Meta-analysis of distinct and shared neural networks in reading, mathematical processing, and working memory in school-age children

Abstract

Reading and mathematical processing are important academic skills for children to learn. Behavioural observations suggest the existence of distinct brain regions for reading and mathematical processing (Kersey et al., 2019), as well as a shared network of general executive processes (Ashkenazi et al., 2013), with working memory being a potential candidate for this shared network due to its relationship with children’s performance in the two academic domains (Alloway & Alloway, 2010; St Clair-Thompson & Gathercole, 2006). This meta-analysis investigated the involvement of working memory in the common regions of reading and mathematical processing, as well as domain-specific processes in the distinct regions of the two academic domains. Activation Likelihood Estimation (ALE) meta-analysis was used on fMRI studies involving the three domains of interest. Results found common regions in the left parietal and right insular areas, corresponding to general executive functions. Additional overlapping activation in the right inferior parietal lobule was explained by the involvement of visuospatial working memory functions in mathematical processing. Distinct reading-related activations in the left inferior frontal and temporal regions were associated with semantic processes, whereas distinct mathematics-related activations in the parietal areas were associated with numerical processes. Findings revealed common executive function regions that may be responsible for the comorbidity in some learning disabilities, as well as distinct regions that may account for the variability in children’s academic abilities.