Thanks for the reference, which includes the neutral point of view, "The paper's methodology has been criticized by David Moody[5]. He draws attention to the fact that different tests were used to evaluate the control and test groups. Due to the way in which the tests were administered, he also questions whether they were valid tests of fluid intelligence."

K. Anders Ericsson has achieved astounding results through training in "digit span," long reputed to correlate well with "general intelligence."

http://en.wikipedia.org/wiki/Working_memory

"Whereas most adults can repeat about seven digits in correct order, some individuals have shown impressive enlargements of their digit span--up to 80 digits." The further discussion at this reference about whether the tested task correlates with anything else of interest would apply as well to the n-back task.

Do you mean the dual n-back task? Jaeggi et al? That one was interesting because it appeared the gains were transferable.