Developmental language disorder (DLD) is a commonly observed developmental disorder but its neurobiological basis is still unknown. Problems are often persistent, especially in children with problems in speech perception. Despite its prevalence and the handicap it causes, DLD has been given very little attention in brain imaging. In this work, we investigate children's brain activation patterns during language processing. We will measure magnetoencephalographic (MEG) responses in patients and healthy controls. We utilize a novel machine learning model, developed in our research group, to investigate whether DLD children show impaired tracking of speech in their cortical activation. We will also investigate longitudinally how the processing and learning of speech sound patterns change during development in DLD and control children. The aim is to discover the fundamental underlying deficits to support the diagnosis and rehabilitation of this common developmental disorder.