Friday, May 02, 2008

My brilliant brain.
























At the moment I'm reconsidering the use of Chess Tempo as a means to improve my scanning skills. On the one hand the problems are rather simple, on the other hand it is obvious that my scanning skills leave a lot to be desired, even with simple problems. I rewatched this video from National Geographic about Susan Polgar for the 4th time. She undergoes a fMRI-scan by prof. Joy Hirsch from the New York Neurological institute (around 40th minute).

Susan is presented with chesspositions from her youth and has to treat them as a normal chessposition, and she has to think about the next move. Prof. Hirsch shows that Susan has hijacked here fusiform face area (FFA) in her brain , which is commonly used to process face recognition. Susans FFA processes chess positions too. A chess position is checked against an internal database in about 0.8 seconds.

To me this suggests that pictures are important. I take great pains to see the input-output (see previous post) of a chessposition as a picture.

When I'm solving the problems of Chess Tempo then I often can solve them in 10-30 seconds when they are simple. From time to time I recognize a position "immediately", that is to say in under 3 seconds or so. Are those processed by my FFA and is that the way to go?

9 comments:

  1. I just realized that I never really thought about it like that. I mean, the idea of neural reorganizing has been the basis behind pretty much all training I've done, but I always thought that when I instantly recognize a tactic, I'm sort of cheating by "knowing the questions before the exam", that it's 'too easy' and hence has insufficient training effect.

    but now, thinking about how faces are recognized, we don't really have to use any effort to recognize faces. we don't learn them by consciously measuring the slight differences in the proportions of the face, we just zap it into memory, and that's it. and from my art background I already know the differences are extremely small, and even 1mm offset in where a crucial feature should be makes the face look nothing like the person we're trying to draw. (also, it's almost impossible to communicate to another person why a line should be "like this" and NOT "like that", you just "know it's the right way to the draw a feature in this kind of situation". -sound familiar? yeah, exactly what happens with chess.)

    but then there's of course the fact that the region in brain is biologically 'meant to' process facial features, where as its capabilities for position processing is sort of forced into it. which might make a huge difference in how 'easy' the processing should be and how much mental strain yields the best result.

    ReplyDelete
  2. heh. I just went over to bdk after a long break, and found out you guys have been talking about all this already. and here I thought I had something fresh to say. :)

    ReplyDelete
  3. About face recognition: Take an intense look at the dog of the post. I bet you can recognize that face among the persons in your neighbourhood!

    ReplyDelete
  4. nah, there are only cool cats living in these parts. :)

    I just tried KNB on ICC (bot) for the first time in maybe a year. it appears that I was right at least about the knight/bishop vision. -I used to be able to blitz it through in 15s, but now it took me 8 tries (using minutes every time) until I got it right. and even after that I failed a couple of times. -I just couldn't see the controlled squares anymore, and I had hard time visualizing even the last few moves well. obviously I need to drill it for a while, to get back the visualisation & coordination skills I remember having.

    ReplyDelete
  5. Interesting on how they link Susan's ability to the face recognition activity in the brain. I'll bet that there are other parameters involved too. Emotional triggers also add to memory catalogging capabilities. For instance, if a position is Quiet and not all that stimulating, you won't recall it as quickly as one with a lot of dynamics like Anderssen vs Kiesritzky's Immortal casual game.

    Just like facial recognition happens more often when the face is presented with either a happy memory ( a pretty face for instance) or a jarring memory ( Someone screaming at us from the car next to us when we cut them off. How well does one recognize a face that doesn't provide any additional stimulus?

    ReplyDelete
  6. Very cool, I hadn't seen that and it gives me a good surge in my confirmation bias module. :)

    ReplyDelete
  7. In general: the lower the startrating, the higher the improvement (surprise)

    ReplyDelete
  8. BP,
    maybe we should play Wagner when solving dull positions:)

    ReplyDelete