by Sam Kean
Milner remembers the star test as a eureka. Before this, neuroscientists thought of memory as monolithic: the brain stored memories all over, and all memory was essentially the same. But Milner had now teased apart two distinct types of memory. There’s declarative memory, which allows people to remember names, dates, facts; this is what most of us mean by “memory.” But there’s also procedural memory—unconscious memories of how to pedal a bicycle or sign your name. Tracing the stars proved that H.M., despite his amnesia, could form new procedural memories. Procedural memories must therefore rely on distinct structures within the brain.
This distinction between procedural and declarative memories (sometimes called “knowing how” versus “knowing that”) now undergirds all memory research. It also sheds light on basic mental development. Infants develop procedural memory early, which explains why they can walk and talk fairly quickly. Declarative memory develops later, and its initial weakness prevents us from remembering much from early childhood.
Another distinct type of memory emerged from Milner’s tests as well. One day Milner asked H.M. to remember a random number, 584, for as long as possible. She then left him alone for fifteen minutes while she had a cup of coffee. Contrary to her expectation, he still knew the number when she returned. How? He’d been repeating it under his breath, over and over. Similarly, H.M. could remember the words “nail” and “salad” for several minutes by imagining a nail piercing some salad greens and reminding himself over and over not to eat the impaled leaves. Any distraction during those minutes would have ejected the words clean out of H.M.’s mind, and five minutes after the test ended, even the memory of having to remember something had vanished. Nevertheless, as long as H.M. concentrated and kept refreshing his memory, he could hold on. This was the first clue that short-term memory exists; moreover, it showed that short-term memory (which H.M. had) and long-term memory (which he lacked) must utilize different brain structures.
After Milner’s discoveries, H.M. became a scientific celebrity, and other neuroscientists began clamoring to explore his unique mind. He did not disappoint. In April 1958, five years after the operation, H.M. and his parents moved into a small Hartford bungalow. In 1966 a few American neuroscientists asked him to draw the home’s floor plan from memory. He succeeded. He didn’t know the bungalow’s address, but walking through its six rooms over and over had tattooed the layout into his brain. This proved that our spatial memory systems, while normally reliant on the hippocampus, can circumvent it if need be (probably via the parahippocampus, a nearby navigation center).
Scientists also discovered that time worked differently for H.M. Up to about twenty seconds, he reckoned time as accurately as any normal person. After that, things veered wildly. Five minutes lasted, subjectively, just forty seconds for him; one hour lasted three minutes; one day fifteen minutes. This implies that the brain uses two different timekeepers—one for the short term and one for everything beyond twenty seconds, with only the latter suffering damage in H.M. Once again, H.M. allowed scientists to break a complex mental function down into different components and to link those components to structures in the brain. Eventually more than one hundred neuroscientists examined H.M., making his probably the most studied mind in history.
All the while H.M. got older, at least physically. Mentally, he remained stuck in the 1940s. He remembered not a single birthday or funeral after that time; the Cold War and sexual revolution never registered; new words such as granola and Jacuzzi remained forever undefined. Worse, a vague sense of uneasiness often bubbled up inside him, and he could never quite shake it. The feeling, Milner reported, was “like that fraction of a second in the morning, when you are in a strange hotel room, before it all falls in[to] place.” Only for H.M. it never did.
In 1980, after H.M.’s father died and his mother got too sick to care for him, he moved into a nursing home. He walked a little gimpily by that point: years of taking heavy-duty epilepsy drugs had withered his cerebellum, and his wide, shuffling gait resembled that of kuru victims. He also got pretty portly after too many forgotten second helpings of cake and pudding. But overall he was a fairly normal patient and lived a (mostly) placid life. He loafed through the nontesting days reading poems or gun magazines, watching trains rumble by, and petting the dogs, cats, and rabbits the facility owned. He learned how to use a walker, thanks to his intact motor memories, and he even attended his thirty-fifth high school reunion in 1982. (Although he recognized no one there, other attendees reported the same problem.) When he dreamed at night, he often dreamed of hills—not of struggling up them, but cresting them and being at the top.
Still, the old, volatile H.M. did flare up now and again. He sometimes refused to take his meds—at which point his nurses scolded him, warning him that Dr. Scoville would get angry if he disobeyed. (That Scoville had died in a car crash didn’t matter. H.M. always fell for it.) He got into fights with other residents as well. One harpy at the nursing home would erase his bingo card midgame and taunt him. H.M. sometimes responded by running to his room and either banging his head on the wall or grabbing his bed and shaking it like a gorilla would its cage. One fit got so violent that his nurses called the police. These were moments of pure animal frustration—and yet in some ways they seem like his most human moments. For a few seconds a real person broke through the dull, bovine exterior. He was reacting the way we’d all want to if dealt his fate: he raged.
As soon as a nurse distracted H.M., he forgot his torment, of course. And aside from those flare-ups he lived a quiet life, albeit in declining health. He finally died in 2008, aged eighty-two, of respiratory failure—at which point scientists revealed him to the world as Henry Gustav Molaison.
The world of neuroscience mourned Molaison: his death led to numerous tributes about his patience and kindness, as well as scores of puns about his being unforgettable. And his brain is still providing insight today. Before his death, his nursing home had started stockpiling ice packs in preparation; when he passed, employees ringed his skull with them to keep his brain cool. Doctors soon arrived to claim the body, and that night they scanned his brain in situ and then liberated it. After two months hardening in formalin, it was flown cross-country in a cooler (which got the window seat) to a brain institute in San Diego. Scientists there soaked it in sugar solutions to draw out excess water, then froze it to solidify it. Finally, they used the medical equivalent of a deli slicer to shave Molaison’s brain into 2,401 slices, each of which they mounted on a glass plate and photographed at 20x magnification, to form a digital, zoomable map down to the level of individual neurons. The slicing process was broadcast live online, and 400,000 people tuned in to say goodbye to H.M.
The brain of H.M., the unforgettable amnesiac, being sliced in preparation for future study. (Courtesy Jacopo Annese, the Brain Observatory, San Diego)
Although H.M. dominated the scientific literature and popular imagination, plenty of other amnesiacs have contributed to our understanding of memory. Take K.C., an amnesiac in suburban Toronto. During a wild and extended adolescence, K.C. jammed in rock bands, partied at Mardi Gras, played cards till all hours, and got into fights in bars; he was also knocked unconscious twice, once in a dune buggy accident, once when a bale of hay konked him. Finally, in October 1981, at age thirty, he skidded off an exit ramp on his motorcycle. He spent a month in intensive care and lost, among other structures, both hippocampi.
After the accident a neuroscientist named Endel Tulving determined that K.C. could remember certain things just fine. But everything he remembered fell within one restricted category: it was all stuff you could look up in reference books, like the difference between stalactites and stalagmites or between spares and strikes in bowling. Tulving called these bare facts “semantic memories,” memories devoid of all context and emotion.
At the same time K.C. had zero “episodic memory”—no memories of things he’d personally done or felt or seen. For instance, in 1979 K.C. surprised his family the night before
his brother’s wedding by getting a perm. To this day he knows his brother got married and can recognize family members in the wedding album (the facts), but he doesn’t remember being at the wedding and has no idea how his family reacted to his curly hair (the personal experiences). The little that K.C. did retain about his preaccident life sounds like something he looked up in a particularly dry biography of himself. Even pivotal moments have been reduced to bulleted points in an index. He knows his family had to abandon his childhood home because a train derailed and spilled toxic chemicals nearby; he knows a beloved brother died two years before his own accident. But these events have no emotional import anymore. They’re just stuff that happened.
These details, along with scans of K.C.’s brain, provided strong evidence that our episodic and semantic memories rely on different brain circuits. The hippocampus helps record both types of memories initially, and it helps retain them for the medium term. The hippocampus probably also helps us access old personal memories in long-term storage. But to access old semantic memories, the brain seems to use the parahippocampus, an extension of the hippocampus on the brain’s southernmost surface. K.C., whose parahippocampi survived, could therefore remember to sink the eight ball last in pool (semantic knowledge), even though every last memory of playing pool with his buddies had disappeared (personal knowledge).*
What’s more, while a healthy hippocampus will usually take responsibility for recording new semantic memories, the parahippocampus can—albeit excruciatingly slowly—absorb new facts if it has to. For instance, after years of shelving books as a volunteer at a local library, K.C.’s parahippocampus learned the Dewey decimal system, even though he had no idea why he knew it. Similarly, H.M.’s healthy parahippocampus picked up a few choice facts after his 1953 surgery. After seeing the crossword clue a thousand times he dimly recalled that “Salk vaccine target” equaled P-O-L-I-O. And through incessant references, he retained a sliver of information about the 1969 moon landing and 1963 Kennedy assassination. Contra the cliché, he couldn’t recall where he was when he learned those things—that’s episodic memory. And his knowledge of the events remained weak and fragmentary, since the parahippocampus cannot learn very well. He nevertheless absorbed that they’d happened.
Along these same lines, K.C. helped neuroscience come to grips with another important distinction in memory research, between recollection and familiarity. Colloquially, recollection means I specifically remember this, while familiarity means this sounds familiar, even if the details are fuzzy. And sure enough, the brain makes the same distinction. In one test K.C.’s doctors compiled a list of words (El Niño, posse) that entered the common parlance after his accident in 1981. They then sprinkled those words into a list of pseudo-words—strings of letters that looked like plausible words but that meant nothing. Time and again K.C. picked out the real word, and did so with confidence. But when asked to define the word, he shrugged. From a list of common names he picked out the people who’d become famous after 1981 (e.g., Bill Clinton). But he had no inkling what Clinton had done. In other words, K.C. found these terms familiar, even though specific recollection eluded him. This indicates that recollection once again requires the hippocampus, while a feeling of familiarity requires only certain patches of cortex.
A final type of memory that amnesiacs have helped illuminate is emotional memory—which makes sense, given that the hippocampus belongs to the limbic system. Possibly because he had no amygdalae, H.M. was always pretty affable around the scientists who visited him, despite never recognizing them. (Not even Milner, who worked with him for a half century.) Other amnesiacs lacked his easygoing manner, though, and a few got outright snarly. In 1992 herpes simplex—the same bug that knocked out people’s ability to recognize fruits, animals, and tools—hollowed out the hippocampi and other structures inside the brain of a seventy-year-old San Diego man named E.P. He started repeating the same anecdotes over and over, verbatim, and eating up to three breakfasts each day. And despite being a former sailor who lived less than two miles from the coast, he suddenly couldn’t remember even the general direction of the Pacific Ocean.
Doctors arranged to test E.P., but he grew suspicious of the “strangers”—really the same woman each time—invading his home. Every visit, he dug in his heels, and every visit, his wife had to talk him into playing nice and drag him to the kitchen table to start testing. Eventually, though, after more than a hundred visits, E.P. let his guard down. He started greeting the tester warmly, despite maintaining that he’d never seen her; he even started moving toward the kitchen table on his own to start testing. Somehow, even though his mind was telling him otherwise, his emotions remembered to trust his tester. Amnesiacs can retain negative emotional memories, too. When H.M. learned that his father had died, his conscious brain of course forgot that fact within minutes. But his emotional brain remembered, and took the news so hard that he plunged into a months-long funk, even though he couldn’t explain why he felt so low. In another example, from around 1911, a Swiss doctor named Édouard Claparède concealed a pin between his fingers before greeting a middle-aged amnesic woman; when they shook hands, he pricked her. Although she remembered nothing of this, she always withdrew her hand, and eyed him, on subsequent meetings.
Taken as a whole, this alphabetic soup of amnesiacs (q.v., e.g., H.M., K.C., E.P.) helped scientists sort out how the brain divides up responsibility* for memories. Nondeclarative memories (like motor memories) rely on the cerebellum and on certain internal clusters of gray matter such as the striatum. Episodic (personal) memories lean heavily on the hippocampus, while semantic (factual) memories utilize the parahippocampus to a much larger degree, especially for retrieval. The frontal lobes contribute as well, both in searching for memories and in double-checking that the brain has grabbed the right memory from long-term storage in the cortex. Sensory and limbic circuits also kick on to reanimate the moment in our minds. Meanwhile, the parietal and frontal lobes whisper to us that we’re reviewing old information, so we don’t get terrified or amorous all over again. Each step works independently, and each one can malfunction without affecting other mental faculties in the slightest.
That’s the theory, at least. In reality it seems impossible to tear out any one aspect of memory—especially our episodic memories, memories of holidays and lovers and times we fell short—without tearing out so much more. K.C. knows how to play solitaire and change a tire, but he can never recall a moment of contentment, peace, loneliness, or lust. And however paradoxical it might seem, losing his past wiped out his future as well. The ultimate biological purpose of memory isn’t to recall the past per se, but to prepare for the future by giving us clues about how to act in certain situations. As a result, when K.C. lost his past self, his future self died along with it. He cannot tell you what he’ll do over the next hour, the next day, the next year; he cannot even imagine these things. This loss of his future self doesn’t pain K.C.; he doesn’t suffer or rue his fate. But in some ways that lack of suffering seems sad in and of itself. However unfair, it’s hard not to see him as reduced, diminished.
In our minds, we more or less equate our identities with our memories; our very selves seem the sum total of all we’ve done and felt and seen. That’s why we cling to our memories so hard, even to our detriment, and that’s why diseases like Alzheimer’s, which rob us of memories, seem so cruel. Indeed, most of us wish that we could cling to our memories more securely—they seem the only bulwark against the erosion of the self that K.C. and H.M. experienced. That’s why it’s such a shock to realize that the opposite burden—a hoarding, avaricious memory that cannot forget—can crush people’s identities in the selfsame way.
Each morning when Moscow reporter Solomon Shereshevsky got to work, his editor assigned him and the other reporters their daily stories, telling them where to go, what to look for, and whom to interview. Despite the intricacy of the instructions, Shereshevsky never took notes, and according to some accounts he never took notes during interv
iews, either. He just remembered. Still, Shereshevsky wasn’t a great reporter, and at one morning meeting in the mid-1920s his editor’s fuse went off when he saw Shereshevsky blithely nodding at him, no pencil in hand. He called Shereshevsky out, challenging him to repeat his instructions. Shereshevsky did, verbatim—and then repeated every other word the editor had said that morning, too. When his fellow reporters stared, Shereshevsky’s brow knit in confusion. Didn’t everyone have complete recall? Half amazed, half creeped out, the editor sent Shereshevsky to a local neuroscientist, Aleksandr Luria.
Although a young man then, Luria had already started down the path that would make him one of the most celebrated neuroscientists of the twentieth century. He championed the romantic side of neuroscience, neuroscience that encompassed more than just cells and circuits. He wanted to capture how people actually experienced life, even the messy bits. In doing so, he swam against the current of modern science, which tends to dismiss anecdotal accounts (the plural of anecdote, after all…). But individual case studies have always been crucial to neuroscience: as with the best fiction, it’s the particulars of people’s lives that unveil the universal truths. Indeed, Luria’s book-length case reports have been called “neurological novels,” and he wrote one of his finest on Shereshevsky.