Cerebrum Article

Memory: Pandora’s Hippocampus?

Greater knowledge of the human brain has enabled us to begin devising therapies to rescue or modify memory for the afflicted, such as Alzheimer’s patients or post-traumatic stress disorder victims. This same knowledge could also allow us to alter how normal, healthy memory operates; we may become able to enhance memory and learning through biological intervention. But the brain consists of complex, interactive networks, and unintended consequences could easily occur. Moreover, memory is woven into our individuality. Altering our memory processes therefore risks altering us fundamentally. We may not be able to resist opening this neuroscientific Pandora’s Box, John Gabrieli writes, but we must proceed with all the wisdom we can muster.

Published: October 1, 2004

Who we are” is what we know: We know the events of our lives, facts about the world, values that guide our choices, languages to communicate with one another, and skills to make our actions effective.

During the past 50 years, our knowledge about how the human brain learns and remembers has exploded, and along with that have come many implications. Most obvious are potential therapies for identity-destroying brain diseases, but also emerging quickly is the realization that this knowledge may allow us to alter how memory operates, even in the absence of actual disease. The possibility of such optional uses of our newfound understanding may prove to be a neuroscientific Pandora’s Box. We will not be able to resist opening it, but we should not fail to think about how tinkering with memory may change who we are and affect others we care about.

Lifting the Lid

Research has provided a new appreciation of the organization of memory in the human brain and provoked new ideas about how memory may be therapeutically rescued or altered. Many people will have a medical need to enhance memory capacity, such as the 14 million Americans expected to have Alzheimer’s disease by the year 2050. To help others, we may be able to meet a humane need to block specific memories—for example, in treating people who have experienced traumatic events and are reliving those horrors in post-traumatic stress disorder. In yet other cases, a valuable goal will be to weaken specific patterns of learning and memory, such as learned responses that trigger craving in recovering addicts or provoke rumination about sad experiences that further deepens the depression of depressed patients.

More controversial are possibilities of enhancing normal memory and learning through biological intervention above and beyond the morning cup of coffee. Although one can certainly argue for caution in developing superhuman memory, we will face more complex and urgent challenges in thinking about the continuum from healthy to diseased individuals. For example, we now appreciate that older people go through a long, slow process of brain and mental changes before crossing the diagnostic threshold of Alzheimer’s disease. Most researchers believe that early intervention may be essential to treat or postpone the disease, because so much brain injury has already occurred by the time an individual can meet diagnostic criteria for probable Alzheimer’s. When, then, would experimental treatment be justified? Would it be appropriate to try to alleviate any age-associated memory loss? Similarly, what if we know that an individual has genes that make him or her especially susceptible to a psychiatric disease? When would it be appropriate to intervene in memory mechanisms in an effort to prevent depression, post-traumatic stress disorder, and other psychiatric diseases now treated only after a person is deeply and persistently miserable? An additional challenge is that, in all of these cases, any treatment would initially have to be experimental with no certain efficacy and safety outcome.

Another important concern is the unintended consequences of any intervention. Modern research has shown that the brain functions in complex, interactive networks where activity in one brain region has great consequences for how other brain regions work. Thus, our humanity reflects these interactions, not only crossing boundaries of brain regions but also integrating the elements of human nature, such as memory, emotion, and personality.

Memory Systems of the Human Brain

In the 1950s, it was thought that the brain recorded, retained, and retrieved memories holistically, with the biological record of an experience etched throughout the central nervous system in an undifferentiated fashion. Memory for a specific experience or a particular fact would be like a drop of milk well-stirred into a glass of clear water—the whole glass of water would change color and the milk would be no more in one part of the water than in any other part. This view was overturned by the study of a single patient, arguably the most famous neurological patient of the twentieth century, known by his initials, H.M.

H.M. had severe and pharmacologically intractable epilepsy, a neurological disorder in which neurons fire without apparent reason and cause seizures. Many patients with epilepsy control their seizures with medications, but a minority respond poorly to medications and become candidates for surgical removal of the brain tissue where the seizures originate. For some patients, the procedure is highly effective. In the case of H.M., the origin of his seizures could not be determined, and physicians assumed it was the hippocampus, the most common, but not the only, brain region from which epileptic seizures originate. Therefore, in 1953, H.M. underwent removal of both the left and right hippocampi, as well as nearby structures such as the amygdala.

In terms of treatment for epilepsy, the surgery was successful. Although he continued taking medications, H.M. had very few seizures thereafter. However, the operation produced a devastating unintended consequence: From that day, H.M. was not able to form a lasting memory for any new experience or new fact. Although he was a man with average mental abilities, he could not remember information for more than a few moments. He did not know his age, the year, any historical event since 1953 (despite many hours of watching television), or that his parents had passed away. He did not remember any experience, no matter how emotionally powerful, and no fact, no matter how important or how many times repeated to him, for more than a few seconds. H.M.’s memories of his life before his surgery remained (he knew his name, his childhood and adolescent history), but he was virtually a blank slate for consciously accessible memories of events and facts since 1953. H.M.’s case was described as being one of global amnesia, because his inability to remember was so severe and so pervasive.

Fifty years of human and animal research have supported what H.M.’s doctors observed in the aftermath of his surgery—that the hippocampus and other structures located in the medial temporal lobe are critical for the formation of memories in the everyday sense of memory. Although pure global amnesia is rare, the consequences of hippocampal injury are frequent and profound. The vast majority of patients with Alzheimer’s disease have initial pathology in the same brain region, which is why memory loss is the most common and severe early difficulty in that disease. Furthermore, research has found indications that many diseases, including schizophrenia and post-traumatic stress disorder, affect the hippocampus.

So complete was H.M.’s amnesia that scientists were initially surprised to discover that some kinds of learning remained fully preserved in H.M. and similar amnesiac patients. In subsequent years, studies documented that various kinds of learning—including perceptual, motor, and cognitive skills and other forms—are normal in amnesia. Further research indicated that several such forms of learning are mediated by other neural circuits, including the basal ganglia, cerebellum, and neocortex. We now think the entire human brain has learning capacities, with each brain region highly specialized for learning specific kinds of information— not unlike a symphony orchestra, in which each instrument makes a specific contribution to the music. The hippocampus and related structures stand out, however, because of their critical importance for the everyday sense of memory and their susceptibility to injury in diseases that affect many people.

Emotions and Memory

Emotions have a powerful influence on memory. Psychological experiments verify our personal observations that we remember emotionally intense experiences more often and in more detail than less intense experiences. (Although emotionally fueled memories seem as susceptible to error and misremembering as neutral memories, emotional experiences have a better chance of not being forgotten.) It is striking that the limbic area of our brains (the structures that form a border around the brain stem) includes in a close anatomical neighborhood circuitry that is essential for both emotion and memory. Emotionally powerful experiences, be they fearful or delightful, may merit special consideration for being remembered.

H.M. and patients like him have injuries to structures adjacent to the hippocampus, including the entorhinal cortex (where Alzheimer’s pathology is believed to originate) and the amygdala. A remarkable convergence of animal and human studies, however, has identified one structure, the amygdala, as a specific link between emotion and memory. When healthy research subjects are shown films or slides (scientists in the laboratory are not allowed to induce truly powerful emotional situations), they remember emotionally intense material better than neutral material. By contrast, patients with injury to only the amygdala remember neutral material normally, but they specifically fail to enhance their memory of emotionally powerful material. In brain imaging studies with healthy adults, amygdala activation during the viewing of visual material correlates with subjective experience —the more intense a person perceives a picture to be, the more amygdala activation occurs. Further, the greater the amygdala activation during the viewing of an emotionally intense picture, the greater the likelihood that the person will remember the picture weeks later. But this applies to emotionally intense experiences only—the amygdala appears to have little role in any but the most intense events. Thus, the amygdala appears to adjust the formation of enduring memories on the basis of emotion.

Amygdala enhancement of memory appears to occur for both negative and positive experiences, although we have more evidence for negative experiences (perhaps because they are easier to induce intensively in the laboratory). Interestingly, amygdala-driven memories in humans appear to have a trade-off: The key or central aspects of an emotional experience are better remembered, but the peripheral aspects of the experience are less well remembered than they are for more neutral experiences. It is as if the emotionally charged information overshadows other aspects of the same experience.

Functional brain imaging has recorded amygdala dysfunction in many psychiatric disorders, including depression, social phobia, and anxiety. However, we do not know whether the amygdala dysfunction is part of the cause of the disease or the consequence of the disease. That is, if another part of the brain were transmitting dysfunctional information to the amygdala, the amygdala could be working normally but appear pathological in response to the other brain region.

Individual Differences in Personality, Sex, Age, and Genes

The brain is not only the physical basis for aspects of human nature we all share, such as dependence on the hippocampus to form new memories, but also the basis for the neurology of individuality—how we are unique. We differ from one another on many dimensions, including personality, sex, age, and genes. Research during the past decade has begun to uncover how these seeds of uniqueness influence what we remember, and thus who we are.

“Personality” refers to stable psychological characteristics such as extroversion or introversion that influence an individual’s behavior across different situations and over time. These enduring traits or predispositions are not the same as the fleeting states of feelings. Personality researchers have developed questionnaires that reliably measure specific personality traits on a continuum (how you score today is similar to how you score next week or next year), and these traits correlate with various behaviors and health outcomes. However, considerable debate continues among psychologists about the relative power of personality versus situations in influencing how we behave—to what extent is a shy person shy across all situations, or instead shy among strangers at a party but outgoing and aggressive at work?

Brain imaging studies have begun to show the neural mechanisms by which personality, emotion, and memory may interact with one another. When viewing an equal number of negative and positive pictures, the more extroverted a person is, the greater the amygdala response to positive pictures (the more introverted, the greater the response to negative pictures). One can imagine that an extrovert is outgoing and sociable, in part because she or he more powerfully remembers positive experiences, whereas the introvert may be more aloof because she or he more powerfully remembers negative experiences. Thus, if one’s personality filters experience, we may learn quite different lessons about life depending on what we remember as positive or negative.

Another dimension of difference is whether we are women or men. This starts with a genetic difference, but powerful socializing influences enter as we learn about our expected gender roles. Unexpectedly, two studies found that activation in the right amygdala predicted what negative pictures men would later remember, whereas activation in the left amygdala predicted the same thing for women. One of the studies also found that, while men and women were largely equal in memory performance, women had superior memory for the most intensely negative pictures and both men and women increasingly activated the left amygdala as they found pictures to be increasingly intense. Speculatively, the brain imaging evidence suggested that emotional evaluation and memory formation are more tightly coupled in the brains of women than of men. (Overall, men and women had a similar amount of brain activation for emotional versus neutral pictures).

Age is another source of uniqueness, in how we differ not only from one another but even from our former, younger selves. Memory formation declines mildly in healthy aging, but some research shows that older people better retain memory for emotionally positive than negative experiences. This increasing emphasis on maintaining a positive emotional disposition can be interpreted as a sort of emotional wisdom. Brain imaging has begun to specify how the older brain may wisely emphasize the positive. For example, in one study, younger (around age 20) and older (around age 70) adults viewed positive, negative, and neutral pictures. In the young adults, amygdala activation was greater for both negative and positive pictures than for neutral ones, but the amygdala of older adults showed a selective reduction in response to negative pictures (and the older adults had far less memory for the negative pictures). Thus, it appears that a lifetime’s experience encourages older people to disengage emotion-driven memory formation for negative experiences, whereas younger adults form emotional memories equally for positive and negative experiences.

Finally, everything we know comes from only two sources, our genes and our experiences; our brains are formed under genetic instructions, then shaped by what we experience. Both genes and experience exert their influence on our behavior by sculpting our brains. The revolution in genetics now allows us to characterize specific single nucleotide polymorphisms (SNPs) that vary from one healthy person to another. When brain-imaging studies have grouped people on the basis of these single genetic variations, they have found activation differences in the hippocampus (and in memory performance) for one gene and in the amygdala for another gene. These exciting studies have opened up a new frontier in which we may begin to explore the relations between genes, brains, and minds.

In H.M., we saw a man mentally frozen since 1953 because he could not retain new information. He is evidence that memories make us who we are. But, perhaps equally important, who we are— men or women, extroverted or introverted, younger or older, having this or that variant of gene—may also determine what we remember. The brain imaging that has uncovered the varied strength and content of individuals’ memories depending on their sex, age, and genetic endowment is showing us that memory is not an add-on to who we are (as when we add memory capacity to a computer) but woven into a fabric of our individuality. Thus, when we talk about altering memory processes we may be talking about altering our fundamental individuality.

Can You Erase Memories?

In the movies, memories are erased from the human brain for both benign and nefarious reasons. In certain situations, therapeutically blocking access to a memory may be desirable, such as memory for traumatic experiences or learned responses to cues for drug craving. We have known for some time that certain medications or electrical stimulation (or head injury) can obliterate recent memories, but these methods cannot be targeted to a specific memory and are difficult to control precisely in time. Thus, it may be possible to pharmacologically block access to a traumatic memory from a year ago, but these methods would also block access to all other memories from the past year (or more).

In functional brain imaging studies of selective suppression for a specific memory, healthy young adults could learn to suppress specific memories of neutral word pairs. This suppression was characterized by selective activation of prefrontal cortex (an area of the brain important for goal-oriented control of cognition) and deactivation of the hippocampus. How this experimental study relates to real-life traumatic memories is still unknown, but what we know about the human brain suggests that the intentional suppression of unwanted memories would involve the turning up of control of one’s thoughts (prefrontal cortex) and the turning down of memories (hippocampus). Several interesting studies have shown that people can use real-time feedback from brain imaging to learn to increase or decrease activation in specific brain regions and, presumably, the mental operations supported by those regions. One can imagine that people could learn, using brain feedback, how to suppress specific memories. Furthermore, certain drugs may be able to augment this process.

Currently, these ways of suppressing memory are voluntary, but increased understanding of such mechanisms may point towards involuntary methods by which others may choose to suppress our memories. Even well-intentioned uses of memory suppression may have profound consequences. For example, emergency treatment of a rape victim or battlefield treatment of a soldier may preempt emotional suffering if a traumatic memory is immediately suppressed. The unintended consequence of such memory suppression may be the victim’s inability to remember information required to convict the rapist or the soldier’s inability to learn from a disturbing experience.

Can You Have Too Good a Memory?

The student may not exist who, preparing for an examination, would not want at that moment a pill to confer photographic memory. More generally, traumatic and embarrassing memories aside, most of us are more irked by things we forget than by things we remember. But can one have too good a memory for one’s own good?

Although we know a great deal about how crippling the loss of memories can be in amnesia and Alzheimer’s disease, we know very little about the consequences of a memory that is too good. Truly photographic memory is very rare (most people with superior memory use long-known mnemonic devices that require considerable effort to apply). In the 1920s, the noted Russian psychologist A.R. Luria carefully studied one case of truly photographic memory, in a person known as S. This man performed virtually perfectly on all memory tests. Indeed, he could recall random lists of numbers months and years after seeing them. He appeared to have a photographic and unlimited memory. Until others studied him, S. was unaware that his memory was unusual, but once he grasped his unique gift, S. became a mnemonic performer, dazzling audiences with his perfect memory.

However, S. could not control his memory. When he was reading or talking with others, words would evoke visual memories so powerfully that S. would have a difficult time attending to the meaning of the words. He could not control powerful memories from rising up in his mind and blocking more abstract interpretations, the idea of what was going on. So intrusive were these vivid memories that S. tried desperately to erase them from his memory by writing them down and then throwing out or burning those papers. He could not, however, throw out the memories that continued to flood his mind.

We now understand that our memories (except for the case of S.) are abstract in nature. We remember the gist of what we see, hear, or read, not the specific visual or auditory details of each experience. This sort of abstraction has rewards and risks. The major reward is that we instantly relate our abstract knowledge of the world to the situation at hand—a lifetime of experience is used to instantly translate the significance of a physical experience into an abstract interpretation of what is going on. The major risk is that the physical details of an experience are thrown quickly away in favor of an efficient interpretation. This makes us prone to the dangers of interpretation— such as false or illusory memories or self-serving biases in which we inadvertently substitute our interpretations for our actual experiences. This natural blending of reality and interpretation extends over time, so that each reconsideration of a memory, be it spontaneous or through interviews or in therapy, alters the memory itself. This is why eyewitness testimony in the court can be more riveting than accurate, and why it has proved to be excruciatingly difficult to validate recovered memories of childhood abuse.

Memory is often thought of as a fragile power, because its balance of recording and interpretation allows us to remember all that we remember, but also to easily forget or misremember. The case of S. demonstrates the risk of memory becoming too powerful —memories of the past flood our minds and drown a clear sense of the present. While a world full of people with perfect memories seems far away, great efforts are being made to develop medications that can boost memory in Alzheimer’s disease. To date, these medications have had, at best, modest benefits for patients with Alzheimer’s, but these patients offer a severe challenge for treatment because their brain injuries demand a remarkable drug effect. But could healthy older people use some of the same prescription medications to boost their memory, with the medications having potentially greater effect in a healthy brain? Or could healthy young people assist their mental performance and encourage others to do so just to keep up (as we have seen with steroid and other drug use in athletes)? What effect would such drug-facilitated memory enhancement have on identities? It is possible that a modest gain in the accuracy of our memories would be more than offset by an imbalance with our emotions, personality, or age—for example, an older person may remember more distressing information and become less happy, without gaining other sources of happiness that are part of youth.

Lifting the Lid Cautiously

Pandora discovered that opening her box satisfied her curiosity but unleashed misery upon the world. Scientists have studied the brain basis of human memory, and especially hippocampus-dependent memory for events and facts, in order to understand how we remember and forget. Our understanding of H.M.’s amnesia fueled our insight into the memory failure of Alzheimer’s disease. Functional neuroimaging in human studies has allowed for visualization of memory functions in health and in many diseases, and the scientific literature on memory abounds with the spectacular progress in animal and molecular neuroscience.

We have an ethical imperative to use this research knowledge to inform treatment of those with diseases of memory. These exciting opportunities, however, are associated with risks of abuse or unintended consequences. Some potential consequences are striking, such as how being able to extinguish a memory might thwart desirable societal goals. Others would be more subtle and perhaps more troubling: a treatment that simply boosted memory could make older people less happy; learning to easily enhance or suppress memory raises the specter of life narratives rewritten for nefarious or unworthy reasons; and, most important, the power to manipulate memory at will, our own or another’s, is the power to alter our sense of who we are as men and women, or in terms of our personality. The potential gains of improving or therapeutically altering memory are compelling, but each step we take to put these advances to use must be accompanied by forethought about ethical dimensions of the brain basis of human individuality and about unintended consequences of manipulating the evolved balance of memory processes. Thinking ahead may allow us to open this box more wisely.