Years later, inspired in part by Henry’s story, neurobiologist Eric R. Kandel took up the subject of the cellular neurobiology of short- and long-term memory. In the late 1960s, Kandel and his colleagues began to study an invertebrate with a simple nervous system, Aplysia (sea snail), to see how it transformed short-term memories into long-term memories. The researchers focused on two simple forms of implicit learning: habituation, the process whereby organisms cease to respond to conspicuous but unimportant stimuli after many exposures to them, and sensitization, the process whereby experiencing a powerful stimulus leads to an increased reaction to a later stimulus that otherwise would have elicited a weaker response. In our daily lives, these unconscious mechanisms operate in the background to protect us and keep us focused. In habituation, we learn to tune out music blaring in the apartment next door, and in sensitization, being bitten by a neighbor’s dog makes us fearful and tense the next time we hear a dog bark.
To study these simple forms of learning, Kandel and his collaborators focused on the snail’s gill-withdrawal reflex, which protects its breathing apparatus. The gill is usually relaxed, but when something touches its siphon, the tube that ejects liquid from the snail’s body, the siphon and gill withdraw into an opening. Kandel and his colleagues trained this simple response to demonstrate habituation and sensitization. In one experiment, they repeatedly delivered a mild touch to the siphon; the snail eventually became habituated to this touch, and its gill-withdrawal reflex weakened. In another experiment, the researchers gave the same mild touch to the siphon, but this time they simultaneously shocked the snail’s tail. In this case, the snail became sensitized and produced a strong gill-withdrawal reflex, even when the weak touch was unaccompanied by the shock. Habituation and sensitization lasted anywhere from one day to several weeks, depending on the training protocol.
Because of the simplicity of Aplysia’s central nervous system, Kandel and colleagues could anatomically map the neural circuitry of the gill-withdrawal reflex and identify the synaptic connections among the cells in this circuit. They then inserted electrodes and recorded the activity of individual sensory and motor neurons. These experiments were possible because in Aplysia the cells are relatively large—up to one millimeter in diameter at the cell body. The electrophysiological recordings made it possible to identify the sensory neurons that were activated by touching the siphon and the motor neurons that initiated the reflex. With these methods, Kandel showed that learning was related to an increase in the electrical strength of the connections at synapses, with the result that one cell could communicate more effectively with its targets. This important study was one of the first to draw attention to the signaling properties of neurons, the cellular and molecular biology of how the connections between neurons are affected by learning.4
Crucially, in the same series of experiments, Kandel and his colleagues demonstrated that the mechanisms underlying short-term memory and long-term memory are different. Short-term memory, they learned, is associated with changes in synaptic function, as opposed to structure. During the course of learning, existing connections may become stronger or weaker, without any ostensible change in structure. By contrast, long-term memory requires physical changes at the synapse. Long-term memory requires protein synthesis; short-term memory does not. Kandel’s experiments supported and amplified Hebb’s insight that two distinct memory processes exist side by side.
Hebb had put the concept of synaptic plasticity on the table by proposing that the repeated stimulation of a group of neurons during learning gradually strengthens their connections, and in doing so establishes lasting memories. Twenty years later, Kandel provided strong support for Hebb’s theory when he linked the recurrent activity of single neurons to particular kinds of learning in Aplysia. His discovery that different proteins are necessary for short- and long-term memory represented an important initial step toward uncovering the molecular basis of memory. Following in Hebb’s and Kandel’s footsteps, neuroscience researchers today focus on identifying the proteins and genes that will tell us how cells talk to one another and support learning.5
Discovering the molecular machinery that supports short- and long-term memory was crucial to understanding the roots of Henry’s amnesia. Similarly, examining his amnesia in behavioral studies became an opportunity for researchers to investigate how these two different types of memory are organized in the human brain. If the single-process theory were correct, then Henry’s short-term memory should have been compromised. As it was, his short-term memory remained intact, while his long-term memory disappeared, suggesting that not only were they separate processes, but they also depended on different areas of the brain.
Having studied under Hebb, Milner was influenced by his dual-process theory of memory. She saw that Henry could provide experimental evidence to address the single-process versus dual-process debate. During Henry’s 1962 visit to Milner’s lab, her graduate student Lilli Prisko gathered data on his short-term memory. She asked Henry to compare two simple, nonverbal stimuli that were separated by a brief delay; the challenge for him was to hold the first item in his memory long enough to say whether it was the “Same” or “Different” from the second item. Prisko chose several different kinds of test items, enabling her to collect data from Henry in multiple experiments. Drawing conclusions based on a single experiment or task is risky, so Prisko avoided that problem by assessing Henry’s short-term memory with complementary tasks. Some used sounds such as clicks and tones, and others used visual images such as light flashes, colors, and non-geometric nonsense figures. She also intentionally chose stimuli that were difficult to verbalize. In testing memory for colors, for instance, Prisko could not have used patches of red, orange, yellow, green, blue, and violet as test stimuli because Henry could have repeated the color names to himself during the delay intervals and outsmarted the test. Instead, she chose five different shades of red to minimize the possibility of verbal rehearsal.6
Henry lay on a stretcher in a quiet, dark area, separated by a screen from where Prisko sat in the main laboratory. They were the only ones present. She called out “First one” to indicate that a trial was starting. The stimulus on this trial was a series of flashes from a strobe light at a rate of three per second. After a brief delay, another set of flashes appeared, moving faster now at about eight per second. Henry had to say “Different” to indicate that there had been a change between the two stimuli. On other trials, the flashes appeared at the same rate, and he had to say “Same.”
Testing Henry in this experiment was challenging, particularly at the beginning. Sometimes he talked instead of sitting quietly, needed prompting to give his responses, or responded after the first stimulus rather than waiting for the second. Every few minutes, Prisko repeated the instructions so that Henry would know what he was supposed to do. She also had to repeat several botched trials to complete the experiment.
The results of the first experiment provided an important basis for understanding Henry’s capacity for perceiving and retaining information. He could easily and accurately perform the task when there was no delay between stimuli, making only one error in twelve trials. He clearly had no problem with the instructions or perceiving the test stimuli: he was perfectly capable of appreciating the difference between them when they were close together in time. With that knowledge, Prisko could assume that any problems Henry encountered during trials with longer delay intervals were a direct result of an inability to remember.
Prisko next tested Henry with the same sets of flashes, this time separated by fifteen, thirty, or sixty seconds. The ability to differentiate between stimuli becomes harder for everyone as the gap between them gets longer and as short-term memory weakens. In Henry’s case, however, the differential was extreme. With a fifteen-second interval, Henry still performed well, making two errors out of twelve trials; when the delay was thirty seconds, his errors increased to four. At sixty seconds, his answers were wrong on six of the twelve trials, which is conside
red chance performance, no better than if Henry had flipped a coin each time to choose his answer. In comparison, normal control subjects made an average of one error out of twelve during the sixty-second delay between stimuli, even if Prisko distracted them.
The abrupt breakdown in Henry’s performance showed that his short-term memory lasted less than sixty seconds. Somewhere between thirty and sixty seconds, his memory of what he had seen or heard crept away. With shorter delays, he performed better than chance: his mind held onto test stimuli, as long as they occurred within this narrow sliver of time. Henry’s results were consistent with those of F.C. and P.B., both of whom Prisko tested soon after. They had fewer errors than Henry, but showed the same pattern of increasing errors as the time between the presentation of the two stimuli was extended, causing the memory of the first to fade away.
To Prisko’s surprise, Henry showed some ability to retain certain kinds of information. After the test with flashes, she let him rest for a few minutes and then began the next task, this time using audible clicks. By now, Henry seemed to have improved as a test subject; although he still talked, he no longer called out after the first stimulus. She gave him a one-hour break before the next test, which was with colors. When he returned, Henry had completely forgotten who she was. Yet after being given his instructions, he seemed to understand the testing scheme better, talked less, and followed the instructions correctly. When she tested him again the next day, it was enough to give Henry the instructions once at the beginning of each new test. His scores were just as bad, and he had no memory of having done the test before, but somehow Henry knew what was expected of him.
How was it that Henry was able to learn the correct procedures—the “how to do it”—but remained unable to retain for more than a few seconds what the specific test stimuli had been? In 1962, no one could explain this strange distinction, but it gave all of us in Milner’s lab a sense that Henry had a great deal to teach us about the nature of memory and learning.
Prisko’s results from testing dealt a blow to the theory that memory was a single process. Around the same time, another case was emerging that posed a similar challenge. A patient in England, known by his initials K.F., sustained massive damage to the left side of his head and brain in a motorcycle accident. He was unconscious for ten weeks, and although he showed gradual improvement over the next few years, he began to have seizures. Like Henry, K.F. had a massive memory deficit, but of exactly the opposite pattern; remarkably, he could form new long-term memories despite his lack of short-term retention. He had a digit span of just two, and could repeat only one number, letter, or word reliably. If an examiner spoke pairs of words at the rate of one word per second, he could repeat both words correctly only half of the time. His short-term store had a very limited capacity. Nevertheless, K.F. performed normally on four different long-term memory tests, indicating that his long-term store for this information was intact.7
Taken together, the findings for Henry and K.F. indicate the existence of two independent memory circuits serving short-term memory and long-term memory, respectively, strongly supporting the dual-process theory. The two circuits have different anatomical locations: cortical processes mediate short-term memory, and medial temporal-lobe processes modulate long-term memory.8
K.F.’s case suggested that short-term memory processes are based in the cerebral cortex, the outer layers of the brain. Scoville did not touch this part of Henry’s brain, so all of Henry’s cortical functions were preserved, available for holding information online for brief periods of time, thus sparing his short-term memory. Further research has shown that short-term memories are scattered in different parts of the cortex, depending on the type of information they represent. Increasing evidence indicates that different areas of the brain are responsible for temporarily retaining memories of faces, bodies, places, words, and so on. These memories are not distributed randomly; instead, they tend to cluster near areas related to how the information was first perceived. The right parietal lobe, for example, is devoted to spatial abilities, so short-term memories related to spatial knowledge are maintained in that area. Similarly, the left side of the brain controls language, and short-term verbal memories are rooted predominantly in the left side of the cortex. With a better understanding of short-term memory as a separate process, we can explore in depth what happens in this brief period of time. We now know much more about the content, capabilities, and limits of short-term memory.9
Information lingers in short-term stores for less than a minute, but we can maintain information indefinitely by rehearsing it in our thoughts. Rehearsal effectively refreshes short-term traces, making them new again. This is a good example of a control process. Control processes underlie our ability to manage thoughts in the interest of achieving a goal. We engage these processes constantly in everyday life, helping us focus on the task at hand, switch from one task to another, and block out unwanted intrusions.10
Imagine a businessman on a flight from Boston to San Francisco, connecting with a flight to Honolulu. After the plane lands in San Francisco, the flight attendant announces the gate numbers for flights to connecting cities. The man listens attentively as each city is called, and once he hears “Honolulu” and the gate number, he starts rehearsing it, repeating it over and over as he leaves the plane and successfully navigates through the crowds to the designated gate, willfully ignoring distractions along the way so that he does not forget where he is going. His gate number will likely disappear from his mind once he has boarded his flight; he kept it alive in his short-term memory just long enough for it to serve its purpose. We regularly juggle complex processes like these to help us remember and guide our behavior to achieve our goals.
Because Henry could rely only on short-term memory, he harnessed cognitive control processes to compensate for his memory deficit. By mentally rehearsing information he was asked to remember, he could sometimes keep thoughts fresh in his mind until asked to retrieve them. Milner noticed this ability during her first testing session with Henry in 1955 in Scoville’s office. She gave Henry these instructions: “I want you to remember the numbers five, eight, four.” She then left the office and had a cup of coffee with Scoville’s secretary. Twenty minutes later, she returned and asked Henry, “What were the numbers?”
“Five, eight, four,” he replied. Milner was impressed; it seemed Henry’s memory was better than she realized.
“Oh, that’s very good!” she said. “How did you do that?”
“Well, five, eight, and four add up to seventeen,” Henry answered. “Divide by two, you have nine and eight. Remember eight. Then five—you’re left with five and four—five, eight, four. It’s simple.”
“Well, that’s very good. And do you remember my name?”
“No, I’m sorry. My trouble is my memory.”
“I’m Dr. Milner, and I come from Montreal.”
“Oh, Montreal, Canada,” Henry said. “I was in Canada once—I went to Toronto.”
“Oh. Do you still remember the number?”
“Number? Was there a number?”
The complex calculations Henry had devised to keep the number in his head were gone. As soon as his attention was diverted to another topic, the content was lost. His forgetting that he had been rehearsing a number is unusual, but even people with intact brains lose information when they are distracted. Consider the airport example: if, while making his way through the airport, the businessman were distracted by a breaking news story on a television monitor, he would likely forget the gate number he had been working to retain in his short-term memory. If he did remember the gate number after being distracted, it would be because he drew on the resources of his long-term memory—a capacity that Henry lacked.11
Henry relied on his short-term memory in collaboration with his control processes. In conversation, he seemed normal because he could easily answer a question that was just asked of him. In this way, he could maintain a seemingly easy back-and-forth, staying sure-foote
d as long as nothing distracted his attention. He could keep names, words, or numbers in mind for a few seconds and could repeat this information, but only when the memory load was small and nothing else intervened to wipe the slate clean. If I were talking with Henry and another person started a different conversation with him, not only would he forget what I had just told him, he would also not remember that I had told him anything at all.
After I started my own lab at MIT in 1977, we had the opportunity to test the effect of distraction on Henry and four other patients with amnesia from other causes, all of whom had significant defects in their long-term memory and had only their short-term memory to rely on. We gave them the Brown-Peterson distractor task, which tests how quickly subjects forget information they have just absorbed. They look through a viewing window and see pairs of consonants followed by pairs of digits, with each pair visible for about three quarters of a second. The subjects might see, for example, VG and then SZ, followed by 83 and 27. They read the consonants and digits, but are asked to remember only the consonants. Because the subjects are busy reading the digits, they cannot rehearse the consonants. By preventing rehearsal, the task measures how much information people forget and how quickly they forget it during a period of about five seconds, before being asked to recall the letters. When psychologist John Brown first introduced the distrator task in 1958, he found that healthy people could recall only one pair of consonants accurately. Subjects forgot the other information—the second pair—in less than five seconds when they could not rehearse it.12
In 1959, psychologists Margaret and Lloyd Peterson built on Brown’s experiment by studying how accuracy changes when delay times are manipulated. In their version of Brown’s test, the examiner said three consonants, such as MXC, and then a three-digit number, such as 973. Participants had to count backward from the number by threes—973, 970, 967, 964—until given a signal to repeat the three consonants, MXC. They received the signal after different periods of time—three, six, nine, twelve, fifteen, and eighteen seconds. Peterson and Peterson found that the more time the participants spent on the distracting activity of counting backward, the fewer consonants they recalled. After the fifteen-and eighteen-second delays, they could remember almost nothing. This study showed that short-term memory lasts less than fifteen seconds under the influence of distraction.13
Permanent Present Tense Page 8