The visual feedback in BW
Gwern Branwen
article: http://community.haskell.org/~gwern/static/N-back%20FAQ.html
I reason that separate articles meant that no one was seeing the other
and the content was growing increasingly stale. Hopefully this won't
annoy anyone.
----
More on topic, I recently switched the option for visual feedback (n
stimuli to go, correct/incorrect visual/audio press, help docs etc.)
during a round to be off; I'd always done it with on, with I think is
the default for Brain Workshop? I figured that it would make my scores
plummet, and was curious how much it would hurt me. To my great
surprise, the effect seems to have been just the opposite. Here's a
week of my stats:
2010-08-26 05:49:06,D4B,61,2,4,45,60,1,1,58,63,0,0,0,0,0,0,0
2010-08-26 06:49:46,D4B,61,2,4,45,60,1,2,100,46,0,0,0,0,0,0,0
2010-08-26 07:29:07,D4B,56,2,4,45,60,1,3,72,41,0,0,0,0,0,0,0
2010-08-27 06:17:57,D4B,60,2,4,45,60,1,1,71,52,0,0,0,0,0,0,0
2010-08-27 06:30:51,D4B,73,2,4,45,60,1,2,90,58,0,0,0,0,0,0,0
2010-08-27 07:07:12,D4B,61,2,4,45,60,1,3,70,57,0,0,0,0,0,0,0
2010-08-27 07:30:22,D4B,81,2,4,45,60,1,4,84,77,0,0,0,0,0,0,0
2010-08-27 07:40:58,D4B,66,2,4,45,60,1,5,100,50,0,0,0,0,0,0,0
2010-08-30 03:04:19,D4B,66,2,4,45,60,1,1,81,52,0,0,0,0,0,0,0
2010-08-30 04:03:55,D4B,76,2,4,45,60,1,2,91,64,0,0,0,0,0,0,0
2010-08-30 05:58:14,D4B,55,2,4,45,60,1,3,72,44,0,0,0,0,0,0,0
2010-09-01 03:35:55,D4B,51,2,4,45,60,1,1,80,38,0,0,0,0,0,0,0
2010-09-01 04:22:08,D4B,63,2,4,45,60,1,2,87,50,0,0,0,0,0,0,0
2010-09-01 05:04:25,D4B,60,2,4,45,60,1,3,100,37,0,0,0,0,0,0,0
2010-09-01 05:28:41,D4B,50,2,4,45,60,1,4,81,31,0,0,0,0,0,0,0
2010-09-01 05:57:40,D4B,76,2,4,45,60,1,5,100,64,0,0,0,0,0,0,0
2010-09-01 07:20:14,D4B,83,2,4,45,60,1,6,92,72,0,0,0,0,0,0,0
2010-09-01 07:39:57,D4B,50,2,4,45,60,1,7,71,38,0,0,0,0,0,0,0
2010-09-01 07:52:09,D4B,77,2,4,45,60,1,8,84,71,0,0,0,0,0,0,0
# switched to no feedback
2010-09-02 01:20:32,D4B,50,2,4,45,61,1,1,64,35,0,0,0,0,0,0,0
2010-09-02 04:06:35,D4B,87,2,4,45,61,1,1,100,78,0,0,0,0,0,0,0
2010-09-02 05:59:55,D4B,74,2,4,45,61,1,2,87,54,0,0,0,0,0,0,0
2010-09-02 06:24:15,D4B,63,2,4,45,61,1,3,88,46,0,0,0,0,0,0,0
2010-09-02 06:45:13,D4B,70,2,4,45,61,1,4,92,57,0,0,0,0,0,0,0
2010-09-02 07:38:42,D4B,57,2,4,45,61,1,5,77,47,0,0,0,0,0,0,0
2010-09-03 05:55:23,D4B,61,2,4,45,61,1,1,72,53,0,0,0,0,0,0,0
2010-09-03 06:34:08,D4B,95,2,4,45,61,1,2,100,90,0,0,0,0,0,0,0
2010-09-03 06:59:10,D4B,60,2,4,45,61,1,3,68,52,0,0,0,0,0,0,0
2010-09-03 07:43:05,D4B,73,2,4,45,61,1,4,92,50,0,0,0,0,0,0,0
2010-09-03 07:54:02,D4B,82,2,4,45,61,1,5,100,73,0,0,0,0,0,0,0
Perhaps I coincidentally improved at D4B such that I could get a 95%
right (my highest ever in D4B) and pick up a bunch of 70s and 80s, but
I suspect it's the lack of feedback. I had weakly endorsed turning off
the feedback in the FAQ:
http://community.haskell.org/~gwern/static/N-back%20FAQ.html#and-the-flashing-rightwrong-feedback
But I wonder if I should change that section. Does anyone who didn't
comment in the 2 linked threads (2008 & 2009) have a fresh opinion on
the matter?
--
gwern
likeprestige
I'm unsure as to what you're talking about, can you elaborate? Sorry.
regards,
likeprestige
> comment in the 2 linked threads (2008 & 2009) have a fresh opinion on
> the matter?
>
> --
> gwern
>
> 144KViewDownload
Jonathan Toomim
labmates that this question has been investigated empirically (at
least in the context of visual psychophysics), and that the consensus
in the field is that using feedback reduces immediate performance but
improves learning rates. I haven't looked up the research to confirm
their opinion, but it sounds plausible to me. I would also expect it
to apply to Brain Workshop. The idea, as I see it, is that feedback
reduces performance because, when you get an answer wrong and you know
it, your brain goes into an introspective mode to analyze the reason
for the error and (hopefully) correct it, but while in this mode your
brain will be distracted from the task at hand and will be more likely
to miss subsequent trials.
Maybe if I find the time, I'll try to look for the papers on the
subject which my labmates and PI alleged existed. You guys might find
a glance at the abstracts worthwhile.
Anyway, I think it's generally a good idea to leave feedback on, since
the point of doing DNB is the learning process that the task provides,
not the task itself. However, it might make sense to turn feedback
off temporarily if you're struggling at a new level. Removing the
distraction of feedback might be the all you need in order to be able
to perform above the fallback threshold (i.e., 50%) long enough to get
used to the higher difficulty.
One thing in particular about how BW does feedback that kinda bothers
me: I think that the visual feedback causes people to tend to move
their gaze away from the fixation cross right before the next trial
when they miss a match, and that's probably bad. When doing
psychophysics, I've usually used auditory feedback. Since there are
so many different modalities in BW, though, it would probably be
difficult to create auditory feedback that would tell the user which
match was missed. Maybe there's something else that can be done to
minimize the problem, such as not changing the font to bold for misses
so that there's no motion in the text, just a color change. Hmm...
Jonathan
> --
> 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
> .
>
> <stats.txt>
likeprestige
Not sure if it would be beneficial but I do like the idea of including
visual/auditory feedback on the target set (information to be
recalled) as well as or as opposed to what is currently implemented.
Just runs along the line of your _interference_ theory a bit more so.
"The idea, as I see it, is that feedback reduces performance
because, when you get an answer wrong and you know it, your brain
distracted from the task at hand and will be more likely to miss
subsequent trials"
negatively reinforce the 'brain training participants' subsequent
trials ... But I dunno. Thoughts?
likeprestige
Vassilis P
the dnb task easier in certain cases because it helps you
"recalibrate". When you have negative feedback you understand that you
didn't remember the sequence correctly and you can start over. At
lumosity dnb where you have no grid and no feedback things are much
more difficult, at least for me.
I would have to agree though that at other tasks at lumosity the sound
feedback is distracting and in several cases I have thought of turning
it off.
sutur
feedback. now, if you want to see the feedback, you also have to turn
on the text on the top of the screen.
likeprestige
and/or their 'brain training' products. If anything, I advise against
the idea of becoming affiliated with them through the form of a
membership or through some other means.
likeprestige
Reece
that, I am certainly going to give it a try following what Jonathan
has posted. My guess would be that those of us who are most easily
distracted probably have the most to gain by playing with the feedback
on.
polar
groups.google.com/group/brain-training/browse_thread/thread/
775fba46e8c163f1/920b8fa0d8598e85?#920b8fa0d8598e85 ). Pro-feedback
argument goes, that you need evaluation to learn. Anti-feedback
argument is, that evaluation disturbs your attention. Personally,
after more informations from successful n-backers and Dr. Jaeggi
herself (http://groups.google.com/group/brain-training/browse_thread/
thread/955524caaf2e9001/c49c3061eff02f29?#c49c3061eff02f29), I
switched feedback off.
It is of course true, that feedback is vital for learning. But the
point is not to learn n-back task - the point is to stimulate a
general ability we dont even "know".
- First thing, performance on n-back task was not correlated to iq
gain. Actually only thing that was correlated to iq gain, was time
spent by training (not even pretest iq or wm level). So you really
dont need to go for best n-back performance possible, to get the most
of it.
- Second thing, before you hit the key, you already know *fairly
well*, if you've made a hit (whether your memory representation is
clear) or not. So getting evaluation is redundant to some level.
- It really DOES break attention. Of course you can argue that you are
"raising the cognitive load, stroop / multitasking type". But N-BACK
shows the best transfer ever (actually Jaeggi proved taht single type
n-back is at least as effective as dual or triple) , so you dont want
to dillute its effectivity by combining it with other types of load,
which in addition are interferring with the task. You need the
attentional flow - my personal view.
> comment in the 2 linked threads (2008 & 2009) have a fresh opinion on
> the matter?
>
> --
> gwern
>
> 144KViewDownload
likeprestige
it a go and thought...
1. No feedback - required more focus because you are relying less on
it telling you whether or not you got it right.
For some reason it makes me think of a person's level of independence.
Those that have high levels of independence perhaps have a low need to
seek approval from others. The reverse for low levels of independence.
I suppose you could describe the same scenario in regards to a persons
level of confidence.
Does lack of feedback build "trust in oneself"?
2. No gridlines. Same principle. Much harder to rehearse position
stimuli.
I have another psychological scenario that I can talk about in regards
to grid lines v. no grid lines however, I imagine that you guys can
already anticipate what that scenario would entail.
Conclusion = going to give it a go for a few days, see how it effects
may actual results and perceived difficulty level of the game.
likeprestige
Gwern Branwen
> I've just turned the feedback option as well as the grid lines. Gave
> it a go and thought...
>
> 1. No feedback - required more focus because you are relying less on
> it telling you whether or not you got it right.
>
> For some reason it makes me think of a person's level of independence.
> Those that have high levels of independence perhaps have a low need to
> seek approval from others. The reverse for low levels of independence.
> I suppose you could describe the same scenario in regards to a persons
> level of confidence.
>
> Does lack of feedback build "trust in oneself"?
>
>
> 2. No gridlines. Same principle. Much harder to rehearse position
> stimuli.
>
> I have another psychological scenario that I can talk about in regards
> to grid lines v. no grid lines however, I imagine that you guys can
> already anticipate what that scenario would entail.
>
>
> Conclusion = going to give it a go for a few days, see how it effects
> may actual results and perceived difficulty level of the game.
>
> likeprestige
My first impression is that turning off the gridlines is as brutal as
turning off the feedback.
2010-09-08 02:41:17,D4B,80,2,4,45,61,1,1,90,73,0,0,0,0,0,0,0
2010-09-08 04:01:32,D4B,50,2,4,45,61,1,2,71,33,0,0,0,0,0,0,0
2010-09-08 04:34:16,D4B,63,2,4,45,61,1,3,75,57,0,0,0,0,0,0,0
2010-09-08 05:40:44,D4B,81,2,4,45,61,1,4,100,70,0,0,0,0,0,0,0
2010-09-08 06:12:36,D4B,70,2,4,45,61,1,5,66,72,0,0,0,0,0,0,0
2010-09-08 06:33:35,D4B,64,2,4,45,61,1,6,71,57,0,0,0,0,0,0,0
2010-09-08 07:17:50,D4B,79,2,4,45,61,1,7,92,68,0,0,0,0,0,0,0
2010-09-08 07:36:57,D4B,69,2,4,45,61,1,8,92,46,0,0,0,0,0,0,0
2010-09-08 07:47:10,D4B,72,2,4,45,61,1,9,100,56,0,0,0,0,0,0,0
2010-09-09 06:05:35,D4B,62,2,4,45,61,1,1,71,53,0,0,0,0,0,0,0
2010-09-09 06:19:22,D4B,70,2,4,45,61,1,2,80,62,0,0,0,0,0,0,0
2010-09-09 06:54:19,D4B,77,2,4,45,61,1,3,100,60,0,0,0,0,0,0,0
2010-09-09 07:33:36,D4B,76,2,4,45,61,1,4,84,69,0,0,0,0,0,0,0
2010-09-09 07:48:03,D4B,68,2,4,45,61,1,5,93,42,0,0,0,0,0,0,0
# disabled gridlines
2010-09-10 07:02:42,D4B,34,2,4,45,61,1,1,75,16,0,0,0,0,0,0,0
2010-09-10 07:24:13,D4B,61,2,4,45,61,1,2,90,25,0,0,0,0,0,0,0
2010-09-10 07:42:24,D4B,60,2,4,45,61,1,3,78,42,0,0,0,0,0,0,0
2010-09-10 11:55:14,D4B,64,2,4,45,61,1,4,63,64,0,0,0,0,0,0,0
2010-09-10 12:19:17,D4B,50,2,4,45,61,1,5,58,43,0,0,0,0,0,0,0
I won't be getting any 95%s with gridlines off any time soon, that's for sure.
--
gwern