Give people fair warning that you are about to trick them into recalling something that never happened and most will still fall prey to the deception, creating "illusory" or "false" memories that sometimes include vivid details, according to new research from Washington University in St. Louis.

"Fully informing people and warning them about the possibility of illusory memories does not permit people to control their thought processes and avoid having them," said lead investigator Kathleen McDermott, Ph.D. "It's clear that people have difficulty suppressing false memories. The key questions now are how and when are these mistaken memories generated and can they be avoided."

Published in the October 1998 issue of the Journal of Memory and Language, the study is another step forward in a recent spate of research aimed at unlocking the mysteries of human memory. It also sheds new light on the especially elusive enigmas of false memories.

The study is the latest in a series of illusory memory experiments conducted by McDermott and Henry L. "Roddy" Roediger III, Ph.D., James S. McDonnell Distinguished University Professor and chair of psychology in Arts and Sciences at Washington University. McDermott, a former doctoral student of Roediger's, is now a research associate in radiology and a research assistant professor in psychology at Washington University.

Roediger, a leading authority on how the mind stores and retrieves knowledge, has spent nearly three decades studying the intricacies of human memory. He is perhaps best known for research on implicit memory, how past experience can be expressed in behavior without a person's intention or awareness; and on memory illusions, why people sometimes remember events quite differently from the way they happened, and in dramatic cases, how people can come to have vivid memories of events that never happened.

"The idea that our memories hold a literal record of our past like a video recorder is wrong," said Roediger. "Rather, remembering is a constructive process and illusions of memory are the result of our struggle to weave the remembered pieces of our past into a coherent narrative story."

Roediger and McDermott helped spark the study of illusory memories with a 1995 paper that showed how lists of associated words could be used simply and precisely to measure levels of false memory creation under various conditions. Psychologists had been using word association tests - What is the first word that comes to mind when I say the word dog? - to construct lists of related words since early this century, but Roediger and McDermott were the first to use these lists to study false memories.

In recent experiments, for instance, Roediger and McDermott have shown that undergraduate students exhibit "remarkable levels of false recall and false recognition" when asked to identify words that were included in a previously viewed list of associated words. Asked, for instance, whether the word sleep was included in a list of 15 related words, such as bed, dream, blanket, doze and pillow, as many as half of the students will incorrectly answer in the affirmative.

Advance warnings
Their latest study takes the word-association experiment a step further to determine if providing explicit advance warnings about the deceptive nature of the memory test has any influence on a person's ability to avoid false memories.

Once again, researchers elicited false memories by presenting students with a list of 15 related words, all of which had been shown to be strongly linked to a specific target or linking word. Students were presented with 20 such lists, in half of which the critical linking word was substituted into the list in place of one of the related words. Students were asked to listen to an audio recording of one of the word lists being read and then asked whether the critical linking word had been included in the list.

In previous studies, students had been told that they were participating in a memory study, but not told that false memories were likely or that this was the topic of the investigation. But in this study, Roediger and McDermott altered pre-study instructions, providing as much information as possible to help students avoid making false inferences about whether a word had or had not been included on the list. Not only were subjects informed about the possibility of a false recognition of an associated word, they were given concrete examples of the phenomenon; told, for instance, that when given a list of 15 related words, such as queen, throne and monarch, people often remember a specific, related but non-presented word, king, even though the word was never part of the list.

Subjects were told to be very careful not to let this memory error happen to them and to pay careful attention to whether or not the critical linking word was presented on the list. To make it even easier, their recognition of the critical linking word was tested immediately after the presentation of each 15-word list.

"Results suggest that subjects are quite poor at performing this straightforward task, even when explicitly informed of the false recognition phenomenon," McDermott said. "This illusion of memory appears to be remarkably robust and little affected by instructional manipulations."

The study included three experiments structured to shed light on how false memories are generated. The experiments were motivated by emerging evidence that memory illusions originate when the non-presented but closely associated word actually pops into the conscious mind as the list of words is being presented. An alternative explanation, said McDermott, is that the mind processes the list on a deeper, non-conscious level, priming itself to retrieve closely associated words but never actively introducing them into the brain's "on-line" attempt to rehearse and remember words actually presented in the list.

McDermott and Roediger reasoned that if people were informed that they may have trouble distinguishing what they heard (externally) from what they generated or thought (internally) and if they were told to pay close attention to this distinction and to be careful not to confuse the two sources of information, they might be able to focus their efforts and eliminate their tendencies to create false memories. But even under these conditions, people in this study continued to confuse what they had heard and what they had thought, creating false memories in more than a third of the tests.

Strong compulsion
The results show, once again, that the mind has a strong compulsion to make inferences and fill in blanks as it processes incomplete data. It appears, McDermott explains, that when we hear a list of words, such as hill, valley, climb and molehill, that a web of neurons in our mind naturally retrieves the word mountain and inserts it into our memory of words actually presented on the list. In fact, research by McDermott and Roediger has shown the illusory memory phenomenon to be so strong that some subjects, when pushed, will actually recall vivid details of how the non-presented word was spoken or where it occurred in the list.

Although the power of illusory memories is becoming more widely known, this study has been surprising to many researchers, McDermott said, because it documents the relative inability of forewarned subjects to effectively suppress the creation of false memories.

"It appears to be a very natural process," McDermott said. "I guess it should not be so surprising because that's how we normally perceive the world. For example, a movie is actually a series of individual frames, but we perceive it as one continuous picture. The mind handles memory in the same fashion, using inferences to piece together a sometimes imperfect recollection of reality."