The diagnosis of autism spectrum disorders (ASD) relies on three core features: impairments in language, impairments in social skills, and the presence of stereotyped or restricted interests. Of these three features, an individual's language ability distinguishes itself as one of the key factors predicting their long-term outcome. However, there exists a large range of language outcomes, with estimates of up to 50% of children with ASD having little to no language. Standardized clinical examinations have been successful at characterizing these children, but they are unable to explain the development of, or the underlying biological causes of, language dysfunction. The application of a more brain-based indicator, also known as a biomarker, holds great promise to characterize the various types of language function and to educate scientists on the different pathways through which language progresses. Electroencephalography (EEG) represents a promising biomarker, as it records the electrical activity within the brain and provides exquisite information about brain processing in real-time both at rest and while a child is engaged in a task. This study will couple behavioral testing with EEG to identify biomarkers of language function in children with ASD. The researchers will apply two measures that should provide insight into language ability. The first measurement, called event related potentials, estimates the brain's electrical activity while an individual learns an artificial language through pattern recognition. The second analysis measures patterns of the brain's activity while a child is at rest (watching bubbles). Each measure has been shown to predict language function in typical developing children, but neither has been studied in children with ASD. These biomarkers of language function will facilitate our characterization of subtypes within the autism spectrum and, ultimately, will allow for the design and implementation of targeted therapies that are specific to these subtypes.