Many science fiction books and films discuss the possibility of memory alteration or implantation. For example, in the cult movie Blade Runner, the human replicant Rachael was implanted with memories from a real person and believed that she was human. Is it actually possible to alter one’s memory or even implant new recollections? If you were an eyewitness, could someone corrupt your memory on purpose without you noticing?
Memory is often represented as something fixed, like an episode shot on a videotape. However, it is seldom like that: many details disappear later, after memory acquisition, or even are remembered incorrectly. Remembering events that never happened is one extreme that was first studied by Loftus and Pickrell. They gave participants brief descriptions of childhood events, provided by their family members. Among them, one event was fake, that is, it was completely made up. Participants were asked to elaborate on these stories, and approximately a third of participants was making up stories prompted by fake events. In 2002, Wade and colleagues used a different strategy to implant false memories: they used childhood photographs. Among them, one was edited: it depicted subjects in the basket during a hot-air balloon ride, although their family members said it never actually happened. Half of the subjects reported at least some memories of the hot-air-balloon ride.
This displays that if we remember details of an event, we need to choose a correct option among many alternatives. Thus, memory recollection resembles a decision-making process. Frequently, when our recollection is faint and we struggle to decide, we simply rely on the opinion of other people. For example, when we are not sure about the deadline for an essay submission, we normally ask our fellow students. But what if we are pretty sure, while everyone else says something different? Can we resist the social pressure?
It has been known for a long time that social environment can affect our ability to make decisions. The most famous study on social conformity is probably the Asch experiment (performed in 1956), by the name of American social psychologist Solomon Asch. In his experiment, participants did a simple perceptual match to sample task: they were presented with a standard line on the left, and three comparison lines on the right, and they had to decide which of the comparison lines matched the standard one by length. There was only one correct answer, and other lines were obviously different from the standard.
The study was framed as a visual discrimination task (as showed below). In fact, there was only one real participant in the study: the others were confederates, who were instructed beforehand to answer according to the experimenter’s instructions. In most trials, confederates responded correctly, but in critical trials, confederates unanimously gave a wrong answer. Confederates gave their answers out loud one after each other, and the actual participant was the last one to decide. Around three quarters of participants conformed to the wrong answer at least once, and overall, about a third of responses were wrong. As the individual error rate was less than 1%, Asch concluded that the urge to conform with the group made participants choose a wrong answer.
In Asch’s experiment, information was perceived directly. It’s harder when the information has to be recollected from memory and is not immediately available in front of one’s eyes. So what if people need to decide based on what they remember?
Long-term memory conformity
Indeed, people often alter their recollections due to peer pressure. This phenomenon is known as ‘memory conformity’. Sometimes people conform only to please others within a certain situation, while retaining the original memory and returning to it when social context is no longer relevant. This is called public conformity. On the other hand, a phenomenon of private conformity also exists: under social pressure, people can genuinely change their memories and keep conforming to it, even when the social pressure is no longer present.
In terms of behaviour, the two types of memory conformity are hard to tell apart. However, since these are different cognitive processes, researchers suggested that brain circuits involved in them are different too. To test this, social neuroscientist Micah Edelson and colleagues conducted the following study.
Participants viewed an eyewitness-style documentary in groups of five. Three days later, participants returned to the lab and individually did a memory test. Participants also rated the confidence of their answers. This was to set up a baseline before the manipulation stage, which occurred four days later. Participants were re-invited to the lab and had to answer the same memory questions in a magnetic resonance imaging (MRI) machine, that recorded participants’ brain activity. Participants observed false answers by the other four participants from the initial group of five, whose photographs were also shown next to their answer. In the trials for which a participant initially had a confident correct memory (as revealed by the first test), experimenters manipulated answers and put false answers next to the names of the group members. In matched control trials, there was just a letter X instead of answers. The results showed that indeed, the manipulation worked: participants conformed to the majority opinion in 68% of the trials.
A week later, participants came back to the lab and were told that the answers given by the co-observers during previous MRI session were determined randomly, with the intent to make participants believe that the information of their fellow co-observers was irrelevant. Then participants did the memory test again, to see what memory error persisted after the social manipulation was removed. Participants returned to their initial answer only in 60% of cases, however, in 40%, they stuck to the wrong answer, which showed that they incorporated the false memory. So even with no sophisticated procedures, with no brain helmet devices as pictured in sci-fi movies, people’s memories could be changed without their awareness. However, it should be noted that some people actually noticed the experimental manipulation.
The functional magnetic resonance data showed that long-lasting memory changes were predicted by enhanced amygdala activity and enhanced connectivity between amygdala and hippocampus. Amygdala is well-known for its role in social cognition, and hippocampus is one of the main areas involved in memory.
Long-term memory conformity and oxytocin
What if we could make the social context more powerful, for example, by means of pharmacology? It has been known for a while that oxytocin, a hormone which is involved in social behaviours such as bonding and trust, and is sometimes even referred as a ‘love hormone’ (you can find more on oxytocin and its role in making contexts more salient in Antje’s blog post). So what if people are given oxytocin and then expected to do a task as in the experiment described earlier? Would people be more trustful and compliant?
Recently, Edelson and colleagues actually tested this. Participants underwent the same procedure as in the previous experiment, while divided into two groups: one of them had used an oxytocin nasal spray some time before the second memory test. The control group received a placebo nasal spray. The results showed that indeed, participants were more compliant after the oxytocin application.
However, this effect was not long-lasting: in the final test, when participants were told that the answers of their co-observers were false, the level of compliance was decreased. The possible explanation for this might be that people more readily agreed with the opinion of their social group under the effect of oxytocin, but it reduced the conflict between the individual’s belief and group’s opinion. This, in turn, reduces the attention to the social information and consequently encoding into long-term memory is attenuated.
This study was the first to highlight how pharmacological substance affects public and private conformity. The neural substrates of this effect are still unknown, and should be a target of future inquiry. As increased activation in amygdala was found to be enhancing the encoding of the persistent memory error under social pressure, oxytocin putatively affects the connectivity between amygdala and hippocampus.
Why is it important?
Private memory conformity effects can have a large impact on eyewitness testimonies. When other witnesses report observations that conflict with one’s own perception or memory one might eventually comply with the others. What’s more debilitating, after some time, individuals may start to believe that this was what they actually saw. An American organization The Innocence Project helps prisoners who can be proven innocent through DNA testing. In pardoned cases, eyewitness misidentification has been identified as a major factor that contributes to wrongful convictions, estimated at more than 70% of exoneration cases (that is 235 cases since 1992).
Moreover, social pressure can be a very powerful tool for implanting false beliefs. The use of oxytocin to induce a short-term compliance in eyewitnesses, for example, just before elections, sounds too sci-fi to you?
Like in Orwell’s ‘1984’, if everyone were to say that 2+ 2 =5 or black is white, some people might actually embrace this as truth. In fact, one of the versions of Asch’s experiment conducted in Soviet union in 1971 was exactly making people saying black is white: people were presented with two pyramids, black and white, and under the pressure of a group saying that both are white, participants complied to this obviously wrong majority opinion by stating that both pyramids were white.
In the current world situation, when humanity is faced with many challenges: wars, the refugee crisis, and aggravation of international relations, people interpret events according to what they see and hear in the media. Through information dynamics that are induced by the Internet and other mass media, manipulation of opinions and world views seems as easy as declaring black as white. Just as in Blade Runner, it is difficult to discern truth from lies. Even in this post.