Huh, what?

In that daily quest for blog topics, er knowledge, I was perusing jamesh's blog and found a fascinating post on memory. Mr. Hindle described an experience where a recent experience seemed like it happened in the distance past just minutes later. He characterized this as an experience jumping right into long term memory. I found this most interesting as the psychology courses I took in university were among the most enjoyable courses I took there. And to think I never would have taken them if I hadn't had some, *ahem*, issues during first year.

The course on memory was particularly interesting and also relevant to me. Specifically the time between recalls that constituted a "long term memory" was surprisingly small. Something like 5 or 10 minutes is typical, but there are other factors. Short term memory is a severely limited resource and something that is easily testable. The brain takes time to encode things from short to long term memory, so asking someone to remember a sequence of numbers will identify how large this resource is. Typical values range from 7-12 digits (as far as I remember). Famous experiments have shown how to increase this size by employing various methods, such as mnemonics, to group the information and relate it to things that are already in long-term memory. The most well-known example is when a volunteer was trained to remember sequences of 40 or 80 digits long. The volunteer happened to be a long-distance runner and knew the statistics of famous runs and runners. He increased his short-term recall by grouping digits into times, speeds or other running-related values.

There are theories on how information is stored in the brain which try and explain a related observation. That is, when presented with new information that is related in some way to knowledge a person already has committed to long-term memory. These theories have names like "schemas" or "scripts". I believe that the people that are able to recall vast amounts of newly-presented information are basically able to directly store the new information into long-term memory.

Another set of research also shows that the human mind has trouble encoding time in a memory. This is one of the reasons that eye-witness testimony is not that reliable - your memory of an event does not implicit record how long it took for the event to occur or even exactly when. Think about memories that are tied to a particular time in your life - the exact time and date aren't simply available - you likely use other clues in the memory to establish the date. You'll here "We were living at address X, and we lived there from 1972 to June of 1978 so that was probably in the spring of 1978." This is why that one event from long-term memory seems as distant as any other.

Myself, I remember as a child that my mother would be going out shopping with me and my sister in tow. On the way to the first store, (we were ~1/2 hour drive away from any useful shopping), she would say "Remind me to buy when we get to the grocery store." I'd remember that and bring up at the appropriate time. I find it hard to do now, but I suspect that I had fewer things to keep track of then. But to be able to do that at all meant I must have stored it in long-term memory. Even now, I wrote half this posting around 11am and the rest around 3:30pm so I had to remember the outline to finish it.

I've gotten a distinct feeling of deja vu when an event embeds itself straight into long term memory like James described. The event occurs and then someone mentions an aspect of it 2 minutes later and I feel like I've always known the event and the question too.

I think it follows that any event or situation that you can recall 5 or 10 minutes later has been stored in long-term memory, but also that there is no guarantee that you'll be able to recall it after some arbitrary time. Unless that memory gets associated with other memories, in particular something that has a easily activated trigger, it may be difficult to recall. It may be expunged later as well - testing this theory is fairly difficult. Memories can "fade" or get crowded out by newer more "important" information. There must be some limit, although testing the limits would be difficult. Personally, I know that I've said out loud "That was awesome - I'll never forget that ever again." Usually I can't recall it within a few months. At the time of occurance, I'm positive it will be hard to forget, that it is well associated within my own mind.

I was going to mention something about flash-bulb memories, but I think that false memories are more relevant to this topic. False memories are information that is introduced into long-term memory in such a way that it seems like they were always present, or created at the same time as a target memory. This is exploiting the mechanism I described as being responsible for what James recalled - that is that memories don't have inherent time encodings. This allows someone to create a new long-term memory (through suggestion, discussion or even indirectly) that appears to be genuine and from the same time as another memory. Leading questions would be an example of this. By posing a series of questions about an event over a suitable length of time, the questioner can inject information. One could ask a question and include a made-up fact about the event. Repeat that made-up information enough, the target could store it in long-term memory and then recall it later as genuine. The test that proved that was scenario where college students were told that their families had been interviewed and they wanted them to recall as much information as possible about the time they were separated from their mother in a department store. They were interviewed weeks later and asked about the "separation event". Several stated they were lost in department store as a child.

As an exercise in recall, I wrote up the description of the test before I found a reference to it. In fact, the experiment was written up as part this article by University of Washington researchers. The actual participants were 18 to 53, they were given a booklet with 3 events in their past, two authentic and the "lost in the mall" story as the third. Relatives confirmed that none of the participants had such an experience. The subjects were asked to write out whether or not they recalled the experience after reading the booklet. They were asked in subsequent interviews if they remembered the events in the booklet (cued with partial excerpts from the booklets) and by the 3rd interview, about 25% remembered the false memory as genuine. Overall, they recalled about 68% of the actual events.

This is further proof that long-term memories have no time component - that one recall is much like any other. The time distance from the initial storage is irrelevant. This concludes today's article on "memory". Brought to you by the letter "Q" and the number 27 (it's prime!)