Posted: June 8th, 2012 – Developmental Psychology has recently published a paper called “Is Working Memory Training Effective?” (May 2012). It is a meta-analysis, calculating results based on data from already published studies, not on new data.
The authors essentially conclude that working memory training is not effective, in sweeping terms when looking at various interventions. Since this conclusion has received some attention in the media, we want to point out how the conclusions in the paper relate to Cogmed training and how they do not.
The analysis in the paper includes a very wide range of populations (different kinds of people who do the training): young children to adults; some with various serious cognitive problems, and some cognitively in fine shape.
The paper also includes several very different interventions: n-back training, memory strategy training, playing various working memory games, Cogmed training, and several different training tools built only for research. All are grouped together as “working memory training”.
Unfortunately, this makes it impossible to draw any useful conclusions about Cogmed training from the analysis. The headlines may suggest that the article is about Cogmed training, it is not.
The meta-analysis includes data from 30 different studies. Eight of them are referred to as “CogMed”, and only four of those studies refer to relevant populations (individuals constrained by their working memory capacity). In other words, only four of the 30 studies in the entire meta-analysis are relevant for how Cogmed is used in real life, in schools and in health care practices around the world.
These four studies show strong results, just as the rest of the 20+ set of published research papers on Cogmed training do. But that cannot offset the results from the other studies that have used other kinds of working memory training, on other groups and in other ways. This is why the authors sweepingly conclude that working memory training does not work: that when all programs, protocols and populations are looked at together, the results are not strong. This is a somewhat obvious finding, and again, does not say much about how effective Cogmed training is for those people it is designed for.
When you review the research case for evidence on how the Cogmed training specifically helps individuals with working memory problems (affecting 10-20% of all children/adults), a very different conclusion is reached showing strong benefits for this group.
It is essential to distinguish the somewhat obvious finding from the Hulme paper, that any kind of working memory training will not significantly help any kind of person, from the more appropriate question that has been and continues to be addressed in research. That is, which forms of cognitive training have what effects, on which clinical populations, at what age levels, depending on what implementation conditions and with what duration of benefits?
Working memory training deserves thoughtful review. It also deserves careful and discerning analysis. Unfortunately, the Developmental Psychology paper is disappointing in this regard.
It is good to see how the field of cognitive training in general, and working memory training in particular, now merits this degree of scrutiny. Our field is growing and becoming more than just a promising possibility; it is natural and necessary that it be challenged and analyzed by skeptical researchers.
Hopefully, the article will ignite conversations on working memory training and its effectiveness as shown in research. This is a discussion that we warmly welcome; it is how we show how Cogmed training stands out in a great way.
In conclusion, while other working memory training programs may have failed to generate solid research data, published research shows that the Cogmed program is effective, when used correctly and with the right populations.
Feel free to write us with your comments.