Slashdot Mirror

← Back to Stories (view on slashdot.org)

Forgetting May be Part of the Remembering Process

Posted by samzenpus on from the what-was-I-talking-about dept.

CFTM writes "The New York Times is running an interesting article about how human memory works and the theorized adaptive nature of forgetfulness". From the article, "Whether drawing a mental blank on a new A.T.M. password, a favorite recipe or an old boyfriend, people have ample opportunity every day to curse their own forgetfulness. But forgetting is also a blessing, and researchers reported on Sunday that the ability to block certain memories reduces the demands on the brain when it is trying to recall something important. The study, appearing in the journal Nature Neuroscience, is the first to record visual images of people's brains as they suppress distracting memories. The more efficiently that study participants were tuning out irrelevant words during a word-memorization test, the sharper the drop in activity in areas of their brains involved in recollection. Accurate remembering became easier, in terms of the energy required."

5 of 191 comments (clear)

  1. The question I've always had about memory... by Scoth · · Score: 5, Interesting

    The question I've always had is more along the lines of the filing system - there are times that I can't remember any part of something until someone reminds me of some small part, and it all comes flooding back. That means it was all in there somewhere, I just couldn't find it. I'm wondering what might cause that, and what might be done to improve it. Or, as the article is saying, perhaps we're not meant to?

    1. Re:The question I've always had about memory... by buswolley · · Score: 5, Interesting
      I happen to be a memory researcher at a major University. I also happen to be on a project very similar to the one in the article. However, we are doing the fMRI imaging with children of different ages, as a developmental study. We also piloted adults, and replicating results similar to the ones in the article. Interesting. Of course, I cannot speak about the research in much detail. Journals don't like that much.

      As to your question, I could tell you a lot about why this is so. 1st, cued recall is much easier than free recall. The cue helps stimulate the appropriate associative networks facilitating recall. In particular, a primary focus of mine is cued recall, or recognition. I use the dual process model of recognition: Recollection and Familiarity.

      Familiarity, as experienced, is the feeling of familiarity we get when we see something that we've seen before, aside from actually remembering anything about it, which is recollection.

      I highly recommend the seminal: Yonelinas. A.P. (2002). The nature of recollection and familiarity: A review of 30 years of research. Journal of Memory and Language, 46, 441-517.

      You can get it here: http://psychology.ucdavis.edu/labs/Yonelinas/index _files/page0003.htm

      --

      A Good Troll is better than a Bad Human.

    2. Re:The question I've always had about memory... by SeekerDarksteel · · Score: 5, Interesting

      In artificial neural networks, there are structures called auto-associative memory networks. The networks are "trained" on certain patterns, then when it receives one of those patterns as input, it outputs a pattern closer to the pattern it was trained on. If you make it recursive (and your network is good enough), you can take as input a pattern that contains only a fragment of one of the patterns it was trained on and get as output the pattern you trained on. It's quite likely that something like that is going on inside our brains to store memories in some fashion, but on a far more complex scale than we can describe at this point.

      --
      The laws of probability forbid it!
    3. Re:The question I've always had about memory... by replicant108 · · Score: 4, Interesting

      In any mnemonic system, linking is a key component.

      The assumption is that any given item of information can only be reliably retrieved if it is linked to something already known.

      In computer science, the concept of the linked mist is probably most analogous.

      Clearly an index plays a vital role in such a system.

  2. Francis Crick: REM sleep like simulated annealing by MarkWatson · · Score: 4, Interesting

    In the late 1980s, I participated for about a year on the DARPA neural network tools panel. If I remember correctly (ha :-) it was Francis Crick who suggested that REM sleep was like simulated annealing; that is, serving the function of adding some randomness to a neural network so that we could forget meaningless things that happened to us during the day.