The Memory Doctor

Training Humans

Better living through memory modification.

In 1998, when Elizabeth Loftus published her first behavior-therapy proposal, she stopped short of editing memories. The proposal, Expert Personalized Suggestion, was designed to exploit the self-fulfilling power of imagination, not to alter recollections. Loftus and her coauthor described memory modification as an unfortunate risk of the procedure. It might happen, they wrote, but only as a tolerable “cost” and “side effect.”

To become a memory doctor, Loftus needed one more tweak. She needed to set aside her compunctions about memory tampering and redirect her mind-control skills from the future to the past.

It happened by accident. A few years after she published her EPS proposal, she began a new line of research, investigating whether false memories could affect behavior. The behavior had to be simple, measurable, influenced by memory, and testable in the lab. One obvious candidate was eating. Food aversions were common and powerful. You could plant a memory of a bad experience with a certain food. Then you could test whether this memory affected the subject’s eating behavior.

Loftus’ collaborators tried the experiment first with pickles and hard-boiled eggs. They asked subjects to fill out questionnaires about their food preferences. A week later, they handed out computer-generated reports telling each subject that he had, among other things, gotten sick as a child from eating one of the specified foods. The subject was told that the computer had drawn this conclusion from an analysis of his answers to the questionnaire. The procedure was like EPS, except that the expert was now a computer, and the suggestion was a false memory. After reading the reports, more than a quarter of the subjects said they believed or remembered the incident. When they were asked later about foods they might eat at a barbecue, those who were successfully duped indicated that they were more likely to avoid pickles or eggs.

Loftus performing her food experiment on Alan Alda

When Loftus saw the results, an idea came to her. What if you could produce similar aversions to fattening foods? Could false memories help people eat a healthier diet? She tried the experiment again, first with potato chips, then with chocolate chip cookies and strawberry ice cream. The potato-chip version failed, apparently because potato chips were too familiar as a tasty snack. But the ice-cream version worked. More than 40 percent of subjects were persuaded that they had gotten sick eating strawberry ice cream, and these subjects were significantly more likely to say they would avoid it.

In February 2005, Loftus and her coauthors published the egg study, concluding that “humans can be trained to avoid food.” Four months later, they published the ice-cream study under the title, “False Beliefs About Fattening Foods Can Have Healthy Consequences.” The diet-improvement rationale, originally an afterthought, was now central. The bottom line, they wrote, was that “we can, through suggestion, manipulate nutritional selection and possibly even improve health.”

In the food experiments, all the threads of Loftus’ career came together. Instead of training a rat, she was training people. Instead of using a reward, she was using the techniques she had learned from the recovered-memory therapists. And instead of planting bad memories, she was planting healthy ones. She was a real-life memory doctor.

But she, too, was responding to behavioral reinforcement. The fattening-food experiment was the most widely celebrated study she had ever done. It made headlines all over the world and earned her a place on the New York Times Magazine’s list of the year’s most innovative ideas. She was being rewarded for doing something socially useful. She reveled in the attention and acclaim. She quoted her press clippings in speeches.

In reality, the work she had done on memories of sexual abusewas more important. But that work had hurt and angered people. It had brought condemnation and contempt on her. In diet therapy, she found no such punishment. Again, she was meeting people who desperately needed her. But this time, they were fighting obesity, not deluded daughters. “I wish I could be of more immediate help to these people, but I’m forced to tell them that the line of research is at its very earliest stages,” she wrote in 2007. “Still, it is refreshing to be working on a topic that could indeed be of genuine help to people and one that is not as anger inspiring and dangerous as the topic of sex abuse.”

Loftus speaking at Beyond Belief, a conference of The Science Network, November 2006

All those years, Loftus had been vilified for attacking therapy. Now she was inventing a therapy of her own. But hers was different. It was good for people, and she could make it better. She began to think of ways to magnify its power: ratcheting up the false feedback, perhaps, or showing photos or videos of people being sickened by the designated foods. Soon she was convincing students that they had loved asparagus as children. “Healthier Eating Could Be Just a False Memory Away,” said the published journal article.

Loftus talked about the food experiments the same way she had talked about EPS. The concept, she explained, was to “tap into people’s imagination and mental thoughts to influence their food choices.” But two crucial elements had changed. First, she was now trying to influence behavior through memory, not just through imagination. In EPS, memory modification had been a kind of collateral damage. In the food experiments, memory modification was the whole idea.

Second, she was proposing permanent deception. Psychologists who misled people in experiments were ethically obliged to tell them the truth when the experiments were over. That was why Loftus had abandoned her original idea to plant a memory of being told by your grandmother that you were her favorite grandchild. Now Loftus was severing that obligation. Therapeutic memory modification, as she envisioned it, would be left uncorrected.

In her EPS proposal, Loftus had sketched an informed-consent script that she thought could be defended as truthful. You could tell the patient that imagination exercises would improve her behavior, and, sure enough, her behavior would improve. But that self-fulfilling trick worked only because the future was unformed. Claims about the past were different. To revise the past, you had to lie.

Loftus speaking at the Beckman Center, March 2007

Loftus didn’t flinch at this step. “A therapist isn’t supposed to lie to clients,” she conceded. “But there’s nothing to stop a parent from trying something like this with an overweight child or teen.” Parents already lied to kids about Santa Claus and the tooth fairy, she observed. To her, it was a no-brainer: “A white lie that might get them to eat broccoli and asparagus vs. a lifetime of obesity and diabetes: Which would you rather have for your kid?”

She was correct: The memories she was planting would do more good than harm. The same could be said for EPS and for her old proposal to limit side-effect warnings. But her ideas were becoming steadily bolder. She was thinking less like a student and more like an engineer. She was becoming more willing to deceive and to redefine consent. She was trying not just to shape the future but to reshape the past. And she was noticing none of this. Memory therapy was coming from a place she had forgotten, and it was going somewhere she couldn’t foresee.

Where would it stop? Should a line be drawn against altering the past? If so, where? Those questions were open to debate. But something unsettling had already happened: Loftus had crossed her own line without recognizing it.

Many years earlier, she had worked with survey researchers on a common problem: false behavioral reports. (Read Memory, Sex, and Food: The Price of Error.) Deluded by rosy memories, people misrepresented their conduct. They underreported their sex partners, exaggerated their consistency in using condoms, and claimed better eating habits than they had really practiced. These delusions, in turn, skewed medical diagnoses and the surveys that informed nutritional science. To illustrate the danger, Loftus had sketched a scenario: Suppose researchers were interviewing women to identify dietary factors in breast cancer. If some of the women claimed to have eaten healthier food than they had really consumed, correlations between their outcomes and their actual diets would be obscured. The Myth of Repressed Memory, p. 177.


Now Loftus was planting precisely such memories. The problem wasn’t just where to draw the line against memory modification. The problem was remembering where you had drawn it.

From food to alcohol: Loftus explores new frontiers in memory therapy

Undeterred, she plowed ahead. If memory therapy could change eating habits, why not drinking habits? With grant money from the National Institute on Alcohol and Alcohol Abuse, she and her collaborators moved on to alcohol, using the same method. In a report on their experiment, submitted in March 2010 to Alcoholism: Clinical and Experimental Research, they called their work “a first step in exploring the idea of using memory implantation techniques for the purposes of reducing or eliminating an unwanted behavior.”  

The list of unwanted behaviors was just getting started.

Next: The Future of the Past

Like Slate on Facebook. Follow us on Twitter. William Saletan’s latest short takes on the news, via Twitter: