Brain Workshop 4.7 test release

[Paul]

Jonathan and I have been working on some fixes and improvements for
Brain Workshop 4.7.

* Added Dual Audio n-back modes. Go to the Sound Selection screen to
choose the sound set and channel (left, right, center) for each sound
stimulus.
* Number of trials per session now increases automatically with higher
n-back levels. The calculation can be adjusted from the config file.
(Jonathan Toomim)
* Timing resolution has been increased to 0.1 seconds, with a maximum
speed of 0.3 seconds per trial.
* Toggling Manual Mode no longer reverts to default settings.
* Title screen graphic is now colored inversely when a black
background is selected.
* Stats file may be specified on the command line with --statsfile,
complementing the --configfile command line parameter.
* Various minor fixes.

Try it out and let me know of any bugs. This is just a test release -
I'll make an official release in a week or two.

http://brainworkshop.sourceforge.net/misc/brainworkshop-4.7-

[Paul]

Windows installer:
http://brainworkshop.sourceforge.net/misc/brainworkshop-4.7-win32-setup.exe

Windows zip version (no installer):
http://brainworkshop.sourceforge.net/misc/brainworkshop-4.7-win32.zip

Linux / OS X Source version (requires python 2.5+):
http://brainworkshop.sourceforge.net/misc/brainworkshop-4.7.zip


Please note: This version will overwrite your existing config file.

Paul

[argumzio]

This is simply stunning work!

Thank you for developing and implementing the ideas we very grateful
users have conceived, Paul!

My initial suggestion would be to set the config default for the two
separate channels as L=left and ;=right so users are confused by the
center-center setting as I was. Other than that I currently see no
problems on my end.

argumzio

[argumzio]

On Nov 29, 6:26 pm, argumzio <argum...@gmail.com> wrote:
> separate channels as L=left and ;=right so users are confused by the

I mean so users are NOT confused. ;-)

Cheers,
argumzio

[argumzio]

On second thought, no, it is probably a better idea to keep them both
as center, since people may want to resume their favored tasks. I
suppose my excitement got the better of me.

An interesting observation I have made, at any rate, is that while
playing A-A, my perception sometimes blends or even switches the
origin of the stimuli. This is definitely going to augment some neural
networks.

argumzio

On Nov 29, 6:26 pm, argumzio <argum...@gmail.com> wrote:

[Paul]

Currently in single audio mode it plays in both speakers regardless of
the channel setting. I'll implement your suggestion for the 4.8
release.

Paul

> --
>
> 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.
>
>
>

[Pheonoxia]

Game Mode A, which is dual auditory only, A-A, is hard. I sucked so
bad at 2-back I got bumped down to 1-back. I split the channels
between left and right. For left I have letters, numbers, and the
phonetic alphabet enabled. On the right I have piano notes enabled.
I'm going to be training this exclusively for a while.

Thanks, Paul and Jonathan! You both rock!

[Paul]

By the way, if there's anyone who experienced sound problems in BW 4.4
(sound cutting out intermittently), please try this version and let me
know if there's any difference.

Paul

[exigentsky]

This is fantastic!

Thank you,
Alex

[Jonathan Toomim]

OS X Application bundle:
http://jtoomim.org/code/brainworkshop-4.7-OSX10.3.zip

[David]

Hi Paul,

Firstly, thank you for your excellent work.

Secondly, have you considered qualitative mechanisms as well for
expanding the audio n-back complexity space? For example, you could
have numbers recited in six different languages (say, English, French,
Spanish, Italian, German, Russian). One n-back task would then be
repetition of number, another repetition of language. Same with
numbers recited in varying pitches. This obviously could be combined
in varying ways with split-channel dual n-back to form triple, quad,
and pent audio-only n-back tasks.

[frozen yog]

Hi Paul,

FYI

I over wrote 4.5, and the config file had the 4.5 details.

I deleted the folder and reinstalled 4.7, and there is no config.ini
file.

I'm on XP, and the error message when trying to run straight after
installing is
"Create process failed do to code 14001 application configuration is
incorrect. Reinstalling may fix the problem"
....

Downloading python 2.6.4 fixed this issue.

On Nov 30, 8:56 am, Paul <plh...@gmail.com> wrote:
> Windows installer:http://brainworkshop.sourceforge.net/misc/brainworkshop-4.7-win32-set...

[Paul]

The config file is not created/overwritten until the first time the
new version is launched.

Thanks for letting me know about the error message. Upon further
investigation it seems like the easiest solution would be to use
Python 2.5 for building the installer instead of 2.6.

Paul

[Paul]

David,

Thanks for the feedback. Implementing complex audio tasks would not be
too difficult from a programming standpoint; the difficult part is
finding/creating the high-quality sounds that would make these modes
possible. (It's easy to generate graphics in a program, but not so
easy to generate sounds.) Numbers recited in 8 languages would be
required to continue the trend of each modality having at least 8
possible values.

To have numbers recited in various pitches, I would need at least one
set of numbers in a singing voice (constant pitch) which I could
pitch-shift using audio editing software, but this might sound too
artificial. I could record the sounds myself but I believe the results
would be less than ideal. :) Another idea would be to find wavs of 8
different musical instruments playing each note of a C scale. (Perhaps
recording a MIDI synthesizer would be the easiest option.) In each
case the challenge is in acquiring the array of 64 sounds required for
a minimal 2 dimensional audio task.

Would the end result be much better than the current implementation in
BW 4.7, which is to play two sounds simultaneously? The current method
could even be extended to play 2 or more sounds in each ear for quad
auditory n-back, assuming the sound sets are sufficiently distinct
from each other. For quad audio n-back using a single sound, 8^4 or
4096 different sounds would be needed.

The most practical direction for BW would be to add a 3rd and 4th
audio modality along the current lines, and diversify the selection of
available sound sets. (Something to work on for the 4.8 release, time
permitting)

Paul

[Bewater]

Hi Paul, i have a bug here. In this new version i cant open the
program, it sais that it has to be something wrong with the
instalation and doesnt let me open it.
I tryed with the zip version, and with the setup version, and i tried
in my pc and on a notebook. But i get this bug all the time.

it has to be something in common between my pc´s. I have a desktop pc
and an asus eee pc, both with xp

I have the zip old version and im having no problem with it. It´s only
in the new one that this happens.

Sorry for my english, im from Argentina

Oh by the way, i have a time reading this group and i love reading
your discussions, keep this working guys its SO interesting =)

[Paul]

When did you download it? I changed something a couple of days ago
which I thought would solve the problem and re-uploaded copies of 4.7
to the same location. Please try re-downloading the newest setup file.

Paul

[Bewater]

It works! Thank you.


by the way i have another bug with my video device in my pc, when it´s
turned on, the only thing i see is a grey squeare in the middle of
brainworkshop´s window. Where the brain image should be, and when i
start a session, i can hear the sounds, and the grid appear, but not
the squeares. It happens in the old version too.

But it´s an onboard card of a pc chips motherboard, so im pretty sure i
´ll be the only one with this bug.


it is not important cos it´s an uncommon device (VIA S3G Uni/Crhome
Pro IGP), and i play in my notebook and it works very well. I tell you
cos you wanted to know about the bugs and all that stuff.

Bewater

[exigentsky]

I tried the Jaeggi mode again and it appears that 4.7pre has the same
issue as 4.6. On Snow Leopard, there are some Vsync problems if full-
screen (otherwise it's fine). You can see the images get distorted
every few seconds as a line appears to scan from top to bottom. It's a
little like watching a TV on camera. You can see the scanning. I've
never had this issue on Tiger so I'm not sure what's going on.

[Paul]

4.4 had VSYNC on by default but I disabled it in 4.7 because many
systems would give an error message. I re-enabled it on the Mac
platform in SVN (linux and Windows users can enable it using the
--vsync command line option). Hopefully this will resolve the error.

Paul

[exigentsky]

Ah, great. I hope that fixes it. :)

- Alex

On Dec 7, 7:01 pm, Paul <plh...@gmail.com> wrote:
> 4.4 had VSYNC on by default but I disabled it in 4.7 because many
> systems would give an error message. I re-enabled it on the Mac
> platform in SVN (linux and Windows users can enable it using the
> --vsync command line option). Hopefully this will resolve the error.
>
> Paul
>

[exigentsky]

Is there a way I can reenable it manually like on Linux/Windows? Full-
screen helps me focus better.

Thanks,
Alex

On Dec 7, 7:01 pm, Paul <plh...@gmail.com> wrote:

> 4.4 had VSYNC on by default but I disabled it in 4.7 because many
> systems would give an error message. I re-enabled it on the Mac
> platform in SVN (linux and Windows users can enable it using the
> --vsync command line option). Hopefully this will resolve the error.
>
> Paul
>

[Paul]

Full-screen can still be enabled in the config as before. only vsync
has changed (which can affect flickering on some systems).

>>> trai...@googlegroups.com.

[exigentsky]

I know but the flickering, image distortion and scanning lines makes
full-screen too distracting. Although, it is weird that I get none of
this outside of full-screen,

On Dec 29, 10:13 am, Paul <plh...@gmail.com> wrote:
> Full-screen can still be enabled in the config as before. only vsync  
> has changed (which can affect flickering on some systems).
>

[Erik]

On 30 Nov 2009, 00:53, Paul <plh...@gmail.com> wrote:
> * Number of trials per session now increases automatically with higher
> n-back levels. The calculation can be adjusted from the config file.
> (Jonathan Toomim)

I just joined the group and tried the 4.7 release.
The previous versions had default number of trials for n-back set to
20+n, which gives 20 tests per session. 4.7 defaults to 20+n^2, which
gets massive quite fast.

Bug? :)

[sutur]

i really like the x+n^2 formula. try setting x to 10! that way you can
always start with dual 2-back and do a few warm up sessions without
wasting too much time on the lower levels. i used to set the number of
sessions to 60 anyway to have fewer but more intense sessions.

[Jonathan Toomim]

It's a feature, not a bug. The idea is half to train your attention
span as well as your working memory, and half to make sure that most
trials require you to both store to and retrieve from memory.

If you want to play with the number of trials per session, check the
config.ini file for these lines:

----------------------------------------------
# The number of trials per session equals
# NUM_TRIALS + NUM_TRIALS_FACTOR * n ^ NUM_TRIALS_EXPONENT,
# where n is the current n-back level.

# Default base number of trials per session.
# Must be greater than or equal to 1.
# Default: 20
NUM_TRIALS = 20

NUM_TRIALS_FACTOR = 1
NUM_TRIALS_EXPONENT = 2
-----------------------------------------------

If 20+n^2 grows too fast for you, you can try 20+0.5n^2 or 20+n^1.5.
(BrainWorkshop will round down if a non-integer number of trials is
produced by this formula.)

If you want the old behavior, set NUM_TRIALS_EXPONENT to 1.

Jonathan

[exigentsky]

Sorry to be off-topic, but I'm curious about the official 4.7 release.
It seemed that it would be released in mid-December. Moreover, I
didn't notice any blocker bugs when I was testing it. Although, I only
tested a small subset of the features. Did you guys decide to expand
the feature set or update some core library like Pyglet to 1.1.4?

[J Wolfgang Goerlich]

Bug report: On Windows 2008 and Windows 7, 64-bit, the squares do not
display approximately 25% of the time. When exiting, I receive a
message that an error has occurred. The following is in the log:

Traceback (most recent call last):
File "\loewis\25\python\Modules\_ctypes\callbacks.c", line 274, in
'calling callback function'
File "pyglet\app\win32.pyc", line 84, in _timer_func
File "pyglet\app\__init__.pyc", line 187, in idle
File "pyglet\clock.pyc", line 700, in tick
File "pyglet\clock.pyc", line 303, in tick
File "brainworkshop.pyw", line 4063, in update
File "brainworkshop.pyw", line 2957, in update
File "pyglet\text\__init__.pyc", line 269, in _set_text
File "pyglet\text\document.pyc", line 285, in _set_text
File "pyglet\text\document.pyc", line 425, in insert_text
File "pyglet\event.pyc", line 340, in dispatch_event
File "pyglet\text\layout.pyc", line 971, in on_insert_text
File "pyglet\text\layout.pyc", line 960, in _init_document
File "pyglet\text\layout.pyc", line 894, in _update
File "pyglet\text\layout.pyc", line 870, in _get_lines
File "pyglet\text\layout.pyc", line 992, in _get_glyphs
File "pyglet\text\document.pyc", line 603, in get_font_runs
File "pyglet\text\document.pyc", line 613, in get_font
File "pyglet\font\__init__.pyc", line 624, in load
AttributeError: 'NoneType' object has no attribute 'object_space'