DNB progress so far

[christopher lines]

This is kind of a selfish post (arnt they all?)

Anyways I just made it up to level 8 today and here is my chart of
progress

DNB progess chart.JPG


I have been using basic mode for couple of months, only change was
increasing the start level to 4 a couple of weeks ago. If I go to
level 8 a few more times then I expect I will move the start level to
5.

I began by doing maybe 50 sessions a day but now I do 23 (thats
intended to give me 20 sessions at my peak level)

I did a couple of the online IQ tests after about 10 days (scored 126
in one of them and 106 in another; I repeated the same tests about a
month later ( about 1 month ago) and scored (133 and 109). I have no
idea why the tests gave such big differences in scores but I
definately think its easier the second time you do the tests because I
remembered the strategies for solving the problems which took some
time to figure out when I first did the tests. I am kind of against
keep re-doing the tests because of learning effects and a bit truobled
that different test produce such different results.

So how does it compare with other progress charts?

Should my IQ by tracking up with the DNB score?

Overall I do "feel" fairly good cognitively but I cant say I have any
evidence of mental improvement subjective or otherwise. I intend to
keep using the DNB game for at least the next month or so until either
I feel there is no benefit or I can be sure there is.

[Gwern Branwen]

On Wed, Dec 2, 2009 at 8:44 PM, christopher lines
<christop...@gmail.com> wrote:
> I did a couple of the online IQ tests after about 10 days (scored 126
> in one of them and 106 in another; I repeated the same tests about a
> month later (  about 1 month ago) and scored (133 and 109). I have no
> idea why the tests gave such big differences in scores but I
> definately think its easier the second time you do the tests because I
> remembered the strategies for solving the problems which took some
> time to figure out when I first did the tests. I am kind of against
> keep re-doing the tests because of learning effects and a bit truobled
> that different test produce such different results.

Tests aren't supposed to compare with each other, but with themselves;
so it doesn't matter that one says 130 and the other 110. The
interesting thing is that each was similar to its previous test but
higher. I would chalk a fair bit of it up to learning, yes. Why I
suggest leaving multiple months between tests if the test doesn't
randomize itself or otherwise change.

> So how does it compare with other progress charts?
>
> Should my IQ by tracking up with the DNB score?

We dunno. There's some sort of linear correlation in Jaeggi's study
between time spent and IQ increases, but that's not DNB score.

--
gwern

[Pontus Granström]

Still you seemed to have gained apporximatley 6 points on both tests. That you cannot compare tests with each other is both true and not true. If the test measure the same thing and have the same amount of load on Fluid intelligence then you can indeed compare them if you know on which group it was normed. Having a 110 IQ in a above average group gives you roughly an IQ of 120.  It is important to note that Jaeggi only claim to raise the Gf and not verbal nor mathematics or similar, even if they usually correlate they technically doesnt have too.

What kind of IQ-tests were they just curious!

--

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.

[christopher lines]

The tesst I did were the first two off the faq MEASURING iq

iqout.com
iqtest.dk


the first-iqout test was giving 106/109 and the second 126/133.

Incidently am now getting malware warning from the first site.

> > brain-trainin...@googlegroups.com<brain-training%2Bunsubscribe@go oglegroups.com>
> > .

[christopher lines]

When I did the IQ tests for the second time I could remember the
strategies for solving the problems which I had to create when I first
did the tests.

Partly for this reason I dont credit much the improvement in scores to
DNB. Maybe it is a cause.

I find this whole area a bit unscientific. There is no theory being
tested so effectively your looking for correlation but thats something
different.

I am not the only one to spot the lack of theories in this
area....anyways a link to a very interesting video lecture I found on
cognitivefun.net

http://www.youtube.com/watch?v=cCdbZqI1r7I


On Dec 3, 6:24 pm, christopher lines <christopherli...@gmail.com>
wrote:

[christopher lines]

I suspect I was hoping someone would comment on the graph, is it
normal or too regular or is the rising DNB score besides the point?

I think we have covered this before that DNB score is secondary to
amount of time spent exercising at your differential level. In which
case maybe the DNB program can also plot the number of sessions on the
graph as well?

Is there a correlation between DNB score and IQ test score?

On Dec 3, 8:31 pm, christopher lines <christopherli...@gmail.com>

[Reece]

There would probably be a much stronger correlation between DNB/IQ
test score if we were to give everyone an IQ test after they had
trained for the same "X" number of hours with DNB so as to better
control for familiarity with the task.

[wzeller]

Wouldn't the strongest correlation between IQ and DNB be found if the
DNB scores were taken from subjects who had never trained at all?
That would control for motivation, training effect, etc., and better
approximate their native WM capacity. This seems analogous to the
connection between IQ and chess ability. Although a person with a
higher IQ would be expected to play a better first game of chess than
a person with a lower IQ, if the latter practiced and the former did
not, the correlation between IQ and chess ability would soon be
overcome by experience.

[Pheonoxia]

Right now I'm training DNB w/ 1.0 second trials. Every so often I get
to 7-back, but usually can't beat 6-back.

[Reece]

Yes, I believe the strongest correlation would likely be between
performance at DNB with no previous training and IQ test score,
although I have sadly been unable to come across any such DNB/IQ
research.

With working memory capacity tied to so many other factors that one
may argue influence both initial and future DNB performance (and are
strongly correlated with IQ) such as motivation, focus, concentration,
etc, it may prove difficult (if not impossible) to adequately control
for said variables regardless of whether we test on Day 1 or Day 100.
For people with ADHD, it wouldn't surprise me if there was a stronger
correlation between DNB performance and IQ after many days/weeks of
training.

One problem with the chess example is that there are many ways to get
better at chess that require nothing in the way of gains in either
working memory capacity or intelligence (eg. mindlessly memorizing
certain configurations and strategies). If one controls for chunking
by drastically decreasing the speed of trials, there does not seem to
be much in the way of strategies which could be applied to DNB -- this
is especially true for more difficult variants such as QNB and PNB
which are incredibly difficult with < 1s intervals even at very low n-
back levels.

It would be interesting to see how chess grandmasters would fare if we
were to randomly generate where each chess piece appeared at the
beginning of the game so as to better control for strategies. I'm sure
they'd still win against an average player the overwhelming majority
of the time, however we might see much less of a difference at higher
skill levels and perhaps some factors such as IQ playing a larger role
beyond a certain number of hours spent learning to play.

[wzeller]

One problem with the chess example is that there are many ways to get
better at chess that require nothing in the way of gains in either
working memory capacity or intelligence (eg. mindlessly memorizing
certain configurations and strategies). If one controls for chunking
by drastically decreasing the speed of trials, there does not seem to
be much in the way of strategies which could be applied to DNB -- this
is especially true for more difficult variants such as QNB and PNB
which are incredibly difficult with < 1s intervals even at very low n-
back levels.

The reason that I brought up the example of chess skill acquisition was actually because practice quickly overwhelms native ability.  I wanted to make the point that the association between mental tasks and IQ is higher before the person begins practicing.  The only difference between DNB and chess is that, perhaps, the skill DNB hones is itself WM or fluid intelligence, whereas chess skill may not be so generalizable.  

It would be interesting to see how chess grandmasters would fare if we
were to randomly generate where each chess piece appeared at the
beginning of the game so as to better control for strategies. I'm sure
they'd still win against an average player the overwhelming majority
of the time, however we might see much less of a difference at higher
skill levels and perhaps some factors such as IQ playing a larger role
beyond a certain number of hours spent learning to play.

It's interesting that you brought this up.  In fact, Bobby Fisher invented just such a game, which is called Fisher Random Chess (or Chess 960, because there are 960 different starting positions).  In this version, the players chose a card that shows how the back row of their pieces should be set up (with the only constant being that the king is between the rooks).  Fisher's goal was to create a game in which pure reasoning ability would trump rote memorization of openings.  It's actually quite a fun game.  Because there are far too many different starting positions to memorize openings for them, you are "playing chess" from move one.

Another interesting experiment, suggesting chess ability is not generalizable (and that chess masters probably don't have preternatural memories) involved showing chess masters and beginners a chess position and asking them to recreate it on a new board.  When the positions were taken from real games, the masters were much, much better than the beginners at recreating the position, often doing so almost exactly.  When the chess positions were random, however, the masters and the beginners were equally bad at recreating the position.  Thus the difference between a master and a novice is not native memory power, or even some general memory skill honed after years of practice, but rather a specific skill for recognizing patterns in chess configurations and treating them as "chunks" in working memory.

[wzeller]

I agree with your statement that a DNB score obtained after training would exhibit greater gains in those with WM deficits but high processing power than more "neuro-typical" people.  I was making a more general argument: i.e., I expect that the average correlation between DNB scores and whole-scale IQ would be higher before training than after training.  

One should consider that bright people with fairly weak
WM's make up University campuses throughout the world -- so it's not
surprising that targeted WM training would sharpen their fluid
intelligence.

This is an interesting statement.  Why would you think that such people would gravitate to universities?  Why wouldn't those with both high WM and high general intelligence be more typical?  I have some ideas, but I'm interested to hear your thoughts.

[Reece]

wzeller,

Thanks for sharing that very interesting information regarding Bobby
Fischer.

Having large working memory deficits myself, I must say that certain
university courses I've taken in the past have been utter nightmares
and I have a very difficult time learning even material which is
relatively easy in a classroom environment.

Free from distractions and able to proceed at my own pace, I have no
trouble academically and actually do quite well. On the other hand,
put me in a laboratory environment with a bunch of extroverts and give
me several pages of instructions on material I'm not familiar with...
Disaster is ascertained.

I am not aware of many NTs with high IQs and low WM -- seems mostly
people with ADHD and/or AS who fit that criteria.

[Gwern Branwen]

On Mon, Dec 7, 2009 at 11:59 AM, wzeller <wze...@gmail.com> wrote:

>> One should consider that bright people with fairly weak
>> WM's make up University campuses throughout the world -- so it's not
>> surprising that targeted WM training would sharpen their fluid
>> intelligence.
>
> This is an interesting statement.  Why would you think that such people
> would gravitate to universities?  Why wouldn't those with both high WM and
> high general intelligence be more typical?  I have some ideas, but I'm
> interested to hear your thoughts.

Here's an off-the-cuff theory. We're looking into a possible draw for
low WM high IQ towards universities and away from their other natural
home, professional occupations & business. High IQ crossed with low WM
would seem to lead to people low in conscientiousness: they are bad at
keeping calendars, remembering about meetings, and running errands.
The corporate world and law for example seem to thrive on reliable
clock-punchers; erratic people tend to be confined to special uses
when they can make up for their ineptitude. So LWHI people would
simply tend to do better in universities where deadlines are looser
and the hardest appointments are classes which run regularly over
large chunks of the year, and then network effects such as assortative
matching could take care of the rest. Perhaps also LWHI will have
lower latent inhibition due to lack of focus and so they are more
'creative' than an equally intelligent HWHI who has instead become a
hardcharging Philadelphia lawyer.

(There might be network effects there too; one of the most interesting
things about Charles Murray, to me, is not his race stuff but his
arguments that society is becoming ever more discriminatory towards
the stupid by creating ever more complex & detailed laws and
regulations. Lawyers who are HWHI may have a very distinct advantage
in being able to juggle endless precedents and laws and clauses than
LWHI; here would be an effect driving LWHI *out* of non-academic
professions.)

--
gwern

[milestones]

You're quoting a post I deleted because I wanted to edit certain parts
of it (not content so much as breaking up paragraphs)...here it is
again with small modifications:

"Wouldn't the strongest correlation between IQ and DNB be found if
the DNB scores were taken from subjects who had never trained at
all?"

It's hard to generalize. I'd say in some cases yes, some cases no.
I'll give my ideas on why one size doesn't fit all.

I would say that if someone's IQ is lowered by poor working memory by
way of attention problems, and a month or so of training
significantly
increases their fluid intelligence, then there is likely to be higher
correlation between post training DNB score/fluid IQ since the
training has remedied attention problems that brought their fluid IQ
level down. One should consider that bright people with fairly weak

WM's make up University campuses throughout the world -- so it's not
surprising that targeted WM training would sharpen their fluid
intelligence.

It appears what's happening in dual N back training is
converting innate processing speed into short term memory functions.
The Jaeggi study shows it can be done. However, this does not
convince
me that general intelligence can be raised. So this gets into the
debate of what underlies the G factor -- processing speed or working
memory?

I'd say WM is subsumed under processing speed and that Dual N
back combines processing power with the "attentional" realm/quadrant/
factor. In effect, then, IQ is not being raised by DNB training but
only is being fully realized or optimized by the category of people
most apt to benefit from it (such as bright, quasi-ADD college
students). I imagine that if you took a group with very fast reaction
and inspection times (which taken together have a high correlation
with G) but with low WM scores, then you would see greater gains in
fluid IQ after training with DNB than those who have WM and reaction/
inspection times that are evenly matched. I think this is why it's
hard to generalize since some people really do benefit from the
training more than others.

Given sufficient processing power, one can improve WM and fluid
intelligence but, the thing is, they are not improving G, but only
transferring extant processing power into a more attentionally
dependent domain. This, unfortunately, seems less possible for people
who lack such raw processing power...also, I bet training is less apt
to ignite the intellectual functioning of those who, prior to
training, have already transferred their processing power into the
realm of WM by other means.

It may be that dual n back is the best game we know of to date that
can convert raw processing power into WM ability and thus voila,
higher "fluid IQ." I do think that people with the right sort of
profile can truly benefit from DNB
training (and not just improve scores but improve latent ability);
however, this is all based on the sample that's tested. I may be
wrong...

WM training may improve the underlying G factor but I just don't see
that it really does. I see how it can
improve cognitive functioning for a certain subset of the general
population...but how it can benefit others to an equal degree is less
clear. It is an open question at this point. That said, I'd argue
that
the gains made by those primed to benefit from the DNB training are
real...though, how lasting the gains are, remains a question mark

[Pheonoxia]

> That said, I'd argue that the gains made by those primed to benefit from the DNB training are real...though, how lasting the gains are, remains a question mark

That's inaccurate. Gains from working memory training lost at least
six months:
http://brain-training.googlegroups.com/web/holmes2009-sustained-enhancement-of-wm-in-children.pdf?gda=3T0iR2gAAACc82UqlqVRqdWaMfnBZ6CDMmZ3J-aT5H4ro8uRY08_GyYyAmSVwKQ8QBSGA7hhAIZ7dMq1SZmA8YSt01IDud488VvkJaFeZCbKxY1VtLJoxabDCaaNjyUsSSbkax1xal8aNWB3A2L4HCtTMEZPaE7O

[milestones]

"This is an interesting statement. Why would you think that such
people would gravitate to universities? Why wouldn't those with
both
high WM and high general intelligence be more typical? I have some
ideas, but I'm interested to hear your thoughts."

Well, high WM and high G people are probably are more typical than any
other but there does seem to be a subset of low wm high and high IQ
people in universities. Why do they (if they really do, I have
suspiciouns but no data) exist in universities in fairly large
numbers? This is a good question -- I will venture some guesses.

One possible reasons is the special talent in a domain that doesn't
require wm at a premium -- art, drama, poetry -- that draws talented
students to the university. You might call these "right brained" types
and it's wrong to say these students tend to be low wm but it's likely
that wm is not as necessary for their talent to be realized as it is
in other academic areas. These fields, however, obviously do require
high intelligence and creativity. Contrast this with math and hard
science fields though, where high wm is an absolute necessity for
achievement along with intelligence and specialized abilities.

Also, humanities fields; they do require good analytical ability but
WM is not at an absolute necessity for high achievement. WM helps a
lot with writing essays, especially structuring but it seems WM helps
with speed of organization of one's points but not the insight. In
other words, then, if one is not out partying, then I bet a high IQ
low WM person can write just a good of an essay as their intellectual
equivalent with high WM, except that it will take the low WM person
more time to structure and executre their ideas. Either type may or
may not be conscientious to this end -- obviously, though, the low wm
type will be at disadvantage as far time goes.

As for why some students end up with high WM and others with low WM
despite having similar levels of G is a complicated question. I'd say
part of this issue goes to the roots of grade school and self
perception about one's own ability. Students who
are bright but feel week in wm (due to high creativity/low latency
inhibition) may gravitate toward less structure. They may also find
that, as they get older, their WM catches up to their intellect
without any focused training (like the DNB) but just by living life.
This gets out of psychometrics and into Neo-Piagetian theories about
self actualization and broader development issues which I think add to
our understanding.

[cev]

I'm confused as to why I find triple-n-back easier than dual. I just
struggled through a session of d9b at 40-50% before posting a typical
75% on t9b. Any theories?

[argumzio]

A compensatory effect by which your superior visual processing masks
your inferior auditory processing, an effect which can be demonstrated
by observing that your auditory performance is poorer than your
(combined) visual performance.

If that isn't the case, then it could be that you simply had a
contingent and transitory instance of poor performance on DNB,
followed by a prepared brain for the TNB task.

If that isn't the case, then it could be that you have been practicing
TNB (recently) more often than you have DNB.

If that isn't the case, then it could be that you experienced an
instance in which CHANCE_OF_GUARANTEED_MATCH made TNB easier even
though its value wasn't changed between DNB and TNB.

Or it could be all/many of the above.

Null hypothesis: there is no reason that you did poorly on DNB
compared to TNB.

Catch-all hypothesis: I haven't provided the correct explanation.

argumzio

[cev]

Hmm yes, good points.

My digit spans give equal scores for auditory and visual (80% at 14
for both), but my dnb scores are biased towards the visual, that is
true.

btw my digit span scores have been quite stubbornly stuck at that
level for the last 8 months or so, despite continuing to train.

[argumzio]

Would that be vis&aud forward dig spans or reverse dig span?

The latter more crucially relies on WM function/capacity than the
former.

argumzio

[cev]

Yeah, those are forward. Backward visual: 13 @ 75%; bwd auditory: 14 @
75%.

No obvious recent improvements at those either but, then again, I
haven't
taken those tests as many times as the forwards.

[argumzio]

Interesting. I wonder to what extent one could improve them. Lately,
I've seen some improvement in backward auditory digit span (11 @
97.73% ≈ 14 @ 76.79%), but I haven't bothered playing with backward
visual yet. After my initial shock (because I previously viewed myself
as being acoustically unsound), I have come to question the degree to
which this can be unreservedly declared an improvement. I probably
wasn't patient enough to go that far. (Let's admit it, it's very
boring and thankless work.)

Nevertheless, this gets me thinking about capacity and functionality
issues. It is expected that backwards spans are lower than forwards
spans, and yet some do just as well in either direction with no
obvious discrepancy. This would suggest to me that those who perform
in this way are performing at—for all intents and purposes—their
attention-capacity limit (without resorting to memorization
techniques).

Here's some background information on this interesting area of study:
http://www.fullpotentials.com/wst_page4.html

To highlight an interesting segment: "For visual digit spans, a span
of 3 or four show an average G.P.A. in the 2 point region. Spans of 5
through 10 show an average G.P.A. of either 3.1, 3.3 or 3.4 with the
order varying as digit span increased. A digit span of 11 show an
average G.P.A. of 4.0. For auditory digit spans, a span of 3 show an
average G.P.A. of 3.2. Average G.P.A scores vary between 3.4 and 3.5
between digit spans of 4 and 9."

Presumably, at our level there's nothing more to gain. Maybe when I
feel a little daring, I'll see if I can move my backward auditory span
any higher. I mean, I have memorized 16 digits of pi, my credit card
number (and its CVV code and PIN), a half-dozen phone numbers, SSN,
addresses, etc., but those aren't digit span tests of short-term
memory...

argumzio

[Aman Idle]

Hey just wondering, is brainworkshop still for free or you have to purchase it now? My version don't work no more for some reason. Can someome give me a direct link to the original software used in the jaeggi studies please?
cheers

[argumzio]

On Jan 18, 6:13 pm, Aman Idle <aman.i...@googlemail.com> wrote:
> Hey just wondering, is brainworkshop still for free or you have to purchase
> it now? My version don't work no more for some reason. Can someome give me a
> direct link to the original software used in the jaeggi studies please?
> cheers

I'm not sure what the trouble is, but the test release of 4.7 can be
found here: http://groups.google.com/group/brain-training/msg/8c719b95ff8a4dca

BW has been, and likely will always be, free.

argumzio