"How to conquer n-back hill" by Ashirgo :)

[Ashirgo]

Hi folks!

I decided to post the "technique" I use to conquer the n-back hill in
that separate thread. The only proof of her efficacy is that (till
now) I reached 12-back with it.

1. Be focused at all cost. The fluid intelligence itself is sometimes
called "the strength of focus".

2. You had better not rehearse the last position/sound . It will
eventually decrease your performance! I mean the rehearsal "step by
step": it will slow you down and distract. The only rehearsal allowed
should be nearly unconscious and "effortless" (you will soon realize
its meaning :)

3. Both points 1 & 2 thus imply that you must be focused on the most
current stimulus as strongly as you can. Nevertheless, you cannot
forget about the previous stimuli. How to do that? You should hold the
image of them (image, picture, drawing, whatever you like) in your
mind. Notice that you still do not rehearse anything that way.

4. Consider dividing the stream of data (n) on smaller parts. 6-back
will be then two 3-back, for instance.

5. Follow square with your eyes as it changes its position.

6. Just turn on the Jaeggi mode with all the options to ensure your
task is closest to the original version.

6. Consider doing more than 20 trials. I am on my way to do no less
than 30 today. It may also help.

7. You may lower the difficulty by reducing the fall-back and advance
levels from >75 and =<90 to 70 and 85 respectively (for instance).

It is all I have to say by now.

You should know that any comments, challenges and doubts are most
welcome here!

Wishing you well,

Ash

[Ashirgo]

And one more thing:

You may even consider disabling the fall-back.

For me it is a bit counter-productive and ruins my spirits, just as
the immediate feedback.

Ash

[Paul]

Thanks for posting this, Ashirgo. 

I wonder if increasing the number of cues per trial would have a beneficial effect. 20 cues is not much especially because once you reach 5-back there are only 4 "cycles" of cues, 10-back is 2 cycles, etc. This might also help to train for longer periods of unbroken concentration.

Regarding the technical ramifications of increasing the number of cues in BW, in Jaeggi mode the thresholds would have to be adjusted because it would still be limited to 4 sound, 4 audio and 2 simultaneous matches despite the increase in cues. Normal mode should scale nicely because the probability of a match is constant for each cue, so the thresholds would not need to be adjusted in this case.

Personally my favourite BW settings might be Normal mode but with all the aesthetic changes of Jaeggi mode (no text/feedback, larger squares, etc). It seems to me that Jaeggi mode, with its strictly defined number of matches, can be less random and more predictable than Normal mode. I didn't see anything in the study that supports the superiority of having a set number of matches instead of a fully randomized sequence. It might be better from a research perspective because it makes the scores easier to compare between various sessions and experimental subjects, but in a personal training context I find the extra randomness of Normal mode to be refreshing. In the next version of BW I'll make it easier to change the appearance of Normal mode to match that of Jaeggi mode.

Paul

[Ron Williams]

Sounds like good advice, especially 1,2&3

4 though seems a lot like 'chunking', which is a trick you use in some
of the usual memory schemes to *get around* the limited size of our
working memory. That's not what we're trying to do here.

The idea of making progression easier is one that I tentatively agree
with. I would place a caveat on this that we should only consider an N
level to have been assimilated when *no* mistakes are made, on, say, 9
out of 10 runs (like 1 or 2 back. We hardly ever fail to get 100% on
those except through gross inattention, right?)

[Ashirgo]

Paul, I agree with that the increase of number of clues would be
appropriate. At 11 back there is only one full cycle and 9/11 of the
last one - and it leaves me unfulfilled:)

As to the fourth advice, that is also right notion. But for me it is
still a necessity if I want to perform higher n-back levels. But in
time I can handle the whole n as a continuous part.

Well, the restriction of 100% of correctness required would be counter-
productive if you consider how annoying it would be for some folks -
at least in a long run.


Regards,

Ash

[putumayo]

Hey! I agree with Ron Williams - "chunking" data into small pieces is
good for your score in dual n back but not so much for your working
memory. It is well known that when you chunk the data into small
pieces (for example a telephone number) you can remember it more
easily. But then you don't use 7 places in your working memory (for a
7 digit number) but 3 or 2, so it is not what we want to do. That why
I like the "intuitive" approach (it is basically the 1 and 2 point).
When you do that you don't have to go far (10 - 12 n back) to
exerscise your working memory at its limit - it is sufficient to be on
5 n back, which is nearly 10 chunks of data in you WM.

Good luck!
putumayo

[putumayo]

I also think that the rehearsing is bad. I there's more: I think that
even having an image of the last stimuli is not the best approach,
because in the real life (for example in the studies) you don't hold
the last stimuli conciously in a form of image or anythig else - you
just let the information flow. That's my approach.

Best regards.
putumayo.

[chinmi04]

You're saying all this like it's a fact?

Unless I misunderstood, what you're saying is that the purpose of dual-
n-back training is to maximize the amount of 'working places' we can
actively use. Better to play intuitively at dual 5-back than reaching
10-back by chunking of information?

Then why are you playing dual-n-back at all? Because what it forces us
to do is essentially the chunking of sound and text into one digit. By
your reasoning you would prefer playing a version of single-n-back
(perhaps with alternating sound and text) than playing what you're
playing right now, namely dual-n-back.

[Ashirgo]

That is right, chinmi04:)

Putumayo, the underlying message of my guidelines is hidden in the
lack of word "intuitive", for I am sceptical about the intuitive
approach to its very essence.

In my opinion the whole dual n-back story is about how to consciously
turn a specific state of mind on; in which you are more focused and
generally benefit from all these things we connect with n-back. It is
a very conscious effort and undoubtedly (as such) must be controlled
by one's mind.

"Chunking" is a very natural technique, since many people actually use
it catching whole tel. numbers and so on. I would not say it will
reduce our expected benefits. In a sense it can be called "too
tricky", but it also could be argued that we should not use any active
tricks.

And, needless to say, these two parts of the exercise will become an
inalienable part of your thinking and processing of data so that you
will no longer notice them.

Please consider my stance.

Regards,

Ash

[VinceD]

Besides.. I could see chunking being bad if we stopped at say, N=5 or
something, but as long as we are continually pushing N to greater
numbers, we are taxing our brains, Chunking or not.

It would seem this pushes beyond any benifits from chunking.. eh?

[darkalex]

can you add some detailes in your conquering of n-back so i can
compare and relate as such:
1.time needed to you to reach 12 n back since first opening of BW.
2.Age
3.proffesion
4.iq
just curious
P.S. i am at 8 n-back myself ussing cell-stragedy with something like
glowing line with fast-rewinded sounds as precaution. took me 3 weeks
normal mode, 21yo;student-programmist/ 120-150

[Confuzedd]

About the chunking, I don't think it really matters. In my opinion,
just think of it as scaling; you'll just end up being able to chunk
even more groups. I.e. if you we can chunk 7 data piles at baseline,
then maybe having a ridiculous n-back level of 18 will be equivalent
to that +3 extra data piles you can work with.

[AshDog]

Why are so many people hung up on the "intuitive" method when Jaeggi
herself clarified what she meant by it (there is a post here with her
response in an email). The normal methods, including typical forms of
chunking, for holding all the information in our heads are perfectly
acceptable in her eyes.

[Ron Williams]

"The normal methods, including typical forms of
chunking, for holding all the information in our heads are perfectly
acceptable in her eyes."

Actually, she explicitly says that they wanted to avoid the use of
tricks such as are used by memory systems.

[Ashirgo]

9. If you feel tired during the exercise, it is not what should
happen. You should feel relaxed and refreshed due to the task. An
excessive tension or pressure in the regions of your brain, forehead,
stretched muscles are not welcome.

Ash

[Ashirgo]

I train rather regularly since 9th of November 2008. I am eighteen
years old. My profession? Starving student of literature in foreign
language(namely, English). IQ? I have no bloody idea, but is must be
around 130+.

Please, you must not underestimate the very first advice. The focus on
the task is crucial. Fluid intelligence is also called "the strength
of focus", therefore I conclude it is the first thing you must
consider performing the task. I realized that yesterday, having done
10-back with 90% result (which allowed me to jump, of course:D ) - it
was my extensive focus that gave me the chance to get higher.

8. Disabling the fall-back is also worth considering. I do not see why
should my well earned 11-back drop down because of my poor
performance, if I eventually improve as much as to advance up to
another level.

As to the number of cycles, indeed it would do better to ensure a
certain (minimal, fixed) level of them during the task. As I have said
somewhere, while doing 11-back, I have one full cycle and 9/11 of the
another and it is not fulfilling! :) Maybe in the future version, what
would you say, Paul?

[Paul]

Perhaps the best form of the option would be a simple True/False to enforce "whole" numbers of cycles. Enabling the option would change things as follows (assuming the default 20 cues per session is still in effect):

 2-back: 20 cues, 10 cycles

 3-back: 21 cues, 7 cycles

 4-back: 20 cues, 5 cycles

 5-back: 20 cues, 4 cycles

 6-back: 24 cues, 4 cycles

 7-back: 21 cues, 3 cycles

 8-back: 24 cues, 3 cycles

 9-back: 27 cues, 3 cycles

10-back: 20 cues, 2 cycles

11-back: 22 cues, 2 cycles

Paul

[Benjamano]

> 2. You had better not rehearse the last position/sound.

What exactly do you mean by rehearse?

A couple of days ago I came up with a technique which allowed me a
significant jump from "3-back with only a couple of successful 4-
backs" to "all successful 4-backs". Instead of relying on "pure
intuition" when a new stimulus presents itself (ie only paying
attention to the most-recent-stimulus) I decided to try to "linking"
the most-recent-stimulus to the one it was replacing - without paying
attention the items in between.

So for each stimulus I consciously do this (in order):
1. concentrate on the on the most-recent-stimulus - repeating its
sound and fixing its location
2. consciously recall the n-back-item - repeating its sound-and-
location
3. imagining a link 1 to 2
4. make my decision

This made 3-back harder for a couple of sessions but then suddenly
both 3-back and 4-back got much easier. Would you consider this
"rehearsing"? I think its valid to immediately repeat the most-recent-
stimulus in step 1 - and step 2 is discarded and of no use for the
next-stimulus. Previously I was trying to avoid "repeating and
sequencing" old-stimulus but it was slow going - my "intuition sense"
was not really kicking in. The most interesting thing with my new
technique is that when I lose focus half through a trial and can no
longer remember the "exact" details for step 2, my "intuition sense"
suddenly kicks in quite strong.

cheers, Ben

[Ashirgo]

Rehearse - to repeat. It is not necessary if you are focused enough.
All you need to do is to create an image of visual data and recording
of auditory data.

I can compare it to two following tasks:

1. You read sth silently and then try to repeat it aloud. You do not
have the chance to repeat it until the caught information is read
aloud. It is for visual part.

2. You listen to the audiobook and then try to repeat last sentence
from your memory.

Notice that normally we do not always need to repeat anything in our
minds to "catch" it and then write down or sth. The same goes with n-
back. It is still the the sufficient focus that makes this possible.

Paul, that pattern you presented seems very logical and thus may be
proper :)

Ash

[putumayo]

Hi! I agree with Ron Williams:

>"The normal methods, including typical forms of
>chunking, for holding all the information in our heads are perfectly
>acceptable in her eyes."

>Actually, she explicitly says that they wanted to avoid the use of
>tricks such as are used by memory systems.

But everyone has his approach and it's OK. I like mine because I know
that it works for me (I already benefited a lot from dual n back - I'm
much better in my studies and I'm studying mathematics so it's not
easy).

Stay with your way of playing dual n back if it's benefiting you.

Best regards.

[esar...@gmail.com]

I don't think that disabling the automatic fallback is a good idea.
Why? Because that way you can get used to the n-task at hand. You want
your brain to adapt to new situations, not to become used to doing 8-
back for instance.

[Ashirgo]

That is a good attitude, you know, putumayo.

As to esaramago, I also have some doubts. First of all because it
might be that I am getting used to doing a certain level (but, anyway,
still demanding!). The only thing that motivates me is the following
quote:

"10-back - 65%
10-back - 95%"

Needless to say, in one row. With fallback enabled it would have
degraded me one level down. What for, if another result is nearly
perfect?

Regards, Ash

[Ashirgo]

13-back reached today, after 34 trials in a row:)

Ash

[DarwinsBane]

Hi Ash,

You're clearly a n-back champ.

It would be instructive if you would take complete a round or two of the working memory tests over at http://www.cambridgebrainsciences.com/.

This would somewhat control for technique and help us assess whether your working memory is truly as phenomenal as your n-back results would indicate.

Cheers mate!

[whoisbambam]

Ashirgo does not come here often......I will be surprised if she
receives this per dnb forum........

Whatever test u use for her, she prefers SPEED.........i think she
achieved dnb14-15 or whatever with the settings at like 1second or
something crazy......

i think her brain would get very bored and distracted if she could not
control the speed to keep her absolutely concentrated on the task.

On Mar 26, 12:55 pm, DarwinsBane <matthewj.c...@gmail.com> wrote:
> Hi Ash,
>
> You're clearly a n-back champ.
>
> It would be instructive if you would take complete a round or two of the

> working memory tests over athttp://www.cambridgebrainsciences.com/.

[Aman Idle]

wow she a lady, i wanna marry her lol

--
You received this message because you are subscribed to the Google Groups "Dual N-Back, Brain Training & Intelligence" group.

To post to this group, send email to brain-t...@googlegroups.com.
To unsubscribe from this group, send email to brain-trainin...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/brain-training?hl=en.