by Sam Kean
Korsakoff focused on the psychology of confabulation, but other scientists extended his work in the early 1900s and started linking these psychological symptoms to specific brain damage. In particular, they discovered tiny hemorrhages in the brains of victims, as well as patches of dead neurons. Pathologists also linked Korsakoff’s syndrome to another, related disease called Wernicke’s syndrome. In fact, because Wernicke’s syndrome often turns into Korsakoff’s, the two were eventually yoked together as Wernicke-Korsakoff syndrome.
The underlying cause of Wernicke-Korsakoff syndrome took longer to suss out, but by the later 1930s a few scientists had linked it to a lack of B1. As doctors now know, alcohol prevents the intestines from absorbing the thiamine in food. This shortage then causes changes inside the brain, especially to glial cells. Among other jobs, glial cells sponge up excess neurotransmitters from the synapses between neurons. And without thiamine, the glia cannot sop up glutamate, which stimulates neurons. As a result of this excess, neurons get overstimulated and eventually exhaust themselves, dying of excitotoxicity.
Because they seemed to share a common root—B1 deficiency—beriberi and Wernicke-Korsakoff syndrome should have caused similar symptoms and similar destruction inside the brain. But through the 1940s no one had any hard evidence to link them. This was partly because Wernicke-Korsakoff remained rare and associated primarily with alcoholics, and partly because doctors who studied beriberi focused on nerve and heart damage, not brain damage. The net result was confusion: was this one disease or two? More important, it highlighted a growing concern over efforts to link physiology and psychology: many doctors frankly doubted that the lack of a simple vitamin—a molecular problem—could leap up so many levels of scale and cause complex mental troubles like confabulation.
Changi proved it could. Among the thousand-plus beriberi victims there, several dozen also came down with symptoms of Wernicke-Korsakoff syndrome, including confabulation. As an example, de Wardener asked one far-gone man, just to test his mental state, “Do you remember when we met in Brighton? I was riding a white horse and you a black horse, and we rode on the beach.” This was bunk, but the man answered that of course he remembered, and filled in the details. Such imaginings often became the patients’ reality, sadly, and a few men died in this state—their last “memories” nothing but vapors and fabrications. Medically, the fact that beriberi always preceded Wernicke-Korsakoff, and that those with the worst beriberi got the worst Wernicke-Korsakoff, implied a common cause. Autopsies then cemented the link: even without a microscope, Lennox, a trained pathologist, could see the characteristic hemorrhages and patches of dead neurons in the brains of victims. Beriberi and Wernicke-Korsakoff seemed to be two stages, chronic and acute, of the same underlying disease.
As further evidence, treating victims with pure thiamine (some doctors had tiny stashes) usually relieved the symptoms of both Wernicke-Korsakoff and beriberi, sometimes within hours: de Wardener remembers a few men roaring to life and consuming whole mountains of rice to combat their sudden hunger. (Mental symptoms such as confabulation might take several weeks to dissipate.) For less acute cases doctors might add Marmite to meals (however unappetizing, this yeast-based extract is lousy with B1) or ferment rice and potatoes to cultivate wild yeast, also chock-full of B1. Some doctors sent men to gather thiamine-rich hibiscus leaves as well. The smarter doctors lied to the men and claimed that hibiscus would pump up their libidos for when they got back home to their gals. After that, troops no doubt couldn’t consume enough hibiscus.
In tandem, the fact that consuming too little thiamine provoked Wernicke-Korsakoff, and that restoring thiamine to the diet relieved it, convinced Lennox and de Wardener that the lack of a simple nutrient could indeed destroy something as profound as our memories, even our sense of truth. But the duo still had to make their case to the medical world—which meant not only surviving the camps but preserving their autopsy files. This wasn’t easy in a war zone, and as de Wardener discovered, about the only way to conceal such things was to bury them, and pray to God they survived.
After V-J Day, de Wardener received mysterious orders to report to Bangkok. Although anxious to start searching for his files, he remembers enjoying the journey: “I took a victorious ride across Siam in a Jeep, with all the Nips bowing… which was very satisfying.” To his surprise, he found his records waiting for him at Bangkok HQ. Apparently a friend had returned to Changi with a shovel not long before, scrabbled through the dirt above the dead sentinel’s body, and liberated the bundle. It was a close thing: the cape had rotted away and the solder sealing up the tin had disintegrated. But the papers had survived, perhaps by a matter of days. Lennox and de Wardener finally published this, well, groundbreaking work in 1947.*
Since World War II, neuroscientists have continued to mine confabulation for insight into how memory works, and it has proved a rich vein indeed. For example, confabulations reveal that each memory seems to have a distinct time stamp, like a computer file. And just like computer files, that time stamp can be corrupted. Most confabulators tell plausible lies; in fact, many of their false “memories” did happen to them at some point. But confabulators often mistake when the memory happened: the scenes in their lives have been shuffled wrong. So while they claim they ate truffled duck last night, in truth they did that thirty years ago while honeymooning in Paris. In some sense, then, confabulation is a breakdown in the ability to tell a coherent story about our lives.
The fact that virtually all confabulators have frontal lobe damage also tells us something. The frontal lobes help coordinate multistep processes, and despite how effortless memory seems, remembering something specific (say, the worst Christmas present you ever got) is complicated. The brain has a fraction of a second to search for the memory, retrieve it, replay it, and summon up the proper sensations and emotions—and that’s assuming you recorded the memory accurately in the first place. If the frontal lobes suffer damage, any one of those steps can go awry. Perhaps confabulators simply retrieve the wrong memory each time they “recall” something, and don’t recognize their error.
Some scientists trace confabulations to shame and a need to cover up deficiencies. Confabulators don’t generally blurt things out unprovoked; you have to ask questions to elicit the lie. And according to this theory, admitting they don’t know something upsets and embarrasses people, so they pretend. For example, most doctors ask at intake how many children someone has. Having to admit “I don’t know” could be catastrophic to a person’s well-being, since what kind of monster doesn’t remember his own children? In short, confabulations could be a defense mechanism, a way for people to hide their brain damage, even from themselves.
As another defense mechanism, some confabulators invent fictional characters and foist their personal failings onto them. One alcoholic confabulator raved to his doctor about imps who kept breaking into his apartment, even after he changed the locks, and stealing things like his remote control. He eventually heaved the imps outdoors on a brutal January night. But, feeling guilty, he braved the weather and draped clothes over them later that night, then called an ambulance. In reality medical workers had discovered him outside that winter, stone drunk and mostly naked. In telling the story he was basically confabulating an allegory on the fly. That’s a remarkable deed for someone with brain damage, and the ruse allowed him to ponder his own flaws more objectively, without implicating himself.
As that last case shows, it’s not always clear whether confabulators understand that they’re lying. Most seem blithely unaware, and many neuroscientists insist that Korsakoff patients don’t realize what’s happening. But is that possible? Covering up a memory gap, even subconsciously, implies that they know on some level that the gap exists. Which means they know and don’t know at the same time. It’s a doozy of a conundrum, and it raises all sorts of stoner questions about whether you can truly deceive yourself, and more broadly about the nature of truth and falsehood. Consider asking a confabulator what she ate f
or breakfast. If she hasn’t the foggiest, she might blurt, “Leftover pizza.” But of course it’s possible she did have cold pizza for breakfast, in which case she would be telling the truth—even though her brain tried, consciously or not, to put one over on you. What on earth would you call that? Neither lying nor telling the truth quite encompasses it. It’s slipperier, and some neuroscientists have taken to calling it “honest lying.”
Philosophical conundrums aside, work on confabulation helped make memory a proper object of neuroscientific study last century, since scientists could finally link memory to the brain and its biology. That said, the biggest breakthrough in memory research in the past hundred years didn’t spring from the minds of confabulators. Indeed, most memory work until the 1950s relied on a flawed assumption—that all parts of the brain contribute equally to forming and storing memories. It’s an idea that took something drastic, a botched operation by a lobotomist, to overturn.
In the early 1930s a bicyclist in Connecticut struck a small boy, who tumbled and cracked his skull. No one knows whether the accident alone caused his epilepsy—three cousins had it, so he might have been predisposed—but the blow probably precipitated it, and at age ten he started having seizures. Each lasted around forty seconds, during which time his mouth flopped open, his eyes slipped shut, and his arms and legs crossed and uncrossed as if curled by an invisible puppeteer. He suffered his first grand mal on, of all days, his fifteenth birthday, while riding in the car with his parents. More followed, in class and at home and while shopping—up to ten seizures a day, with at least one major episode per week. So at an age when most people are struggling to find an identity, he was saddled with one he didn’t want: the kid who shook, who bit his tongue, who slumped over and blacked out and pissed himself. The mockery got so bad he dropped out of high school, and he earned his diploma only at twenty-one, from a different school. He ended up living at home and working in a motor shop.
Finally the desperate young man—soon immortalized as H.M.—decided to try surgery. When younger, H.M. had dreamed of practicing neurosurgery himself and studying how the brain works. But while H.M. did end up contributing, profoundly, to neuroscience, his affliction ensured that he would never grasp his own importance.
H.M. started seeing Dr. William Scoville around 1943. A noted daredevil—before a medical conference in Spain once, he’d stripped off his jacket and mixed it up with the toros in the bullring—Scoville liked risky surgeries, too, and had jumped onto the American lobotomy bandwagon* early. But he disliked the drastic changes in his patients’ personalities, so he began experimenting with “fractional” lobotomies, which destroyed less tissue. Over the years he basically worked his way around the brain, carving out this piece or that and checking the results, until he finally reached the hippocampus.
Because it was part of the limbic system, scientists at the time believed that the hippocampus helped process emotions, but its exact function remained unknown. Rabies often destroyed it, and James Papez had singled it out for attention. (A poetaster, Papez even penned a ditty to his wife that read: “It’s Pearl, my girl on Broad Street / that I miss… My hippocampus tells me this.”) Scoville was less enamored: he’d seen the mental turmoil that hippocampus damage could cause. So in the early 1950s he started removing the hippocampi (you have one in each hemisphere) from a few psychotics. Although it was hard to be sure in people with such disturbed minds, they seemed to suffer no ill effects, and two women in particular showed a marked reduction in seizures. Unfortunately Scoville neglected to do careful follow-up tests until November 1953—after he’d convinced H.M. to try the surgery.
H.M.’s operation took place in Hartford, Connecticut, on September 1, 1953. Scoville peeled back his patient’s scalp, then used a hand crank and one-dollar drill saw from a local hardware store to remove a bottle cap’s worth of bone from above each eye. As cerebrospinal fluid drained away, the brain settled down in its cavity, giving Scoville more room to work. With what looked like an elongated shoehorn, he nudged aside H.M.’s frontal and temporal lobes and peered inside.
The hippocampus sits at ear level and has the rough shape and diameter of a curled thumb. Hoping to remove as little tissue as possible, Scoville first sparked each hippocampus with wires to find the origin of H.M.’s seizures. No luck, so he grabbed a long metal tube and began cutting and sucking out tissue gram by gram; he eventually removed three inches’ worth of hippocampus on each side. (Two nubs of hippocampal tissue remained behind, but because Scoville also removed the connections between those nubs and other parts of the brain, the nubs were useless, like unplugged computers.) For good measure, Scoville removed H.M.’s amygdalae and other nearby structures as well. Given how deeply all these structures are embedded in the brain, only a neurosurgeon could have destroyed them with such precision.
Post-op, H.M. remained drowsy for a few days, but he could recognize his family and carry on a seemingly normal conversation. And by many measures, the operation succeeded. His personality never changed; the seizures all but disappeared (two attacks per year at most); and when the fog of epilepsy lifted, his IQ jumped from 104 to 117. Just one problem: his memory was shot. Aside from a few small islands of recollection—like the fact that Dr. Scoville had operated on him—an entire decade’s worth of memories from before the surgery had vanished. Equally terrible, he couldn’t form new memories. Names escaped him now, as did the day of the week. He repeated the same comments over and over, verbatim, and while he might remember directions to the bathroom long enough to get there, he always had to ask again later. He’d even consume multiple lunches or breakfasts if no one stopped him, as if his appetite had no memory, either. His mind had become a sieve.
In light of modern knowledge, H.M.’s deficit makes sense. Memory formation involves several steps. First, neurons in the cortex jot down what our sensory neurons see and feel and hear. This ability to record first impressions still worked in H.M. But like messages scrawled on the beach, these impressions erode quickly. It’s the next step, involving neurons in the hippocampus, that makes memories last. These neurons produce special proteins that encourage axon bulbs to swell in size. As a result, the axons can stream more neurotransmitter bubbles toward their neighbors. This in turn strengthens the synapse connections between those neurons before the memory decays. Over months and years—provided the first impression was strong enough, or we think about the event from time to time—the hippocampus then transfers the memory to the cortex for permanent storage. In short, the hippocampus orchestrates both the recording and the storage of memories, and without it, this “memory consolidation” cannot occur.
Scoville couldn’t have known all this, but he’d clearly sabotaged H.M.’s memory, and he didn’t know what to do. So a few months later, when he saw that Wilder Penfield was about to publish a report on hippocampus damage, Scoville called the renowned surgeon and confessed.
Penfield had recently operated on two patients with hippocampal epilepsy. To be safe, he’d removed the structure on just one side, but unbeknownst to him, the seizures had already destroyed the other hippocampus in each person. So removing the one left both patients without a working hippocampus, and they developed the purest amnesia Penfield had ever seen. Although he was still puzzling through the cases, a graduate student was going to present them at a scientific meeting in Chicago in 1954.
When Scoville called, Penfield reportedly flipped out, berating him for his recklessness. After calming down, though, the scientist in Penfield realized (much as the beriberi doctors had) that Scoville had actually performed an invaluable experiment: here was a chance to determine what the hippocampus did. As part of its mission Penfield’s clinic in Montreal tracked the psychological changes that patients experienced after psychosurgery. So Penfield dispatched a Ph.D. student from the Neuro, Brenda Milner, down to Connecticut to investigate the hippocampusless H.M.
After his memory vanished, H.M. lost his job and had no choice but to keep living with his parents. He spok
e in a monotone now and had no interest in sex, but otherwise seemed normal. To the neighbors, it probably just looked like he was loafing his life away. He took a part-time job packing rubber balloons into plastic bags, and did odd chores around the house. (Although his parents had to remind him where they kept the lawn mower every single time, he could actually mow just fine, since he could see what grass he hadn’t cut.) His temper did flare up occasionally: his mother tended to nag, and he cuffed her a few times and kicked her shins. Another time, when an uncle removed a few choice rifles from the family’s gun collection, he flew into a rage. (Despite his amnesia he retained a lifelong love of guns, and always remembered to renew his NRA membership.) But he whiled away most days peacefully, either doing crossword puzzles—working through the clues methodically, in order—or flopping in front of the television and watching either Sunday Mass or the old movies that, to him, would never become classics. It was like early retirement, except for the days Milner arrived to test him.
Milner would take the night train down from Montreal to Hartford, arriving at 3 a.m. and spending the next few days with H.M. Her battery of tests confirmed Scoville’s basic observations pretty quickly: H.M. had little memory of the past and no ability to form new memories going forward. This was already a big advance—proof that some parts of the brain, namely the hippocampus, contribute more to forming and storing memories than other parts. And what Milner discovered next redefined what “memory” even meant.
Rather than keep asking him questions he couldn’t answer, she started testing H.M.’s motor skills. Most important, she gave him a piece of paper with two five-pointed stars on it, one nested inside the other: . The outer star was about six inches wide, and there was a half-inch or so gap between them. The test required H.M. to trace a third star between the two with a pencil. The catch was, he couldn’t see the stars directly: Milner had shielded the diagram, and he had to look at them in a mirror instead. Left was right, right was left, and every natural instinct about where to move his pencil was wrong. Anyone taking this mirror test for the first time makes a mess—the pencil line looks like an EKG—and H.M. proved no exception. Somehow, though, H.M. got better. He didn’t remember any of the thirty training sessions Milner ran him through. But his unconscious motor centers did remember, and after three days he could trace the star in the mirror fluently. He even commented near the end, “This is funny… I would have thought it would be rather difficult, but it seems I’ve done pretty well.”
