By William J. Cromie
The split-second birth of a memory has been captured with the aid of new brain-visualization techniques. The resulting images can be used to predict whether a specific experience will be remembered or forgotten.
"This has never been done before," said Anthony Wagner, a psychologist at Harvard University. "We actually recorded what the brain was doing during the fraction of a second that a memory is created. The levels of brain activity we saw predicted whether or not an experience would become part of long-term memory."
Wagner worked with other researchers from Harvard, Massachusetts General Hospital in Boston, and Washington University in St. Louis to record brain activity of people reading single words. The readers had to decide whether a word represented a concrete object or an abstract concept, i.e., "chair" or "charisma." They were not told to memorize the words, or that their memories of the words would be tested.
Twenty minutes later, however, they were shown groups of words and asked which ones they saw during the brain scanning. Their answers indicated which words they remembered and which they forgot. Researchers then compared levels of activity in various parts of their brains during remembering and forgetting.
"This is the first work to tie the creation of a simple verbal memory to specific levels of activity in certain areas of the brain," notes Daniel Schacter, head of Harvard's Department of Psychology and a participant in the research.
Most of the activity occurred in the left frontal lobe, located behind the temple, and in the left temporal lobe, further back in the brain. Specifically, much of the action took place in the inner wall of the left temporal lobe in a structure called the parahippocamal cortex. That structure lies on the pathway to the hippocampus, a part of the brain that is vital for storing and retrieving memories.
"Even when two experiences are similarly novel, differences in parahippocampal activity can predict whether one or both will be remembered or forgotten," Wagner notes.
It takes less than a second for a simple memory like this to form, he says.
In complementary experiments at Stanford University in California, researchers found that activity in the right frontal lobe and both left and right temporal lobes predicts which seen scenes will be remembered.
"The two studies give us a more complete picture of how memories form," Wagner comments. "In the simplest terms, you can say verbal parts of an experience appear to be encoded in left frontal regions; visual and spatial parts of an experience appear to be encoded in right frontal regions."
Word memory is intimately related to language ability and the latter function rests in the left side of the brain for a majority of people. In some left-handers, however, language sits on the right side; yet others split language functions between the two sides. All of the people in the Harvard and Stanford experiments were right-handed.
Details of both studies were reported in a recent issue of Science magazine.
The big unanswered question is what causes the difference in brain activity in the first place. Why did people in the Harvard study remember some words and forget others? Why did individuals in the Stanford study remember some scenic photos and not others?
The short answer, says Wagner, is "We don't know."
But being scientists, they have a theory. "It may be an interaction between a person's prior experience and the time it takes to analyze a word," Wagner speculates.
Some words, in other words, require retrieving more information from memory than others. This is related to the time it takes a person to decide whether a word represents something concrete or abstract. "The more you attend to a word and think about its meaning, the greater the level of brain activity and the more likely the word will be remembered," Wagner says.
James Brewer of Stanford adds: "Perhaps one person sees a photograph of Zion National Park and thinks, 'Hey I just visited that place on my way to California.' Another person might think, 'Outdoor desert scene. . . What's the next picture going to be?' The first individual's memory of the picture is likely to be stronger."
Wagner believes this kind of memory imaging could be used to predict the onset of Alzheimer's disease. One of the first areas affected by the malady is the middle of the temporal lobes, and that results in memory loss. Changes in activity in this region might warn of coming problems and enable treatment to begin at the earliest stages of the disease.
At present, no sure way exists to distinguish Alzheimer's from other types of senile dementia. Autopsy provides the only definitive diagnosis.
"Our experiment provides the most direct evidence to date about the role of temporal regions in memory formation," Wagner says, "so it might help spot signs of early brain changes that signal the onset of Alzheimer's."
The new memory-catching method may also provide a way to study natural repair in a brain damaged by stroke. When a stroke strikes the left frontal area, victims may lose all or part of their speech. Many, however, regain at least some speaking ability when the right side of the brain takes over this function. (There is evidence that women regain lost language faster because they are less dependent on the left side than men for language.)
"We believe that the right frontal area is wired-up in such a way that it can take over functions on the left," Wagner explains. "If so, scans of frontal region activity taken just after a stoke, and progressively thereafter, should enable us to watch any functional reorganization that occurs."
These methods might also be applied to learning about learning. Wagner has begun a follow-up study to explore how the brain is involved in learning and, perhaps, how to enhance the process.
"We know that attending to the meaning of something rather than its appearance increases the probability that it will be remembered," Wagner points out. He and his colleagues scanned the brains of people who were asked to determine either the meaning of a word or whether it appeared in lower or upper case. Concentrating on the meaning, they found, produces more cerebral activity and results in better memory.
Other research shows that when people practice repeatedly to learn a new skill, they do better if they take breaks often. "We want to find out why spacing is so fruitful," says Wagner, "and to investigate how to best structure learning and training sessions."
Copyright 1998 President and Fellows of Harvard College