Boring article which suggests combination N-Back is not utterly worthless.

[A.Four.Sigma]

http://mc.psychonomic-journals.org/content/36/8/1470.full.pdf+html
Abstract
" Performance on task switching, a paradigm commonly used to measure
executive function, has been shown to improve with practice. However,
no study has tested whether these benefits are specific to the tasks
learned or are transferable to new situations. We report evidence of
transferable improvement in a cued, randomly switching paradigm as
measured by mixing cost, but we report no consistent improvement for
switch cost. Improvement in mixing costs arises from a relative
reduction in time to perform both switch and nonswitch trials that
immediately follow switch trials, implicating the ability to recover
from unexpected switches as the source of improvement. These results
add to a growing number of studies demonstrating generalizable
improvement with training on executive processing."


Now I personally don't have to task-switch very often in my daily
life; in light of this, I hope that task switching improves overall
cognitive ability in some abstract way, as opposed to just helping one
be a speedy task switcher. However, by extrapolating the researcher's
findings, I surmise that training in task switching should improve
overall processing speed (based on the following sentence [which is
long enough to rival parts of the Federalist Papers]:

"Improvement in mixing costs arises from a relative reduction in time
to perform both switch and nonswitch trials that immediately follow
switch trials, implicating the ability to recover from unexpected
switches as the source of improvement.")

I interpret this as faster "processing" ability, which I would guess
is a very "general" cognitive function, with the implication that
task-switching could make one a faster thinker, overall. This is pure
conjecture, though.u


PS this seems to be the first mention of this journal on our
forum;it's worth perusal as it is free, and has some research that
hasn't been discussed here, such as WM possibly predicting individuals
abilities to solve the classic "9-dot" problem. It's most current
issue even talks about our beloved Stroop task (albeit not in
reference to its relation to Gf, wm, or any of our main areas of
interest)

[Gwern Branwen]

So... n-backing could improve processing/speed, but nothing about
solving harder problems. Not so strange, but good to have something
specific. I'm adding it to the FAQ in the section about the more
elaborate Brain Workshop modes.

--
gwern
http://www.gwern.net

[Pontus Granström]

To be fair it doesn't say anything about n-backing but rather task switching.

> --
> 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.
>
>

[A.Four.Sigma]

Pontus, not just n-backing, but combination mode specifically..

On Oct 27, 2:23 pm, Pontus Granström <lepon...@gmail.com> wrote:
> To be fair it doesn't say anything about n-backing but rather task switching.
>
> On Wed, Oct 27, 2010 at 9:15 PM, Gwern Branwen <gwe...@gmail.com> wrote:

[Samples]

It seems rather couterintuitive to suggest that single n-back would
not help with solving harder problems, but rather just helps with
processing speed. Even if one were to just that single entity, that
you are proposing, into consideration.

"n-backing could improve processing/speed, but nothing about
solving harder problems. "

Do you think increasing someones processing speed would lead to a
decrease or increase in there problem solving ability? I tend to lean
towards the latter.


The higher single n-back you can achieve, without some artificial or
material strategy (as suggested by Jaeggi, 2010), the higher one's
problem solving abilities (fluid/reasoning abilities, g, gf) are. The
same goes for dual-n-back but not to the same degree. I'm specifically
focusing on single n-back to avoid the task-switching hypothesis that
may applied if I were to strictly talk about dual. In Jaeggi's
study single n-back was a greater
predictor of performance on both IQ test (from recollection I
think it was the BOMAT and RAPM) than dual n-back.


Single-n-back = seems to have nothing to do with working memory
capacity (Jaeggi 2010)
Dual-n-back = only partially connected to working memory
capacity. Virtually no correlation between performance on dual (and
single) n-back and OSPAN.

Other brain workshop variants = Don't think anyone here can
confidently suggest explanations. No evidence. One can only use
logical intuition, paradoxical I know, in regards to this
determination.

Crab = It seems highly suggestive that crab is any more
correlated than dualn-back to working memory capacity.

combination = seems logical that theres a lot of task switching
involved, which relates to A4S study I guess.

Variable = Less confident about it's role. Would seem logical that it
involves more working memory capabilities than dual and crab


Extra modalities = So convoluted, one could argue either way. But it
does involve a lot more task switching, or monitorization of
information.


These are just assumptions. We really can't confidently suggest their
roles. Considering Jaeggi herself made the admission that
she does not fully understand the possible role that dual-n-
back plays.

Who are we to think we can come up with any better connections???




Sameil Plairz





On Oct 28, 6:15 am, Gwern Branwen <gwe...@gmail.com> wrote:

> gwernhttp://www.gwern.net- Hide quoted text -
>
> - Show quoted text -

[Thomasthetankengine]

Gwern,

Some questions about your post.

How is this study related to n-back?

How did you deduce that as a result of this study one can assume that
n-back improves processing speed?

How did you come to the conclusion that n-back does not increase a
persons ability to solve complex problems?

How is at all "specific" information?


Please do your best to answer these questions, I'm so confused atm. A
lot of conflicting information. It's like trying to interpret the
truths and untruths that may be written in a bible.


Much obliged!

On Oct 28, 6:15 am, Gwern Branwen <gwe...@gmail.com> wrote:

[Pontus Granström]

From the file the 20-minutes version of the rapm

Could the timed version of the APM additionally involve a speed factor as well?
Our results suggest this might not be the case. The correlation between the 20-minute
timed version of the APM and the score on the six tests did not differ from the correlation
between the untimed APM score at completion and the score on the six tests.

Administering the APM as a speed test, even with a limit of 20 minutes, is better
than administering a selection of 12 items from the APM, even if this subset represents
the APM quite well in a psychometric sense, because the selection also represents a
quite different task for the participant, leaving the intermittent items out.

Might be worth adding to the FAQ!

[A.Four.Sigma]

Sorry Pontus, the above correction was pointed at Gwern. I make typos
when tired. ThomasTheTankEngine: This posting was meant to show that
combination N back may improve "reaction time", which I hazard to
guess is directly linked to processing speed. This has nothing to do
with "single n back" or apparently anything else that's been posted as
a response to the article. Needless to say, this thread has gone off
on several tangents. To summarize, it mainly says that if you want to
think faster, Combination N back MAY allow you to do that.

On Oct 28, 4:57 am, Pontus Granström <lepon...@gmail.com> wrote:
> From the file the 20-minutes version of the rapm
>
> Could the timed version of the APM additionally involve a speed factor as
> well?
> Our results suggest this might not be the case. The correlation between the
> 20-minute
> timed version of the APM and the score on the six tests did not differ from
> the correlation
> between the untimed APM score at completion and the score on the six tests.
>
> Administering the APM as a speed test, even with a limit of 20 minutes, is
> better
> than administering a selection of 12 items from the APM, even if this subset
> represents
> the APM quite well in a psychometric sense, because the selection also
> represents a
> quite different task for the participant, leaving the intermittent items
> out.
>
> Might be worth adding to the FAQ!
>
> On Thu, Oct 28, 2010 at 11:44 AM, Thomasthetankengine <
>

> > brain-trainin...@googlegroups.com<brain-training%2Bunsu...@googlegroups.com>
> > .

[Gwern Branwen]

On Wed, Oct 27, 2010 at 8:26 PM, Samples <xallge...@hotmail.com> wrote:
> It seems rather couterintuitive to suggest that single n-back would
> not help with solving harder problems, but rather just helps with
> processing speed. Even if one were to just that single entity, that
> you are proposing, into consideration.
>
> "n-backing could improve processing/speed, but nothing about
> solving harder problems. "
>
> Do you think increasing someones processing speed would lead to a
> decrease or increase in there problem solving ability? I tend to lean
> towards the latter.

My computer science background means that I find it perfectly
straightforward that speed can be unconnected to what class of
problems one can solve.

For example, take Turing machines. The class of problems a universal
Turing machine can solve if it doesn't have unbounded memory is set by
how much memory it *does* have. If its tape is X units long, then any
problem which involves X+1 markings is forever beyond that Turing
machine. It simply can't do it. And any problem involving X or X-1
markings is forever solvable by that Turing machine. It doesn't matter
whether the Turing machine executes operations at 1 terahertz or 0.1
hertz. It either can solve the problem or cannot.

Or to put it a different way, it's like someone who wants to parse
Wikipedia's full history for some statistical project. 'Yes', they
say, 'I realize that the tarball alone fills up 100% of my hard drive
and RAM, but I just bought a new processor which 20% faster!' Doesn't
matter.

More abstractly is the idea of complexity and constant factors (see
http://en.wikipedia.org/wiki/Big_O_notation ); in this analogy,
n-backing improves one's speed (reduces the constant factor) and thus
one solves problems faster, but the performance still degrades the
same as the size/difficulty of the problem goes to infinity.

> The higher single n-back you can achieve, without some artificial or
> material strategy (as suggested by Jaeggi, 2010), the higher one's
> problem solving abilities (fluid/reasoning abilities, g, gf) are. The
> same goes for dual-n-back but not to the same degree. I'm specifically
> focusing on single n-back to avoid the task-switching hypothesis that
> may applied if I were to strictly talk about dual. In Jaeggi's
> study       single       n-back was a        greater
> predictor     of performance on both IQ test (from recollection I
> think it was the BOMAT and RAPM) than         dual        n-back.

Yes, I still find the SNB result very odd and am not sure what to make
of it. Doesn't it seem as if single n-back, with no interfering
modalities or choice to make between which set to think about, ought
to be even more of a WM task than DNB? Yet, there you have it.

> These are just assumptions. We really can't confidently suggest their roles. Considering Jaeggi herself made the admission that she does not fully understand the possible role that dual-n-back plays.

> Who are we to think we can come up with any better connections???

This would seem to be an argument for agnosticism, skepticism, and
refraining from an extreme confidence that DNB increases IQ and the
results are not explainable by any other mechanism.

--
gwern
http://www.gwern.net

[Pontus Granström]

You cannot compare the human brain with a turing machine. As I also got a degree in computer science I realize it's limitations. As you know a human brain easily can solve a problem that a TM never can. Given the complexity
of the human brain and it's processing, it seems quite simplistic to only look on "face value" things like "n-back is just moving a square" or "remembering stuff".RAPM does not require this hence it cannot improve IQ, just speed, because speed is a neurological construct that is not so "dangerous to improve". Al though speed is quite heavily dependent on things like myelin and so on.  There might be more issues to the problem and given the amazing complexity of our brain it only seems naive to give oversimplified abstract-models-of-computer-programs-kind-of-explanations to this.

For example, running activates an area that improves stroop performance. Why would moving my legs at a rapid rate has anything to do with my cognitive functioning? Why would be so stupid to believe that? Another example: when
brain scanning cats that learns simple recognition tasks they found that neurons are used for multiple purposes, because this simple tasks required 10% of it's neurons. If we had neurons that were isolated to do one task we couldn't learn much.
This was also what Jaeggi suspected that n-back activates neural networks shared by G. My point is this, there might be so many complex and subtle issues that only can be understood if we understand the brain, which no one really does. Therefore a open mind might be recommended.
 

--
gwern
http://www.gwern.net

--

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.

[Samples]

Gwern,

On some points I empathise with your analogies, however, people are
not computers or computers are not people
(at least not yet).

"More abstractly is the idea of complexity and constant factors (see
http://en.wikipedia.org/wiki/Big_O_notation ); in this analogy,
n-backing improves one's speed (reduces the constant factor) and thus
one solves problems faster, but the performance still degrades the
same as the size/difficulty of the problem goes to infinity."

What I agree with:

- As one's processing speed increases so does the rate at which they
can complete
problems that are in their reasoning capabilities. However, as the
complexity of
each problem increases, the rate at which they complete the problems
decreases to the point where speed eventually becomes negligible,
because the problems are now beyond their reasoning capacities.


What I do not agree with:

- That n-back increases processing speed. There is no evidence for
this, only assumptions.
- That n-back does _not_ increase one's ability to solve problems that
are of
greater complexity. There is no evidence that it doesn't, however,
there is room to suggest that it does. The improved scores in Matrix
reasoning tests can vouch for this. As to the significance of the
evidence, that is currently an unanswered question (but I don't think
we'll have to wait for - too long).

--> 4 examples (there could be more or future evidence of less- who
knows) of some effects
of n-back that may have led to increased scores.

1. Increased processing speed = leads to improvenment in RAPM test
scores because they can solve more problems that are of similar (or
within their reasoning capacity) complexity in a fixed time.
2. Increased attention span = same consequence as above because they
can focus their attention better and for longer periods.
3. Increased recognition abilities = same consequence because they can
make connections faster and more accurately.
4. All of the above

In relation to point 2 and 3, specifically, this would mean that they
are able to solve the tougher problems on Matrix test.


Bear in mind, this is all said under the assumption that the time in
which a participant is given to complete an 'intelligence' test
effects the accuracy to which an assessment of intelligence can be
made. Meaning, less time = less accurate measure.

However, recent knowledge has been brought to my attention (per Pontus
Granstom) that suggests otherwise. I am yet to read the article.

See link below:
http://brain-training.googlegroups.com/web/20min+version+rapm.pdf?gda=eOWvkUgAAAAXk78ja_2pBjj4akcEzk5FAy3xjOA9TMWM8QRPXdiVhCqPwt78wthZDJJad1zHLnCmwwmmjY8lLEkm5GsdcWpfGjVgdwNi-BwrUzBGT2hOzg

"Yes, I still find the SNB result very odd and am not sure what to
make
of it. Doesn't it seem as if single n-back, with no interfering
modalities or choice to make between which set to think about, ought
to be even more of a WM task than DNB? Yet, there you have it."

It is odd, as is also exclaimed by Jaeggi. But...

RAPM measures non-verbal reasoning. If this is the case and it is not
a strong predictor of verbal reasoning then it understandable that
they found this result.

If they used 2 position modalities instead of 1 position and 1
auditory modality, would they of obtained less obscure findings in
relation to dual v. single n-back?

I am leaning towards yes.

"This would seem to be an argument for agnosticism, skepticism, and
refraining from an extreme confidence that DNB increases IQ and the
results are not explainable by any other mechanism."

It's inappropriate to describe the role that n-back plays on certain
components of human intelligence, when researchers do not yet fully
understand the brain and its ability to manifest intelligence. We
don't know whether long term use of n-back will produce the same
effects
either.

On another note, not that I care about the the following argument (I
don't),
it's actually similar to the God no God scenario. Some people like to
postulate the existence of something that is unknown when what is
known or obeserved is still not fully understood or is misunderstood.



Sameil Pleirz



On Oct 30, 10:32 am, Gwern Branwen <gwe...@gmail.com> wrote:

> More abstractly is the idea of complexity and constant factors (seehttp://en.wikipedia.org/wiki/Big_O_notation); in this analogy,

[Pontus Granström]

Big O-notation? You mean that the person who scores 14 correct answers and the person scoring 7 are just as intelligent the only difference lies in their processing speed? Where as the lowest scorer went from problem 14 to problem 18, the highest scorer went from problem 28 to 36, but these problems are also harder and takes more time. It seems highly unlikely that we all are born with suboptimal "g-speed". Besides there's no support what so ever that speed is a
factor in timed RAPM. I wonder why you do not add this to your FAQ as counterweight to Moody's claims. You could for example write:

"Even though there might be tempting to suggest that speed plays a dominant role in timed RAPM there's no evidence to suggest that this is the case based on the correlation with other tests. That is, if speed was a constant that we added to the performance we would see a lower correlation with other tests. Due to lowering the average correlation (ref: 20 min rapm)"

"There's also quite substantial research on timed procedures beyond the reference given in the first study".

[Pontus Granström]

You could also add:"
In contrary to what Moody suggest, there's no data on that RAPM problems are monotonically harder, rather it measures accuracy over 4-5 levels of difficulty. It might be that n-back increase our accuracy and thereby the total amount of correct answers, hence by the definition that more correct answers makes us more intelligent, we have become so."

[A.Four.Sigma]

This whole debate sounds foolish, without any real points being made.
Don't confuse yourselves: _None of this_ is talking about DNB per se,
it's TASK SWITCHING, which we in the group surmise is trained by
COMBINATION N BACK, and if it affects PROCESSING SPEED. The question
posed by this thread is does _CNB_ improve processing speed.

I don't think that processing speed improves non-verbal reasoning
(RAPM). My processing speed is extremely high (on Lumosity, I quickly
went above the 99th percentile in ability to quickly perform
arithmetic operations. Yet my non-verbal reasoning ability is slightly
above average. But my verbal score on WISC IQ for children was just
under the fourth sigma (std dev). I have always been an extremely fast
thinker, as Lumosity shows, much faster, I predict, than most anyone
I've met. Yet I perform abominably when it comes to putting together
tangrams or complex puzzles and matrices, which is why I play DNB.
Processing speed doesn't improve non-verbal reasoning; in my
experience it has allowed me the ability to fail several times at the
mental rotations in a time horizon that would allow a normal person to
fail once. Id est, being UNABLE to see any pattern whatever is not
improved by thinking faster. You simply are able to think faster about
the fact that you can't see any pattern.

On Oct 30, 9:34 am, Pontus Granström <lepon...@gmail.com> wrote:
> You could also add:"
> In contrary to what Moody suggest, there's no data on that RAPM problems are
> monotonically harder, rather it measures accuracy over 4-5 levels of
> difficulty. It might be that n-back increase our accuracy and thereby the
> total amount of correct answers, hence by the definition that more correct
> answers makes us more intelligent, we have become so."
>

> On Sat, Oct 30, 2010 at 4:26 PM, Pontus Granström <lepon...@gmail.com>wrote:
>
> > Big O-notation? You mean that the person who scores 14 correct answers and
> > the person scoring 7 are just as intelligent the only difference lies in
> > their processing speed? Where as the lowest scorer went from problem 14 to
> > problem 18, the highest scorer went from problem 28 to 36, but these
> > problems are also harder and takes more time. It seems highly unlikely that
> > we all are born with suboptimal "g-speed". Besides there's no support what
> > so ever that speed is a
> > factor in timed RAPM. I wonder why you do not add this to your FAQ as
> > counterweight to Moody's claims. You could for example write:
>
> > "Even though there might be tempting to suggest that speed plays a dominant
> > role in timed RAPM there's no evidence to suggest that this is the case
> > based on the correlation with other tests. That is, if speed was a constant
> > that we added to the performance we would see a lower correlation with other
> > tests. Due to lowering the average correlation (ref: 20 min rapm)"
>
> > "There's also quite substantial research on timed procedures beyond the
> > reference given in the first study".
>

> > On Sat, Oct 30, 2010 at 7:44 AM, Samples <xallgeared...@hotmail.com>wrote:
>
> >> Gwern,
>
> >> On some points I empathise with your analogies, however, people are
> >> not computers or computers are not people
> >> (at least not yet).
>
> >> "More abstractly is the idea of complexity and constant factors (see

> >>http://en.wikipedia.org/wiki/Big_O_notation); in this analogy,
> >> n-backing improves one's speed (reduces the constant factor) and thus
> >> one solves problems faster, but the performance still degrades the
> >> same as the size/difficulty of the problem goes to infinity."
>

> >>http://brain-training.googlegroups.com/web/20min+version+rapm.pdf?gda...

> ...
>
> read more »

[Pontus Granström]

One thing that struck me is why we never have discussed why RAPM-performance would transfer to performance in other areas. Why problems that takes 1 minutes to solve should transfer
into problems that takes hours,days or years. Isn't this a "speed test" relative to such problems? I also had an analogy that people who scores well on high school exams in math tend to score well
on university exams, why is this?

[moe]

"But my verbal score on WISC IQ for children was just
under the fourth sigma (std dev)."

Wow!! you're quite the monster. Awesome score on the arithmetic game
also. I'm not too good at lumosity's arithmetic despite making it to
PASAT 8 a few times and successfully multiplying 6 digit numbers in my
head on a few occasions (ie 387,287 x 987,486). I would give anything
to think faster. Whenever I'm solving a problem it's like the
information sits in my mind for a good period of time before being
processed as if I have some sort of severe white matter deficit or
something (maybe I should exercise more iunno).

Any suggestions on improving processing speed other than using n-back
variants (n-back doesn't work for me)? Has training on task switching
tasks been successful for anyone?

> > >> (seehttp://en.wikipedia.org/wiki/Big_O_notation);inthis analogy,

> > >> > n-backing improves one's speed (reduces the constant factor) and thus
> > >> > one solves problems faster, but the performance still degrades the
> > >> > same as the size/difficulty of the problem goes to infinity.
>
> > >> > > The higher single n-back you can achieve, without some artificial or
> > >> > > material strategy (as suggested by Jaeggi,
>

> ...
>
> read more »

[A.Four.Sigma]

Interesting you should bring that up;
In college I trained 2 times per day, tons of lifting, and tons of
cardio, including teaching myself to swim and play water polo. I doubt
that more than a handful of Lumosity users trained as much as I or
were as healthy. This probably explains at least 30% of why I easily
surpassed all but a few of their scores. If you undertake such a
training routine, you will probably improve processing speed.

Secondly, I also notice the logical issues you talk about. It is as if
my thinking just stops. It's quick at most things, but when it comes
to matrices or figuring out how much everyone should pay at the dinner
table, all of a sudden I go dumb and need complete silence for a
minute or so to calculate the proper answer. This is more of a WM
issue than processing speed, and pattern recognition, I think. I
experience the same issues, even with high processing speed. Maybe the
issue is less speed than the number of "cores" one has, for example.
At any rate, I think combination N back creates the same feeling of
dumbness that I get in these situations, so I assume that playing it
might help attentuate my dumbness at processing certain brands of
abstract logic. (I guess that means processing speed, but again, I
think there are different processors in the brain for different types
of problems. Extreme speed in one domain does not mean extremity of
speed in all; this has been a classic problem I have struggled with
(How can I be so quick at everything EXCLUDING nonverbal reasoning/
matrices). This is one of the reasons I've embarked on studying DNB.

This was more or less a free-write on the topic. Hopefully there are a
few pieces of good information to be found in it.

> > > >>http://en.wikipedia.org/wiki/Big_O_notation);inthis analogy,
> > > >> n-backing improves one's speed (reduces the constant factor) and thus
> > > >> one solves problems faster, but the performance still degrades the
> > > >> same as the size/difficulty of the problem goes to infinity."
>

> ...
>
> read more »- Hide quoted text -

[Dom]

Hello fellow Brainworkshop users,

This is a quote made by Samples. Does anyone have any thoughts on
this? I'm really bad at non-verbal reasoning. Trying to to come up
with the best way to train. I understand that no one can provide a
clearly defined direction that is backed up by evidence on this, but
it would be really helpful if I could get someones opinion on this.

"RAPM measures non-verbal reasoning. If this is the case and it is
not
a strong predictor of verbal reasoning then it understandable that
they found this result. "


"If they used 2 position modalities instead of 1 position and 1
auditory modality, would they of obtained less obscure findings in
relation to dual v. single n-back?"

Thanks everyone,

Dom

On Oct 30, 4:44 pm, Samples <xallgeared...@hotmail.com> wrote:
> Gwern,
>
> On some points I empathise with your analogies, however, people are
> not computers or computers are not people
> (at least not yet).
>

> "More abstractly is the idea of complexity and constant factors (seehttp://en.wikipedia.org/wiki/Big_O_notation); in this analogy,
> n-backing improves one's speed (reduces the constant factor) and thus
> one solves problems faster, but the performance still degrades the
> same as the size/difficulty of the problem goes to infinity."
>

> See link below:http://brain-training.googlegroups.com/web/20min+version+rapm.pdf?gda...

[A.Four.Sigma]

Obscure? I thought it was conclusive that there was apparently no
difference between the two. Either one should tax wm to the maximum,
since a user would reach higher n theoretically, in Snb, to offset the
absence of audio stimuli. Any type of Nb should improve nonverbal
skill, according to this. Just be consistent with training and notice
the benefits for yourself.

> > > More abstractly is the idea of complexity and constant factors (seehttp://en.wikipedia.org/wiki/Big_O_notation);inthis analogy,

> > > n-backing improves one's speed (reduces the constant factor) and thus
> > > one solves problems faster, but the performance still degrades the
> > > same as the size/difficulty of the problem goes to infinity.
>
> > > > The higher single n-back you can achieve, without some artificial or
> > > > material strategy (as suggested by Jaeggi, 2010), the higher one's
> > > > problem solving abilities (fluid/reasoning abilities, g, gf) are. The
> > > > same goes for dual-n-back but not to the same degree. I'm specifically
> > > > focusing on single n-back to avoid the task-switching hypothesis that
> > > > may applied if I were to strictly talk about dual. In Jaeggi's
> > > > study       single       n-back was a        greater
> > > > predictor     of performance on both IQ test (from recollection I
> > > > think it was the BOMAT and RAPM) than         dual        n-back.
>
> > > Yes, I still find the SNB result very odd and am not sure what to make
> > > of it. Doesn't it seem as if single n-back, with no interfering
> > > modalities or choice to make between which set to think about, ought
> > > to be even more of a WM task than DNB? Yet, there you have it.
>
> > > > These are just assumptions. We really can't confidently suggest their roles. Considering Jaeggi herself made the        admission      that she does not fully  understand the possible role that       dual-n-back     plays.
> > > > Who are we to think we can come up with any better connections???
>
> > > This would seem to be an argument for agnosticism, skepticism, and
> > > refraining from an extreme confidence that DNB increases IQ and the
> > > results are not explainable by any other mechanism.
>
> > > --

> > > gwernhttp://www.gwern.net-Hide quoted text -

[Dom]

Thanks A4S,

But what's your opinion on this:

"If they used 2 position modalities instead of 1 position and 1
auditory modality, would they of obtained less obscure findings in
relation to dual v. single n-back?"

Thx again...

> > > > gwernhttp://www.gwern.net-Hidequoted text -
>
> > > - Show quoted text -- Hide quoted text -

[A.Four.Sigma]

As stated above, I don't understand your question since I'm not aware
that there is anything obscure at all about the findings. The results
are plainly stated: that there seems no real difference between SNB
and Dual NB. If you're asking whether the findings would have showed
DNB to be superior if twin visual stimuli were used, I obviously can't
answer that due to a lack of data, but to guess, I would apply
identical logic to that of my previous post: the difference between
single and multiple stimuli is adjusted for by the max n-level
attainable in each mode. Less stimuli = higher max n, so theoretically
an identically difficult challenge, ceteris paribus.

Second, while DNB yields identical, non-verbal, RAPM improvements, it
was not tested if this improved any other abilities as far as I
know.It could provide benefits not yet tested for. The only way to
know is to do fMRI of players using the different modes and see what
the difference is

In fact (eureka moment ) I think we should propose that Jaeggi do a
new study, utilizing each of the BW modes and testing how they each
affect the brains/performance of individuals. If they are all
identical, it would be very interesting, but that is not likely to be
the case, except apparently for SNB and DNB. Unexplored frontiers....

> ...
>
> read more »

[Dom]

A4S,

If you think there are no real differences between n-back and dual-n-
back what do you make of the following findings in the Jaeggi (2010)
study? I don't mean to be rude or anything I'm just trying to paint my
self the clearest picture. Thanks for your time.

"Despite the apparent process overlap in single and dual-n-back
performance, we still observed differential cognitive processes
mediating performance in single or dual n-back tasks: whereas single n-
back performance was mostly predicted by matrix reasoning, dual-task
performance was mediated by working memory capacity in addition to
Gf."

= Difference 1

"Further, the single n-back task was the only predictor for both
matrix reasoning measures."

= Difference 2


"Considering the rationale that transfer is more likely to happen for
tasks that share considerable variance, we can conclude that training
on both single and dual n-back tasks should yield transfer to matrix
reasoning, but that transfer to workingmemory capacity should be less
likely, especially in the case of single n-back training."

= Conclusion on difference 1



"Considering the variance explained in the matrix reasoning tasks, our
data suggest that the single n-back task might be an even better
training vehicle than the dual n-back task."

= Conclusion on difference 2

> ...
>
> read more »

[Pontus Granström]

What we learn from this is that we should be careful in what tasks we lump together or arbitrary consider equal. There's more to the puzzle than meets the eye.

> ...
>
> read more »

[A.Four.Sigma]

Dom,

Thanks for the clarification of your question; I didn't see those
parts of the article. I was more or less parroting what I have read
from time to time on the board, being that there is not a significant
difference between the two. My new thoughts on this are that this
seems
intuitive. SNB forces one to recall many visual stimuli, thus it seems
it would take strong visual spatial ability/matrix skill, especially
at high N.

DNB is a mixture of modalities, but in reducing the overall visual
intensity to make room for audio, it seems to detract from some of the
matrix skill development.

Given this conclusion, I would hazard to infer that a person who did
Multi-Stim DNB (2 visual stimuli, no audio) would see greater benefits
on matrices than someone doing Audio/Visual DNB. From the conclusions
above, it seems that the visual component of the Nback task is more
important for improving matrix reasoning than audio or otherwise.

Next implication: Color: position and image seem to be key to matrix
development. In the modes that include color, I'm not sure that color
improves matrix reasoning because it is not quite a recognition of
spatial patterns the way position and image are.

Finally, I am glad to have seen this data, because it seems to suggest
that more stimuli may be more taxing on 1) Gf, 2)WM. This seems to
vindicate those of us who have been doing QNB and PNB. I think most
people could care less about scoring higher on matrices, instead
wanting to become more intelligent overall, so DNB would be the better
choice (or even QNB/PNB).

Caveat Emptor: Saying that more stimuli is more taxing on WM/GF may be
spurious, a study on tripleNback and quadNback is necessary to prove
that WM/GF taxation doesn't go down as one adds more than 2 stimuli
(i.e. DNB==>TNB==>QNB etc..)

I personally would be interested in training matrices ability for IQ
test purposes, so I will probably spend some time doing 1)DNB no audio
(aka MultiStim), and SNB.

Second Caveat Emptor: A major gray area caused by this study is what
is the difference between Gf (DNB training) and matrix reasoning ( SNB
training), since matrix reasoning is supposed to predict GF, how and
why can/does Jaeggi separate Gf from Matrix reasoning?

A4S

> ...
>
> read more »

[A.Four.Sigma]

Dom,

Hopefully this helps (I posted it in another thread here):
" I found the following article, which says something very
interesting:
http://learningdisabilities.about.com/od/glossar1/g/nonverbalintell.htm

It says nonverbal IQ is a function of ones ability to remember visual
sequences of information. That sounds exactly like single N back to
me. This definitely provides insight as to why SNB is more effective
at improving matrix ability than DNB."

As I hypothesized above, longer strings of N visual stimuli should tax
nonverbal intelligence processes more. Apparently audio input does not
tax non-verbal systems (see following article):
http://learningdisabilities.about.com/od/glossar1/g/verbalintellige.htm

It says that ability to recall auditory information is a function of
verbal intelligence!


I think that this information is fundamental to what we are trying to
understand by Nbacking: More visual stimuli, IE multi stim Nback, will
tax non verbal IQ more. Audio stimuli will rely on verbal processing
ability. Armed with this information, one can model their BW training
program to target the area they wish to improve most.


However, we have to keep in mind we dont understand the impact of Crab
or Variable N back. Combination Nback is less of an unknown to us: the
article this thread is based on vaguely seems to say it improves
"executive function". I predict that the combination of these modes
with other NB modes amplifoes improvements in verbal and nonverbal
intelligence insofar as they increase the complexity of verbal and
nonverbal patterns that must be recognized during the course of
training.

> ...
>
> read more »- Hide quoted text -