The Problem with Long-term Memory

Long-term memory is a funny thing. Generally, we think that we have either a very good long-term memory or our perception is that it’s more akin to collecting water with a sieve: nothing sticks.

For those who think that they belong to the first category: heed this word of caution that follows.

Science tells us that many people’s long-term memory is lousy in terms of accurate recall. How so? Psychiatrist Offer carried out a fascinating investigation. He surveyed a bunch of fourteen year olds in 1962 and asked them a whole load of pretty personal questions involving home life, sexuality, religion, parental discipline and general activities. That in itself is not remarkable but this next point is. He not only kept the results of the survey but then re-surveyed the same people 34 years later in 1996. The purpose was to discover how robust the long-term memories were.

The results were interesting. There were significant differences between the adult memories of what happened as an adolescent and what was reported as happened as an adolescent. Accurate, long-term memory recall was generally no better than chance.

What’s going on that would cause such poor memory recall? The act of learning something is called consolidation. This means that the memory makes its way to long-term memory. However, in recalling the memory, the brain must then reconsolidate the memory. Yes, reconsolidate the memory. Many scientists believe that the act of recalling a long-term memory converts it to short term (more accurately called working memory) and that this occurs every time the long-term memory is retrieved. This means that the working memory must be reconsolidated every time it’s recalled to change it back in to long-term memory.

If that wasn’t bad enough, there’s another, even greater problem. In recalling the memory other memories bleed into it, as the brain actively engages in pattern matching the recalled memory with any other memories it thinks are associated with it an any way. The problem is: they may not be. The result? Long-term memory is very suspect to influence.

The authors of the study made the significant conclusion that even more care is required in obtaining accurate historical information that was medical or psychological. This point could be applied to law enforcement offers (who seek to make sure witness statements are taken at the earliest opportunity after the crime is committed) and anyone requiring historically accurate information.

By the way, if we extrapolate the point about memory bleed, then the only way to prevent previous long-term memories having other memories bleed into them is to not recall the long-term memories in the first place! Oh, that’s helpful: not!

You can read more about the study here.