The goal of this observational study is to test a reciprocal relationship between statistical learning and the development of language and literacy in first-graders with autism and their non-autistic peers. The main questions it aims to answer are:

1. whether children's statistical learning abilities can predict their long-term improvement of language and literacy skills in school;
2. how children's brains automatically learn patterns from speech and prints;
3. whether children's learning in the lab reflects the language patterns they have learned over the years from their native language.

First-grade students will participate in the study twice across three months.

During Time 1, children will complete

• a battery of language, reading, and cognitive assessments
• a series of computer-based statistical learning games both inside and outside of functional MRI scanner.

During Time 2, children will complete a battery of language and reading assessments to detect the growth in three months.

Researchers will compare the autistic and the non-autistic groups to see if statistical learning plays a similar or different role in predicting children's language and literacy growth.

In this one-year R56 project, the investigators will design a mini-longitudinal project to test how children's linguistic statistical learning predicts their growth in spoken and written language in three months. The investigators predict that weaker linguistic statistical learning underlies the exacerbation of language and literacy delay in children with ASD. In this project, the investigators will gamify an existing study protocol, which has led to promising discoveries motivating the original R01 proposal. With an enriched platform, the investigators expect a low attrition rate (<10%) and few missing data (>90% task completion rate) in this longitudinal sample of six-year-olds both with and without ASD, matched on age and sex ratio. This project will lead to critical feasibility and preliminary data to support the key hypotheses of all three aims. Due to a shortened timeline, the investigators plan to recruit 25 children per group, sampled from a diverse population, to ensure the team's capacity for a three-month follow-up session before the end of a school year.

Aim 1: Establish the longitudinal relationships between SL and language/literacy development in both TD and ASD. In this aim, the investigators will evaluate the role of SL in both concurrent and the growth of language/literacy skills assessed in three months during the school year. The investigators expect that linguistic SL (embedded pattern learning of syllables and letters) is associated with concurrent language and literacy skills and will prove to be crucial predictors of growth in language and literacy skills in both groups. Given the small sample size in this one-year project, the investigators will build prediction models both across the two groups and within each group. Sex, concurrent language and reading skills, and nonverbal IQ will be included as key covariables.

Aim 2: Determine the longitudinal relationships between neural bases of SL and developing language networks in the brains of children with ASD. In this aim, the investigators will test whether altered brain functions during linguistic SL are related to poorly functioning language networks (language comprehension and phonological working memory), defined within each individual child. The investigators will also examine whether neural bases of linguistic SL predict future language and literacy development in children with ASD. Participants from Aim 1 will complete linguistic SL tasks and two well-validated language tasks in the fMRI scanner at Time 1. The investigators expect to find reduced engagement of language networks during SL in children with ASD and to provide a neurobiological explanation for the cascading effect impaired linguistic SL casts on future language development.

Aim 3: Test whether linguistic SL is a proxy for children's sensitivity to real-world language statistics. The investigators ask whether the specific weakness in linguistic SL in ASD generalizes to real-world language statistics using serial recall tasks, a well-validated instrument for individuals' sensitivity to native-language statistical patterns. The same participants in the previous aims will complete a phonological and an orthographic serial recall task containing both high- and low-frequency English bigrams/trigrams at Time 1 and Time 2. The investigators predict that these tasks will be associated with children's linguistic SL performance in Aim 1 and children's linguistic SL at earlier times will predict the future growth of sensitivity to natural language statistics. These findings will establish the critical link between linguistic SL and children's natural language skills.

Inclusion Criteria:

• current first grader (6;0 - 7;6)
• Geographically located within the Greater Boston Metropolitan Area
• Native English speakers
• Normal hearing
• Normal or corrected vision

Inclusion criteria for children with ASD (N = 25):

• Children with a professional diagnosis of autism according to expert clinical judgment
• Capable of speaking sentences with three or more words
• Social Communication Parent Questionnaire score > 15
• Autism diagnosis confirmed by ADOS

Inclusion criteria for typically developing controls (N = 25):

• Neurotypical: with no known cognitive, neurological, or psychiatric disorders
• Social Communication Parent Questionnaire score < 11
• Receive a score within 1 SD of the population mean for age on all assessments.

Exclusion Criteria:

• non-native speakers of English
• More than 30 hours of exposure to a language other than English per week
• history of brain injuries and head injuries
• intellectual disability, mutism, motor delay, or developmental coordination disorder
• metal in body
• claustrophobic
• history of prior neurosurgical procedure
• substance abuse
• signs of increased intracranial pressure

The demographic and behavioral information will be made available in the form of a comma delimited text file and a code book. We will disguise individuals' identities by redacting birthdates and date of examination from the available records and by replacing names with coded alphanumeric values. When permission is granted, we will deposit all raw, de-identified behavioral and neuroimaging data into the National Institute of Mental Health Data Archive (NDA). The fMRI data will be uploaded to a public repository to improve meta-analyses and data-sharing (NeuroVault, OpenNeuro). To protect the privacy and confidentiality of the participants, raw video or audio files will not be shared outside of the research team. We do not commit to processing additional data beyond our own needs for the purpose of sharing. Any data we have processed we are willing to share. We will submit electronic versions of all accepted manuscripts to the NIH National Library of Medicine PubMed database.

The data will be made available to other users within 12 months of the final data point being collected or at the time of an associated publication, whichever comes first.

All research participants will be consented for broad data sharing. To protect the privacy and confidentiality of the participants, raw video or audio files will not be shared outside of the research team. We do not commit to processing additional data beyond our own needs for the purpose of sharing. Any data we have processed we are willing to share.

To request access of the data, researchers will use the standard processes at NDA, and the NDA Data Access Committee will decide which requests to grant. The standard NDA data access process allows access for one year and is renewable.

The NDA GUID tool allows researchers to aggregate data from the same research participant without different laboratories having to share personally identifiable information about that research participant. The NDA data dictionaries do not permit personally identifiable information to be shared. NDA maintains a Certificate of Confidentiality.