Recently the journal Developmental Psychology published a meta-analysis of research on the value of Working Memory Training. A meta-analysis is not original research but combines multiple studies and uses statistical adjustments to try to draw an overall conclusion from disparate studies. In this case, the authors concluded that there is little to no benefit of Working Memory Training.
The meta-analysis is interesting for what it includes and for what it doesn’t address. There are two particular aspects of the studies included in the meta-analysis that are important to understand.
First, the average number of training sessions in the studies was 12. Most therapists (vision, speech and language, other educational specialists) will recognize that patients or clients may only just be beginning to show some changes within 12 sessions. As the meta-analysis points out, programs that claim to change your memory (or other specific abilities) in “10 minutes a day” or “in 12 sessions” are unlikely to create lasting change that will transfer beyond the specific skill you practice in the training. And, after all, we don’t want to just get better at remembering a bunch of numbers; we want to get better at remembering all kinds of other things.
Second, the studies included in the meta-analysis had to specifically (and narrowly) address working memory in order to meet the criteria for inclusion. In the programs used in the studies, players work through a few games that target the ability to hold more information, with the challenges becoming more and more difficult. The games used in these studies had no purpose but exercising a specific aspect of working memory. While working memory is part of our capacity to learn and is needed in many everyday mental processes, it does not work in isolation from other functions like visual span, visual selective attention, auditory selective attention, sequential memory or a whole host of other processes. Since working memory is the process of holding multiple other functions active while processing takes place, the integration with other functions is essential. The programs in the studies included in the meta-analysis are too narrow to show transfer beyond short-term gains and into broader application because they develop only part of the mental processes needed.
As for what the meta-analysis did not include, we have to look at studies that didn’t meet the criteria for inclusion. For example, the review did not include any research on BrainWare Safari. BrainWare Safari develops multiple cognitive skills, including working memory, in an integrated and holistic manner. The recommended protocol for BrainWare Safari is 3 to 5 sessions per week for 10 to 12 weeks, yielding 30 to 60 sessions (far more than the 12 sessions averaged for the studies in the meta-analysis). Published studies with BrainWare have shown both cognitive and academic improvement when used according to the recommended protocol. Retention of the improvement and even further growth after completing the time in BrainWare has been shown as well.
The conclusion from this latest article should not be that working memory training doesn’t work, but that working memory training must be done right to create lasting and transferable change.