Neuroscience

Remembering a Detail That Was Never There

·

A commuter carrying a bag on a railway platform.

Someone takes a bag at a railway station. Later, another witness mentions a red strap. By the time an interviewer asks about the strap, red is a familiar detail. Where did that familiarity come from? A witness can remember a perfectly real conversation and still give the wrong answer about the bag.

The strap is an invented example of a problem that memory researchers can actually measure. A laboratory can keep a record of what appeared in a scene, arrange for a misleading detail to turn up afterward, and then ask what the witness saw. When cannabis enters the experiment, the interesting question is which part of that sequence becomes more vulnerable.

A railway platform with a record of everything

In a double-blind study led by Lilian Kloft, volunteers received cannabis containing THC or a placebo. Virtual reality put them inside events that researchers could replay identically. In one scenario a participant witnessed an incident. In another the participant played the role of a perpetrator. A virtual railway station lacks the stakes of a real crime, but it gives the experiment an extraordinary advantage over an actual investigation. The research team knows exactly which details existed.

The study’s interview materials reveal a small but consequential choice. Volunteers answered yes or no, and if they were unsure, they had to guess. Some questions concerned things that really appeared. Others introduced or repeated a suggestion. Still others asked about things that had neither appeared nor been suggested beforehand. That last category made it possible to spot a tendency to say yes even when nobody had previously planted the detail.

Two ways to get the same answer wrong

Return to the strap. One witness may picture it vividly because the later conversation supplied both the color and an opportunity to imagine the bag. Another may barely remember the bag and agree that red sounds plausible. Both people tick the same box. An interviewer who sees only the tick cannot tell how they got there.

Across its tasks, Kloft’s team found greater susceptibility to false memories under acute cannabis intoxication. Errors could extend to details with little connection to what a participant had actually experienced. The distinction matters because a simple account in which THC makes people especially receptive to someone else’s story leaves out an important possibility. Uncertainty about the scene, combined with an obligation to choose, can also produce a false report.

That makes correct answers informative, too. Imagine two groups that recognize the same number of genuine objects. If one group also endorses many more objects that were missing, its members may have used a more permissive rule for deciding what counted as familiar. Looking only at the genuine objects would conceal that difference. Looking only at the mistakes would conceal what the witnesses remembered accurately.

What survives a week?

The researchers interviewed volunteers soon after the simulations and again roughly a week later. The second interview asks a different question. The original event, an intervening suggestion and the first interview are now all in the past. A participant can recognize the wording of a question from last week without being able to retrieve the original scene. A separate experiment at the University of Chicago isolated retrieval by giving THC when volunteers were asked to remember previously encountered material. That design allowed the researchers to investigate errors introduced when a memory was called back. The pattern of differences in Kloft’s study depended on the task and the delay, so an error rate from the first interview cannot simply be carried over to the second.

People who have taken no drug are also susceptible to information encountered after an event, as Elizabeth Loftus describes in her review of experiments on misinformation. That broader history explains why the placebo comparison is so important. It also puts the laboratory result in perspective for anyone listening to an account of a real incident. Before attaching weight to the apparent precision of a detail, it helps to establish when it first entered the conversation. The red strap might tell us something quite specific about an interview that happened later.