A new theory of the neuroscientific basis for the human instinct for fair play.

Associated with every religious system I have read about, across continents and centuries, is a rubric, known in Christianity as the Golden Rule, that requires me to do unto you as I would have you do unto me. This rule is so ingrained in how we behave toward each other that we rarely stop to question its source. If pressed, we might speculate that its origins are lost in the mists of time—for example, when the first high priests figured out how to satisfy a sovereign’s demand for social stability. But suppose, for the moment, that the Golden Rule is even older, that it is as old as our own biology. Suppose that, although the Golden Rule might have acquired all sorts of sociopolitical decorations over the course of human history, it nonetheless is traceable to identifiable processes in our brains. If this were so, we could understand why this rule and its many variations have survived so widely in human ethical systems, philosophies, and religions. In his recent book, The Ethical Brain, Michael S. Gazzaniga, Ph.D., challenged us not only to face the ethical questions inherent in neurobiology but also to try to understand how our brains govern our ethical responses themselves. Here I want to explore a new theory of the neuroscientific basis for the human instinct for fair play. FAIR PLAY: HUMAN, ANIMAL, AND COMPUTER The evidence for this universal principle of fair behavior is almost overwhelming; it comes not only from religion but also from anthropology, archeology, the study of animal behavior, and even computer science. South African scholars David Lewis-Williams, Ph.D., and David Pearce, Ph.D., argue from archeologic and anthropologic evidence that neural patterns of activity hard-wired into the human brain help explain the range of religious art and social practices they believe were produced and exhibited by Neolithic people. They write, “The commonalities we highlight cannot be explained in any other way than by the functioning of the universal human nervous system.” Similar images in the art of widely separated tribes suggested overwhelmingly to those authors that a common mentality underlay their religious beliefs. Biologic forces in the form of common neuronal patterns of activity have helped shape human behavior in society, including learning not to treat others in ways we would not want to be treated ourselves. My suspicion that basic neurobiological mechanisms foster fair play is deepened by realizing that animals also behave this way. Animal behaviorists call it “reciprocal altruism.” Individual animals take risks to protect the group; one animal does things that help other members of its group survive, just as it would want those other members to act for its benefit. A small bird sounds its alarm call in a manner that alerts and saves the rest of the flock from the approaching hawk, even though by doing so it reveals its own location to the hawk. A baboon shares food with a hungry member of its troop, even though by doing so it has less food for itself. Moreover, the mechanisms involved are not mystical; even computers can be programmed to behave “fairly.” Robert Axelrod, Ph.D., and William Hamilton, Ph.D., at the University of Michigan programmed computers to display mutual cooperation and, by running a computer tournament showed that such behavior is an evolutionarily stable strategy. In other words, mutual cooperation in these computer games started spontaneously, thrived, and resisted opposition. They started the computer tournament with each “player” able to exhibit either mutual cooperation or selfish defection by following these two rules: 1. “On the first move, cooperate” and 2. “On all subsequent moves, do whatever the other computer did on the first move.” Thus, all computers cooperated on the first move and, if they followed the second rule, also cooperated on subsequent moves, a “tit for tat” strategy. If computers can display cooperative responses, then we know that straightforward physical mechanisms must be sufficient to explain the result. In the research by Axelrod and Hamilton, an ongoing interaction between computers was necessary for cooperation to thrive. Human beings, too, are more likely to treat another person properly if they expect to interact with that person again. We obey the Golden Rule when we sense that we inhabit the same “space” as the other person. Conversely, we do not murder, because we do not want to be murdered ourselves. We see that the people to whom we would do harm actually share fates with us. FEAR, FROM THE THALAMUS TO THE AMYGDALA Evidence from every direction points to an underlying neurobiology equally stable over time and across cultures and species—yes, even including the computer. What I propose is that the mechanisms in our brains that guide behavior adhering to the Golden Rule are those that govern our fear responses. I further propose that the process involves a blurring of identity, in which one person’s envisioned fates and fears are merged with another’s. We share not only our fates, but our fears with others. To understand why this might be so, I should explain a little about the complicated processes involved in fear, from the neural pathways in the brain to genes and hormones, emotions and memories. Some stimuli from the environment produce fear, others do not. The difference in our brains derives from the ability of fearsome stimuli to trigger activity in an ancient part of our forebrain called the amygdala. There, certain genes and a few messenger chemicals are the crucial factors. A series of landmark discoveries in several laboratories has shown how stimuli produce fear. The two basic kinds of fear, unlearned and learned, depend on the amygdala—at least 10 ancient groups of neurons lying roughly parallel to our ears—which participates in primitive neural circuits deep in the forebrain. Unlearned fear, often manifest by animals stopping all movements (“freezing”) to see what is going on and to assess their risks, depends on a cell group called the central nucleus of the amygdala. Joseph LeDoux, Ph.D., and his colleagues at New York University showed that learned fear depends on storage of an emotional memory trace in a different part of the amygdala, the lateral amygdala. To paraphrase James McGaugh, Ph.D., at the University of California, Irvine, neurons in the lateral amygdala make emotionally significant events memorable. Scientists argue about what fears are innate, that is, unlearned. Some think that human beings are born with the fear of only a small number of events, such as loud noises and falling. Learned fear, however, develops when emotionally neutral stimuli and neutral environments— those with no particular emotional meaning—are associated with innately fearful experiences such as severe pain or other terrible events. For example, we could learn to fear a particular kind of house if we were subjected to deeply frightening experiences in a similar house. Stimuli that produce fear are not recognized as fearful from the microsecond they touch our sensory receptors. Instead, they enter our nervous systems value free, through the skin and up the spinal cord, on our tongues or through our noses, from our eyes or ears, and toward the forebrain. They reach a forebrain region called the thalamus. At that point, signals split and travel to the amygdala by two different routes, as shown in the illustration on the following page. One set of signals ascends toward the sensory areas of the cerebral cortex. In fact, the word “thalamus” comes from the Greek word meaning “antechamber” because, neuroanatomically, the thalamus acts as an antechamber to the cortex, that magnificent covering of our brains. From the cerebral cortex, those signals travel to the amygdala. The other set of signals goes directly to the amygdala from the thalamus. LeDoux’s laboratory has shown that the messages that travel these two routes are not identical to each other, and currently neuroscientists endeavor to explain how and why. I would like to posit an answer. Let me divide all responses that we call “fear” into primitive emotional reactions and more sophisticated cognitive reactions. I theorize that innately fearful events trigger electrical discharges in nerve cells located in the amygdala, straightaway, having used the direct route from the thalamus. These electrical signals lead to our experience of emotional fear. Learned fear, involving memories of initially neutral events that have become associated with fear, is more complex. I propose that our fearful thoughts depend on the information-processing power of the cerebral cortex. To appreciate the emotional component of fear, let us first concentrate on the amygdala itself. Michael Fanselow, Ph.D., at UCLA, and others have shown that its lateral parts encode memories of stimuli associated with aversive events. Then, from the amygdala, emotionally loaded signals are sent to many other parts of the forebrain, in particular an ancient part of the forebrain called the septum, and to the hypothalamus. These forebrain cell groups are both necessary and sufficient for the unique experience of fear. A more modern part of the forebrain, the frontal cortex, also receives fear signals, but it functions primarily to suppress sheer fright. For a person to avoid panic and long-term conditions such as post-traumatic stress disorder, the frontal cortex neurons must function normally. In addition, outputs from the amygdala also influence memory processes in many brain regions. Learned fear employs the same pathways, except that the stimuli associated with negative experiences start as ineffective for activating the amygdala. Only after repeated associations with pain and other negative events do these stimuli produce a reaction in the amygdala. These circuits in the central nervous system, whether for unlearned or learned fear, produce responses that allow an animal or a human being to avoid the source of the fear. Some of these responses, such as those I have just described, are very rapid. But in addition to these quick electrical and behavioral responses, brain mechanisms for fear make use of slow effects that involve gene expression. These slow effects depend on genes that make large proteins that bind stress hormones and on genes that code for a small molecule responsive to stress, corticotrophin-releasing hormone (CRH). As an intriguing aside, both of these modes of action—rapid/electrical and slow/genomic—provide possible targets for pharmaceuticals that could be used therapeutically to reduce fear. THREE STEPS IN AN AMYGDALA NEURON What happens in amygdaloid neurons that constitutes the mechanism of fear? In neurochemical and genetic terms, at least three steps can be identified: the arrival of specific chemical signals at the amygdaloid neuron; the signaling within the neuron consequent to that arrival; and the eventual effect of that signal, whether it is an electrical discharge or a change in gene expression. Let me give one example. The neuron transmitter called glutamate is involved in fear signaling. As soon as glutamate binds to its receptors on the cell membranes of amygdaloid neurons, calcium is permitted to enter the neuron. Then, the calcium unleashes a long series, what scientists call a “cascade,” of complex signals within the neuron. This cascade uses specialized proteins that “decorate” neurons with appendages containing phosphorus and oxygen. Such a cascade is required for the memory of fear. A second example uses a different type of biochemical cascade. A neuropeptide called brain-derived neurotrophic factor, known to participate in stress and fear, binds to its specific receptor on amygdaloid neurons. Once binding occurs, a small set of proteins—called “G-proteins” because they are organized around the chemical guanine—change their state of activation. That change triggers a new set of biochemical reactions that, as in our previous example, add appendages containing phosphorus and oxygen to the neurons. So far I have covered two steps: the arrival of particular neurotransmitters and neuropeptides at an amgdaloid neuron and the biochemical signaling within the cytoplasm of that neuron, beneath the membrane but outside the cell nucleus. The third step determines the consequences of this signaling, which can be either fast or slow. The fast effects are triggered if the last reaction of the biochemical cascade phosphorylates (adds phosphorus and oxygen to) a protein that participates in an ion channel in the amygdaloid nerve cell membrane. That process causes a rapid electrical change in the cell. The slow actions occur in the cell nucleus, where the last reaction of the biochemical cascade would phosphorylate either a genomic transcription factor or a chromatin protein that is part of the covering of the cell’s DNA. Therefore, both neurotransmitters and neuropeptides make use of series of phosphorylation reactions to transmit their signals both to membrane ion channels (for rapid electrophysiological changes signaling fear) and to the cell nucleus, for slow adaptations in the genes related to enduring fear states. HORMONE AND GENE CONNECTIONS The neurotransmitters and neuropeptides I have discussed affect the physiology of the amygdala. But to complete the story, I must add a fascinating—and complex—hormonal connection. Consider corticotrophin releasing hormone (CRH), mentioned earlier. CRH, which is produced by neurons in the central nucleus of the amygdala, not only fosters behavioral responses to fear but also triggers the stress hormone axis that runs from the hypothalamus to the pituitary to the adrenal gland. When this axis is triggered, steroidal hormones such as cortisol pour out of the adrenal gland and help the entire body deal with stress and fear. But the effects of cortisol on behavior depend on whether levels of adrenal stress hormones are chronically and repeatedly elevated. If they are not, the effect of cortisol is restorative, bringing the body’s systems back from an emergency state toward a normal one. For example, such stress hormones acting in the rat amygdala would be essential for the extinction of a learned fear response. But if an animal was subjected to chronic fear and stress, such that the adrenal hormones were called on again and again, then cortisol could have the opposite effect, actually amplifying behavioral responses to fearful stimuli. In humans, hormonal mechanisms that originate with CRH in the amygdala can either relieve our stress or heighten our anxiety, depending on our emotional history. The implications of these complex dynamics for the management of anxiety states, post-traumatic stress disorder, obsessive-compulsive disorders, and other illness are clear. New research continues to discover genes that affect how fear operates in the amygdala, but scientists do not yet know exactly what these genes do in amygdaloid neurons. For example, in a recent paper in Cell, a team led by Gleb Shumyatsky, Ph.D., at Rutgers University, reported a gene coding for a protein named stathmin that helps to make a mouse timid. Removing that gene through a biochemical process in the laboratory made the mouse more daring. Such mice failed to avoid innately fearful environments and had less memory of fearful events. In cellular terms, how this happens is still unknown, but electrical recordings from cells in the lateral amygdala showed deficits in responses to inputs from the cortex and the thalamus. For frightening emotions and emotional memories to operate correctly, in a biologically adaptive fashion, the entire central nervous system must be aroused. In addition to this “generalized arousal,” James McGaugh, Ph.D., and his colleagues at the University of California, Irvine, have reported that the proper operation of amygdaloid mechanisms related to fear depend on two specific arousal-related neurotransmitters, norepinephrine and dopamine. These chemical messengers coming from the lower brain stem energize amygdaloid neurons, so they can properly process inputs from the thalamus and cortex related to fear. An average person, or a “jumpy” one, would have enough norepinephrine and dopamine coming from the lower brain stem to the amygdala that he would respond briskly to fearful stimuli. But a person without enough norepinephrine and dopamine might not be afraid at all, even though signals from the thalamus and cortex tell his amygdala that he should be. This scenario may correspond to our experience of the proverbial unflappable “cool guy,” who does not seem to react dramatically to situations that deserve a full-blown fear response. HOW SHARED FEARS MAKE FAIR PLAY Based on the neurobiology of fear I have explained, I theorize that fear, learned and unlearned, helps a person avoid acts of violence and other behavior that could harm someone. That is, the genes and neural circuits that manage fear provide the crucial biological components of a process that leads to behavior obeying this universal ethical principle. This process is, of course, not all of ethics, but it is certainly central to the explanation of how human beings behave according to one important principle. From a scientific point of view, the best explanations of complex phenomena are the most parsimonious. Scientists try to use as few and as simple explanatory forces as possible. By referring to the basic, almost primitive, brain mechanisms governing fear, I can achieve this type of scientific explanation for reciprocal altruism. The explanation does not require fancy tricks of learning and memory but instead invokes the easiest step of all: the loss of information—a term I have coined for the blurring of identity between the actor and his target, a distinction that his brain normally would make readily but that is lost in the service of altruistic behavior. Loss of information takes place in the constant neurobiological process by which we distinguish self from other. Put simply, fear in a high-stakes situation can evoke an ethical response by impairing the sharpness of this distinction, suddenly making it unclear who is at risk. Consider a person’s action toward another. Before this action occurs, it is represented in the person’s brain, as every act must be. Moreover, this act will have consequences for the other individual that the person can understand, foresee, and remember. Now comes the crucial step. The person blurs the difference between the other individual and himself. Instead of foreseeing the consequences of his act solely for the other individual, he sees them for himself, as well. Let me give an extreme example, posed for absolute clarity. If someone is planning to knife another person in the stomach—with gruesome effects on the other person’s guts and blood—he loses the difference between the other person’s blood and guts and his own. He is less likely to carry through with the knifing he planned because he shares the other person’s fear, indeed, experiences it in his own body. This is easy to explain, because the loss or impairment of any one of the many complex steps involved in the identification of other individuals would lead to the loss of information my theory supposes. This loss of information—the person’s blurring of the distinction between himself and the intended target of his action—is an essential element of my theory. How could it happen? The recognition of another person depends on a long series of electrophysiological and biochemical reactions to the stimuli particular to that person. These stimuli include seeing the other person’s face, hearing his voice, feeling his touch, and smelling his personal odors. Every one of our senses works hard to identify that someone is that very person, there, not another person, and not ourselves. Reduction of the ability to make such discriminations in any of these sensory pathways will result in a blurring of the target’s identity. He is less easy to discriminate from others and, in fact, from oneself. This theory is not unreasonable. After all, explaining sensory discrimination is a daunting task, much harder than explaining how a computer works. Now, think of how easy it is to make a computer not work so efficiently. Improving the performance of a complex device (or biological process) is difficult. Degrading it—as my theory supposes— is ridiculously easy. Therefore, blurring the target’s identity and thus making it more like your own is, in terms of brain mechanisms, reasonable to suppose. In fact, making a person’s identity your own is essential to empathy. THE ROLE OF SOCIAL RECOGNITION If my theory posits a loss of social distinction, it would be useful to understand how social recognition works. How do we recognize another person for who he is, as distinct from ourselves? Scientists are beginning to piece together the molecular basis of social recognition through brain research on laboratory animals. Consider laboratory mice, which rely on smell to recognize others. Because virtually all social odors are signaled through forebrain pathways that lead to the amygdala, this brain region once again comes into play. Pheromones— odors that inform members of a species about other members of the same species—are recorded by specific parts of the amygdala. What then happens in the amygdala that is important for social recognition was studied by Jennifer Ferguson, Ph.D., and her collaborators Larry Young, Ph.D., and Tom Insel, M.D., at Emory University and by my laboratory at Rockefeller University. When influenced by sex hormones, larger amounts of the neuropeptide oxytocin meet larger concentrations of oxytocin receptors in the amygdala. Such a development is crucial, because Ferguson has shown that the administration of oxytocin to the amygdala fosters better social recognition. Young has also implicated a different neuropeptide, vasopressin, as important for fostering friendly social behaviors in laboratory animals. Working together, oxytocin and vasopressin encourage experimental animals to act toward each other with positive, friendly behaviors. This emerging neurobiology for normal harmonious social interactions in animals, including humans, provides a posi- tive set of mechanisms for ethical responses in a neutral or friendly situation. This adds to the mechanisms described earlier that apply to fearful situations. Even facing a high-stakes threatening situation, the great majority of humans will make an ethical response, because fear acts to do the job when the friendly hormones are not adequate to it. SHARED FEAR, AND BEYOND Usually, we have to recognize and remember differences between ourselves and others. For neuroscientists, my proposed explanation of an ethical decision by the would-be knifer is attractive because it involves only the loss of information, not its acquisition or storage. Learning and remembering complex information are difficult processes to understand. But the loss of information requires only the breakdown of any single part of the complex memory-storage processes, whether they are intricate biochemical adaptations, tricky synaptic growths, or precise temporal patterns of electrical activity. Eliminating or altering any one of the many mechanisms involved in social recognition or memory allows us to identify with the person toward whom we are about to act. Moreover, the specific mechanism that is affected could differ from person to person and from occasion to occasion. In mechanistic terms, therefore, it is incredibly easy to achieve a sense of shared fate with another. In my example of the potential knifing, because of a blurring of identity—a loss of individuality—the knifer temporarily puts himself in the other person’s place. Because that person would be afraid, so is he. He avoids an unethical act because of shared fear. Is that all there is to explaining why we obey the Golden Rule when the chips are down? No, a more positive approach also exists. While I have emphasized an avoidance of the negative through shared fears, I have also invoked brain mechanisms, such as the action of the neuropeptides mentioned above, to explain positive “affiliative” behaviors toward others. That is a story for another article, but it would begin with the brain and behavior research in animals that asks how people recognize and tell each other apart, as sketched above. Friendly behavior begets compassion and sympathy—words derived from the Latin and Greek for “feeling with”—and we can envision a brain-based explanation for positive behaviors that also involves shared fates. I have sought to explain behavior that follows an ethical principle when a just decision is to be made. But, as we know all too well, not everyone behaves ethically all the time. Why do some people find it harder to behave ethically than others? Why do some people have positive, friendly dispositions and others do not? Can neurobiology add to our understanding of what goes wrong when an ethical universal is violated whether as a petty unkindness or a horrible act of aggression? These questions, too, await further exploration by scientists. It may seem almost incredible that modern neurobiology has begun to explain the virtually universal ethical principle of treating others as we would wish to be treated. If this scientific explanation does provide an understanding of the underlying brain mechanisms, we should nevertheless remember that neuroscience is a new partner, not a replacement, for other intellectual approaches to the discipline of ethics. Certainly, the more we know about the mechanistic basis of empathy and ethical behaviors, the more effectively we can deal with situations in which we are saddened by their absence.

brain, ethics, gazzaniga, neuroimaging, imaging, ethical, network, circuit, thalamus, amygdala