Aim The neurobiological contributions of childhood language disorder are not well understood. Yet there is increasing evidence that language disorder is associated with differences in brain structure and/or function in core language regions. A key hypothesis has been that children with language disorder do not show the same degree of leftward asymmetry of these regions as observed in typically developing children. We aimed to systematically review structural and functional magnetic resonance imaging (fMRI) studies to examine brain commonalities and differences between children with language disorder and typically developing controls; and differences in leftward asymmetry between these groups.

Method: A systematic review was conducted using MeSH terms synonymous with childhood language disorder and brain MRI methods. The search identified 1443 papers, and 18 articles met the criteria and were appraised for level and quality of evidence.

Results: Atypical brain structure and function was reported within traditionally recognized language regions across studies, including the inferior frontal gyrus, posterior superior temporal gyrus, and caudate nucleus. The direction of difference (e.g. increased/decreased) was variable, however, likely because of differences in language disorder groups examined and magnetic resonance data acquisition and analysis approaches. As regards asymmetry, there was some evidence of reduction of the anticipated structural and functional leftward asymmetry in frontal language regions in language disorder groups.

Interpretation: Mounting evidence suggests that children with language disorder have atypical brain structure and function within neural regions integral to language. There is limited support for the hypothesis that children with language disorder show a reduction of leftward structural and/or functional asymmetry in frontal language regions. Interpretation is limited, however, by a high degree of variability in language disorder assessment and phenotype, and in magnetic resonance methodologies. A large‐scale magnetic resonance study of brain structure and function is required in a well‐defined language disorder population cohort, with replication, to provide confirmatory data on the neural correlates of childhood language disorder.