Dual-n-back experiments

I have been doing some dual-n-back on the phone - got a good app named just "N-back" whose logo is a little orange goldfish.

At first I did single-n-back for numbers and got up to about 80% of 4. The strategy I used was after each new number comes, you repeat the 4 outstanding ones, but with some inflection - the first one you just sort of put to the side, and prepare the second one to become the new start of a sequence. So it's like a,BCD which converts to BCDE after you see E and compare it to a. It works pretty well, but it seems like cheating.

I also found myself using finger gestures to point to the sequence, or repeating the sequence to myself to perhaps store it in my audio memory.

All of that seems like cheating - because sometimes, in moments of stress while playing, you can just immediately jump to what you remembered the sequence was at the right point, without using any tricks. That ability seems like the real point of doing this type of training.

Anyway after getting to level 4 in single n-back for both shape (number) and position, I combined them and did dual 2-back, which is pretty easy to pass. Then on to 3, which took a day or to. And then i really started to scale the mountain with 4. That took a few weeks of about 5-8 min per day (usually 3 plays of 30) to get. I have just now started level dual-5-back, which is killing me. I can hold it together for a while, but am lost a lot.

One thing you are forced to do here is patch up a broken sequence - so say you remember ABCDE but then get lost. It's tempting to just quit there, but if you can just remember DE you can greatly reduce your possible errors.

Another challenge is if you can remember the position but not the shape, it's worth it to just remember the position. Remembering either factor separately is tough, though.

Anyway I want to get to level 6 or 7 of dual-n-back before quitting.

What is this actually doing? Not sure. The first few days I dreamed like crazy after doing it, but that's not happening anymore. I am still making progress though - it might be due to real new memory registers being made, or maybe just to improvements in concentration. That is a big part of it too - not freaking out, and being able to slice the threads of memory together well even when some are lost. So overall I think it is a good exercise, and the claims supporting it may actually have something to them. I first heard about it from this site: http://www.bulletproofexec.com/how-to-add-2-75-iq-points-per-hour-of-training/