cognitive fun! talk RSS http://cognitivefun.net Cognitive neuroscience for everyone! Sun, 26 May 2013 06:35:02 +0000 Hi your case is very similar to mine, well except that both STM and LTM were slow and LTM was a lot slower, but with dual n back I feel like STM has improved a lot! both in size and speed, I also have a high anxiety level ( I don't know about the epilepsy) as a matter of fact I sometimes suffer from episodes of panic or I don't know how to call them but I get very anxious and afraid for no reason, it is supposed that it is caused by clinical depression ( lack of neurotransmitters). So I don't know if it's epilepsy or what but thank God it is not so often. and I a software developer (DB and shell programming and excel sheets etc) and I really feel like I can finish things a lot faster with 4 months of dual n back (january 2010 to may the 4th )and some games from one site in internet (lumosity) I haven't been so constant in dual n back though in this 4 months I have only played it about 50 days. So the main point here is that maybe we could compare results (at least subjectively but it can be objectively if we agree to some sort of test) and tips that work for each other etc so we can both advance faster in our brain training Anyone is invited also to pass tips of methods that have worked for you (sorry if my english is not good or something I am not a native english speaker I speak spanish) my mail is alexman128@gmail.com Wed, 05 May 2010 04:04:59 +0000 http://cognitivefun.net/talk/post/4428#r_20411 You can still use one object for multiple digits in backwards span! Tue, 26 Aug 2008 05:28:50 +0000 http://cognitivefun.net/talk/post/4428#r_7703 Using mnemonics (presumably method of loci) do you find backwards and forwards digit span tasks equally difficult? Though I imagine the way to increase forward digit span (24=dominoes instead of 2=swan 4=door) makes backwards digit span more difficult, when you are only putting one object in each locus, are the two tests equally hard (or easy)? Tue, 26 Aug 2008 05:10:44 +0000 http://cognitivefun.net/talk/post/4428#r_7696 Oh, and the ability to manipulate and rechunk items in working memory might be quite useful too, causing some extra associations to be activated. Example: I may initially chunk 1-9-8-4-6-4 as 198-464 Then I realise that first digits form "1984", a title of a novel by Orwell and make it 1984-64, where 64 is 8^2 so it's also a distinctive chunk which is hard to mix up with other digits. So in some instances we have both temporarily increased STM load (reinvoking the digits in the middle of the series, reencoding as new chunks) and some extra LTM load (invoking associations to digit groups). Not bad at all. Sat, 21 Jun 2008 09:10:25 +0000 http://cognitivefun.net/talk/post/4428#r_4473 Well, if retrieval from long-term memory improved in a general sense (and not only for digits-symbol encoding), it would be very useful for me. My working memory is OK-ish (it only "resets" randomly), but both storage, recognition and retrieval from LTM are quite slow and unreliable, due to damage to hippocampus visible in MRI (from lifelong high anxiety level or temporal lobe epilepsy or both). If there was a technique that would improve that part, it would be quite useful for me - unfortunately, as memory is rather scattered around the brain, and as I have slight structural damage, I suspect improving LTM in general and not domain-specific sense would require some very complex exercise design. On the other hand, I don't know if it's a coincidence or training effect, but my last 2 or 3 weeks at work were much more productive (at complex software refactoring tasks which I previously delayed for several weeks, I'm a chronic procrastinator). Exactly the type of tasks that would benefit from improved short-term memory and executive function. I mostly do digit span tasks and n-back (up to 5), as well as low-N PASAT and reaction time. Anecdotal evidence, and quite unreliable too, but that's all I have ;) Sat, 21 Jun 2008 08:58:27 +0000 http://cognitivefun.net/talk/post/4428#r_4472 High digit span is domain-specific. That is, improving your digit span does not imply a general increase in WM function. I agree with your suspicions that improvements result from chunking and pattern recognition. People with high domain-specific task performance usually show brain activation for episodic memory retrieval. Essentially, they've become very quick at accessing related long term memories and use them as short cuts instead of relying completely on the working memory buffer. While this is a bit unfortunate, I think the digit span task is a bit more special, because even if it is domain-specific, it is still very useful in daily life :) Sat, 21 Jun 2008 05:24:24 +0000 http://cognitivefun.net/talk/post/4428#r_4470 The same principles can be used to memorise shapes or letters. You are right however that using the mnemonic techniques means you are not using general working memory in the way that you would if you were not using the techniques. I'm not aware of any evidence that practicing the digit span task leads to improvements in working memory (although I would be interested to hear if any exists; the dual n-back study of Jaeggi-Buschkuehl suggests that this sort of thing might be possible). If someone shows improvements in digit span through practice, I suspect it is far more likely to be because they are getting better at using chunking, pattern recognition, and other mnemonic strategies, rather than because their general working memory is improving. Sat, 21 Jun 2008 02:38:26 +0000 http://cognitivefun.net/talk/post/4428#r_4466 I've practiced the techniques quite a bit; I've been using them on and off for the past ten years or so, although mainly for things like learning foreign language vocabulary rather than numbers. Even so, it took a bit more practice before I was able to apply the mnemonics to the digit span task. Using mnemonic techniques for the digit span task is difficult because you have to memorise quickly, and it is essential to be able to translate the numbers into images quickly to do this. So you need to be very well drilled in associating the numbers with images, which takes a lot of practice. Sat, 21 Jun 2008 02:24:19 +0000 http://cognitivefun.net/talk/post/4428#r_4465 it is very nice but if I gave you 20 letters or 20 shapes instead of digits it wouldn't work. My goal is to improve the general working memory span and to do that I can't use any chunking or mnemonic strategy. Fri, 20 Jun 2008 03:18:58 +0000 http://cognitivefun.net/talk/post/4428#r_4462 That's really awesome, thanks! How much have you practiced these techniques? Fri, 20 Jun 2008 04:36:09 +0000 http://cognitivefun.net/talk/post/4428#r_4456 Sure! I am using a memory system, with numbers represented by mental images. The system I use is very similar to that described by former world memory champion Dominic O'Brien in a number of books he has written. I group the digits into groups of four; each four digit group is represented by a mental image of a person performing an action using a particular code (for example, for 3644 I vividly imagine Michael Jackson riding a horse). I imagine these images in a "journey" or "locus", such as a building I know well. Without the mnemonic systems, I would only be getting a digit span of 8 or so. With it (and with quite a bit of practice) I have so far got up to 20. But a lot more is possible with a trained memory - the world record for memorising a spoken number (spoken at one digit per second) now stands at 188 digits. As noted on the description of the test on cognitive fun site, the mental processes used with a trained memory like this are quite different to the processes you would use without the training. Using the mnemonics, you are using "longer term" storage rather than short term working memory. Wed, 18 Jun 2008 09:56:13 +0000 http://cognitivefun.net/talk/post/4428#r_4435 Care to share how you are doing it? Wed, 18 Jun 2008 04:46:46 +0000 http://cognitivefun.net/talk/post/4428#r_4429 TimM is getting a 20-digit span! Wed, 18 Jun 2008 04:46:46 +0000 http://cognitivefun.net/talk/post/4428#r_4428