Duke University Alumni Magazine

MINING THE MEANING OF MEMORIES Cognitive Science by Dennis Meredith Rubin's clues: using smells to unlock the mysteries of remembering Photo: Bruce Feeley Psychologist David Rubin uses sophisticated machines, immobilizing masks, and modest children's rhymes to explore how we store the memories of our lives.

ooking like a horror movie prop, the white mesh mask envelopes the young man's face, clamping his head unmoving in the cavity of the great thumping machine. Periodically, a puff of air enters his nose, smelling of coconut, salami, strawberry . . . or sometimes nothing at all.

"Suntan lotion . . ." he says to the researcher after the coconut smell. ". . . I remember putting on suntan lotion when we went to the beach."

Duke experimental psychologist David Rubin carefully records the young man's reminiscence, as he will for those sparked by some three dozen other odors over two hours. Once the young man is extricated from the machine, Rubin will have more data points to inform his explorations of the nature of memory. Rubin and his colleagues use the medical center's functional Magnetic Resonance Imaging (fMRI) machine to try to eavesdrop on the brain as it undergoes the mysterious process of remembering. Rubin hopes that the fMRI images--which light up where the brain shows increased blood flow that presumably reflect increased activity--will yield “movies” showing how remembering progresses across the brainscape.

But Rubin is not interested merely in the bare neural machinery of recall, as it functions in rote exercises at recalling lists of words. Rather, he is fascinated by the rich process of “autobiographical memory” by which we seem to relive the ghost-memories of our lives, the very fabric of our identity woven over decades of living.

In his explorations, Rubin and his colleagues employ not only the latest technology, but the most venerable tool of human ingenuity. Over the years, Rubin and a cadre of undergraduate and graduate students have conducted a multitude of clever experiments to tease out new insights into the nature of memory. They have tested how people remember images on coins, Academy Award-Winning movies, and even the modest children’s counting rhyme, “Eenie, Meenie, Miney, Mo.”

The very sophisticated analysis of these seemingly simple experiments has revealed important, often startling insights into how our memories work. For example, Rubin has quantified the striking phenomenon of the “memory bump,” in which we remember most vividly those events that happened between the ages of 10 and 30. Such studies go to the heart of many profoundly important human questions, he says.

“For one thing, memory is something that people lose in many diseases, including Alzheimer’s disease, amnesia and head injuries,” he says. “It’s deeply upsetting to them, and the hope is that the understanding we can develop might eventually help alleviate such problems. But more broadly, memory is peoples’ lives.” he says. “It’s what people tell you about themselves. When people sit around and talk, they tell stories from their lives. The natural human way that people present themselves socially is through their memories, and they often base their behavior on what they remember.”

Thus, says Rubin, his studies seek insights that help people understand themselves, as well as advance scientific understanding about cognitive function. “It’s a way of taking hard-nosed, quantitative laboratory research and applying it in an area that people can understand,” he says.

Like his fellow cognitive scientists, Rubin understands that he is chipping away at a massive, profound mystery.

“There’s really no coherent theory of memory,” he says. “Even though researchers have located many memory functions in the brain, that is a far cry from figuring out the memory process itself. It’s a standard error to believe that if you locate something in the brain and name it, you control and understand it. We’ve done that with memory.”

The popular concept that human memory is like computer data storage reveals most dramatically how woefully inadequate the understanding of memory is, says Rubin.

“This theory implies that when a piece of information is stored in long-term memory it’s right, it’s accurate and it never changes by itself. But that’s not the way biological systems work. Some nerve cells may die, for example, changing a memory a little bit. The computer theory also holds that all information is the same, that imagery isn’t distinct from narrative,” he says.

But memories are actually woven from all kinds of different sensory experience, each with a different neural circuit. These kinds of experience include narrative, imagery, rhythm and motor movement, all of which integrate to provide a unified memory.

“As a baby you put things in your mouth and ran your tongue around them and built images of them that became memories. Or, if I put you in a room with a blindfold on, after a while you’d still develop a minds-eye image of the room, as you would if you could see it. So, memory has a multi-modal spatial aspect, too.

“Memory is really what happens when the whole brain works together,” he says. “When you have a vivid memory of the high school prom or when you remember a song or a poem, the memory sort of takes over and moves you from where you are back into a state of recall. A memory has a bigger effect than some sort of a computer access. It really involves the whole body in a reconstructive emotional and sensory experience, not just data retrieval.”

Such rich biological concepts of memory, even though more accurate, can be profoundly disquieting for people who depend on the computer-theory, says Rubin. “Lawyers really don’t want memory to be a reconstruction of subjective experiences by a changing biological system,” he says.

“They want memories to be accurate because the legal system depends on it. But memory of a visual scene, for instance, is a process in which photons hit the eye, something happens in the brain, and months later when somebody asks you what you saw, you reconstruct it using a biological system. We know the photons didn’t just go up little tubes and get stored, or that people don’t have a videotape player in their head.”

The fMRI studies, supported by the Olfactory Research Fund, represent at least a geographical approach to understanding the how memories are evoked, says Rubin. “First, we’re hoping that the onset of the odor stimulus makes olfactory areas become active. Then, for those vivid memories, we'd like to see visual areas become active, because when people have a sense of reliving a memory, it often means an accompanying image.”

Other researchers have found a ten-second lag between a stimulus like an odor and a vivid autobiographical memory. Rubin’s aim is to learn more about the search process that apparently goes on during this lag. “We’re asking subjects how arduous the search process was,” says Rubin. “Did they really feel they were going after something? And where does that go on in the brain?”

However, Rubin and his colleagues recognize that geography doesn’t necessarily reveal mechanism. “If we see the brain lighting up in one place, it could be a center of activity or it could be an inhibitory center, or it could merely be a way-station in processing. And maybe a small undetected bit of activity could represent a critically important processing step. But at least we’re beginning to break up the big black box of memory into small black boxes.”

The psychologist and his students have also tested undergraduates in experiments exploring the quirks and fallibility of normal memory. For example, in a paper entitled “A Schema for Common Cents,” Rubin and then-undergraduate Theda Kontis described how they asked 125 students to draw from memory a penny, nickel, dime and quarter.

The analysis showed that they could recall little of the coins they used every day (Click here to take a coin-memory test yourself.). The finding offered an intriguing insight into the spotty nature of memory, says Rubin. And, it offers a telling lesson for the designers of coins, such as the ill-fated Susan B. Antony dollar, which for recognition relied mainly on the images on the coin, rather than size or color.

“People don’t learn things they see over and over again,” says Rubin. “They learn things enough to deal with the world.”

Similarly, few people could tell where the letters and numbers are on telephone buttons, notes Rubin. Nor can most people describe where the moles or hairs are on the back of their hand, he says, despite the popular expression “I know it like the back of my hand.”

Perhaps Rubin’s most intriguing finding has been the phenomenon of the “memory bump,” in which people remember most vividly those events that happened to them between ages 10 and 30. In numerous studies, Rubin and his students tested older people’s recall of events such as Academy Award winning movies, World Series-winning baseball teams, top news stories, presidential campaigns, or important events from their own lives. The studies revealed that the ages between 10 and 30 provided the richest trove of memories.

“Everybody seemed to know about this phenomenon, but nobody bothered to quantify it,” says Rubin. The explanations for the bump are likely complex and intertwined, he theorizes.

One possible explanation for the bump is that the novelty of experiences during these early adult years leads to deeper memory encoding, he says. Or, a young adult’s self-definition of identity that happens during those years may better crystallize memories. Also, young people may just have sharper mental faculties--perhaps to increase their fitness at choosing a mate--which would contribute to the more vivid memory formation.

But whatever the explanation, the idea of a memory bump could prove clinically useful, says Rubin. “For example, if we knew what caused this phenomenon in healthy people, it would help explain why patients with Alzheimer’s disease experience the kind of memory degradation they do. A standard anecdote you get about Alzheimer’s patients is that they remember the old things, but not the new things,” he says. “And eventually as they get near the end, they jump generations. So, their daughter comes to visit them and they see her as their sister.”

In one set of experiments, Rubin and research associate Matthew Schulkind and a team of undergraduates are playing big band music to groups of volunteer senior citizens to try to understand the bump phenomenon. “We play them old songs and ask them how the songs make them feel, and whether they can complete the words,” says Rubin. “The music may work because it resonates with memory that involves large parts of the brain, including motor movements and emotions.” According to Rubin, some music therapists already use the golden-oldies technique with residents of nursing homes who suffer Alzheimer’s and other dementias, to enliven them and get them to socialize more.

Of all Rubin’s studies, though, the most prodigious has been his exploration of the psychology of epic poems, North Carolina ballads and counting-out rhymes. His integration of the folklore and history of these oral recitations with cognitive science resulted in his award- winning 1996 book, Memory in Oral Traditions, which Contemporary Psychology called “a landmark contribution for both scientists and scholars.”

In the book, Rubin sought to explain how such ballads, and even the seemingly trivial rhymes that children chant to choose have survived almost unchanged for centuries.

In his studies, Rubin found a wealth of insight, even in the simple “Eenie, Meenie, Miney, Mo . . ." and its fellow rhymes.

“No psychologists studying memory had really explored these oral ballads and rhymes,” says Rubin, “Even though reciting them is a remarkable mental skill, involving remembering a lot of knowledge in a way that doesn’t change because of the structure of the ballad.”

Thus, in his ballad studies Rubin explored how the rhythm, words, images and story intertwined to make it possible for balladeers to accurately recall even hours of song verse.

“For example, these oral traditions are high imagery, with many changes in location,” he says. “In ballads, the character goes from location to location, with the ballad never spending more than three verses without changing place. And in ballads, people don’t just sit around and mope. They jump off bridges, they get buried in shallow graves, they cut off people’s heads--it’s all high imagery.”

Rubin also explored how the oral ballads and rhymes were carefully crafted with multiple constraints in form and subject, like an intricate puzzle that fits together in only one way. Such constraints make the recitations smooth progressions of verse that lead a performer almost unavoidably from one element to the next.

Each stanza, Rubin points out, is a rhythmic, musical unit that must contain a complete idea, must follow the rhyme scheme and must avoid words larger than one or two syllables. Rubin has interwoven his ballad and rhyme studies with his research on autobiographical memory, just as our memories are interwoven with our lives.

“I find it fascinating studying these things that people do all the time, doing careful quantitative work to understand them in a scientific sense. And then contributing that understanding to society,” he says.

Rubin believes that the next decade of science will no doubt witness an enormous leap in understanding how genes build brains, and how tiny splashes of brain chemicals play among the labyrinth of brain cells to create memories. And that same decade will also see a forging of even richer partnerships among biologists, psychologists and humanists to apply that knowledge to medical treatment and to our everyday lives.

The result will certainly be memorable.

Why You Don't Forget “Eenie, Meenie . . . Eenie, Meenie, Miney Mo, Catch a tiger by the toe. If he hollers let him go. Eenie Meenie Miney Moe This modest little rhyme has survived for centuries basically unchanged. Why? Because of its “multiple constraints” that combine to limit choices and to cue memory--a key feature of other oral ballads and rhymes, say David Rubin and his colleagues. Here’s the explanation, from a recent paper by Rubin, Violeta Ciobany of Bucharest University and William Langston of Denison University: “Most of the words contain a repeated sound pattern, usually word repetition, rhyme or alliteration, and all the words not involved in the meaning are involved in one of these poetic devices. “Consider the first line, which has remained stable without any deep structure. The first word, eenie is part of the second word meenie. Meenie, miney and mo alliterate. Eenie, meenie and miney rhyme with a sound that repeats as the first vowel of eenie and meenie. Mo rhymes with toe and go. The first lines also contain a progression of front-to-back middle vowels--e, i, o- -as in the fee fie, fo of fee fie fo fum, or the ee, eye, ee, eye oh of “Old McDonald had a Farm.” Therefore, meenie, miney, mo, sounds better than miney, meenie, mo and the order is unlikely to change. “The remaining sound /n/ repeats in the same location in three Words. The whole line repeats as the last line, where the single syllable Word mo, coincides with the person who is chosen. The change from the two-syllable pattern adds to the closure of the piece . . . “Thus, there is not a phoneme or even a distinctive feature in the first line that can change without breaking some pattern. The middle two lines offer more flexibility and do change more over time and over retellings. (An historical note: The pre-tiger victim of the toe-catching is a term now considered a racial epithet, but was not so when the rhyme originated. Rather, the word referred to the River Niger, and was a neutral term for a person from that region.)