According to Freud’s theory of repression, the mind hides memories of traumatic events in places where they cannot easily be retrieved, in order to prevent overwhelming anxiety. It is these “repressed memories” that the memory-recovering techniques beloved of some psychiatrists aim to unearth.
The existence of repressed memories is taken as a truism by psychiatry. Unfortunately, it has never been verified by rigorous scientific experiment. And that is not a matter of mere academic interest, since memories apparently recovered by psychiatric techniques such as hypnosis—particularly memories of childhood abuse—have sometimes been enough to put people in prison, even when there has not been any corroborating evidence. Moreover, even in cases where an individual has undoubtedly witnessed something traumatic, the reliability of his memories can be critical to convicting the true perpetrator. Witnesses frequently disagree, and this may reflect the way memory forms. Some actual data on the relationship between unpleasant experiences and memory would therefore be welcome.
In this week’s Proceedings of the National Academy of Sciences Bryan Strange, of University College, London, and his colleagues provide some. Rather than abuse their experimental subjects, though, they merely showed them streams of words on a computer screen.
Some of these words (murder, massacre and so on) had bad connotations. Others (meeting, gathering and conference, for example) were emotionally neutral. The subjects of the experiment, who did not know in advance what was required of them, were asked to look at the stream, which was presented one word at a time. Then, when they had been shown it, they were asked to recall the words in it. In the past, this technique has showed that emotionally charged words are more likely to be recalled than neutral ones. What Dr Strange wanted to look at was how well people remember neutral words adjacent to the emotionally charged ones in the stream. He discovered that words immediately preceding emotionally charged ones were less likely to be remembered than normal.
Intrigued, he pushed a little further. Previous work had established that emotion-associated enhancement of memory is caused, at least in part, by the action of stress hormones, in particular norepinephrine, on a part of the brain called the amygdala. He wondered if a similar mechanism was at work in the emotion-associated memory loss the team discovered.
The action of norepinephrine on the amygdala can be blocked by a drug called propranolol. When the researchers repeated their experiments on volunteers who had been dosed with this drug, they found, as expected, that those volunteers did not remember emotional words any better than neutral ones. In addition, however, they found that memory for neutral words which preceded emotional ones improved.
The team was also able to draw on evidence from a patient who suffers from Urbach-Wiethe disease, a rare genetic disorder that can cause damage to the amygdala. They used brain-imaging techniques to confirm that her amygdalas (people actually have two, one in each hemisphere of the brain) were, indeed, damaged. They also measured her cognitive functions—intelligence, attention and both short-term and long-term memory—and found that these were normal. But her memory was not affected by emotion; she remembered emotionally charged and neutral words equally well, regardless of the order they were presented in.
The memory gap
The kind of memory Dr Strange studied is called explicit memory. It concerns facts and experiences—knowledge that can be recalled by conscious effort and can be reported verbally. Researchers believe that explicit memory is formed in several steps. The first is translating newly learned information into so-called neural correlates. This does not involve permanent changes to the brain’s structure. In the second stage, consolidation, structural changes such as the formation and destruction of connections between nerve cells take place. This process involves the expression of genes and the synthesis of new proteins, and Dr Strange suspects that emotion interferes with these biochemical events. As a result, no memory is formed.
Another line of evidence that supports this interpretation is work on post-traumatic stress disorder (PTSD) carried out by Roger Pitman, of Harvard University. Dr Pitman recently conducted a trial to see if propranolol could prevent the development of this disorder, which afflicts those who have been exposed to horrific events, such as battles or plane crashes, with emotionally disturbing flash-back memories. He reasoned that excessive amounts of stress hormones released at the time of a traumatic event might be responsible for overly strong memory formation. Because memory takes time to form, he conjectured that drugs which block the action of these hormones soon after the trauma might decrease the intensity of the memory. This turned out to be true: a course of propranolol started shortly after an acute traumatic event was able to reduce the symptoms of PTSD one month later.
On the face of it, there is something slightly contradictory about these results. It is odd that the amnesia observed by Dr Strange is for events just before an emotionally charged incident, when what is actually desirable is to wipe away any recollection of the incident itself. But a simple laboratory experiment using what are, after all, ultimately harmless words, is not the same as a case of child abuse or the horrors of war. And it seems clear that the amnesia, as well as the memory formation, is in some way a result of the stress hormones.
What is undoubtedly true is that memory, like everything else in biology, is an evolved, functional response. If individuals tend to be better off by not remembering certain things, natural selection will tend to construct their brains that way. Indeed, the existence of post-traumatic stress disorder suggests that individuals are better off without those memories. And in fact, most people do come out of trauma with their psyches intact, so it is possible that what has happened to PTSD sufferers is that the memory-prevention mechanism has gone wrong.
Freud might thus have been right about the reason for what he thought he had observed about trauma and memory. But it looks as though he was wrong about the mechanism. The evidence, though limited at the moment, suggests that memories are not repressed. Rather, they are never formed in the first place. Obviously, no psychiatric technique can recover something that was not there to start with. That is something of which the courts should be acutely aware when they assess the credibility of witnesses. It is also something psychiatrists may care to ponder when they are trying to dredge up “forgotten” childhood memories.