Make Brainworkshop Harder and Easier

[whoisbambam]

I was messing with the config.ini file. I set it for manual mode, to
start at my preferred Nback level, locked in so it can not change
levels up or down.

On the screen it says press F5 to decrease the speed--I do not
understand how to change this in the ini file. If possible, can
somebody paste an example or two of decreasing and increasing speed?

I set the speed becasue it is easier for me to slow it down a bit when
increasing levels.

I am also interested in knowing how to make each level harder--so if i
am on dnb4, i want to slowly increase the difficulty manually--making
it most difficult before going to dnb5 (and then making dnb5 easiest).

What are the steps to achieve this in dnb?

As it stands now, I just lock myself into a given dnb level, i adjust
the speed up and down, and i first set the
DEFAULT_CHANCE_OF_INTERFERENCE =0, then go to 0.250, then i guess
to ???

any other settings to make it more difficult (not super fast speed) as
i do my daily training, in preparation to weaning into the next level?

For some reason my anxiety gets too high if i am on say dnb3 getting
10+ 100% and all others above 80%, then i go to dnb4 or whatever, and
perform abysmally (say 25%)--i know i am probly supposed to challenge
myself like that, but there are other processes i guess that probly
concern 'fear of failure' that causes emotions to raise (frustration)
that probly is something i should try to minimize via the
beforementioned info.

thanks.

[Jonathan Toomim]

To change the speed in the config.ini file, take a look at this section:

# Number of 0.1 second intervals per trial.
# Must be greater than or equal to 4 (ie, 0.4 seconds)
# Look above to find the corresponding mode number. Add a line for
the mode
# if it doesn't already exist. Modes not specifically listed here will
# use TICKS_DEFAULT instead.
#
# Crab and multi-modes will default to the ticks associated with the
modes
# they're based on, *plus an optional bonus*, unless you add a line
here to
# give it a specific value. Any bonuses will be ignored for specified
modes.
TICKS_DEFAULT = 30
TICKS_4 = 35
TICKS_5 = 35
TICKS_6 = 35
TICKS_7 = 40
TICKS_8 = 40
TICKS_9 = 40

# Tick bonuses for crab and multi-modes not listed above. Can be
negative
# if you're a masochist.

BONUS_TICKS_CRAB = 0
BONUS_TICKS_MULTI_2 = 5
BONUS_TICKS_MULTI_3 = 10
BONUS_TICKS_MULTI_4 = 15

> --
> 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
> .
>

[Jonathan Toomim]

More specific instructions:

Time intervals in BW are calculated as a number of ticks. Each tick
lasts 100 ms. If you want the trials to go faster, you want each one
to last for a smaller number of ticks.

Each game mode (e.g. dual n-back, crab triple stim PIS position n-
back, etc) has a mode number associated with it. For DNB, it's 2. To
change the length of DNB sessions to 2 seconds without changing any
other modes, you would want to add a line that says:

TICKS_2 = 20

If you wanted to slow down all game modes by 1 second, you would
change TICKS_DEFAULT:

TICKS_DEFAULT = 40

If you just want to slow down crab modes, you can do that by changing
the bonuses:

BONUS_TICKS_CRAB = 10

would slow down crab modes by 1 second compared to their base non-crab
modes.

More on the mode numbers: For vanilla DNB, it's 2. For most other
modes, you can find a list in the config.ini file. For crab modes,
add 128. For 2x, 3x, and 4x-stim modes, add 256, 512, and 768,
respectively. Another way to find out the mode number for a mode is
to play one session in that mode, then check the last line of your
[USERNAME-]stats.txt file. The second number after the short mode
name string (e.g. D5B) is the mode number (should be 2 for DNB
sessions; use this to make sure you're using the right column).

Jonathan

[whoisbambam]

thank you.

ok, so right here:

# give it a specific value. Any bonuses will be ignored for
specified
modes.
TICKS_DEFAULT = 30
TICKS_4 = 35
TICKS_5 = 35
TICKS_6 = 35
TICKS_7 = 40
TICKS_8 = 40
TICKS_9 = 40

since i play dnb, that would change to this:

# give it a specific value. Any bonuses will be ignored for
specified
modes.

TICKS_2 = 20

TICKS_DEFAULT = 30
TICKS_4 = 35
TICKS_5 = 35
TICKS_6 = 35
TICKS_7 = 40
TICKS_8 = 40
TICKS_9 = 40

What is dnb set at already, this?:
TICKS_2 = 10

since that is 10ticks, each tick 100ms, that is 1000ms, which is 1
second, and the minimum interval is 4, but after that, it can be any
number greater than that (5, 6, 14, etc)

thanks.

[whoisbambam]

sorry, i think default is 3seconds.
:)
thank all of you for your patience as i sort thru this.

so increasing to 35 for 3.5seconds may help in the lower levels of dnb
probably.

besides changing the times back and forth, and the interference lower
and higher (i am assuming higher interference makes it harder?), are
there any ways to adjust the 'difficulty' of dnb?

i am trying to make it so i can first make the current level i am
easiest to hardest before progressing to the next level............

[Jonathan Toomim]

> TICKS_2 = 20

That would make DNB use 2.0 seconds per trial, which is 33% faster
than the default setting.

> What is dnb set at already, this?:
> TICKS_2 = 10

No, the default for DNB comes from the TICKS_DEFAULT = 30 setting. The
default is 3 seconds between trials.

Jonathan

[likeprestige]

Hey Jonathan,

Is there any way to reduce the amount of time stimuli were displayed
while still maintaining the set time interval between trials?

Thanks Jonathan,

likeprestige

[Jonathan Toomim]

You'd have to run it from the source code and edit the source. Around
line 4200 or so, in the function update(dt), there's a couple of lines
that look like this:

> if mode.tick == 6 or mode.tick == mode.ticks_per_trial - 1:
> for visual in visuals: visual.hide()

That 6 there is what determines when the square(s) disappear(s). A
value of six means that it will disappear after (6-1)*100ms = 500ms.
So if you want them to show up for 300ms, use a value of 4 there.

Jonathan

[likeprestige]

Thanks Jonathan, appreciated!

likeprestige

P.S "Helen Wills Neuroscience Institute" = Certainly not a bad thing
to have on your resume!

[whoisbambam]

I was reading thru the config.ini file.

How does changing the chance of guaranteed match make it significantly
easier? doesnt repetitive sequences (like a square showing up in spot
0, 7, then 0 again) make it more confusing?:


# The chance that a match will be generated by force, in addition to
the
# inherent 1/8 chance. High settings will cause repetitive sequences
to be
# generated. Increasing this value will make the n-back task
significantly
# easier. The value must be a decimal from 0 to 1.
# Note: this option has no effect in Jaeggi mode.
# Default: 0.125
CHANCE_OF_GUARANTEED_MATCH = 0.125


Is there some example somebody can come up with that shows the effect
of guaranteed match and how that is supposed to be easier?


thanks.

[Jonathan Toomim]

It's mostly a function of how the scoring works. If you have
JAEGGI_SCORING set to True, then higher CHANCE_OF_GUARANTEED_MATCH
doesn't make things easier.

(By the way, I'm a huge fan of JAEGGI_SCORING = True. I'm considering
making it the default in the future. I'd be interested in hearing your
input on this.)

So for each trial and each modality, there are four possibilities for
a response:

True positive (TP): you pressed a button when there was a match
True negative (TN): you didn't press a button when there was no match
False positive (FP): you pressed a button when there was no match
False negative (FN): you didn't press a button when there was a match

With JAEGGI_SCORING = True, both TPs and TNs are considered correct
responses.
With JAEGGI_SCORING = False, only TPs are considered correct
responses. TNs are ignored.

In both modes, both FPs and FNs are considered correct responses.

Thus, with JAEGGI_SCORING = False, the number of chances you have to
make a correct response (TP) equals the number of matches in the
sequence. Your "percent correct" score is TP/(TP+FP+FN), so trading
TNs for TPs (by increasing the number of matches possible) will
improve your score, even if you still make the same number of errors
in a session.

With JAEGGI_SCORING = True, your "percent correct" score is (TP+TN)/(TP
+TN+FP+FN), so trading TNs for TPs does absolutely nothing.

Does that make sense?

Jonathan

[whoisbambam]

This one seems strange:

True negative (TN): you didn't press a button when there was no match

I mean, the rest seem pretty obvious:
tp: there was a match, i pressed the button, i get a 'point' (score
raises)
fn: there was a match, i didnt press the button, i didnt get a point
(score lowers)
fp: there was no match, i pressed the button, i didnt get a point.
(score lowers)

tn seems strange.



On another note, it seems going from dnb3 to dnb4 is more difficult
than going from dnb2 to dnb3.

Is this my imagination, or is there some sort of 'reason' for this
seemingly 'nonlinear' change in difficulty?

does this mean going from dnb4 to dnb5 will be even MORE substantially
difficult than it was going from dnb3 to dnb4 (for me, anyway--i am
thinking this may have something to do with my own inherent working
memory--that perhaps i can only hold 6 items in memory or something
(dnb3?), and dnb4 requires 8 or something, and exceeds my working
memory (inherent).


If i am to understand it correctly, it really doesnt matter, cause it
is not the progression of levels that makes the difference, but the
'struggle' playing the game day to day.

If I am stuck on dnb4 forever, it doesnt matter--i would still get the
benefits of improvement as long as i was struggling day to day, in
other words..........

[Joe G]

Most people find 3->4 and 4->5 to be the hardest jumps. As to why that
is, I am not sure. It may be related to the magic number of 7+-2 which
solidly fits around in this region. D(3-8)B ranges from 6-8 elements:
squarely in the middle of the average range. This -- to me -- implies
that one will not surpass surpass this level without (a) naturally and/
or subconsciously developing chunking techniques or (b) legitimate,
functional changes to the structure of one's brain.

[Joe G]

I meant to say "D(3-4)B"

[Jonathan Toomim]

One reason why TNs might be strange to you is that with JAEGGI_SCORING
= False, they're ignored by BW. (Or maybe the default is for TNs to be
ignored because Paul knew they would be strange to many people.)

At any given point, you have the option of either pressing a key
(positive response) or not pressing a key (negative response). Your
response can either be correct (true) or incorrect (false). Thus, you
have a 2x2 matrix of possible response types.

You could think of it using this analogy: Imagine you're on a jury,
and a man is on trial for allegedly murdering his business partner. He
either killed him or he didn't. You can either proclaim him guilty or
not guilty. Thus, you have a 2x2 matrix of possible verdicts. If you
correctly proclaim him guilty, that's a TP. If you correctly proclaim
him not guilty, that's a TN. If you throw an innocent man in prison,
that's a FP. If you let a murderer run the streets unchecked, that's a
FN. Acquitting the innocent is at least as important as convicting the
guilty.

With JAEGGI_SCORING = False, BW only rewards you for convicting the
guilty. That's why you'll do better if there are more murderers in the
world. So to speak.

Jonathan

[whoisbambam]

Ok, I admit it--I am having much difficulty with dnb4 (i find it
embarassing to discuss these things publicly, lol. i need to get over
it).

As I ease my way thru it, I was wondering if there was a way to almost
completely eliminate interference as much as possible (it may be
close, considering dnb4 needs 5 squares or something before i press
the button, and there are only 9 squares, so when the pattern repeats,
i think it must overwrite one of the squares--if not, we are close to
this problem).

I am trying to just keep 4 positions and 4 sounds straight in my head/
mind, but once it starts blinking a square over the top of a previous
square that it showed me, i seem to drop all the information out of my
mind, lose track of everything, and literally have to start over with
the matching process.

:(

I have no idea why interference screws me up so much--most annoying.

I am thinking it is a good exercise just to keep the 4 positions and 4
sounds clear (and then learn the 5th matches without losing the 2nd
pair of matches).

If this can not be done, is there some other program i can practice
on, just to try to remember 4 progressing pairs of visual/auditory
matches?


I hope this made sense............

I tried to get 'used' to interference by jacking it way up in dnb3--i
was doing pretty good with it, altho i am not sure i was actually
remembering it, but instead i knew some 'pattern' once interference
ocurred.