Does Working Memory Training Lead to Better Student Learning Two Years Later? The Memory Maestros Trial

Working memory is an important cognitive function that enables children to learn foundational skills such as maths, reading and language as well maintain attention in the classroom. In the past decade, there has been an increase in the number of adaptive computer-based programs, such as Cogmed, which claim to improve working memory and thus benefit children’s learning and classroom attention. Not surprisingly, these programs have been implemented in an increasing number of schools in Australia and internationally as a way to support children with low working memory. However, there remains a substantial gap in the education research evidence as to whether such programs are more beneficial than usual classroom teaching.This paper discusses the findings from the Memory Maestros randomized controlled trial. This trial aimed to determine whether providing a computerized working memory training program, Cogmed, to children with low working memory in Grade 1 (aged 6-7 years), compared with receiving usual classroom teaching, leads to better academic achievement 1 and 2 years later. Forty-four Melbourne schools participated, where 1723 Grade 1 children underwent working memory assessments using the Automated Working Memory Assessment to identify those with low working memory. Overall, 452 children were identified as having low working memory, with 226 randomised to receive the computerized program and 226 received usual classroom teaching. All children with low working memory were followed-up at 6 months, 1 and 2 years later, with 88% of children remaining in the study 2 years later. At each time point, children repeated the Automated Working Memory Assessment and academic achievement (primary outcome) was measured by the Wide Range Achievement Test-4 at 1 and 2 years. Parents reported on the child’s behavior and health-related quality of life, whilst teachers reported on the child’s classroom behavior, approach to learning, and ability in maths and literacy. Overall, compared to the children who received usual classroom teaching, intervention children had better visual-spatial short term memory at 6 months. However, these benefits attenuated by 12 months and were no longer present at 24 months. Importantly, our findings also show that there was no benefit in other outcomes; in fact, intervention children actually performed worse in maths at 2 years. Although implementing such a program is highly feasible for 6-7 year olds, given the loss of classroom time, cost and lack of lasting benefit, our findings suggest we cannot recommend population-based delivery of Cogmed within a screening paradigm.