Review of the Evidence on, and Fundamental Questions About, Efforts to Improve Executive Functions, Including Working Memory

403 articles screened, 179 studies (193 papers) included, all ages, published in English peer-reviewed journals through 2015

All ages; every method claiming to improve executive functions.

A systematic review with eleven explicit inclusion criteria, including a requirement for an objective behavioural EF measure, a control group, at least eight participants per condition, and outcomes beyond the trained task. It separates active-control from no-treatment-control studies explicitly, describing the latter as a lower bar. Its importance to this archive is that it is the field's leading advocate conceding the transfer limit in her own systematic review.

Two stated principles carry the review. First, executive functions can be improved across the lifespan and by different methods. Second - and this is the concession - transfer of training is narrow, benefits are fairly specific rather than general, and they rarely transfer to untrained skills; the authors describe this as the nearly universal conclusion of reviews for the past 25 years. On Cogmed, 79% of the 15 studies examining NEAR transfer found at least suggestive evidence, but only 33% of the 12 studies examining FAR transfer did. For n-back the figures are 40% of 10 near-transfer studies and 30% of 11 far-transfer studies. Their summary is that there is only extremely spotty evidence for far transfer to untrained cognitive skills for any training method, and that the rare claims of far transfer have not held up to scrutiny. They also note that no study has tested transfer from working-memory training to reasoning while controlling for the working-memory components of the reasoning test.

Not applicable as a review.

Not applicable. Notable as the authors' own partial retraction of their 2016 optimism.