Dual 10-Back and some thoughts about Dual N-Back

[Kevin Ann]

Just started Dual 10-Back today since I felt like I was starting to get the hang of Dual 8-Back and wanted to push myself.  Was chunking 2x4 items (can't really help it), so I figured I'd stress my working memory + focus even more by bumping up the difficulty.

Parameters:

- 2.5 sec/item

- 120 items/set

- 5 minute sets

One thing I wonder is how tasks that are at first fundamentally and purely fluid intelligence and working memory tasks start to shift more to crystallized intelligence and long-term memory. You actually start getting good at the specific parts of Dual N-Back and multiple stimuli get chunked together to form "atoms" themselves. For example, if you have the sequential "T T T T" in each of the four corners of a 3x3 grid, then you are no longer really remembering four separate stimuli, but rather grouping them as a single stimulus. Despite your being mindful of not chunking, it's almost irresistible to chunk in this way.

I am tentatively taking the position that enough practice and experience makes the Dual N-back task less about fluid intelligence and short-term/working memory and increasingly more about crystallized intelligence and long-term memory. Seems like the a way to keep Dual N-Back (more) purely a working memory task is to continuously bump up the difficulty so that chunking doesn't diminish the amount of stress placed on working memory.

I read a very insightful and Zen-like post that encapsulates my outlook (paraphrased): "The point of Dual N-Back is never to master Dual N-Back."

[Brandon Woodson]

You are making good progress. I am interested to see how far you will go and will stay posted. Kudos.

-Brandon

[jotaro]

you put a good point by saying the multi independent stimuli becomes as one 

but i wonder because when i play i dont chunk ,i dont think at all

i just let it be and watch and hope to deal with it when the time comes.

however it might get to a group even if i dont chunk. automatically by my long term memory

that in that case my method will get dropped to the ground dead.

and it will better just chunk and rise difficulty or rise difficulty more rapidly without the chunk as it as now.

 

[Kevin Ann]

I think the name of the game is stressing your working memory. 

Chunking circumvents that process so you have to challenge yourself more to counteract the use of crystallized intelligence and long-term memory.

Here's an incomplete list of auditory and spatial chunks I use. Many ways to increase this, for example, S can be appended to make all chunks plural and some letters can be repeated to make the sequence longer (SHT goes to SHHT).

Auditory

SQRT = square root 

LLC = limited liability company

RCKT = rocket

QRST = alphabetical

SRT = sort

SHT = shit

QRTC = quartic

STR = stars

RTS = real time strategy (starcraft)

CRT = cart

Spatial

Corners

Repeated

Crosses

Alternating

Ladders (up and down)

Rows

Columns

Edges

Diagonals

[.h2o.]

I guess the main question that comes to mind is whether you are seeing any benefits outside of dual n back scores. Some of these techniques seem quite elaborate and very task specific even for the version of dual n back you are using. For example many versions of dual n back use different letters, and I can see your SQRT chunking technique breaking down so that you are back at your "real" working memory ability. I don't know about you, but for me n back games are not the most entertaining and would rather spend the time perfecting some other task specific skill like... Guitar Hero for instance. Instead, I keep coming back to dual n back because I see it as a way to increase working memory - the master skill that helps you master all the other skills.

I see the main task of dual n back ( and n back in general) is learning to think sequentially and store information sequentially. The people who have the best memory for details are sequential thinkers. In the game I try to get my attention as steady as clockwork and try to process the stimuli as it is without further associations. Sometimes when doing n back I cannot help but think about meanings behind acronyms such as PG (movie rating), HP (Hewlett Packard), PHP (coding language) or see a group of squares as a triangle but I would never go as far as to creating forced associations. Of course, I never got as far as you so what do I know...

How about you try a game without your chunking associations and see what your working memory gains are.

[Kevin Ann]

Oh, what you say is very true. The very reason I'm pushing myself up to higher N is because I really can't help chunking in this manner, which is a very task-specific ability.  I'm not going to fool myself that getting higher scores at a given N necessarily mean that my working memory is getting better - not even considering the more controversial issue of transfer and improvement of fluid intelligence.

As far as I'm concerned, the best way to know I'm stressing my working memory is to feel I'm struggling. 8-Back was starting to feel easy with chunking, so I'm jumping up to 10-back since that still feels hard even if I do use chunking. Perhaps as I get to even higher N, not only will this be a game of short-/long-term memory and fluid/crystallized intelligence, but also one of creativity to come up with chunks and weave a story together on the fly. Like, say if it became N = 20 or N = 30, I really think you need to be creative in thinking of how to store and access that information.

As for thinking linearly, remembering, and forgetting... I try to remember 10 at a time, answer while I remember the next 10, and so on. This worked with other N too. At first, you go totally blank, but with practice and rest, memory fades less and less until it becomes very clear and you can traverse your memory more and more.

I'm not sure how to test definitively how my working memory and fluid intelligence are getting better, since the following are very difficult to answer and may even be ill-defined:
- What other confounding factors are there?  Example: motivation, re-test biases, statistical fluctuation, etc.

- How well can a test be devised to track such amorphous ideas as "working memory" (what is the time-scale of working memory vs. short-term memory vs. long-term memory)?

- How well-defined are concepts like: working memory, intelligence, fluidity, crystallization, task specificity, different types of memory, etc.?

At the end of the day, I enjoy doing Dual N-Backing because it gets me to relax and into Flow, and that is an excellent ends/benefit in itself. It's awesome that there are other benefits too with varying degrees of scientific evidence, such as improvements in working memory, focus, fluid intelligence, etc.

At the end of the day, I agree mostly with what Payman Sagafhi has written here, especially his thoughts about how crystallized intelligence is very important too.

Anyways, interesting subject.  I can't wait to get my neural prostheses and hope I live long enough for the advent of technology that allows for a radical expansion of my biological brain through technological means in order to experience vast new vistas of experience, consciousness, and cognitive ability.

I'm having trouble posting due to a bad internet connection, so I will end it here...