The person being interviewed on the quirky NPR radio program This American Life told a story that happened to him many years back. He had been walking with his wife in New York City when he saw Jackie Kennedy across the street waving at him. Since he did not know her, he looked around to see if she was waving at someone behind him but there was no one there. Not wanting to snub a former first lady, he waved back genially just before a taxi halted before her and he realized that she had merely been hailing a cab.

This was a mildly amusing anecdote, but what is more interesting is that his wife interrupted him at that point to say that he had not been present on that occasion at all and that the incident had actually happened to her when she had been walking alone. But she had recounted the story to friends many times in her husband’s presence and somehow it had got embedded in his own memory as his own story. Her husband took some persuading that he had imagined his role because he remembered very vivid details about the incident but his wife pointed out that he could not have possibly seen some of things he said (like the kind of buttons on Jackie Kennedy’s coat) because she was too far away.

I think all of us have experienced the feeling of astonishment when something that we remembered quite clearly turned out to be not true or quite different. I have this vivid memory that when I was about six years old, a major fire broke out in a house a few doors away from where we lived. I remember clearly the flames shooting up into the dark night sky. We watched the blaze for some time but at some point my father decided it was coming too close and put us in the car and drove us away to safety. Our house was spared and we returned later that night. But recently, I happened to ask my mother and older sister about this and neither of them could remember the event at all.

How could this happen? The answer is that memory is a tricky business. When we experience something, we tend to think that the event is recorded by some kind of VCR in our brains to be played back later. In such a model of the brain, there may be some degradation in the recording and playback modes but there should be no major distortions. The main story should not change, and characters, chunks of dialogue, and events should not appear or disappear mysteriously.

But in actuality, the brain is not at all like a VCR, it is more like a computer. It appears that events are broken up into pieces and stored in multiple locations. Then when we “remember”, the event is not recalled like a videotape playback, it is reconstructed from the separate stored bits. It is similar to the way that a computer stores a program in its hard drive and then runs it later.

Of course, this means that there must be some algorithm in the brain that breaks up the experience into bits and stores them, another algorithm for retrieving these bits and reconstructing the memory, and some other mechanism that is constantly implementing these algorithms. The problem is that these things don’t seem to function with 100% accuracy, with the result that when we recall something, it is possible (if not likely) that the memory is not a faithful reconstruction of the original. Even a slight malfunctioning in the algorithm can cause a major distortion in the memory. The distortions become more pronounced with greater time lags and if we have a strong emotional investment in the event.

(I have noticed this again when writing this blog. Many times I recall some news item or quote that is relevant to a posting. But when I look up the original source, I often find that the story is not quite what I thought it was. This is why academics place such a premium on reading and citing the originalsources of anything, as far as possible. It minimizes the chances of distortions creeping in.)

So maybe my fire memory was due to an algorithm malfunction that erroneously coupled images of a fire (from some other source) with my father taking us out in his car, or something like that. This is why trusting to memory alone is a dangerous business, especially if the stakes are high, and why notes of something taken contemporaneously of an event are more valuable than verbal recollections.

At one time in the 1980s, there was a rash of high profile court cases involving the sexual abuse of children by workers in day-care centers. There was a huge media frenzy about it with several day care center workers going to jail. Many of those convictions were subsequently overturned but the lives of those day care workers were already ruined. The recent flurry of cases involving abuses by Catholic priests follow a familiar pattern.

Sexual abuse of minors is a serious problem in America (and anywhere it occurs) and it is one of the most despicable of crimes. But while the abuse of children is a horrible thing, we have to be careful to not let our disgust destroy the lives of innocent people who are convicted without having sound evidence against them. This is particularly relevant when adults (as witnesses) recall events that happened to them as children and were supposedly suppressed because of the trauma associated with them. The website of the National Center for Reason and Justice argues that some of the well-publicized cases of child sexual abuse were based on shaky science and shaky memories.

So how can we reliably distinguish false memories from real ones, since so much can depend on doing so? The book The Art of Changing the Brain by my colleague James Zull of the Biology Department (p. 84) explains how brain scanning science may be able to help.

[T]he brain behaves differently when we recall something that really happened and when we recall something that didn’t happen. [Schacter’s] studies showed that the part of the brain that is needed for memory formation, the parts around the hippocampus, was activated by both false memories and true ones. But when a true memory was recalled, a second part of the brain also became more active. This second part of the brain was the sensory cortex that was involved in the real event when it was sensed by the brain. In these experiments, people were asked to remember spoken words, so recall of words that had actually been spoken activated the hearing part of the cortex, the auditory cortex. This part of the brain was silent for the false memories.

(emphasis in original)

In other words, real events usually are accompanied by actual sights and sounds, and possibly also tastes, smells, and physical contact. These physical stimuli are stored in a different part of the brain from that where the memory is stored and should be activated in the memory reconstruction process. False memories do not have such recorded physical sensations.

I am not sure if brain scanning data will become reliable enough to be part of criminal trials. I am also not sure if the sensory cortex retains actual sounds and sights from very long ago. But it would be nice if we had help in identifying which memories are reliable and which are not.