The perfect N-back training

Blog on Brain Training

The perfect N-back training

2015-03-22 13:03:27

There are some strong evidence that N-back training improve fluid intelligence (Au et al., 2014). However not every user experiences such positive effect. In some cases N-back training doesn't work. Why?

The answer is probably the way we train. Some guys do their N-backing sessions hard and intense and they improve intelligence. But others try to make it easier and that is not good.

You need to be highly motivated to do daily, short, intense sessions and it could be achieved, when you compete with other users on Brain Scale, and when your sessions are about 20 minutes long. It is sufficient on the beginning but it is not enough later.

One of the first research on modern N-back task revealed, that participants with more training, improve their fluid intelligence more. As Susanne Jaeggi and her colleagues observed (2008), the more you train, the more you gain. Don't expect much, if you train only one or two weeks. You need to train for prolonged time.

But there is the problem. After long training your brain gains experience and develops strategies, which make training easier and less efficient. The most common strategy is chunking, which allow you to reach the high level of N-back. Our working memory has capacity of three or four items, and reaching level eighth or higher of N-back is only possible with chunking. Chunking is not so bad, but you should avoid it, especially chunking with mnemonics. Don't worry if you chunk as long, as your session is intense.

With chunking came some other strategies, which you should rather avoid:

  • Playing the odds - when you guessing square locations or letters. Sometimes it allows you to promote to higher level, but it also make training less intense.
  • Jumping attention - when you intentionally omit some hard strings for easier promotion. Especially hard strings are those with interference, the overlapping of letters or squares locations. The string could look like: TQRTQRATQR... and it is hard to follow it, when you are for example on 4-back level because you need to suppress that irrelevant information, which interference gives. However a few researches suggest, that interference control is the most important part of working memory training (Burgess et al., 2014). So if you experience interference don't jump your attention, but try to focus yourself. It is probably the moment, when your fluid intelligence is improving.
  • Rehearsal - when you sub-vocalizing or repeating the strings of letters etc. It is highly hypothetical, but this strategy seems to interrupt the interference control process, making the training easy and less effective (Juvina & Taatgen, 2007).


Au, Jacky, et al. "Improving fluid intelligence with training on working memory: a meta-analysis." Psychonomic bulletin & review (2014): 1-12.

Burgess, Gregory C., et al. "Neural mechanisms of interference control underlie the relationship between fluid intelligence and working memory span." Journal of Experimental Psychology: General 140.4 (2011): 674.

Jaeggi, Susanne M., et al. "Improving fluid intelligence with training on working memory." Proceedings of the National Academy of Sciences 105.19 (2008): 6829-6833.

Juvina, Ion, and Niels A. Taatgen. "Modeling control strategies in the n-back task." Proceedings of the 8th international conference on cognitive modeling. 2007.

Leave a message
To leave a comment please log in or register.


30 Dec 2015 at 12:04

I am confused by the exact way you are referring to sub-vocalization. Are you saying not to sub-vocalize (or repeat with one's inner voice) each individual letter immediately after it is delivered to us or are you saying not to "sub-vocally" review the entire chain of letters as it then exists in between each new square/letter delivery? I believe it was on an article I found on the Brain Workshop site a few years ago that said to focus entirely on each individual square/letter combination immediately after its delivery (and not to consciously keep a running tally). Under that advice, sub-vocalizing each individual square letter combo in isolation was fine; it was merely the sub-vocalizing of the tally that was discouraged. Are you agreeing with the method I just described or are you saying no sub-vocalizing at all, whatsoever (even if it is just each individual combo in isolation)?

01 Jan 2016 at 21:02

You have to know, that there is almost none evidence reffering for particular techniques. So I can't say for sure, if method You've described is good or not. Try to use generic advises for training and try which one seems to be best for You. I don't know other way.