Assessing far transfer
I am wondering if there should be transfer of performance from picture n back to dual n back? I have been improving nicely on picture n back with 100% performance on 3 and 4 back and around 70-80 on 5 back after about a week of practice. It is difficult for me to practice much on dual n back since I need to use headphones and can't always get on a computer where I can use them.
? | 3 years ago
Reply
My guess is yes, to a small extent. Dual tasks have more complex executive processing requirements than normal tasks. Assuming that a regular n-back task stresses your attentional control, and you master that, then it seems logical that similar attentional control skills can be applied to the dual task.
But since you didn't practice concurrently managing another source of input, that task would have to be practiced anew.
Here's an analogy. If you are trying to learn a piano piece, you can master the left hand first. Then the right hand. Will this help you on playing with both hands? Yes (it beats being completely clueless), but you'll have to practice both-hand play as a new task. Now, add pedals to that, and vocals!
But since you didn't practice concurrently managing another source of input, that task would have to be practiced anew.
Here's an analogy. If you are trying to learn a piano piece, you can master the left hand first. Then the right hand. Will this help you on playing with both hands? Yes (it beats being completely clueless), but you'll have to practice both-hand play as a new task. Now, add pedals to that, and vocals!
cognitivefun | 3 years ago
Reply
Assessing far transfer
Will play dual n-back til September 5th, whereupon I will try these tasks again (leaving them be until then).
Don't quite know what would constitute statistically significant improvements, but hope to feel, for instance, that 15 digits on the aural forward span is a piece of cake.
Today's efforts:
Aural forward digit span: 14 digits 77.38% 7.74s
Aural reverse digit span 11 digits 93.94% 8.53s
PASAT 4-back 95.00% time: 2,482 ms
PASAT 5-back 50.00% time: 4,386 ms
Picture n-back 5-back 73.33% 598.16 ms
Spatial Updating 3 boxes, unsimultaneous 76.67% 1.28 s
Fare thee well, beloved Speed PASAT - til we meet again on the 5th :(
Don't quite know what would constitute statistically significant improvements, but hope to feel, for instance, that 15 digits on the aural forward span is a piece of cake.
Today's efforts:
Aural forward digit span: 14 digits 77.38% 7.74s
Aural reverse digit span 11 digits 93.94% 8.53s
PASAT 4-back 95.00% time: 2,482 ms
PASAT 5-back 50.00% time: 4,386 ms
Picture n-back 5-back 73.33% 598.16 ms
Spatial Updating 3 boxes, unsimultaneous 76.67% 1.28 s
Fare thee well, beloved Speed PASAT - til we meet again on the 5th :(
cevapcici | 3 years ago
Reply
The n-back and digit span tasks seem to measure different, partially independent "components" of IQ. Also, simple (forward) span is a bit different than complex (backward) span - the complex one involves much larger part of prefrontal cortex, for example. I don't have the abstracts or book excerpts at hand, but might try to find them if you want (or maybe they're already on resources page, I can't remember where did I read about it).
14 digits on auditory digit span is damn good in my opinion - do you use any mnemonic system (I think you said you did some time ago, but I don't really remember)?
PASAT 4 with 95% accuracy is *insanely* good, IMHO.
Interestingly, my "best ever" backward digit span is 12 (very unreliable, too), while forward is only 10. Weird.
14 digits on auditory digit span is damn good in my opinion - do you use any mnemonic system (I think you said you did some time ago, but I don't really remember)?
PASAT 4 with 95% accuracy is *insanely* good, IMHO.
Interestingly, my "best ever" backward digit span is 12 (very unreliable, too), while forward is only 10. Weird.
bovinebrain | 3 years ago
Reply
How fast are you improving with the intuitive approach as compared to the various rehearsal systems from before? When you started with this new approach did you start at the same n where you were before or did you have to start again from the beginning?
Whoopska | 3 years ago
Reply
I've been up to 9-back here with my former approach, tho' could only just manage 50%. I was okay with 6-back this morning, playing intuitively, but couldn't get more than 50% consistently on 7-back. I've continued from where I left off, task-wise.
I'm finding it a little odd - I have to tell myself to shut up if I begin to verbalise anything and I also encourage myself to lose the thread ie. lose any conscious awareness of patterns that I have noticed and just try to respond to the individual stimulus in front of me.
I've given up with doing the strategies in a parallel effort, as I cannae be bothered.
Give it a couple of weeks and I should be able to report any improvements.
I'm finding it a little odd - I have to tell myself to shut up if I begin to verbalise anything and I also encourage myself to lose the thread ie. lose any conscious awareness of patterns that I have noticed and just try to respond to the individual stimulus in front of me.
I've given up with doing the strategies in a parallel effort, as I cannae be bothered.
Give it a couple of weeks and I should be able to report any improvements.
cevapcici | 3 years ago
Reply
"I have to tell myself to shut up if I begin to verbalise anything and I also encourage myself to lose the thread ie. lose any conscious awareness of patterns that I have noticed and just try to respond to the individual stimulus in front of me."
You won't have to do this when there is a test that is resistant to chunking :)
cognitivefun | 3 years ago
Reply
"14 digits on auditory digit span is damn good in my opinion - do you use any mnemonic system (I think you said you did some time ago, but I don't really remember)?"
I was using a digit/consonant substitution system for the visual digit spans but have since forgotten the correspondences. 12 digits forward on the auditory digit span *feels* pretty easy - I can afford not to do any rehearsal and recall the numbers no probs. In any case, rehearsal seems to be a double-edged sword: when done *during* an ongoing task, it prevents one from attending fully to new information.
The only technique I use is to group the digits in 3s - I have tried grouping in 4s and found it slightly less effective.
Your backward/forward span difference is interesting: I guess it's because of the anticipation of the reversal you will have to perform.
The only technique I use is to group the digits in 3s - I have tried grouping in 4s and found it slightly less effective.
Your backward/forward span difference is interesting: I guess it's because of the anticipation of the reversal you will have to perform.
cevapcici | 3 years ago
Reply
You might have some difficulty determining what gains are from transfer and what gains are from practice of the specific task. As an example of the clouding that I'm talking about, I remember reading how when a control group and an experimental group both take an initial measure (IQ for example) and then after a period of training or waiting take a second measure, the control groups performance increases because they have become more familiar with the task.
Since you are pretty high up in the I-scores list, this effect may not really apply to you, but what I would look out for is a similar level of increase in the simpler tasks (e.g. digit span) and a greater increase in harder tasks (e.g. reverse digit span).
A similar explanation was used in "The Bell Curve" to contravene the "lack of motivation" explanation for the difference in Black/White test performance. "...because the backward form [RDS] is twice as g-loaded as the forward form [DS], it is a much better measure of general intelligence" (The Bell Curve 283).
If visible, this doubled effect would be some nice confirmation of the Jaeggi conclusion that the dual n-back increases 'Gf' (from some perspective different and from others the same as 'g').
Since you are pretty high up in the I-scores list, this effect may not really apply to you, but what I would look out for is a similar level of increase in the simpler tasks (e.g. digit span) and a greater increase in harder tasks (e.g. reverse digit span).
A similar explanation was used in "The Bell Curve" to contravene the "lack of motivation" explanation for the difference in Black/White test performance. "...because the backward form [RDS] is twice as g-loaded as the forward form [DS], it is a much better measure of general intelligence" (The Bell Curve 283).
If visible, this doubled effect would be some nice confirmation of the Jaeggi conclusion that the dual n-back increases 'Gf' (from some perspective different and from others the same as 'g').
Whoopska | 3 years ago
Reply
"what I would look out for is a similar level of increase in the simpler tasks (e.g. digit span) and a greater increase in harder tasks (e.g. reverse digit span).
"
Ah, yes - this is interesting.
cevapcici | 3 years ago
Reply
I am also, under my brother (the king of subitizing)'s id, doing the dual n-back whilst attempting to use NO strategies - merely 'feeling' whether there is a match or not.
The strategies that I have come up with so far:
- vocalised or sub-vocalised rehearsal of auditory stimuli
- chunking of stimuli into groups
- grouping visual stimuli as 3D shapes (triangles, usually)
- sequence-spotting
- using the trial number count as a prompt
are thus DISCARDED.
Early reports suggest that attempting to do the task intuitively is actually more fun - at least, I don't get that horrible feeling of loss of control when my rehearsals break down and I have to make 'shots in the dark' - everything is a shot in the dark! ;-)
25 mins on each id; will compare and report outcomes come September.
The strategies that I have come up with so far:
- vocalised or sub-vocalised rehearsal of auditory stimuli
- chunking of stimuli into groups
- grouping visual stimuli as 3D shapes (triangles, usually)
- sequence-spotting
- using the trial number count as a prompt
are thus DISCARDED.
Early reports suggest that attempting to do the task intuitively is actually more fun - at least, I don't get that horrible feeling of loss of control when my rehearsals break down and I have to make 'shots in the dark' - everything is a shot in the dark! ;-)
25 mins on each id; will compare and report outcomes come September.
cevapcici | 3 years ago
Reply
Update: actually trying to do the task intuitively is very dull - it's a kind of estranged-feeling concentrating-without-thinking.
cevapcici | 3 years ago
Reply
"concentrating-without-thinking"
This is an intriguing observation. Have you ever played a game of chess/go without deliberately thinking about strategy? If you suppress these thoughts, perhaps you would be playing "intuitively," but expert performance at these tasks requires ruthless reading, pruning, and chunking of memory. Spread activation could be passive, but executive processes do seem to require sustained attention and directed action.
This also reminds of the two kinds of television watching: passive watching and "mindful watching." The latter, if memory serves, actually lead to learning effects, but not so the former.
My hunch regarding "intuitive" is that pattern matching and fast memory management become intuitive: more of an executive, less perceptual. If this is correct, then assuming that two people begin with different strategies: one doing it the "intuitive way," and one just being deliberate about it, if they both reach expert performance, one would actually expect comparable cortical processes.
In any case, this is just one opinion; everybody is bound to be different. I suppose a dual n-back task strategy, or lack thereof, would also be part of one's learning style.
This also reminds of the two kinds of television watching: passive watching and "mindful watching." The latter, if memory serves, actually lead to learning effects, but not so the former.
My hunch regarding "intuitive" is that pattern matching and fast memory management become intuitive: more of an executive, less perceptual. If this is correct, then assuming that two people begin with different strategies: one doing it the "intuitive way," and one just being deliberate about it, if they both reach expert performance, one would actually expect comparable cortical processes.
In any case, this is just one opinion; everybody is bound to be different. I suppose a dual n-back task strategy, or lack thereof, would also be part of one's learning style.
cognitivefun | 3 years ago
Reply
I always play chess/go without thinking about strategy - that's why I'm so bad at them! ;-)
Strategy in those cases is used to reduce complexity and demands on the brain - my worry is that, when it comes to dual n-back - maintaining these demands is precisely what one wants.
If I was to make a board games analogy, then perhaps I would use Scrabble training. Many players practise anagramming sets up to a maximum of only 7 letters. They then use mnemonics to help them unscramble longer words, for instance they would learn BESTIAL+K = BALKIEST rather than testing themselves on ABEIKLST. By using these short-cuts, they are perhaps limiting the extent to which their anagramming could be improved. They reach a point at which intuitive thinking is seen as more demanding than conscious strategising and so plump for the latter option.
It might be the case that, with a task like dual n-back, once you reach a certain point, memory aids pay off more readily than improvement of the basic skill - especially as, with a greater quantity of information, there is a greater opportunity for pattern spotting etc, as there is a larger number of connections that could be drawn between stimuli.
Strategy in those cases is used to reduce complexity and demands on the brain - my worry is that, when it comes to dual n-back - maintaining these demands is precisely what one wants.
If I was to make a board games analogy, then perhaps I would use Scrabble training. Many players practise anagramming sets up to a maximum of only 7 letters. They then use mnemonics to help them unscramble longer words, for instance they would learn BESTIAL+K = BALKIEST rather than testing themselves on ABEIKLST. By using these short-cuts, they are perhaps limiting the extent to which their anagramming could be improved. They reach a point at which intuitive thinking is seen as more demanding than conscious strategising and so plump for the latter option.
It might be the case that, with a task like dual n-back, once you reach a certain point, memory aids pay off more readily than improvement of the basic skill - especially as, with a greater quantity of information, there is a greater opportunity for pattern spotting etc, as there is a larger number of connections that could be drawn between stimuli.
cevapcici | 3 years ago
Reply
"I see only one move ahead, but it is always the correct one. – Jose R. Capablanca"
(actual quotation is a little suspect as you can read in http://burncastleburn.blogspot.com/2007/12/how-deep-do-your-look.html comments)
(actual quotation is a little suspect as you can read in http://burncastleburn.blogspot.com/2007/12/how-deep-do-your-look.html comments)
Whoopska | 3 years ago
Reply
Am only doing the task intuitively now - will see if it has any effect when I retake the tests mentioned herein.
cevapcici | 3 years ago
Reply
More stats for later comparison:
Visual reverse digit span: 10, 86.95%
Visual forwards digit span: 10, 93.33%; 11, 87.45%
No mnemonics used
Visual reverse digit span: 10, 86.95%
Visual forwards digit span: 10, 93.33%; 11, 87.45%
No mnemonics used
cevapcici | 3 years ago
Reply
