Thanks for the Memory


The human brain has been described as having the consistency of tofu or soft butter, and as being like a three-pound Brie. It has been compared to a computer, though that’s a misguided analogy since the brain does not operate through digital logic. Nor is its content—what we call knowledge—discrete. The brain is dynamic and plastic, changing in response to whatever comes its way. This is not a metaphor. Encounter something once and it is foreign to you. Encounter it many times and it is familiar. The thing itself hasn’t changed; your brain has. Experience has laid down new neural pathways. They are biochemical and electrical. They are real. Within limits, they can be observed and measured.

Looking at brains, cell by cell, is a relatively new phenomenon. When the neuroscientist Eric Kandel was a medical intern in the 1950s it hardly had been done. Nor was it Kandel’s ambition—not directly. Rather, he wanted to find the places in the brain where the ego, the id, and the superego reside. When one of his advisers, Harry Grundfest, suggested that to really understand the mind, one had to study the brain “one cell at a time,” Kandel took up the gambit. In so doing, he became, over the next half-century, one of the preeminent neuroscientists in the world. In 2000, he shared the Nobel Prize in Physiology or Medicine for his work on the cellular basis of learning and memory.

Those fifty years saw exponential leaps in the understanding of how the brain works. Concepts were worked out, like “the neuron doctrine,” that the nerve cell is the building block of the brain, and “the ionic hypothesis,” which describes how charged atoms called ions, traveling across the cell membrane, can generate electrical signals that carry information within and between tissues. New technologies that allowed doctors and scientists to see brains in vivo were invented, and the armamentarium of psychotropic drugs was developed, all coincident with Kandel’s rising career. As a consequence, his memoir, In Search of Memory, is an intimate tour of modern neuroscience. It is also a kind of intellectual joke: here is a book about the discovery of the biological basis of memory that has been written, essentially, out of one man’s prodigious recollections. To read it is to appreciate what he has accomplished.

Kandel is the author of six earlier books, including the standard textbook on neuroscience, an encyclopedic volume that contains surprisingly felicitous writing for what it is. As with that book, Kandel uses In Search of Memory as a podium, a place from which to deliver a series of lessons on the basic science of mind, but to a more general audience. As a consequence, In Search of Memory is largely an exercise in translation, and there is a limit to how much of Kandel’s native language, a patois of cell biology, genetics, biochemistry, neurology, psychiatry, pharmacology, and electrophysiology, can find expression in ours. Most of the time this doesn’t matter—it’s Kandel’s enthusiasm that’s driving the story—but sometimes, just as he’s getting to the critical moment, the moment where everything is revealed, words fail him, and the reader, too.* Thankfully the book comes with an extensive glossary, though its definitions can also be arcane.

Eric Kandel was born in Vienna in 1929, the son of a toy shop owner. Jewish and well off, a son of the bourgeoisie, he was listening on a homemade shortwave radio when Hitler marched into Austria on March 12, 1938. Two days later Hitler reached Vienna where, Kandel recalls, he was greeted by wildly enthusiastic crowds. Kandel and his brother were made to withdraw from their school and to attend one for Jews only. Their father’s store was taken from him, and the family forced out of their home. Within days of Hitler’s invasion of Austria, Kandel’s parents made plans for the family to leave the country. It took a year for them to acquire the documents they needed, but in April 1939 Kandel and his brother emigrated to the United States. Their parents came a few months later. Eric Kandel was ten years old.

In Brooklyn Kandel was enrolled first in a public elementary school, where he picked up the language, and then a Jewish day school, where he became fluent in Hebrew as well. His father, meanwhile, went to work in a toothbrush factory, saved enough money to open his own clothing store, and eventually bought the building in which the store and the family apartment were located. It was one kind of American success story. Eric Kandel’s was another. From the Yeshiva of Flatbush he went to Erasmus Hall High School, and then to Harvard, where he studied history. At Harvard he met Anna Kris, whose parents, Ernst and Marianne Kris, were Cambridge psychoanalysts, and whose grandfather, Oskar Rie, was a close friend of Freud. Kandel began to spend time in the company of the Krises and their colleagues, and “was converted to their view that psychoanalysis offered a fascinating approach, perhaps the only approach, to understanding mind.” Under their influence, he decided to attend medical school, in order to become a psychiatrist. Then he did an elective course in the neurophysiology lab of Harry Grundfest at Columbia University where he encountered the brain and the study of memory. “It was clear to me that learning and memory are central to psychoanalysis and psychotherapy,” he writes, looking back. “After all, aspects of many psychological problems are learned, and psychoanalysis is based on the principle that what is learned can be unlearned. In a large sense, learning and memory are central to our very identity. They make us who we are.” Still, when he tells his story, it is the pursuit of pure science that animates him—hearing for the first time the pop-pop as a neuron fires in the brain of a crayfish, finding his “voice” as a scientist, solving the biggest of puzzles with the smallest pieces (cells, molecules, genes).

Just as his personal story, like anyone’s, is a random mix of chance and intention, so, too, the story he tells about the development of brain science. Here is Luigi Galvani, in 1791, hanging frogs’ legs on copper hooks from an iron balcony, the legs twitching, and Galvani’s discovery of electricity in the body—and here, more than a century later, is its descendant, the ionic hypothesis. Here is HM, a young man whose bike accident at the age of nine left him so severely epileptic that a surgeon felt compelled to remove a huge section of his temporal lobe and hippocampus in 1953 in a desperate effort to control his seizures—an operation that also excised his ability to remember new things. (HM is still alive, living in a Connecticut nursing home.) And here is Eric Kandel, even today, exploring questions raised by HM’s condition: By what biochemical processes is a memory made, and how might those processes be modified?

Brenda Milner, the Canadian neuropsychologist who worked most closely with HM, showed that memory is not just one thing—that there are various kinds of memory, each of which accomplishes very different tasks and resides in distinct parts of the brain. There is, for instance, declarative memory, the memory for facts, for people, for temporal events—the kind of memory that HM lost when his hippocampus was removed—and there is procedural memory, the muscle memory, say, that allows us to ride a bike or drive a car, the memory that comes from habit, which HM retained. For Kandel, who had taken a job at the National Institute of Mental Health (NIMH) around the same time the analyses of HM were first published, Milner’s findings suggested a place to begin his research: in the hippocampus, looking for cells that store specific memories. Milner had discovered the where. Kandel was searching for the how.

In collaboration with Alden Spencer, another young NIMH researcher, Kandel began research on hippocampal neurons, trying to understand how they stored memories. Decades later he recalls the twenty-four-hour days, each experiment, all of its iterations. (He seems to have a convenient hole in his memory, however, for the failures, the inevitable boredom, the egos, the more than sporting competitiveness. This is, for the most part, a sunny tale.) Twelve months into their pursuit, Kandel and Spencer made a seminal observation: that neurons in the hippocampus

were not sufficiently different from those of spinal motor neurons to account for the ability of the hippocampus to store memories. It took us a year to realize what should have been obvious from the start: the cellular mechanisms of learning and memory reside not in the special properties of the neuron itself, but in the connections it receives and makes with other cells in the neuronal circuit to which it belongs.

Those connections were reinforced when either chemicals called neurotransmitters or an electrical charge jumped the gap between neurons, binding them. The gap, called the synapse, was bridged, typically, from an offshoot at the end of one neuron called an axon, to a receptor in the other called a dendrite. The better the connection, the more secure the circuit, the stronger the memory.

The observation that memory existed between cells rather than inside them brought Kandel to a new objective: to understand how sensory information reaches the hippocampus and what happens to it when it does. Kandel turned to the giant marine snail, Aplysia, an animal that first entered recorded history in the work of Pliny the Elder, to try to figure it out. With a mere 20,000 neurons (as opposed to our 100 billion), clustered together in groups called ganglia, some of which are visible to the naked eye, the sea snail was an ideal reductionist laboratory in which Kandel could carry on research. Taking a cue from Pavlov’s work on behavioral responses to stimuli, he decided to look at how an individual brain cell would change in response to particular patterns of electrical pulses. One kind of pulse simulated habituation—as when a dog, upon hearing a bell ring again and again, learns that it is harmless. Another replicated sensitization, the opposite of habituation—as when the dog is given a shock to its paw and pulls it away. The third possibility, classical conditioning, involved the pairing of a neutral and an averse stimulus—as when the shock is administered at the same time as the bell is rung, and the dog comes to associate the sound of a bell with a shock to his paw and recoils upon hearing it, even in the absence of the shock. Kandel writes:

I wanted these neural analogs to answer one key question: How are synapses changed by different patterns of carefully controlled electrical stimuli?… The experiments clearly showed that synaptic strength is not fixed—it can be altered in different ways by different patterns of activity. Specifically, the neural analogs of sensitization and averse classical conditioning strengthened a synaptic connection, whereas the analog of habituation weakened the connection.

Reductionism served Kandel well. (It also inspired a host of other neuroscientists to look for clues about human cognition in animals with very simple nervous systems.) Sifting through the two thousand neurons in a single Aplysia ganglion, he and his collaborators made the novel observation that the strength of synaptic connections, rather than being fixed and constant, is augmented or weakened by experience, as with his observation that habituation weakens the connection and sensitization enhances it. Plasticity—the capacity to change—is built into the molecular architecture of the synapse. In other words, who we are, our essence, is mutable by learning. Kandel began to make an inventory of Aplysia’s different behaviors, looking for one that he could manipulate, in an effort to see, firsthand, the synaptic effect of learning. He ended up choosing the snail’s most basic response, its gill-withdrawal reflex: touch its body near the gill—its breathing mechanism—and Aplysia quickly and defensively retracts its gill. Touch its body near the gill numerous times, though, and it stops withdrawing the gill; the snail has gotten habituated to being stroked, which it no longer perceives to be threatening. Shock it with an electrical pulse, however—Pavlov’s sensitization—and it quickly and emphatically pulls back again.

Kandel and his colleagues also found that to establish long-term memory in Aplysia—to make the learning stick—it was best to train the snail slowly, over several days, rather than intensely, at once. Touching it forty times in a row, for instance, caused the snail to become habituated for a single day. When those forty touches were spread out over ten days, however, the snail stayed habituated—it didn’t withdraw its gill—for an entire week. So how did that happen? How did that short-term memory become a long-term memory? What was going on in the snail’s brain cells?

It had taken Kandel fifteen years to be in the position to answer those questions. He knew from his research that long-term memory was not simply enhanced short-term memory, that not only did the synaptic changes in long-term memory last longer, but the number of synapses changed in response to stimuli. (The number went down in habituation and up in sensitization.) Kandel and his group had a hunch—that the release of one neurotransmitter, serotonin, augmented a second neurotransmitter, glutamate, causing a kind of biochemical domino effect in the cell. In a paper published in 1971 in the Journal of Neurophysiology, laying out their theory, they further speculated that a molecule called cyclic AMP was causing the cells to retain those changes.

Kandel and his associates did not come across cyclic AMP by chance—it was already known to amplify the biochemical signaling in fat and muscle cells. They surmised that it might be doing something similar in the brain as well. Already, in 1968, a researcher at the University of Washington had shown how that amplification worked: cyclic AMP activated an enzyme called protein kinase A, which in turn acted like a switch, turning some proteins on and some proteins off. In the case of Aplysia, Kandel found that a shock to its tail not only released the neurotransmitter serotonin, but that serotonin then stimulated the production of cyclic AMP. Later, when a postdoctoral student in Kandel’s lab injected cyclic AMP directly into Aplysia’s sensory cells, not only did the amount of glutamate increase, but the synapse between the sensory neuron—the neuron registering the shock—and the motor neuron—the one that signaled the gill to retract—was noticeably stronger. Here, at last, “were the first links in the chain of biochemical events leading to short-term memory storage,” Kandel recalls. Twenty-four years later, that work won him the Nobel Prize.

From cyclic AMP, Kandel proceeded to study the genes that throw the switch on long-term memory formation, and through working with those genes, he became interested in making genetically altered mouse models of memory dysfunction to study diseases like Alzheimer’s and schizophrenia. Mice brains are anatomically similar to human brains, so from his work with mouse models he began to devise ways to study consciousness, looking at how a mouse navigates the physical world, and how the physical world is represented in its brain and affected by attention. It is not the ego, the super-ego, or the id, but it is, he writes, one of “the larger questions that had attracted me to psychoanalysis at the beginning of my career.” So Grundfest was right after all.

Still, he could not have known. Any life story relies on a false linearity: one thing happens, and then another, and then the next, and when those events are laid out in that order they seem logical and determined. But little of what Kandel has accomplished in his career could have been sought, or even conceived, when he set out half a century ago, since the science he has relied on along the way typically has been only a few steps ahead of him. Neuroscience—a field that did not exist when he began his career, and which his career has helped define—is an amalgam of disciplines, molecular biology, cognitive psychology, genetics, electrophysiology, and biochemistry among them. Kandel’s experiments have piggybacked on the remarkable breakthroughs in many of those fields, such as the discovery of how to insert genes into cells and how to make transgenic mice, i.e., mice whose genome has been altered. It has been his particular gift to take these new tools and figure out how to apply them to the study of memory.


Kandel’s lasting excitement at making the cyclic AMP finding, and the great joy he has found in pursuit of neurological truths, stand in stark contrast to Katrina Firlik’s wan account of her work as a neurosurgeon, Another Day in the Frontal Lobe: A Brain Surgeon Exposes Life on the Inside. Maybe this is because Firlik has been working at her profession for less than a decade. Or maybe it has something to do with the difference between doing science and practicing medicine, especially surgery, which has the reputation, at least, of being more mechanical than thoughtful, a stereotype Firlik does little to dispel. Firlik, clearly, is a well-educated doctor. She was the first woman to train in the University of Pittsburgh’s neurosurgery department and she was chief resident there. Now, in addition to practicing surgery at Greenwich Hospital, in suburban Connecticut, she teaches at Yale Medical School. Her academic credentials are top-notch—a reminder that insight can’t be taught.

Why Firlik decided to become a neurosurgeon is rather vague. Her father was a surgeon, she writes, so she gravitated in that direction—but what kind of surgeon? “The brain was definitely more interesting than the kidney (or the heart, or the bones, or the skin…),” she writes, and

the stories of neurologist Oliver Sacks gently tapped me over the edge…. I asked other neurologists what they thought of Oliver Sacks. Some felt he was a good storyteller but a run-of-the-mill neurologist. I found these comments cruel and unfair, first because these neurologists didn’t really know him and, second, because they were probably just jealous.

But in the end neurology didn’t interest her much, either, so neurosurgery—with its seven years of training—was it.

And here she is, all these years later, picking out bullets and removing brain tumors and fixing spines. (It turns out that a lot of neurosurgery is performed on the back.) Yes, she admits, she was once reduced to tears by her diagnosis of a patient—never again—and the cost of malpractice insurance is outrageous; and there have been some thorny issues, such as operating on neurologically impaired babies without having a good sense if they will be able to have any semblance of a sentient life; but all in all Firlik reports that neurosurgery has worked out just fine for her. The operating room has turned out to be a congenial workplace, where people listen to music and talk sports and trade stock tips. “I remember listening to a live radio program during a case,” she writes.

Everyone wanted to catch the O.J. Simpson verdict in real time. There was a collective gasp in the room at the pronouncement of innocence, and then back to work. (By the way, there is no need to worry about the patient’s safety here. There are key moments of concentration during certain cases when everyone in the room knows not to talk—such as when an aneurysm at the base of the brain has just been isolated and is about to be clipped—but surgery does not otherwise demand absolute silence, and discussion does not have to be limited to immediate surgical concerns. The OR can be a lively social environment, unless the surgeon is in a rotten mood.)

All of this is vaguely amusing, until Firlik stumbles onto the subject of bioethics, a topic that, for neurosurgeons in the ever-nearing future, is destined to involve the possibility of performing a host of surgical enhancements to the brain, including what she blithely calls “brain lifts” to tone an aging brain. (Some procedures, like transcranial magnetic stimulation, are already being done.) Content, herself, to stay on the sidelines of whatever debates these procedures generate—will the operations be regulated, who gets to decide, does distributive justice matter here, should doctors treat normal conditions?—her willful passivity is chilling. Still it is no more so than her stated point of view, that taking account of the social implications of neurosurgery misses the more important point: these operations make patients happy. Writing of a successful Manhattan neurosurgeon she knows whose practice includes implanting electrical stimulators under the scalp to provide constant low-grade stimulation in an attempt to boost memory function (he also offers patients massage and pedicures), Firlik points out that

the ethics behind cognitive enhancement is the one deepening wrinkle…. Academicians—many of whom have never even spoken to satisfied clients…—claim that cognitive enhancement threatens to broaden the socioeconomic gaps in society…. Plastic surgery triggered similar debates years ago, but the debates didn’t last.

Unlike Katrina Firlik, Eric Kandel is not glib about the future of cognitive enhancement; but in In Search of Memory, at least, he has surprisingly little to say about it. He raises the issue, then begs off, citing a paper he and a number of colleagues published in Nature Reviews Neuroscience two years ago, whose conclusions he then fails to share. Readers who make the effort to track it down will find that even the best scientific minds are perplexed at how, if at all, neurocognitive enhancements should be controlled. The paper’s authors suggest that there should be some measure of regulation, but they are unsure what it should be, or how far it should go. Though they recognize that doing nothing is, essentially, doing something, Kandel and his fellow authors end their discussion by proposing that the best course of action is to keep talking.

Like many neuroscientists who patent their scientific findings in the hopes of using them to develop drugs and other therapies, Kandel has more than a passing interest in discussions about neuroethics. In 1996 he and some colleagues founded Memory Pharmaceuticals, a publicly traded company that has four drugs in development, two of which exploit Kandel’s cyclic AMP work to treat memory and psychiatric disorders. Because nearly everyone is a possible candidate for drugs that promise to enhance cognitive functioning and improve memory, the potential of such drugs is tremendous.

In the case of Memory’s cyclic AMP– based drugs, however, science might intervene before the market does. Last year, somewhat mysteriously, the giant pharmaceutical company Roche pulled out of its partnership to develop Memory Pharmaceuticals’ two cyclic AMP drugs (though it retained the option to jump back in if the company continued to pursue them and got good results). Though Roche offered no explanation for its action, an article in the journal Neuron in November 2003 offers a clue. It suggests the limits of relying on reductionist science when making the leap from describing molecular pathways to manipulating them. According to one of the authors of the Neuron paper, Amy Arnsten, a professor of neurobiology at Yale (who has her own drug company affiliation with the British company Squire), relying on a simple animal like Aplysia, which has no hippocampus and no prefrontal cortex (the part of the brain that, in humans, largely makes us who and what we are), provides only a partial view of what’s going on with cyclic AMP in the brain. Arnsten, who works with mice, rats, and primates, whose brains more accurately resemble ours, found that the same compounds that boosted cyclical AMP in their hippocampus had devastating effects on their prefrontal cortex, which controls attention, working memory (memory for, say, telephone numbers and addresses), and executive functions like organization and planning.

What this suggests is that a drug that may enhance one part of the brain could easily and equally impair another. It is also a reminder that the brain, while composed of constituent parts, is a global and holistic system. Search it and you will find memory in the hippocampus and also in the prefrontal cortex and in the dentate gyrus and in the amygdala. Memory is not one thing and it does not reside in a single corner of the brain. If we are sometimes—often-times—haunted by memory, it is because it inhabits us.


They Were in New York December 21, 2006

  1. *

    For example, in trying to explain how the molecules called cyclic AMP affect memory, he writes as follows:

    When one molecule of chemical messenger binds to a metabotropic receptor, that receptor stimulates adenylyl cyclase to make a thousand molecules of cyclic AMP. Cyclic AMP then binds to key proteins that trigger a whole family of molecular responses throughout the cell. Finally, adenylyl cyclase continues making cyclic AMP for minutes. The actions of metabotropic receptors therefore tend to be more powerful, more widespread, and more persistent than the actions of ionotropic receptors. Whereas ionotropic actions typically last milliseconds, metabotropic actions can last from seconds to minutes—one thousand to ten thousand times longer.