Trust your memory? Maybe you shouldn't

Story highlights

  • Elizabeth Loftus is a psychologist and expert on the malleability of human memory
  • Loftus is most known for her work on memory distortion and false memory implantation
  • Her research could also be used to fight obesity, addiction with false memories

This is part of CNN's "Life's Work" series, which features innovators and pioneers who are making a difference in the world of health and medicine.

(CNN)You probably feel pretty attached to your memories -- they're yours, after all. They define who you are and where you came from, your accomplishments and failures, your likes and dislikes.

Your memories help you separate friends from enemies. They remind you not to eat too much ice cream or drink cheap tequila because you remember how horrible it felt the last time you indulged.
Or do you?
    One conversation with Elizabeth Loftus may shake your confidence in everything you think you remember. Loftus is a cognitive psychologist and expert on the malleability of human memory. She can, quite literally, change your mind.
    Her work is reminiscent of films like "Memento" and "Eternal Sunshine of the Spotless Mind," where what you believe happened is probably far from the truth -- whether you're the eyewitness to a crime or just trying to move past a bad relationship.
    "She's most known for her important work on memory distortion and false memories," says Daniel Schacter, a psychology professor at Harvard University who first met Loftus in 1979 and describes her as energetic, smart and passionate. "It's made people in the legal system aware the memory does not work like a tape recorder."
    In fact, Loftus' research shows your memory works more like a Wikipedia page -- a transcription of history created by multiple people's perceptions and assumptions that's constantly changing.
    Eyewitness testimony
    Elizabeth Loftus is a cognitive psychologist at the University of California Irvine.
    One of Loftus' first experiments, published in 1974, involved car accidents. In the lab she played videos of different incidents and then asked people what they remembered seeing. Their answers depended greatly on how she phrased the question.
    For instance, if she asked how fast the cars were going when they "smashed" into each other, people estimated, on average, that the cars were going 7 mph faster than when she substituted the word "hit" for "smashed." And a week after seeing the video, those who were asked using the word "smashed" remembered seeing broken glass, even though there was none in the film.
    Even a seemingly less important word in the sentence can make a difference in an eyewitness account, Loftus found. In a subsequent study she asked people if they saw "a broken headlight" or "the broken headlight." Those who were asked about "the" broken headlight were more likely to remember seeing it, though it never existed.
    Police officers' biggest mistake is talking too much, Loftus says. "They don't, you know, wait and let the witness talk. They are sometimes communicating information to the witness, even inadvertently, that can convey their theory of what happened, their theory of who did it."
    This is particularly troubling when witnesses are identifying a perpetrator in a lineup. One of Loftus' studies found even facial recognition can be "contagious" -- if a witness overhears another witness or police officer describe a misleading facial feature, they are more likely to describe the criminal with that feature.
    It's not all the cops' fault. "Misinformation is out there in the real world, everywhere," Loftus says. "Witnesses talk to each other ... they turn on the television or read the newspaper if it's a high-publicity event. They see other witnesses' account. All of these situations provide opportunities for new information to supplement, distort or contaminate their memories."
    Loftus has testified in and consulted on hundreds of trials over the past several decades, usually for the defense. Many were high-profile cases, including those of the Hillside Strangler, Michael Jackson, Martha Stewart, Oliver North and Phil Spector.
    She's not bothered by defending people others sometimes see as vicious criminals.
    "DNA testing ... has revealed that there are hundreds and hundreds of people who have been convicted in crimes, and they're completely innocent," she says, noting that they're often convicted because of unreliable eyewitness testimony.
    Repressed memories
    Perhaps Loftus' most powerful -- and controversial -- work came in the 1990s when she first began manufacturing false memories.
    In 1990, Loftus got an intriguing call from the defense attorney for George Franklin, father of Eileen Franklin. In her mid-20s, Eileen Franklin claimed she remembered seeing her father rape and murder her best friend as a child. The prosecution said she had repressed the memory up until that point.
    Loftus testified at the trial about the fallibility of memories but could not say whether she had ever studied repressed memories such as Eileen Franklin was maintaining. George Franklin was convicted, and Loftus went back to the lab.
    After doing some research, she became convinced a therapist might have led Eileen Franklin to suspect her father in the murder. Therapists were essentially guiding patients to remember false events, Loftus believed -- asking leading questions and telling their patients to imagine an event that might have happened.
    For example, if a woman came in with an eating disorder, her therapist might say "80% of patients with an eating disorder were abused. Were you?" Then the therapist might ask the patient to think about who might have abused her and when.
    While Loftus couldn't definitively prove that repressed memories weren't real, she could show that it was possible to implant a memory of a traumatic event that never happened.
    Loftus recruited 24 students and their close family members for her 1995 study "The Formation of False Memories." She asked each family member to provide her with three real childhood memories for their student, and then sent these memories in a packet, along with one false memory, to the study participants. The false memories were about getting lost on a shopping trip and included real details, such as the name of a store where they often shopped and siblings they were likely with.
    The students were told all four memories were real and had been supplied by their family member. After receiving the packet, the students identified whether they remembered each event and how confident they were that it had happened to them. In follow-up interviews the researchers asked them to recall details from the events they remembered.
    Seven of the 24 students "remembered" the false event in their packets. Several recalled and added their own details to the memory.
    "It was pretty exciting to watch these normal, healthy individuals pick up on the suggesti