correlation between activation of the medial prefrontal cortex and establishment of extinction.

(Based on data from Phelps et al., 2004)

Damage to the amygdala interferes with the effects of emotions on memory. Normally, when people encounter events that produce a strong emotional response, they are more likely to remember these events. Cahill et al. ( 1995 ) studied a patient with bilateral degeneration of the amygdala. The investigators narrated a story about a young boy walking with his mother on his way to visit his father at work. To accompany the story, they showed a series of slides. During one part of the story, the boy was injured in a traffic accident, and gruesome slides illustrated his injuries. When this slide show is presented to normal subjects, they remember more details from the emotion-laden part of the story. However, a patient with amygdala damage showed no such increase in memory. In another study (Mori et al., 1999 ), researchers questioned patients with Alzheimer’s disease who had witnessed the devastating earthquake that struck Kobe, Japan, in 1995. They found that memory of this frightening event was inversely correlated with amygdala damage: The more a patient’s amygdala was degenerated, the less likely it was that the patient remembered the earthquake.

As we saw in Chapter 7 , Patient I. R., a woman who had sustained damage to the auditory association cortex, was unable to perceive or produce melodic or rhythmic aspects of music (Peretz et al., 2001 ). She could not even tell the difference between consonant (pleasant) and dissonant (unpleasant) music. However, she was still able to recognize the mood conveyed by music. Gosselin et al. ( 2005 ) found that patients with damage to the amygdala showed the opposite symptoms: They had no trouble with musical perception but were unable to recognize scary music. They could still recognize happy and sad music. Thus, amygdala lesions impair recognition of a musical style that is normally associated with fear.

Anger, Aggression, and Impulse Control

Almost all species of animals engage in aggressive behaviors, which involve threatening gestures or actual attack directed toward another animal. Aggressive behaviors are species typical; that is, the patterns of movements (for example, posturing, biting, striking, and hissing) are organized by neural circuits whose development is largely programmed by an animal’s genes. Many aggressive behaviors are related to reproduction. For example, aggressive behaviors that gain access to mates, defend territory needed to attract mates or to provide a site for building a nest, or defend offspring against intruders can all be regarded as reproductive behaviors. Other aggressive behaviors are related to self-defense, such as that of an animal threatened by a predator or an intruder of the same species.

Aggressive behavior can consist of actual attacks, or they may simply involve threat behaviors , which consist of postures or gestures that warn the adversary to leave or it will become the target of an attack. The threatened animal might show defensive behaviors —threat behaviors or an actual attack against the animal that is threatening it—or it might show submissive behaviors —behaviors that indicate that it accepts defeat and will not challenge the other animal. In the natural environment most animals display far more threats than actual attacks. Threat behaviors are useful in reinforcing social hierarchies in organized groups of animals or in warning intruders away from an animal’s territory. They have the advantage of not involving actual fighting, which can harm one or both of the combatants.

threat behavior A stereotypical species-typical behavior that warns another animal that it may be attacked if it does not flee or show a submissive behavior.

defensive behavior A species-typical behavior by which an animal defends itself against the threat of another animal.

submissive behavior A stereotyped behavior shown by an animal in response to threat behavior by another animal; serves to prevent an attack.

Predation is the attack of a member of one species on a member of another, usually because the latter serves as food for the former. While engaged in attacking a member of the same species or defending itself against the attack, an animal appears to be extremely aroused and excited, and the activity of the sympathetic branch of its autonomic nervous system is high. In contrast, the attack of a predator is much more “cold blooded”; it is generally efficient and not accompanied by a high level of sympathetic activation. A predator is not angry with its prey; attacking the prey is simply a means to an end.

predation Attack of one animal directed at an individual of another species on which the attacking animal normally preys.

RESEARCH WITH LABORATORY ANIMALS

Neural Control of Aggressive Behavior.

The neural control of aggressive behavior is hierarchical. That is, the particular muscular movements that an animal makes in attacking or defending itself are programmed by neural circuits in the brain stem. Whether an animal attacks depends on many factors, including the nature of the eliciting stimuli in the environment and the animal’s previous experience. The activity of the brain stem circuits appears to be controlled by the hypothalamus and the amygdala, which also influence many other species-typical behaviors. And, of course, the activity of the limbic system is controlled by perceptual systems that detect the status of the environment, including the presence of other animals.

A series of studies by Shaikh, Siegel, and their colleagues (reviewed by Gregg and Siegel, 2001 ) investigated the neural circuitry involved in aggressive attack and predation in cats. The investigators placed electrodes in various regions of the brain and observed the effects of electrical stimulation of these regions on the animals’ behavior. In some cases the electrode was actually a stainless-steel cannula, coated with an insulating material except for the tip. These devices (called cannula electrodes) could be used to infuse chemicals into the brain as well as to stimulate it. The investigators found that aggressive attack and predation can be elicited by stimulation of different parts of the PAG and that the hypothalamus and the amygdala influence these behaviors through excitatory and inhibitory connections with the PAG. They found that the three principal regions of the amygdala and two regions of the hypothalamus affect defensive rage and predation, both of which appear to be organized by the PAG. (They assessed predation by presenting the cats with an anesthetized rat, so no pain was inflicted.) A possible connection between the lateral hypothalamus and the ventral PAG has not yet been verified. Rather than listing the connections and their effects, I will direct you to Figure 11.5 .

Role of Serotonin.

An overwhelming amount of evidence suggests that the activity of serotonergic synapses inhibits aggression. In contrast, destruction of serotonergic axons in the forebrain facilitates aggressive attack, presumably by removing an inhibitory effect (Vergnes et al., 1988 ).

A group of researchers has studied the relationship between serotonergic activity and aggressiveness in a free-ranging colony of rhesus monkeys (reviewed by Howell et al., 2007 ). They assessed serotonergic activity by capturing the monkeys, removing a sample of cerebrospinal fluid (CSF), and analyzing it for 5-HIAA, a metabolite of serotonin (5-HT). When 5-HT is released, most of the neurotransmitter is taken back into the terminal buttons by means of reuptake, but some escapes and is broken down to 5-HIAA, which finds its way into the CSF. Thus, high levels of 5-HIAA in the CSF indicates an elevated level of serotonergic activity. The investigators found that young male monkeys with the lowest levels of 5-HIAA showed a pattern of risk-taking behavior, including high levels of aggression directed toward animals that were older and much larger than themselves. They were much more likely to take dangerous, unprovoked long leaps from tree to tree at a height of more than 7 m (27.6 ft). They were also more likely to pick fights that they could not possibly win. Of forty-nine preadolescent male monkeys that the investigators followed for four years, 46 percent of those with the lowest 5-HIAA levels died, while all of the monkeys with the highest levels survived. (See Figure 11.6 . ) Most of the monkeys that died were killed by other monkeys. In fact, the first monkey to be killed had the lowest level of 5-HIAA and was seen attacking two mature males the night before his death. Clearly, serotonin does not simply inhibit aggression; rather, it exerts a controlling influence on risky behavior, which includes aggression.

FIGURE 11.5 Neural Circuitry in Defensive Behavior and Predation

The diagram shows interconnections of parts of the amygdala, hypothalamus, and periaqueductal gray matter (PAG) and their effects on defensive rage and predation in cats, based on the studies by Shaikh, Siegel, and their colleagues. Black arrows indicate excitation; red arrows indicate inhibition.

FIGURE 11.6 Serotonin and Risk-Taking Behavior

The graph shows the percentage of young male monkeys alive or dead as a function of 5-HIAA level in the cerebrospinal fluid, measured four years previously.

(Based on data from Higley et al., 1996.)

Genetic studies with other species confirm the conclusion that serotonin has an inhibitory role in aggression. For example, selective breeding of rats and silver foxes have yielded animals that display tameness and friendly responses to human contact. These animals showed increased brain levels of serotonin and 5-HIAA (Popova, 2006 ).

RESEARCH WITH HUMANS

Human violence and aggression are serious social problems. Consider the following case histories:

· Born to an alcoholic teen mother who raised him with an abusive alcoholic stepfather, Steve was hyperactive, irritable, and disobedient as a toddler. . . . After dropping out of school at age 14, Steve spent his teen years fighting, stealing, taking drugs, and beating up girlfriends. . . . School counseling, a probation officer, and meetings with child protective service failed to forestall disaster: At 19, several weeks after his last interview with researchers, Steve visited a girlfriend who had recently dumped him, found her with another man, and shot him to death. The same day he tried to kill himself. Now he’s serving a life sentence without parole. (Holden, 2000 , p. 580)

· By the time Joshua had reached the age of 2, . . . he would bolt out of the house and into traffic. He kicked and head-butted relatives and friends. He poked the family hamster with a pencil and tried to strangle it. He threw regular temper tantrums and would stage toy-throwing frenzies. “At one point he was hurting himself—banging his head against a wall, pinching himself”, not to mention leaping off the refrigerator [said his mother]. . . . Showering Joshua with love . . . made little difference: By age 3, his behavior got him kicked out of his preschool. (Holden, 2000 , p. 581)

Role of Heredity.

Early experiences can certainly foster the development of aggressive behavior, but studies have shown that heredity plays a significant role. For example, Viding et al. ( 2005 , 2008 ) studied a group of same-sex twins at the ages of 7 years and 9 years and found a higher correlation between monozygotic twins than dizygotic twins on measures of antisocial behavior and levels of callous, unemotional behavior, which indicates a genetic component in development of these traits. No evidence was found that a shared environment played a role.

Role of Serotonin.

Several studies have found that serotonergic neurons play an inhibitory role in human aggression. For example, a depressed rate of serotonin release (indicated by low levels of 5-HIAA in the CSF) are associated with aggression and other forms of antisocial behavior, including assault, arson, murder, and child beating (Lidberg et al., 1984 , 1985 ; Virkkunen et al., 1989 ). Coccaro et al. ( 1994 ) studied a group of men with personality disorders (including a history of impulsive aggression). They found that the men with the lowest serotonergic activity were most likely to have close relatives with a history of similar behavior problems.

If low levels of serotonin release contribute to aggression, perhaps drugs that act as serotonin agonists might help to reduce antisocial behavior. In fact, a study by Coccaro and Kavoussi ( 1997 ) found that fluoxetine (Prozac), a serotonin agonist, decreased irritability and aggressiveness, as measured by a psychological test. Joshua, the little boy described in the introduction to this subsection, came under the care of a psychiatrist who prescribed monoaminergic agonists and began a course of behavior therapy that managed to stem Joshua’s violent outbursts and risk-taking behaviors.

Role of the Ventromedial Prefrontal Cortex.

Many investigators believe that impulsive violence is a consequence of faulty emotional regulation. For most of us, frustrations may elicit an urge to respond emotionally, but we usually manage to calm ourselves and suppress these urges. As we shall see, the ventromedial prefrontal cortex plays an important role in regulating our responses to such situations. The analysis of social situations involves much more than sensory analysis; it involves experiences and memories, inferences and judgments. In fact, the skills involved include some of the most complex ones we possess. These skills are not localized in any one part of the cerebral cortex, although research does suggest that the right hemisphere is more important than the left. But the ventromedial prefrontal cortex—which includes the medial orbitofrontal cortex and the subgenual anterior cingulate cortex—plays a special role.

FIGURE 11.7 The Ventromedial Prefrontal Cortex

As we saw earlier in the discussion of the process of extinction, the ventromedial prefrontal cortex (vmPFC) plays a role in inhibiting emotional responses. The vmPFC is located just where its name suggests, at the bottom front of the cerebral hemispheres. (See Figure 11.7 . ) The vmPFC receives direct inputs from the dorsomedial thalamus, temporal cortex, ventral tegmental area, olfactory system, and amygdala. Its outputs go to several brain regions, including the cingulate cortex, hippocampal formation, temporal cortex, lateral hypothalamus, and amygdala. Finally, it communicates with other regions of the prefrontal cortex. Thus, its inputs provide it with information about what is happening in the environment and what plans are being made by the rest of the frontal lobes, and its outputs permit it to affect a variety of behaviors and physiological responses, including emotional responses organized by the amygdala.

FIGURE 11.8 Phineas Gage’s Accident

The steel rod entered his left cheek and exited through the top of his head.

The fact that the vmPFC plays an important role in control of emotional behavior is shown by the effects of damage to this region. The first—and most famous—case comes from the mid-1800s. Phineas Gage, the foreman of a railway construction crew, was using a steel rod to ram a charge of blasting powder into a hole that had been drilled in solid rock. Suddenly, the charge exploded and sent the rod into his cheek, through his brain, and out the top of his head. (See Figure 11.8 . ) He survived, but he was a different man. Before his injury he was serious, industrious, and energetic. Afterward, he became childish, irresponsible, and thoughtless of others. His outbursts of temper led some people to remark that it looked as if Dr. Jekyll had become Mr. Hyde. He was unable to make or carry out plans, and his actions appeared to be capricious and whimsical. His accident had largely destroyed the vmPFC bilaterally (Damasio et al., 1994 ).

People whose vmPFC has been damaged by disease or accident are still able to accurately assess the significance of particular situations but only in a theoretical sense. For example, Eslinger and Damasio ( 1985 ) found that a patient with bilateral damage of the vmPFC (produced by a benign tumor, which was successfully removed) displayed excellent social judgment. When he was given hypothetical situations that required him to make decisions about what the people involved should do—situations involving moral, ethical, or practical dilemmas—he always gave sensible answers and justified them with carefully reasoned logic. However, his own life was a different matter. He frittered away his life’s savings on investments that his family and friends pointed out were bound to fail. He lost one job after another because of his irresponsibility. He became unable to distinguish between trivial decisions and important ones, spending hours trying to decide where to have dinner but failing to use good judgment in situations that concerned his occupation and family life. (His wife finally left him and sued for divorce.) As the authors noted, “He had learned and used normal patterns of social behavior before his brain lesion, and although he could recall such patterns when he was questioned about their applicability, real-life situations failed to evoke them” (p. 1737). Evidence suggests that the vmPFC serves as an interface between brain mechanisms involved in automatic emotional responses (both learned and unlearned) and those involved in the control of complex behaviors. This role includes using our emotional reactions to guide our behavior and in controlling the occurrence of emotional reactions in various social situations.

Mr. V., described in the opener to this chapter, had brain damage that impaired his judgment without affecting traditional measures of verbal intelligence. This damage included the right frontal and parietal lobes, so we cannot attribute his symptoms to any single region.

Damage to the vmPFC causes serious and often debilitating impairments of behavioral control and decision making. These impairments appear to be a consequence of emotional dysregulation. Anderson et al. ( 2006 ) obtained ratings of emotional behaviors of patients with lesions of the vmPFC, such as frustration tolerance, emotional instability, anxiety, and irritability, from the patients’ relatives. They also obtained ratings of the patients’ real-world competencies, such as judgment, planning, social inappropriateness, and financial and occupational status, from both relatives and clinicians. They found a significant correlation between emotional dysfunction and impairments in real-world competencies. There was no relationship between cognitive abilities and real-world competencies, which strongly suggests that emotional problems lie at the base of the real-world difficulties exhibited by people with vmPFC damage.

An interesting functional imaging study by Nili et al. ( 2010 ) suggests that the vmPFC plays a role in brain mechanisms of courage. Nili and his colleagues scanned the brains of people who were or were not afraid of snakes. While the people were in the scanner, they could press buttons that controlled the action of a conveyer belt that brought a live snake toward or away from them. People who were not afraid of snakes brought the snake near them and showed no signs of fear. However, people who were afraid of snakes did show signs of fear as the snake approached. Some of the fearful people pressed the button that moved the snake away from them, but others brought the snake near to them, even though they were plainly afraid. In other words, they showed courage—they overcame their fear. A display of courage was accompanied by activation of a region of the vmPFC, the subgenual anterior cingulate cortex (sgACC). Subjects who succumbed to their fear—that is, those who moved the snake away—did not show activation of the sgACC.

Evidence suggests that emotional reactions guide moral judgments as well as decisions involving personal risks and rewards and that the prefrontal cortex plays a role in these judgments as well. In previous years, moral judgments were believed to be a product of conscious, rational decision making. However, recent research on the role of neural mechanisms of emotion suggest that emotions play an important role—perhaps the most important role—in the formation of moral judgments.

Consider the following moral dilemma (Thomson, 1986 ): You see a runaway trolley with five people aboard hurtling down a track leading to a cliff. Without your intervention these people will soon die. However, you are standing near a switch that will shunt the trolley off to another track, where the vehicle will stop safely. But a worker is standing on that track, and he will be killed if you throw the switch to save the five helpless passengers. Should you stand by and watch the trolley go off the cliff, or should you save them—and kill the man on the track?

Most people conclude that the better choice would be to throw the switch; saving five people justifies the sacrifice of one man. This decision is based on conscious, logical application of a rule that it is better to kill one person than five people. But consider a variation of this dilemma. As before, the trolley is hurtling toward doom, but there is no switch at hand to shunt it to another track. Instead, you are standing on a bridge over the track. An obese man is standing there too, and if you give him a push, his body will fall onto the track and stop the trolley. (You are too small to stop the trolley, so you cannot save the five people by sacrificing yourself.) What should you do?

Most people feel repugnance at the thought of pushing the man off the bridge and balk at the idea of doing so, even though the result would be the same as the first dilemma: one person lost, five people saved. Whether we kill someone by sending a trolley his way or by pushing him off a bridge into the path of an oncoming trolley, he dies when the trolley strikes him. But somehow, imagining yourself pushing a person’s body and causing his death seems more emotionally wrenching than throwing a switch that changes the course of a runaway trolley. Thus, moral judgments appear to be guided by emotional reactions and are not simply the products of a rational, logical decision-making processes.

In a functional-imaging study, Greene et al. ( 2001 ) presented people with moral dilemmas such as the one I just described and found that thinking about them activated several brain regions involved in emotional reactions, including the vmPFC. Making innocuous decisions, such as whether to take a bus or train to some destination, did not activate these regions. Perhaps, then, our reluctance to push someone to his death is guided by the unpleasant emotional reaction we feel when we contemplate this action.

Let’s reconsider the contrast between the decision to throw a switch to save four lives and the decision to shove someone onto the tracks to accomplish the same goal. Consideration of the first dilemma raises a much smaller emotional reaction than consideration of the second dilemma does, and consideration of only the second dilemma strongly activates the vmPFC. We might expect that people with vmPFC damage, who show impairments of emotional reaction, might chose to push the man onto the track in the second dilemma. In fact, that is exactly what they do. They demonstrate utilitarian moral judgment: Killing one person is better than letting five people get killed. Koenigs et al. ( 2007 ) presented nonmoral, impersonal moral, and personal moral scenarios to patients with vmPFC lesions, patients with brain damage not including this region, and normal controls. For example, the switch-throwing scenario we just considered is an impersonal moral dilemma, and the person-pushing dilemma is a personal moral dilemma. Table 11.1 lists examples of the scenarios that the investigators of the study presented to their subjects. (See Table 11.1 . )

Koenigs and his colleagues predicted that patients with vmPFC lesions should make the same decisions as subjects in the other two groups on the nonmoral and impersonal moral judgments, because these decisions are normally solved rationally, without a strong emotional component. Only the outcome, or utility, of the choice need be considered. However, the emotional deficits of patients with prefrontal damage would be expected to lead to utilitarian judgments even in the case of personal moral judgments—and that is precisely what was seen. Figure 11.9 shows the proportion of the subjects from each of the three groups who endorsed a decision to act in high-conflict personal moral dilemmas, such as the lifeboat scenario. As you can see, the patients with vmPFC lesions were much more likely to say “yes” to the question posed at the end of the scenarios. (See Figure 11.9 . )

TABLE 11.1 Examples of Scenarios Involving Nonmoral, Impersonal Moral, and Personal Moral Judgments from the Study by Koenigs et al. (2007)

Brownies (Nonmoral scenario)

You have decided to make a batch of brownies for yourself. You open your recipe book and find a recipe for brownies. The recipe calls for a cup of chopped walnuts. You don’t like walnuts, but you do like macadamia nuts. As it happens, you have both kinds of nuts available to you.

Would you substitute macadamia nuts for walnuts in order to avoid eating walnuts?

Speedboat (Impersonal moral scenario)

While on vacation on a remote island, you are fishing from a seaside dock. You observe a group of tourists board a small boat and set sail for a nearby island. Soon after their departure you hear over the radio that there is a violent storm brewing, a storm that is sure to intercept them. The only way that you can ensure their safety is to warn them by borrowing a nearby speedboat. The speedboat belongs to a miserly tycoon who would not take kindly to your borrowing his property.

Would you borrow the speedboat in order to warn the tourists about the storm?

Lifeboat (Personal moral scenario)

You are on a cruise ship when there is a fire on board, and the ship has to be abandoned. The lifeboats are carrying many more people than they were designed to carry. The lifeboat you’re in is sitting dangerously low in the water—a few inches lower, and it will sink. The seas start to get rough, and the boat begins to fill with water. If nothing is done it will sink before the rescue boats arrive, and everyone on board will die. However, there is an injured person who will not survive in any case. If you throw that person overboard the boat will stay afloat and the remaining passengers will be saved.

Would you throw this person overboard in order to save the lives of the remaining passengers?

I mentioned earlier that our reluctance to push someone to his death even though that action would save the lives of others might be caused by the unpleasant thought of what it would feel like to commit that action and see the man fall to his death. If this is true, then perhaps people with vmPFC lesions say that they are willing to push the man off the bridge because thinking about doing so does not evoke an unpleasant emotional reaction. In fact, Moretto et al. ( 2009 ) found that people without brain damage showed physiological signs of an unpleasant emotional reaction when they contemplated pushing the man off the bridge—and said that they would not push him. People with vmPFC lesions did not show signs of this emotional reaction—and said that they would push him.

FIGURE 11.9 Moral Decisions and the vmPFC

The graph shows the percentage of people with lesions of the ventromedial prefrontal cortex and normal controls who endorse decisions of nonmoral, impersonal moral, and personal moral scenarios like the ones listed in Table 11.1 .

(Based on data from Keonigs et al., 2007.)

It might seem that I have been getting away from the topic of this section: anger and aggression. However, recall that many investigators believe that impulsive violence is a consequence of faulty emotional regulation. The amygdala plays an important role in provoking anger and violent emotional reactions, and the prefrontal cortex plays an important role in suppressing such behavior by making us see its negative consequences. The amygdala matures early in development, but the prefrontal cortex matures much later, during late childhood and early adulthood. As the prefrontal cortex matures, adolescents show increases in speed of cognitive processing, abstract reasoning ability, ability to shift attention from one topic to another, and ability to inhibit inappropriate responses (Yurgelun-Todd, 2007 ). In fact, a structural imaging study by Whittle et al. ( 2008 ) found that aggressive behavior during parent–child interactions during adolescence was positively related to the volume of the amygdala and negatively related to the relative volume of the right medial prefrontal cortex.

Raine et al. ( 1998 ) found evidence of decreased prefrontal activity and increased subcortical activity (including the amygdala) in the brains of convicted murderers. These changes were primarily seen in impulsive, emotional murderers. Cold-blooded, calculating, predatory murderers—whose crimes were not accompanied by anger and rage—showed more normal activity. Presumably, increased activation of the amygdala reflected an increased tendency for display of negative emotions, and the decreased activation of the prefrontal cortex reflected a decreased ability to inhibit the activity of the amygdala and thus control the people’s emotions. Raine et al. (2002) found that people with antisocial personality disorder showed an 11 percent reduction in volume of the gray matter of the prefrontal cortex.

Earlier in this chapter we saw that decreased activity of serotonergic neurons is associated with aggression, violence, and risk taking. As we saw in this subsection, decreased activity of the prefrontal cortex is also associated with antisocial behavior. These two facts appear to be linked. The prefrontal cortex receives a major projection of serotonergic axons. Research indicates that serotonergic input to the prefrontal cortex activates this region. For example, a functional-imaging study by Mann et al. ( 1996 ) showed that fenfluramine, a drug that stimulates the release of 5-HT, increases the activity of the prefrontal cortex, which presumably inhibits the activity of the amygdala and suppresses aggressive behavior. Crockett et al. ( 2010 ) found that a single high dose of a 5-HT agonist decreased the likelihood of subjects making a decision to cause harm in scenarios that presented moral dilemmas. In other words, the increased serotonergic activity made them less likely to make utilitarian decisions. It seems likely, therefore, that an abnormally low level of serotonin release can result in decreased activity of the prefrontal cortex and increased likelihood of utilitarian judgments or, in the extreme, antisocial behavior.

Several studies have found evidence for deficits in serotonergic innervation of the ventromedial prefrontal cortex. New et al. ( 2002 ) found that a serotonin-releasing drug increased the activity of the orbitofrontal cortex in normal, nonviolent subjects but failed to do so in subjects with a history of impulsive aggression. A functional-imaging study found evidence for lower levels of serotonin transporters in the medial prefrontal cortex of people with impulsive aggression (Frankle et al., 2005 ). Because serotonin transporters are found in the membrane of serotonergic terminal buttons, this study suggests that the medial prefrontal cortex of these people contains decreased serotonergic input.

As we saw earlier, impulsive aggression has been successfully treated with specific serotonin reuptake inhibitors such as fluoxetine (Prozac). A functional-imaging study by New et al. (2004) measured regional brain activity of people with histories of impulsive aggression before and after 12 weeks of treatment with fluoxetine. They found that the drug increased the activity of the prefrontal cortex and reduced aggressiveness.

Hormonal Control of Aggressive Behavior

As we saw, many instances of aggressive behavior are in some way related to reproduction. For example, males of some species establish territories that attract females during the breeding season. To do so, they must defend the territories against the intrusion of other males. Even in species in which breeding does not depend on the establishment of a territory, males may compete for access to females, competition that also involves aggressive behavior. Females, too, often compete with other females for space in which to build nests or dens in which to rear their offspring, and they will defend their offspring against the intrusion of other animals. As you learned in Chapter 10 , most reproductive behaviors are controlled by the organizational and activational effects of hormones; therefore, we should not be surprised that many forms of aggressive behavior are, like mating, affected by hormones.

AGGRESSION IN MALES

Adult males of many species fight for territory or access to females. In laboratory rodents, androgen secretion occurs prenatally, decreases, and then increases again at the time of puberty. Intermale aggressiveness also begins around the time of puberty, which suggests that the behavior is controlled by neural circuits that are stimulated by androgens. Indeed, many years ago, Beeman ( 1947 ) found that castration reduced aggressiveness and that injections of testosterone reinstated it.

In Chapter 10 we saw that early androgenization has an organizational effect. The secretion of androgens early in development modifies the developing brain, making neural circuits that control male sexual behavior become more responsive to testosterone. Similarly, early androgenization has an organizational effect that stimulates the development of testosterone-sensitive neural circuits that facilitate intermale aggression. (See Figure 11.10 . )

The organizational effect of androgens on intermale aggression (aggressive displays or actual fights between two males of the same species) is important, but it is not an all-or-none phenomenon. Prolonged administration of testosterone will eventually induce intermale aggression even in rodents that were castrated immediately after birth. Data reviewed by vom Saal ( 1983 ) show that exposure to androgens early in life decreases the amount of exposure that is necessary to activate aggressive behavior later in life. Thus, early androgenization sensitizes the neural circuits: The earlier the androgenization, the more effective is the sensitization.

FIGURE 11.10 Organizational and Activational Effects of Testosterone on Social Aggression

We also saw in Chapter 10 that androgens stimulate male sexual behavior by interacting with androgen receptors in neurons located in the medial preoptic area (MPA). This region also appears to be important in mediating the effects of androgens on intermale aggression. Bean and Conner ( 1978 ) found that implanting testosterone in the MPA reinstated intermale aggression in castrated male rats. Presumably, the testosterone directly activated the behavior by stimulating the androgen-sensitive neurons located there. The medial preoptic area, then, appears to be involved in several behaviors related to reproduction: male sexual behavior, maternal behavior, and intermale aggression.

Males readily attack other males but usually do not attack females. Their ability to discriminate the sex of the intruder appears to be based on the presence of particular pheromones. Bean ( 1982 ) found that intermale aggression was abolished in mice by cutting the vomeronasal nerve, which deprives the brain of input from the vomeronasal organ. And if the urine of female mice is painted on a male mouse, he will not be attacked if he is introduced into another male’s cage (Dixon and Mackintosh, 1971 ; Dixon, 1973 ). Stowers et al. ( 2002 ) found that a targeted mutation against a protein that is essential for the detection of pheromones by the vomeronasal organ abolishes a male mouse’s ability to discriminate between males and females. Because male intruders were not recognized as rival males, they were not attacked. In fact, the mice with the targeted mutation attempted to copulate with the intruders. ( Simulatepheromones and sexual recognition on MyPsychLab.)

AGGRESSION IN FEMALES

Two adult female rodents that meet in a neutral territory are less likely than males to fight. But aggression between females, like aggression between males, appears to be facilitated by testosterone. Van de Poll et al. ( 1988 ) ovariectomized female rats and then gave them daily injections of testosterone, estradiol, or a placebo for 14 days. The animals were then placed in a test cage, and an unfamiliar female was introduced. As Figure 11.11 shows, testosterone increased aggressiveness, whereas estradiol had no effect. (See Figure 11.11 . )

Androgens have an organizational effect on the aggressiveness of females, and a certain amount of prenatal androgenization appears to occur naturally. Most rodent fetuses share their mother’s uterus with brothers and sisters, arranged in a row like peas in a pod. A female mouse may have zero, one, or two brothers adjacent to her. Researchers refer to these females as 0M, 1M, or 2M, respectively. (See Figure 11.12 . ) Being next to a male fetus has an effect on a female’s blood levels of androgens prenatally. Vom Saal and Bronson ( 1980 ) found that females located between two males had significantly higher levels of testosterone in their blood than did females located between two females (or between a female and the end of the uterus). When they are tested as adults, 2M females are more likely to exhibit interfemale aggressiveness.

FIGURE 11.11 Effects of Estradiol and Testosterone on Interfemale Aggression in Rats

(Based on data from van de Poll et al., 1988.)

Females of some primate species (for example, rhesus monkeys and baboons) are more likely to engage in fights around the time of ovulation (Carpenter, 1942 ; Saayman, 1971 ). This phenomenon is probably caused by their increased sexual interest and consequent proximity to males. As Carpenter noted, “She actively approaches males and must overcome their usual resistance to close association, hence she becomes an object of attacks by them” (p. 136). Another period of fighting occurs just before menstruation (Sassenrath, Powell, and Hendrickx, 1973 ; Mallow, 1979 ). During this time, females tend to attack other females.

FIGURE 11.12 0M, 1M, and 2M Female Mouse Fetuses

(Adapted from vom Saal, F. S., in Hormones and Aggressive Behavior, edited by B. B. Svare. New York: Plenum Press, 1983.)

EFFECTS OF ANDROGENS ON HUMAN AGGRESSIVE BEHAVIOR

Boys are generally more aggressive than girls. Clearly, most societies tolerate assertiveness and aggressive behavior from boys more than they do from girls. Without doubt, the way we treat boys and girls and the models to which we expose them play important roles in sex differences in aggressiveness in our species. The question is not whether socialization has an effect (certainly, it does) but whether biological influences, such as exposure to androgens, have an effect too.

Prenatal androgenization increases aggressive behavior in all species that have been studied, including primates. Therefore, if androgens did not affect aggressive behavior in humans, our species would be exceptional. After puberty, androgens also begin to have activational effects. Boys’ testosterone levels begin to increase during the early teens, at which time aggressive behavior and intermale fighting also increase (Mazur, 1983 ). Of course, boys’ social status changes during puberty, and their testosterone affects their muscles as well as their brains, so we cannot be sure that the effect is hormonally produced or, if it is, that it is mediated by the brain.

As we just saw, the small amount of prenatal exposure to androgens that a 2M female receives has measurable organizational effects on aggressive behavior. Cohen-Bendahan et al. ( 2005 ) compared the proneness to aggression in 13-year-old female dizygotic twins who had shared the uterus with a brother (1M females) with those who had shared it with a sister (1F females). They found a modest but statistically significant increase in aggressiveness in the 1M girls. The testosterone levels of the 1M and 1F girls did not differ, so the increased aggressiveness presumably was a result of increased prenatal exposure to androgens. Of course, we cannot rule out the possibility that being raised with a same-aged brother might have an effect on a girl’s proneness to aggression.

As we saw in Chapter 10 , girls with congenital adrenal hyperplasia (CAH) are exposed to abnormally high levels of androgens—produced by their own adrenal glands—during prenatal development. The effects of this exposure include a preference for boys as playmates, interest in toys and games that boys typically prefer, and, in adulthood, an increased prevalence of sexual attraction to other women. Berenbaum and Resnick ( 1997 ) found that women and adolescent girls with CAH displayed higher levels of aggression, as measured by parents’ ratings and paper-and-pencil tests.

Scientifically rigorous evidence concerning the activational effects of androgens on aggression in adulthood is difficult to obtain in humans. Obviously, we cannot randomly castrate some men to find out whether their aggressiveness declines. In the past, authorities attempted to suppress sex-related aggression by castrating convicted male sex offenders. Investigators have reported that both heterosexual and homosexual aggressive attacks disappear, along with the offender’s sex drive (Hawke, 1951 ; Sturup, 1961 ; Laschet, 1973 ). However, the studies typically lack appropriate control groups and usually do not measure aggressive behavior directly.

Some cases of aggressiveness, especially sexual assault, have been treated with synthetic steroids that inhibit the production of androgens by the testes. Clearly, treatment with drugs is preferable to castration, because the effects of drugs are not irreversible. However, the efficacy of treatment with antiandrogens has yet to be established conclusively. According to Walker and Meyer ( 1981 ), these drugs decrease sex-related aggression but have no effect on other forms of aggression. In fact, Zumpe et al. ( 1991 ) found that one of these drugs decreased sexual activity and aggression toward females when administered to male monkeys but actually increased intermale aggression.

Another way to determine whether androgens affect aggressiveness in humans is to examine the testosterone levels of people who exhibit varying levels of aggressive behavior. However, even though this approach poses fewer ethical problems, it presents methodological ones. First, let me review some evidence. In a review of the literature, Archer ( 1994 ) found that most studies found a positive relationship between men’s testosterone levels and their level of aggressiveness. For example, Dabbs and Morris ( 1990 ) studied 4462 U.S. military veterans. The men with the highest testosterone levels had records of more antisocial activities, including assaults on other adults and histories of more trouble with parents, teachers, and classmates during adolescence. The largest effects were seen in men of lower socioeconomic status.

Mazur and Booth ( 1998 ) suggest that the primary social effect of androgens may be not on aggression, but on dominance. If androgens enhance motivation to dominate others, that motivation may sometimes lead to aggression but not in all situations. For example, a person might strive to defeat others symbolically (through athletic competition or acquisition of symbols of status) rather than through direct aggression.

In any event we must remember that correlation does not necessarily indicate causation. A person’s environment can affect his or her testosterone level. For example, losing a tennis match or a wrestling competition causes a fall in blood levels of testosterone (Mazur and Lamb, 1980 ; Elias, 1981 ). Even winning or losing a simple game of chance carried out in a psychology laboratory can affect participants’ testosterone levels: Winners feel better afterward and have a higher level of testosterone (McCaul, Gladue, and Joppa, 1992 ). Bernhardt et al. ( 1998 ) found that basketball and soccer fans showed an increase in testosterone levels if their team won and a decrease if it lost. Thus, we cannot be sure in any correlational study that high testosterone levels cause people to become dominant or aggressive; perhaps their success in establishing a position of dominance increases their testosterone levels relative to those of the people they dominate.

As news reports have publicized, some athletes take anabolic steroids to increase their muscle mass and strength and, supposedly, to increase their competitiveness. Anabolic steroids include natural androgens and synthetic hormones with androgenic effects. Thus, we might expect that these hormones would increase aggressiveness. Indeed, several studies have found exactly that. For example, Yates, Perry, and Murray ( 1992 ) found that male weight lifters who were taking anabolic steroids were more aggressive and hostile than those who were not. But, as the authors note, we cannot be certain that the steroid is responsible for the increased aggressiveness; it could simply be that the men who were already more competitive and aggressive were the ones who chose to take the steroids.

An interesting set of experiments with another species of primates might have some relevance to human aggression. Evidence suggests that the effects of alcohol may interact with those of androgens. Winslow and Miczek ( 1985 , 1988 ) found that alcohol increases intermale aggression in dominant male squirrel monkeys, but only during the mating season, when their blood level of testosterone is two to three times higher than that during the nonmating season. These studies suggest that the effects of alcohol interact both with social status and with testosterone. (See Figure 11.13 . ) This suggestion was confirmed by Winslow, Ellingoe, and Miczek ( 1988 ), who tested monkeys during the nonmating season. They found that alcohol increased the aggressive behavior of dominant monkeys if the monkeys were also given injections of testosterone. However, these treatments were ineffective in subordinate monkeys, which presumably had learned not to be aggressive. The next step will be to find the neural mechanisms that are responsible for these interactions.

FIGURE 11.13 Alcohol, Mating, and Aggressive Behavior in Monkeys

The graph shows the effect of alcohol intake on frequency of aggressive behavior of dominant and subordinate male squirrel monkeys during the mating season and the nonmating season.

(Based on data from Winslow and Miczek, 1988.)

SECTION SUMMARY: Emotions as Response Patterns

The word emotion refers to behaviors, physiological responses, and feelings. This section has discussed emotional response patterns, which consist of behaviors that deal with particular situations and physiological responses (both autonomic and hormonal) that support the behaviors. The amygdala organizes behavioral, autonomic, and hormonal responses to a variety of situations, including those that produce fear, anger, or disgust. In addition, it is involved in the effects of odors and pheromones on sexual and maternal behavior. It receives inputs from the olfactory system, the association cortex of the temporal lobe, the frontal cortex, and the rest of the limbic system. Its outputs go to the frontal cortex, hypothalamus, hippocampal formation, and brain stem nuclei that control autonomic functions and some species-typical behaviors. Electrical recordings of single neurons in the amygdala indicate that some of them respond when the animal perceives particular stimuli with emotional significance. Stimulation of the amygdala leads to emotional responses, and its destruction disrupts them. Pairing of neutral stimuli with those that elicit emotional responses results in classically conditioned emotional responses. Learning of these responses takes place primarily in the amygdala. Extinction of conditioned emotional responses involves inhibitory control of amygdala activity by the ventromedial prefrontal cortex.

Studies of people with amygdala lesions and functional-imaging studies with humans indicate that the amygdala is involved in emotional reactions in our species, too. However, many of our conditioned emotional responses are acquired by observing the responses of other people or even through verbal instruction. Study of people with amygdala lesions and functional-imaging studies indicate that the amygdala is involved in the effects of emotions on learning.

Aggressive behaviors are species-typical and serve useful functions most of the time. In addition, animals may exhibit threat or submissive behaviors, which may avoid an actual fight. The periaqueductal gray matter appears to be involved in defensive behavior and predation. These mechanisms are modulated by the hypothalamus and amygdala.

The activity of serotonergic neurons appears to inhibit risk-taking behaviors, including aggression. Destruction of serotonergic axons in the forebrain enhances aggression, and administration of drugs that facilitate serotonergic transmission reduces it. Low CSF levels of 5-HIAA (a metabolite of serotonin) are correlated with increased risk-taking and aggressive behavior in monkeys and humans. The presence of short alleles of the serotonin transporter gene increase the likelihood of a reactive amygdala and development of depression and anxiety disorders. Genetic factors play a role in people’s level of aggression and antisocial behavior.

The ventromedial prefrontal cortex plays an important role in emotional reactions. This region communicates with the dorsomedial thalamus, temporal cortex, ventral tegmental area, olfactory system, amygdala, cingulate cortex, lateral hypothalamus, and other regions of the frontal cortex, including the dorsolateral prefrontal cortex. People with ventromedial prefrontal lesions show impulsive behavior and often display outbursts of inappropriate anger. They are able to explain the implications of complex social situations but often respond inappropriately when they find themselves in these situations. The activity of the vmPFC increases where people show courageous behavior—letting a snake approach them even though they fear snakes.

Evidence indicates that the ventromedial prefrontal cortex is involved in making moral judgments. When people make judgments that involve conflicts between utilitarian judgments (one person dies but five people live) and personal moral judgments (are you willing to push a man to his death to save others?), the ventromedial prefrontal cortex is activated. People with damage to the vmPFC display utilitarian moral judgments. Evidence suggests that nonpsychopathic people are reluctant to harm others because thinking about doing so produces an unpleasant emotional reaction. The release of serotonin in the prefrontal cortex activates this region, and some investigators believe that the serotonergic input to this region is responsible for the ability of serotonin to inhibit aggression and risky behavior. A single high dose of a 5-HT agonist decreases the likelihood that a person will make utilitarian decisions in a moral dilemma task. The ventromedial prefrontal cortex of people with impulsive aggression contains less dense serotonergic input.

Because many aggressive behaviors are related to reproduction, these behaviors are influenced by hormones, especially sex steroid hormones. Androgens primarily affect offensive attack; they are not necessary for defensive behaviors, which are shown by females as well as males. In males, androgens have organizational and activational effects on offensive attack, just as they have on male sexual behavior. The effects of androgens on intermale aggression appear to be mediated by the medial preoptic area.

Female rodents will fight when they meet in neutral territory but less often than males. Female rodents that have been slightly androgenized (2M females) are more likely to attack other females. Female primates are most likely to fight around the time of ovulation, perhaps because their increased sexual interest brings them closer to males.

Androgens apparently promote aggressive behavior in humans, but this topic is more difficult to study in our species than in laboratory animals. Evidence from girls with congenital adrenal hyperplasia and female dizygotic twins who shared the uterus with a brother suggests that prenatal androgens promote aggressive behavior later in life. Research suggests that the primary effect of androgens may be to increase motivation to achieve dominance and that increased aggression may be secondary to this effect. In any case we cannot be sure whether higher androgen levels promote dominance or whether successful dominance increases androgen levels. Studies with monkeys suggest that testosterone and alcohol have synergistic effects, particularly in dominant animals. (Synergy, from a Greek word meaning “working together,” refers to combinations of factors that are more effective than the sum of their individual actions.) Perhaps these effects are related to our observations that some men with a history of violent behavior become more aggressive when they drink.

▪ THOUGHT QUESTIONS

Phobias can be seen as dramatic examples of conditioned emotional responses. These responses can even be contagious; we can acquire them without direct experience with an aversive stimulus. For example, a child who sees a parent show signs of fright in the presence of a dog may also develop a fear reaction to the dog. Do you think that some prejudices might be learned in this way, too?

Suppose a man sustained brain damage that destroyed his ventral prefrontal cortex, and soon afterward he began exhibiting antisocial behavior. One day, while he was in his car waiting for a red light to change, he saw a man with whom he had previously had a violent argument cross the street in front of him. He suddenly stepped on the accelerator, struck the man, and killed him. Should the fact of his brain damage play any role in his prosecution and judgment in court? Why or why not?

From the point of view of evolution, aggressive behavior and a tendency to establish dominance have useful functions. In particular, they increase the likelihood that only the healthiest and most vigorous animals will reproduce. Can you think of examples of good and bad effects of these tendencies among members of our own species?

Communication of Emotions

The previous section described emotions as organized responses (behavioral, autonomic, and hormonal) that prepare an animal to deal with existing situations in the environment, such as events that pose a threat to the organism. For our earliest premammalian ancestors that is undoubtedly all there was to emotions. But over time, other responses, with new functions, evolved. Many species of animals (including our own) communicate their emotions to others by means of postural changes, facial expressions, and nonverbal sounds (such as sighs, moans, and growls). These expressions serve useful social functions: They tell other individuals how we feel and—more to the point—what we are likely to do. For example, they warn a rival that we are angry or tell friends that we are sad and would like some comfort and reassurance. They can also indicate that a danger might be present or that something interesting seems to be happening. This section examines such expression and communication of emotions.

Facial Expression of Emotions: Innate Responses

Charles Darwin ( 1872/1965 ) suggested that human expressions of emotion have evolved from similar expressions in other animals. He said that emotional expressions are innate, unlearned responses consisting of a complex set of movements, principally of the facial muscles. Thus, a person’s sneer and a wolf’s snarl are biologically determined response patterns, both controlled by innate brain mechanisms, just as coughing and sneezing are. (Of course, people can sneer and wolves can snarl for quite different reasons.) Some of these movements resemble the behaviors themselves and may have evolved from them. For example, a snarl shows one’s teeth and can be seen as an anticipation of biting.

Darwin obtained evidence for his conclusion that emotional expressions were innate by observing his own children and by corresponding with people living in various isolated cultures around the world. He reasoned that if people all over the world, no matter how isolated, show the same facial expressions of emotion, then these expressions must be inherited instead of learned. The logical argument goes like this: When groups of people are isolated for many years, they develop different languages. Thus, we can say that the words people use are arbitrary; there is no biological basis for using particular words to represent particular concepts. However, if facial expressions are inherited, then they should take approximately the same form in people from all cultures, despite their isolation from one another. Darwin did, indeed, find that people in different cultures used the same patterns of movement of facial muscles to express a particular emotional state.

Research by Ekman and his colleagues (Ekman and Friesen, 1971 ; Ekman, 1980 ) tends to confirm Darwin’s hypothesis that facial expression of emotion uses an innate, species-typical repertoire of movements of facial muscles (Darwin, 1872/1965 ). For example, Ekman and Friesen ( 1971 ) studied the ability of members of an isolated tribe in New Guinea to recognize facial expressions of emotion produced by Westerners. They had no trouble doing so and they themselves produced facial expressions that Westerners readily recognized. Figure 11.14 shows four photographs taken from videotapes of a man from this tribe reacting to stories designed to evoke facial expressions of happiness, sadness, anger, and disgust. I am sure that you will have no trouble recognizing which is which. (See Figure 11.14 . )

Because the same facial expressions were used by people who had not previously been exposed to each other, Ekman and Friesen concluded that the expressions were unlearned behavior patterns. In contrast, different cultures use different words to express particular concepts; production of these words does not involve innate responses but must be learned.

FIGURE 11.14 Facial Expressions in a New Guinea Tribesman

The tribesman made faces when he heard the following stories: (a) “Your friend has come and you are happy.” (b) “Your child had died.” (c) “You are angry and about to fight.” (d) “You see a dead pig that has been lying there a long time.”

(From Ekman, P., The Face of Man: Expressions of Universal Emotions in a New Guinea Village. New York: Garland STPM Press, 1980. Reprinted with permission.)

Other researchers have compared the facial expressions of blind and normally sighted people. They reasoned that if the facial expressions of the two groups are similar, then the expressions are natural for our species and do not require learning by imitation. In fact, the facial expressions of young blind and sighted children are very similar (Woodworth and Schlosberg, 1954 ; Izard, 1971 ). In addition, a study of the emotional expressions of people competing (and winning or losing) athletic events in the 2004 Paralympic Games found no differences between the expressions of congenitally blind, noncongenitally blind, and sighted athletes (Matsumoto and Willingham, 2009 ). Thus, both the cross-cultural studies and the investigations of blind people confirm the naturalness of these facial expressions of emotions.

A study by Sauter et al. ( 2010 ) reached similar conclusions. The investigators carried out a vocal version of the study by Ekman and Friesen. They presented European English-speakers and natives of isolated northern Namibian villages with recordings of sounds of nonverbal vocalizations to situations that would be expected to produce the emotions of anger, disgust, fear, sadness, surprise, or amusement. The participants were told a story and then heard two different vocalizations (sighs, groans, laughs, and so on), one of which would be appropriate for the emotion produced by the story. Members of both cultures had no difficulty choosing the correct vocalizations of members of their culture and the other culture. (See Figure 11.15 . )

FIGURE 11.15 Identification of Nonverbal Vocal Expressions of Emotions in a Different Culture

The graph shows the mean number of correct responses in identifying audio recordings of nonverbal emotional expressions of English speakers and Himba speakers (residents of northern Namibia) by speakers of English and Himba. Accuracy for cross-cultural identification was almost as good as intra-cultural identification.

(Based on data from Sauter et al., 2010.)

Neural Basis of the Communication of Emotions: Recognition

Effective communication is a two-way process. That is, the ability to display one’s emotional state by changes in expression is useful only if other people are able to recognize them. In fact, Kraut and Johnston ( 1979 ) unobtrusively observed people in circumstances that would be likely to make them happy. They found that happy situations (such as making a strike while bowling, seeing the home team score, or experiencing a beautiful day) produced only small signs of happiness when the people were alone. However, when the people were interacting socially with other people, they were much more likely to smile. For example, bowlers who made a strike usually did not smile when the ball hit the pins, but when they turned around to face their companions, they often smiled. Jones et al. ( 1991 ) found that even 10-month-old children showed this tendency. (No, I’m not suggesting that infants have been observed while bowling.)

Recognition of another person’s facial expression of emotions is generally automatic, rapid, and accurate. Tracy and Robbins ( 2008 ) found that observers quickly recognized brief expressions of a variety of emotions. If they were given more time to think about the expression they had seen, they showed very little improvement.

LATERALITY OF EMOTIONAL RECOGNITION

We recognize other people’s feelings by means of vision and audition—seeing their facial expressions and hearing their tone of voice and choice of words. Many studies have found that the right hemisphere plays a more important role than the left hemisphere in comprehension of emotion. For example, Bowers et al. ( 1991 ) found that patients with right hemisphere damage had difficulty producing or describing mental images of facial expressions of emotions. Subjects were asked to imagine the face of someone who was very happy (or very sad, angry, or afraid). Then they were asked questions about the facial expression—for example, Do the eyes look twinkly? Is the brow raised? Are the corners of the lips raised up? People with right hemisphere damage had trouble answering these questions but could easily answer questions about nonemotional images, such as What’s higher off the ground: a horse’s knee or the top of its tail? or What number from one to nine does a peanut look like?

Several functional-imaging studies have confirmed these results. For example, George et al. ( 1996 ) had subjects listen to some sentences and identify their emotional content. In one condition the subjects listened to the meaning of the words and said whether they described a situation in which someone would be happy, sad, angry, or neutral. In another condition the subjects judged the emotional state from the tone of the voice. The investigators found that comprehension of emotion from word meaning increased the activity of the prefrontal cortex bilaterally, the left more than the right. Comprehension of emotion from tone of voice increased the activity of only the right prefrontal cortex. (See Figure 11.16 . )

Heilman, Watson, and Bowers ( 1983 ) recorded a particularly interesting case of a man with a disorder called pure word deafness, caused by damage to the left temporal cortex. (This syndrome is described in Chapter 14 .) The man could not comprehend the meaning of speech but had no difficulty identifying the emotion being expressed by its intonation. This case, like the functional-imaging study by George et al. ( 1996 ), indicates that comprehension of words and recognition of tone of voice are independent functions.

ROLE OF THE AMYGDALA AND PREFRONTAL CORTEX

As we saw in the previous section, the amygdala plays a special role in emotional responses. It plays a role in emotional recognition as well. For example, several studies have found that lesions of the amygdala (the result of degenerative diseases or surgery for severe seizure disorders) impair people’s ability to recognize facial expressions of emotion, especially expressions of fear (Adolphs et al., 1994 , 1995 ; Young et al., 1995 ; Calder et al., 1996 ; Adolphs et al., 1999 ). In addition, functional-imaging studies (Morris et al., 1996 ; Whalen et al., 1998 ) have found large increases in the activity of the amygdala when people view photographs of faces expressing fear but only small increases (or even decreases) when they look at photographs of happy faces. However, although amygdala lesions impair visual recognition of facial expressions of emotion, several studies have found that they do not appear to affect people’s ability to recognize emotions in tone of voice (Anderson and Phelps, 1998 ; Adolphs and Tranel, 1999 ).

FIGURE 11.16 Perception of Emotions

The PET scans indicate brain regions activated by listening to emotions expressed by meanings of words (red) or tone of voice (green).

(Tracings of brain activity from George et al., 1996.)

Several studies suggest that the amygdala receives visual information that we use to recognize facial expressions of emotion directly from the thalamus and not from the visual association cortex. Adolphs ( 2002 ) notes that the amygdala receives visual input from two sources: subcortical and cortical. The subcortical input (from the superior colliculus and the pulvinar, a large nucleus in the posterior thalamus) appears to provide the most important information for this task. In fact, some people with blindness caused by damage to the visual cortex can recognize facial expressions of emotion even though they have no conscious awareness of looking at a person’s face, a phenomenon known as affective blindsight (de Gelder et al., 1999 ; Anders et al., 2004 ). Tamietto et al. ( 2009 ) confirmed that “emotional contagion” can take place even without conscious awareness. They presented photographs of faces expressing happiness or fearfulness to the sighted and blind fields of people with unilateral visual cortex lesions. Although the people did not report seeing an emotional expression (or even a face) in their blind fields, they automatically made facial expressions of their own that matched those of the faces in the photographs. Figure 11.17 shows the contraction of a “smile muscle” and a “frown muscle” under these conditions. Note that the appropriate muscles contracted when the happy or fearful face was presented to either the sighted field (“seen stimuli”) or the blind field (“unseen stimuli”). (See Figure 11.17 . )

affective blindsight The ability of a person who cannot see objects in his or her blind field to accurately identify facial expressions of emotion while remaining unconscious of perceiving them; caused by damage to the visual cortex.

People can express emotions through their body language, as well as through muscular movements of their face (de Gelder, 2006 ). For example, a clenched fist might accompany an angry facial expression, and a fearful person might run away. The sight of photographs of bodies posed in gestures of fear activates the amygdala, just as the sight of fearful faces does (Hadjikhani and de Gelder, 2003 ). Meeren, van Heijnsbergen, and de Gelder ( 2005 ) prepared computer-modified photographs of people showing facial expressions of emotions that were either congruent with the person’s body posture (for example, a facial expression of fear and a body posture of fear) or incongruent (for example, a facial expression of anger and a body posture of fear). The investigators asked people to identify the facial expressions shown in the photos and found that the ratings were faster and more accurate when the facial and body expressions were congruent. In other words, when we look at other people’s faces, our perception of their emotion is affected by their body posture as well as their facial expression.

As we saw in Chapter 6 , the visual cortex receives information from two systems of neurons. The magnocellular system (named for layers of large cells in the lateral geniculate nucleus of the thalamus that relay visual information from the eye to the visual cortex) provides information about movement, depth, and very subtle differences in brightness in the scene before our eyes. This system appeared early in evolution of the mammalian brain and provides most mammals (dogs and cats, for example) with a monochromatic, somewhat fuzzy image of the world. The parvocellular system (named for layers of small cells in the lateral geniculate nucleus) is found only in some primates, including humans. This system provides us with color vision and detection of fine details. The part of the visual association cortex responsible for recognition of faces, the fusiform face area, receives information primarily (but not solely) from the parvocellular system. The information that the amygdala receives from the superior colliculus and the pulvinar has its origin in the more primitive magnocellular system.

FIGURE 11.17 Unconscious Imitation of Facial Expressions of Emotions

When patients with unilateral damage to the visual cortex saw photographs of happy or fearful faces, they smiled or frowned when the photographs were presented to their sighted or blind field, which indicates that visual information concerning emotional expressions can take place without conscious awareness.

(Based on data from Tamietto et al., 2009.)

An ingenious functional-imaging study by Vuilleumier et al. ( 2003 ) presented people with pictures of faces showing neutral or fearful expressions. Some of the pictures were normal, some had been filtered with a computer program so that they showed only high spatial frequencies, and some had been filtered to show only low spatial frequencies. ( Chapter 6 described the concept of spatial frequencies.) As Figure 11.18 shows, high spatial frequencies show fine details of transitions between light and dark, and low spatial frequencies show fuzzy images. As you may have deduced, these photos primarily stimulate the parvocellular and magnocellular systems, respectively. (See Figure 11.18 . )

Vuilleumier and his colleagues found that the fusiform face area was better at recognizing individual faces and primarily used high spatial frequency (parvocellular) information to do so. In contrast, the amygdala (and the superior colliculus and pulvinar, which provide it with visual information) was able to recognize an expression of fear based on low spatial frequency (magnocellular) information but not on high spatial frequency information.

Krolak-Salmon et al. ( 2004 ) recorded electrical potentials from the amygdala and visual association cortex through electrodes that had been implanted in people who were being evaluated for neurosurgery to alleviate a seizure disorder. They presented the people with photographs of faces showing neutral expressions or expressions of fear, happiness, or disgust. They found that fearful faces produced the largest response and that the amygdala showed activity before the visual cortex did. The rapid response supports the conclusion that the amygdala receives visual information from the magnocellular system (which conducts information very rapidly) that permits it to recognize facial expressions of fear.

FIGURE 11.18 Involvement of Magnocellular and Parvocellular Systems in Emotional Perception

The figure shows the stimuli used by Vuilleumier et al. ( 2003 ). The more primitive magnocellular system is sensitive to low spatial frequencies (SF), and the more recently evolved parvocellular system is sensitive to high spatial frequencies.

(From Vuilleumier, P., Armony, J. L., Driver, J., and Dolan, R. J. Nature Neuroscience, 2003, 6, 624–631.)

So far, the evidence suggests that the amygdala plays an indispensable role in recognition of facial expressions of fear. However, a study by Adolphs et al. ( 2005 ) suggests that, under the appropriate conditions, other regions of the brain can perform this task. Adolphs and his colleagues discovered that S. M., a woman with bilateral amygdala damage, failed to look at the eyes when she examined photographs of faces. Spezio et al. ( 2007 ) conducted a similar study, but this one measured S. M.’s eye movements while she was actually conversing with another person. Like the study by Adolphs et al., this one found that S. M. failed to direct her gaze to the other person’s eyes. In contrast, she spent an abnormally large amount of time looking at the other person’s mouth. (See Figure 11.19 . )

By themselves, eyes are able to convey a fearful expression. (See Figure 11.20 . ) A functional-imaging study by Whalen et al. ( 2004 ) found that viewing the fearful eyes shown in Figure 11.20 activated the ventral amygdala, the region that receives the majority of the cortical and subcortical inputs to the amygdala. So the fact that S. M. did not look at eyes suggests a cause for her failure to detect only this emotion. In fact, when Adolphs et al. ( 2005 ) instructed S. M. to look at the eyes of the face she was examining, she was able to recognize an expression of fear. However, unless she was reminded to do so, she soon stopped looking at eyes, and her ability to recognize a fearful expression disappeared again. It will be interesting to learn whether other people with amygdala damage can also recognize expressions of fear if they are instructed to look at eyes.

FIGURE 11.19 Eye Fixations After Amygdala Damage

The figure shows the numbers of fixations on a person’s face made by a patient with bilateral amygdala damage (Patient S. M.) and a normal subject. Warmer colors indicate increasing numbers of fixations. Note that Patient S. M. does not look at the other person’s eyes.

FIGURE 11.20 Role of the Whites of Eyes in Emotional Perception

The stimuli used in the study by Whalen et al. ( 2004 ) show that the whites of the eyes alone can convey the impression of a fearful expression.

PERCEPTION OF DIRECTION OF GAZE

Perrett and his colleagues (see Perrett et al., 1992 ) have discovered an interesting brain function that may be related to recognition of emotional expression. They found that neurons in the monkey’s superior temporal sulcus (STS) are involved in recognition of the direction of another monkey’s gaze—or even that of a human. They found that some neurons in this region responded when the monkey looked at photographs of a monkey’s face or a human face but only when the gaze of the face in the photograph was oriented in a particular direction. For example, Figure 11.21 shows the activity level of a neuron that responded when a human face was looking upward. (See Figure 11.21 . )

Why is gaze important in recognition of emotions? First, it is important to know whether an emotional expression is directed toward you or toward someone else. For example, an angry expression directed toward you means something very different from a similar expression directed toward someone else. And if someone else shows signs of fear, the expression can serve as a useful warning, but only if you can figure out what the person is looking at. In fact, Adams and Kleck ( 2005 ) found that people more readily recognized anger if the eyes of another person were directed toward the observer and fear if they were directed somewhere else. As Blair ( 2008 ) notes, an angry expression directed toward the observer means that the other person wants the observer to stop what he or she is doing.

FIGURE 11.21 A Gaze-Direction Cell

The graph shows the responses of a single neuron in the cortex lining the superior temporal sulcus of a monkey’s brain. The cell fired most vigorously when the monkey was presented a photograph of a face looking upward.

(From Perrett, D. I., Harries, M. H., Mistlin, A. J., et al. International Journal of Comparative Psychology, 1992, 4, 25–55.)

The neocortex that lines the STS seems to provide such information. Lesions there disrupt monkeys’ ability to discriminate the direction of another animal’s gaze, but they do not impair the monkeys’ ability to recognize other animals’ faces (Campbell et al., 1990 ; Heywood and Cowey, 1992 ). As we saw in Chapter 6 , the posterior parietal cortex—the endpoint of the dorsal stream of visual analysis—is concerned with perceiving the location of objects in space. A functional-imaging study by Pelphrey et al. ( 2003 ) had people watch an animated cartoon of a face. When the direction of gaze changed, increased activity was seen in the right STS and posterior parietal cortex. Presumably, the connections between neurons in the STS and the parietal cortex enable the orientation of another person’s gaze to direct one’s attention to a particular location in space.

ROLE OF IMITATION IN RECOGNITION OF EMOTIONAL EXPRESSIONS: THE MIRROR NEURON SYSTEM

Adolphs et al. ( 2000 ) discovered a possible link between somatosensation and emotional recognition. They compiled computerized information about the locations of brain damage in 108 patients with localized brain lesions and correlated this information with the patients’ ability to recognize and identify facial expressions of emotions. They found that the most severe damage to this ability was caused by damage to the somatosensory cortex of the right hemisphere. (See Figure 11.22 . ) They suggest that, when we see a facial expression of an emotion, we unconsciously imagine ourselves making that expression. Often, we do more than imagine making the expressions—we actually imitate what we see. Adolphs et al. suggest that the somatosensory representation of what it feels like to make the perceived expression provides cues that we use to recognize the emotion being expressed in the face we are viewing. In support of this hypothesis, Adolphs and his colleagues report that the ability of patients with right hemisphere lesions to recognize facial expressions of emotions is correlated with their ability to perceive somatosensory stimuli. That is, patients with somatosensory impairments (caused by right-hemisphere lesions) also had impairments in recognition of emotions.

FIGURE 11.22 Brain Damage and Recognition of Facial Expressions of Emotion

This computer-generated image shows performance of subjects with localized brain damage on recognition of facial expressions of emotion. The colored areas outline the site of the lesions. Lesions that resulted in good performance are shown in shades of blue; those that resulted in poor performance are shown in red and yellow.

(From Adolphs, R., Damasio, H., Tranel, D., Cooper, G., and Damasio, A. R. The Journal of Neuroscience, 2000, 20, 2683–2690.)

Hussey and Safford ( 2009 ) review a considerable amount of evidence that supports this hypothesis (the so-called simulationist hypothesis). For example, neuroimaging studies have shown that brain regions that are activated when particular emotional expressions are observed are also activated when these expressions are imitated. In addition, a study by Pitcher et al. ( 2008 ) used transcranial magnetic stimulation to disrupt the normal activity of brain regions involved in visual perception of faces or perception of somatosensory feedback from one’s own face. They found that disruption of either region impaired people’s ability to recognize facial expressions of emotion. Finally, a study by Oberman et al. ( 2007 ) had people hold a pen between their teeth in a way that interfered with smiling. When they did so, they had difficulty recognizing facial expressions of happiness, but not expressions of disgust, fear, and sadness, which involve the upper part of the face more than smiling does.

We are beginning to understand the neural circuit that provides this form of feedback. In Chapter 8 I described the role of mirror neurons in the control of movement. Mirror neurons are activated when an animal performs a particular behavior or when it sees another animal performing that behavior. Presumably, these neurons are involved in learning to imitate the actions of others. These neurons, which are located in the ventral premotor area of the frontal lobe, receive input from the superior temporal sulcus and the posterior parietal cortex. As we saw in Chapter 8 , this circuit is activated when we see another person perform a goal-directed action, and feedback from this activity helps us to understand what the person is trying to accomplish. Carr et al. ( 2003 ) suggest that the mirror neuron system, which is activated when we observe facial movements of other people, provides feedback that helps us to understand how other people feel. In other words, the mirror neuron system may be involved in our ability to empathize with the emotions of other people. (I will have more to say about empathy in the last section of this chapter.)

A neurological disorder known as Moebius syndrome provides further support for this hypothesis. Moebius syndrome is a congenital condition that involves defective development of the sixth (abducens) and seventh (facial) cranial nerves and results in facial paralysis and inability to make lateral eye movements. Because of this paralysis, people affected with Moebius syndrome cannot make facial expressions of emotion. In addition, they have difficulty recognizing the emotional expressions of other people (Cole, 2001 ). Perhaps their inability to produce facial expressions of emotions makes it impossible for them to imitate the expressions of other people, and the lack of internal feedback from the motor system to the somatosensory cortex may make the task of recognition more difficult.

In Chapter 8 I described research on audiovisual neurons—neurons that respond to the sounds of particular actions and to the sight of those actions. Warren et al. ( 2006 ) obtained evidence that audiovisual neurons play a role in communication of emotions, too. The investigators asked volunteers to make emotional sounds in response to written scenarios that presented situations expected to evoke triumph, amusement, fear, and disgust. The volunteers were asked not to make verbal responses such as “yuck” or “yippee” but to restrict themselves to nonverbal vocal responses. These sounds were presented to subjects while they underwent fMRI scanning. The scans showed that hearing the emotional vocalizations activated the same regions of the brain that were activated by facial expressions of these emotions. In other words, when we hear other people make nonverbal emotional sounds, our mirror neuron system is activated, and the feedback from this activation may contribute to our recognition of the emotions being expressed by these sounds.

DISGUST

And now for something completely different. Several studies have found that damage to the insular cortex and basal ganglia impairs people’s ability to recognize facial expressions of disgust (Sprengelmeyer et al., 1996 , 1997 ; Calder et al., 2000 ). In addition, a functional-imaging study by Wicker et al. ( 2003 ) found that both smelling a disgusting odor and seeing a face of a person showing an expression of disgust activate the insular cortex. Disgust (literally, “bad taste”) is an emotion provoked by something that tastes or smells bad—or by an action that we consider to be in bad taste (figuratively, not literally). Disgust produces a very characteristic facial expression; if you want to see a good example, refer back to Figure 11.14d . As we saw in Chapter 7 , the insula contains the primary gustatory cortex, so perhaps it is not a coincidence that this region is also involved in recognition of “bad taste.”

A functional-imaging study by Thielscher and Pessoa ( 2007 ) asked subjects to press one of two levers to indicate whether the facial expression they saw was one of disgust or fear. The expressions varied in intensity, and one of them was actually neutral, indicating neither disgust nor fear. Nevertheless, the subjects were asked to press one of the levers on every trial, indicating disgust or fear. When the subjects saw faces expressing disgust, the insular cortex and part of the basal ganglia were activated. What was particularly interesting was that even when the subjects were watching a neutral expression, if they pressed the “disgust” lever, the “disgust” regions of the brain were activated.

The results of an online survey presented on the British Broadcasting Corporation Science web site suggests that the emotion of disgust has its origins in avoidance of disease. The survey presented pairs of photos and asked people to indicate which photos were more disgusting. The people who responded indicated that the one that appeared to hold a potential threat of disease was more disgusting. For example, a yellow liquid that has soaked a tissue looks more like a body fluid than a blue liquid does. (See Figure 11.23 . )

FIGURE 11.23 Disease and Disgust

The figure shows pairs of photographs with high and low relation to the threat of disease used in the online survey presented on the BBC Science web site. The numbers in red or green indicate the mean ratings (range = 1–5) made by people who completed the survey.

(From Curtis, V., Aunger, R., and Rabie, T. Biology Letters, 2004, 271, S131–S133.)

Neural Basis of the Communication of Emotions: Expression

Facial expressions of emotion are automatic and involuntary (although, as we saw, they can be modified by display rules). It is not easy to produce a realistic facial expression of emotion when we do not really feel that way. In fact, Ekman and Davidson have confirmed an early observation by a nineteenth-century neurologist, Guillaume-Benjamin Duchenne de Boulogne, that genuinely happy smiles, as opposed to false smiles or social smiles people make when they greet someone else, involve contraction of a muscle near the eyes, the lateral part of the orbicularis oculi—now sometimes referred to as Duchenne’s muscle (Ekman, 1992 ; Ekman and Davidson, 1993 ). As Duchenne put it, “The first [zygomatic major muscle] obeys the will but the second [orbicularis oculi] is only put in play by the sweet emotions of the soul; the . . . fake joy, the deceitful laugh, cannot provoke the contraction of this latter muscle” (Duchenne, 1862/1990 , p. 72). (See Figure 11.24 . ) The difficulty actors have in voluntarily producing a convincing facial expression of emotion is one of the reasons that led Constantin Stanislavsky to develop his system of method acting, in which actors attempt to imagine themselves in a situation that would lead to the desired emotion. Once the emotion is evoked, the facial expressions follow naturally (Stanislavsky, 1936 ).

FIGURE 11.24 An Artificial Smile

The photograph shows Dr. Duchenne electrically stimulating muscles in the face of a volunteer, causing contraction of muscles around the mouth that become active during a smile. As Duchenne discovered, however, a true smile also involves muscles around the eyes.

This observation is confirmed by two neurological disorders with complementary symptoms (Hopf, Mueller-Forell, and Hopf, 1992 ; Topper, Kosinski, and Mull, 1995 ; Urban et al., 1998 ; Michel et al., 2008 ). The first, volitional facial paresis , is caused by damage to the face region of the primary motor cortex or to the fibers connecting this region with the motor nucleus of the facial nerve, which controls the muscles responsible for movement of the facial muscles. (Paresis, from the Greek “to let go,” refers to a partial paralysis.) The interesting thing about volitional facial paresis is that the patient cannot voluntarily move the facial muscles but will express a genuine emotion with those muscles. For example, Figure 11.25(a) shows a woman trying to pull her lips apart and show her teeth. Because of the lesion in the face region of her right primary motor cortex, she could not move the left side of her face. However, when she laughed ( Figure 11.25b ), both sides of her face moved normally. (See Figure 11.25a and 11.25b . ) In contrast, emotional facial paresis is caused by damage to the insular region of the prefrontal cortex, to the white matter of the frontal lobe, or to parts of the thalamus. This system joins the system responsible for voluntary movements of the facial muscles in the medulla or caudal pons. People with this disorder can move their face muscles voluntarily but do not express emotions on the affected side of the face. Figure 11.25(c) shows a man pulling his lips apart to show his teeth, which he had no trouble doing. Figure 11.25(d) shows him smiling; as you can see, only the left side of his mouth is raised. He had a stroke that damaged the white matter of the left frontal lobe. (See Figure 11.25c and 11.25d . ) These two syndromes clearly indicate that different brain mechanisms are responsible for voluntary movements of the facial muscles and automatic, involuntary expression of emotions involving the same muscles.

volitional facial paresis Difficulty in moving the facial muscles voluntarily; caused by damage to the face region of the primary motor cortex or its subcortical connections.

emotional facial paresis Lack of movement of facial muscles in response to emotions in people who have no difficulty moving these muscles voluntarily; caused by damage to the insular prefrontal cortex, subcortical white matter of the frontal lobe, or parts of the thalamus.

Several studies have investigated the brain mechanisms involved in laughter, an expression of emotion more intense than smiling. Arroyo et al. ( 1993 ) reported the case of a patient who had seizures that were accompanied by mirthless laughter—that is, the patient laughed but was neither happy nor amused. Recordings made with depth electrodes indicated that the seizure began in the left anterior cingulate gyrus. Removal of a noncancerous tumor located nearby ended both the seizures and the mirthless laughter. The authors suggest that anterior cingulate cortex may be involved in the muscular movements that produce laughter. Shammi and Stuss ( 1999 ) found that damage to the right vmPFC impaired people’s ability to understand—and be amused by—jokes. For example, consider the following joke:

· Mr. Smith’s neighbor approaches him and asks, “Say, are you using your lawnmower this afternoon?” “Yes, I am,” replies Mr. Smith.

Which alternative below finishes the joke?

· a. “Oops!” as he steps on a rake that barely misses his face.

· b. “Fine, then you won’t be wanting your golf clubs—I’ll just borrow them.”

· c. “Oh well, can I borrow it when you’re done, then?”

· d. “The birds are always eating my grass seed.”

FIGURE 11.25 Emotional and Volitional Paresis

(a) A woman with volitional facial paresis caused by a right hemisphere lesion tries to pull her lips apart and show her teeth. Only the right side of her face responds. (b) The same woman shows a genuine smile. (c) A man with emotional facial paresis caused by a left-hemisphere lesion shows his teeth. (d) The same man smiles. Only the left side of his face responds.

(From Hopf, H. C., Mueller-Forell, W., and Hopf, N. J., Neurology, 1992, 42, 1918–1923.)

FIGURE 11.26 Humor and Violation of Social Norms

The graph shows the activation of the right ventromedial prefrontal cortex and the left orbitofrontal cortex, as measured by fMRI, by exposure to humorous cartoons with increasing funniness and increasing violation of social norms.

(Based on data from Goel and Dolan, 2007.)

The funny alternative is, of course, (b). But people with ventromedial prefrontal damage usually chose (a), presumably because its slapstick aspect reminded them of humor that they had seen in the past. Clearly, they did not get the point of the joke.

A functional-imaging study by Goel and Dolan ( 2001 ) found that different types of jokes activated different regions of the brain, but all of them activated one region: the right ventromedial prefrontal cortex. Another functional-imaging study by the same authors (Goel and Dolan, 2007 ) presented subjects with socially appropriate and socially inappropriate cartoon jokes. (The inappropriate jokes had strong sexual content that some subjects found offensive.) The investigators found that increasingly funny jokes produced increasing activation of several regions, including the nucleus accumbens (a region involved in reinforcement and reward) and the right vmPFC, while jokes with increasing violation of social norms produced increasing activation of several regions, including the right amygdala and the left orbital frontal cortex. (See Figure 11.26 . )

As we saw in the previous subsection, the right hemisphere plays a more significant role in recognizing emotions in the voice or facial expressions of other people—especially negative emotions. The same hemispheric specialization appears to be true for expressing emotions. When people show emotions with their facial muscles, the left side of the face usually makes a more intense expression. For example, Sackeim and Gur ( 1978 ) cut photographs of people who were expressing emotions into right and left halves, prepared mirror images of each of them, and pasted them together, producing so-called chimerical faces (from the mythical Chimera, a fire-breathing monster, part goat, part lion, and part serpent). They found that the left halves were more expressive than the right ones. (See Figure 11.27 . ) Because motor control is contralateral, the results suggest that the right hemisphere is more expressive than the left.

Moscovitch and Olds ( 1982 ) made more natural observations of people in restaurants and parks and found that the left side of their faces appeared to make stronger expressions of emotions. They confirmed these results in the laboratory by analyzing videotapes of people telling sad or humorous stories. A review of the literature by Borod et al. ( 1998 ) found forty-eight other studies that obtained similar results.

Using the chimerical faces technique, Hauser ( 1993 ) found that rhesus monkeys, like humans, express emotions more strongly in the left sides of their faces. Analysis of videotapes further showed that emotional expressions also begin sooner in the left side of the face. These findings suggest that hemispherical specialization for emotional expression appeared before the emergence of our own species. Figure 11.28 shows six videotape frames of a monkey’s fear grimace expressed during the course of an interaction with a more dominant monkey. (See Figure 11.28 . )

FIGURE 11.27 Chimerical Faces

(a) Original photo. (b) Composite of the right side of the man’s face. (c) Composite of the left side of the man’s face.

FIGURE 11.28 Emotional Expression and the Right Hemisphere

Successive frames from a videotape of a rhesus monkey show a fear grimace in response to an interaction with a more dominant monkey. The movement begins in the left side of the face, controlled by the right hemisphere.

Left hemisphere lesions do not usually impair vocal expressions of emotion. For example, people with Wernicke’s aphasia (described in Chapter 14 ) usually modulate their voice according to mood, even though the words they say make no sense. In contrast, right-hemisphere lesions do impair expression of emotion, both facially and by tone of voice.

We saw in the previous subsection that the amygdala is involved in the recognition of facial expression of emotions. Research indicates that it is not involved in emotional expression. Anderson and Phelps ( 2000 ) reported the case of S. P., a 54-year-old woman whose right amygdala was removed to treat a serious seizure disorder. Because of a preexisting lesion of the left amygdala, the surgery resulted in a bilateral amygdalectomy. After the surgery, S. P. lost the ability to recognize facial expressions of fear, but she had no difficulty recognizing individual faces, and she could easily identify male and female faces and accurately judge their ages. What is particularly interesting is that the amygdala lesions did not impair S. P.’s ability to produce her own facial expressions of fear. She had no difficulty accurately expressing fear, anger, happiness, sadness, disgust, and surprise. However, when she saw a photograph of herself showing fear, she could not tell what emotion her face had been expressing.

SECTION SUMMARY: Communication of Emotions

We (and members of other species) communicate our emotions primarily through facial gestures. Darwin believed that such expressions of emotion were innate—that these muscular movements were inherited behavioral patterns. Ekman and his colleagues performed cross-cultural studies with members of an isolated tribe in New Guinea. Their results supported Darwin’s hypothesis.

Recognition of other people’s emotional expressions involves the right hemisphere more than the left. Studies with normal people have shown that people can judge facial expressions or tone of voice better when the information is presented to the right hemisphere than when it is presented to the left hemisphere. Functional imaging indicates that when people judge the emotions of voices, the right hemisphere is activated more than the left. Studies of people with left- or right-hemisphere brain damage corroborate these findings. In addition, studies show that recognition of particular faces involves neural circuits different from those needed to recognize facial expressions of emotions. Finally, the amygdala plays a role in recognition of facial expressions of fearfulness; lesions of the amygdala disrupt this ability, and functional-imaging studies show increased activity of the amygdala while the subject is engaging in this task. The ability to judge emotions by a person’s tone of voice is not affected.

The amygdala receives magnocellular (primitive) visual information from the superior colliculus and pulvinar, and this information is used in making judgments about fearful expressions. Because of this input, people with damage to the visual cortex that leads to blindness in part of the visual field can nevertheless recognize facial expressions of emotions presented there, a phenomenon called affective blindsight. We can also recognize emotions expressed in a person’s body posture or movement, and the amygdala receives and processes this input as well. One of the reasons that bilateral amygdala damage impairs recognition of fearful facial expressions appears to be failure to look at people’s eyes.

The direction of the gaze of a person expressing an emotion has information value. Neurons in the superior temporal sulcus are sensitive to direction of gaze and transmit this information to other parts of the brain, including the amygdala. Mirror neurons in the ventral premotor cortex receive visual information concerning the facial expression of other people that activate the neural circuits responsible for these expressions. Feedback from this activity, which may be transmitted to the somatosensory cortex, helps us to comprehend the emotional intentions of other people. Damage to the basal ganglia and insular cortex disrupts recognition of facial expressions of disgust, and functional-imaging studies show increased activity in the insular cortex (which contains the primary gustatory cortex) when people smell disgusting odors or see faces displaying disgust.

Facial expression of emotions (and other stereotypical behaviors such as laughing and crying) are almost impossible to simulate. For example, only a genuine smile of pleasure causes the contraction of the lateral part of the orbicularis oculi (Duchenne’s muscle). The anterior cingulate gyrus appears to play a role in the motor aspects of laughter, while the appreciation of humor appears to involve the right ventromedial prefrontal cortex. Genuine expressions of emotion are controlled by special neural circuits. The best evidence for this assertion comes from the complementary syndromes of emotional and volitional facial paresis. People with emotional facial paresis can move their facial muscles voluntarily but not in response to an emotion, whereas people with volitional facial paresis show the opposite symptoms. In addition, the left halves of people’s faces—and the faces of monkeys—tend to be more expressive than the right halves.

▪ THOUGHT QUESTIONS

Do you think it is important to be able to recognize other people’s emotions? Why? Suppose that you could no longer recognize people’s facial expressions of emotions. What consequences would that loss have for you?

Novelists will sometimes say that a person’s smile did not reach his or her eyes. What do they mean by that?

Feelings of Emotions

So far, we have examined two aspects of emotions: the organization of patterns of responses that deal with the situation that provokes the emotion and the communication of emotional states with other members of the species. The final aspect of emotion to be examined in this chapter is the subjective component: feelings of emotion.

The James-Lange Theory

William James (1842–1910), an American psychologist, and Carl Lange (1834–1900), a Danish physiologist, independently suggested similar explanations for emotion, which most people refer to collectively as the James-Lange theory (James, 1884 ; Lange, 1887 ). Basically, the theory states that emotion-producing situations elicit an appropriate set of physiological responses, such as trembling, sweating, and increased heart rate. The situations also elicit behaviors, such as clenching of the fists or fighting. The brain receives sensory feedback from the muscles and from the organs that produce these responses, and it is this feedback that constitutes our feeling of emotion.

James-Lange theory A theory of emotion that suggests that behaviors and physiological responses are directly elicited by situations and that feelings of emotions are produced by feedback from these behaviors and responses.

James said that our own emotional feelings are based on what we find ourselves doing and on the sensory feedback we receive from the activity of our muscles and internal organs. For example, when we find ourselves trembling and feel queasy, we experience fear. Where feelings of emotions are concerned, we are self-observers. Thus, the two aspects of emotions reported in the first two sections of this chapter (patterns of emotional responses and expressions of emotions) give rise to the third: feelings. (See Figure 11.29 . )

James’s description of the process of emotion might strike you as being at odds with your own experience. Many people think that they experience emotions directly, internally. They consider the outward manifestations of emotions to be secondary events. But have you ever found yourself in an unpleasant confrontation with someone else and discovered that you were trembling, even though you did not think that you were so bothered by the encounter? Or did you ever find yourself blushing in response to some public remark that was made about you? Or did you ever find tears coming to your eyes while you watched a film that you did not think was affecting you? What would you conclude about your emotional states in situations like these? Would you ignore the evidence from your own physiological reactions?

FIGURE 11.29 The James-Lange Theory of Emotion

An event in the environment triggers behavioral, autonomic, and endocrine responses. Feedback from these responses produces feelings of emotions.

A well-known physiologist, Walter Cannon, criticized James’s theory. He said that the internal organs were relatively insensitive and that they could not respond very quickly, so feedback from them could not account for our feelings of emotions. In addition, he observed that cutting the nerves that provide feedback from the internal organs to the brain did not alter emotional behavior (Cannon, 1927 ). However, subsequent research indicated that Cannon’s criticisms are not relevant. For example, although the viscera are not sensitive to some kinds of stimuli, such as cutting and burning, they provide much better feedback than Cannon suspected. Moreover, many changes in the viscera can occur rapidly enough that they could be the causes of feelings of emotion.

Cannon cited the fact that cutting the sensory nerves between the internal organs and the central nervous system does not abolish emotional behavior in laboratory animals. However, this observation misses the point. It does not prove that feelings of emotion survive this surgical disruption—only that emotional behaviors do. We do not know how the animals feel; we know only that they will snarl and attempt to bite if threatened. In any case, James did not attribute all feelings of emotion to the internal organs; he also said that feedback from muscles was important. The threat might make the animal snarl and bite, and the feedback from the facial and neck muscles might constitute a “feeling” of anger, even if feedback from the internal organs was cut off. But we have no way to ask the animal how it felt.

James’s theory is difficult to verify experimentally because it attempts to explain feelings of emotion, not the causes of emotional responses, and feelings are private events. Some anecdotal evidence supports the theory. For example, Sweet ( 1966 ) reported the case of a man in whom some sympathetic nerves were severed on one side of the body to treat a cardiovascular disorder. The man—a music lover—reported that the shivering sensation he felt while listening to music now occurred only on the unoperated side of his body. He still enjoyed listening to music, but the surgery had altered his emotional reaction.

In one of the few tests of James’s theory, Hohman ( 1966 ) collected data from people with spinal cord damage. He asked these people about the intensity of their emotional feelings. If feedback is important, one would expect that emotional feelings would be less intense if the injury were high (that is, close to the brain) than if it were low, because a high spinal cord injury would make the person become insensitive to a larger part of the body. In fact, this result is precisely what Hohman found: The higher the injury, the less intense the feeling was. As one of Hohman’s subjects said:

· I sit around and build things up in my mind, and I worry a lot, but it’s not much but the power of thought. I was at home alone in bed one day and dropped a cigarette where I couldn’t reach it. I finally managed to scrounge around and put it out. I could have burned up right there, but the funny thing is, I didn’t get all shook up about it. I just didn’t feel afraid at all, like you would suppose. (Hohman, 1966 , p. 150)

Another subject showed that angry behavior (an emotional response) does not appear to depend on feelings of emotion. Instead, the behavior is evoked by the situation (and by the person’s evaluation of it) even if the spinal cord damage has reduced the intensity of the person’s emotional feelings:

· Now, I don’t get a feeling of physical animation, it’s sort of cold anger. Sometimes I act angry when I see some injustice. I yell and cuss and raise hell, because if you don’t do it sometimes, I’ve learned people will take advantage of you, but it doesn’t have the heat to it that it used to. It’s a mental kind of anger. (Hohman, 1966 , p. 151)

Feedback from Emotional Expressions

James stressed the importance of two aspects of emotional responses: emotional behaviors and autonomic responses. As we saw earlier in this chapter, a particular set of muscles—those of the face—helps us to communicate our emotional state to other people. Several experiments suggest that feedback from the contraction of facial muscles can affect people’s moods and even alter the activity of the autonomic nervous system.

Ekman and his colleagues (Ekman, Levenson, and Friesen, 1983 ; Levenson, Ekman, and Friesen, 1990 ) asked subjects to move particular facial muscles to simulate the emotional expressions of fear, anger, surprise, disgust, sadness, and happiness. They did not tell the subjects what emotion they were trying to make them produce, but only what movements they should make. For example, to simulate fear, they told the subjects, “Raise your brows. While holding them raised, pull your brows together. Now raise your upper eyelids and tighten the lower eyelids. Now stretch your lips horizontally.” (These movements produce a facial expression of fear.) While the subjects made the expressions, the investigators monitored several physiological responses controlled by the autonomic nervous system.

The simulated expressions did alter the activity of the autonomic nervous system. In fact, different facial expressions produced somewhat different patterns of activity. For example, anger increased heart rate and skin temperature, fear increased heart rate but decreased skin temperature, and happiness decreased heart rate without affecting skin temperature.

Why should a particular pattern of movements of the facial muscles cause changes in mood or in the activity of the autonomic nervous system? Perhaps the connection is a result of experience; in other words, perhaps the occurrence of particular facial movements along with changes in the autonomic nervous system leads to classical conditioning, so that feedback from the facial movements becomes capable of eliciting the autonomic response—and a change in perceived emotion. Or perhaps the connection is innate. As we saw earlier, the adaptive value of emotional expressions is that they communicate feelings and intentions to others. The research presented earlier in this chapter on the role of mirror neurons and the somatosensory cortex suggests that one of the ways we communicate feelings is through unconscious imitation.

A study by Lewis and Bowler ( 2009 ) found that interfering with muscular movement associated with a particular emotion decreased people’s ability to experience that emotion. As you know, injections of a very dilute solution of botulinum toxin (Botox) into facial muscles can reduce wrinkling of the skin caused by chronic contraction of facial muscles. Lewis and Bowler studied people who had been treated with injections of Botox into the corrugator muscle, whose contraction is responsible for a large part of the facial expression of frowning, which is associated with negative emotions. They found that these people showed significantly less negative mood, compared with people who had receive other forms of cosmetic treatment. These results, like those described earlier in this subsection, suggest that feedback from a person’s facial expressions can affect his or her mood.

A functional-imaging study by Damasio et al. ( 2000 ) asked people to recall and try to reexperience past episodes from their lives that evoked feelings of sadness, happiness, anger, and fear. The investigators found that recalling these emotions activated the subjects’ somatosensory cortex and upper brain stem nuclei involved in control of internal organs and detection of sensations received from them. These responses are certainly compatible with James’s theory. As Damasio et al. put it,

· Emotions are part of a neural mechanism based on structures that regulate the organism’s current state by executing specific actions via the musculoskeletal system, ranging from facial and postural expressions to complex behaviors, and by producing chemical and neural responses aimed at the internal milieu, viscera and telencephalic neural circuits. The consequences of such responses are represented in both subcortical regulatory structures . . . and in cerebral cortex . . . , and those representations constitute a critical aspect of the neural basis of feelings. (p. 1049)

I suspect that if James were still alive, he would approve of these words.

The tendency to imitate the expressions of other people appears to be innate. Field et al. ( 1982 ) had adults make facial expressions in front of infants. The infants’ own facial expressions were videotaped and were subsequently rated by people who did not know what expressions the adults were displaying. Field and her colleagues found that even newborn babies (with an average age of 36 hours) tended to imitate the expressions they saw. Clearly, the effect occurs too early in life to be a result of learning. Figure 11.30 shows three photographs of the adult expressions and the expressions they elicited in a baby. Can you look at them yourself without changing your own expression, at least a little? (See Figure 11.30 . )

FIGURE 11.30 Imitation in an Infant

The photographs show happy, sad, and surprised faces posed by an adult and the responses made by the infant.

(From Field, T., in Development of Nonverbal Behavior in Children, edited by R. S. Feldman. New York: Springer-Verlag, 1982. Reprinted with permission.)

Perhaps imitation provides one of the channels by which organisms communicate their emotions—and evoke feelings of empathy. For example, if we see someone looking sad, we tend to assume a sad expression ourselves. The feedback from our own expression helps to put us in the other person’s place and makes us more likely to respond with solace or assistance. And perhaps one of the reasons we derive pleasure from making someone else smile is that their smile makes us smile and feel happy. In fact, a functional-imaging study by Pfeifer et al. ( 2008 ) found that when normal 10-year-old children watched and imitated emotional expressions, increased activity was seen in the frontal mirror neuron system. In addition, the level of neural activation was positively correlated with measures of the children’s empathetic behavior and their interpersonal skills.

SECTION SUMMARY: Feelings of Emotions

From the earliest times, people recognized that emotions were accompanied by feelings that seemed to come from inside the body, which probably provided the impetus for developing physiological theories of emotion. James and Lange suggested that emotions were primarily responses to situations. Feedback from the physiological and behavioral reactions to emotion-producing situations gave rise to the feelings of emotion; thus, feelings are the results, not the causes, of emotional reactions. Hohman’s study of people with spinal cord damage supported the James-Lange theory; people who could no longer feel the reactions from most of their body reported that they no longer experienced intense emotional states.

Ekman and his colleagues have shown that even simulating an emotional expression causes changes in the activity of the autonomic nervous system. Perhaps feedback from these changes explains why an emotion can be “contagious”: We see someone smile with pleasure, we ourselves imitate the smile, and the internal feedback makes us feel at least somewhat happier. The tendency to mimic the facial expression of others appears to be a consequence of activity in the brain’s system of mirror neurons.

▪ THOUGHT QUESTIONS

Some people are more conscious of their own bodily reactions than most people. What effect would this sensitivity have on their mood?

Can you think of any evolutionary advantage of the fact that human infants can mimic the facial emotional expressions of adults?

Review Questions

Study and Review on MyPsychLab

Discuss the behavioral, autonomic, and hormonal components of an emotional response and the role of the amygdala in controlling them.

Discuss the nature, functions, and neural control of aggressive behavior.

Discuss the role of the ventral prefrontal cortex in the analysis of social situations and the effects of damage to this region.

Discuss the hormonal control of aggression in males and aggression in females.

Discuss the effects of androgens on human aggressive behavior.

Discuss cross-cultural studies on the expression and comprehension of emotions.

Describe the neural control of the recognition of emotional expression in normal people and people with brain damage.

Discuss the neural control of emotional expression in normal people and people with brain damage.

Discuss the James-Lange theory of feelings of emotion and evaluate relevant research.

Explore the Virtual Brain in MyPsychLab

▪ EMOTION AND STRESS

Emotion and stress have profound effects on cognition. Although distinct, the neural circuits underlying emotion and stress interact. The Emotion and Stress module of the virtual brain will help you visualize the brain regions involved in the generation and perception of emotion. It will also help you visualize those regions involved in the generation of the stress response.